NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Sobieraj DM, White CM, Kluger J, et al. Adjunctive Devices for Patients With Acute Coronary Syndrome Undergoing Percutaneous Coronary Intervention [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2011 Dec. (Comparative Effectiveness Reviews, No. 42.)

Cover of Adjunctive Devices for Patients With Acute Coronary Syndrome Undergoing Percutaneous Coronary Intervention

Adjunctive Devices for Patients With Acute Coronary Syndrome Undergoing Percutaneous Coronary Intervention [Internet].

Show details

Results

Results of Literature Search

Upon conducting the original literature search to identify articles that evaluated the impact of thrombectomy or embolic protection devices on final health or intermediate outcomes, we retrieved 1056 unique citations. After duplicates were removed, 978 articles remained. During title and abstract review, 571 articles were excluded and during full text review 244 articles were excluded. Upon updating the literature search in March 2011, a total of 121 citations were retrieved. Of those, 10 citations were duplicates leaving 111 unique citations. Eighty-two and 19 citations were excluded at the abstract and full text level, respectively, leaving 10 citations which were added to the original literature base. All citations excluded at the full text level are listed in Appendix D along with the reason for exclusion. A total of 175 articles were found to match our inclusion criteria. A summary of search results is presented in Figure 2.

Figure 2 shows the flow of study identification and selection. The initial database search resulted in 697 records from Medline, 88 records from Central, 5 records from Cochrane Systematic Reviews and 227 records from Web of Science. An additional 41 records were identified from other sources. 121 records were identified through an updated literature search performed in March, 2011 using the same search criteria. After duplicates were removed, there were 1091 unique citations eligible for title and abstract screening. The first phase of screening excluded 653 records for the following reasons: 128 records were narrative reviews, editorials or letters to editors; 3 records were reports conducted outside of humans; 162 records were reports not evaluating an adjunctive device in PCI; 5 records were reports that were outside of ACS; 270 records were uncontrolled studies; 46 records were studies conducted outside of native vessels; 39 records were not an RCT or observational studies with >500 patients. This process left 438 records to be assessed for eligibility by screening the full-text articles. The second phase of screening excluded 263 records for the following reasons: 52 records were narrative reviews, editorials or letters to editors; 1 record was a report conducted outside of humans; 21 records were reports not evaluating an adjunctive device in PCI; 3 records were reports that were outside of ACS; 100 records were uncontrolled studies; 3 records were studies conducted outside of native vessels; 72 records were not an RCT or observational studies with >500 patients; 7 records did not report outcomes of interest; 4 records had incomplete data with no response from authors. The full text screening left 175 records to be included in the report. Of the 175 records included, 127 records were included in quantitative synthesis and 48 in qualitative synthesis. Of the 127 records included for quantitative synthesis, 43 records were unique trials, 10 records were sub-studies, 35 records were supplemental data to original trials, 39 records were data duplicating original trials. Of the 48 studies included in qualitative synthesis, 19 records were unique studies, 18 records were systematic reviews and meta-analyses, 1 record was pooled analysis, 7 records were duplicates of study data, 3 records were duplicate systematic reviews and meta-analyses data. Of the 43 unique trials included in quantitative synthesis, 37 studies were in STEMI population including 17 studies evaluating catheter aspiration devices; 5 studies evaluating mechanical thrombectomy devices; 9 studies evaluating distal balloon embolic protection devices; 5 studies evaluating distal filter embolic protection devices; 1 study evaluating proximal balloon embolic protection devices and 6 studies were in mixed ACS population including 1 study evaluating catheter aspiration devices; 3 studies evaluating distal balloon embolic protection devices; 2 studies evaluating distal filter embolic protection devices. Of the 37 unique studies included in qualitative synthesis, 3 studies were in mixed ACS population, 2 studies were in NSTEMI/UA population, 3 studies were in STEMI population with unique comparisons, 2 studies were in STEMI population with direct comparisons, 9 studies were observational studies, 18 studies were reviews and meta-analyses and 1 study was a pooled analysis.

Figure 2

PRISMA flow diagram for the search for KQs 1-3. Abbreviations: ACS=acute coronary syndrome; n=number; NSTEMI= nonST segment elevation myocardial infarction; PCI=percutaneous coronary intervention; PRISMA=preferred reporting items for systematic reviews (more...)

Key Question 1. In patients with ACS who are undergoing PCI of native vessels, what are the comparative effects of adjunctive devices from different classes (e.g., catheter aspiration, mechanical thrombectomy, distal balloon embolic protection, distal filter embolic protection, proximal balloon embolic protection, embolic protection devices combined) on intermediate outcomes (e.g., ST-segment resolution, MBG-3, TIMI-3 blood flow, ejection fraction, distal embolization and no reflow) and terminal outcomes (mortality, myocardial infarction, stroke, target revascularization, MACE, and health-related quality-of-life)?

Key Points

Fifty RCTs and 7 controlled observational studies were included.

Direct Comparative Trials Assessing Final Health Outcomes in ACS

  • Two direct comparative randomized trials were available that assessed final health outcomes in ACS.
    • One direct comparative randomized trial compared the use of catheter aspiration devices to distal balloon embolic protection devices in patients undergoing STEMI. In this controlled trial, no significant differences in mortality, myocardial infarction, stroke, target revascularization, or MACE were found at the longest duration of followup.
    • One direct comparative randomized trial compared the use of one catheter aspiration device to another catheter aspiration device in patients with STEMI. In this controlled trial, no significant differences in myocardial infarction, target revascularization, or MACE were found at the longest duration of followup with the other final health outcomes not being evaluated.

Direct Comparative Trials Assessing Intermediate Health Outcomes in ACS

  • Three direct comparative randomized trials were available that assessed intermediate health outcomes in ACS.
    • Two direct comparative randomized trials compared the use of catheter aspiration devices to distal balloon embolic protection devices in patients undergoing STEMI. In these RCTs, no significant differences were found between groups for ST-segment resolution (one trial), ejection fraction (two trials), MBG-3 (one trial), TIMI-3 blood flow (one trial), or no reflow (one trial) with insufficient data for other intermediate endpoints.
    • One direct comparative randomized trial compared the use of one catheter aspiration device to another catheter aspiration device in patients with STEMI. In this controlled trial, no significant differences in ST segment resolution, MBG-3, or TIMI- 3 blood flow occurred with insufficient data for other intermediate endpoints.

RCTs / Controlled Observational Studies in Patients with STEMI Assessing Final Health Outcomes

  • Thirty-five RCTs and five controlled observational studies evaluated patients with STEMI undergoing PCI and compared a thrombectomy or embolic protection device versus control using the maximal duration of followup. Five final health outcomes [mortality, myocardial infarction, stroke, target revascularization and MACE] were evaluated.
    • In RCTs, the use of catheter aspiration devices significantly decreased the risk of MACE but did not significantly impact mortality, myocardial infarction, stroke, or target revascularization versus control using the maximal duration of followup.
      • When the clinical trials eligible for pooling were limited to higher quality trials, the risk for MACE was significantly reduced when catheter aspiration devices were used versus control but the other endpoints were nonsignificantly impacted.
      • When the clinical trials eligible for pooling were evaluated at different time periods, mortality was significantly reduced at 365 days and target revascularization and MACE were significantly reduced at 180 days, but no other significant effects were seen for these or other final health outcomes at other time periods.
      • Three controlled observational studies were generally supportive of findings from RCTs as no significant differences were found at 30 days and 365 days for all five final health outcomes, with exception of 30-day stroke where one of two studies found an increased risk with catheter aspiration use.
    • In RCTs, the use of mechanical thrombectomy devices did not significantly impact mortality, myocardial infarction, stroke, target revascularization or MACE versus control using the longest duration of followup.
      • When the clinical trials eligible for pooling were limiting to higher quality trials, no significant impact on mortality, myocardial infarction, stroke, target revascularization or MACE occurred versus control.
      • When the clinical trials eligible for pooling were evaluated at different time periods, target revascularization and MACE was significantly reduced at 180 days and 365 days (one trial), respectively, but no other significant effects were seen for these or other final health outcomes at other time periods.
      • A controlled observational study was supportive of the myocardial infarction, stroke, target revascularization, and MACE findings.
    • In RCTs, the use of distal filter, distal balloon, proximal balloon, or the use of any one of these embolic protection devices did not significantly impact mortality, myocardial infarction, stroke, or MACE versus control using the longest duration of followup. However the use of a distal filter device or any one of the embolic protection devices significantly increased the risk of target revascularization although this was not seen with distal balloon or proximal balloon devices.
      • Limiting the trials to higher quality trials did not result in any changes in the significance of findings for any final health outcome.
      • When clinical trials eligible for pooling were evaluated at different time periods, stroke was significantly reduced at 30 days (one trial) with the use of a distal balloon embolic protection device versus control and target revascularization and MACE were significantly increased at 365 days (1 trial) with the use of a distal filter embolic protection device versus control or any embolic protection device versus control. No other significant effects were seen for these or other final health outcomes at other time periods.
      • In one controlled observational study the use of an embolic protection device did not significantly impact MACE versus control.

RCTs / Controlled Observational Studies in Patients with STEMI Assessing Intermediate Health Outcomes

  • Thirty-seven RCTs and four controlled observational studies evaluated patients with STEMI undergoing PCI and compared a thrombectomy or embolic protection device versus control. Six intermediate health outcomes (ST-segment resolution, MBG-3, TIMI-3 blood flow, ejection fraction, distal embolization and no reflow) were evaluated.
    • In RCTs, the use of catheter aspiration devices significantly increased the occurrence of ST-segment resolution, achievement of a MBG-3 and TIMI-3 blood flow while significantly reducing the risk of distal embolization and the occurrence of no reflow versus control. In RCTs, ejection fraction was not significantly impacted by catheter aspiration use versus control in the majority of trials (9 of 11) while one controlled observational study found a decreased ejection fraction in the catheter aspiration group versus control.
      • When the clinical trials eligible for pooling were limited to higher quality trials, significant benefits were again seen for the aforementioned intermediate outcomes. No impact on ejection fraction was seen versus control.
      • A controlled observational study was supportive of the findings for distal embolization although the use of a catheter aspiration device did not significantly impact the risk of resolving ST-segment elevation or attaining TIMI-3 blood flow (two studies).
    • In RCTs, the use of mechanical thrombectomy devices did not significantly impact ST-segment resolution, MBG-3, TIMI-3 blood flow, distal embolization, or no reflow versus control. In RCTs, ejection fraction was not impacted by mechanical thrombectomy devices versus control.
      • When the clinical trials eligible for pooling were limiting to higher quality trials, no significant impact was seen on any of the aforementioned intermediate health outcomes versus control.
      • In a controlled observational study the use of a mechanical thrombectomy device was associated with a significantly reduced rate of TIMI-3 blood flow versus control.
    • In RCTs, the use of distal filter embolic protection devices did not significantly impact ST-segment resolution, ejection fraction, MBG-3, TIMI-3 blood flow, distal embolization, or no reflow versus control.
      • When the clinical trials eligible for pooling were limiting to higher quality trials, no significant impact was seen on any of the aforementioned intermediate health outcomes versus control.
    • In RCTs, the use of distal balloon embolic protection devices significantly increased the occurrence of a MBG-3 and TIMI-3 blood flow but did not significantly impact ST-segment resolution, ejection fraction, distal embolization, or no reflow versus control.
      • When the clinical trials eligible for pooling were limiting to higher quality trials, significant increases in the occurrence of achieving a MBG-3 and TIMI-3 blood flow were still seen but no significant impact was seen on any of the other aforementioned intermediate health outcomes versus control.
    • In RCTs, the use of proximal balloon embolic protection devices did not significantly impact ST-segment resolution, MBG-3, TIMI-3 blood flow, distal embolization or ejection fraction versus control with no data on the other intermediate health outcomes.
      • Only one trial was available for the aforementioned intermediate health outcomes versus control and it was determined to be of good methodological quality.
    • In RCTs, the use of embolic protection devices combined significantly increased the occurrence of a MBG-3 and TIMI-3 blood flow but did not significantly impact ST-segment resolution, ejection fraction, distal embolization, or no reflow versus control.
      • When the clinical trials eligible for pooling were limiting to higher quality trials, significant increases in the occurrence of achieving MBG-3 and TIMI-3 blood flow were still seen but no significant impact was seen on any of the other aforementioned intermediate health outcomes versus control.

RCTs / Controlled Observational Studies in Mixed or Other ACS Populations Assessing Final Health Outcomes

  • Five RCTs and two controlled observational studies evaluated patients with mixed ACS (STEMI or NSTEMI or UA) undergoing PCI and compared a thrombectomy or embolic protection device versus control. Five final health outcomes [mortality, myocardial infarction, stroke, target revascularization and MACE] were evaluated.
    • In a RCT, the use of catheter aspiration devices did not significantly impact the risk of in-hospital mortality.
      • In a controlled observational study, the use of a catheter aspiration device significantly reduced the risk of 30-day mortality compared to control.
      • No trials or studies evaluated myocardial infarction, stroke, target revascularization, or MACE at any time period or mortality at additional time periods versus control.
    • In RCTs, the use of mechanical thrombectomy devices did not impact the risk of 30-day mortality (one trial), 30-day target revascularization (one trial), or 30-day MACE (one trial).
      • In an controlled observational study, the use of a mechanical thrombectomy device has no impact on the risk of 180-day mortality, myocardial infarction, target revascularization or MACE.
      • No trials or studies evaluated stroke or other aforementioned final health outcomes at other time points versus control.
    • In RCTs, the use of a distal filter embolic protection device did not impact the risk of 30-day mortality (one trial) or 180-day MACE (one trial) and there was insufficient data to analyze other final health outcomes. No additional trials or studies evaluated final health outcomes at additional time periods.
    • In RCTs, the use of distal balloon embolic protection devices did not impact the risk of mortality using the maximal duration of followup. Neither trial was determined to be of higher methodological quality.
      • Evaluating the clinical trials at different time periods of followup did not result in any significant findings for mortality, although each analysis was based on a single trial.
      • In a single trial, the risk of 180-day MACE was not impacted by the use of a distal balloon embolic protection device versus control.
      • No trials or studies evaluated stroke, target revascularization or aforementioned final health outcomes at individual time points.
    • In RCTs, the use of an embolic protection device (distal or proximal; filter or balloon) did not impact the risk of mortality using the longest duration of followup.
      • Limiting the pooled analysis to trials of higher methodological quality resulted in one trial and therefore pooling was not possible.
      • Evaluating the trials at ≤ 30 days did not significantly impact mortality.
      • No additional data for embolic protection devices combined was available in addition to what was reported in the individual embolic protection device categories.
  • Two RCTs and no controlled observational studies evaluated patients with other ACSs (NSTEMI or UA) undergoing PCI and compared a thrombectomy or embolic protection device versus control. Five final health outcomes [mortality, myocardial infarction, stroke, target revascularization and MACE] were evaluated.
    • In RCTs, the use of a distal filter embolic protection device did not impact the risk of 30-day mortality (one trial), in-hospital (one trial) or 30-day MACE (one trial) versus control.
    • No trials or studies evaluated stroke and there was insufficient data to analyze myocardial infarction or target revascularization.
    • No other device categories were evaluated.

RCTs / Controlled Observational Studies in Mixed or Other ACS populations Assessing Intermediate Health Outcomes

  • Six RCTs and one controlled observational study evaluated patients with mixed ACS (STEMI or NSTEMI or UA) undergoing PCI and compared a thrombectomy or embolic protection device versus control on intermediate health outcomes. Six intermediate health outcomes (ST-segment resolution, MBG-3, TIMI-3 blood flow, ejection fraction, distal embolization and no reflow) were evaluated.
    • In RCTs, the use of catheter aspiration devices did not significantly impact the risk of attaining TIMI-3 blood flow.
      • In a RCT, the use of a catheter aspiration device significantly increased the risk of attaining a MBG-3.
      • No trials or studies evaluated ST-segment elevation, ejection fraction, distal embolization or no reflow.
    • In RCTs, the use of mechanical thrombectomy devices significantly increased the risk of resolving ST-segment elevation (one trial) and had no impact on attaining TIMI-3 blood flow (one trial) versus control.
      • In an controlled observational study, the use of a mechanical thrombectomy device was associated with a significantly lower rate of TIMI-3 blood flow versus control.
      • No trials or studies evaluated ejection fraction, MBG-3, distal embolization or no reflow.
    • In RCTs, the use of distal filter embolic protection devices did not impact ejection fraction (one trial) or TIMI-3 blood flow (one trial) versus control.
      • No trials or studies evaluated resolution of ST-segment elevation, MBG-3, distal embolization or no reflow.
    • In RCTs, the use of a distal balloon embolic protection device significantly increased the risk of attaining a MBG-3 and did not impact the risk of attaining TIMI-3 blood flow. The trials included were not determined to be of higher methodological quality therefore sensitivity analysis was not possible.
      • In RCTs, the use of a distal balloon embolic protection device led to a significantly increased risk of resolving ST-segment elevation (one trial), significantly higher ejection fraction (one trial) and a significantly reduced risk of no reflow (one trial).
      • No trials or studies evaluated distal embolization.
    • No studies or trials evaluated the use of proximal balloon embolic protection devices in patients with mixed ACS.
    • In RCTs, the use of an embolic protection device did not impact the risk of attaining TIMI-3 blood flow.
      • In RCTs, the use of embolic protection devices increased ejection fraction in one trial and had no impact on ejection fraction in another trial.
      • For the resolution of ST-segment elevation, MBG-3, distal embolization, and no reflow no additional data the results are presented in the respective embolic protection device group and no additional data for embolic protection devices combined was available in addition to what was reported in the individual embolic protection device categories.
  • Two RCTs and no controlled observational studies evaluated patients with other ACSs (NSTEMI or UA) undergoing PCI and compared a thrombectomy or embolic protection device versus control. Six intermediate health outcomes (ST-segment resolution, MBG-3, TIMI-3 blood flow, ejection fraction, distal embolization and no reflow) were evaluated.
    • In RCTs, the use of a distal filter embolic protection device did not impact the risk of attaining TIMI-3 blood flow (one trial) versus control.
      • In a RCT, the use of a distal filter embolic protection device did not impact the risk of distal embolization (one trial) versus control.
      • There was insufficient data to evaluate no reflow and no trials or studies evaluated resolution of ST-segment elevation, ejection fraction, MBG-3 or distal embolization.
    • No other device categories were evaluated within this population.

Detailed Analysis

Study Design and Population Characteristics

Overall, 53 RCTs and 9 controlled observational studies have evaluated the impact of thrombectomy or embolic protection devices in ACS. Catheter aspiration, mechanical thrombectomy, distal filter embolic protection, distal balloon embolic protection and proximal balloon embolic protection devices have been evaluated for at least one endpoint but no studies evaluating proximal filter embolic protection devices met our inclusion and exclusion criteria.

One-hundred and twenty-seven publications of RCTs, which represent 43 unique trials (n=8185) met the inclusion criteria 11-21,27-141 for the quantitative analysis. Of the 127 publications, 50 were full articles11-18,27,29,40,44,62-64,66,68,69,71,72,74,75,83,84,88-90,95,98,103,107,111,112,114,115,119,123,125,133-141 48 were abstracts,20,21,28,30-33,52-60,70,76-79,85-87,91,93,96,99-102,104,105,108,109,116,117,120,122,124,127-132,134 and 29 were slide presentations.15,19,34-37,39,41-43,45-50,61,67,73,80-82,92,93,97,110,118,121 Of the 43 unique trials, 37 were in patients with STEMI and six were in patients with mixed ACS. The trial characteristics, trial quality assessment, and baseline and procedural characteristics can be found in Appendix C and Appendix E.

Thirty-seven unique RCTs evaluated the impact of thrombectomy or embolic protection devices versus control on final, intermediate, or adverse health outcomes when used as an adjunct to PCI as compared to PCI alone in patients with STEMI. Of the 37 trials, 17 trials12-16,19-21,62,68,69,71,74,83,85-87,138 (n=3355) evaluated the impact of catheter aspiration devices, five trials11,27,29,40,44 (n=1374) evaluated the impact of mechanical thrombectomy devices, five trials89,95,98,101,137 (n=962) evaluated the impact of distal filter embolic protection devices, nine trials17,103,107,111,112,119,133,135,136 (n=1479) evaluated the impact of distal balloon embolic protection devices and one trial 18 (n=284) evaluated the impact of proximal balloon embolic protection devices.

Amongst the 37 trials, the earliest trial was published in 200344 and the latest was published in 2010.11,135-138 The duration of followup of the trials ranged from “in-hospital”85,86,111,119,136 to 450 days.139 One trial reported a followup duration of 450 days, one of 240 days, one of 270 days, 13 trials reported a followup duration of 180 days, one trial reported a followup duration of 90 days, eight trials reported a followup duration of 30 days, and two trials reported a followup duration of 5-8 days. Fourteen trials received funding from industry, of which three reported additional funding from a university or clinical research grant. One trial reported a hospital as the funding source while 20 trials did not report a funding source and two trials were reported to be unfunded.

The mean age of patients enrolled in the 37 trials ranged from 55 to 69 years presenting within 6 to 48 hours of symptom onset. Twenty-one of the 34 trials included patients presenting within 12 hours of symptom onset. Males constituted at least half of the patients in the trials, ranging from 55.1 to 95 percent of the total population. The mean ischemic time reported in the 37 trials ranged from 120 to 510 minutes. The percent of patients presenting with TIMI 0/1 at baseline ranged from 54.8 to 100 percent. Of the 37 trials, 24 trials included patients with no prior fibrinolysis before the index PCI. Five trials included patients with prior fibrinolysis as well as primary PCI and eight trials did not report whether patients who received prior fibrinolysis were included or not.

Six unique trials evaluated the impact of thrombectomy or embolic protection devices versus control on final or intermediate health outcomes when used as an adjunct to PCI as compared to PCI alone in patients with mixed ACS.125-128,130,131 Of these six trials, two evaluated catheter aspiration devices,127,128 one evaluated a distal filter embolic protection device,126 and three evaluated distal balloon embolic protection devices.125,130,132 The earliest trial was published in 2003132 and the most recent trial was published in 2008.125 The duration of followup ranged from in-hopsital128 to 730 days .125 Three trials reported followup duration of in-hospital, two trials reported followup duration of 180 days, and one trial reported followup duration of 730 days. One trial received funding from industry while the other 5 trials did not report a funding source.

The mean age of patients enrolled in the six trials ranged from 55.17 years to 65.9 years. The percentage of males ranged from 76.79 to 95 percent. Two trials reported mean ischemic time which ranged from 372 to 474 minutes. One trial reported the percent of patients with TIMI 0/1 blood flow at baseline which ranged from 57 to 64 percent. Two trials did not include patients who previously failed fibrinoltyic therapy while the other four trials did not report this statistic.

Forty-eight publications met inclusion criteria for the qualitative synthesis.142-189 Of these publications, thirty-eight publications represented 20 unique studies (n=14771) 142-168 and twenty-one publications represented eighteen unique systematic reviews with meta-analysis (n=80181). 169-189 Of the 20 unique studies, 17 were full articles,143-146,148,149,153,154,156,158,160-166 nine were abstracts,142,147,150-152,157,159,167,168 and one was a slide presentation.155 Of the 20 unique studies, one study was a RCTs evaluating thrombectomy or embolic protection devices in patients with mixed ACS,166 nine studies were controlled observational studies,142,144-147,149,152-154 two studies were RCTs evaluating thrombectomy or embolic protection devices in patients with UA or NSTEMI,155,156 two studies were direct comparative RCTs,158,160 two studies were RCTs with selective inclusion/exclusion criteria in patients with STEMI,162,163 one study was a RCT with unique comparison in patients with STEMI,165 two studies were RCTs with unique comparison in patients with mixed ACS,164,168 and one study was a pooled analysis in STEMI patients.143 The characteristics of the studies, study quality assessment, and baseline and procedural characteristics can be found in Appendix C and Appendix E.

Amongst the 20 unique studies, the earliest study was published in 2002 and the latest was published in 2010. The duration of followup of the studies ranged from “in-hospital” to 365 days. The mean age group of the patients in the 23 studies ranged from 49.3 to 68 years presenting within 3 hours to 12 hours of symptom onset. Males constituted at least half of the patients in the studies, ranging from 50 to 100 percent of the total population. Two studies used an active control as a comparator. One trial compared the use of the catheter aspiration device Thrombuster® along with the use of mutant tissue plasminogen activator versus the use of the catheter aspiration device alone.165 The other trial compared the use of thrombectomy, distal protection and stenting versus thrombectomy and stenting alone.168 These two studies are therefore not discussed any further.

Of the 18 unique systematic reviews with meta-analysis (n=80181), 11 were full text articles169-171,176-179,189 and seven were abstracts.180-187 The earliest systematic review was published in 2006 and the latest was published in 2010. The number of studies included in each systematic review ranged from seven to 90 studies. The characteristics, quality assessment and results of these systematic reviews can be found in Appendix C. Although several recent systematic reviews have conducted meta-analyses, the majority are limited to patients with STEMI and do not evaluate adjunctive devices in other ACS, few included the analysis of adverse events which are further limited to procedure time and coronary perforation, and the most recent analyses did not evaluate embolic protection devices. Therefore, an updated analysis will more accurately reflect contemporary practice.

Specifically for key question 1, we present direct comparative data between agents first and subsequently present the comparisons of each type of device versus control for each endpoint.

Numerous endpoints of interest are evaluated at different time points and several trials report the endpoints at multiple time points. We present data for each endpoint at numerous time points as specified: maximum duration of followup (data using the longest reported time point evaluating that endpoint in the trial), ≤30 days (data using the shortest reported time point evaluating the endpoint in the trial up to and including 30 days), 365 days (data from a trial evaluating the endpoint for ≥365 days), 180 days (data from a trial evaluating the endpoint for 180 to 364 days), 30-days (data from a trial evaluating the endpoint for 30 to 179 days), and in-hospital (data from a trial evaluating the endpoint during the initial hospitalization).

Outcome Evaluation

A summary of the results for final health outcomes evaluated at the maximal duration of followup for each device category versus control can be found in Table 7 to Table 12 while the results for evaluations of intermediate outcomes in each device category versus control can be found in Table 13 to Table 26.

Table 7. Final health outcomes using the maximal duration of followup in randomized controlled trials evaluating catheter aspiration devices in patients with ST-segment elevation myocardial infarction.

Table 7

Final health outcomes using the maximal duration of followup in randomized controlled trials evaluating catheter aspiration devices in patients with ST-segment elevation myocardial infarction.

Table 8. Final health outcomes using the maximal duration of followup in randomized controlled trials evaluating mechanical thrombectomy devices in patients with ST-segment elevation myocardial infarction.

Table 8

Final health outcomes using the maximal duration of followup in randomized controlled trials evaluating mechanical thrombectomy devices in patients with ST-segment elevation myocardial infarction.

Table 9. Final health outcomes using the maximal duration of followup in randomized controlled trials evaluating distal filter embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 9

Final health outcomes using the maximal duration of followup in randomized controlled trials evaluating distal filter embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 10. Final health outcomes using the maximal duration of followup in randomized controlled trials evaluating distal balloon embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 10

Final health outcomes using the maximal duration of followup in randomized controlled trials evaluating distal balloon embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 11. Final health outcomes using the maximal duration of followup in randomized controlled trials evaluating proximal balloon embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 11

Final health outcomes using the maximal duration of followup in randomized controlled trials evaluating proximal balloon embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 12. Final health outcomes using the maximal duration of followup in randomized controlled trials evaluating embolic protection devices combined in patients with ST-segment elevation myocardial infarction.

Table 12

Final health outcomes using the maximal duration of followup in randomized controlled trials evaluating embolic protection devices combined in patients with ST-segment elevation myocardial infarction.

Table 13. Ejection fraction of direct comparative randomized controlled trials in ST-segment elevation myocardial infarction.

Table 13

Ejection fraction of direct comparative randomized controlled trials in ST-segment elevation myocardial infarction.

Table 14. Ejection fraction in randomized controlled trials evaluating catheter aspiration devices in patients with ST-segment elevation myocardial infarction.

Table 14

Ejection fraction in randomized controlled trials evaluating catheter aspiration devices in patients with ST-segment elevation myocardial infarction.

Table 15. Ejection fraction in randomized controlled trials evaluating mechanical thrombectomy devices in patients with ST-segment elevation myocardial infarction.

Table 15

Ejection fraction in randomized controlled trials evaluating mechanical thrombectomy devices in patients with ST-segment elevation myocardial infarction.

Table 16. Ejection fraction in randomized controlled trials evaluating distal filter embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 16

Ejection fraction in randomized controlled trials evaluating distal filter embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 17. Ejection fraction in randomized controlled trials evaluating thrombectomy or embolic protection devices in patients with mixed acute coronary syndromes.

Table 17

Ejection fraction in randomized controlled trials evaluating thrombectomy or embolic protection devices in patients with mixed acute coronary syndromes.

Table 18. Ejection fraction in randomized controlled trials evaluating distal balloon embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 18

Ejection fraction in randomized controlled trials evaluating distal balloon embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 19. Ejection fraction in randomized controlled studies with unique comparison in patients with mixed acute coronary syndromes.

Table 19

Ejection fraction in randomized controlled studies with unique comparison in patients with mixed acute coronary syndromes.

Table 20. Ejection fraction in randomized controlled trials evaluating proximal balloon embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 20

Ejection fraction in randomized controlled trials evaluating proximal balloon embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 21. Intermediate health outcomes in randomized controlled trials evaluating catheter aspiration devices in patients with ST-segment elevation myocardial infarction.

Table 21

Intermediate health outcomes in randomized controlled trials evaluating catheter aspiration devices in patients with ST-segment elevation myocardial infarction.

Table 22. Intermediate health outcomes in randomized controlled trials evaluating mechanical thrombectomy devices in patients with ST-segment elevation myocardial infarction.

Table 22

Intermediate health outcomes in randomized controlled trials evaluating mechanical thrombectomy devices in patients with ST-segment elevation myocardial infarction.

Table 23. Intermediate health outcomes in randomized controlled trials evaluating distal filter embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 23

Intermediate health outcomes in randomized controlled trials evaluating distal filter embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 24. Intermediate health outcomes in randomized controlled trials evaluating distal balloon embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 24

Intermediate health outcomes in randomized controlled trials evaluating distal balloon embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 25. Intermediate health outcomes in randomized controlled trials evaluating proximal balloon embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 25

Intermediate health outcomes in randomized controlled trials evaluating proximal balloon embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 26. Intermediate health outcomes in randomized controlled trials evaluating embolic protection devices combined in patients with ST-segment elevation myocardial infarction.

Table 26

Intermediate health outcomes in randomized controlled trials evaluating embolic protection devices combined in patients with ST-segment elevation myocardial infarction.

Mortality

Direct Comparative Trials

Catheter aspiration device versus distal balloon embolic protection device in patients with STEMI. One direct comparative randomized trial evaluated the impact of the Diver™ CE catheter aspiration device versus the GuardWire™ Plus distal balloon embolic protection device on mortality.160 In this trial, there was no difference in the risk of 30-day mortality [RR 1.00 (0.18, 5.54)].

Trials Versus Control

Catheter aspiration devices in patients with STEMI. Eleven RCTs evaluated the impact of catheter aspiration devices versus control on mortality using the maximal duration of followup.14-16,19,49,62,64,68,69,71,74,82,83,85,138 One trial was excluded from the pooled analysis of relative risk because no events occurred in either group during the prespecified time period.74 In the 10 trials suitable for pooling, the use of catheter aspiration devices did not significantly impact the risk of mortality [RR 0.69 (0.47, 1.02)] (Figure 3). The weighted-mean followup for mortality using the maximal duration of followup was 7.92 months. Statistical heterogeneity and publication bias were not detected (I2=0 percent, Egger's P=0.64).

This figure depicts the meta-analysis of the impact of catheter aspiration devices versus control on mortality using the maximal duration of followup. The first trial by Dudek and colleagues in 2010 provided a relative risk of 1.28 with a 95 percent confidence interval of 0.33 to 5.01. The second trial by Sardella and colleagues in 2009 provided a relative risk of 0.11 with 95 percent confidence interval of 0.00 to 0.93. The third trial by Chao and colleagues in 2008 provided a relative risk of 2.76 with a 95 percent confidence interval of 0.24 to infinity. The fourth trial by Chevalier and colleagues, also in 2008, provided a relative risk of 0.86 with a 95 percent confidence interval of 0.25 to 2.89. The fifth trial by Ikari and colleagues, also in 2008, provided a relative risk of 1.86 with a 95 percent confidence interval of 0.25 to 14.12. The sixth trial by Svilaas and colleagues, also in 2008, provided a relative risk of 0.61 with a 95 percent confidence interval of 0.38 to 0.98. The seventh trial by De Luca and colleagues in 2006 provided a relative risk of 0.70 with a 95 percent confidence interval of 0.16 to 2.95. The eighth trial by Kaltoft and colleagues, also in 2006, provided a relative risk of 0.33 with a 95 percent confidence interval 0.00 to 3.78. The ninth trial by Silva-Orrego and colleagues, also in 2006, was excluded from the meta-analysis because no deaths occurred in either the treatment or control group and therefore a relative risk could not be calculated. The tenth trial by Burzotta and colleagues provided a relative risk of 1.00 with a 95 percent confidence interval of 0.24 to 4.16. The eleventh trial by Noel and colleagues provided a relative risk of 0.36 with a 95 percent confidence interval of 0.00 to 4.03. The combined effect of the ten trials showed a relative risk of 0.69 with a 95 percent confidence interval of 0.47 to 1.02. The Cochran Q p-value was 0.870, the I-squared value was zero percent and the Egger's p-value was 0.638.

Figure 3

Impact of catheter aspiration devices versus control on mortality using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.870 I2: 0 percent

When limiting the pooled analysis to only trials of good methodological quality14-16,19,49,62,64,68,69,71,74,82,83,138 the risk of mortality using the maximal duration of followup was in the catheter aspiration device group compared to control [RR 0.70 (0.47, 1.03)]. The weighted mean duration of followup for this analysis was 8.08 months. Statistical heterogeneity (I2=0 percent) was not detected.

When the impact of catheter aspiration devices versus control was assessed in hospital [RR 0.81 (0.23, 2.86)], ≤30 days [RR 0.65 (0.39, 1.10)], 30-days [RR 0.61 (0.35, 1.07)], and 180-days [RR 0.89 (0.31, 2.51)] (Appendix Figures 1-4); no significant difference in the risk of mortality were seen in each analysis. In the 365-day analysis, there was a significant reduction in the risk of mortality with the use of catheter aspiration devices versus control in the two trials with available data [RR 0.62 (0.39, 0.98)] (Appendix Figure 5). Using the risk difference for the analysis [RD -0.03 (-0.06, -0.002), (CER 0.08, 0.14)], 33 patients would need to be treated to prevent one death.

Three controlled observational studies evaluated the association between the use of catheter aspiration devices during PCI and 30-day mortality142,144,152 and 365-day mortality.144 In the first study, the Export® aspiration catheter was compared to control.144 There was no significant difference in 30-day or 365-day mortality between the groups (4.9 percent versus 4.6 percent, p=0.82, 5.8 percent versus 7.4 percent, p=0.70, respectively).144 The second two studies did not report the names of the devices used. In the first study, catheter aspiration was compared to control and the rate of 30-day hospitalization was not significantly different between the groups (5.1 percent versus 4.4 percent, p=0.749). Authors report that after regression analysis, the results remained nonsignificant, although details were not provided.142 In the second study, there was no difference in the 30-day mortality rate with use of a catheter aspiration device during PCI versus PCI without catheter aspiration (2.6 percent versus 2.4 percent, p=0.74).152

Catheter aspiration devices in other ACS populations. One RCT evaluated the impact of the catheter aspiration device Rescue™ PT versus control on mortality in patients with acute myocardial infarction.128 The risk of in-hospital mortality was not significantly different between the catheter aspiration device group and control [RR 1.00 (0.18, 5.60)].

One controlled observational study of patients with acute myocardial infarction evaluated the association between the use of catheter aspiration devices and 30-day mortality.149 The following catheter aspiration devices were included in this study: RESCUE™ catheter, Thrombuster® catheter, Transvascular Aspiration Catheter™ and Export® PercuSurge system. In univariate analysis, the use of a catheter aspiration device was associated with a significantly lower rate of 30-day mortality compared to PCI without catheter aspiration [HR 0.64 (0.45, 0.93)] although upon adjustment for baseline characteristics, there was no longer a significant benefit associated with catheter aspiration devices [HR 0.66 (0.36, 1.19)].

Mechanical thrombectomy devices in patients with STEMI. Five RCTs evaluated the impact of mechanical thrombectomy devices versus control on mortality using the maximal duration of followup.11,27,29,40,44 One trial was excluded from the pooled analysis of relative risk because no deaths occurred within the prespecified time period in either group.27 In the four trials eligible for pooling, the use of a mechanical thrombectomy device did not significantly impact the risk of mortality [RR 1.19 (0.51, 2.76)]11,29,40,44 (Figure 4). The weighted-mean followup for mortality using the maximal duration of followup was 7.80 months. A higher level of statistical heterogeneity was detected (I2=54.9 percent) and publication bias was not be detected (Egger's P=0.736). All trials were determined to be of good methodological quality.11,27,29,40,44

Narrative: This figure depicts the meta-analysis of the impact of mechanical thrombectomy devices versus control on mortality using the maximal duration of followup. The first trial by Migliorini and colleagues in 2010 provided a relative risk of 0.55 with 95 percent confidence interval of 0.21 to 1.17. The second trial by Ali and colleagues in 2006 provided a relative risk of 2.80 with a 95 percent confidence interval of 1.07 to 7.38. The third trial by Lefèvre and colleagues in 2005 provided a relative risk of 1.52 with a 95 percent confidence interval of 0.47 to 4.88. The fourth trial by Antoniucci and colleagues in 2004 was excluded from the meta-analysis because no deaths occurred in either the treatment or control group and therefore a relative risk could not be calculated. The fourth trial by Napodano and colleagues in 2003 provided a relative risk of 1.00 with a 95 percent confidence interval of 0.24 to 4.16. The combined effect of the four trials showed a relative risk of 1.19 with a 95 percent confidence interval of 0.51 to 2.76. The Cochran Q p-value was 0.084, the I-squared value was 54.9 percent and the Egger's p-value was 0.736.

Figure 4

Impact of mechanical thrombectomy devices versus control on mortality using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.084 I2: 54.9 percent

When the impact of mechanical thrombectomy devices versus control was assessed during hospitalization [RR 1.00 (0.24, 4.16)], ≤30 days [RR 1.25 (0.47, 3.32)], 30-days [same results as the ≤ 30 days analysis], 180-days [RR 1.35 (0.53, 3.44)] and 365-days [RR 0.50 (0.21, 1.17)], (Appendix Figures 6-7); no significant changes were seen although the in-hospital and 365-day analyses were each based on a single trial.

One controlled observational study evaluated the association between the use of a mechanical thrombectomy device and mortality.145 Patients undergoing PCI with a mechanical thrombectomy device, either the AngioJet® XMI or XVG catheter, were compared to patients undergoing PCI without mechanical thrombectomy. The use of a mechanical thrombectomy device was not associated with a significant impact on the risk of in-hospital mortality compared to PCI without a mechanical thrombectomy device (2.9 percent versus 5.4 percent, p=0.11). After adjustment for baseline and angiographic characteristics, the use of a mechanical thrombectomy device did not significantly impact the odds of in-hospital mortality [OR 0.58 (0.26, 1.32)] compared to PCI without a mechanical thrombectomy device.

Mechanical thrombectomy devices in other ACS populations. One RCT evaluated the impact of mechanical thrombectomy devices versus control on mortality in patients with STEMI or UA.166 In this trial, the X-Sizer® device was compared to control. The use of a mechanical thrombectomy device did not significantly impact the risk of 30-day mortality [RR 2.00 (0.27, 14.89)] compared to control.

One controlled observational study evaluated the association between the use of mechanical thrombectomy devices and mortality.153 The types of ACSs included in this study were not reported. Patients undergoing PCI with the mechanical thrombectomy device AngioJet® were compared to patients undergoing PCI without mechanical thrombectomy and mortality was evaluated at 270 days. The use of a mechanical thrombectomy device was associated with a significant impact on 180-day mortality compared to PCI without a mechanical thrombectomy device (5.0 percent versus 6.5 percent, p=0.53).

Distal filter embolic protection devices in patients with STEMI. Five RCTs evaluated the impact of distal filter embolic protection devices versus control on mortality using the maximal duration of followup.89,95,98,101,137 In these trials, the use of distal filter embolic protection devices did not significantly impact the risk of mortality [RR 0.97 (0.54, 1.75)] (Figure 5). The weighted-mean followup for mortality using the maximal duration of followup was 10.84 months. Statistical heterogeneity and publication bias were not detected (I2=0 percent, Egger's P=0.739). Limiting the pooled analysis to only trials of good methodological quality,89,95,98,137 the risk of mortality using the maximal duration of followup remained nonsignificant [RR 0.97 (0.53, 1.79)]. The weighted mean duration of followup was 11.49 months. Statistical heterogeneity was not detected (I2=0 percent).

This figure depicts the meta-analysis of the impact of distal filter embolic protection devices versus control on mortality using the maximal duration of followup. The first trial by Ito and colleagues in 2010 provided a relative risk of 0.30 with a 95 percent confidence interval of 0.00 to 3.30. The second trial by Kelbaek and colleagues in 2008 provided a relative risk of 1.01 with 95 percent confidence interval of 0.40 to 2.56. The third trial by Cura and colleagues in 2007 provided a relative risk of 1.25 with a 95 percent confidence interval of 0.38 to 4.16. The fourth trial by Guetta and colleagues, also in 2007, provided a relative risk of 4.81 with a 95 percent confidence interval of 0.51 to infinity. The fifth trial by Lefevre and colleagues in 2004 provided a relative risk of 0.88 with a 95 percent confidence interval of 0.09 to 8.17. The combined effect of the five trials showed a relative risk of 0.97 with a 95 percent confidence interval of 0.54 to 1.75. The Cochran Q p-value was 0.760, the I-squared value was zero percent and the Egger's p-value was 0.739.

Figure 5

Impact of distal filter embolic protection devices versus control on mortality using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.760 I2: 0 percent

When the impact of distal filter embolic protection devices versus control was assessed at ≤30 days [RR 1.02 (0.50, 2.08)], 30-days [same results as the ≤30 days analysis],180-days [RR 1.25 (0.38, 4.16)] (Appendix Figure 8) and 365-days [RR 0.87 (0.43, 1.78)] no significant changes in the risk of mortality were seen in each analysis, although the 180-day analysis was based on a single trial.

No controlled observational trials were conducted that evaluated the impact of distal filter embolic protection devices on mortality.

Distal filter embolic protection devices in other ACS populations. Two RCTs evaluated the impact of distal filter embolic protection devices versus control on mortality in other ACS populations using the maximal duration of followup. However, the trials were not suitable for pooling because one trial was conducted in patients with either STEMI or NSTEMI126 and the other trial was conducted in patients with either NSTEMI or UA.156 In the first trial, the impact of a distal filter embolic protection device (FilterWire EX™) on 30-day mortality versus control126 in patients with STEMI or NSTEMI was evaluated. The use of a distal filter embolic protection device did not significantly impact the risk of 30-day mortality [RR 0.67 (0.14, 3.27)] compared to control. This trial was determined to be of good methodological quality. In the second trial, the impact of a distal filter embolic protection device (AngioGuard™) on 30-day mortality156 in patients with NSTEMI or UA was evaluated. The risk of 30-day mortality was not significantly different between the distal filter embolic protection device group and control [RR 1.00 (0.24, 2.45)]. This trial was determined to be of fair methodological quality.

No controlled observational studies of distal filter embolic protection devices assessed this outcome.

Distal balloon embolic protection devices in patients with STEMI. Four RCTs evaluated the impact of distal balloon embolic protection devices versus control on mortality using the maximal duration of followup.17,103,110,112,133 The use of a distal balloon embolic protection device did not significantly impact the risk of mortality using the maximal duration of followup [RR 0.82 (0.45, 1.51)] (Figure 6). The weighted-mean followup for mortality using the maximal duration of followup was 6 months. A lower level of statistical heterogeneity was found (I2=2.5 percent) and publication bias was detected (Egger's P=0.023). All trials were determined to be of good methodological quality.17,103,110,112,133

This figure depicts the meta-analysis of the impact of distal balloon embolic protection devices versus control on mortality using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. The first trial by Tahk and colleagues in 2008 provided a relative risk of 0.19 with 95 percent confidence interval of 0.00 to 1.81. The second trial by Hahn and colleagues in 2007 provided a relative risk of 0.12 with a 95 percent confidence interval of 0.00 to 0.91. The third trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 0.97 with a 95 percent confidence interval of 0.44 to 2.14. The fourth trial by Stone and colleagues in 2005 provided a relative risk of 0.96 with a 95 percent confidence interval of 0.38 to 2.43. The combined effect of the four trials showed a relative risk of 0.82 with a 95 percent confidence interval of 0.45 to 1.51. The Cochran Q p-value was 0.380, the I-squared value was 2.5 percent and the Egger's p-value was 0.023.

Figure 6

Impact of distal balloon embolic protection devices versus control on mortality using the maximal duration of followup versus control in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.380 I2: 2.5 percent

When the impact of distal balloon embolic protection devices versus control was assessed at in-hospital [RR 0.69 (0.24, 2.03)], ≤30 days [RR 0.64 (0.30, 1.39)], 30-days [same results as the ≤30 day analysis], and 180-days [RR 0.86 (0.48, 1.57)] (Appendix Figures 9-10); no significant changes were seen in the risk of mortality in each analysis, although the in-hospital analysis is based on a single trial.

No controlled observational studies of distal balloon embolic protection devices assessed this outcome.

Distal balloon embolic protection devices in other ACS populations. Two RCTs evaluated the impact of distal balloon embolic protection devices versus control on mortality in patients with acute myocardial infarction using the maximal duration of followup.125,131 The use of a distal balloon embolic protection device did not significantly impact the risk of mortality using the maximal duration of followup [RR 0.31 (0.10, 1.77)] (Figure 7). The weighted mean duration of followup was 10.99 months for this analysis. Neither trial was determined to be of good methodological quality.125,132

This figure depicts the meta-analysis of the impact of distal balloon embolic protection devices versus control on mortality using the maximal duration of followup in patients with mixed acute coronary syndrome. The first trial by Parikh and colleagues in 2008 provided a relative risk of 0.31 with 95 percent confidence interval of 0.05 to 1.91. The second trial by Matsushita and colleagues in 2003 provided a relative risk of 0.33 with a 95 percent confidence interval of 0.00 to 2.79. The combined effect of the two trials showed a relative risk of 0.31 with a 95 percent confidence interval of 0.06 to 1.77. The Cochran Q p-value was 0.976 and too few strata were available to calculate the I-squared value and the Egger's p-value.

Figure 7

Impact of distal balloon embolic protection devices versus control on mortality using the maximal duration of followup in patients with mixed acute coronary syndromes. Cochran Q: P=0.976 I2: Too few strata

When the impact of distal balloon embolic protection devices versus control was assessed during hospitalization [RR 0.33 (0.00, 2.79)]132 and at 365-days [RR 0.31 (0.05, 1.91)],125 no significant changes in risk were seen, although each analysis was based on a single trial.

One RCT evaluated the impact of distal balloon embolic protection devices versus abciximab therapy on 180-day mortality in patients with acute myocardial infarction.164 In this trial, the PercuSurge device was used. The risk of 180-day mortality could not be calculated because no events occurred in either group within the prespecified time period.

No controlled observational studies of distal balloon embolic protection devices assessed this outcome.

Proximal balloon embolic protection devices in patients with STEMI. One RCT evaluated the impact of the proximal balloon embolic protection device Proxis™ versus control on mortality.18,141 The use of a proximal balloon embolic protection device did not significantly impact the risk of 30-day mortality [RR 1.01 (0.18, 5.69)] or 180-day mortality [RR 0.51 (0.11, 2.33)] versus control.

Proximal balloon embolic protection devices in other ACS populations. No studies or trials were available that were evaluating the impact of proximal balloon embolic protection devices versus control on mortality in the population.

Embolic protection devices combined in patients with STEMI. Ten RCTs evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) versus control on the occurrence of mortality using the maximal duration of followup.17,18,89,95,98,101,103,112,133,137 In these trials, the use of embolic protection devices combined did not significantly impact the risk of mortality [RR 0.87 (0.58, 1.30)] (Figure 8). The weighted-mean followup for mortality using the maximal duration of followup was 8.11 months. Statistical heterogeneity and publication bias were not detected (I2=0 percent, Egger's P=0.254).

This figure depicts the meta-analysis of the impact of embolic protection devices combined versus control on mortality using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. The first trial by Ito and colleagues in 2010 provided a relative risk of 0.30 with a 95 percent confidence interval of 3.30. The second trial by Haeck and colleagues in 2009 provided a relative risk of 0.51 with 95 percent confidence interval of 0.11 to 2.33. The third trial by Kelbaek and colleagues in 2008 provided a relative risk of 0.87 with a 95 percent confidence interval of 0.43 to 1.78. The fourth trial by Takh and colleagues in 2008 provided a relative risk of 0.19 with 95 percent confidence interval of 0.00 to 1.81. The fifth trial by Cura and colleagues in 2007 provided a relative risk of 1.25 with a 95 percent confidence interval of 0.38 to 4.16. The sixth trial by Guetta and colleagues, also in 2007, provided a relative risk of 4.81 with a 95 percent confidence interval of 0.51 to infinity. The seventh trial by Hahn and colleagues, also in 2007, provided a relative risk of 0.12 with a 95 percent confidence interval of 0.00 to 0.91. The eight trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 0.97 with a 95 percent confidence interval of 0.44 to 2.14. The ninth trial by Stone and colleagues in 2005 provided a relative risk of 0.96 with a 95 confidence interval of 0.38 to 2.43. The tenth trial by Lefevre and colleagues in 2004 provided a relative risk of 0.88 with a 95 percent confidence interval of 0.09 to 8.17. The combined effect of the ten trials showed a relative risk of 0.87 with a 95 percent confidence interval of 0.58 to 1.30. The Cochran Q p-value was 0.798, the I-squared value was zero percent and the Egger's p-value was 0.254.

Figure 8

Impact of embolic protection devices combined versus control on mortality using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.798 I2: 0 percent

When limiting the pooled analysis to only trials of good methodological quality,17,18,89,95,98,103,112,133,137 the risk of mortality using the maximal duration of followup remained nonsignificant in the embolic protection devices combined group compared to control [RR 0.87 (0.57, 1.31)]. The weighted mean followup for mortality using the maximal duration of followup was 8.31 months. No statistical heterogeneity was found (I2=0 percent).

When the impact of embolic protection devices combined versus control was assessed at in-hospital [RR 0.69 (0.24, 2.03)], ≤30 days [RR 0.84 (0.50, 1.39)], 30-days [same results as the ≤30 day analysis],180-days [RR 0.87 (0.52, 1.46)], and 365-days [RR 0.87 (0.43, 1.78)], (Appendix Figures 11-12); no significant changes in the risk of mortality were seen in each analysis, although the in-hospital and 365-day analyses were based on a single trial.

Embolic protection devices combined in other ACS populations. Three RCTs evaluated the impact of embolic protection device (distal or proximal; filter or balloon) versus control in patients with mixed ACS on mortality using the maximal duration of followup.125,126,132 The use of an embolic protection device did not significantly impact the risk of mortality [RR 0.59 (0.18, 1.89)] versus control (Figure 9). The weighted mean duration of followup was 8.12 months for this analysis. Statistical heterogeneity was not detected (I2=0 percent) but publication bias could not be evaluated. One trial was determined to be of good methodological quality,134 therefore a pooled analysis limited to trials of higher methodological quality was not possible.

This figure depicts the meta-analysis of the impact of embolic protection devices combined versus control on mortality using the maximal duration of followup in patients with mixed acute coronary syndrome. The first trial by Parikh and colleagues in 2008 provided a relative risk of 0.31 with 95 percent confidence interval of 0.05 to 1.91. The second trial by Gick and colleagues in 2005 provided a relative risk of 1.00 with a 95 percent confidence interval of 0.24 to 4.25. The third trial by Matsushita and colleagues in 2003 provided a relative risk of 0.33 with 95 percent confidence interval of 0.00 to 2.79. The combined effect of the three trials showed a relative risk of 0.59 with a 95 percent confidence interval of 0.18 to 1.89. The Cochran Q p-value was 0.619, the I-squared value was zero percent and there were too few strata to calculate the Egger's p-value.

Figure 9

Impact of embolic protection devices combined versus control on mortality using the maximal duration of followup in patients with other acute coronary syndromes. Cochran Q: P=0.619 I2: 0 percent

When the impact of embolic protection devices combined versus control was evaluated at ≤30days, no significant impact on the risk of mortality was found [RR 0.55 (0.12, 2.50)]. (Appendix Figure 13).

Myocardial Infarction

Direct Comparative Trials

Catheter aspiration device versus distal balloon embolic protection device in patients with STEMI. One direct comparative randomized trial evaluated the impact of the Diver™-Invatec catheter aspiration device versus the Export®-Medtronic catheter aspiration device on myocardial infarction using the maximum duration of followup which in this case was 365 days.158 Patients with either Q-wave or nonQ-wave myocardial infarctions were evaluated. In this trial, the use of Diver™-Invatec did not significantly impact the risk of 365-day Q-wave myocardial infarction [RR 2.88 (0.25 to infinity)] or the risk of 365-day nonQ-wave myocardial infarction [RR 0.32 (0.00, 3.63)] compared to Export®-Medtronic. This trial was determined to be of good methodological quality.

Catheter aspiration device versus catheter aspiration device in patients with STEMI. One direct comparative randomized trial evaluated the impact of the Diver™ CE catheter aspiration device versus the Guardwire™ Plus distal balloon embolic protection device on 30-day myocardial infarction.160 In this trial, the use of Diver™ CE did not significantly impact the risk of 30-day myocardial infarction [RR 3.00 (0.26, infinity)] compared to Guardwire™ Plus.

Trials Versus Control

Catheter aspiration devices in patients with STEMI. Ten RCTs evaluated the impact of catheter aspiration devices versus control on the occurrence of myocardial infarction over the maximal duration of followup.12,14-16,19,49,62,64,69,71,74,82,83,138 In these trials, the use of catheter aspiration devices did not significantly impact the risk of myocardial infarction using the maximal duration of followup [RR 0.61 (0.36, 1.04)] (Figure 10). The weighted-mean followup for myocardial infarction in this analysis was 8.80 months. Statistical heterogeneity and publication bias were not detected (I2=0 percent, Egger's P=0.651).

This figure depicts the meta-analysis of the impact of catheter aspiration devices versus control on myocardial infarction using the maximal duration of followup. The first trial by Dudek and colleagues in 2010 provided a relative risk of 0.32 with a 95 percent confidence interval of 0.05 to 2.19. The second trial by Liistro and colleagues in 2009 provided a relative risk of 1.02 with 95 percent confidence interval of 0.24 to 4.26. The third trial by Sardella and colleagues, also in 2009, provided a relative risk of 0.33 with a 95 percent confidence interval of 0.00 to 3.76. The fourth trial by Chevalier and colleagues in 2008 provided a relative risk of 2.15 with 95 percent confidence interval of 0.28 to 16.30. The fifth trial by Ikari and colleagues, also in 2008, provided a relative risk of 0.31 with a 95 percent confidence interval of 0.00 to 3.55. The sixth trial by Svilaas and colleagues, also in 2008, provided a relative risk of 0.52 with a 95 percent confidence interval of 0.27 to 1.03. The seventh trial by De Luca and colleagues in 2006 provided a relative risk of 3.25 with a 95 percent confidence interval of 0.29 to infinity. The eighth trial by Kaltoft and colleagues, also in 2006, provided a relative risk of 0.33 with a 95 confidence interval of 0.00 to 3.78. The ninth trial by Silva-Orrego and colleagues, also in 2006, provided a relative risk of 0.32 with a 95 percent confidence interval of 0.00 to 3.60. The tenth trial by Burzotta and colleagues in 2005 provided a relative risk of 1.00 with a 95 percent confidence interval of 0.18 to 5.50. The combined effect of the ten trials showed a relative risk of 0.61 with a 95 percent confidence interval of 0.36 to 1.04. The Cochran Q p-value was 0.915, the I-squared value was zero percent and the Egger's p-value was 0.651.

Figure 10

Impact of catheter aspiration devices versus control on myocardial infarction using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.915 I2: 0 percent

When limiting the pooled analysis to only trials of good methodological quality12,14-16,19,49,62,64,69,71,74,82,83,138 the risk of myocardial infarction using the maximal duration of followup remained nonsignificantly impacted in the catheter aspiration device group compared to control [RR 0.61 (0.36, 1.04)]. The weighted mean duration of followup was 8.80 months. Statistical heterogeneity (I2=0 percent) was not detected.

When the impact of catheter aspiration device use versus control was assessed at in-hospital [RR 0.32 (0.03, 3.06)], <30 days [RR 0.55 (0.24, 1.25)], 30 days [RR 0.60 (0.25, 1.45)], 180 days [RR 0.70 (0.24, 1.99)], and 365 days [RR 0.51 (0.26, 1.00)] (Appendix Figures 14-18); no significant difference in the risk of myocardial infarction were seen versus control in each analysis.

Two controlled observational studies evaluated the association between the use of catheter aspiration devices during PCI and 30-day myocardial infarction144,152 and 365-day myocardial infaction.144 In the first study, the Export® aspiration catheter was compared to control. There was no significant difference in 30-day or 365-day myocardial infarction (1.2 percent versus 0.5 percent, p=0.59, 3.9 percent versus 1.4 percent, p=0.10, respectively).144 The second study did not report the catheter aspiration devices studied.152 The use of a catheter aspiration device was not associated with a significantly different rate of 30-day myocardial infarction compared to control (1.3 percent versus 1.9, p=0.44).

Catheter aspiration devices in other ACS population. No trials or studies evaluated the impact of catheter aspiration devices versus control in this population.

Mechanical thrombectomy devices in patients with STEMI. Five RCTs evaluated the impact of mechanical thrombectomy devices versus control on myocardial infarction using the maximal duration of followup.11,27,29,40,44 Two trials were excluded from the pooled analysis of relative risk because no myocardial infarctions occurred within the prespecified time period in either treatment group.27,40 In the three trials eligible for pooling, the use of a mechanical thrombectomy device did not significantly impact the risk of myocardial infarction [RR 0.71 (0.27, 1.85)]11,29,44 (Figure 11). The weighted-mean followup for myocardial infarction using the maximal duration of followup was 8.98 months. Statistical heterogeneity was not detected (I2=0 percent) and publication bias could not be evaluated. All of the trials in the pooled analysis were determined to be of good methodological quality.11,29,44

This figure depicts the meta-analysis of the impact of mechanical thrombectomy devices versus control on myocardial infarction using the maximal duration of followup. The first trial by Migliorini and colleagues in 2010 provided a relative risk of 0.66 with 95 percent confidence interval of 0.13 to 3.29. The second trial by Ali and colleagues also in 2006 was excluded from the meta-analysis because no myocardial infarctions occurred in either the treatment or the control groups and therefore a relative risk could not be calculated. The third trial by Lefevre and colleagues in 2005 provided a relative risk of 0.51 with 95 percent confidence interval of 0.11 to 2.31. The fourth trial by Antonuicci and colleagues in 2004 was excluded from the meta-analysis because no myocardial infarction occurred in either treatment arm and therefore a relative risk could not be calculated. The fifth trial by Napodano and colleagues in 2003 provided a relative risk of 1.00 with a 95 percent confidence interval of 0.24 to 4.16. The combined effect of the three trials showed a relative risk of 0.71 with a 95 percent confidence interval of 0.27 to 1.85. The Cochran Q p-value was 0.838, the I-squared value was zero percent and there were too few strata to calculate Egger's p-value.

Figure 11

Impact of mechanical thrombectomy devices versus control on myocardial infarction using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.838 I2: 0 percent

When the impact of mechanical thrombectomy device use versus control was assessed at in-hospital [RR 1.00 (0.11, 9.41)], ≤30 days [RR 0.63 (0.21, 1.96)], 30 days [same results as the ≤30 days analysis], 180-days [RR 0.57 (0.17, 1.92)], 365-days [RR 0.66 (0.13, 3.29)] (Appendix Figures 19-20); no significant difference in the risk of myocardial infarction was seen in each analysis, although the 365-day analysis was based on a single trial.

One controlled observational study evaluated the association between the use of a mechanical thrombectomy device and in-hospital myocardial infarction.145 Patients undergoing PCI with a mechanical thrombectomy device, either the AngioJet® XMI or XVG catheter, were compared to patients undergoing PCI without mechanical thrombectomy. The use of a mechanical thrombectomy device was not associated with a significant difference in the rate of in-hospital myocardial infarction compared to PCI without a mechanical thrombectomy device (1.0 percent versus 2.5 percent, p=0.10).

Mechanical thrombectomy devices in other ACS populations. One controlled observational study evaluated the association between the use of mechanical thrombectomy devices and myocardial infarction.153 The types of ACSs included in this study were not reported. Patients undergoing PCI with the AngioJet® mechanical thrombectomy device were compared to patients undergoing PCI without mechanical thrombectomy. The use of a mechanical thrombectomy device was not associated with a significantly different rate of 180-day myocardial infarction compared to PCI without a mechanical thrombectomy device (4.0 percent versus 2.1 percent, p=0.14).

Distal filter embolic protection devices in patients with STEMI. Five RCTs evaluated the impact of distal filter embolic protection devices versus control on the occurrence of myocardial infarction using the maximal duration of followup versus control.89,95,98,101,137 One trial was excluded from the analysis because no events occurred in the groups compared.137 In these four remaining trials, the use of distal filter embolic protection devices did not significantly impact the risk of myocardial infarction using the maximal duration of followup [RR 0.72 (0.15, 3.34)] (Figure 12). The weighted-mean followup for myocardial infarction using the maximal duration of followup was 11.22 months. A lower level of statistical heterogeneity was detected (I2=39.8 percent) but publication bias was not detected (Egger's P=0.128).

This figure depicts the meta-analysis of the impact of distal filter embolic protection devices versus control on myocardial infarction using the maximal duration of followup. The first trial by Ito and colleagues in 2010 was excluded from the analysis becasue no events occurred in the groups compared. The second trial by Kelbaek and colleagues in 2008 provided a relative risk of 5.03 with 95 percent confidence interval of 0.79 to 32.40. The third trial by Cura and colleagues in 2007 provided a relative risk of 0.09 with 95 percent confidence interval of 0.00 to 0.74. The fourth trial by Guetta and colleagues, also in 2007, provided a relative risk of 0.32 with 95 percent confidence interval of 0.00 to 3.63. The fifth trial by Lefevre and colleagues in 2004 provided a relative risk of 0.88 with a 95 percent confidence interval of 0.09 to 8.17. The combined effect of the five trials showed a relative risk of 0.72 with a 95 percent confidence interval of 0.15 to 3.34. The Cochran Q p-value was 0.173, the I-squared value was 39.8 percent and Egger's p-value was 0.128

Figure 12

Impact of distal filter embolic protection devices versus control on myocardial infarction using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.173 I2: 39.8 percent

Limiting the pooled analysis to only trials of good methodological quality89,95,98,137 the risk of myocardial infarction using the maximal duration of followup remained nonsignificant [RR 0.56 (0.06, 5.02)]. The weighted mean duration of followup was 11.93 months. A higher level of statistical heterogeneity was detected (I2=60 percent).

When the impact of distal filter embolic protection devices use versus control was assessed at ≤30 days [RR 0.73 (0.12, 4.44), 30-days [same result as ≤30 days], 180-days [RR 0.09 (0.00, 0.74)], and 365-days [RR 2.35 (0.61, 9.00)] (Appendix Figure 21), no significant difference in the risk of myocardial infarction were observed, although the 180-day and 365-day results are each based on a single trial.

There were no available controlled observational studies evaluating this endpoint.

Distal filter embolic protection devices in other ACS populations. Two RCTs evaluated the impact of distal filter embolic protection devices versus control in patients with other ACSs on myocardial infarction using the maximal duration of followup.126,156 These trials were not suitable for pooling because one trial evaluated patients with either STEMI or NSTEMI126 and the other trial evaluated patients with UA.156 Additionally, the risk of myocardial infarction could not be calculated in either case because no events occurred in either trial during the specified time period.

There were no available controlled observational studies evaluating this endpoint.

Distal balloon embolic protection devices in patients with STEMI. Five RCTs evaluated the impact of distal balloon embolic protection devices versus control on myocardial infarction using the maximal duration of followup.17,103,107,110,112,133 The use of a distal balloon embolic protection device did not significantly impact the risk of myocardial infarction [RR 0.67 (0.29, 1.57)] (Figure 13). The weighted-mean followup for myocardial infarction using the maximal duration of followup was 6 months. Statistical heterogeneity and publication bias were not detected (I2=0 percent, Egger's P=0.820). All trials were determined to be of good methodological quality.17,103,107,110,112,133

This figure depicts the meta-analysis of the impact of distal balloon embolic protection devices versus control on myocardial infarction using the maximal duration of followup. The first trial by Tahk and colleagues in 2008 provided a relative risk of 0.96 with 95 percent confidence interval of 0.10 to 9.09. The second trial by Hahn and colleagues in 2007 provided a relative risk of 0.35 with 95 percent confidence interval of 0.00 to 3.88. The third trial by Matsuo and colleagues, also in 2007, provided a relative risk of 2.78 with 95 percent confidence interval of 0.24 to infinity. The fourth trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 0.32 with a 95 percent confidence interval of 0.00 to 3.17. The fifth trial by Stone and colleagues in 2005 provided a relative risk of 0.64 with a 95 percent confidence interval of 0.24 to 1.70. The combined effect of the five trials showed a relative risk of 0.67 with a 95 percent confidence interval of 0.29 to 1.57. The Cochran Q p-value was 0.877, the I-squared value was zero percent and Egger's p-value was 0.820.

Figure 13

Impact of distal balloon embolic protection devices versus control on myocardial infarction using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.877 I2: 0 percent

When the impact of distal balloon protection device use versus control was assessed at in-hospital [RR 0.32 (0.00, 3.71)], ≤30 days [RR 0.85 (0.32, 2.23)], 30 days [same results as the ≤30 days analysis], and 180 days [same results as maximal duration of followup analysis] (Appendix Figure 22-23); no significant differences in the risk of myocardial infarction were seen in each analysis, although the in-hospital analysis is based on a single trial.

There were no available controlled observational studies that evaluated this endpoint.

Distal balloon embolic protection devices in other ACS populations. One RCT evaluated the impact of the distal balloon embolic protection device PercuSurge versus abciximab therapy on myocardial infarction in patients with acute myocardial infarction.164 The use of a distal balloon embolic protection device did not significantly impact the risk of 180-day myocardial infarction [RR 1.66 (0.34, 8.10)] compared to abciximab therapy.

There were no controlled observational studies that evaluated this endpoint.

Proximal balloon embolic protection devices in patients with STEMI. One RCT evaluated the impact of the proximal balloon embolic protection device Proxis™ versus control on myocardial infarction.18,141 The use of a proximal balloon embolic protection device did not significantly impact the risk of having a myocardial infarction over 30 days [RR 0.68 (0.14, 3.34)] or 180 days [RR 1.0.1 (0.24, 4.33)].

Proximal balloon embolic protection devices in other ACS populations. No trials or studies were available that evaluated the impact of proximal balloon embolic protection devices versus control on myocardial infarction in this population.

Embolic protection devices combined in patients with STEMI. Eleven RCTs evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) versus control on the occurrence of myocardial infarction using the maximal duration of followup.17,18,89,95,98,101,103,107,112,133,137 One trial was excluded from the pooled analysis because no events occurred in the groups compared.137 In the remaining ten trials, the use of embolic protection devices combined did not significantly impact the risk of myocardial infarction [RR 0.83 (0.45, 1.53)] (Figure 14). The weighted-mean followup for myocardial infarction using the maximal duration of followup was 8.08 months. Statistical heterogeneity and publication bias were not detected (I2=0 percent, Egger's P=0.372).

This figure depicts the meta-analysis of the impact of embolic protection devices combined versus control on myocardial infarction using the maximal duration of followup. The first trial by Ito and colleagues in 2010 was excluded from the analysis becasue no events occured in teh groups compared. The second trial by Haeck colleagues in 2009 provided a relative risk of 1.01 with 95 percent confidence interval of 0.24 to 4.33. The third trial by Kelbaeck and colleagues in 2008 provided a relative risk of 2.53 with 95 percent confidence interval of 0.67 to 8.28. The fourth trial by Tahk and colleagues, also in 2008, provided a relative risk of 0.96 with 95 percent confidence interval of 0.10 to 9.09. The fifth trial by Cura and colleagues in 2007 provided a relative risk of 0.09 with a 95 percent confidence interval of 0.00 to 0.74. The sixth trial by Guetta and colleagues, also in 2007, provided a relative risk of 0.32 with a 95 percent confidence interval of 0.00 to 3.63. The seventh trial by Hahn and colleagues, also in 2007, provided a relative risk of 0.35 with a 95 percent confidence interval of 0.00 to 3.88. The eight trial by Matsuo and colleagues, also in 2007, provided a relative risk of 2.78 with a 95 percent confidence interval of 0.24 to infinity. The ninth trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 0.32 with a 95 percent confidence interval of 0.00 to 3.71. The tenth trial by Stone and colleagues in 2005 provided a relative risk of 0.64 with a 95 percent confidence interval of 0.24 to 1.70. The eleventh trial by Lefevre and colleagues in 2004 provided a relative risk of 0.88 with a 95 percent confidence interval of 0.09 to 8.17. The combined effect of the eleven trials showed a relative risk of 0.83 with a 95 percent confidence interval of 0.45 to 1.53. The Cochran Q p-value was 0.689, the I-squared value was zero percent and Egger's p-value was 0.372.

Figure 14

Impact of embolic protection devices combined versus control on myocardial infarction using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.689 I2: 0 percent

When limiting the pooled analysis to only trials of good methodological quality,17,18,89,95,98,103,107,112,133,137 the risk of myocardial infarction using the maximal duration of followup remained nonsignificant in the embolic protection devices combined group compared to control [RR 0.83 (0.45, 1.55)]. The weighted-mean followup for myocardial infarction using the maximal duration of followup was 8.27 months. No statistical heterogeneity was found (I2=0 percent).

When the impact of embolic protection devices combined versus control was assessed at in-hospital [RR 0.32 (0.00, 3.71)], ≤30 days [RR 0.83 (0.41, 1.69)], 30-days [same results as the ≤30 days analysis], 180-days [RR 0.65 (0.31, 1.33)], and 365-days [RR 2.35 (0.67, 8.28)] (Appendix Figures 24-25); no significant differences in the risk of myocardial infarction were seen in each analysis, although the in-hospital and 365-day analyses were based on a single trial each.

There were no available controlled observational studies that evaluated this endpoint.

Embolic protection devices combined in other ACS populations. No trials or studies were available that evaluated the impact of any embolic protection device versus control on myocardial infarction in addition to the three trials reported above. Pooling was not suitable because each trial evaluated a different ACS.

Stroke

Direct Comparative Trials

Catheter aspiration devices versus distal balloon embolic protection devices in patients with STEMI. One direct comparative randomized trial evaluated the impact of the Diver™ CE catheter aspiration device versus the Guardwire™ Plus distal balloon embolic protection device on stroke.160 The risk of 30-day stroke could not be calculated because no events occurred in either group during the specified time period.

Trials Versus Control

Catheter aspiration devices in patients with STEMI. Five RCTs evaluated the impact of catheter aspiration devices versus control on stroke using the maximal duration of followup.14,15,49,71,74,82,83 One trial was excluded from the pooled analysis because no events occurred in either group during the prespecified time period.74 In the four trials eligible for pooling, the use of catheter aspiration devices did not significantly impact the risk of stroke [RR 3.18 (0.73, 13.88)] (Figure 15). The weighted-mean followup for stroke using the maximal duration of followup was 0.79 months. There was no statistical heterogeneity (I2=0 percent) but publication bias was detected (Egger's P=0.001). All of the trials included in the pooled analysis were determined to be of good methodological quality.14,15,49,71,74,82,83

This figure depicts the meta-analysis of the impact of catheter aspiration devices versus control on stroke using the maximal duration of followup. The first trial by Sardella and colleagues in 2009 provided a relative risk of 4.94 with 95 percent confidence interval of 0.52 to infinity. The second trial by Chevalier and colleagues in 2008 provided a relative risk of 5.37 with 95 percent confidence interval of 0.57 to infinity. The third trial by Kaltoft and colleagues in 2006 provided a relative risk of 4.95 with 95 percent confidence interval of 0.52 to infinity. The fourth trial by Silva-Orrego and colleagues, also in 2006, was excluded from the meta-analysis because no stroke occurred in either the treatment or control group and therefore a relative risk could not be calculated. The fifth trial by Burzotta and colleagues in 2005 provided a relative risk of 1.00 with a 95 percent confidence interval of 0.11 to 9.42. The combined effect of the four trials showed a relative risk of 3.18 with a 95 percent confidence interval of 0.73 to 13.88. The Cochran Q p-value was 0.807, the I-squared value was zero percent and Egger's p-value was 0.001.

Figure 15

Impact of catheter aspiration devices versus control on stroke using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.807 I2: 0 percent

The four trials which evaluated stroke using the maximal duration of followup are the same trials and data included in the analysis of ≤ 30 day stroke above14,15,71,83 because the maximal duration of followup for stroke in the four trials was ≤ 30 days. The use of a catheter aspiration device did not significantly impact the risk of in-hospital stroke [RR 4.94 (0.52, infinity)] versus control in a single trial and 30 days stroke occurrence in three others [RR 2.77 (0.51, 14.98)] (Appendix Figure 26). One trial evaluated the impact of catheter aspiration devices on 180-day stroke.74 In this trial, the use of the Pronto™ extraction catheter was compared to control. No stroke events occurred in either treatment arm, therefore a relative risk and risk difference could not be evaluated.

Two controlled observational studies evaluated the association between the use of catheter aspiration devices during PCI and 30-day stroke144,152 and 365-day stoke.144 In the first study, the Export® aspiration catheter was compared to control. There was no significant difference in 30-day or 365-day stroke (0 percent versus 0 percent, p=1.00, 0 percent versus 0.06 percent, p=0.10, respectively).144 In the second study, the name of the catheter aspiration device name was not reported.152 The use of a catheter aspiration device was associated with a significantly higher rate of 30-day stroke compared to control (1.3 percent versus 0.4 percent, p=0.03).

Catheter aspiration devices in other ACS populations. There were no trials or studies that evaluated catheter aspiration devices in other ACS populations.

Mechanical thrombectomy in patients with STEMI. Five RCTs evaluated the impact of mechanical thrombectomy devices versus control on stroke using the maximal duration of followup.11,27,29,40,44 One trial was excluded from the pooled analysis of relative risk because no strokes occurred within the prespecified time period in either treatment group.44 In the four trials eligible for pooling, the use of a mechanical thrombectomy device did not significantly impact the risk of stroke [RR 2.42 (0.75, 7.78)]11,27,29,40 (Figure 16). The weighted-mean followup for this analysis was 5.79 months. Statistical heterogeneity and publication bias were not detected (I2=0 percent, Egger's P=0.227). All of the pooled trials were determined to be of good methodological quality.11,27,29,40

This figure depicts the meta-analysis of the impact of mechanical thrombectomy devices versus control on stroke using the maximal duration of followup. The first trial by Migliorini and colleagues in 2010 provided a relative risk of 1.99 with 95 percent confidence interval of 0.26 to 15.14. The second trial by Ali and colleagues in 2006 provided a relative risk of 2.00 with 95 percent confidence interval of 0.43 to 9.28. The third trial by Lefevre and colleagues in 2005 provided a relative risk of 5.05 with 95 percent confidence interval of 0.53 to infinity. The fourth trial by Antonuicci and colleagues in 2004 provided a relative risk of 3.00 with 95 percent confidence interval of 0.26 to infinity. The fifth trial by Napodano and colleagues in 2003 was excluded from the meta-analysis because no strokes occurred in either the treatment or the control group and therefore a relative risk could not be calculated. The combined effect of the four trials showed a relative risk of 2.42 with a 95 percent confidence interval of 0.75 to 7.78. The Cochran Q p-value was 0.956, the I-squared value was zero percent and Egger's p-value was 0.227.

Figure 16

Impact of mechanical thrombectomy devices versus control on occurrence of stroke using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.956 I2: 0 percent

When the impact of mechanical thrombectomy versus control was assessed at ≤30 days [RR 1.89 (0.55, 6.48)], 30-days [same results as the ≤30 days analysis], 180-days [RR 2.05 (0.27, 15.78)], and 365-days [RR 1.99 (0.26, 15.14)] (Appendix Figures 27-28); no significant differences in the risk of stroke were seen in each analysis, although the 365-day analysis is based on a single trial.

One controlled observational study evaluated the association between the use of a mechanical thrombectomy device and in-hospital stroke.145 Patients undergoing PCI with a mechanical thrombectomy device, either the AngioJet® XMI or XVG catheter, were compared to patients undergoing PCI without mechanical thrombectomy. The use of a mechanical thrombectomy device was not associated with a significantly different rate of in-hospital stroke compared to PCI without a mechanical thrombectomy device (0.5 percent versus 0.4 percent, p=1.00).

Mechanical thrombectomy devices in other ACS populations. No trials or studies assessed the use of mechanical thrombectomy devices in other ACS populations.

Distal filter embolic protection devices in patients with STEMI. One RCT evaluated the impact of a distal filter embolic protection device versus control on the occurrence of stroke using the maximal duration of followup.89 In this trial, the use of a distal filter embolic protection device did not significantly impact the risk of long-term occurrence of stroke [RR 1.51 (95 percent CI=0.30 to 7.52)]. The duration of followup for stroke was 1 month. The trial was determined to be of good methodological quality.89

Distal filter embolic protection devices in other ACS populations. One RCT evaluated the impact of the distal filter embolic protection device FilterWire versus control on stroke126 in patients with NSTEMI or STEMI. The risk of 30-day stroke could not be calculated because no events occurred in either group during the specified time period.

No controlled observational studies assessed for this endpoint.

Distal balloon embolic protection devices in patients with STEMI. One RCT evaluated the impact of distal balloon embolic protection devices versus control on stroke using the maximal duration of followup.112 In this trial, the use of the GuardWire™ Plus was compared to control therapy. The use of a distal balloon embolic protection device did not significantly impact the risk of stroke at 180 days [RR 0.48 (0.10, 2.22)].112 The impact of distal balloon embolic protection devices on stroke was also evaluated in this trial at 30 days.112 The use of a distal balloon embolic protection device significantly decreased the risk of ≤30-day stroke [RR 0.11 (0.00, 0.94)] and 30-day stroke [same results as the ≤30 day analysis] versus control. This trial was determined to be of good methodological quality.

No controlled observational studies assessed for this endpoint.

Distal balloon embolic protection devices in other ACS populations. No trials or studies were available that evaluated the impact of distal balloon embolic protection devices versus control on stroke in this population.

Proximal balloon embolic protection devices in patients with STEMI. One RCT evaluated the impact of the proximal balloon embolic protection device Proxis™ versus control on stroke.18,141 The use of a proximal balloon embolic protection device did not significantly impact the risk of having a stroke over 30 days [RR 0.34 (0.01, 3.81)] or 180 days [RR 0.20 (0.00, 1.92)].

Proximal balloon embolic protection devices in other ACS populations. No trials or studies were available that evaluated the impact of proximal balloon embolic protection devices versus control on stroke in this population.

Embolic protection devices combined in patients with STEMI. Three RCTs evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) versus control on the occurrence of stroke using the maximal duration of followup.18,89,112,141 In these trials, the use of embolic protection devices combined did not significantly impact the risk of stroke [RR 0.68 (0.22, 2.11)] (Figure 17). The weighted mean followup for stroke using the maximal duration of followup was 3.74 months. Statistical heterogeneity was not detected (I2=0 percent) and publication bias could not be calculated due to the number of studies available. All of the trials were determined to be of good methodological quality.18,89,112

This figure depicts the meta-analysis of the impact of embolic protection combined devices versus control on stroke using the maximal duration of followup. The first trial by Haeck and colleagues in 2009 provided a relative risk of 0.20 with 95 percent confidence interval of 0.00 to 1.93. The second trial by Kelbaeck and colleagues in 2008 provided a relative risk of 1.51 with 95 percent confidence interval of 0.30 to 7.52. The third trial by Stone and colleagues in 2005 provided a relative risk of 0.48 with 95 percent confidence interval of 0.10 to 2.22. The combined effect of the three trials showed a relative risk of 0.68 with a 95 percent confidence interval of 0.22 to 2.11. The Cochran Q p-value was 0.459, the I-squared value was zero percent and there were too few strata to calculate the Egger's p-value.

Figure 17

Impact of embolic protection devices combined versus control on stroke using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.459 I2: 0 percent

When the impact of embolic protection devices combined was assessed at ≤30 days [RR 0.56 (0.11, 2.84)] and 180-days [RR 0.39 (0.09, 1.71)] (Appendix Figures 29-30); no significant difference in the risk of stroke were seen versus control.

Embolic protection devices combined in other ACS populations. No trials or studies were available that evaluated the impact of any embolic protection device versus control on stroke in this population in addition to the one trial reported above, and therefore pooling was not possible.

Target Revascularization

Direct Comparative Trials

Catheter aspiration device versus catheter aspiration device in patients with STEMI. One direct comparative randomized trial evaluated the impact of the Diver™-Invatec catheter aspiration device versus the Export®-Medtronic catheter aspiration device on 365-day target revascularization.158 In this trial, the use of Diver™-Invatec did not significantly impact the risk of 365-day target revascularization [RR 1.44 (0.30, 7.00)] compared to Export®-Medtronic. In this trial, no events occurred in either group at 30-days.

Catheter aspiration device versus distal balloon embolic protection device in patients with STEMI. One direct comparative trial evaluated the impact of the catheter aspiration device Diver™ CE versus the distal balloon embolic protection device Guardwire™ Plus on 30-day target revascularization.160 In this trial, there was no difference in the risk of 30-day target revascularization [RR 1.00 (0.11, 9.45)].

Trials Versus Control

Catheter aspiration devices in patients with STEMI. Nine RCTs evaluated the impact of catheter aspiration devices versus control on target revascularization using the maximal duration of followup.12,14-16,19,49,62,64,68,74,82,83,138 In these trials, the use of catheter aspiration devices did not significantly impact the risk of target revascularization [RR 0.79 (0.61, 1.02)] (Figure 18). The weighted-mean followup for target revascularization using the maximal duration of followup was 9.48 months. Statistical heterogeneity and publication bias were not detected (I2=0 percent, Egger's P=0.548).

This figure depicts the meta-analysis of the impact of catheter aspiration devices versus control on target revascularization using the maximal duration of followup. The first trial by Dudek and colleagues in 2010 provided a relative risk of 1.92 with a confidence interval of 0.26 to 14.53. The second trial by Liistro and colleagues in 2009 provided a relative risk of 1.02 with 95 percent confidence interval of 0.29 to 3.56. The third trial by Sardella and colleagues, also in 2009, provided a relative risk of 0.79 with 95 percent confidence interval of 0.24 to 2.64. The fourth trial by Chao and colleagues in 2008 provided a relative risk of 0.75 with 95 percent confidence interval of 0.20 to 2.82. The fifth trial by Chevalier and colleagues, also in 2008, provided a relative risk of 2.15 with 95 percent confidence interval of 0.28 to 16.30. The sixth trial by Ikari and colleagues, also in 2008, provided a relative risk of 0.60 with a 95 percent confidence interval of 0.36 to 1.00. The seventh trial by Svilaas and colleagues, also in 2008, provided a relative risk of 0.87 with 95 percent confidence interval of 0.63 to 1.20. The eighth trial by Silva-Orrego and colleagues in 2006 provided a relative risk of 0.47 with 95 percent confidence interval of 0.06 to 3.54. The ninth trial by Burzotta and colleagues in 2005 provided a relative risk of 1.00 with a confidence interval of 0.11 to 9.42.The combined effect of the nine trials showed a relative risk of 0.81 with a 95 percent confidence interval of 0.62 to 1.04. The Cochran Q p-value was 0.936, the I-squared value was zero percent and Egger's p-value was 0.534.

Figure 18

Impact of catheter aspiration devices versus control on target revascularization using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.948 I2: 0 percent

When limiting the pooled analysis to only trials of good methodological quality12,14-16,49,62,64,68,74,82,83,138 the risk of target revascularization using the maximal duration of followup remained nonsignificant with the use of a catheter aspiration device group compared to control [RR 0.79 (0.61, 1.02)]. The weighted mean duration of followup was 9.48 months. Statistical heterogeneity was not detected (I2=0 percent).

When the impact of catheter aspiration devices versus control was assessed at 180 days [RR 0.61 (0.39, 0.94)] (Appendix Figure 31) a significant reduction in the risk of target revascularization versus control was seen. Using the risk difference for the analysis [RD −0.03 (-0.06, 0.002), (CER 0.01, 0.20)] 33 patients would need to be treated to prevent one target revascularization. However, at in-hospital [RR 1.35 (0.26, 6.94)], ≤30 days [RR 0.85 (0.53, 1.38)], 30 days [RR 0.82 (0.50, 1.35)] and 365 days [RR 0.87 (0.63, 1.19)] (Appendix Figures 32-35); no significant differences in the risk of target revascularization were seen versus control in each analysis.

Two controlled observational studies evaluated the association between the use of catheter aspiration devices during PCI and 30-day target revascularization144,152 and 365-day target revascularization.144 In the first study, the Export® aspiration catheter was compared to control. There was no significant difference in 30-day or 365-day target revascularization (2.4 percent versus 1.9 percent, p=0.936, 7.8 percent versus 7.1 percent, p=0.923 , respectively).144 In the second study, the name of the catheter aspiration device name was not reported.152 The use of a catheter aspiration device was not associated with a significantly different rate of 30-day target revascularization compared to control (1.9 percent versus 2.5 percent, p=0.46).

Catheter aspiration devices in other ACS populations. No trials or studies assessed target revascularization in other ACS populations.

Mechanical thrombectomy in patients with STEMI. Five RCTs evaluated the impact of mechanical thrombectomy devices versus control on target revascularization using the maximal duration of followup.11,27,29,40,44 Two trials were excluded from the pooled analysis of relative risk because no target revascularizations occurred within the prespecified time period in either treatment group.27,44 In the three trials eligible for pooling, the use of a mechanical thrombectomy device did not significantly impact the risk of target revascularization [RR 0.87 (0.36, 2.10)]11,29,40 (Figure 19). The weighted-mean followup for target revascularization using the maximal duration of followup was 6.22 months. A lower level of statistical heterogeneity was detected (I2=39.2 percent) and publication bias could not be evaluated. All of the pooled trials were determined to be of good methodological quality.11,29,40

This figure depicts the meta-analysis of the impact of mechanical thrombectomy devices versus control on target revascularization using the maximal duration of followup. The first trial by Migliorini and colleagues in 2010 provided a relative risk of 0.68 with 95 percent confidence interval of 0.41 to 1.13. The second trial by Ali and colleagues in 2006 provided a relative risk of 5.00 with 95 percent confidence interval of 0.78 to 32.16. The third trial by Lefevre and colleagues in 2005 provided a relative risk of 0.61 with 95 percent confidence interval of 0.16 to 2.24. The fourth trial by Antonuicci and colleagues in 2004 and the fifth trial by Napodano and colleagues in 2003 were excluded from the meta-analysis because no target revascularizations occurred in either the treatment or control group and therefore a relative risk could not be calculated. The combined effect of the three trials showed a relative risk of 0.87 with a 95 percent confidence interval of 0.36 to 2.10. The Cochran Q p-value was 0.193, the I-squared value was 39.2 percent and there were too few strata to calculate the Egger's p-value.

Figure 19

Impact of mechanical thrombectomy devices versus control on target revascularization using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.193 I2: 39.2 percent

When the impact of mechanical thrombectomy devices versus control was assessed at ≤30 days [RR 1.62 (0.21, 12.55)], 30 days [same results as the ≤30 days analysis], 180-days [RR 0.55 (0.33, 0.92)], and 365-days [RR 0.68 (0.41, 1.13)] (Appendix Figures 36-37); no significant differences in the risk of target revascularization were seen versus control in each analysis. The 365-day analysis is based on a single trial.

One controlled observational study evaluated the association between the use of a mechanical thrombectomy device and in-hospital target revascularization.145

Patients undergoing PCI with a mechanical thrombectomy device, either the AngioJet® XMI or XVG catheter, were compared to patients undergoing PCI without mechanical thrombectomy. The use of a mechanical thrombectomy device was not associated with a significant difference in the rate of in-hospital target revascularization compared to PCI without a mechanical thrombectomy device (2.7 percent versus 2.1 percent, p=0.57).

Mechanical thrombectomy devices in other ACS populations. One RCT evaluated the impact of the X-Sizer® mechanical thrombectomy device versus control on 30-day target revascularization in patients with STEMI or UA.166. The use of a mechanical thrombectomy device did not significantly impact the risk of 30-day target revascularization [RR 0.33 (0.00, 3.75)] compared to control.

One controlled observational study evaluated the association between the use of mechanical thrombectomy devices and 180-day target revascularization.153 The types of ACSs included in this study were not reported. Patients undergoing PCI with the mechanical thrombectomy device AngioJet® were compared to patients undergoing PCI without mechanical thrombectomy and target revascularization was evaluated at 270 days. The use of a mechanical thrombectomy device was not associated with a significantly different rate of 180-day target revascularization compared to PCI without a mechanical thrombectomy device (5.5 percent versus 4.8 percent, p=0.72).

Distal filter embolic protection devices in patients with STEMI. Three RCTs evaluated the impact of distal filter embolic protection devices versus control on target revascularization using the maximal duration of followup.89,94,95,137 One trial was excluded from the analysis because no events occurred in the groups compared.137 In the two remaining trials, the use of distal filter embolic protection devices significantly increased the risk of target revascularization using the maximal duration of followup [RR 1.61 (1.03, 2.54)] (Figure 20). The weighted-mean followup for target revascularization using the maximal duration of followup was 13.36 months. Using the risk difference [RD 0.04 (-0.0006, 0.08), (CER 0 to 0.09)], one case of target revascularization would occur with the use of a distal filter embolic protection device in 25 cases. All three trials were determined to be of good methodological quality.89,94,95,137

This figure depicts the meta-analysis of the impact of distal filter embolic protection devices versus control on target revascularization using the maximal duration of followup. The first trial by Ito and colleagues in 2010 was excluded from the analysis becasue no events occurred in teh groups compared. The second trial by Kelbaeck and colleagues in 2008 provided a relative risk of 1.78 with 95 percent confidence interval of 1.09 to 2.93. The third trial by Cura and colleagues in 2007 provided a relative risk of 1.00 with 95 percent confidence interval of 0.35 to 2.82. The combined effect of the two trials showed a relative risk of 1.61 with a 95 percent confidence interval of 1.03 to 2.54. The Cochran Q p-value was 0.341 and there were too few strata to calculate I-squared value and Egger's p-value.

Figure 20

Impact of distal filter embolic protection devices versus control on target revascularization using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.341 I2: Too few strata

Target revascularization at 365-days was significantly increased with the use of mechanical thrombectomy devices versus control [RR 1.78 (1.09, 2.93)] although this was based on a single trial. Using the risk difference [RD 0.01 (-0.005, 0.03), (CER 0.07] one case of target revascularization would occur with the use of a distal filter embolic protection device in 100 cases. Target revascularization at ≤30 days [RR 3.02 (0.61, 14.84)], 30-days [RR 3.02 (0.70, 13.01)], and 180-days [RR 1.00 (0.35, 2.82)] was not significantly impacted although each analysis is based on a single trial.

No controlled observational studies were available that assessed for this endpoint.

Distal filter embolic protection devices in other ACS populations. Two RCTs evaluated the impact of distal filter embolic protection devices versus control on target revascularization in patients with other ACSs using the maximal duration of followup.126,156 These trials were not suitable for pooling because the first trial evaluated patients with either NSTEMI or STEMI126 and the second trial evaluated patients with UA.156 Both trials evaluated target revascularization at 30-days although the risk could not be calculated because no events occurred in either trial during the specified time period.126,156

Distal balloon embolic protection devices in patients with STEMI. Five RCTs evaluated the impact of distal balloon embolic protection devices versus control on target revascularization using the maximal duration of followup.17,103,107,110,112,133 The use of a distal balloon embolic protection device did not significantly impact the risk of target revascularization [RR 0.93 (0.61, 1.42)] (Figure 21). The weighted-mean followup for target revascularization using the maximal duration of followup was 6 months. Statistical heterogeneity and publication bias were not detected (I2=0 percent, Egger's P=0.369). All of the trials were determined to be of good methodological quality. 17,103,107,110,112,133

This figure depicts the meta-analysis of the impact of distal balloon embolic protection devices versus control on target revascularization using the maximal duration of followup. The first trial by Tahk and colleagues in 2008 provided a relative risk of 1.44 with 95 percent confidence interval of 0.30 to 7.04. The second trial by Hahn and colleagues in 2007 provided a relative risk of 0.21 with 95 percent confidence interval of 0.00 to 1.89. The third trial by Matsuo and colleagues, also in 2007, provided a relative risk of 0.51 with 95 percent confidence interval of 0.19 to 1.39. The fourth trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 1.03 with 95 percent confidence interval of 0.55 to 1.95. The fifth trial by Stone and colleagues in 2005 provided a relative risk of 1.11 with 95 percent confidence interval of 0.55 to 2.24. The combined effect of the five trials showed a relative risk of 0.93 with a 95 percent confidence interval of 0.61 to 1.42. The Cochran Q p-value was 0.597, the I-squared value was zero percent and the Egger's p-value was 0.369.

Figure 21

Impact of distal balloon embolic protection devices versus control on target revascularization using maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.597 I2: 0 percent

When the impact of distal balloon embolic protection devices versus control was assessed at in-hospital [RR 0.32 (0, 3.71)].≤30 days [RR 1.38 (0.55, 3.50)], 30 days [same results as the ≤30 days analysis], and 180 days [RR 0.93 (0.61, 1.42)] (Appendix Figures 38-39); no significant differences in the risk of target revascularization were seen versus control in each analysis, although the in-hospital analysis is based on a single trial.

No controlled observational studies assessed for this outcome.

Distal balloon embolic protection devices in other ACS populations. One RCT evaluated the impact of the distal balloon embolic protection device PercuSurge versus abciximab therapy on target revascularization in patients with acute myocardial infarction.164 The use of a distal balloon embolic protection device did not significantly impact the risk of 180-day target revascularization [RR 1.11 (0.46, 2.67)] compared to abciximab therapy.

No controlled observational studies assessed for this outcome in this population.

Proximal balloon embolic protection devices in patients with STEMI. One RCT evaluated the impact of the proximal balloon embolic protection device Proxis™ versus control on target revascularization.18,141 The use of a proximal balloon embolic protection device did not significantly impact the risk of target revascularization over 30 days [RR 0.51 (0.14, 1.81)] or 180 days [RR 0.71 (0.29, 1.75)].

Proximal balloon embolic protection devices in other ACS populations. No trials or studies were available that evaluated the impact of proximal balloon embolic protection devices versus control on target revascularization in the population.

Embolic protection devices combined in patients with STEMI. Nine RCTs evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) versus control on target revascularization using the maximal duration of followup.17,18,89,95,103,107,112,133,137 One trial was excluded from the analysis because no events occurred in the groups compared.137 In the eight remaining trials, the use of embolic protection devices combined did not significantly impact the risk of long-term occurrence of target revascularization [RR 1.11 (0.80, 1.52)] (Figure 22). The weighted mean followup for target revascularization using the maximal duration of followup was 8.60 months. A lower level of statistical heterogeneity was detected as was a trend towards publication bias (I2=10 percent, Egger's P=0.066). All of the trials were determined to be of good methodological quality.17,18,89,95,103,107,112,133,137

This figure depicts the meta-analysis of the impact of embolic protection devices combined versus control on target revascularization using the maximal duration of followup. The first trial by Ito and colleagues was excluded from the analysis becasue no events occurred in the groups compared. The second trial by Haeck and colleagues in 2009 provided a relative risk of 0.71 with 95 percent confidence interval of 0.29 to 1.75. The third trial by Kelbaeck and colleagues in 2008 provided a relative risk of 1.78 with 95 percent confidence interval of 1.09 to 2.93. The fourth trial by Tahk and colleagues, also in 2008, provided a relative risk of 1.44 with 95 percent confidence interval of 0.30 to 7.04. The fifth trial by Cura and colleagues in 2007 provided a relative risk of 1.00 with 95 percent confidence interval of 0.35 to 2.82. The sixth trial by Hahn and colleagues, also in 2007, provided a relative risk of 0.21 with 95 percent confidence interval of 0.00 to 1.89. The seventh trial by Matsuo and colleagues, also in 2007, provided a relative risk of 0.51 with 95 percent confidence interval of 0.19 to 1.39. The eight trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 1.03 with 95 percent confidence interval of 0.55 to 2.24. The ninth trial by Stone and colleagues in 2005 provided a relative risk of 1.11 with 95 percent confidence interval of 0.55 to 2.24. The combined effect of the nine trials showed a relative risk of 1.11 with a 95 percent confidence interval of 0.80 to 1.52. The Cochran Q p-value was 0.353, the I-squared value was zero percent and the Egger's p-value was 0.066.

Figure 22

Impact of embolic protection devices combined versus control on target revascularization using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.353 I2: 10 percent

When the impact of embolic protection devices combined versus control was assessed at 365-days the risk of target revascularization was significantly increased with the use of embolic protection devices versus control [RR 1.78 (1.09, 2.93)] although this was based on a single trial. Using the risk difference [RD 0.05 (0.009, 0.10), (CER 0.07)] one case of target revascularization would occur for every 25 patients who undergo surgery with an embolic protection device. At in-hospital [RR 0.32 (0.00 to 3.71)], <30 days [RR 1.24 (0.62, 2.48)] 30 days [same results as the <30 days analysis] and 180 days [RR 0.90 (0.63, 1.30)], (Appendix Figures 40-41) no significant differences in risk of target revascularization were seen versus control, although the in-hospital analysis is based on a single trial.

No controlled observational studies assessed for this endpoint.

Embolic protection devices combined in other ACS populations. No trials or studies were available that evaluated the impact of any embolic protection device versus control on target revascularization in addition to the 3 trials reported above. Pooling was not suitable because each trial evaluated a different ACS.

Combined MACE

MACE was reported as a composite outcome in trials and the definition used in each trial corresponding to the extracted data can be found in Appendix Tables 87-98. Overall, the definitions of MACE within each analysis were found to be similar and appropriate for meta-analysis.

Direct Comparative Trials

Catheter aspiration device versus catheter aspiration device in patients with STEMI. One direct comparative randomized trial evaluated the impact of the Diver™-Invatec catheter aspiration device versus the Export®-Medtronic catheter aspiration device on 365-day MACE.158 In this trial, the use of Diver™-Invatec did not significantly impact the risk of 365-day MACE [RR 2.40 (0.57, 10.41)] compared to Export®-Medtronic. This same trial evaluated the impact of the Diver™-Invatec versus the Export®-Medtronic device on 30-day MACE.158 The use of Diver™-Invatec did not significantly impact the risk of 30-day MACE [RR 0.65 (0.13, 3.16)] compared to Export®-Medtronic.

Catheter aspiration device versus distal balloon embolic protection device in patients with STEMI. One direct comparative randomized trial evaluated the impact of the Diver™ CE catheter aspiration device versus the Guardwire™ Plus distal balloon embolic protection device on 30-day MACE.160 In this trial, the use of Diver™ CE did not significantly impact the risk of 30-day MACE [RR 1.33 (0.35, 5.16)] compared to Guardwire™ Plus.

Trials Versus Control

Catheter aspiration devices in patients with STEMI. Eleven RCTs evaluated the impact of catheter aspiration devices versus control on MACE of maximal duration of followup. 12,14-16,19,49,54,62,64,68,69,71,83,85,138 In these trials, the use of a catheter aspiration device significantly reduced the occurrence of MACE using the maximal duration of followup [RR 0.73 (0.61, 0.88)] (Figure 23). The weighted-mean followup for MACE using the maximal duration of followup was 12.43 months. Statistical heterogeneity and publication bias were not detected (I2=0 percent, Egger's P=0.965). Given the risk difference [RD -0.03 (-0.01, 0.001), CER (0.02 to 0.35)], 33 people would need to be treated with a catheter aspiration device to prevent one MACE.

This figure depicts the meta-analysis of the impact of catheter aspiration devices versus control on MACE using the maximal duration of followup. The first trial by Dudek and colleagues in 2010 provided a relative risk of 0.80 with 95 percent confidence interval of 0.27 to 2.40. The second trial by Liistro and colleagues in 2009 provided a relative risk of 1.16 with 95 percent confidence interval of 0.47 to 2.91. The third trial by Sardella and colleagues, also in 2009, provided a relative risk of 0.33 with 95 percent confidence interval of 0.12 to 0.93. The fourth trial by Chao and colleagues in 2008 provided a relative risk of 0.50 with 95 percent confidence interval of 0.19 to 3.05. The fifth trial by Chevalier and colleagues, also in 2008, provided a relative risk of 1.25 with 95 percent confidence interval of 0.45 to 3.47. The sixth trial by Ikari and colleagues, also in 2008, provided a relative risk of 0.56 with 95 percent confidence interval of 0.40 to 0.80. The seventh trial by Svilaas and colleagues, also in 2008, provided a relative risk of 0.82 with 95 percent confidence interval of 0.64 to 1.05. The eighth trial by De Luca and colleagues in 2006 provided a relative risk of 0.81 with 95 percent confidence interval of 0.21 to 3.05. The ninth trial by Kaltoft and colleagues, also in 2006, provided a relative risk of 0.99 with 95 percent confidence interval of 0.18 to 5.54. The tenth trial by Burzotta and colleagues in 2005 provided a relative risk of 1.00 with 95 percent confidence interval of 0.33 to 3.91. The eleventh trial by Noel and colleagues, also in 2005, provided a relative risk of 0.54 with 95 percent confidence interval of 0.07o 3.91. The combined effect of the eleven trials showed a relative risk of 0.73 with a 95 percent confidence interval of 0.61 to 0.88. The Cochran Q p-value was 0.545, the I-squared value was zero percent and the Egger's p-value was 0.965.

Figure 23

Impact of catheter aspiration devices versus control on MACE using maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.645 I2: 0 percent

When limiting the pooled analysis to only trials of good methodological quality12,14-16,19,49,54,62,68,69,71,83,138 the risk of MACE using the maximal duration of followup remained significantly reduced in the catheter aspiration device group compared to control [RR 0.73 (0.61, 0.88)]. The weighted mean duration of followup was 12.66 months. Statistical heterogeneity was not detected (I2=0 percent). Given the risk difference [RD -0.03 (-0.07, 0.003), (CER 0.02 to 0.35)], 34 people would need to be treated with a catheter aspiration device to prevent one MACE.

When the impact of catheter aspiration devices versus control was assessed at in-hospital [RR 0.97 (0.36, 2.58)], ≤30 days [RR 0.80 (0.57, 1.12)], 30 days [RR 0.79 (0.56, 1.13)], and 365 days [RR 0.61 (0.26, 1.41)] (Appendix Figures 42-45); no significant differences in the risk of MACE were seen versus control in each analysis while a significant decrease in risk at 180 days [RR 0.66 (0.47, 0.94)] (Appendix Figure 46) was seen with the use of a catheter aspiration device versus control. Given the risk difference [RD -0.04 (-0.10, -0.003), (CER 0.06 to 0.27)], 25 people would need to be treated with a catheter aspiration device to prevent one MACE.

Two controlled observational studies evaluated the association between the use of catheter aspiration devices during PCI and 30-day MACE and 365-day MACE.144,152 In the first study, the Export® aspiration catheter was compared to control.144 The use of catheter aspiration was not associated with a significant difference in the rate of MACE at 30-days or 365-days versus control (8.5 percent versus 6.8 percent, p=0.47, 12.8 percent versus 14.1 percent, p=0.79, respectively).144 The catheter aspiration devices included in the second study was not reported.152

The use of a catheter aspiration device was not associated with a significant difference in the rate of 30-day MACE compared to control (5.5 percent versus 5.3 percent, p=0.81). The use of a catheter aspiration device was not associated with a significant difference in the rate of 30-day MACE [HR 0.96 (0.56, 1.52)] or 365-day MACE [HR 1.03 [0.68, 1.55)].

Catheter aspiration devices in other ACS populations. No trials or studies assessed for this endpoint in this population.

Mechanical thrombectomy in patients with STEMI. Four RCTs evaluated the impact of mechanical thrombectomy devices versus control on MACEs using the maximal duration of followup.11,27,29,40 One trial was excluded from the pooled analysis of relative risk because there were no MACE at the prespecified time-point in either treatment groups.27 In the three trials eligible for pooling, the use of a mechanical thrombectomy device did not significantly impact the risk of MACE using the maximal duration of followup [RR 1.23 (0.50, 3.01)]11,29,40 (Figure 24). The weighted mean followup for MACE was 6.22 months. A higher level of statistical heterogeneity was found (I2=79.9 percent) and publication bias could not be evaluated. The three pooled trials were all determined to be of good methodological quality.11,29,40

This figure depicts the meta-analysis of the impact of mechanical thrombectomy devices versus control on MACE using the maximal duration of followup. The first trial by Migliorini and colleagues in 2010 provided a relative risk of 0.66 with 95 percent confidence interval of 0.44 to 0.97. The second trial by Ali and colleagues in 2006 provided a relative risk of 1.43 with 95 percent confidence interval of 1.43 to 11.29. The third trial by Lefevre and colleagues in 2005 provided a relative risk of 1.01 with 95 percent confidence interval of 0.50 to 2.04. The fourth trial by Antonuicci and colleagues in 2004 was excluded from the meta-analysis because no MACE occurred in either the treatment or the control group and therefore the relative risk could not be calculated. The combined effect of the three trials showed a relative risk of 1.23 with a 95 percent confidence interval of 0.50 to 3.01. The Cochran Q p-value was 0.007, the I-squared value was 79.9 percent and there were too few strata to calculate the Egger's p-value.

Figure 24

Impact of mechanical thrombectomy devices versus control on MACE using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.007 I2: 79.9 percent

When the impact of mechanical thrombectomy devices versus control was assessed at ≤30 days [RR 1.28 (0.37, 4.38)], 30 days [same results as the ≤30 days analysis] and 180-days [RR 0.71 (0.41, 1.20)] (Appendix Figures 47-48), no significant difference in the risk of MACE were seen versus control. One trial evaluated the impact of mechanical thrombectomy devices on 365-day MACE versus control.11 In this trial, the use of the AngioJet® rheolytic thrombectomy system was compared to control therapy and significantly decreased the risk of 365-day MACE [RR 0.66 (0.44, 0.97)] versus control. Given the risk difference for 365-day MACE [RD -0.10 (-0.15, -0.01), (CER 0.23)], 10 people would need to be treated with a catheter thrombectomy device in order to prevent one occurrence of MACE.

One controlled observational study evaluated the association between the use of a mechanical thrombectomy device and in-hospital MACE.145 Patients undergoing PCI with a mechanical thrombectomy device, either the AngioJet® XMI or XVG catheter, were compared to patients undergoing PCI without mechanical thrombectomy. The use of a mechanical thrombectomy device was not associated with a significant difference in the rate of in-hospital MACE compared to PCI without a mechanical thrombectomy device (7.5 percent versus 9.0 percent, p=0.47) and remained nonsignificant after adjustment for baseline and angiographic characteristics [OR 0.83 (0.48, 1.42)].

Mechanical thrombectomy devices in other ACS populations. One RCT evaluated the impact of the mechanical thrombectomy device X-Sizer® versus control on 30-day MACE in patients with STEMI or UA.166 The risk of 30-day MACE was not significantly different between the mechanical thrombectomy device group and control [RR 1.00 (0.18, 5.43)].

One controlled observational study evaluated the association between the use of mechanical thrombectomy devices and 180-day MACE.153 The types of ACSs included in this study were not reported. Patients undergoing PCI with the mechanical thrombectomy device AngioJet® were compared to patients undergoing PCI without mechanical thrombectomy and MACE was evaluated at 270 days. The use of a mechanical thrombectomy device was not associated with a significant difference in the rate of 180-day MACE compared to PCI without a mechanical thrombectomy device (14.0 percent versus 11.6 percent, p=0.35).

Distal filter embolic protection devices in patients with STEMI. Five RCTs evaluated the impact of distal filter embolic protection devices versus control on the occurrence of MACE using the maximal duration of followup.89,95,98,101,137 In these trials, the use of distal filter embolic protection devices did not significantly impact the risk of MACE using the maximal duration of followup [RR 1.34 (0.97, 1.86)] (Figure 25). The weighted-mean followup for MACE was 10.84 months. Statistical heterogeneity and publication bias were not detected (I2=0 percent, Egger's P=0.419).

This figure depicts the meta-analysis of the impact of distal filter embolic protection devices versus control on MACE using the maximal duration of followup. The first trial by Ito and colleagues in 2010 provided a relative risk of 0.30 with a 95 percent confidence interval of 0.00 to 3.30. The second trial by Kelbaeck and colleagues in 2008 provided a relative risk of 1.48 with 95 percent confidence interval of 1.03 to 2.15. The third trial by Cura and colleagues in 2007 provided a relative risk of 0.91 with 95 percent confidence interval of 0.42 to 1.96. The fourth trial by Guetta and colleagues, also in 2007, provided a relative risk of 2.88 with 95 percent confidence interval of 0.43 to 19.79. The fifth trial by Lefevre and colleagues in 2004 provided a relative risk of 0.88 with 95 percent confidence interval of 0.16 to 4.74. The combined effect of the five trials showed a relative risk of 1.34 with a 95 percent confidence interval of 0.97 to 1.86. The Cochran Q p-value was 0.601, the I-squared value was zero percent and the Egger's p-value was 0.419.

Figure 25

Impact of distal filter embolic protection devices versus control on MACE using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.601 I2: 0 percent

When limiting the pooled analysis to only trials of good methodological quality89,95,98,137 the risk of MACE remained nonsignificant with distal filter embolic protection devices compared to control [RR 1.36 (0.98, 1.89)]. The weighted mean duration of followup was 11.49 months. Statistical heterogeneity was not detected (I2=0 percent).

When the impact of distal filter embolic protection devices versus control was assessed at 365-days, there was a significant increase in the risk of MACE [RR 1.48 (1.03, 2.15)] although this was based on a single trial. Using the risk difference [RD 0.06 (0.004, 0.12), (CER 0.13)], one case of MACE would occur with the use of a distal filter embolic protection device in 17 cases. MACE at ≤30 days [RR 1.29 (0.77, 2.15)], 30 days [same results as the ≤30 day analysis], and 180 days [RR 1.10 (0.68, 1.78)] (Appendix Figures 49-50) was not significantly different versus control in each analysis.

No controlled observational studies were available that assessed for this endpoint.

Distal filter embolic protection devices in other ACS populations. Two RCTs evaluated the impact of distal filter embolic protection devices versus control in patients with other ACSs on MACE using the maximal duration of followup.126,155 These trials were not suitable for pooling because the first trial evaluated patients with either NSTEMI or UA155 and the second trial evaluated patients with either STEMI or NSTEMI.126 In the trial evaluating patients with NSTEMI or UA,155 the FilterWire EZ™ device was compared to control. The use of a distal filter embolic protection device was not associated with a significant impact on the risk of MACE at in-hospital [RR 1.24 (0.50, 3.06)] and at 30-days [RR 1.08 (0.45, 2.59)] compared to control.155 In the trial evaluating patients with either STEMI or NSTEMI,126 the FilterWire EX™ device was compared to control. The use of a distal filter embolic protection device did not significantly impact the risk of 180-day MACE [RR 1.08 (0.53, 2.23)] compared to control.

Distal balloon embolic protection devices in patients with STEMI. Six RCTs evaluated the impact of distal balloon embolic protection devices versus control on MACE using the maximal duration of followup.17,103,107,111,112,133 One study was excluded from the pooled analysis of relative risk because there were no MACE at the prespecified time point in either treatment group.111 In the five studies eligible for pooling, the use of a distal embolic protection device did not significantly impact the risk of MACE [RR 0.87 (0.64, 1.19)]17,103,107,112,133 (Figure 26). The weighted-mean followup for MACE was 6 months. Statistical heterogeneity was not detected (I2=0 percent) but publication bias was detected (Egger's P=0.032). All of the trials were determined to be of good methodological quality.17,103,107,111,112,133

This figure depicts the meta-analysis of the impact of distal balloon embolic protection devices versus control on MACE using the maximal duration of followup. The first trial by Tahk and colleagues in 2008 provided a relative risk of 0.77 with 95 percent confidence interval of 0.23 to 2.52. The second trial by Hahn and colleagues in 2007 provided a relative risk of 0.12 with 95 percent confidence interval of 0.00 to 0.91. The third trial by Matsuo and colleagues, also in 2007, provided a relative risk of 0.77 with 95 percent confidence interval of 0.36 to 1.65. The fourth trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 0.97 with 95 percent confidence interval of 0.60 to 1.56. The fifth trial by Zhou and colleagues, also in 2007, was excluded from the meta-analysis because no MACE occurred in either the treatment or control group and therefore a relative risk could not be calculated. The sixth trial by Stone and colleagues in 2005 provided a relative risk of 0.89 with 95 percent confidence interval of 0.53 to 1.49. The combined effect of the five trials showed a relative risk of 0.87 with a 95 percent confidence interval of 0.64 to 1.19. The Cochran Q p-value was 0.685, the I-squared value was zero percent and the Egger's p-value was 0.032

Figure 26

Impact of distal balloon embolic protection devices versus control on MACE using the maximal duration of followup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.685 I2: 0 percent

When the impact of distal embolic protection devices was assessed at ≤30 days [RR 0.74 (0.44, 1.23)], 30 days [same results as the ≤30 day analysis], and 180 days [RR 0.87 (0.64, 1.19)] (Appendix Figures 51-52); no significant differences in the risk of MACE were seen versus control in each analysis.

No controlled observational studies assessed for this endpoint.

Distal balloon embolic protection devices in other ACS population. One RCT evaluated the impact of the distal balloon embolic protection device GuardWire® PercuSurge versus control on MACE in patients with acute myocardial infarction.132 The use of a distal filter embolic protection device did not significantly impact the risk of 180-day MACE [RR 0.33 (0.05, 1.87)] compared to control.

No controlled observational studies evaluated this endpoint in this population.

Proximal balloon embolic protection devices in patients with STEMI. One RCT evaluated the impact of the proximal balloon embolic protection device Proxis™ versus control on MACE.18,141 The use of a proximal balloon embolic protection device did not significantly impact the risk of MACE over 30 days [RR 0.34 (0.01, 8.23)] or 180 days [RR 0.74 (0.36, 1.54)].

Proximal balloon embolic protection devices in other ACS populations. No trials or studies were available that evaluated the impact of proximal balloon embolic protection devices versus control on MACE in this population.

Embolic protection devices combined in patients with STEMI. Twelve RCTs evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) versus control on MACE using the maximal duration of followup.17,18,89,95,98,101,103,107,111,112,133,137 The trial by Zhou et al was excluded from the pooled analysis of relative risk because no events occurred within the prespecified time period in either control or treatment group. In the 11 trials suitable for pooling, the use of embolic protection devices combined did not significantly impact the risk of long-term occurrence of MACE [RR 1.04 (0.84, 1.29)] (Figure 27). The weighted mean followup for MACE using the maximal duration of followup was 7.97 months. Statistical heterogeneity and publication bias were not detected (I2=0 percent, Egger's P=0.0.084). The analysis was then limited to only trials of good methodological quality17,18,89,95,98,103,107,111,112,133,137 although one trial was excluded from the analysis because no events occurred in either group during the prespecified time period.111 In the ten trials of good methodological quality suitable for pooling, the risk of MACE remained nonsignificant in the combined embolic protection device group compared to control [RR 1.03 (0.82, 1.29)]. The weighted mean followup for MACE using the maximal duration of followup was 8.15 months. A lower level of statistical heterogeneity was detected (I2=4 percent).

This figure depicts the meta-analysis of the impact of embolic protection devices combined versus control on MACE using the maximal duration of followup. The first trial by Ito and colleagues provided a relative risk of 0.30 with a 95 percent confidence interval of 0.00 to 3.30. The second trial by Haeck and colleagues in 2009 provided a relative risk of 0.74 with 95 percent confidence interval of 0.36 to 1.54. The third trial by Kelbaeck and colleagues in 2008 provided a relative risk of 1.48 with 95 percent confidence interval of 1.03 to 2.15. The fourth trial by Tahk and colleagues, also in 2008, provided a relative risk of 0.77 with 95 percent confidence interval of 0.23 to 2.52. The fifth trial by Cura and colleagues in 2007 provided a relative risk of 0.91 with a 95 percent confidence interval of 0.42 to 1.96. The sixth trial by Guetta and colleagues, also in 2007, provided a relative risk of 2.88 with 95 percent confidence interval of 0.43 to 19.79. The seventh trial by Hahn and colleagues, also in 2007, provided a relative risk of 0.12 with 95 percent confidence interval of 0.00 to 0.91. The eight trial by Matsuo and colleagues, also in 2007, provided a relative risk of 0.77 with 95 percent confidence interval of 0.36 to 1.65. The ninth trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 0.97 with 95 percent confidence interval of 0.60 to 1.56. The tenth trial by Zhou and colleagues, also in 2007, was excluded from the meta-analysis because no MACE occurred in either the treatment or control group and therefore a relative risk could not be calculated. The eleventh trial by Stone and colleagues in 2005 provided a relative risk of 0.89 with 95 percent confidence interval of 0.53 to 1.49. The twelfth trial by Lefevre and colleagues in 2004 provided a relative risk of 0.88 with 95 percent confidence interval of 0.16 to 4.74. The combined effect of the twelve trials showed a relative risk of 1.04 with a 95 percent confidence interval of 0.84 to 1.29. The Cochran Q p-value was 0.494, the I-squared value was zero percent and the Egger's p-value was 0.084.

Figure 27

Impact of embolic protection devices combined versus control on MACE using the maximal duration of folloup in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.494 I2: 0 percent

When the impact of distal embolic protection devices versus control was assessed at 365-days, the risk of MACE was significantly increased with the use of embolic protection devices versus control [RR 1.48 (1.03, 2.14)] although this was based on a single trial. Using the risk difference [RD 0.06 (0.005, 0.12), (CER 0.13)] one case of MACE would occur for every 17 patients who undergo surgery with an embolic protection device. At ≤30 days [RR 0.92 (0.66, 1.30)], 30 days [same results as the ≤30 day analysis], and 180 days [RR 0.91 (0.71, 1.16)] (Appendix Figures 53-54), no significant differences in the risk of MACE were seen versus control in each analysis.

One controlled observational study evaluated the association between the use of a distal protection device and 365-day MACE in patients with STEMI.147 In this study, the device name was not reported nor was the distinction between distal balloon and distal filter. There was no significant difference in the adjusted rate of 365-day MACE when comparing the distal protection group with those who did not receive distal protection during PCI [HR 0.85 (0.59, 3.48)].

Embolic protection devices combined in other ACS populations. No trials or studies were available that evaluated the impact of any embolic protection device versus control on MACE in addition to the three trials reported above, and pooling was not suitable because each trial evaluated a different ACS.

Health-Related Quality of Life

Direct Comparative Trials

No direct comparative trials evaluated the impact of catheter aspiration, mechanical thrombectomy or embolic protection devices on this endpoint.

Trials Versus Control

Catheter aspiration devices. No trials or studies evaluated the impact of catheter aspiration devices on this endpoint.

Mechanical thrombectomy devices. No trials or studies evaluated the impact of catheter aspiration devices on this endpoint.

Distal Balloon Embolic Protection Devices. No trials or studies evaluated the impact of distal balloon embolic protection devices on this endpoint.

Proximal Balloon Embolic Protection Devices. No trials or studies evaluated the impact of proximal balloon embolic protection devices on this endpoint.

Embolic Protection Devices Combined. No trials or studies evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) on this endpoint.

ST-Segment Resolution

ST-segment resolution was defined in different ways in different trials. We defined ST-segment resolution as ≥70 percent resolution at 60 minutes if reported, ≥50 percent resolution at 60 minutes if ≥70 percent resolution at 60 minutes data was not reported, or ≥70 percent resolution postPCI or at 90 minutes if 60 minute data was unavailable.

Direct Comparative Trials

Catheter aspiration device versus catheter aspiration device in STEMI. One direct comparative randomized trial evaluated the impact of the Diver™-Invatec catheter aspiration device versus the Export®-Medtronic catheter aspiration device on ST-segment resolution.158 In this trial, ST-segment resolution was defined as resolution great than or equal to 70 percent at 90 minutes. The use of Diver™-Invatec did not significantly impact the risk of resolving ST-segment elevation [RR 0.79 (0.61, 1.00)] compared to Export®-Medtronic.

Catheter aspiration device versus distal balloon protection device in ACS. One direct comparative randomized trial evaluated the impact of the Diver™ CE catheter aspiration device versus the Guardwire™ Plus distal balloon embolic protection device on ST-segment resolution.160 In this trial, ST-segment resolution was defined as greater than or equal to 70 percent up to 6 hours postprocedure (measured immediately after the procedure and at 90 minutes and 6 hours postprocedure). The use of Diver™ CE did not significantly impact the risk of resolving ST-segment elevation [RR 0.97 (0.72, 1.32)] compared to Guardwire™ Plus.

Trials Versus Control

Catheter aspiration devices in patients with STEMI. Fifteen RCTs evaluated the impact of catheter aspiration devices versus control on ST-segment resolution and were included in the pooled analysis.12,14-16,19,20,62,69,71,74,83,85-87,87,138,176 The use of a catheter aspiration device significantly increased the risk of resolving ST-segment elevation versus control [RR 1.51 (1.32, 1.73)] (Figure 28). A higher level of statistical heterogeneity was found (I2=64.2) as was the presence of publication bias (Egger's P=0.041). Given the risk difference [RD 0.22 (0.15, 0.30), (CER 0.11 to 0.65)], five people would need to be treated with a catheter aspiration device to allow one person to experience ST-segment resolution.

This figure depicts the meta-analysis of the impact of catheter aspiration devices versus control on ST-segment resolution. The first trial by Dudek and colleagues in 2010 provided a relative risk of 1.53 with a 95 percent confidence interval of 1.11 to 2.14. The second trial by Liistro and colleagues in 2009 provided a relative risk of 1.80 with 95 percent confidence interval of 1.27 to 2.65. The third trial by Lipiecki and colleagues, also in 2009, provided a relative risk of 1.11 with 95 percent confidence interval of 0.62 to 1.97. The fourth trial by Moura and colleagues, also in 2009, provided a relative risk of 2.03 with 95 percent confidence interval of 1.58 to 2.71. The fifth trial by Sardella and colleagues, also in 2009, provided a relative risk of 2.10 with a 95 percent confidence interval of 1.60 to 2.84. The sixth trial by Chevalier and colleagues in 2008 provided a relative risk of 1.13 with 95 percent confidence interval of 0.95 to 1.34. The seventh trial by Ikari and colleagues, also in 2008, provided a relative risk of 1.21 with 95 percent confidence interval of 0.80 to 1.83. The eighth trial by Svilaas and colleagues, also in 2008, provided a relative risk of 1.28 with 95 percent confidence interval of 1.13 to 1.45. The ninth trial by De Luca and colleagues in 2006 provided a relative risk of 1.48 with a 95 percent confidence interval of 1.08 to 2.10. The tenth trial by Kaltoft and colleagues, also in 2006, provided a relative risk of 1.04 with 95 percent confidence interval of 0.72 to 1.50. The eleventh trial by Lee and colleagues, also in 2006, provided a relative risk of 1.64 with 95 percent confidence interval of 1.14 to 2.42. The twelfth trial by Silva-Orrego and colleagues, also in 2006, provided a relative risk of 1.35 with 95 percent confidence interval of 1.03 to 1.80. The thirteenth trial by Burzotta and colleagues in 2005 provided a relative risk of 1.72 with 95 percent confidence interval of 1.13 to 2.68. The fourteenth trial by Noel and colleagues, also in 2005, provided a relative risk of 4.33 with 95 percent confidence interval of 1.55 to 13.21. The fifteenth trial by Dudek and colleagues in 2004 provided a relative risk of 2.70 with a 95 percent confidence interval of 1.51 to 5.24. The combined effect of the fifteen trials showed a relative risk of 1.51 with a 95 percent confidence interval of 1.32 to 1.73. The Cochran Q p-value was less than 0.001, the I-squared value was 64.2 percent and the Egger's p-value was 0.041.

Figure 28

Impact of catheter aspiration devices versus control on ST-segment resolution in patients with ST-segment elevation myocardial infarction. Cochran Q: P < 0.001 I2: 64.2 percent

When limiting the pooled analysis to only trials of good methodological quality12,14-16,62,69,71,74,83,138,176 the risk of resolving ST-segment elevation remained significantly increased in the catheter aspiration device group compared to control [RR 1.39 (1.21, 1.61)]. A higher level of statistical heterogeneity was detected (I2=60.4 percent). Given the risk difference [RD 0.18 (0.10, 0.26), (CER 0.27 to 0.65)], six people would need to be treated with a catheter aspiration device to allow one person to experience ST-segment resolution.

One RCT evaluated the impact of the catheter aspiration device Diver™ CE versus control on ST-segment resolution although was not included in the pooled analysis. In this trial patients were only included in if they attained TIMI-3 blood flow postprocedure, therefore it was not included in the pooled analysis of ST-segment resolution. The use of a catheter aspiration device did not significantly impact the risk of resolving ST-segment elevation [RR 0.93 (0.52, 1.62)] compared to control.

One controlled observational study evaluated the association between the use of catheter aspiration devices during PCI and resolution of ST-segment elevation.152 The catheter aspiration devices included in this study were not reported. The use of a catheter aspiration device was not associated with significant difference in the rate of resolution of ST-segment elevation (48.2 percent versus 50.3 percent, p=0.51).

Catheter aspiration devices in other ACS populations. No trials or studies assessed for this endpoint in this population.

Mechanical thrombectomy devices in patients with STEMI. Five RCTs evaluated the impact of mechanical thrombectomy devices versus control on ST-segment resolution.11,27,29,40,44 The use of a mechanical thrombectomy device did not significantly impact the risk of resolving ST-segment elevation [RR 1.16 (0.99, 1.36)] (Figure 29). A higher level of statistical heterogeneity was found (I2=75.1 percent) but publication bias was not detected (Egger's P=0.402). All of the trials in the pooled analysis were determined to be of good methodological quality.11,27,29,40,44

This figure depicts the meta-analysis of the impact of mechanical thrombectomy devices versus control on ST-segment resolution. The first trial by Migliorini and colleagues in 2010 provided a relative risk of 1.09 with 95 percent confidence interval of 1.00 to 1.19. The second trial by Ali and colleagues in 2006 provided a relative risk of 0.88 with 95 percent confidence interval of 0.75 to 1.04. The third trial by Lefevre and colleagues in 2005 provided a relative risk of 1.29 with 95 percent confidence interval of 1.02 to 1.65. The fourth trial by Antonuicci and colleagues in 2004 provided a relative risk of 1.25 with a 95 percent confidence interval of 1.04 to 1.56. The fifth trial by Napodano and colleagues in 2003 provided a relative risk of 1.58 with 95 percent confidence interval of 1.19 to 2.21. The combined effect of the five trials showed a relative risk of 1.16 with a 95 percent confidence interval of 0.99 to 1.36. The Cochran Q p-value was 0.003, the I-squared value was 75.1 percent and the Egger's p-value was 0.402.

Figure 29

Impact of mechanical thrombectomy devices versus control on ST-segment resolution in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.003 I2: 75.1 percent

No controlled observational studies assessed for this endpoint.

Mechanical thrombectomy devices in other ACS populations. One RCT evaluated the impact of the mechanical thrombectomy device X-Sizer® versus control on ST-segment resolution in patients with STEMI or UA.166 ST-segment resolution was defined as resolution greater than 50 percent after the procedure. The use of a mechanical thrombectomy device significantly increased the risk of resolving ST-segment elevation [RR 1.58 (1.05, 2.57)] compared to control. Given the risk difference for ST-segment resolution [RD 0.30 (0.03, 0.54), (CER 0.52)], three people would need to be treated with a mechanical thrombectomy device in order to have one person experience ST-segment resolution. This trial was determined to be of good methodological quality.

No controlled observational studies assessed for this endpoint in this population.

Distal filter embolic protection devices in patients with STEMI. Five RCTs evaluated the impact of distal filter embolic protection devices versus control on ST-segment resolution.89,95,98,101,137 In these trials, the use of distal filter embolic protection devices did not significantly impact the risk of resolving ST-segment elevation [RR 1.05 (0.97, 1.15)] (Figure 30). Statistical heterogeneity and publication bias were not detected (I2=0 percent, Egger's P=0.279).hen limiting the pooled analysis to only trials of good methodological quality89,95,98,137 the risk of resolving of ST-segment elevation remained nonsignificant [RR 1.05 (0.96, 1.15)]. Statistical heterogeneity was not detected (I2=0 percent).

This figure depicts the meta-analysis of the impact of distal filter embolic protection devices versus control on ST-segment resolution. The first trial by Ito and colleagues in 2010 proviede a relative risk of 2.01 with a 95 percent confidence interval of 0.97 to 1.15. The second trial by Kelbaeck and colleagues in 2008 provided a relative risk of 1.05 with 95 percent confidence interval of 0.96 to 1.16. The third trial by Cura and colleagues in 2007 provided a relative risk of 1.02 with 95 percent confidence interval of 0.78 to 1.34. The fourth trial by Guetta and colleagues, also in 2007, provided a relative risk of 0.99 with 95 percent confidence interval of 0.74 to 1.33. The fifth trial by Lefevre and colleagues in 2004 provided a relative risk of 1.24 with a 95 percent confidence interval of 0.81 to 1.98. The combined effect of the five trials showed a relative risk of 1.05 with a 95 percent confidence interval of 0.97 to 1.15. The Cochran Q p-value was 0.651, the I-squared value was zero percent and the Egger's p-value was 0.279.

Figure 30

Impact of distal filter embolic protection devices versus control on ST-segment resolution in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.651 I2: 0 percent

No controlled observational studies assessed for this endpoint in this population.

Distal filter embolic protection devices in other ACS populations. No trials or studies assessed for this endpoint in this population.

Distal balloon embolic protection devices in patients with STEMI. Four RCTs evaluated the impact of distal balloon embolic protection devices versus control on ST-segment resolution.103,107,112,133 The use of a distal balloon embolic protection device did not significantly impact the risk of resolving ST-segment elevation [RR 1.08 (0.91, 1.29)] (Figure 31). A lower level of statistical heterogeneity was found (I2=41.2 percent) but publication bias was not detected (Egger's P=0.311). All of the trials were determined to be of good methodological quality.103,107,112,133

This figure depicts the meta-analysis of the impact of distal balloon embolic protection devices versus control on ST-segment resolution. The first trial by Hahn and colleagues in 2007 provided a relative risk of 1.87 with 95 percent confidence interval of 1.16 to 3.34. The second trial by Matsuo and colleagues, also in 2007, provided a relative risk of 0.97 with 95 percent confidence interval of 0.72 to 1.32. The third trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 1.07 with 95 percent confidence interval of 0.81 to 1.41. The fourth trial by Stone and colleagues in 2005 provided a relative risk of 1.02 with a 95 percent confidence interval of 0.89 to 1.18. The combined effect of the four trials showed a relative risk of 1.08 with a 95 percent confidence interval of 0.91 to 1.29. The Cochran Q p-value was 0.164, the I-squared value was 41.2 percent and the Egger's p-value was 0.311.

Figure 31

Impact of distal balloon embolic protection devices versus control on ST-segment resolution in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.164 I2: 41.2 percent

No controlled observational studies assessed for this endpoint in this population.

Distal balloon embolic protection devices in other ACS populations. One RCT evaluated the impact of the distal balloon embolic protection device Guardwire™ Plus versus control on early resolution of ST-segment elevation in patients with acute myocardial infarction.130 The use of a distal balloon embolic protection device significantly increased the risk of resolving ST-segment elevation compared to control [RR 1.58 (1.10, 2.46)]. Given the risk difference [RD 0.29 (0.10, 0.50), (CER 0.50)], three people would need to be treated with a distal balloon embolic protection device to have one patient experience an ST segment resolution. This trial was determined to be of poor methodological quality.130

One RCT evaluated the impact of the distal balloon embolic protection device PercuSurge versus abciximab therapy on ST-segment resolution in patients with acute myocardial infarction.164 ST-segment resolution was defined as ≥70 percent at 60 minutes. The use of a distal balloon embolic protection device did not significantly impact the risk of resolving ST-segment elevation [RR 1.28 (0.86, 1.92)] compared to abciximab therapy.

No controlled observational studies assessed for this endpoint in this population.

Proximal balloon embolic protection devices in patients with STEMI. One RCT evaluated the impact of the proximal balloon embolic protection device Proxis™ versus control on ST-segment resolution.18 The use of a proximal balloon embolic protection device did not significantly impact the risk of resolving ST-segment elevation [RR 1.11 (0.97, 1.28)]. The trial was determined to be of good methodological quality.18

Proximal balloon embolic protection devices in other ACS populations. No trials or studies assessed for this endpoint in this population.

Embolic protection devices combine in patients with STEMI. Ten RCTs evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) versus control on ST-segment resolution.18,89,95,98,101,103,107,112,133,137 In these trials, the use of embolic protection devices combined did not significantly impact the risk of resolving ST-segment elevation [RR 1.06 (1.00, 1.13)] (Figure 32). Statistical heterogeneity and publication bias were not detected (I2=0 percent, Egger's P=0.117).

This figure depicts the meta-analysis of the impact of embolic protection devices combined versus control on ST-segment resolution. The first trial by Ito and colleagues in 2010 provided a relative risk of 2.01 with a 95 percent confidence interval of 0.81 to 5.42. The second trial by Haeck and colleagues in 2009 provided a relative risk of 1.11 with 95 percent confidence interval of 0.97 to 1.28. The third trial by Kelbaeck and colleagues in 2008 provided a relative risk of 1.05 with 95 percent confidence interval of 0.96 to 1.16. The fourth trial by Cura and colleagues in 2007 provided a relative risk of 1.02 with 95 percent confidence interval of 0.78 to 1.34. The fifth trial by Guetta and colleagues, also in 2007, provided a relative risk of 0.99 with a 95 percent confidence interval of 0.74 to 1.33. The sixth trial by Hahn and colleagues, also in 2007, provided a relative risk of 1.87 with 95 percent confidence interval of 1.16 to 3.34. The seventh trial by Matsuo and colleagues, also in 2007, provided a relative risk of 0.97 with 95 percent confidence interval of 0.72 to 1.32. The eighth trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 1.07 with 95 percent confidence interval of 0.81 to 1.41. The ninth trial by Stone and colleagues in 2005 provided a relative risk of 1.02 with a 95 percent confidence interval of 0.89 to 1.18. The tenth trial by Lefevre and colleagues in 2004 provided a relative risk of 1.24 with 95 percent confidence interval of 0.81 to 1.98. The combined effect of the ten trials showed a relative risk of 1.06 with a 95 percent confidence interval of 1.00 to 1.13. The Cochran Q p-value was 0.534, the I-squared value was zero percent and the Egger's p-value was 0.117.

Figure 32

Impact of embolic protection devices combined versus control on ST-segment resolution in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.534 I2: 0 percent

When limiting the analysis to only trials of good methodological quality18,89,95,98,103,107,112,133,137 the risk of resolving ST-segment elevation remained nonsignificant in the combined embolic protection device group compared to control [RR 1.06 (1.00, 1.13)]. Statistical heterogeneity was not detected (I2=0 percent).

Embolic protection devices combine in other ACS populations. No trials or studies were available in addition to the two trials reported above that evaluated the impact of any embolic protection device versus control on ST-segment resolution in this patient population. Pooling was not suitable because a different comparator was used in each trial.

Ejection Fraction

Direct Comparative Trials

Catheter aspiration device versus distal balloon embolic protection device in patients with STEMI. One direct comparative randomized trial evaluated the impact of the Diver™ CE catheter aspiration device versus the Guardwire™ Plus distal balloon embolic protection device on left-ventricular ejection fraction (Table 13).160 There was no significant difference in the mean left-ventricular ejection fraction between Diver™ CE and Guardwire™ Plus groups at baseline (45 percent ±11 versus 46 percent ±10, p=0.56) or at 30 days postprocedure (54 percent ±12 versus 54 percent ±11, p=0.60), respectively.

Catheter aspiration device versus distal balloon embolic protection device versus control in patients with STEMI. One direct comparative randomized trial evaluated the impact of catheter aspiration devices and distal balloon embolic protection devices on 6-month ejection fraction (Table 13).163 In this trial, patients were randomized to one of three groups, catheter aspiration with Rescue™ or Thrombuster® devices, distal balloon embolic protection with PercuSurge or GuardWire devices, or to control therapy. Patients were excluded from the trial if they had coronary no reflow or slow flow. Ejection fraction at 180-days did not differ significantly amongst the three groups (50 percent ±8 versus 54 percent ±11 versus 52 percent ±12, p=NS).

Trials Versus Control

Catheter aspiration devices in patients with STEMI. Eleven RCTs evaluated the impact of catheter aspiration devices versus control on ejection fraction but were not amenable for statistical pooling therefore results are reported qualitatively (Table 14).12-14,16,21,53,68,69,71,82,83,87,162 In the first trial the mean left ventricular ejection fraction at baseline did not differ between the two groups (p=0.60).12 When baseline mean LVEF values were compared to mean LVEF at 6 months, a greater improvement was noted in the catheter aspiration group compared to control (48 percent ±6 to 55 percent ±6 versus 48 percent ±7 to 49 percent ±8, p<0.001), respectively. 12 In the second trial there was no significant difference in the left ventricular ejection fraction at 7 days between the catheter aspiration group and control (48 percent ±12 versus 45 ±11, p=0.04).13 In the third trial a subset of patients with anterior myocardial infarction from the original trial were randomized to evaluate ejection fraction.14 No difference in the mean ejection fraction was found at 3-5 days postprocedure (46.3 percent ±8.6 versus 44.3 percent ±9.5, p=0.06) or at 3 months (49.0 percent ±9.3 versus 46.7 percent ±10.6, p=0.30) between the catheter aspiration and control groups, respectively.14 In the fourth trial, patients were only included in the trial if they achieved a TIMI-3 blood flow postprocedure.162 In this trial, the mean left ventricular ejection fraction was not significantly different at 7 days postprocedure between the catheter aspiration device group and control (50.1 percent ±8.4 versus 46.5 percent ±7.9, p=NS). In the fifth trial there was no significant difference in the left ventricular ejection fraction at 5-8 days between the catheter aspiration device group and control (46.7 percent ±11 versus 42.5 percent ±10, p=0.16).21 In the sixth trial there was no significant difference between the catheter aspiration group and control in mean left ventricular ejection fraction at baseline (51.3 ±11.9 versus 51.3 ±11.9, p=0.99) or at 6 months (57.1 ±12.5 versus 56.7 ±12.3, p=0.77).16 In the seventh trial there was no significant difference in the mean left ventricular ejection fraction at 28 days between the catheter aspiration device group and control (56 percent ±10 versus 57 percent ±10, p=0.51).68 In the eighth trial the mean ejection fraction was reported in a figure and with use of Engauge Digitizer Version 2.0 to read the figure the values for ejection fraction were obtained.69 There was no significant difference between the catheter aspiration group and control in mean left ventricular ejection fraction immediately postprocedure (37.29 percent ±9.97 versus 36.67 percent ±3.03, p=NS) and at 6 months (42.97 percent ±9.97 versus 41.28 percent ±3.37, p=NS).69 In the ninth trial there was no significant difference in the median left ventricular ejection fraction at 30 days between the catheter aspiration device group and control (51 percent (43-57) versus 53 percent (47-58), p=0.13). In a substudy of 50 participants from the trial by Burzotta et al. ejection fraction was reported in a figure.83,84 Enguage Digitizer, Version 2.0 was used to read the figure and obtain values for ejection fraction. Mean ejection fraction was significantly greater in the catheter aspiration group compared to control at 24 hours (50.36 percent ±8.76 versus 45.75 percent ±7.49, p<0.05), 1 week (53.34 percent ±10.99 versus 48.09 percent ±9.4, p<0.05), and 6 months (53.28 percent ±10.04 versus 47.72 percent ±8.28, p<0.05). Mean ejection fraction at 1 week and at 6 months was significantly greater than mean ejection fraction at 24 hours in the catheter aspiration group (p<0.05).83 In the eleventh trial the mean left ventricular ejection fraction did not differ significantly between the catheter aspiration group and control in-hospital (56.5 percent ±9.1 versus 52.8 ±12.8, p=NS) or at 3 months (60.3 percent ±9.2 versus 55.3 percent ±14.7, p=NS).87

One controlled observational study evaluated the impact of catheter aspiration devices versus control on left ventricular ejection fraction.142 The use of a catheter aspiration device significantly decreased left ventricular ejection fraction versus control postPCI (49±11 versus 53±11, p<0.0005).

Catheter aspiration devices in other ACS populations. No trials or studies assessed for this endpoint in this population.

Mechanical thrombectomy in patients with STEMI. Two RCTs evaluated the impact of mechanical thrombectomy devices versus control on ejection fraction but were not amenable for statistical pooling therefore results are reported qualitatively (Table 15).40,44 In the first trial there was no significant difference in ejection fraction at 14 to 28 days postprocedure between the mechanical thrombectomy device group and control (51.3 percent ±11.53 versus 52.3 ±10.89, p=0.38).40 In the second trial the mean ejection fraction significantly improved in the mechanical thrombectomy device group (49.3 percent ±7.6 to 51.9 percent ±7.9, p=0.02) and in control (48.8 percent ±5.9 to 49.9 percent ±8.9, p=0.04) from baseline to 30 days.44 There was no significant difference in ejection fraction between the mechanical thrombectomy device group and control at baseline (p=0.50) or at 30 days (p=0.26). Ejection fraction was also measured at discharge and did not differ significantly between the mechanical thrombectomy device group and control (51.0 percent ±7.7 versus 48.7 percent ±10.9, p=0.29).44

Mechanical thrombectomy devices in other ACS populations. No trials or studies assessed for this endpoint in this population.

Distal filter embolic protection devices in patients with STEMI. Two RCTs evaluated the impact of distal filter embolic protection devices versus control on ejection fraction but were not amenable for statistical pooling therefore results are reported qualitatively (Table 16).95,98 In the first trial there was no significant difference in ejection fraction measured at 48 to 72 hours postprocedure between the distal filter embolic protection device group and control (47.4 percent ±9.9 versus 45.3 percent ±7.3, p=0.29).95 In the second trial left ventricular ejection fraction measured after the procedure did not differ significantly between the distal filter embolic protection device group and control (47 percent versus 44 percent, p=0.56).98

No controlled observational studies assessed for this endpoint in this population.

Distal filter embolic protection devices in other ACS populations. One RCT evaluated the impact of the distal filter embolic protection device FilterWire EX™ versus control on ejection fraction in patients with either NSTEMI or STEMI (Table 17).126 In this trial, ejection fraction values were reported in a figure, therefore Engauge Digitizer, Version 2.0 was used to read the figure and obtain values for ejection fraction. There was no significant difference in the ejection fraction measured at 3 days postprocedure between the distal filter embolic protection device group and control (47.57 percent ±10.94 versus 51.22 percent ±11.75, p= 0.26).

No controlled observational studies assessed for this endpoint in this population.

Distal balloon embolic protection devices in patients with STEMI. Six RCTs evaluated the impact of distal balloon embolic protection devices versus control on ejection fraction but were not amenable for statistical pooling therefore results are reported qualitatively (Table 18).17,103,107,110,119,133,135 In the first trial, there was a significantly higher ejection fraction at 3 and 6 months postPCI in the distal balloon embolic protection device group versus control (51.6 ±3.6 versus 49.3±5.3 percent and 53.0±3.7 percent versus 50.8±5.2 percent, respectively, p<0.05 for both comparisons).135 Authors reported that the difference between the two groups at 1 month was not significantly different although values were not reported. In the second trial there was no significant difference in the mean ejection fraction at baseline (52.1 percent ±9.4 versus 49.0 percent ±11.2, p=0.10) or at 6 months (58.1 percent ±11.4 versus 54.6 percent ±10.3, p=0.24) between the distal balloon embolic protection device group and control.17 The change in left ventricular ejection fraction from baseline to 6 months did not differ significantly between the distal balloon embolic protection device group and control (6.18 percent ±9.46 versus 5.65 percent ±8.64, p=0.83), respectively.17 In the third trial there was no significant difference in left ventricular ejection fraction at 3 days postprocedure (50 percent ±9 versus 49 percent ±13, p=0.60) or at 6 months (48 percent ±16 versus 50 percent ±9, p=0.74) between the distal balloon embolic protection device group and control.103 In the fourth trial there was no significant difference in left ventricular ejection fraction after the procedure (46.1 percent ±9.5 versus 55.4 percent ±13.9, p= .99) or at 6 months (61.9 percent versus 62.7 percent, p=0.36) between the distal balloon embolic protection device group and control.107 In the fifth trial there was no significant difference in left ventricular ejection fraction postprocedure (54.0 percent versus 53.8 percent, p=0.90), at 1 month (55.3 percent versus 55.4 percent, p=NS) or at 6 months (57.1 percent versus 57.1 percent, p=NS) between the distal balloon embolic protection device group and control.110,133 In the sixth trial there was no significant difference in mean left ventricular ejection fraction at discharge (47 percent ±9 versus 48 percent ±8, p=0.89) between the distal balloon embolic protection device group and control.119

No controlled observational studies assessed for this endpoint in this population.

Distal balloon embolic protection devices in other ACS populations. One RCT evaluated the impact of the distal balloon embolic protection device Guardwire™ Plus versus control on ejection fraction in patients with acute myocardial infarction (Table 17).130 The distal balloon embolic protection device group had a significantly higher post procedural mean left ventricular ejection fraction compared to control (51.2±14.5 percent versus 46.7±12.2 percent, p=0.02).

One RCT evaluated the impact of the distal balloon embolic protection device PercuSurge versus abciximab therapy on ejection fraction in patients with acute myocardial infarction (Table 19).164 There was no significant difference in median left ventricular ejection fraction upon admission between the distal balloon embolic protection device group and the abciximab group [43 percent (39-45) versus 40 (38-44), p=NS], respectively. Left ventricular ejection fractions increased in both groups at 6 months (46 percent (45-49) versus 46 percent (44-50), p=NS), although the changes were not significantly different between groups.

No controlled observational studies assessed for this endpoint in this population.

Proximal balloon embolic protection devices in patients with STEMI. One RCT evaluated the impact of the proximal balloon embolic protection device Proxis™ versus control on ejection fraction in patients with STEMI190 (Table 20). There was no significant difference in ejection fraction at 4 to 6 months postPCI in the Proxis™ group versus control (50 percent ±11 versus 50 percent ±12, p=0.46).

Proximal balloon embolic protection devices in patients with other ACS. No RCT or controlled observational studies evaluated the impact of proximal balloon embolic protection devices versus control on this outcome.

Embolic protection devices combined. No additional studies evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) aside from those reported in their respective device categories.

Myocardial Blush Grade

Direct Comparative Trials

Catheter aspiration device versus catheter aspiration device in patients with STEMI. One direct comparative randomized trial evaluated the impact of the Diver™-Invatec catheter aspiration device versus the Export®-Medtronic catheter aspiration device on MBG.158 The use of Diver™-Invatec did not significantly impact the risk of attaining a MBG-3 [RR 0.71 (0.42, 1.18)] compared to Export®-Medtronic.

Catheter aspiration device versus distal balloon embolic protection device in patients with STEMI. One direct comparative randomized trial evaluated the impact of the catheter aspiration device Diver™ CE versus the distal balloon embolic protection device Guardwire™ Plus on MBG.160 The use of Diver™ CE did not significantly impact the risk of attaining a MBG of 2 [RR 0.97 (0.77, 1.23)] compared to Guardwire™ Plus.

Trials Versus Control

Catheter aspiration devices in patients with STEMI. Thirteen RCTs evaluated the impact of catheter aspiration devices versus control on MBG and were included in the pooled analysis.12-16,19,20,62,69,74,83,86,87,138,175,176 The use of a catheter aspiration device significantly increased the risk of attaining a MBG-3 [RR 1.61 (1.41, 1.84)] (Figure 33). A higher level of statistical heterogeneity was found (I2=55.4 percent) but publication bias was not detected (Egger's P=0.117). Give the risk difference [RD 0.22 (0.16, 0.28), (CER 0.12 to 0.71)], five people would need to receive the catheter aspiration device to cause one person to experience a MBG-3.

This figure depicts the meta-analysis of the impact of catheter aspiration devices versus control on myocardial blush grade of 3. The first trial by Dudek and colleagues in 2010 provided a relative risk of 1.32 with a 95 percent confidence interval of 1.07 to 1.66. The second trial by Liistro and colleagues in 2009 provided a relative risk of 1.30 with 95 percent confidence interval of 1.10 to 1.60. The third trial by Lipiecki and colleagues, also in 2009, provided a relative risk of 1.03 with 95 percent confidence interval of 0.42 to 2.49. The fourth trial by Moura and colleagues, also in 2009, provided a relative risk of 1.45 with 95 percent confidence interval of 1.21 to 1.79. The fifth trial by Sardella and colleagues, also in 2009, provided a relative risk of 2.45 with a 95 percent confidence interval of 1.74 to 3.55. The sixth trial by Chevalier and colleagues in 2008 provided a relative risk of 1.40 with 95 percent confidence interval of 0.96 to 2.05. The seventh trial by Ikari and colleagues, also in 2008, provided a relative risk of 2.25 with 95 percent confidence interval of 1.62 to 3.16. The eighth trial by Slivaas and colleagues, also in 2008, provided a relative risk of 1.42 with a 95 percent confidence interval of 1.21 to 1.67. The ninth trial by De Luca and colleagues in 2006 provided a relative risk of 2.80 with 95 percent confidence interval of 1.18 to 6.95. The tenth trial by Lee and colleagues, also in 2006, provided a relative risk of 1.97 with 95 percent confidence interval of 0.93 to 4.25. The eleventh trial by Silva-Orrego and colleagues, also in 2006, provided a relative risk of 2.03 with 95 percent confidence interval of 1.58 to 2.72. The twelfth trial by Burzotta and colleagues in 2005 provided a relative risk of 1.87 with a 95 percent confidence interval of 1.04 to 3.48. The thirteenth trial by Dudek and colleagues in 2004 provided a relative risk of 1.47 with 95 percent confidence interval of 0.89 to 2.55. The combined effect of the thirteen trials showed a relative risk of 1.61 with a 95 percent confidence interval of 1.41 to 1.84. The Cochran Q p-value was 0.008, the I-squared value was 55.4 percent and the Egger's p-value was 0.117.

Figure 33

Impact of catheter aspiration devices versus control on myocardial blush grade of 3 in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.008 I2: 55.4 percent

When limiting the pooled analysis to trials of only good methodological quality,12,14-16,62,69,74,83,138,175,176 the risk of attaining a MBG-3 remained significantly increased [RR 1.75 (1.44, 2.14). A higher level of statistical heterogeneity was detected (I2=69.2) and a trend towards publication bias was detected (Egger's P=0.07). Given the risk difference for attaining a MBG-3 [RD 0.25 (0.16, 0.33), (CER 0.13 to 0.71)], four people would need to receive the catheter aspiration device to cause one person to experience a MBG-3.

One RCT evaluated the impact of the catheter aspiration device Diver™ CE versus control on MBG although was not included in the pooled analysis.162 In this trial, patients were only included if they attained TIMI-3 blood flow postprocedure, therefore it was not included in the pooled analysis of MBG. The use of a catheter aspiration device did not significantly impact the risk of attaining a MBG-3 [RR 1.04 (0.62, 1.69)] compared to control.

No controlled observational studies assessed for this endpoint in this population.

Catheter aspiration devices in other ACS populations. One RCT evaluated the impact of the catheter aspiration device Diver™-Invatec versus control on MBG in patients with acute myocardial infarction.127 The use of a catheter aspiration device significantly increased the risk of attaining a MBG-3 [RR 4.45 (1.51, 13.88] compared to control. Given the risk difference for MBG-3 [RD 0.30 (0.10, 0.51), (CER 0.09)], three people would need to be treated with a catheter aspiration device to cause one person to achieve a MBG-3. This trial was determined to be of poor methodological quality.

No controlled observational studies assessed for this endpoint in this population.

Mechanical thrombectomy devices in patients with STEMI. Four RCTs evaluated the impact of mechanical thrombectomy devices versus control on MBG.11,29,40,44 The use of a mechanical thrombectomy device did not significantly impact the risk of attaining a MBG-3 [RR 1.07 (0.80, 1.43)] (Figure 34). A higher level of statistical heterogeneity was found (I2=76.5 percent) but publication bias was not detected (Egger's P=0.408). All trials were determined to be of good methodological quality. 11,29,40,44

This figure depicts the meta-analysis of the impact of mechanical thrombectomy devices versus control on myocardial blush grade of 3. The first trial by Migliorini and colleagues in 2010 provided a relative risk of 0.91 with 95 percent confidence interval of 0.82 to 1.01. The second trial by Ali and colleagues in 2006 provided a relative risk of 0.84 with 95 percent confidence interval of 0.64 to 1.12. The third trial by Lefevre and colleagues also in 2005 provided a relative risk of 1.02 with 95 percent confidence interval of 0.67 to 1.57. The fourth trial by Napodano and colleagues in 2003 provided a relative risk of 1.94 with a 95 percent confidence interval of 1.31 to 3.02. The combined effect of the four trials showed a relative risk of 1.07 with a 95 percent confidence interval of 0.80 to 1.43. The Cochran Q p-value was 0.005, the I-squared value was 76.5 percent and the Egger's p-value was 0.408.

Figure 34

Impact of mechanical thrombectomy devices versus control on myocardial blush grade of 3 in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.005 I2: 76.5 percent

No controlled observational studies assessed for this endpoint in this population.

Mechanical thrombectomy devices in other ACS populations. No trials or studies assessed for this endpoint in this population.

Distal filter embolic protection devices in patients with STEMI. Two RCTs evaluated the impact of distal filter embolic protection devices versus control on MBG.95,98 In these trials, the use of distal filter embolic protection devices did not significantly impact the risk of attaining a MBG-3 [RR 0.97 (0.81, 1.15)] (Figure 35). Publication bias could not be evaluated. Both of the trials were determined to be of good methodological quality.95,98

This figure depicts the meta-analysis of the impact of distal filter embolic protection devices versus control on myocardial blush grade of 3. The first trial by Cura and colleagues in 2007 provided a relative risk of 0.94 with 95 percent confidence interval of 0.75 to 1.18. The second trial by Guetta and colleagues, also in 2007, provided a relative risk of 1.01 with 95 percent confidence interval of 0.76 to 1.35. The combined effect of the two trials showed a relative risk of 0.97 with a 95 percent confidence interval of 0.81 to 1.15. The Cochran Q p-value was 0.692 and there were too few strata to calculate the I-squared value and the Egger's p-value.

Figure 35

Impact of distal filter embolic protection devices versus control on myocardial blush grade of 3 in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.692 I2: Too few strata

No controlled observational studies assessed for this endpoint in this population.

Distal filter embolic protection devices in other ACS populations. No trials or studies assessed for this endpoint in this population.

Distal balloon embolic protection devices in patients with STEMI. Six RCTs evaluated the impact of distal balloon embolic protection devices versus control on MBG.17,103,107,111,112,133 The use of a distal balloon embolic protection device significantly increased the risk of attaining MBG-3 [RR 1.39 (1.15, 1.69)] (Figure 36). A lower level of statistical heterogeneity was found (I2=43.5 percent) but publication bias was not detected (Egger's P=0.203). Given the risk difference [RD 0.15 (0.10, 0.24), (CER 0.20 to 0.53)], seven people would need to be treated with a distal balloon embolic protection device to cause one person to experience a MBG-3. All of the trials were determined to be of good methodological quality.17,103,107,111,112,133

This figure depicts the meta-analysis of the impact of distal balloon embolic protection devices versus control on myocardial blush grade of 3 in patients with ST-segment elevation myocardial infarction. The first trial by Tahk and colleagues in 2008 provided a relative risk of 1.82 with 95 percent confidence interval of 1.25 to 2.75. The second trial by Hahn and colleagues in 2007 provided a relative risk of 1.26 with 95 percent confidence interval of 0.48 to 3.39. The third trial by Matsuo and colleagues, also in 2007, provided a relative risk of 1.33 with 95 percent confidence interval of 0.97 to 1.85. The fourth trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 1.24 with a 95 percent confidence interval of 0.83 to 1.86. The fifth trial by Zhou and colleagues, also in 2007, provided a relative risk of 1.96 with 95 percent confidence interval of 1.32 to 2.99. The sixth trial by Stone and colleagues in 2005 provided a relative risk of 1.16 with 95 percent confidence interval of 0.98 to 1.36. The combined effect of the six trials showed a relative risk of 1.39 with a 95 percent confidence interval of 1.15 to 1.69. The Cochran Q p-value was 0.115, the I-squared value was 43.5 percent and the Egger's p-value was 0.203.

Figure 36

Impact of distal balloon embolic protection devices versus control on myocardial blush grade of 3 in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.115 I2: 43.5 percent

No controlled observational studies assessed for this endpoint in this population.

Distal balloon embolic protection devices in other ACS populations. Two RCTs evaluated the impact of distal balloon embolic protection devices versus control on MBG in patients with acute myocardial infarction.125,130 The use of a distal balloon embolic protection device significantly increased the risk of attaining a MBG-3 [RR 3.22 (1.03, 10.10)] compared to control (Figure 37). Given the risk difference [RD 0.51 (0.18, 0.84), (CER 0.14 to 0.37)], two people would need to be treated with a distal balloon embolic protection device in order to cause one to achieve a MBG-3. Neither trial was determined to be of good methodological quality.125,130

This figure depicts the meta-analysis of the impact of distal balloon embolic protection devices versus control on myocardial blush grade of 3 in patients with mixed acute coronary syndrome. The first trial by Parikh and colleagues in 2008 provided a relative risk of 5.92 with 95 percent confidence interval of 2.77 to 13.74. The second trial by Nanasato and colleagues in 2004 provided a relative risk of 1.96 with 95 percent confidence interval of 1.19 to 3.34. The combined effect of the two trials showed a relative risk of 3.22 with a 95 percent confidence interval of 1.03 to 10.08. The Cochran Q p-value was 0.020 and there were too few strata to calculate the I-squared value and the Egger's p-value.

Figure 37

Impact of distal balloon embolic protection devices versus control on myocardial blush grade of 3 in patients with mixed acute coronary syndrome. Cochran Q: P=0.020 I2: Too few strata

One RCT evaluated the impact of the distal balloon embolic protection device PercuSurge versus abciximab therapy on MBG in patients with acute myocardial infarction.164 The use of a distal balloon embolic protection device did not significantly impact the risk of attaining a MBG-3 [RR 0.94 (0.71, 1.25)] versus abciximab therapy.

No controlled observational studies assessed for this endpoint in this population.

Proximal balloon embolic protection devices in patients with STEMI. One RCT evaluated the impact of the proximal balloon embolic protection device Proxis™ versus control on MBG.18 The use of a proximal balloon embolic protection device did not significantly impact the risk of attaining MBG-3 [RR 0.98 (0.88, 1.10)]. Limiting the analysis to trials of good methodological quality18 did not change the results.

No controlled observational studies assessed for this endpoint in this population.

Proximal balloon embolic protection devices in other ACS. No trials or studies assessed for this endpoint in this population.

Embolic protection devices combined in patients with STEMI. Nine RCTs evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) versus control on MBG.17,18,95,98,103,107,111,112,133 In these trials, the use of embolic protection devices combined significantly increased the risk of attaining a MBG-3 [RR 1.20 (1.02, 1.40)] (Figure 38). A high level of statistical heterogeneity was detected (I2=68.2 percent) but publication bias was not detected (Egger's P=0.055). Given the risk difference [RD 0.09 (0.02, 0.17), (CER 0.20 to 0.82)], eleven people would need to be treated with an embolic protection device to cause one person to achieve a MBG-3. All of the trials were determined to be of good methodological quality.17,18,95,98,103,107,111,112,133

This figure depicts the meta-analysis of the impact of embolic protection devices combined versus control on myocardial blush grade of 3. The first trial by Haeck and colleagues in 2009 provided a relative risk of 0.98 with 95 percent confidence interval of 0.87 to 1.10. The second trial by Tahk and colleagues in 2008 provided a relative risk of 1.82 with 95 percent confidence interval of 1.25 to 2.75. The third trial by Cura and colleagues in 2007 provided a relative risk of 0.94 with 95 percent confidence interval of 0.75 to 1.18. The fourth trial by Guetta and colleagues, also in 2007, provided a relative risk of 1.01 with a 95 percent confidence interval of 0.76 to 1.35. The fifth trial by Hahn and colleagues, also in 2007, provided a relative risk of 1.26 with 95 percent confidence interval of 0.48 to 3.39. The sixth trial by Matsuo and colleagues, also in 2007, provided a relative risk of 1.33 with 95 percent confidence interval of 0.97 to 1.85. The seventh trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 1.24 with 95 percent confidence interval of 0.83 to 1.86. The eighth trial by Zhou and colleagues, also in 2007, provided a relative risk of 1.96 with a 95 percent confidence interval of 1.32 to 2.99. The ninth trial by Stone and colleagues in 2005 provided a relative risk of 1.16 with a 95 percent confidence interval of 0.98 to 1.36. The combined effect of the nine trials showed a relative risk of 1.19 with a 95 percent confidence interval of 1.02 to 1.40. The Cochran Q p-value was 0.002, the I-squared value was 68.2 percent and the Egger's p-value was 0.055

Figure 38

Impact of embolic protection devices combined versus control on myocardial blush grade of 3 in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.002 I2: 68.2 percent

Embolic protection devices combined in other ACS populations. No trials or studies were available that evaluated the impact of an embolic protection device versus control on MBG-3 in this patient population in addition to the two trials pooled and reported in the distal balloon embolic protection device section above.

TIMI-3 Blood Flow

Direct Comparative Trials

Catheter aspiration device versus catheter aspiration device in patients with STEMI. One direct comparative randomized trial evaluated the impact of the Diver™-Invatec catheter aspiration device versus the Export®-Medtronic catheter aspiration device on TIMI-3 blood flow.158 The use of Diver™-Invatec did not significantly impact the risk of attaining TIMI-3 blood flow [RR 0.89 (0.71, 1.10)] compared to Export®-Medtronic.

Catheter aspiration device versus distal balloon embolic protection device in patients with STEMI. One direct comparative randomized trial evaluated the impact of the catheter aspiration device Diver™ CE versus the distal balloon embolic protection device Guardwire™ Plus on TIMI-3 blood flow.160 The use of Diver™ CE did not significantly impact the risk of attaining TIMI-3 blood flow [RR 0.98 (0.89, 1.08)] compared to Guardwire™ Plus.

Trials Versus Control

Catheter aspiration devices in patients with STEMI. Thirteen RCTs evaluated the impact of catheter aspiration devices versus control on TIMI-3 blood flow.12-16,19,62,69,71,74,83,85,87,138,175,176 The use of a catheter aspiration device significantly increased the risk of attaining TIMI-3 blood flow [RR 1.08 (1.04, 1.12)] (Figure 39). A lower level of statistical heterogeneity was found (I2 =11.5 percent) but no publication bias was detected (Egger's P=0.585). Given the risk difference [RD 0.06 (0.03, 0.10), (CER 0.68 to 0.88)], 17 people would need to be treated with a catheter aspiration device to cause one person to achieve TIMI-3 blood flow.

This figure depicts the meta-analysis of the impact of catheter aspiration devices versus control on TIMI-3 in patients with ST-segment elevation myocardial infarction. The first trial by Dudek and colleagues in 2010 provided a relative risk of 1.08 with a 95 percent confidence interval of 0.96 to 1.23. The second trial by Liistro and colleagues in 2009 provided a relative risk of 1.17 with 95 percent confidence interval of 1.04 to 1.38. The third trial by Lipiecki and colleagues, also in 2009, provided a relative risk of 0.66 with 95 percent confidence interval of 0.40 to 0.98. The fourth trial by Sardella and colleagues, also in 2009, provided a relative risk of 1.19 with 95 percent confidence interval of 1.00 to 1.42. The fifth trial by Chevalier and colleagues in 2008 provided a relative risk of 1.06 with a 95 percent confidence interval of 0.93 to 1.21. The sixth trial by Ikari and colleagues, also in 2008, provided a relative risk of 1.09 with 95 percent confidence interval of 0.99 to 1.10. The seventh trial by Svilaas and colleagues, also in 2008, provided a relative risk of 1.04 with 95 percent confidence interval of 0.99 to 1.10. The eighth trial by De Luca and colleagues in 2006 provided a relative risk of 1.15 with a 95 percent confidence interval of 0.88 to 1.55. The ninth trial by Kaltoft and colleagues, also in 2006, provided a relative risk of 1.02 with a 95 percent confidence interval of 0.92 to 1.14. The tenth trial by Silva-Orrego and colleagues, also in 2006, provided a relative risk of 1.14 with 95 percent confidence interval of 0.99 to 1.33. The eleventh trial by Burzotta and colleagues in 2005 provided a relative risk of 1.18 with 95 percent confidence interval of 0.94 to 1.51. The twelfth trial by Noel and colleagues, also in 2005, provided a relative risk of 1.19 with 95 percent confidence interval of 0.96 to 1.55. The thirteenth trial by Dudek and colleagues in 2004 provided a relative risk of 0.97 with a 95 percent confidence interval of 0.79 to 1.29. The combined effect of the thirteen trials showed a relative risk of 1.08 with a 95 percent confidence interval of 1.04 to 1.12. The Cochran Q p-value was 0.329, the I-squared value was 11.5 percent and the Egger's p-value was 0.585.

Figure 39

Impact of catheter aspiration devices versus control on TIMI-3 blood flow in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.329 I2: 11.5 percent

Limiting the pooled analyses to trials of good methodological quality still resulted in a significantly increased risk of attaining TIMI-3 blood flow [RR 1.07 (1.04, 1.11)]. A lower level of statistical heterogeneity was found (I2=0 percent). Given the risk difference [RD 0.06 (0.03, 0.10), (CER 0.68 to 0.88)], 17 people would need to be treated with a catheter aspiration device to cause one person to achieve TIMI-3 blood flow.12,14-16,62,69,71,74,83,138,175,176

Two controlled observational studies evaluated the impact of catheter aspiration devices versus control on TIMI-3 blood flow.142,144 In the first study, the use of a catheter aspiration device did not significantly impact the achievement of TIMI-3 blood flow versus control (89.1 percent versus 87.6 percent, p=0.67).144 In the second study, the use of a catheter aspiration device did not significantly impact the achievement of TIMI-3 blood flow versus control (88.3 percent versus 86.5 percent, p=0.471).142

Catheter Aspiration Devices in other ACS populations. Two RCTs evaluated the impact of catheter aspiration devices versus control on TIMI-3 blood flow in patients with acute myocardial infarction.127,128 The use of a catheter aspiration device did not significantly impact the risk of attaining TIMI-3 blood flow [RR 1.15 (0.82, 1.62)] compared to contro127,128 (Figure 40). Both trials were determined to be of poor methodological quality.

This figure depicts the meta-analysis of the impact of catheter aspiration devices versus control on TIMI-3 in patients with mixed acute coronary syndrome. The first trial by Sardella and colleagues in 2005 provided a relative risk of 1.39 with 95 percent confidence interval of 1.02 to 1.95. The second trial by Kuni and colleagues in 2004 provided a relative risk of 1.02 with 95 percent confidence interval of 0.95 to 1.10. The combined effect of the two trials showed a relative risk of 1.15 with a 95 percent confidence interval of 0.82 to 1.62. The Cochran Q p-value was 0.027 and there were too few strata to calculate the I-squared value and the Egger's p-value.

Figure 40

Impact of catheter aspiration devices versus control on TIMI-3 blood flow in patients with mixed acute coronary syndrome. Cochran Q: P=0.027 I2: Too few strata

No controlled observational studies assessed for this endpoint in this population.

Mechanical thrombectomy devices in patients with STEMI. Four RCTs evaluated the impact of mechanical thrombectomy devices versus control on TIMI-3 blood flow.11,29,40,44 The use of a mechanical thrombectomy device did not significantly impact the risk of attaining TIMI-3 blood flow [RR 0.98 (0.92, 1.04)] (Figure 41). A high level of statistical heterogeneity was found (I2=67.5 percent) but publication bias was not detected (Egger's P=0.464). All of the trials were determined to be of good methodological quality.11,29,40,44

This figure depicts the meta-analysis of the impact of mechanical thrombectomy devices versus control on TIMI-3. The first trial by Migliorini and colleagues in 2010 provided a relative risk of 0.94 with 95 percent confidence interval of 0.86 to 1.02. The second trial by Ali and colleagues in 2006 provided a relative risk of 0.94 with 95 percent confidence interval of 0.89 to 0.98. The third trial by Lefevre and colleagues in 2005 provided a relative risk of 1.08 with 95 percent confidence interval of 0.99 to 1.18. The fourth trial by Napodano and colleagues in 2003 provided a relative risk of 0.98 with a 95 percent confidence interval of 0.86 to 1.10. The combined effect of the four trials showed a relative risk of 0.98 with a 95 percent confidence interval of 0.92 to 1.04. The Cochran Q p-value was 0.026, the I-squared value was 67.5 percent and the Egger's p-value was 0.464.

Figure 41

Impact of mechanical thrombectomy devices versus control on TIMI-3 blood flow in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.026 I2: 67.5 percent

One controlled observational study evaluated the association between the use of a mechanical thrombectomy device and TIMI-3 blood flow.145 Patients undergoing PCI with a mechanical thrombectomy device, either the AngioJet® XMI or XVG catheter, were compared to patients undergoing PCI without mechanical thrombectomy. The use of a mechanical thrombectomy device was associated with a significantly lower rate of TIMI-3 blood flow compared to PCI without a mechanical thrombectomy device (86 percent versus 90 percent, p=0.04).

Mechanical Thrombectomy Devices in other ACS populations. One RCT evaluated the impact of the mechanical thrombectomy device X-Sizer® versus control on TIMI-3 blood flow in patients with STEMI or UA.166 The use of a mechanical thrombectomy device did not significantly impact the risk of attaining TIMI-3 blood flow [RR 1.07 (0.86, 1.36)] compared to control. This trial was determined to be of good methodological quality.

One controlled observational study evaluated the association between the use of mechanical thrombectomy devices and TIMI-3 blood flow.153 The types of ACSs included in this study were not reported. Patients undergoing PCI with the mechanical thrombectomy device AngioJet® were compared to patients undergoing PCI without mechanical thrombectomy. The use of a mechanical thrombectomy device was associated with a significantly lower rate of TIMI-3 blood flow compared to PCI without a mechanical thrombectomy device (85 percent versus 93 percent, p=0.0003). However, there were significantly more patients with TIMI-3 blood flow in the mechanical thrombectomy device group at baseline compared to the group without mechanical thrombectomy (15 percent versus 27 percent, p=0.0001).

Distal Filter Embolic Protection Devices in patients with STEMI. Five RCTs evaluated the impact of distal filter embolic protection devices versus control on TIMI-3 blood flow.89,93,98,101,137 In these trials, the use of distal filter embolic protection devices did not significantly impact the risk of attaining TIMI-3 blood flow [RR 1.00 (0.90, 1.11)] (Figure 42). A higher level of statistical heterogeneity was detected (I2=69.6 percent) although publication bias was not detected (Egger's P=0.252).

This figure depicts the meta-analysis of the impact of distal filter embolic protection devices versus control on TIMI-3. The first trial by Ito and colleagues in 2010 provided a relative risk of 1.17 with a 95 percent confidence interval of 0.85 to 1.73. The second trial by Kelbaeck and colleagues in 2008 provided a relative risk of 1.11 with 95 percent confidence interval of 1.05 to 1.17. The third trial by Cura and colleagues in 2007 provided a relative risk of 0.92 with 95 percent confidence interval of 0.77 to 1.07. The fourth trial by Guetta and colleagues, also in 2007, provided a relative risk of 0.94 with 95 percent confidence interval of 0.80 to 1.08. The fifth trial by Lefevre and colleagues in 2004 provided a relative risk of 0.94 with a 95 percent confidence interval of 0.78 to 1.11. The combined effect of the five trials showed a relative risk of 1.00 with a 95 percent confidence interval of 0.90 to 1.11. The Cochran Q p-value was 0.011, the I-squared value was 69.6 percent and the Egger's p-value was 0.252.

Figure 42

Impact of distal filter embolic protection devices versus control on TIMI-3 blood flow in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.011 I2: 69.6 percent

When limiting the pooled analysis to only trials of good methodological quality,89,95,98,137 the risk of attaining TIMI-3 blood flow remained nonsignificant in the distal filter embolic protection device group versus control [RR 1.02 (0.90, 1.15)]. A higher level of statistical heterogeneity was detected (I2=70.2 percent).

No controlled observational studies assessed for this endpoint in this population.

Distal filter embolic protection devices in other ACS populations. Three RCTs evaluated the impact of distal filter embolic protection devices versus control in patients with other ACSs on TIMI-3 blood flow although were not suitable for pooling because each trial evaluated a different ACS.126,155,156 In the first trial, the FilterWire EZ™ device was compared to control in patients with NSTEMI.155 The use of a distal filter embolic protection device did not significantly impact the risk of attaining TIMI-3 blood flow [RR 0.99 (0.90, 1.09)]. In the second trial the Angioguard™ device was compared to control in patients with UA.156 The risk of attaining TIMI-3 blood flow could not be calculated because all patients in both groups attained TIMI-3 blood flow after the procedure. In the third trial, the FilterWire EX™ was compared to control in patients with either NSTEMI or STEMI.126 The risk of attaining TIMI-3 blood flow was not different between the distal filter embolic protection device group and control [RR 1.00 (0.92, 1.09)]. Of the three trials, this one trial was determined to be of good methodological quality.

No controlled observational studies assessed for this endpoint in this population.

Distal balloon embolic protection devices in patients with STEMI. Nine RCTs evaluated the impact of distal balloon embolic protection devices versus control on TIMI-3.17,103,107,111,112,119,133,135,136 One study was excluded from the pooled analysis of relative risk because all patients in both groups achieved TIMI-3 blood flow with the same number of participants in each group.119 In the eight trials eligible for pooling, the use of a distal balloon embolic protection device significantly increased the risk of attaining TIMI-3 blood flow [RR 1.11 (1.03, 1.19)]17,103,107,111,112,133 (Figure 43). A higher level of statistical heterogeneity was found (I2=60.4 percent) but publication bias was not detected (Egger's P=0.094). Using the risk difference [RD 0.08 (0.02, 0.14, (CER 0.69 to 1.00)], for every 13 patients who undergo surgery with a distal balloon embolic protection device 1 will achieve TIMI-3 blood flow. When limiting the analysis to trials of good methodological quality, the achievement of TIMI-3 blood flow remained significantly increased in the distal balloon embolic protection device group versus control [RR 1.09 (1.01, 1.17)]. Using the risk difference [RD 0.07 (0.01 to 0.13), (CER 0.75 to 0.96)] for every 15 patients who undergo surgery with a distal balloon embolic protection device one will achieve TIMI-3 blood flow.

This figure depicts the meta-analysis of the impact of distal balloon embolic protection devices versus control on TIMI-3 in patients with ST-segment elevation myocardial infarction. The first trial by Duan and colleagues in 2010 provideed a relative risk of 1.23 with a 95 percent confidence interval of 1.05 to 1.49. The second trial by Pan and colleages in 2010 provided a relative risk of 1.28 with a 95 percent confidence interval of 1.05 to 1.61. The third trial by Tahk and colleagues in 2008 provided a relative risk of 1.26 with 95 percent confidence interval of 1.10 to 1.51. The fourth trial by Hahn and colleagues in 2007 provided a relative risk of 1.00 with 95 percent confidence interval of 0.79 to 1.25. The fifth trial by Matsuo and colleagues, also in 2007, provided a relative risk of 1.06 with 95 percent confidence interval of 0.89 to 1.27. The sixth trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 0.99 with a 95 percent confidence interval of 0.88 to 1.11. The seventh trial by Zhou and colleagues, also in 2007, provided a relative risk of 1.20 with 95 percent confidence interval of 1.05 to 1.42. The eight trial by Okamura and colleagues in 2005 was excluded from the meat-analysis because no TIMI-3 blood flow occurred in either the treatment or control groups and therefore a relative risk could not be calculated. The nineth trial by Stone and colleagues, also in 2005, provided a relative risk of 1.03 with 95 percent confidence interval of 0.97 to 1.09. The combined effect of the eight trials showed a relative risk of 1.11 with a 95 percent confidence interval of 1.03 to 1.19. The Cochran Q p-value was 0.014, the I-squared value was 60.4 percent and the Egger's p-value was 0.094.

Figure 43

Impact of distal balloon embolic protection devices versus control on TIMI-3 blood flow in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.014 I2: 60.4 percent

No controlled observational studies assessed for this endpoint in this population.

Distal balloon embolic protection devices in other ACS populations. Two RCTs evaluated the impact of distal balloon embolic protection devices versus control on TIMI-3 blood flow in patients with acute myocardial infarction.125,130 The use of a distal balloon embolic protection device did not significantly impact the risk of attaining TIMI-3 blood flow [RR 1.36 (0.65, 2.86)] compared to control (Figure 44). Neither trial was determined to be of good methodological quality.

This figure depicts the meta-analysis of the impact of distal balloon embolic protection devices versus control on TIMI-3 in patients with mixed acute coronary syndrome. The first trial by Parikh and colleagues in 2008 provided a relative risk of 1.78 with 95 percent confidence interval of 1.27 to 2.64. The second trial by Nanasato and colleagues in 2004 provided a relative risk of 1.07 with 95 percent confidence interval of 0.93 to 1.25. The combined effect of the two trials showed a relative risk of 1.36 with a 95 percent confidence interval of 0.65 to 2.86. The Cochran Q p-value was less than 0.0001 and there were too few strata to calculate the I-squared value and the Egger's p-value

Figure 44

Impact of distal balloon embolic protection devices versus control on TIMI-3 blood flow in patients with mixed acute coronary syndrome. Cochran Q: P <0.0001 I2: Too few strata

One RCT evaluated the impact of the distal balloon embolic protection device PercuSurge versus abciximab therapy on TIMI-3 blood flow in patients with acute myocardial infarction.164 The use of a distal balloon embolic protection device did not significantly impact the risk of attaining TIMI-3 blood flow [RR 1.01 (0.87, 1.15)] versus abciximab therapy.

No controlled observational studies assessed for this endpoint in this population.

Proximal balloon embolic protection devices in patients with STEMI. One RCT evaluated the impact of the proximal balloon embolic protection device Proxis™ versus control on TIMI-3 blood flow.18 The use of a proximal balloon embolic protection device did not significantly impact the risk of attaining TIMI-3 blood flow [RR 1.06 (0.98, 1.16)]. This trial was determined to be of good methodological quality.18

Proximal balloon embolic protection devices in other ACS populations. No trials or studies assessed for this endpoint in this population.

Embolic protection devices combined in patients with STEMI. Fifteen RCTs evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) versus control on TIMI-3 blood flow.17,18,89,95,98,101,103,107,111,112,119,133,135-137 The trial by Okamura et al was excluded from the pooled analysis of relative risk because no events occurred within the prespecified time period in either control or treatment group. In the trials 14 suitable for pooling, the use of embolic protection devices combined significantly increased the risk of attaining TIMI-3 blood flow [RR 1.06 (1.01, 1.12)] (Figure 45). Using the risk difference [RD 0.05 (0.01, 0.10), (CER 0.69 to 0.96)] 1 patient would attain TIMI-3 blood flow after surgery for every 25 patients who undergo surgery with an embolic protection device. A high level of statistical heterogeneity was detected (I2=58.3 percent) but publication bias was not detected (Egger's P=0.811).

This figure depicts the meta-analysis of the impact of embolic protection devices combined versus control on TIMI-3 in patients with ST-segment elevation myocardial infarction. The first trial by Duan and colleagies in 2010 provided a relative risk of 1.23 with a 95 percent confidence interval of 1.05 to 1.49. The second trial by Ito and colleagues in 2010 provided a relative risk of 1.17 with a 95 percent confidence interval of 0.85 to 1.73. The third trial by Pan and colleagues in 2010 provided a relative risk of 1.28 with a 95 percent confidence interval of 1.05 to 1.51. The fourth trial by Haeck and colleagues in 2009 provided a relative risk of 1.06 with 95 percent confidence interval of 0.98 to 1.16. The fifth trial by Kelbaeck and colleagues in 2008 provided a relative risk of 1.11 with 95 percent confidence interval of 1.05 to 1.17. The sixth trial by Tahk and colleagues, also in 2008, provided a relative risk of 1.26 with 95 percent confidence interval of 1.10 to 1.51. The seventh trial by Cura and colleagues in 2007 provided a relative risk of 0.92 with a 95 percent confidence interval of 0.77 to 1.07. The eight trial by Guetta and colleagues, also in 2007, provided a relative risk of 0.94 with 95 percent confidence interval of 0.80 to 1.08. The ninth trial by Hahn and colleagues, also in 2007, provided a relative risk of 1.00 with 95 percent confidence interval of 0.79 to 1.25. The tenth trial by Matsuo and colleagues, also in 2007, provided a relative risk of 1.06 with 95 percent confidence interval of 0.89 to 1.27. The eleventh trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 0.99 with 95 percent confidence interval of 0.88 to 1.11. The twelfth trial by Zhou and colleagues, also in 2007, provided a relative risk of 1.20 with a 95 percent confidence interval of 1.05 to 1.42. The thirteenth trial by Okamura and colleagues in 2005 was excluded from the meta-analysis because no TIMI-3 blood flow occurred in either the treatment or the control group and therefore a relative risk could not be calculated. The fourteenth trial by Stone and colleagues, also in 2005, provided a relative risk of 1.03 with 95 percent confidence interval of 0.97 to 1.09. The fifteenth trial by Lefevre and colleagues in 2004 provided a relative risk of 1.04 with 95 percent confidence interval of 0.99 to 1.10.The combined effect of the fourteen trials showed a relative risk of 1.06 with a 95 percent confidence interval of 1.01 to 1.12. The Cochran Q p-value was 0.003, the I-squared value was 58.3 percent and the Egger's p-value was 0.811.

Figure 45

Impact of embolic protection devices combined versus control on TIMI-3 blood flow in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.003 I2: 58.3 percent

When the pooled analysis was limited to only trials of good methodological quality,17,18,89,95,98,103,107,111,112,133,135,137 the risk of attaining TIMI-3 blood flow remained significantly increased in the combined embolic protection device group compared to control [RR 1.06 (1.01, 1.12)]. A higher level of statistical heterogeneity was detected (I2=55.4 percent).

Embolic protection devices combined in other ACS populations. Three RCTs evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) versus control in patients with mixed ACS (acute myocardial infarction, not otherwise specified) on attaining TIMI-3 blood flow.125,126,130 The use of an embolic protection device did not significantly impact the risk of attaining TIMI-3 blood flow versus control [RR 1.15 (0.93, 1.41)] (Figure 46). One trial was determined to be of higher methodological quality126 therefore sensitivity analysis was not possible based on trial quality.

This figure depicts the meta-analysis of the impact of embolic protection devices combined versus control on TIMI-3 in patients with mixed acute coronary syndrome. The first trial by Parikh and colleagues in 2008 provided a relative risk of 1.78 with 95 percent confidence interval of 1.27 to 2.64. The second trial by Gick and colleagues in 2005 provided a relative risk of 1.00 with 95 percent confidence interval of 0.92 to 1.09. The third trial by Nanasato and colleagues in 2004 provided a relative risk of 1.07 with 95 percent confidence interval of 0.93 to 1.25. The combined effect of the three trials showed a relative risk of 1.15 with a 95 percent confidence interval of 0.93 to 1.41. The Cochran Q p-value was 0.001, I-squared value was 85.5 percent and there were too few strata to calculate the Egger's p-value.

Figure 46

Impact of embolic protection devices combined versus control on TIMI-3 blood flow in patients with mixed acute coronary syndrome. Cochran Q: P=0.001 I2: 85.5 percent

Distal Embolization

Direct Comparative Trials

No direct comparative trials evaluated the impact of catheter aspiration, mechanical thrombectomy or embolic protection devices on this endpoint.

Trials Versus Control

Catheter aspiration devices in patients with STEMI. Ten RCTs evaluated the impact of catheter aspiration devices versus control on distal embolization.12-16,19,71,74,83,86,138,176 The use of a catheter aspiration device significantly decreased the risk of distal embolization [RR 0.56 (0.39, 0.79)] (Figure 47). A lower level of statistical heterogeneity was found (I2=43.4 percent) but no publication bias was detected (Egger's P=0.161). Given the risk difference [RD -0.09 (-0.17, -0.01), (CER 0.03 to 0.66)], 12 people would need to be treated with a catheter aspiration device to prevent one person from experiencing distal embolization.

This figure depicts the meta-analysis of the impact of catheter aspiration devices versus control on distal embolization. The first trial by Dudek and colleagues in 2010 provided a relative risk of 0.98 with a 95 percent confidence interval of 0.31 to 3.08. The second trial by Liistro and colleagues in 2009 provided a relative risk of 0.29 with 95 percent confidence interval of 0.11 to 1.78. The third trial by Lipiecki and colleagues, also in 2009, provided a relative risk of 0.80 with 95 percent confidence interval of 0.17 to 3.66. The fourth trial by Sardella and colleagues, also in 2009, provided a relative risk of 0.36 with 95 percent confidence interval of 0.24 to 0.54. The fifth trial by Chevalier and colleagues in 2008 provided a relative risk of 0.54 with a 95 percent confidence interval of 0.27 to 1.04. The sixth trial by Ikari and colleagues, also in 2008, provided a relative risk of 0.54 with 95 percent confidence interval of 0.36 to 0.81. The seventh trial by Kaltoft and colleagues in 2006 provided a relative risk of 1.51 with 95 percent confidence interval of 0.58 to 3.97. The eighth trial by Lee and colleagues, also in 2006, provided a relative risk of 2.96 with 95 percent confidence interval of 0.71 to 12.53. The ninth trial by Silva-Orrego and colleagues, also in 2006, provided a relative risk of 0.29 with a 95 percent confidence interval of 0.10 to 0.78. The tenth trial by Burzotta and colleagues in 2005 provided a relative risk of 0.49 with 95 percent confidence interval of 0.16 to 1.43. The combined effect of the ten trials showed a relative risk of 0.56 with a 95 percent confidence interval of 0.39 to 0.79. The Cochran Q p-value was 0.069, the I-squared value was 43.4 percent and the Egger's p-value was 0.161.

Figure 47

Impact of catheter aspiration devices versus control on distal embolization in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.069 I2: 43.4 percent

When limiting the pooled analysis to only trials of good methodological quality, 12,14-16,71,74,83,138,176 the risk of distal embolization remained significantly decreased in the catheter aspiration device group compared to control [RR 0.48 (0.34, 0.66)]. A lower level of statistical heterogeneity was detected (I2=33.7 percent). Given the risk difference [RD -0.14 (-0.23, -0.04), (CER 0.06 to 0.66)], seven people would need to be treated with a catheter aspiration device to prevent one person from experiencing distal embolization.

One controlled observational study evaluated the association between the use of catheter aspiration devices during PCI and distal embolization.152. The catheter aspiration devices included in this study were not reported. The use of a catheter aspiration device was associated with a significantly higher rate of distal embolization (9.0 percent versus 3.2 percent, p<0.0001).

Catheter aspiration devices in other ACS populations. No trials or studies assessed for this endpoint in this population.

Mechanical thrombectomy devices in patients with STEMI. Three RCTs evaluated the impact of mechanical thrombectomy devices versus control on distal embolization.29,40,44 The use of a mechanical thrombectomy device did not significantly impact the risk of distal embolization [RR 0.44 (0.17, 1.12)] (Figure 48). A lower level of statistical heterogeneity was found (I2=41.6 percent) and publication bias could not be evaluated. All of the trials were determined to be of good methodological quality. 29,40,44

This figure depicts the meta-analysis of the impact of mechanical thrombectomy devices versus control on distal embolization. The first trial by Ali and colleagues in 2006 provided a relative risk of 0.84 with 95 percent confidence interval of 0.38 to 1.86. The second trial by Lefevre and colleagues in 2005 provided a relative risk of 0.21 with 95 percent confidence interval of 0.05 to 0.81. The third trial by Napodano and colleagues in 2003 provided a relative risk of 0.29 with 95 percent confidence interval of 0.07 to 1.13. The combined effect of the three trials showed a relative risk of 0.44 with a 95 percent confidence interval of 0.17 to 1.12. The Cochran Q p-value was 0.181, the I-squared value was 41.6 percent and there were too few strata to calculate the Egger's p-value.

Figure 48

Impact of mechanical thrombectomy devices versus control on distal embolization in patients with ST-segment elevation myocardial infarction. Cochran: P=0.181 I2: 41.6 percent

No controlled observational studies assessed for this endpoint in this population.

Mechanical thrombectomy devices in other ACS populations. No trials or studies assessed for this endpoint in this population.

Distal filter embolic protection devices in patients with STEMI. One RCT evaluated the impact of a distal filter embolic protection device versus control on distal embolization.95 In this trial, the use of a distal filter embolic protection device did not significantly impact the risk of distal embolization [RR 0.63 (0.22, 1.73)]. This trial was determined to be of good quality.95

No controlled observational studies assessed for this endpoint in this population.

Distal filter embolic protection devices in other ACS populations. One RCT evaluated the impact of the distal filter embolic protection device FilterWire EX™ versus control on distal embolization in patients with either NSTEMI or STEMI.126 The use of a distal filter embolic protection device did not significantly impact the risk of distal embolization [RR 0.38 (0.11, 1.26)] compared to control. This trial was determined to be of good methodological quality.

No controlled observational studies assessed for this endpoint in this population.

Distal balloon embolic protection devices in patients with STEMI. Four RCTs evaluated the impact of distal balloon embolic protection devices versus control on distal embolization.103,107,112,133 The use of a distal balloon embolic protection device did not significantly impact the risk of distal embolization [RR 1.10 (0.67, 1.81)] (Figure 49). A lower level of statistical heterogeneity was found (I2=5.8 percent) and publication bias was not detected (Egger's P=0.176). All of the trials were determined to be of good methodological quality103,107,112,133 did not change the results.

This figure depicts the meta-analysis of the impact of distal filter embolic protection devices versus control on distal embolization. The first trial by Hahn and colleagues in 2007 provided a relative risk of 0.70 with 95 percent confidence interval of 0.24 to 1.99. The second trial by Matsuo and colleagues, also in 2007, provided a relative risk of 1.16 with 95 percent confidence interval of 0.35 to 3.86. The third trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 0.55 with 95 percent confidence interval of 0.18 to 1.74. The fourth trial by Stone and colleagues in 2005 provided a relative risk of 1.60 with 95 percent confidence interval of 0.85 to 3.03. The combined effect of the four trials showed a relative risk of 1.10 with a 95 percent confidence interval of 0.67 to 1.81. The Cochran Q p-value was 0.364, the I-squared value was 5.8 percent and the Egger's p-value was 0.176.

Figure 49

Impact of distal balloon embolic protection devices versus control on distal embolization in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.364 I2: 5.8 percent

No controlled observational studies assessed for this endpoint in this population.

Distal balloon embolic protection devices in other ACS populations. No trials or studies assessed for this endpoint in this population.

Proximal balloon embolic protection devices in patients with STEMI. One RCT evaluated the impact of the proximal balloon embolic protection device Proxis™ versus control on distal embolization.18 The use of a proximal balloon embolic protection device did not significantly impact the risk of having distal embolization [RR 0.71 (0.37, 1.35)]. This single trial was determined to be of good quality.18

No controlled observational studies assessed for this endpoint in this population.

Proximal balloon embolic protection devices in other ACS populations. No trials or studies assessed for this endpoint in this population.

Embolic protection devices combined in patients with STEMI. Six RCTs evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) versus control on distal embolization.18,95,103,107,112,133 In these trials, the use of embolic protection devices combined did not significantly impact the risk of distal embolization [RR 0.91 (0.64, 1.30)] (Figure 50). A low level of statistical heterogeneity was detected (I2=0.2 percent) but publication bias was not detected (Egger's P=0.409). All of the trials were determined to be of good methodological quality.18,95,103,107,112,133

This figure depicts the meta-analysis of the impact of embolic protection devices combined versus control on distal embolization. The first trial by Haeck and colleagues in 2009 provided a relative risk of 0.71 with 95 percent confidence interval of 0.38 to1.33. The second trial by Cura and colleagues in 2007 provided a relative risk of 0.63 with 95 percent confidence interval of 0.22 to 1.73. The third trial by Hahn and colleagues, also in 2007, provided a relative risk of 0.70 with 95 percent confidence interval of 0.24 to 1.99. The fourth trial by Matsuo and colleagues, also in 2007, provided a relative risk of 1.16 with 95 percent confidence interval of 0.35 to 3.86. The fifth trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 0.55 with 95 percent confidence interval of 0.18 to 1.74. The sixth trial by Stone and colleagues in 2005 provided a relative risk of 1.60 with 95 percent confidence interval of 0.85 to 3.03. The combined effect of the six trials showed a relative risk of 0.91 with a 95 percent confidence interval of 0.64 to 1.30. The Cochran Q p-value was 0.415, the I-squared value was 0.2 percent and the Egger's p-value was 0.409.

Figure 50

Impact of embolic protection devices combined versus control on distal embolization in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.415 I2: 0.2 percent

Embolic protection devices combined in other ACS populations. No trials or studies were available that evaluated the impact of any embolic protection device versus control on distal embolization in addition to the one trial reported above, and therefore pooling was not possible.

No Reflow

Direct Comparative Trials

Catheter aspiration device versus distal balloon protection device in STEMI. One direct comparative randomized trial evaluated the impact of the Diver™ CE catheter aspiration device versus the Guardwire™ Plus distal balloon embolic protection device on no reflow.160 In this study, a composite of no reflow / slow reflow was reported. The use of Diver™ CE did not significantly impact the risk of no reflow / slow reflow [RR 1.25 (0.38, 4.14)] compared to Guardwire™ Plus.

Trials Versus Control

Catheter aspiration devices in patients with STEMI. Eight RCTs evaluated the impact of catheter aspiration devices versus control on no reflow.12,15,16,19,51,74,83,85,86,138 The use of a catheter aspiration device significantly decreased the risk of no reflow [RR 0.52 (0.35, 0.76)] (Figure 51). A low level of statistical heterogeneity was found (I2=15.7 percent) but no publication bias was detected (Egger's P=0.278). Given the risk difference [RD -0.07 (-0.11, -0.03), (CER 0.05 to 0.27)], 15 people would need to be treated with a catheter aspiration device in order to prevent one no reflow event from occurring.

This figure depicts the meta-analysis of the impact of catheter aspiration devices versus control on no reflow. The first trial by Dudek and colleagues in 2010 provided a relative risk of 0.58 with a 95 percent confidence itnerval of 0.28 to 1.17. The second trial by Liistro and colleagues in 2009 provided a relative risk of 0.20 with 95 percent confidence interval of 0.05 to 0.78. The third trial by Chevalier and colleagues in 2008 provided a relative risk of 0.33 with 95 percent confidence interval of 0.12 to 0.93. The fourth trial by Ikari and colleagues, also in 2008, provided a relative risk of 0.64 with 95 percent confidence interval of 0.39 to 1.05. The fifth trial by Lee and colleagues in 2006 provided a relative risk of 1.64 with 95 percent confidence interval of 0.45 to 6.04. The sixth trial by Silva-Orrego and colleagues, also in 2006, provided a relative risk of 0.18 with 95 percent confidence interval of 0.05 to 0.70. The seventh trial by Burzotta and colleagues in 2005 provided a relative risk of 0.68 with 95 percent confidence interval of 0.22 to 2.11. The eight trial by Noel and colleagues, also in 2005, provided a relative risk of 0.31 with 95 percent confidence interval of 0.08 to 1.16.The combined effect of the eight trials showed a relative risk of 0.52 with a 95 percent confidence interval of 0.35 to 0.76. The Cochran Q p-value was 0.307, the I-squared value was 15.7 percent and the Egger's p-value was 0.278.

Figure 51

Impact of catheter aspiration devices versus control on no reflow in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.307 I2: 15.7 percent

When limiting the pooled analysis to only trials of good methodological quality,15,16,19,48,51,74,83,138 the risk of having no reflow remained significantly decreased in the catheter aspiration device group compared to control [RR 0.45 (0.27, 0.75)]. A lower level of statistical heterogeneity was detected (I2=22.3 percent). Given the risk difference [RD -0.08 (-0.12, -0.05), (CER 0.10 to 0.19)], thirteen people would have to be treated with a catheter aspiration device to prevent one no reflow event from occurring.

No controlled observational studies assessed for this endpoint in this population.

Catheter aspiration devices in other ACS populations. No trials or studies assessed for this endpoint in this population.

Mechanical thrombectomy devices in patients with STEMI. Three RCTs evaluated the impact of mechanical thrombectomy devices versus control on no reflow.29,40,44 The use of a mechanical thrombectomy device did not significantly impact the risk of no reflow [RR 0.50 (0.17, 1.48)] (Figure 52). A lower level of statistical heterogeneity was found (I2=41.7 percent). All of the trials were determined to be of good methodological quality. 29,40,44

This figure depicts the meta-analysis of the impact of mechanical thrombectomy devices versus control on no reflow. The first trial by Ali and colleagues in 2006 provided a relative risk of 1.21 with 95 percent confidence interval of 0.40 to 3.67. The second trial by Lefevre and colleagues in 2005 provided a relative risk of 0.31 with 95 percent confidence interval of 0.09 to 1.00. The third trial by Napodano and colleagues in 2003 provided a relative risk of 0.20 with 95 percent confidence interval of 0.03 to 1.23. The combined effect of the three trials showed a relative risk of 0.50 with a 95 percent confidence interval of 0.17 to 1.48. The Cochran Q p-value was 0.180, the I-squared value was 41.7 percent and there were too few strata to calculate the Egger's p-value.

Figure 52

Impact of mechanical thrombectomy devices versus control on no reflow in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.180 I2: 41.7 percent

No controlled observational studies assessed for this endpoint in this population.

Mechanical thrombectomy devices in other ACS populations. No RCTs or controlled observational studies assessed for this endpoint in this population.

Distal filter embolic protection devices in patients with STEMI. Two RCTs evaluated the impact of distal filter embolic protection devices versus control on no reflow.95,101 In these trials, the use of distal filter embolic protection devices did not significantly impact the risk of having no reflow [RR 0.59 (0.14, 2.51)] (Figure 53). Only one of these trials were determined to be of good methodological quality.95 In that trial there was no difference in the risk of no reflow with the use of a distal filter embolic protection device versus control [RR 1.00 (0.18, 5.55)].

This figure depicts the meta-analysis of the impact of distal filter embolic devices versus control on no reflow. The first trial by Cura and colleagues in 2007 provided a relative risk of 1.00 with 95 percent confidence interval of 0.18 to 5.55. The second trial by Lefevre and colleagues in 2004 provided a relative risk of 0.29 with 95 percent confidence interval of 0.04 to 1.92. The combined effect of the two trials showed a relative risk of 0.59 with a 95 percent confidence interval of 0.14 to 2.51. The Cochran Q p-value was 0.409 and there were too few strata to calculate the I-squared value and the Egger's p-value.

Figure 53

Impact of distal filter embolic protection devices versus control on no reflow in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.409 I2: Too few strata

No controlled observational studies assessed for this endpoint in this population.

Distal filter embolic protection devices in other ACS populations. One RCT evaluated the impact of distal filter embolic protection devices on no reflow156 in patients with NSTEMI or UA. In this trial, the Angioguard™ device was compared to control. The risk of no reflow could not be calculated because no events occurred in either group.

No controlled observational studies assessed for this endpoint in this population.

Distal balloon embolic protection devices in patients with STEMI. Four RCTs evaluated the impact of distal balloon embolic protection devices versus control on no reflow.103,107,110,112 The use of a distal balloon embolic protection device did not significantly impact the risk of no reflow [RR 0.51 (0.19, 1.33)] (Figure 54). Statistical heterogeneity and publication bias were not detected (I2=0 percent, Egger's P=0.880). All of the trials were determined to be of good methodological quality. 103,107,110,112

This figure depicts the meta-analysis of the impact of distal balloon embolic protection devices versus control on no reflow. The first trial by Hahn and colleagues in 2007 provided a relative risk of 1.05 with 95 percent confidence interval of 0.11 to 9.67. The second trial by Matsuo and colleagues, also in 2007, provided a relative risk of 1.39 with 95 percent confidence interval of 0.28 to 6.82. The third trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 0.32 with 95 percent confidence interval of 0.08 to 1.38. The fourth trial by Stone and colleagues in 2005 provided a relative risk of 0.17 with 95 percent confidence interval of 0.03 to 1.06. The combined effect of the four trials showed a relative risk of 0.51 with a 95 percent confidence interval of 0.19 to 1.33. The Cochran Q p-value was 0.403, the I-squared value was zero percent and the Egger's p-value was 0.880.

Figure 54

Impact of distal balloon embolic protection devices versus control on no reflow in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.403 I2: 0 percent

Distal balloon embolic protection devices in other ACS populations. One RCT evaluated the impact of the distal balloon embolic protection device PercuSurge GuardWire™ Plus Temporary Occlusion and Aspiration System versus control on no reflow in patients with acute myocardial infarction.125 The use of a distal balloon embolic protection device significantly decreased the risk of no reflow compared to control [RR 0.36 (0.20, 0.59)]. Given the risk difference for no reflow [RD -0.54 (-0.71, -0.31), CER 0.02 to 0.05)], two people would need to be treated with a distal balloon embolic protection device to prevent one person from experiencing no reflow. This trial was determined to be of good methodological quality.

No controlled observational studies assessed for this endpoint in this population.

Proximal balloon embolic protection devices in patients with STEMI. No trials or studies evaluated the impact of proximal balloon embolic protection devices on this endpoint.

Proximal balloon embolic protection devices in other ACS populations. No trials or studies assessed for this endpoint in this population.

Embolic protection devices combined in patients with STEMI. Six RCTs evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) versus control on no reflow.95,101,103,107,112,133 In these trials, the use of embolic protection devices combined did not significantly decreased the risk of having no reflow [RR 0.53 (0.24, 1.18)] (Figure 55). Statistical heterogeneity and publication bias were not detected (I2=0 percent, Egger's P=0.969). When limiting the analysis to only trials of good methodological quality95,103,107,112,133 the risk of no reflow remain nonsignificant in the embolic protection devices combined group versus control [(RR 0.58 (0.25, 1.37)]. No statistical heterogeneity was found (I2=0 percent).

This figure depicts the meta-analysis of the impact of embolic protection devices combined versus control no reflow. The first trial by Cura and colleagues in 2007 provided a relative risk of 1.00 with 95 percent confidence interval of 0.18 to 5,55. The second trial by Hahn and colleagues, also in 2007, provided a relative risk of 1.05 with 95 percent confidence interval of 0.11 to 9.67. The third trial by Matsuo and colleagues, also in 2007, provided a relative risk of 1.39 with 95 percent confidence interval of 0.28 to 6.82. The fourth trial by Muramatsu and colleagues, also in 2007, provided a relative risk of 0.32 with 95 percent confidence interval of 0.08 to 1.38. The fifth trial by Stone and colleagues in 2005 provided a relative risk of 0.17 with 95 percent confidence interval of 0.03 to 1.06. The sixth trial by Lefevre and colleagues in 2004 provided a relative risk of 0.29 with 95 percent confidence interval of 0.04 to 1.92. The combined effect of the four trials showed a relative risk of 0.53 with a 95 percent confidence interval of 0.24 to 1.18. The Cochran Q p-value was 0.603, the I-squared value was zero percent and the Egger's p-value was 0.969.

Figure 55

Impact of embolic protection devices combined versus control on no reflow in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.603 I2: 0 percent

Embolic protection devices combined in other ACS populations. No trials or studies were available that evaluated the impact of any embolic protection device versus control on no reflow in this population in addition to the two trials reported above. Pooling was not suitable because the trials evaluated different ACS.

Discussion

While there were a number of controlled trials where patients undergoing PCI were treated with a thrombectomy or embolic protection device plus standard of care therapy or standard of care therapy alone, the duration of followup, the time points at which they evaluated events and the number of times they evaluated events also varied considerably between trials. For our base case analysis, we used the maximum duration of followup to allow the pooling of a greater number of individual studies. However, we also evaluated for the shortest duration of followup within a trial and at several durational ranges specified a priori. As such, we sought to determine if the effects seen during the maximal duration of followup was representative of the results derived using other time frames. Although several systematic reviews have conducted meta-analysis in the past, the majority are limited to patients with STEMI and did not evaluate adjunctive devices in other ACS and the most recent analyses did not evaluate embolic protection devices. Therefore, applicability of those results to contemporary practice is limited.

Final health outcomes in patients with STEMI. In patients with STEMI, the impact of catheter aspiration devices was directly compared to distal balloon embolic protection devices on final health outcomes in only one direct comparative RCT. In this trial, no significant differences in mortality, myocardial infarction, stroke, target revascularization, or MACE were found at the longest duration of followup. Given limited direct comparative trial data, the superiority of one device over another cannot be directly determined.

In patients with STEMI, the use of catheter aspiration devices significantly reduced the occurrence of MACE versus standard of care by 27 percent using the maximum duration of followup (12.43 months). To prevent one major adverse cardiovascular events, 33 patients would need to be treated with a catheter aspiration device. Using other time cut-offs, the directionality of effect for MACE was similar but statistical significance was only maintained at the 180-days evaluation (studies reporting MACE outcomes from in-hospital to 365 days after the procedure). When we assessed individual components of MACE (mortality, myocardial infarction, or target revascularization) using the maximal duration of followup, no significant reductions were found and limiting the pooled analyses to trials of good quality did not impact the significance of results. When other time periods were assessed, mortality was significantly reduced by 38 percent at the 365-days, target revascularization was significantly reduced by 38 percent at 180-days, and MACE was significantly reduced by 34 percent at 180-days. While there was a significant reduction in MACE with catheter aspiration devices versus control, there is a nonsignificant three-fold increase in the risk of developing stroke using the maximum duration of followup (0.79 months). The direction of effect suggests an increased risk of stroke with catheter aspiration devices regardless of the time point chosen.

In patients with STEMI, the use of mechanical thrombectomy devices did not significantly impact the risk of mortality, myocardial infarction, stroke, target revascularization or MACE at the maximal duration of followup although a higher level of statistical heterogeneity was found in the mortality, target revascularization, and MACE analyses. Like with the catheter aspiration analyses at various time periods, a significant reduction in the risk of 365-day MACE and 180-day target revascularization was found with mechanical aspiration device use versus control. All of the trials included in the pooled analyses were determined to be of higher methodological quality therefore sensitivity analyses based on trial methodological quality did not reduce the observed heterogeneity or impact the overall results. When evaluating each final health outcome by individual time point; statistical heterogeneity could not longer be evaluated in most cases because too few studies were left to evaluate. Therefore it is difficult to say whether the inclusion of various time points in the pooled analysis of final health outcomes contributed to the higher level of statistical heterogeneity when evaluating mechanical thrombectomy devices.

Given this data, we could not make any determinations as to whether catheter aspiration or mechanical thrombectomy are superior strategies versus standard of care or whether one type of device is superior to another.

In patients with STEMI, the use of distal filter, distal balloon, proximal balloon or embolic protection devices combined (distal or proximal; filter or balloon) did not significantly impact the risk of mortality, myocardial infarction, stroke, or MACE at the maximal duration of followup versus control. The risk of target revascularization was significantly increased with the use of distal filter embolic protection devices or embolic protection devices combined versus control using the maximal duration of followup, although this was not seen with the other embolic protection device classes. Pooled analyses of final health outcomes at individual time points were limited within the distal filter and balloon embolic protection device categories because of the few number of trials reporting these outcomes and the rare occurrence of events in the trials which did report results. Therefore, the majority of individual time points could not be evaluated in these device categories or risk was based on a single trial. No significant findings were observed, with few exceptions. Distal filter embolic protection devices and any embolic protection device significantly increased the risk of target revascularization at 365 days (1 trial each) and of MACE at 365 days (1 trial each) versus control. A significant reduction in the risk of 30 day stroke was seen when distal balloon embolic protection devices were compared to control. Pooling of results for proximal balloon embolic protection devices was not possible since only one trial was available with reported outcomes. Final health outcomes were reported at 30 and 180 days were nonsignificant for all analyses. Within any embolic protection device category (distal or proximal; filter or balloon), limiting the pooled analyses to trials determined to be of higher methodological quality did not change the direction or significance of the results pertaining to any of the final health outcomes.

Given this data, we could not make any determinations as to whether one embolic protection device category is a superior strategy versus standard of care or whether one type of device is superior to another, or to catheter aspiration or mechanical thrombectomy devices.

Final health outcomes in patients with other ACSs. In patients with mixed ACS (STEMI, NSTEMI, or UA) trials were identified evaluating the impact of four device categories (catheter aspiration, mechanical thrombectomy, distal filter and distal balloon embolic protection devices) on final health outcomes. Overall data was very limited and only trials evaluating distal balloon and embolic protection devices combined were amenable to pooling. Additionally, the range of time points at which final health outcomes were reported made comparison across device categories difficult. No significant differences were found between any of the device categories and control on all of the final health outcomes. Overall, making comparisons across device categories or within device categories comparing various time points for a single outcome is difficult given the limited number of trials and studies in patients with mixed ACS.

In patients with NSTEMI or UA a limited number of studies which evaluated thrombectomy or embolic protections devices were identified. Two RCTs which evaluated the impact distal filter embolic protection devices versus control on final health outcomes using the maximal duration of followup were identified although were not amenable to pooling. Of the five final health outcomes, MACE and mortality were reported with results which could be evaluated, although no significant difference between the distal filter embolic protection devices and control were found. No other studies or trials were found in this patient population for the other device categories.

Intermediate health outcomes. In patients with STEMI, the impact of catheter aspiration devices was directly compared to distal balloon embolic protection devices on intermediate health outcomes in a single direct comparative RCT. In this trial, none of the intermediate health outcomes reached statistical significance. No other trials or studies were found to directly compare device categories on their impact on final health outcomes.

In patients with STEMI, the use of a catheter aspiration device significantly improved intermediate health outcomes, including resolution of ST-segment elevation, achievement of MBG-3 and TIMI-3 blood flow, and reduction in distal embolization and no reflow. However, the use of a catheter aspiration device does not appear to significantly impact ejection fraction versus control. Although not amenable to pooling, the majority of trials which evaluated ejection fraction showed no significant differences (9 of the 11 trials) and these trials evaluated ejection fraction within a wide range of time points including immediately postPCI up to 6 months postPCI. The use of mechanical thrombectomy devices did not significantly impact any of the intermediate health outcomes. In a controlled observational study, the use of a mechanical thrombectomy device significantly decreased the rate of TIMI-3 blood flow versus control. Although not amenable to pooling, in the two trials which evaluated the impact of mechanical thrombectomy devices on ejection fraction versus control, no significant differences were seen. Overall, it appears that the use of catheter aspiration devices more favorably impacts intermediate health outcomes than the use of mechanical thrombectomy devices, although this is based on indirect comparisons.

Distal filter embolic protection devices, distal balloon embolic protection devices, and embolic protection devices combined did not have significant impact on most intermediate health outcomes versus control. A single trial evaluating the impact of proximal balloon embolic protection devices versus control was identified therefore pooling was not possible. The significant findings included the impact of distal balloon and embolic protection devices combined both significantly increasing the risk of achieving a MBG-3 and TIMI-3 blood flow. In the evaluation of ejection fraction, data was not amenable to pooling. The impact of distal balloon and distal filter embolic protection device on ejection fraction versus control was reported, although only one trial evaluating distal balloon embolic protection devices found a significantly higher ejection fraction versus control at 90 and 180 days.

In patients with mixed ACS (STEMI, NSTEMI, or UA) RCTs sparsely reported intermediate health outcomes comparing thrombectomy or embolic protection devices versus control and most data was not amenable to pooling. One RCT demonstrated a significant increase in the risk of resolving ST-segment resolution with the use of mechanical thrombectomy devices versus control. No other trials evaluated any device categories on ST-segment resolution. Mixed results were observed in the evaluation of the risk of attaining TIMI-3 blood flow. Pooled results evaluating the impact of catheter aspiration devices, distal balloon embolic protection devices, or embolic protection devices combined did not show a significant difference versus control. Both in the evaluation of catheter aspiration devices and distal balloon embolic protection devices, a significant increase in the risk of attaining MBG-3 was seen, although only the analysis of distal balloon embolic protection devices was based on a pooled analysis. A significant reduction in the risk of no reflow in the distal balloon embolic protection device versus control was noted. One trial reported the impact of distal filter embolic protection devices on ejection fraction at 3 days versus control, and no significant change was seen.

In patients with NSTEMI or UA, limited data was available regarding the impact of thrombectomy or embolic protection devices versus control on intermediate health outcomes. The use of distal filter embolic protection devices did not significantly impact the risk of attaining TIMI-3 blood flow. No other trials or studies evaluated other device categories or other intermediate health outcomes, therefore the impact of thrombectomy or embolic protection devices versus control on intermediate health outcomes in this population is difficult to evaluate.

Key Question 2. In patients with ACS who are undergoing PCI of native vessels, how does the rate and type of adverse events (e.g., coronary dissection, coronary perforation, prolonged procedure time) differ between device types when compared to PCI alone?

Key Points

Twenty three RCTs and two controlled observational studies were included.

Direct Comparative Trials in ACS Patients Assessing Adverse Outcomes

  • Two direct comparative randomized trials in patients with STEMI undergoing PCI evaluated adverse outcomes.
    • One direct comparative randomized trial compared a catheter aspiration device to another catheter aspiration device. In this trial, the use of one catheter aspiration device versus another did not significantly impact the risk of coronary dissection. No patients experienced coronary perforation in either group.
    • One direct comparative randomized trial compared a catheter aspiration device to a distal balloon embolic protection device. In this trial, the use of a catheter aspiration device did not impact procedure time compared to a distal balloon embolic protection device.
    • No direct comparative trials evaluated side branch occlusion.

RCTs / Controlled Observational Studies in Patients with STEMI Assessing Adverse Outcomes

  • Twenty RCTs and three controlled observational studies evaluated patients with STEMI undergoing PCI and compared a thrombectomy or embolic protection device versus control. Four adverse events (coronary dissection, coronary perforation, prolonged procedure time, and side branch occlusion) were evaluated.
    • In RCTs eligible for pooling, the use of catheter aspiration devices versus control significant reduced the risk of coronary dissection and did not significantly impact the risk of side branch occlusion. In the one trial in which coronary perforation was assessed, no events occurred in either group. Nine trials evaluated procedure time although were ineligible for pooling. In eight of the nine trials the use of catheter aspiration devices versus control did not significantly prolong procedure time. One controlled observational study found no significant difference in procedure time between catheter aspiration and control.
      • When limited to good quality trials, catheter aspiration device use still reduced the risk of coronary dissection with nonsignificant effects on the other aforementioned adverse events.
      • One controlled observational study found no significant impact of catheter aspiration devices on the risk of coronary dissection versus control.
    • In RCTs, the use of mechanical thrombectomy devices versus control did not significantly impact the risk of coronary dissection, coronary perforation, or side branch occlusion. Three trials evaluated the impact of mechanical thrombectomy devices versus control on procedure time although were ineligible for pooling. In all three trials the procedure time was significantly prolonged in the mechanical thrombectomy device group versus control.
      • When limited to good quality trials, significant increases in procedural time and nonsignificant effects on the risk of coronary dissection, coronary perforation, or side branch occlusion occurred.
      • One controlled observational study found no significant impact of mechanical thrombectomy devices on the risk of coronary perfusion versus control.
    • In RCTs, the use of distal filter embolic protection devices versus control did not significantly impact the risk of side branch occlusion. No coronary dissections and coronary perforations occurred in either group in the one trial reporting these outcomes. Use of a distal filter embolic protection device increased the procedure time versus control in the one trial evaluating this outcome.
      • Limiting to good quality trials yielded the same results.
      • No controlled observational studies were available.
    • In RCTs, the use of distal balloon embolic protection devices versus control did not significantly impact the risk of coronary perforation or side branch occlusion. One trial evaluated the impact of distal balloon embolic protection devices versus control on coronary dissection although no events occurred in either group. Three trials evaluated the impact of distal balloon embolic protection devices versus control on procedure time although were not amenable to pooling. In two of the three trials, procedure time was significantly prolonged with the use of a distal balloon embolic protection device versus control.
      • Limiting to good quality trials yielded the same results.
      • No controlled observational studies were available.
    • In a RCT, the use of a proximal balloon embolic protection device versus control significantly prolonged procedure time. No other trials or studies evaluated the impact of proximal balloon embolic protection devices versus control on adverse events of interest.
      • Limiting to good quality trials yielded the same results.
      • No controlled observational studies were available.
    • In RCTs eligible for pooling, the use of an embolic protection device (distal or proximal; filter or balloon) did not significantly impact the risk of side branch occlusion. In a single trial, the use of an embolic protection device did not significantly impact the risk of coronary perforation versus control. The risk of coronary dissection could not be calculated in the single trial which reported this outcome. Five RCTs evaluated the impact of an embolic protection device on procedure time although were ineligible for pooling. In four of the five trials procedure time was significantly prolonged with the use of an embolic protection device versus control.

RCTs / Controlled Observational Studies in Mixed or Other ACS Populations Assessing Adverse Outcomes

  • One RCT evaluated patients with mixed ACS (STEMI, NSTEMI, or UA) undergoing PCI and comparing thrombectomy or embolic protection devices versus control on adverse events.
    • In a RCT, the use of a distal balloon embolic protection device versus control significantly prolonged the procedure time.
    • No other trials or studies evaluated other device categories or adverse events.
  • No trials or studies evaluating patients with NSTEMI or UA undergoing PCI and comparing catheter aspiration, mechanical thrombectomy, or embolic protection devices versus control on adverse events were identified.

Detailed Analysis

Study Design and Population Characteristics

The study design and population characteristic have been previously described in key question one. Although several systematic reviews have conducted meta-analyses in the past, the majority are limited to patients with STEMI and did not evaluate adjunctive devices in other ACS, only two were identified to evaluate adverse events limited to procedure time and coronary perforation, and the most recent analyses did not evaluate embolic protection devices. Therefore, applicability of those results to contemporary practice is limited.

Specific to key question two, we present direct comparative data between agents first and subsequently present the comparisons of each type of device versus control for each endpoint.

Outcome Evaluation

A summary of the results for adverse events comparing each device category to control can be found in Table 27 to Table 32.

Table 27. Adverse events in randomized controlled trials evaluating catheter aspiration devices in patients with ST-segment elevation myocardial infarction.

Table 27

Adverse events in randomized controlled trials evaluating catheter aspiration devices in patients with ST-segment elevation myocardial infarction.

Table 28. Adverse events in randomized controlled trials evaluating mechanical thrombectomy devices in patients with ST-segment elevation myocardial infarction.

Table 28

Adverse events in randomized controlled trials evaluating mechanical thrombectomy devices in patients with ST-segment elevation myocardial infarction.

Table 29. Adverse events in randomized controlled trials evaluating distal filter embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 29

Adverse events in randomized controlled trials evaluating distal filter embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 30. Adverse events in randomized controlled trials evaluating distal balloon embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 30

Adverse events in randomized controlled trials evaluating distal balloon embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 31. Adverse events in randomized controlled trials evaluating proximal balloon embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 31

Adverse events in randomized controlled trials evaluating proximal balloon embolic protection devices in patients with ST-segment elevation myocardial infarction.

Table 32. Adverse events in randomized controlled trials evaluating embolic protection devices combined in patients with ST-segment elevation myocardial infarction.

Table 32

Adverse events in randomized controlled trials evaluating embolic protection devices combined in patients with ST-segment elevation myocardial infarction.

Coronary Dissection

Direct Comparative Trials

Catheter aspiration device versus catheter aspiration device in patients with STEMI. One direct comparative randomized trial evaluated the impact of the Diver™-Invatec catheter aspiration device versus the Export®-Medtronic catheter aspiration device on coronary dissection.158 In this trial, the use of Diver™-Invatec did not significantly impact the risk of coronary dissection [RR 0.33 (0.00, 3.71)] compared to Export®-Medtronic. This trial was determined to be of good methodological quality.

Trials Versus Control

Catheter aspiration devices in patients with STEMI. Five RCTs evaluated the impact of catheter aspiration devices on coronary dissection versus control.16,62,68,69,74 The use of catheter aspiration devices significantly decreased the risk of coronary dissection [RR 0.30 (0.12, 0.75)] (Figure 56). No statistical heterogeneity or publication bias was found (I2=0 percent, Egger's P=0.626). All of the trials were determined to be of good methodological quality. 16,62,68,69,74 Given the risk difference [RD -0.02 (-0.12, 0.10), (CER 0.0 to 0.1)], 50 people would need to be treated with a catheter aspiration device to prevent one person from experiencing a coronary dissection.

This figure depicts the meta-analysis of the impact of catheter aspiration devices on coronary dissection versus control. The first trial by by Chao and colleagues in 2008 provided a relative risk of 0.33 with a 95 percent confidence interval of 0.00 to 3.76. The second trial by Ikari and colleagues, also in 2008, provided a relative risk of 0.21 with 95 percent confidence interval of 0.09 to 0.72. The third trial by Svilaas and colleagues, also in 2008, was excluded from the meta-analysis because no coronary dissection occurred in either the treatment or the control group and therefore a relative risk could not be calculated. The fourth trial by De Luca and colleagues in 2006 provided a relative risk of 3.00 with a 95 percent confidence interval of 0.27 to infinity. The fifth trial by Silva-Orrego and colleagues, also in 2006, provided a relative risk of 0.11 with a 95 percent confidence interval of 0.00 to 0.94. The combined effect of the four trials showed a relative risk of 0.30 with a 95 percent confidence interval of 0.12 to 0.75. The Cochran Q p-value was 0.464, the I-squared value was zero percent and the Egger's p-value was 0.626.

Figure 56

Impact of catheter aspiration devices on coronary dissection versus control in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.464 I2: 0 percent

One controlled observational study evaluated the association between the use of catheter aspiration devices during PCI and coronary dissection versus control.152 The names of the catheter aspiration devices included in this study were not reported. The use of a catheter aspiration device during PCI was not associated with a significantly different rate of coronary dissection compared to PCI without the use of a catheter aspiration device (6.6 percent versus 5.3 percent, p=0.32).

Catheter aspiration devices in other ACS populations. No trials or studies evaluated the impact of catheter aspiration devices on this outcome.

Mechanical thrombectomy devices in patients with STEMI. One RCT evaluated the impact of a mechanical thrombectomy device on coronary dissection versus control.40 The use of a mechanical thrombectomy device did not significantly impact the risk of coronary dissection [RR 1.51 (0.57, 4.01)]. This trial was determined to be of good methodological quality. 40

No controlled observational studies evaluated this endpoint in this population.

Mechanical thrombectomy devices in other ACS populations. No trials or studies evaluated the impact of mechanical thrombectomy devices on this outcome.

Distal filter embolic protection devices in patients with STEMI. One RCT evaluated the impact of distal filter embolic protection devices on coronary dissection versus control.95 The risk of coronary dissection could not be calculated because no events occurred in either control or treatment group. This trial was determined to be of good methodological quality.

No controlled observational studies evaluated this endpoint in this population.

Distal filter embolic protection devices in other ACS populations. No trials or studies evaluated the impact of distal filter embolic protection devices on this outcome.

Distal balloon embolic protection devices in patients with STEMI. One RCT evaluated the impact of distal balloon embolic protection devices on coronary dissection versus control.111 The risk of coronary dissection could not be calculated because no events occurred in either control or treatment group. This trial was determined to be of good methodological quality.

No controlled observational studies evaluated this endpoint in this population.

Distal balloon embolic protection devices in other ACS populations. No trials or studies evaluated the impact of distal balloon embolic protection devices on this outcome.

Proximal balloon embolic protection devices in patients with STEMI. No trials or studies evaluated the impact of proximal balloon embolic protection devices on this outcome.

Proximal balloon embolic protection devices in other ACS populations. No trials or studies evaluated the impact of proximal balloon embolic protection devices on this outcome.

Embolic protection devices combined in patients with STEMI. Two RCTs evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) on coronary dissection versus control.95,111 In these two trials, the risk could not be calculated because no events occurred in either control or treatment group. Both trials were determined to be of good methodological quality.

No controlled observational studies evaluated this endpoint in this population.

Embolic protection devices combined in other ACS populations. No trials or studies evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) on this outcome.

Coronary Perforation

Direct Comparative Trials

Catheter aspiration devices versus catheter aspiration device in patients with STEMI. One direct comparative randomized trial evaluated the impact of the Diver™-Invatec catheter aspiration device versus the Export®-Medtronic catheter aspiration device on coronary perforation.158 The risk of coronary perforation could not be calculated because no events occurred in either group during this trial. This trial was determined to be of good methodological quality.

Trials Versus Control

Catheter aspiration devices in patients with STEMI. One RCT evaluated the impact of using a catheter aspiration device on coronary perforation versus control.16 The risk of coronary perforation could not be calculated because no events occurred in either control or treatment group.

No controlled observational studies evaluated this endpoint in this population.

Catheter aspiration devices in other ACS populations. No trials or studies evaluated the impact of catheter aspiration devices on this outcome.

Mechanical thrombectomy devices in patients with STEMI. Two RCTs evaluated the impact of mechanical thrombectomy devices on coronary perforation versus control.11,40 The use of mechanical thrombectomy devices did not significantly impact the risk of coronary perforation [RR 1.04 (0.15, 7.04)] (Figure 57). Publication bias could not be evaluated since only two studies were available. Both trials were determined to be of good methodological quality. 11,40

This figure depicts the meta-analysis of the impact of mechanical thrombectomy devices on coronary perforation versus control. The first trial by Migliorini and colleagues in 2010 provided a relative risk of 0.32 with 95 percent confidence interval of 0.00 to 3.67. The second trial by Ali and colleagues in 2008 provided a relative risk of 2.01 with 95 percent confidence interval of 0.26 to 15.27. The combined effect of the two trials showed a relative risk of 1.04 with a 95 percent confidence interval of 0.15 to 7.04. The Cochran Q p-value was 0.366 and there were too few strata to calculate the I-squared value and Egger's p-value.

Figure 57

Impact of mechanical thrombectomy devices on coronary perforation versus control in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.366 I2: Too few strata

One controlled observational study evaluated the association between the use of a mechanical thrombectomy device and coronary perforation versus control.145 Patients undergoing PCI with a mechanical thrombectomy device, either the AngioJet® XMI or XVG catheter, were compared to patients undergoing PCI without mechanical thrombectomy. The use of a mechanical thrombectomy device was not associated with a significant difference in the rate of coronary perforation compared to PCI without a mechanical thrombectomy device (0.0 percent versus 0.2 percent, p>0.99).

Mechanical thrombectomy devices in other ACS populations. No trials or studies evaluated the impact of mechanical thrombectomy devices on this outcome.

Distal filter embolic protection devices in patients with STEMI. One RCT evaluated the impact of a distal filter embolic protection device on coronary perforation versus control.95 The risk of coronary perforation could not be calculated because no events occurred in either control or treatment group.

No controlled observational studies evaluated this endpoint in this population.

Distal filter embolic protection devices in other ACS populations. No trials or studies evaluated the impact of distal filter embolic protection devices on this outcome.

Distal balloon embolic protection devices in patients with STEMI. Two RCTs evaluated the impact of distal balloon embolic protection devices on coronary perforation versus control.111,112 In one trial no events occurred in either the control or treatment group.111 In the other trial the use of a distal balloon embolic protection device did not significantly impact the risk of coronary perforation [RR 5.11 (0.53, infinity)]. Publication bias could not be calculated. Both trials were determined to be of good methodological quality.111,112

No controlled observational studies evaluated this endpoint in this population.

Distal balloon embolic protection devices in other ACS populations. No trials or studies evaluated the impact of distal balloon embolic protection devices on this outcome.

Proximal balloon embolic protection devices in patients with STEMI. No trials or studies evaluated the impact of proximal balloon embolic protection devices on this outcome.

Proximal balloon embolic protection devices in other ACS populations. No trials or studies evaluated the impact of proximal balloon embolic protection devices on this outcome.

Embolic protection devices combined in patients with STEMI. Three RCTs evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) on coronary perforation versus control.95,111,112 In two trials no events occurred in either control or treatment group.95,111 In another trial,112 the use of an embolic protection device did not significantly impact the risk of coronary perforation [RR 5.11 (0.53, infinity)]. All of the trials were determined to be of good methodological quality.95,111,112

No controlled observational studies evaluated this endpoint in this population.

Embolic protection devices combined in other ACS populations. No trials or studies evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) on this outcome.

Prolonged Procedure Time

Direct Comparative Trials

Catheter aspiration device versus catheter aspiration device in patients with STEMI. One direct comparative randomized trial evaluated the impact of the Diver™ CE catheter aspiration device versus the Guardwire™ Plus distal balloon embolic protection device on procedure time.160 In this trial, there was no significant difference in procedure time between the Diver™ CE and Guardwire™ Plus groups (60 min ±24 versus 65 min ±28, p=0.36). This trial was determined to be of good methodological quality.

Trials Versus Control

Catheter aspiration devices in patients with STEMI. Nine RCTs evaluated the impact of catheter aspiration devices on procedure time versus control but were not amenable to pooling.12,15,16,62,68,71,74,83,162 In the first trial, the mean procedure time was not significantly different between the catheter aspiration device group and control (75.7±33.0 min versus 75.9±38.7 min, p=0.90).12 In the second trial, patients were only included in the trial if they achieved a TIMI-3 blood flow postprocedure.162 The mean procedure time was not significantly different between the catheter aspiration device group and control (39.5±10.1 min versus 32.3±18.6 min, p=0.14). 162 In the third trial, the mean procedure time was not significantly different between the catheter aspiration device group and control (36.7±18.0 min versus 34.5±21.5 min, p=0.08).15 In the fourth trial, the mean procedure time was not significantly different between the catheter aspiration device group and control (87.0±32.4 min versus 93.6±78.6 min, p=0.16).16 In the fifth trial, the median procedural time was not significantly different between the catheter aspiration device group and control [28 min (14-42) versus 26 min (12-40), p=0.92].62 In the sixth trial, procedure time (defined as lab to TIMI-3 blood flow time) was not significantly different between the catheter aspiration device group and control (49±18 min versus 53±23 min, p=0.54).68 In the seventh trial, the median procedural time was significantly prolonged in the catheter aspiration device group compared to control [39 minutes (29-48) versus 29 minutes (23-38), p<0.0001]. 71 In the eighth trial, the mean procedure time was not significantly different between the catheter aspiration device group and control (57±19 minutes versus 54±21 minutes, p=0.36).74 In the final trial, the mean procedure time was not significantly different between the catheter aspiration device group and control (81±43 minutes versus 72±34 minutes, p=0.41).83 All included trials were determined to be of good methodological quality.

One controlled observational study evaluated the impact of catheter aspiration versus control on procedure time.144 In this study, the use of a catheter aspiration device was not associated with a prolonged procedure time versus control (41.2 minutes versus 36.5 minutes, p=0.12).144

Catheter aspiration devices in other ACS populations. No trials or studies evaluated the impact of mechanical thrombectomy devices on this outcome.

Mechanical thrombectomy devices in patients with STEMI. Three RCTs evaluated the impact of mechanical thrombectomy devices on procedure time versus control although were not amenable to pooling.11,27,29,40 In the first trial, the median procedure time was significantly prolonged in the mechanical thrombectomy device group compared to control [59.5 minutes (45-70) versus 46 minutes (35-60), p<0.001].11 In the second trial, the mean procedure time was significantly prolonged in the mechanical thrombectomy device group compared to control (75.4±30.9 minutes versus 59.2±26.8 minutes), p<0.001).40 In the third trial, the mean procedure time was significantly prolonged in the mechanical thrombectomy device group compared to control (54±28 minutes versus 45±25 minutes, p=0.009).29 All three trials were determined to be of good methodological quality.

No controlled observational studies evaluated this endpoint in this population.

Mechanical thrombectomy devices in other ACS populations. No trials or studies evaluated the impact of mechanical thrombectomy devices on this outcome.

Distal filter embolic protection devices in patients with STEMI. One RCT evaluated the impact of distal filter embolic protection devices on procedure time versus control.95 In this trial, the SpideRX™ device was used. The median procedure time was significantly prolonged in the distal filter embolic protection device group compared to control [52 minutes (43-70) versus 43.5 minutes (30-54), p<0.001]. This trial was determined to be of good methodological quality.

No controlled observational studies evaluated this endpoint in this population.

Distal filter embolic protection devices in other ACS populations. One RCT evaluated the impact of distal filter embolic protection devices on procedure time.156 The mean procedure time was only reported for the device group (63 minutes ±17).

No controlled observational studies evaluated this endpoint in this population.

Distal balloon embolic protection devices in patients with STEMI. Three RCT evaluated the impact of distal balloon embolic protection devices on procedure time versus control although were not amenable to pooling.107,112,133 In the first trial, mean procedure time was significantly prolonged in the distal balloon embolic protection device group compared to control (75.8±30 minutes versus 53±25 minutes, p<0.01).107 In the second trial, the mean procedure time was not significantly different between the distal balloon embolic protection device group and control (29.7±18.3 minutes versus 29.5±18.2 minutes, p=0.91).133 In the third trial the median procedure time was significantly prolonged in the distal balloon embolic protection device group compared to control [53 minutes (42-69) versus 39 minutes (29-51), p<0.001].112 This trial was determined to be of good methodological quality.

One RCT evaluated the impact of distal balloon embolic protection devices on procedure time versus abciximab therapy.164 In this trial, the PercuSurge device was used. The median procedure time was not significantly different between the distal balloon embolic protection device group and the abciximab group [58 minutes (35-88) versus 43 minutes (25-87), p=NS].164 This trial was determined to be of good methodological quality.

No controlled observational studies evaluated this endpoint in this population.

Distal balloon embolic protection devices in other ACS populations. One RCT evaluated the impact of distal balloon embolic protection devices on procedure time in patients with acute myocardial infarction.125 In this trial, the GuardWire® device was used. The mean procedure time was significantly prolonged in the distal balloon embolic protection device group compared to control (25.01 minutes ±11.89 versus 31.98 minutes ±15.33, p=0.03). This trial was determined to be of good methodological quality.

No controlled observational studies evaluated this endpoint in this population.

Proximal balloon embolic protection devices in patients with STEMI. One RCT evaluated the impact of proximal balloon embolic protection devices on procedure time versus control.18 In this trial, the Proxis™ device was used. The median procedure time was significantly prolonged in the proximal balloon embolic protection device group compared to control [45 minutes (36-58) versus 31 minutes (25-40), p<0.01].18 This trial was determined to be of good methodological quality.

No controlled observational studies evaluated this endpoint in this population.

Proximal balloon embolic protection devices in other ACS populations. No trials or studies evaluated the impact of proximal balloon embolic protection devices on this outcome.

Embolic protection devices combined in patients with STEMI. Five RCTs evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) on prolonged procedure time versus control although were not amenable to pooling.18,95,107,112,133 The procedure time results have been reported in each of the respective embolic protection device categories above. No additional data was available. In four of the five trials, the procedure time was significantly prolonged in the embolic protection device group versus control.18,95,107,112

Embolic protection devices combined in other ACS populations. One RCT evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) on prolonged procedure time versus control125 whose results are reported under distal balloon embolic protection devices in other ACS populations. No additional data was available.

Side Branch Occlusion

Direct Comparative Trials

No direct comparative trials evaluated the impact of thrombectomy or embolic protection devices on this outcome.

Trials Versus Control

Catheter aspiration devices in patients with STEMI. Two RCTs evaluated the impact of catheter aspiration devices on side branch occlusion versus control.15,62 The use of catheter aspiration devices did not significantly impact the risk of side branch occlusion [RR 1.19 (0.40, 3.54)] (Figure 58). Publication bias could not be calculated since only two studies were available. Both of the trials were determined to be of good methodological quality.15,62

This figure depicts the meta-analysis of the impact of catheter aspiration devices on side branch occlusion versus control. The first trial by Chevalier and colleagues in 2008 provided a relative risk of 1.08 with 95 percent confidence interval of 0.19 to 6.02. The second trial by Svilaas and colleagues, also in 2008, provided a relative risk of 1.25 with 95 percent confidence interval of 0.37 to 4.29. The combined effect of the two trials showed a relative risk of 1.19 with a 95 percent confidence interval of 0.40 to 3.54. The Cochran Q p-value was 0.898 and there were too few strata to calculate the I-squared value and Egger's p-value.

Figure 58

Impact of catheter aspiration devices on side branch occlusion versus control in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.898 I2: Too few strata

No controlled observational studies evaluated this endpoint in this population.

Catheter aspiration devices in other ACS populations. No trials or studies evaluated the impact of catheter aspiration devices on this outcome.

Mechanical thrombectomy devices in patients with STEMI. One RCT evaluated the impact of a mechanical thrombectomy device on side branch occlusion versus control.44 In this trial, the use of a mechanical thrombectomy device did not impact the risk of side branch occlusion versus control [RR 1.00 (0.11, 9.41)]. This trial was determined to be of good methodological quality.44

No controlled observational studies evaluated this endpoint in this population.

Mechanical thrombectomy devices in other ACS populations. No studies evaluated the impact of mechanical thrombectomy devices on this outcome.

Distal filter embolic protection devices in patients with STEMI. One RCT evaluated the impact of a distal filter embolic protection device on side branch occlusion versus control.95 In this trial, the use of a distal filter embolic protection device did not significantly impact the risk of side branch occlusion versus control [RR 0.33 (0.00, 3.80)]. This trial was determined to be of good methodological quality. 95

Distal filter embolic protection devices in other ACS populations. No trials or studies evaluated the impact of distal filter embolic protection devices on this outcome.

Distal balloon embolic protection devices in patients with STEMI. Two RCTs evaluated the impact of distal balloon embolic protection devices on side branch occlusion.107,112 The use of distal balloon embolic protection devices did not significantly impact the risk of side branch occlusion versus control [RR 0.93 (0.61, 1.42)] (Figure 59). Publication bias could not be calculated since only two studies were available. Both trials were determined to be of good methodological quality. 107,112

This figure depicts the meta-analysis of the impact of distal balloon embolic protection devices on side branch occlusion versus control. The first trial by Matsuo and colleagues in 2007 provided a relative risk of 1.85 with 95 percent confidence interval of 0.25 to 13.97. The second trial by Stone and colleagues in 2005 provided a relative risk of 0.91 with 95 percent confidence interval of 0.60 to 1.39. The combined effect of the two trials showed a relative risk of 0.93 with a 95 percent confidence interval of 0.61 to 1.42. The Cochran Q p-value was 0.565 and there were too few strata to calculate the I-squared value and Egger's p-value.

Figure 59

Impact of distal balloon embolic protection devices on side branch occlusion versus control in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.565 I2: Too few strata

Distal balloon embolic protection devices in other ACS populations. No trials or studies evaluated the impact of distal balloon embolic protection devices on this outcome.

Proximal balloon embolic protection devices in patients with STEMI. No trials or studies evaluated the impact of proximal balloon embolic protection devices on this outcome.

Proximal balloon embolic protection devices in other ACS populations. No trials or studies evaluated the impact of proximal balloon embolic protection devices on this outcome.

Embolic protection devices combined in patients with STEMI. Three RCTs evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) on side branch occlusion.95,107,112 In these trials, the use of embolic protection devices nonsignificantly decreased the risk of side branch occlusion [RR 0.91 (0.60, 1.39)] (Figure 60). Statistical heterogeneity was not detected (I2=0 percent) and publication bias could not be determined due to the number of studies available. All of the trials were determined to be of good methodological quality. 95,107,112

This figure depicts the meta-analysis of the impact of embolic protection devices combined on side branch occlusion versus control. The first trial by Cura and colleagues in 2007 provided a relative risk of 0.33 with 95 percent confidence interval of 0.00 to 3.80. The second trial by Matsuo and colleagues, also in 2007, provided a relative risk of 1.85 with 95 percent confidence interval of 0.25 to 13.97. The third trial by Stone and colleagues in 2005 provided a relative risk of 0.91 with 95 percent confidence interval of 0.60 to 1.39. The combined effect of the three trials showed a relative risk of 0.91 with a 95 percent confidence interval of 0.60 to 1.39. The Cochran Q p-value was 0.697, I-squared value was zero percent and there were too few strata to calculate the Egger's p-value.

Figure 60

Impact of embolic protection devices combined on side branch occlusion versus control in patients with ST-segment elevation myocardial infarction. Cochran Q: P=0.697 I2: 0 percent

Embolic protection devices combined in other ACS populations. No trials or studies evaluated the impact of embolic protection devices combined (distal or proximal; filter or balloon) on this outcome.

Discussion

In patients with STEMI undergoing PCI and comparing a catheter aspiration, mechanical thrombectomy, or embolic protection device versus control, only a minority of trials reported on the occurrence of the four most important adverse events (coronary dissection, coronary perforation, prolonged procedure time, and side branch occlusion). This made it difficult to determine the balance of benefits to harms for these devices.

The use of catheter aspiration devices versus control significant reduced the risk of coronary dissection and did not significantly impact the risk of side branch occlusion. One trial reported the outcome of coronary perforation although risk could not be calculated since no events occurred in either group. Overall, the use of catheter aspiration devices versus control did not significantly prolong procedure time in eight of nine trials and in one controlled observational study. When evaluated qualitatively, the procedure time were shortened in one trial, prolonged by 5 or less minutes in four trials and in one controlled observational study, and were more than 5 minutes prolonged in another four trials.

The use of mechanical thrombectomy devices versus control appears to be safe overall. In RCTs, the use of mechanical thrombectomy devices versus control did not significantly impact the risk of coronary dissection, coronary perforation, or side branch occlusion. However, mechanical thrombectomy devices appear to prolong the procedure time versus control. Three trials evaluated the impact of mechanical thrombectomy devices versus control on procedure time although were ineligible for pooling. In all three trials the procedure time was significantly prolonged in the mechanical thrombectomy device group versus control. The mean procedure time was prolonged by 9 to 16.2 minutes and one trial reported a median in which the procedure time was prolonged by 13.5 minutes.

Limited data was available to analyze the adverse events associated with the use of distal filter embolic protection devices versus control. In RCTs, the use of distal filter embolic protection devices versus control did not significantly impact the risk of side branch occlusion. The risk of coronary dissection and coronary perforation could not be calculated in the one trial in which it was reported. One trial evaluated procedure time which was significantly prolonged in the distal filter embolic protection device group versus control by a median of 8.5 minutes.

In RCTs, the use of distal balloon embolic protection devices versus control did not significantly impact the risk of coronary perforation or side branch occlusion. The risk of coronary dissection could not be calculated in the one trial which reported this outcome because no events occurred. Three trials evaluated the impact of distal balloon embolic protection devices versus control on procedure time although were not amenable to pooling. In two of the three trials, procedure time was significantly prolonged with the use of a distal balloon embolic protection device versus control. In these two trials, the procedure time was prolonged by a mean of 22.8 minutes and a median of 14 minutes.

The only adverse event which was reported in trials evaluating proximal balloon embolic protection devices versus control was procedure time. In one controlled trial, the procedure time was significantly prolonged in the proximal balloon embolic protection device group versus control by a median of 14 minutes.

When evaluating embolic protection devices combined (distal or proximal; filter or balloon), similar trends were observed as those evaluating the individual embolic protection device categories. The risk of coronary dissection could not be calculated because no events occurred in the trials which reported this outcome. Only one trial reported coronary perforation (distal balloon embolic protection device) therefore results did not change. The majority of trials evaluating embolic protection devices demonstrated a prolonged procedure time (four of five trials). The use of embolic protection devices combined did not significantly impact the risk of side branch occlusion, although a trend towards decreased risk was seen.

No trials or studies evaluating patients with NSTEMI or UA undergoing PCI and comparing thrombectomy or embolic protection devices versus control on adverse events were identified.

One trial evaluated patients with mixed ACS (STEMI, NSTEMI, or UA) undergoing PCI and comparing thrombectomy or embolic protection devices versus control on adverse events. In this trial, the use of a distal balloon embolic protection device versus control significantly prolonged the procedure time by a mean of 6.97 minutes.

Key Question 3. In ACS patients undergoing PCI of native vessels, which patient characteristics (e.g., gender, age, ethnicity, diabetes, smoker, ejection fraction, primary or rescue PCI, use of glycoprotein IIb/IIIa inhibitors, ischemia time, presence of thrombus-containing lesion, infarct-related artery and prePCI TIMI flow, use of direct stenting) affect outcomes?

Key Points

A total of nine RCTs, an individual patient data meta-analysis, a pooled analysis, and five observational studies provided useful data for Key Question 3.

  • RCTs evaluating treatment effect stratified by subgroups found the following:
    • No statistically significant difference in outcomes with catheter aspiration, mechanical thrombectomy or embolic protection devices efficacy based on differences in gender, age, diabetes, smoking status, primary or rescue PCI, presence of thrombus-containing lesion, prePCI TIMI flow, or the use of direct stenting.
    • A trend (P-value for interaction<0.10 between subgroups) towards greater improvements in attaining complete ST-segment resolution with proximal balloon embolic protection in those receiving a glycoprotein IIb/IIIa inhibitor versus those without such therapy.
    • A trend (P-value for interaction<0.10 between subgroups) towards greater improvements in attaining complete ST-segment resolution with proximal balloon embolic protection in those with an anterior infarct-related artery lesions versus lesions in other arteries.
    • Conflicting data was identified regarding the effect of ischemic time on outcomes following the use of catheter aspiration devices.
      • There was a trend (P-value for interaction<0.10 between subgroups) towards greater achievement of a higher MBG with catheter aspiration in those with ischemic times <180 minutes versus longer ischemic times.
      • There was significantly greater improvement (P-value for interaction between subgroups =0.02) in the achievement of TIMI 3 flow with catheter aspiration and a trend (P-value for interaction <0.10 between subgroups) towards greater reductions in slow flow or no reflow in those with prolonged ischemic times (6 to 24 hours from symptom onset) versus those with shorter ischemic times.
  • It should be noted that results of subgroup analyses from RCTs may be prone to type 2 error and false findings resulting from multiple hypothesis testing.
  • No RCTs evaluated the effect of ethnicity or ejection fraction on thrombectomy or embolic protection device efficacy.
  • The individual patient data meta-analysis by Burzotta and colleagues171,172 found that the use of aspiration or mechanical thrombectomy was associated with a survival benefit in the subgroup of patients treated with glycoprotein IIb/IIIa inhibitors but not in those not receiving them.
    • No qualitative differences in mortality were seen when splitting the study population according to the presence or absence of diabetes, earlier or later time to reperfusion, type of vessel (left anterior descending, circumflex, right coronary artery) containing the culprit lesion and lower or higher prePCI TIMI flow.
  • The pooled analysis by DeVita and colleagues143 found that in subgroups of short (≤3 hours) and intermediate (>3 hours to <6 hours) time to treatment (TTT) there was no significant difference between catheter aspiration and control on in-hospital MACE, STSR, MBG 2-3 or TIMI-3. In the subgroup of long TTT (>6 hours and ≤12 hours), catheter aspiration devices significantly increased the rate of STSR and TIMI-3 blood flow compared to control but did not significantly impact other outcomes.
  • The controlled observational study by Nakatani and colleagues149 found Killip class (a correlate to heart failure and ejection fraction) not to be a modifier of 30-day mortality with catheter aspiration device use. This constitutes the only data available to evaluate the potential confounding effect of heart function on outcomes.
  • The controlled observational study by Sardella and colleagues49 found that use of catheter aspiration, age, and symptom to balloon time were significant predictors of cardiac death (no deaths were of noncardiac cause) at 2 years.
  • Observational single arm studies found catheter aspiration and/or embolic protection device effectiveness to be negatively affected by increased age, prolonged ischemic time, female gender, presence of diabetes and absence of baseline thrombus.

Detailed Analysis

Study Design and Population Characteristics

A total of nine RCTs, an individual patient data meta-analysis, a pooled analysis and five observational studies were included in Key Question 3. All RCT data were in patients experiencing STEMI. STEMI was also an inclusion criterion for all trials in the individual patient data meta-analysis and the pooled analysis. Some of the observational studies included a mixed STEMI and NSTEMI population. Of the RCTs, 5, 2, 1, and 1 evaluated catheter aspiration, distal filter embolic protection, distal balloon embolic protection and proximal balloon embolic protection, respectively. None evaluated mechanical thrombectomy devices, although RCTs of these devices were included in the individual patient data meta-analysis. The pooled analysis evaluated catheter aspiration devices used in the three included RCTs. Outcomes evaluated in these trials included MBG, complete (>70 percent) ST-segment resolution, slow- and/or no-reflow, target vessel revascularization, MACE, TIMI blood flow and distal embolization.

Outcomes Results

Two trials provided subgroup results based on gender (Table 33).62,89 In the trial by Svilaas and colleagues, males were significantly less likely to experience a MBG of 0 or 1 if they received catheter aspiration than if they did not [RR 0.60 (0.44, 0.82)]. While females did not experience a significant reduction in achieving a MBG of 0 or 1 when the device was employed [RR 0.74 (0.49, 1.11)], the reductions noted between the genders was not found to be statistically differ (P-value for interaction=0.43 between subgroups).62 In the trial by Kelbaeck and colleagues, males and females both had nonsignificant improvements in ST-segment resolution (≥70 percent at 90 minutes postPCI) when a filter distal embolic protection device was employed and the reductions were not found to differ statistically between genders (P-value for interaction=0.79 between subgroups).89

Table 33. Results of subgroup analysis from randomized controlled trials evaluating the effect of gender on clinical outcome.

Table 33

Results of subgroup analysis from randomized controlled trials evaluating the effect of gender on clinical outcome.

Three trials provided subgroup results stratified by age;62,83,89 however, numerical data was obtainable for only one (Table 34).62 The trial by Svilaas and colleagues demonstrated that both those over 65 years [RR 0.74 (0.55, 0.99)] and 65 years or younger [RR 0.58 (0.39, 0.88)] were less likely to experience a MBG of 0 or 1 if they used a catheter aspiration device, with no differences noted between groups (P-value for interaction=0.34 between subgroups). These findings are supported by results of the trial by Burzotta and colleagues, which also found that a catheter aspiration device was beneficial (obtained both a MBG≥2 and complete ST-segment resolution) in both those greater than 60 and 60 years or younger (no numerical data reported).83 The trial by Kelbaeck and colleagues suggested that age (<70 or ≥70) did not affect the efficacy of filter distal embolic protection (P-value for interaction between subgroups >0.10).89

Table 34. Results of subgroup analysis from randomized controlled trials evaluating the effect of age on clinical outcome.

Table 34

Results of subgroup analysis from randomized controlled trials evaluating the effect of age on clinical outcome.

The impact of ethnicity was not evaluated as part of subgroup analyses and was only sporadically reported in the demographic tables of included trials. Thus we were unable to assess its affect on any outcome.

Two trials evaluated the impact of using filter distal embolic protection devices in patients with diabetes mellitus (Table 35).89,95 One trial provided subgroup results based on the presence or absence of diabetes mellitus while a second trial only provided the results in the diabetic subgroup. In the trial by Kelbaeck and colleagues, there was a nonsignificant reduction in the risk of achieving ST-segment resolution (≥70 percent at 90 minutes postPCI) in diabetic patients [RR 0.81 (0.55 to 1.19)] but a nonsignificant increase in nondiabetic patients [RR 1.07 (0.97 to 1.17)], with a weak trend towards differences between the groups (P-value for interaction=0.17 between subgroups).89 In the trial by Cura and colleagues, those with diabetes had a nonsignificant reduction on the risk of achieving ST segment resolution (≥70 percent at 60 minutes postPCI) [RR 0.91 (0.65 to 1.29)].95 In the total population of the trial by Cura and colleagues, the use of the device did not increase the proportion of patients achieving complete ST-segment resolution at 60 minutes (61 percent versus 60 percent; p=0.91) or any other time point.95 The device and endpoint were similar between trials so a pooled analysis of the diabetic subgroups of these two trials yielded a nonsignificant reduction in the risk of achieving ST-segment resolution [RR 0.86 (0.67 to 1.12)]. Due to the limited number of data points in this analysis statistical heterogeneity could not be assessed. Our literature search also identified a single individual patient data meta-analysis by Burzotta and colleagues.171,172 This meta-analysis pooled data from eleven RCTs of adjunctive thrombectomy devices (catheter aspiration or mechanical thrombectomy) (N=2686 patients) in patients with STEMI. Embolic protection device trials were not included in this meta-analysis. Kaplan–Meier analysis conducted in this meta-analysis showed that randomization to a thrombectomy device was associated with significantly lower risk of all-cause mortality (p=0.049), MACE (p=0.01) and the composite endpoint of death or myocardial infarction (p=0.01). Upon subgroup analysis undertaken in this meta-analysis, no qualitative difference in mortality was seen when splitting the study population according to the presence or absence of diabetes.171,172

Table 35. Results of subgroup analysis from randomized controlled trials evaluating the effect of diabetes mellitus on clinical outcome.

Table 35

Results of subgroup analysis from randomized controlled trials evaluating the effect of diabetes mellitus on clinical outcome.

Three trials evaluated the impact of using embolic protection devices in patients with a history of smoking (Table 36).18,95,112 Two trials provided subgroup results based on the presence or absence of a history of current smoking while a third trial only provided the results in the current smoker subgroup. While the use of a proximal balloon embolic protection device in the trial by Haeck and colleagues significantly increased the risk of achieving ST-segment resolution (>70 percent postPCI) in smokers [RR 1.41 (1.11, 1.80)] but not nonsmokers [RR 1.32 (0.90, 1.95)], the results were similar between subgroups (P-value for interaction=0.78 between subgroups).120 In the trial by Stone and colleagues, the use of a balloon distal embolic protection device did not significantly impact the risk of achieving an ST-segment resolution (>70 percent at 30 minutes postPCI) in current smokers [RR 0.99 (0.81, 1.22)] or nonsmokers [RR 1.05 (0.87, 1.27)] with no difference seen between subgroups (P-value for interaction=0.68 between subgroups).112 In the trial by Cura and colleagues, smoking did not significantly impact the risk of achieving an ST-segment resolution (>70 percent 60 minutes postPCI) [RR 1.12, 0.93, 1.34)].95 As noted above, in the total population of the trial by Cura and colleagues, the use of the device did not increase the proportion of patients achieving complete the ST-segment resolution at 60 minutes (61 percent versus 60 percent; p=0.91) or any other time point. When the current smoker subgroups of the trials were pooled, the risk of achieving an ST-segment resolution from embolic protection devices was nonsignificantly increased [RR 1.16 (0.95, 1.41)], but due to differences in the devices employed and the definitions of ST segment resolution, statistical heterogeneity was high (I2=63.3 percent).

Table 36. Results of subgroup analysis from randomized controlled trials evaluating the effect of smoking on clinical outcome.

Table 36

Results of subgroup analysis from randomized controlled trials evaluating the effect of smoking on clinical outcome.

The impact of ejection fraction on outcomes was not evaluated in subgroup analysis, thus precluding evaluation. Only the trials by Burzotta and colleagues83 and Stone and colleagues112 provided subgroup results based on whether the device was used for primary angioplasty or for rescue angioplasty; however, the trial by Burzotta and colleagues did not provide any numerical data and thus was not included (Table 37). In this trial, a catheter aspiration device was not statistically significantly beneficial (obtained both a MBG≥2 and complete ST-segment resolution) in either the subgroup of patients undergoing primary or rescue angioplasty (no numerical data reported).83 In subgroup analysis within the trial by Stone and colleagues, neither those receiving a balloon distal embolic protection device for primary [RR 1.05 (0.90, 1.22)] nor rescue angioplasty [0.91 (0.64, 1.29)] had significant impact on ST-segment resolution (>70 percent at 30 minutes postPCI) and no statistically significant difference was noted between subgroups (P-value for interaction=0.46 between subgroups).112

Table 37. Results of subgroup analysis from randomized controlled trials evaluating the effect of failed thrombolysis on clinical outcome.

Table 37

Results of subgroup analysis from randomized controlled trials evaluating the effect of failed thrombolysis on clinical outcome.

Only one trial evaluated the effect of concurrent GP IIb/IIIA inhibitor use on a catheter aspiration device's efficacy.83 In this trial by Burzotta and colleagues, the use of a catheter aspiration device was not statistically significantly beneficial (obtained both a MBG≥2 and complete ST-segment resolution) in either the subgroup who did or did not receive a GP IIb/IIIa inhibitor (no numerical data reported). In the aforementioned individual patient data meta-analysis,171,172 subgroup analysis according to administration of GP IIb/IIIa inhibitors showed that randomization to an adjunctive thrombectomy device was associated with a mortality benefit in the subgroup of patients treated with GP IIb/IIIa inhibitors [n=1787 patients; hazard ratio 0.61 (0.38 to 0.90); p=0.045], but not in those without GP IIb/IIIa inhibitors [n=899 patients; hazard ratio 0.93 (0.48 to 1.80); p=0.84]. In addition, two trials evaluated the affect of concurrent GP IIb/IIIa inhibitor use on an embolic protection device's (distal filter and proximal balloon) ability to obtain complete ST-segment resolution (Table 38).18,95 In both the trial by Cura and colleagues and Haeck and colleagues, the subgroup of patients administered GP IIb/IIIa inhibitors achieved statistically significant increased rates of complete (>70 percent) ST-segment resolution [RR 1.36 (1.09 to 1.69) and RR 1.97 (1.17 to 3.32), respectively]. However in the trial by Haeck and colleagues, the subgroup not receiving a GP IIb/IIIa inhibitor did not realize a statistically significant improvement in complete ST-segment resolution [RR 1.20 (0.97 to 1.49)] The P-value for interaction between GP IIb/IIIa inhibitor use and nonuse groups in this trial (proximal embolic balloon protection) was nearing statistical significance (p=0.08), suggesting concomitant GP IIb/IIIa inhibitor use may enhance the ability of embolic protection to achieve complete ST-segment resolution.18

Table 38. Results of subgroup analysis from randomized controlled trials evaluating the effect of glycoprotein IIb/IIIa inhibitor use on clinical outcome.

Table 38

Results of subgroup analysis from randomized controlled trials evaluating the effect of glycoprotein IIb/IIIa inhibitor use on clinical outcome.

A total of eight trials and one pooled analysis evaluated the affect of ischemia time on the efficacy of adjunctive devices to improve postST-segment myocardial infarction outcomes; however, only six of the eight trials provided numerical results (Table 39).16,18,62,68,83,89,95,112 In the trial by Svilaas and colleagues, regardless of total ischemia time (≥180 minutes or <180 minutes) patients were less likely to have a MBG of 0 or 1 postPCI when catheter aspiration was used [RR 0.73 (0.55 to 0.99) and RR 0.45 (0.28 to 0.74), respectively].62 However, the P-value for interaction between subgroups trended towards statistical significance (p=0.09) suggesting catheter aspiration may be more effective in patients undergoing PCI within 180 minutes. The trial by Ikari and colleagues also supported the conclusion that catheter aspiration devices had beneficial effects on MBG in patients undergoing early- (≤6 hours from symptom onset) and late- (6-24 hours) reperfusion [RRs of achieving a MBG-3 were 2.32 (1.50 to 3.58) and 2.34 (1.21 to 4.54)], respectively; with no difference between subgroups (p-value for interaction =0.98 between subgroups).16 However, when looking at the slow/no-reflow or achievement of TIMI-3 blood flow endpoints in this trial, only patients undergoing late perfusion realized statistically significant benefits between longer and shorter ischemic times [RR 0.23 (0.07 to 0.72) and RR 1.45 (1.12 to 1.86)].16 The P-value for interaction between subgroups of effect trended towards statistical significance for slow/no-reflow (p=0.07) and was statistically significant for the TIMI-3 blood flow endpoint (p=0.02). Neither results from the trial by Ikari and colleagues nor from an additional trial by Chao and colleagues demonstrated any ischemia time subgroup to statistically significantly benefit from catheter aspiration in respect to final health outcomes including target lesion or vessel revascularization, mortality, or combined major adverse cardiac events (all crossing the line of unity). Data from Chao did qualitatively appear to suggest decreasing efficacy of catheter aspiration on terminal endpoints as ischemic times increased; however, the effects between subgroups in each of these trials and endpoints were not found to be statistically significantly different (P-values for interaction all >0.25). On their own, the three trials evaluating embolic protection devices (one each of distal balloon, distal filter and proximal balloon) did not suggest embolic protection devices allowed patients to achieve complete ST-segment resolution to a greater or lesser extent in different ischemia time subgroups (P-value for interaction >0.22 for all between subgroups). Only those within the shorter ischemia time subgroup receiving proximal balloon embolic protection were found to have a statistically significantly increased chance of complete ST-segment resolution [RR 1.38 (1.06 to 1.80)]. When results from these three trials were pooled separately by shorter and longer ischemia subgroups, similar results were seen [pooled RR for shorter ischemia time 1.08 (0.85 to 1.38), I2=63.8 percent and pooled RR for longer ischemia time 1.09 (0.95 to 1.24), I2=0 percent). The trial by Burzotta and colleagues found that a catheter aspiration device was beneficial (obtained both a MBG≥2 and complete ST-segment resolution) in both those with ischemia times greater than 250 minutes and 250 minutes or less (no numerical data reported).83 The trial by Kelbaeck and colleagues suggested that ischemic time (stratified at 6 hours) did not affect the efficacy of distal filter embolic protection (P-value for interaction between subgroups >0.10). Upon subgroup analysis undertaken in the individual patient data meta-analysis, no qualitative difference in mortality was seen when splitting the study population according to shorter, intermediate or longer ischemia times.171,172

Table 39. Results of subgroup analysis from randomized controlled trials evaluating the effect of ischemic time on clinical outcome.

Table 39

Results of subgroup analysis from randomized controlled trials evaluating the effect of ischemic time on clinical outcome.

One pooled analysis by De Vita and colleagues pooled data from three RCT which compared catheter aspiration to standard procedure in patients with STEMI.143 Four outcomes were evaluated (in-hospital MACE, STSR ≥70 percent, TIMI-3 and MBG 2 or 3) based on three subgroups of time to treatment (TTT), defined as time from symptoms onset to catheter laboratory [≤3 hours (short TTT), >3 hours to <6 hours (intermediate TTT), and >6 hours to ≤12 hours (long TTT)]. Two hundred-ninety nine patients were analyzed overall with 128 in the short TTT subgroup, 135 in the intermediate TTT subgroup, and 36 in the long TTT subgroup. There was no significant difference between catheter aspiration and control in the outcomes evaluated in the short and intermediate TTT subgroups. In the long TT subgroup, the catheter aspiration group was significantly more likely to achieve STSR (50 percent versus 20 percent, p=0.01) and TIMI-3 blood flow (88 percent versus 60 percent, p=0.01) versus control although none of the other outcomes were significant.

Six trials evaluated the effect of visible thrombus at baseline on the efficacy of adjunctive devices (Table 40) 18,62,89,95,112 The trial by Svilaas and colleagues evaluated the effect of catheter aspiration use on MBG in patients with and without visible thrombus. Regardless of the presence of visible thrombus at baseline, catheter aspiration use resulted in fewer patients having a MBG of 0 or 1 postprocedure [RR with visible thrombus 0.61 (0.43 to 0.87) and RR without visible thrombus 0.70 (0.50 to 0.98)]. A test for interaction between these subgroups showed no statistically significant difference in effect (p=0.58).62 The trial by Burzotta and colleagues found that a catheter aspiration device was beneficial (obtained both a MBG≥2 and complete ST-segment resolution) in the subgroup of patients with a high thrombus burden (thrombus score of 4 to 4), but not those with a lower burden (thrombus score of 1 or 2) (no numerical data reported).83 The remaining four trials evaluated embolic protection devices use on obtainment of complete ST-segment resolution in patients with and without visible thrombus. In the trial by Kelbaeck and colleagues, those patients without visible thrombus at baseline were more likely to achieve complete ST-segment resolution when using a filter distal embolic protection device versus control [RR 1.17 (1.01 to 1.37); however, the same device did not appear to benefit patients with visible thrombus [RR 1.00 (0.89 to 1.12). The difference between these subgroups what not found to be statistically significant (P-value for interaction=0.11 between subgroups).89 The trial by Haeck and colleagues demonstrated contradictory results [RR with 1.31 (1.02 to 1.68) and RR without 1.39 (0.94 to 2.05) baseline thrombus, P-value for interaction=0.80 between subgroups).18 In both the trial by Cura and colleagues and Stone and colleagues, the use of distal embolic protection (filter or balloon) was not found to be statistically significantly beneficial in either the visible thrombus or no thrombus subgroups.95,112 When embolic protection studies were pooled separately by baseline thrombus subgroup, neither the visible thrombus nor no visible thrombus subgroups demonstrated statistical significant effects on complete ST-segment resolution [RR with baseline visible thrombus 1.10 (0.95 to 1.27), I2=24.5 percent and RR without thrombus 1.12 (0.76 to 1.64), I2=not estimable).

Table 40. Results of subgroup analysis from randomized controlled trials evaluating the effect of visible thrombus on clinical outcome.

Table 40

Results of subgroup analysis from randomized controlled trials evaluating the effect of visible thrombus on clinical outcome.

A total of six trials evaluated the effect of the infarct-related artery on the efficacy of adjunctive devices to improve postST-segment myocardial infarction outcome (Table 41) 18,62,83,89,95,112 In the trial by Svilaas and colleagues, catheter aspiration was found to reduce the risk of postprocedure MBG of 0 or 1 in patients with the RCA as the infarct-related artery [RR 0.48 (0.29 to 0.81)] or other arteries [RR 0.71 (0.53 to 0.93)]. A test for interaction between these infarct-related artery subgroups showed no statistically significant difference in effect (p=0.19). The trial by Burzotta and colleagues found that a catheter aspiration device was not statistically significantly beneficial in obtaining both a MBG≥2 and complete ST-segment resolution in either those with a LAD or a RCA/CX as the infarct-related artery (no numerical data reported). Upon subgroup analysis undertaken in the individual patient data meta-analysis,171,172 no qualitative difference in mortality was seen when splitting the study population according to the type of infarct-related artery (left anterior descending or circumflex artery or RCA). The remaining four trials evaluated embolic protection devices. In three of these four trials,89,95,112 distal embolic protection devices (balloon or filter) failed to improve patients' chance of attaining complete ST-segment resolution when evaluating patients by specific infarct-related artery subgroups. In addition, tests for interaction between infract-related artery subgroups showed no statistically significant difference in effect in these three trials (p>0.20 for all). However, in the trial by Haeck and colleagues, proximal balloon embolic protection was found to increase patients chances of achieving complete ST-segment resolution when the lesion was in an anterior artery [RR 2.41 (1.11 to 5.19)], but not in other arteries [RR 1.20 (0.99 to 1.46)].18 A test for interaction between these infarct-related artery subgroups showed a trend towards a statistically significant difference in effect (p=0.09). Due to the heterogeneous nature by which trials divided subgroups, pooling was deemed inappropriate.

Table 41. Results of subgroup analysis from randomized controlled trials evaluating the effect of infarct-related artery on clinical outcome.

Table 41

Results of subgroup analysis from randomized controlled trials evaluating the effect of infarct-related artery on clinical outcome.

In addition to the effect of infarct-related artery, trials have also evaluated whether proximal or nonproximal location of the lesion within an artery affects the efficacy of adjunctive devices to improve postST-segment myocardial infarction outcomes (Table 42). The trial by Svilaas and colleagues demonstrated that catheter aspiration devices work equally well in preventing a postprocedure MBG of 0 or 1 in proximal and nonproximal lesion subgroups [RR 0.60 (0.43 to 0.85) and RR 0.69 (0.49 to 0.97), respectively] (P-value for interaction between subgroups=0.57).62 While, in trial by Haeck and colleagues, only patients with proximally located lesions were shown to achieve a higher rate of complete ST-segment resolution with the use of proximal balloon embolic protection [RR 1.71 (1.14 to 2.55)].18 Those in the nonproximal lesion subgroup did not realize statistically significant benefit [RR 1.18 (0.92 to 1.51)]. A test for interaction between these infract-related artery subgroups showed no statistically significant difference in effect (p=0.12). The trial by Kelbaeck and colleagues suggested that proximal or nonproximal lesion location did not affect the efficacy of filter distal embolic protection (P-value for interaction between subgroups >0.10) (numerical data not reported).

Table 42. Results of subgroup analysis from randomized controlled trials evaluating the effect of lesion location on clinical outcome.

Table 42

Results of subgroup analysis from randomized controlled trials evaluating the effect of lesion location on clinical outcome.

Five trials evaluated the effect of baseline TIMI flow on the efficacy of adjunctive devices to improve postST-segment myocardial infarction outcomes; however, the trial by Burzotta and colleagues did not provide numerical data and is therefore not included in (Table 43) 62,83,89,95,112 In the trial by Svilaas and colleagues, catheter aspiration was found to reduce the risk of postprocedural MBG of 0 or 1 in patients with a preprocedural TIMI blood flow of 0 or 1 [RR 0.72 (0.55 to 0.95)], but fell just shy of significance in those with a TIMI flow graded at 2 or 3 [RR 0.60 (0.36 to 1.10)] (P-value for interaction between subgroups=0.54). Similar results were found in the trial by Burzotta and colleagues, which found that a catheter aspiration device was beneficial (obtained both a MBG≥2 and complete ST-segment resolution) in those with a baseline TIMI flow of 0 or 1, but not those with a TIMI flow of 2 or 3 (no numerical data reported). Upon subgroup analysis undertaken in the individual patient data meta-analysis,171,172 no qualitative difference in mortality was seen when splitting the study population according to preprocedural TIMI flow (0–1 or 2–3). Three trials evaluated distal embolic protection (two filter, one balloon). In each of these trials, no preprocedure TIMI subgroup was found to provide a statistically significant effect on complete ST-segment resolution.

Table 43. Results of subgroup analysis from randomized controlled trials evaluating the effect of baseline thrombolysis in myocardial infarction flow on clinical outcome.

Table 43

Results of subgroup analysis from randomized controlled trials evaluating the effect of baseline thrombolysis in myocardial infarction flow on clinical outcome.

Only one trial evaluated the effect of direct stenting on the efficacy of adjunctive devices to improve postST-segment myocardial infarction outcomes (Table 44).75 In both the direct stenting and no direct stenting patient subgroups, use of catheter aspiration in this trial had no effect on patients' chances of attaining a postprocedure TIMI flow of 3, experiencing distal embolization or no reflow. The P-values for interaction between subgroups was not statistically significant for any of these endpoints (p>0.68). When evaluating the MBG-3 and the complete ST-segment resolution endpoints in this trial, patients not undergoing direct stenting received statistically significant benefit from catheter aspiration use [RR 2.07 (1.33 to 3.22) and RR 1.56 (1.00 to 2.45)], but patients undergoing direct stenting did not [RR 1.41 (0.96 to 2.07) and RR 1.41 (0.81 to 2.47), respectively]. However, the P-value for interaction between subgroups was not statistically significant for either endpoint (p≥0.20 for both).

Table 44. Results of subgroup analysis from randomized controlled trials evaluating the effect of direct stenting on clinical outcome.

Table 44

Results of subgroup analysis from randomized controlled trials evaluating the effect of direct stenting on clinical outcome.

In addition to the results from the above-mentioned RCTs and individual patient data meta-analysis, five observational studies were identified that provide data addressing key question 3.

The largest of these observational studies was the prospective, multicenter Osaka Acute Coronary Insufficiency Study (OACIS).149 Researchers evaluated 3,913 patients who underwent PCI within 24 hours after symptom onset, of which, 990 patients (25.3 percent) were treated with catheter aspiration before PCI. Overall, OACIS found a trend towards 30-day mortality benefit with intracoronary thrombectomy [hazard ratio (HR) 0.658, p=0.17]. Intracoronary thrombectomy was an independent predictor of a lower 30-day mortality risk in patients aged ≥70 years (HR 0.239, p=0.007) and patients with diabetes mellitus (HR 0.275, p=0.039), but not in patients < 70 years of age or nondiabetics. P-value for interaction between subgroups was statistically significant for age (p=0.008), but not diabetes status (p=0.17). Furthermore, baseline TIMI flow, gender, smoking and Killip class (a correlate to heart failure and ejection fraction) were not found to be modifiers of 30-day mortality (P-value for interaction between subgroups >0.24). A second observational study was conducted by the researchers from the EXPIRA RCT which randomized patients who had a STEMI to catheter aspiration (n=88) versus standard PCI (n=87).139 A multivariate Cox proportional hazard regression model was used to identify independent predictors of cardiac death at 2 years. No deaths during the trial period were other than cardiac cause. Randomization to thrombus aspiration [HR 0.12 (0.006 to 0.251), p=0.006], age [HR 1.508 (1.055 to 2.156), p=0.024] and symptom to balloon time [HR 1.322 (1.078 to 1.622), p=0.007] were found to be significant predictors of cardiac death at 2 years. Diabetes, hypertension, and final MBG <2 were not found to significantly predict cardiac death.

The remaining three single-arm observational studies conducted multivariate analysis. Cohen and colleagues evaluated catheter aspiration with the Export® catheter in patients experiencing STEMI and undergoing primary PCI to identify covariates associated with successful thrombectomy (increase in TIMI flow grade of at least 1). Upon multivariate logistic regression analysis, researchers identified ischemic time <6 hours as the only independent predictor of successful thrombectomy (p=0.04).146 Kramer and colleagues evaluated the use of catheter aspiration (Rescue™ or Export®) or proximal balloon embolic protection (Proxis™) in 914 patients experiencing STEMI and undergoing primary PCI. They found that age >60 years [hazard ratio 1.83 (1.14 to 2.93)], female gender [hazard ratio 4.22 (2.29 to 7.76)] and the presence of diabetes mellitus [hazard ratio 1.73 (1.09 to 2.76)] were all independent predictors of increased mortality by four years, whereas, current smoking, total ischemic time and having the LAD as the infarct-related artery were not.154 Ochala and colleagues conducted a multivariate analysis to determine independent predictors of achieving a postprocedure TIMI flow of 2 or 3 in the distal balloon embolic protection (PercuSurge) arm of a RCT of 120 ST-segment elevation patients undergoing primary PCI. In this analysis, the presence of baseline thrombus was found to independently predict increased odds of TIMI 2 or 3 flow in embolic protection device treated patients. LAD as the infarct-related artery, ischemic time greater than or equal to 6 hours and presence of diabetes mellitus were not found to be predictors of TIMI 2 or 3 flow attainment.164

Discussion

While a clinical trial or observational study may demonstrate an overall benefit for an intervention, this benefit may or may not occur to a similar extent across different types of constituents. As such, it is important to determine what, if any, data exists evaluating the impact of an intervention in these important subgroups. For Key Question 3, nine RCTs, an individual patient data meta-analysis, 171,172 a pooled analysis, and five observational studies provided some insight. However, most of the evidence is in the form of subgroup analysis stratified by covariate within RCTs. These subgroup analyses were typically underpowered to demonstrate statistically significant differences within and between subgroups and we cannot be sure that the results attained were due to a lack of impact or lack of power. Clinical trials with larger sample sizes would be needed to draw more definitive conclusions from such analyses. Secondly, many of the included trials and studies conducted subgroup analyses on large numbers of covariates making conclusions susceptible to bias resulting from multiple hypothesis testing.

Finally, the clinical trials provide univariate evaluations and we do not know if the results are due the factor being investigated or due to a confounder that one subgroup has in a differing amount from another subgroup.

Randomized trials and an individual patient data meta-analysis of RCTs have not demonstrated statistically significant effect modification of aspiration, mechanical thrombectomy or embolic protection device efficacy by gender, diabetes, smoking status, primary or rescue PCI, presence of thrombus-containing lesion, prePCI TIMI flow, or the use of direct stenting. Furthermore, no RCTs evaluated the effect of ethnicity or ejection fraction on thrombectomy or embolic protection device efficacy. While randomized trials and the individual patient meta-analysis did not show an affect of age, diabetes, baseline thrombus and gender on aspiration or thrombectomy device efficacy, a limited number of observational studies did.

Individual randomized trials did not demonstrate a modifying effect of glycoprotein IIb/IIIa use on aspiration or mechanical thrombectomy device efficacy. However, the individual patient data meta-analysis found that randomization to aspiration or mechanical thrombectomy was associated with a survival benefit in the subgroup of patients treated with glycoprotein IIb/IIIa inhibitors, but not in those not receiving them. This may suggest a modifying effect of glycoprotein IIb/IIIa inhibitors with these devices. While embolic protection devices were not studied in the individual patient meta-analysis, a single randomized trial of proximal balloon protection demonstrated a similar modifying affect of glycoprotein IIb/IIIa inhibitor use; with greater efficacy in those receiving a glycoprotein IIb/IIIa inhibitor. Limited data exists evaluating the effect of glycoprotein IIb/IIIa inhibitor use on the efficacy of distal embolic protection devices.

It appears doubtful that ischemic time affects the efficacy of aspiration or mechanical thrombectomy devices or embolic protection devices. Data regarding the affect of ischemic time on efficacy of aspiration catheter efficacy (MBG and TIMI 3 flow) was conflicting in randomized trials; while, the OASIS observational study suggested prolonged ischemic time negatively affected the ability of thrombectomy or embolic protection devices to reduce mortality. A pooled analysis suggested prolonged time to treatment lead to greater efficacy (STSR and TIMI-3) of catheter aspiration devices. Neither beneficial nor harmful associations between ischemic time and aspiration or mechanical thrombectomy devices were observed in the individual patient data meta-analysis.

Individual randomized trials and the individual patient meta-analysis suggested no modification of aspiration or mechanical thrombectomy device efficacy based upon infarct-related artery. However, a single trial, found a trend towards statistically significant greater efficacy (complete ST-segment resolution) of proximal balloon embolic protection in those with an anterior infarct-related artery. No studies have evaluated whether distal embolic protection device efficacy is impacted by infarct-related artery location.

Strength of Evidence and Applicability

Strength of Evidence

A summary of the strength of evidence for Key questions 1 and 2 are in Table 45 and Table 46 while the full evaluation of the strength of evidence for each outcome is found in Appendix G.

Table 45. Summary of the strength of evidence for Key Question 1: In patients with acute coronary syndrome who are undergoing percutaneous coronary intervention of native vessels, does the use of an adjunctive device affect final or intermediate health outcomes compared to usual care?

Table 45

Summary of the strength of evidence for Key Question 1: In patients with acute coronary syndrome who are undergoing percutaneous coronary intervention of native vessels, does the use of an adjunctive device affect final or intermediate health outcomes (more...)

Table 46. Summary of the strength of evidence for Key Question 2: In patients with acute coronary syndrome who are undergoing percutaneous coronary intervention of native vessels, does the use of an adjunctive device affect adverse outcomes compared to usual care?

Table 46

Summary of the strength of evidence for Key Question 2: In patients with acute coronary syndrome who are undergoing percutaneous coronary intervention of native vessels, does the use of an adjunctive device affect adverse outcomes compared to usual care? (more...)

A majority of the available evidence was in the STEMI population. In patients with STEMI, there was a high strength of evidence that catheter aspiration devices versus control decreased the risk of MACE, distal embolization and no reflow. The strength of evidence was moderate that catheter aspiration devices increased the attainment of ST-segment resolution, MBG-3, or TIMI-3 blood flow and had no effect on ejection fraction versus control. The strength of evidence was low that catheter aspiration devices had no effect on the risk of mortality, myocardial infarction, or target revascularization and insufficient for stroke, all versus control. Regarding adverse events, the strength of evidence for catheter aspiration devices versus control was high that the risk of coronary dissection was decreased and that there was no effect on prolongation of procedure time. The strength of evidence was insufficient that catheter aspiration devices had no effect versus control on coronary perforation.

The strength of evidence associated with all final health outcomes in the STEMI population undergoing PCI with a mechanical thrombectomy device was insufficient due to limited data available per outcome. No reflow was also graded with insufficient evidence. There was moderate strength of evidence that mechanical thrombectomy devices had no effect on the risk distal embolization, ejection fraction or attainment of TIMI-3 blood blow versus control. The strength of evidence was low that mechanical thrombectomy devices had no effect ST-segment resolution or attainment of a MBG-3. When analyzing different time points for the outcome of MACE, there was a significant reduction in the risk of MACE at 365 days [RR 0.66 (0.44, 0.97)] not seen in evaluations at earlier time periods, although this was based on a single randomized controlled trial. The strength of evidence for prolongation of procedure time was high for mechanical thrombectomy devices versus control, while the strength of evidence was insufficient for coronary dissection or perforation due to the limited amount of data…

For comparisons between distal filter embolic protection devices and control, no evaluation had a high strength of evidence. The strength of evidence was moderate that there was no effect on the risk of MACE, ST-segment resolution, or attainment of a MBG-3. The strength of evidence was low that there was increased risk of target revascularization and no effect on the attainment of TIMI-3 blood flow or ejection fraction. The strength of evidence was insufficient for mortality, myocardial infarction, stroke, distal embolization and no reflow. For adverse outcomes, the strength of evidence was insufficient for all outcomes.

The strength of evidence was high that there was an increased risk of attaining a MBG-3 with the use of a distal balloon embolic protection device versus control. The strength of evidence was moderate that there was no effect on ST-segment resolution and ejection fraction and low that there was increased risk on attainment of TIMI-3 blood flow with the use of distal balloon embolic protection devices versus control. The strength of evidence was insufficient for mortality, myocardial infarction, target revascularization, MACE, distal embolization and no reflow due to the limited amount of data. Regarding adverse outcomes, strength of evidence was low that there was prolonged procedure time while insufficient for other adverse outcomes comparing distal balloon embolic protection devices versus control.

For all final health, intermediate and adverse outcomes the strength of evidence was insufficient for the comparison of proximal balloon embolic protection devices versus control with one exception. The strength of evidence was moderate that proximal balloon embolic protection devices prolong procedure time versus control.

For comparisons between embolic protection devices combined versus control, the strength of evidence was moderate that the attainment of a MBG-3 was increased with the use of an embolic protection device versus control and that there was no effect on the risk of MACE, distal embolization, or ejection fraction. The strength of evidence was low that there was increased risk of attaining TIMI-3 blood flow and that there was no effect on the risk of ST-segment resolution with the use of embolic protection devices combined versus control. All other outcomes were insufficient due to the limited amount of data available. In terms of adverse outcomes, the strength of evidence was moderate that the use of embolic protection devices combined prolong procedure time versus control while insufficient for all other adverse outcomes.

In the mixed ACS population strength of evidence was predominately insufficient or low for all device categories versus control. There was a high strength of evidence that distal balloon embolic protection devices decreased the risk of no reflow versus control, which was propagated into the embolic protection devices combined analysis. There was a moderate strength of evidence that mechanical devices and distal balloon embolic protection devices increased the attainment of ST-segment resolution and that distal balloon embolic protection devices increased the attainment of MBG-3 versus control and prolonged procedure time. Both distal balloon results were propagated into the embolic protection devices combined analyses. The strength of evidence was low that catheter aspiration devices increased the attainment of MBG-3 versus control. All other outcomes for all device categories were insufficient.

In the UA / NSTEMI population the strength of evidence was insufficient for all final health, intermediate, and adverse outcomes due to the limited amount of data available for each comparison and outcome.

Applicability

The applicability of evidence was high for four evaluations: the impact of distal balloon embolic protection devices on stroke versus control and the impact of mechanical thrombectomy devices on coronary dissection, perforation and prolonged procedure time versus control. Applicability of the trials was in the moderate to low range (52.6 percent and 43.3 percent of comparisons, respectively) for all other outcomes because the trials were mostly conducted outside of the United States. The applicability of individual trials, studies, and the body of evidence per outcome assessed can be found in Appendix H along with the description of factors that impacted the applicability of the body of evidence.

Views

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...