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

National Clinical Guideline Centre (UK). Unstable Angina and NSTEMI: The Early Management of Unstable Angina and Non-ST-Segment-Elevation Myocardial Infarction. London: Royal College of Physicians (UK); 2010. (NICE Clinical Guidelines, No. 94.)

Cover of Unstable Angina and NSTEMI

Unstable Angina and NSTEMI: The Early Management of Unstable Angina and Non-ST-Segment-Elevation Myocardial Infarction.

Show details

4Anti–thrombin therapy

Instability of coronary plaque is the pathophysiological substrate for the clinical syndromes of UA and NSTEMI, and is associated with activation of local prothrombotic systems. It is therefore not surprising that considerable research has been undertaken to investigate the role of anticoagulant therapy in the management of patients with these conditions. Heparin and the direct antithrombin agents inhibit the conversion of fibrinogen to fibrin and therefore reduce the likelihood of clot (thrombus) formation. Prescribers should be aware that advanced age, reduced body weight (<50 kg) and impaired renal function increase bleeding risk associated with anticoagulants.

4.1. Heparins

Heparins, both unfractionated and low molecular weight, are indirect thrombin inhibitors which form complexes with antithrombin, and inactivate thrombin, clotting factor Xa (and to a lesser extent, factors XIIa, XIa, and IXa). Low molecular weight heparins (LMWH) have a number of potential advantages over unfractionated heparin (UFH):

  • They can be administered by subcutaneous injection, rather than having to be given by an intravenous bolus or infusion, and they have greater bioavailability.
  • The duration of their anticoagulant effect is greater, allowing once or twice daily administration.
  • Their anticoagulant response is more predictable and is correlated with body weight, making dosage calculation easier.
  • They do not require monitoring by blood testing though the dose may have to be adjusted for patients who are very obese or have renal failure.
  • They have a reduced risk of causing immune-mediated thrombocytopenia.

UFH has been shown to be superior to placebo in patients with NSTEMI and UA118 and in a number of trials has been compared to LMWH. A literature search was therefore performed to compare LMWH and UFH in these patients. Thus the clinical question asked, and upon which the literature was searched, was:

“What is the efficacy and safety of adding a LMWH compound to aspirin (with or without clopidogrel) in the management of patients with UA or NSTMEMI compared to the combination of unfractionated heparin and aspirin (with or without clopidogrel)?”

4.1.1. Clinical methodological introduction

The literature was searched for systematic reviews or RCTs published in 1999 to 2009. The rationale for searching from January 1999 onwards was to reflect current practice, particularly the use of stents for revascularisation. To be included, the population must contain > 60% people with unstable angina/NSTEMI. Seven RCTs 119-125 were identified which compared low molecular weight heparin (LMWH) and unfractionated heparin (UFH) in non ST-segment elevation ACS patients. The follow up period ranged from 6 to 30 days.

Of these trials, two double blind RCTs, ESSENCE (N=3,171) 122 and TIMI IIB (N=3910) 119, compared enoxaparin and UFH on a background of aspirin.

The open label FRIC RCT 125 (N=1499) compared dalteparin and UFH on a background of aspirin.

The double blind RCT, ACUTE II 121 (N=525), and two open label RCTs, INTERACT 124 (N=746) and A-Z 120 (N=3987) compared enoxaparin and UFH on a background of glycoprotein IIb/IIIa inhibitor and aspirin. The open label RCT, SYNERGY (N=10,027) 123 compared enoxaparin and UFH on a background of aspirin or clopidogrel (62% received clopidogrel) with GPIs recommended, but not mandated (57% received GPIs).

One meta–analysis 126 was rejected because it lacked an explanation of how the studies were searched for and assessed for quality.

The NCC–CC conducted a meta-analysis comparing low molecular weight heparins to unfractionated heparin (7 RCTs: ESSENCE, TIMI IIB, ACUTE II, INTERACT, A to Z, SYNERGY, FRIC). Subsequently, a systematic review 127 comparing enoxaparin with UFH was identified in the literature re-runs. The Murphy et al. systematic review contained an extra outcome (death, nonfatal MI, or nonfatal major bleed). Also, the authors contacted trial investigators for data, and were therefore able to include more studies for the outcome of death or nonfatal MI than the NCC meta-analysis had. The results of the Murphy et al meta-analysis and the NCC meta-analysis were similar for other outcomes.

4.1.2. Clinical evidence statements

The NCC–CC meta-analysis (including one dalteparin study) found a non–significant difference between LMWH and UFH for:

  • Death (7 RCTs; OR 0.96 [95% CI 0.75 to 1.23])
  • Urgent revascularization rates (4 RCTs; OR 0.92 [95% CI 0.79, 1.07])
  • Death or MI (4 RCTs; OR 0.88 [95% CI 0.72, 1.06])
  • Death or MI or urgent revascularization (4 RCTs; OR 0.88 [95% CI 0.77, 1.02])
  • Major bleeding (7 RCTs; OR 1.10 [95% CI 0.85, 1.42]); this analysis had significant heterogeneity I2 = 49.8%.
  • Minor bleeding (6 RCTs; OR 1.58 [95% CI 1.00, 2.50]); this analysis had significant heterogeneity I2 = 92.7%.

Evidence Level 1+

The NCC–CC meta-analysis (including one dalteparin study) showed that LMWH significantly reduced the odds of:

  • MI (7 RCTs; OR 0.87 [95% CI 0.79, 0.95]).

Evidence Level 1+

As the majority of the evidence compared enoxaparin with unfractionated heparin, the meta-analysis results were presented separately for this comparison. Enoxaparin versus UFH (refer to Table 4-1):

Table 4-1. Summary of outcomes for enoxaparin versus UFH from the Murphy et al and NCC meta-analyses.

Table 4-1

Summary of outcomes for enoxaparin versus UFH from the Murphy et al and NCC meta-analyses.

There was no significant difference between enoxaparin and UFH for:

  • Death
  • Death, nonfatal MI, or nonfatal major bleed
  • Urgent revascularization rates
  • Major bleeding (this analysis had significant heterogeneity, I2 = 58.1%).

Evidence Level 1+

Compared with UFH, enoxaparin significantly reduced the odds of:

  • Nonfatal MI
  • Death or non fatal MI
  • Death or MI or urgent revascularisation

Evidence Level 1+

Compared with UFH, enoxaparin significantly increased the odds of:

  • Minor bleeding (this analysis had significant heterogeneity; I2 = 94.0%)

Evidence Level 1+

Forest plots for NCC–CC meta-analysis comparing LMWH with UFH

See Figure 4-1, Figure 4-2, Figure 4-3, Figure 4-4, Figure 3-5, Figure 3-6.

DEATH

Figure 4-1

Death.

Figure 4-2. Myocardial Infarction.

Figure 4-2

Myocardial Infarction.

URGENT REVASCULARISATION

Figure 4-3

Urgent Revascularization.

DEATH OR MI OR URGENT REVASCULARISATION

Figure 4-4

Death or MI or Urgent Revascularization.

MAJOR BLEED

Figure 4-5

Major Bleeding.

MINOR BLEED

Figure 4-6

Minor Bleeding.

4.1.3. Health economic methodological introduction

One relevant cost-effectiveness analysis from a UK perspective was identified based on clinical effectiveness data from the ESSENCE study 122(with data from TIMI IIB 119 used in sensitivity analysis). Six studies from non-UK perspectives were also identified but as these used the same clinical effectiveness data as the UK analysis were judged to likely to add little additional information for UK decision making and were not reviewed128-133.

Nicholson et al. 134 reported a decision analysis based primarily on data from the ESSENCE RCT122 (sensitivity analysis did incorporate data from TIMI IIB119). The study compared enoxaparin with UFH in patients with UA or NSTEMI. A UK NHS perspective was taken. Cost and QALYs are estimated at one-year. Outcomes incorporated were death, MI, recurrent angina and quality of life. Costs included were enoxaparin, UFH, drug administration (consumables, IV pump, monitoring, nursing time), hospital length of stay (at 30 days), revascularisation (at one year). An alternative analysis looked at using costs of cardiac events at one year rather than length of stay. Cost effectiveness was expressed in terms of cost per QALY gained. The key potential limitation of the study is the use of data from the ESSENCE trial which reported in 1997 and had a low stent and thienopyridine use relative to current practices. Additionally, a lifetime analysis might be considered more appropriate as mortality was impacted and the quality of life valuation method was not choice-based.

4.1.4. Health economic evidence statements

Nicholson et al. 134 reported that enoxaparin was dominant compared to UFH in people with UA/NSTEMI – costs were reduced by £317 per person with a QALY gain of 0.013. Additional drug costs of enoxaparin were mostly offset by administration costs of UFH (saline, consumables, iv pump, monitoring, nurse time). Additional cost savings came from reduced length of stay and revascularisation avoided. These results are considered applicable to the UK NHS setting. However, there is a potential serious limitation relating to the use of data from the ESSENCE trial, which reported in 1997 and which is noted to have a low revascularisation rate relative to more recent practice (27% in enoxaparin group, and 32.2% in the unfractionated heparin group). Results are reported as being very sensitive to rates of revascularisation, and duration and cost of length of stay. However, in all but one sensitivity analysis enoxaparin remained dominant – when length of stay was used from a UK sub-group of ESSENCE there was a net cost (due to increased length of stay in the enoxaparin group) with an incremental cost-effectiveness ratio of £3,305 per QALY gained.

4.1.5. Evidence summary

The trials evaluating LMWH in patients with UA or NSTEMI show that enoxaparin is at least comparable, and may be superior, to UFH. Evidence for the use of daltaparin is limited. Enoxaparin reduces the rates of composite end points (death, re-infarction, revascularisation, recurrent myocardial ischaemia) and when analysed separately there is a reduction in MI but no evidence of a mortality benefit. Also, treatment with enoxaparin is associated with an increased risk of minor, but not major, bleeding.

A UK NHS perspective economic analysis based on the ESSENCE Trial 134 found that enoxaparin was dominant over UFH (more effective and lower cost) in patients with non-ST elevation MI or UA

4.1.6. Evidence to recommendations

The GDG acknowledged that one of the difficulties in analysing the numerous trials which compare LMWH with UFH is that they have occurred over more than a 10-year time period during which the use of adjunctive therapies, such as clopidogrel and GPIs, has changed. The earlier trials, such as FRIC (1997) 125 and ESSENCE (1997) 122 had background therapy of aspirin alone, whereas more recent trials (ACUTE-II [2002] 121, INTERACT [2003] 124 A-Z [2004])120 had both aspirin and GPIs as adjunctive treatment, and one (SYNERGY [2004]123 had GPIs with aspirin and/or Clopidogrel. In some trials the use of GPIs was mandated, and in others it was left to physician discretion. Given that these agents (aspirin, clopidogrel, GPIs and heparin) can all have an effect on outcome it was difficult for the GDG to dissect out the relative benefits of each individually. All but one of the trials the GDG reviewed involved the use of enoxaparin, and this is reflected in UK clinical practice, where dalteparin is not widely used.

It was noted that the cost of enoxaparin is now lower than used in the ESSENCE trial (£10.80/day versus £12.16/day) and some centres have also reduced the dose of enoxaparin in elderly patients (to 0.75mg/kg as opposed to the usual 1mg/kg) which will also lower drugs costs. It is judged likely therefore that administration costs for UFH will still largely offset the difference in drug costs between enoxaparin and UFH. While the magnitude of the estimates of various clinical effects is lower in the meta–analysis of all enoxaparin studies compared with the ESSENCE study alone, the direction of effect remains the same. As such, it is judged likely that enoxaparin would remain a cost-effective treatment option compared with UFH.

Despite these potentially confounding factors the GDG concluded that: there was insufficient evidence to state that enoxaparin is clearly superior to UFH across an unselected population with UA/NSTEMI, but the following supports its superiority in some respects:

  • The meta-analyses showed that enoxaparin is associated with a significant reduction in MI, a composite endpoint (death, MI, urgent revascularisation) or a composite endpoint of death or nonfatal MI.
  • The increase in the minor bleeding outcome had significant heterogeneity suggesting that pooled analysis of these studies should be regarded with caution.
  • LMWH is easy to administer, has a more predictable anticoagulant effect and does not requiring monitoring.
  • The available health economic evidence suggests the use of enoxaparin is cost–effective compared to UFH.
  • The patient/carer representatives of the GDG favoured subcutaneous over intravenous route of administration and thus strongly preferred the use of low molecular weight heparin.
  • There were insufficient data to allow the GDG to make clear recommendations regarding the use of dalteparin. Therefore, the meta-analysis assessing enoxaparin compared with UFH was used to inform recommendations.

4.2. Fondaparinux

4.2.1. Clinical introduction

Fondaparinux is a synthetic pentasaccharide; the first of a new class of synthetic antithrombotics. It binds to anti–thrombin with greater affinity than either UFH or LMWH, and increases the ability of anti–thrombin to inactivate clotting factor Xa. It has 100% bioavailability after subcutaneous administration and has a half-life much longer than UFH or LMWH. Its effects are not reversed by protamine but may be by recombinant factor VIIa135. It has little effect on the activated partial thromboplastin time (aPTT), prothrombin time or bleeding time, and it does not alter fibrinolysis or platelet function (and thrombocytopenia, sometimes seen with UFH and LMWH, is rare). Monitoring can be achieved via an anti-factor Xa assay calibrated with fondaparinux136.

The standard dose for patients with acute coronary syndromes is considered to be 2.5 mg/day subcutaneously. The majority of an administered dose of fondaparinux is excreted unchanged in the urine, with an elimination half-life of 15 to 17 hours. Patients who had serum creatinine levels >265 μmol/l were excluded from the major ACS clinical trial (OASIS-5); it is contraindicated in those with clearance <20 ml/min86.

The clinical questions posed were:

“What is the efficacy and safety of adding a factor Xa inhibitor (fondaparinux) to aspirin in the management of patients with UA or NSTEMI compared to the combination of LMWH/UFH and aspirin therapy?”

“What is the efficacy and safety of adding a synthetic pentasaccharide (fondaparinux and enoxaparin) to aspirin as adjunct therapy to patients with UA/NSTEMI undergoing PCI compared to the combination of LMWH/UFH and aspirin therapy?”

4.2.2. Clinical methodological introduction

The literature was searched from 1999 to 2009 for systematic reviews and RCTs. The rationale for searching from January 1999 onwards was to reflect current practice, particularly the use of stents for revascularisation. Studies were excluded if the population comprised < 60% of people with a diagnosis of non ST-segment elevation ACS. Outcomes of interest were 30 day survival, re-infarction, LV function, revascularisation, quality of life, and serious complications. Primary outcomes assessed earlier than 30 days were also reported.

In the OASIS-5 double blind RCT, patients presenting with UA or NSTEMI (mean age 66 years) were randomised to fondaparinux (N = 10057; 2.5 mg s.c., mean treatment duration 5.4 days) or enoxaparin (N = 10021; 1 mg/kg, twice daily, s.c.; mean treatment duration 5.2 days). Aspirin (97%) and clopidogrel (67%) were administered in both trial arms. Primary outcomes included major bleeding at 9 days or death, MI, or refractory ischemia at 9 days. Secondary outcomes were measured at 30 days 113.

A prospectively determined subgroup analysis of the OASIS-5 RCT compared the efficacy and safety of fondaparinux (N=3134) with enoxaparin (N=3104) in people undergoing PCI within the first eight days of randomisation. People in the enoxaparin group received unfractionated heparin (UFH) if their last dose of enoxaparin was greater than six hours before PCI (65-100 iu/kg depending on whether a GPI had also been given or not). People receiving fondaparinux within six hours of PCI received no additional fondaparinux if they were also on a GPI, or an additional 2.5 mg if they were not. Those who had fondaparinux for more than six hours prior to the PCI received an additional dose of 2.5 to 5 mg depending on whether they received a GPI or not. A protocol amendment advised the consideration of open-label UFH prior to PCI in both trial arms for the last 1758 people undergoing PCI 137.

4.2.3. Health economic methodological introduction

One relevant cost-effectiveness analysis was identified comparing fondaparinux and enoxaparin in UA/NSTEMI patients114.

Sculpher et al.114 reported a cost–utility analysis undertaken from a US direct medical cost perspective based on 180-day effectiveness and resource use data from the OASIS-5 study. A decision analytic model and additional data sources were used to extrapolate beyond the 180-day trial follow-up to estimate lifetime costs and QALYs. The risk of having experienced any of the following key clinical events at 180 days was calculated based on OASIS-5: death, non-fatal MI, non-fatal stroke, combined stroke and MI, major bleed and minor bleed. The cost associated with having each of these key events at 180-days, and of not having any event, were estimated using regression analysis of resource use data from a US subgroup of OASIS-5 (n=759) and US unit costs. Beyond 180 days, long-term mortality rates adjusted for UA/NSTEMI patients were applied in order to estimate life-time costs and QALYs. Higher mortality rates were attributed to those who experienced a non-fatal MI, stroke or both at 180-days. An annual cost of coronary heart disease was applied while patients remain alive. EQ-5D utility weights (US tariff) were applied to life-years in order to calculated QALYs. Lower utility weights were applied to patients who had an MI or a stroke than other patients. Probabilistic sensitivity analysis was used to evaluate uncertainty and a number of other sensitivity analyses were also undertaken.

The study is judged partially applicable to the UK setting. The key issue being the US perspective, in particular there is uncertainty over the applicability of US resource use data and unit costs; although this is at least partially addressed by a sensitivity analysis that used costs based on resource use from all patients instead of only the US subgroup. In addition the US EQ-5D tariff is used and a discount rate of 3%, as opposed to 3.5% recommended by NICE. The study is judged to be of good methodological quality. Catheter-related thrombosis is not incorporated into the analysis but this was judged to be a minor limitation and one that is somewhat addressed by a sensitivity analysis that incorporates an additional cost associated with fondaparinux. This was based on the fact that when randomised treatment was added as a covariate into the cost regression analysis an additional (although non-significant) cost was associated with fondaparinux independent of the clinical events incorporated.

4.2.4. Clinical evidence statements

Fondaparinux versus enoxaparin in people with NSTE ACS

Compared with enoxaparin, fondaparinux significantly:

  • Reduced the risk of major bleeding at nine days (primary safety outcome) (2.2% in fondaparinux group versus 4.1% in enoxaparin group; HR 0.52 [95% CI 0.44 to 0.61], p <0.001) and at 30 days (3.1% in fondaparinux versus 5.0% in enoxaparin: HR 0.62 [95% CI 0.54 to 0.72], p<0.001). Major bleeding was consistently lower with fondaparinux compared with enoxaparin in all groups assessed, regardless of whether UFH was administered before randomisation or not. 113
  • Reduced the composite risk of death, MI, refractory ischaemia, or major bleeding at nine days (7.3% in fondaparinux group versus 9.0% in enoxaparin group: HR 0.81 [95% CI 0.73 to 0.89], p<0.001) and at 30 days (10.2% in fondaparinux group versus 12.4% in enoxaparin group: HR 0.82 [95% CI 0.75 to 0.89], p<0.001) 113.
  • Reduced the risk of death at 30 days (2.9% for fondaparinux versus 3.5% for enoxaparin; HR 0.83 [95% CI 0.71 to 0.97], p=0.02) 113.

Evidence Level: 1++

There was no significant difference between the fondaparinux and enoxaparin groups for 113:

  • Composite risk of death, MI, or refractory ischaemia at 9 days (primary efficacy outcome) (5.8% in fondaparinux group versus 5.7% in enoxaparin group; HR 1.01 [95% CI 0.90 to 1.13])
  • Composite risk of death, MI, or refractory ischaemia at 30 days (8.0% in fondaparinux group versus 8.6% in enoxaparin group; HR 0.93 [95% CI 0.84, 1.02])
  • Composite risk of death or MI at 30 days (6.2% in fondaparinux group versus 6.8% in enoxaparin group;HR 0.90 [95% CI 0.81, 1.01])
  • Risk of MI at 30 days (3.9% in fondaparinux group versus 4.1% in enoxaparin group;HR 0.94 [95% CI 0.82, 1.08])
  • Risk of refractory ischaemia at 30 days (2.2% in fondaparinux group versus 2.2% in enoxaparin group;HR 0.99 [95% CI 0.82, 1.19])
  • Risk of stroke at 30 days (0.7% in fondaparinux group versus 1.0% in enoxaparin group; HR 0.77 [95% CI 0.57, 1.05])

Evidence Level: 1++

Fondaparinux versus enoxaparin in peoplewith NSTE ACS undergoing PCI within the first eight days of randomisation 137

In people undergoing PCI, fondaparinux significantly reduced the:

  • Composite risk of death, MI, stroke, or major bleeding at nine days (8.2% for fondaparinux versus 10.4% for enoxaparin HR 0.78 [95% CI 0.67, 0.93], p=0.004) and at 30 days (9.5% for fondaparinux versus 11.8 % for enoxaparin: HR 0.80 [95% CI 0.69 to 0.93]) 137.
  • Rate of major bleeding at nine days (2.4% for fondaparinux versus 5.1% for enoxaparin; HR 0.46 [95% CI 0.35 to 0.61], p<0.00001) and at 30 days (2.9% for fondaparinux versus 5.4% for enoxaparin; HR 0.52 [95% CI 0.40 to 0.67], p<0.00001) 137.

Evidence Level: 1+

In people undergoing PCI, there was no significant difference between the fondaparinux and enoxaparin groups for: 137

Evidence Level: 1+

An increase in the rate of guiding-catheter thrombus formation was noted with fondaparinux in OASIS-5 (29 episodes [0.9 percent], versus eight episodes with enoxaparin [0.3 percent]; RR 3.59 [95% CI 1.64 to 7.84], p=0.001) - a difference that was observed both before (1.2% vs. 0.3%) and after (0.7% vs. 0.2%) the protocol amendment using unfractionated heparin 113.

Impact of clopidogrel or GPIs on Major Bleeding

In people undergoing PCI, fondaparinux significantly reduced the risk of major bleeding at 30 days compared with enoxaparin when GPIs were used in both groups (N=1198 fondaparinux; N=1263 enoxaparin; HR 0.56 [95% CI 0.39 to 0.81], p=0.002) as well as in the absence of GPIs (N=1874 fondaparinux; N=1842 enoxaparin; HR 0.43 [95% CI 0.30 to 0.63], p<0.0001) 137.

In people undergoing PCI, fondaparinux significantly reduced the risk of major bleeding at 30 days compared with enoxaparin when clopidogrel was used in both groups (N=912 fondaparinux; N=923 enoxaparin; HR 0.45 [0.28-0.72], p=0.001) as well as in the absence of clopidogrel (N=2060 fondaparinux; N=2086 enoxaparin; HR 0.52 [95% CI 0.39 to 0.71], p<0.0001). 137

Evidence Level: 1+

4.2.5. Health economic evidence statements

Sculpher et al.114 found that fondaparinux was a dominant strategy compared to enoxaparin (that is, it was associated with lower costs and higher QALYs) from a lifetime, US perspective. Mean costs were reduced by £121n per person and mean QALYs were increased by 0.04. In probabilistic sensitivity analysis fondaparinux was cost-saving in 82.4% of simulations. At a £32,266 ($50,000) threshold fondaparinux was cost effective in 99.3% of simulations. Costs were also lower with fondaparinux at 180 days at £353.

In sensitivity analysis the model was run for low risk and high risk patients. Fondaparinux was remained a dominant strategy in both groups. The difference in costs and QALYs were smaller in magnitude in low risk patients and bigger in high risk patients. The probability of fondaparinux being cost saving and cost effective were approximately the same as for the overall analysis.

Re-estimating event costs using all patients instead of only US patients increased the 180-day cost-saving to £473n with fondaparinux compared to enoxaparin and fondaparinux remained dominant in the lifetime analysis. A 25% reduction in enoxaparin cost did not change conclusions.

Including a randomised treatment covariate in the cost regression showed a non-significant additional cost associated with fondaparinux independent of clinical events. When this was incorporated into the lifetime analysis fondaparinux remained dominant in high risk patients and had an ICER of £1510n per QALY gained overall and £4486n per QALY gained in low risk patients.

No UK analyses were identified. In terms of drug costs alone: fondaparinux costs £6.41 per day; enoxaparin costs £10.38 per day (assuming dose of 1mg/kg and weight of 80kg)139. Given that the clinical evidence suggests a benefit of fondaparinux over enoxaparin, it was judged likely that the conclusion that use of fondaparinux is cost effective compared to enoxaparin would be maintained if a UK perspective were taken.

Health economic modelling

As drug costs were lower with fondaparinux than enoxaparin, and the clinical evidence supported improved outcomes with fondaparinux, there was considered to be low uncertainty that fondaparinux would be cost effective and it was judged a low priority to conduct a modelling study to analyse this. However, it was of interest to consider how use of fondaparinux instead of enoxaparin might impact the comparisons made in the model regarding use of GPIs and bivalirudin. As such fondaparinux was incorporated into the cost-effectiveness analysis undertaken for the guideline.

A cost–utility analysis was undertaken with costs and quality-adjusted life-years (QALYs) considered over patients' lifetimes from a UK NHS perspective.

This compared the following treatment strategies in the acute management of UA/NSTEMI (heparin baseline):

In addition the analysis was run as above but with fondaparinux substituted for heparin in the first three arms (fondaparinux baseline). Fondaparinux was not incorporated in the bivalirudin arms in this analysis as there is no experience with these agents combined and so it was not judged appropriate. The analysis incorporated 1-year death, MI and post-acute revascularisation, and inhospital bleeding.

As comparing fondaparinux and enoxaparin was not the primary objective of the analysis, some issues relating to this comparison may not have been captured fully. For example, catheter-related thrombosis was not incorporated into the model. This however is considered unlikely to impact conclusions. In addition, the six-month relative risks from OASIS-5 were assumed to hold at one year as the studies used for the main comparisons in the model all had one-year follow-up. However, it was possible to compare fondaparinux and enoxaparin in the model – costs were reduced and QALYs increased with fondaparinux. This is consistent with the published analysis based on the OASIS-5 study114.

For the full analysis methods, detailed results and discussion see the report in Appendix C. A summary is provided in the GPI and bivalirudin chapters.

4.2.6. Evidence summary

A preliminary investigation (the PENTUA Study140) had shown that fondaparinux and enoxaparin had similar efficacy when used in acute coronary syndromes. This finding led to the large OASIS-5 trial141, which randomised over 20,000 patients with UA or NSTEMI to receive either fondaparinux (2.5 mg, once daily subcutaneously) or enoxaparin (1 mg/kg, twice daily subcutaneously (reduced to 1mg/kg once daily if creatinine clearance <30ml/min), for a mean duration of 5.3 days. People were excluded from the trial if their creatinine was >265 μmol/l. Over 60% of patients underwent cardiac catheterisation, and over 30% had PCI. Aspirin was given to 97% of patients, and clopidogrel to 67% in both arms of the trial. GPIs were given to 41% of those undergoing PCI (N=6239). GPI use was not reported for the entire trial population.

At nine days the composite end point of death, MI or refractory ischaemia was no different between the fondaparinux and enoxaparin groups, indicating non-inferiority of fondaparinux with respect to efficacy. However, fondaparinux was associated with a significantly lower rate of major bleeding (2.2% for fondaparinux and 4.1% for enoxaparin; HR 0.52, 95% CI, 0.44 to 0.61, p<0.001), indicating superiority of fondaparinux with respect to safety. This reduction in bleeding occurred irrespective of whether a GPI was administered or not. The composite end point of death, MI, refractory ischemia, or major bleeding occurred in 7.3% of the patients in the fondaparinux group, compared with 9.0% of the patients in the enoxaparin group (HR 0.81; 95% CI 0.73 to 0.89; P<0.001) at nine days and this difference was shown to persist to 180 days.

Regardless of treatment, patients who had major bleeding in hospital had significantly higher rates of death (13.2% vs. 2.8%), re-infarction (11.9% vs. 3.6%), and stroke (3.5% versus 0.7%) at 30 days (P<0.001), than patients without bleeding. The mortality rate among those who had minor bleeding was also higher at 30 days than among those with no bleeding episodes (6.9% vs. 2.8%), and these higher event rates associated with bleeding persisted after the authors adjusted for the various clinical characteristics associated with bleeding.

A US health economic analysis based on OASIS-5 estimated lifetime costs and QALYs and found that fondaparinux reduced costs and increased QALYs114. No UK analyses were identified but fondaparinux has lower daily drug costs than enoxaparin and improves clinical outcomes. Modelling undertaken for the guideline also found that fondaparinux is likely to be cost-saving and improve clinical outcomes – although it is noted that this comparison was a secondary objective of the analysis.

Fondaparinux and PCI

In a pre-specified sub-group analysis of over 3,000 patients undergoing PCI in OASIS-5137 fondaparinux significantly reduced the risk of major bleeding at nine days (HR 0.48 [0.31–0.72], p<0.0005), and minor bleeding at nine days (HR 0.38 [0.25–0.58] p<0.00001) compared to enoxaparin in people who had their PCI within the first 24 hours of randomization. In this PCI sub-group there was no significant difference in risk of death, MI or stroke at nine days. Major bleeding at 30 days was significantly reduced in the fondaparinux group whether or not clopidogrel, and/or a GPI, were used.

Depending on the timing of the most recent administration of the active agent, some patients in the enoxaparin group of OASIS-5 received additional UFH, with or without a GPI, and some in the fondaparinux group received an additional dose of fondaparinux, the dose of which depended on whether a GPI was given or not. In the enoxaparin group 55% received additional UFH, whereas only 20.8% of the fondaparinux group did so. It is possible that this difference in administration of additional heparin contributed to the observation of higher bleeding in the enoxoparin arm of the trial.

Isolated reports of catheter thrombosis in a small number of cases (0.9% for fondaparinux group vs 0.4% for enoxaparin group) resulted in a protocol amendment that detailed the correct method of administration of the intravenous study drug and emphasized the importance of flushing all catheters and the intravenous line to ensure that the entire bolus of the study drug (which was 0.5 ml for fondaparinux) reached the patient, since it was considered possible that catheter thrombosis may have been due to incomplete administration. In addition, centres were reminded that, at the investigator's discretion, it was permissible to give open-label UFH before PCI in addition to the protocol-mandated study drug142. Unlike UFH and enoxaparin, fondaparinux does not inhibit the contact clotting activation pathway (involving clotting factors XII, XI)143 and this may be a possible explanation for its association with increased catheter thrombosis.

The authors of the OASIS-5 PCI sub-study 142 concluded that upstream fondaparinux is superior to enoxaparin in terms of net clinical benefit, but they recommended that “in fondaparinux-treated patients, UFH rather than intravenous fondaparinux be used as adjunctive therapy at the time of PCI”. They also noted that the protection provided against catheter thrombus by adding conventional doses of UFH to fondaparinux or enoxaparin did not increase the risk of major bleeding in either randomized treatment group, and that the substantial benefit of upstream fondaparinux in reducing bleeding was therefore maintained.

4.2.7. Evidence to recommendations

The GDG noted that the evidence was dependent on the result of a single randomised controlled trial (OASIS-5). However, this involved over 20,000 patients and was felt to be of high quality. It showed benefit of fondaparinux compared to enoxaparin with an overall reduction in major bleeding and mortality, and the reduction in bleeding risk was apparent in various subsets of patients (those undergoing PCI, those with and without clopidogrel, those with and without treatment with concomitant GPIs). Fondaparinux requires once daily administration and does not require weight adjustment, unlike enoxaparin which requires twice daily administration and is weight dependent. A US perspective analysis based on OASIS-5 found fondaparinux to be associated with lower costs and higher QALYs than enoxaparin. While no UK analyses were identified its current price is lower than enoxaparin (fondaparinux £6.41 per day, enoxaparin approximately £10 per day [assuming an average weight of 80kg, the dose is 80mg twice daily])139. One would therefore expect fondaparinux to be dominant over enoxaparin in any cost-effectiveness analysis.

However, the GDG noted the observation that use of fondaparinux alone at the time of a PCI procedure is associated with a small increase in catheter-related thrombosis (that did not translate into an increased risk of clinical events), and the recommendation of the trial's authors to give unfractionated heparin, rather than additional fondaparinux, at the time of a PCI procedure. International guidelines7 have suggested using a bolus of 50-100 units/kg of UFH for those previously given fondaparinux and undergoing PCI, whereas the OASIS investigators suggested 50 to 60 units/kg. There is insufficient evidence to make a recommendation regarding the exact dose of supplemental UFH that should be used. Operators should regard the range of 50-100 units/kg as a guide and decide the dose on an individual patient basis, considering the timing of the most recent dose of fondaparinux (< or > 6 hours), the concomitant use of a GPI, and the balance between underlying ischaemic risk and potential for bleeding. In routine clinical practice it is common for interventionists currently to miss the morning dose of enoxaparin for patients going to the catheter laboratory and to use UFH during PCI (UFH is used during the procedure in most patients) and therefore the addition of UFH to fondaparinux is unlikely to have a significant impact on any cost-benefit assessment

OASIS-5 confirmed the importance of bleeding as a predictor of adverse outcome and the need for clinicians to be aware of this association when patients with UA or NSTEMI are offered combinations of anti–platelet and anti–thrombin agents. It excluded people with a creatinine of >265 μmol/l, and renal dysfunction is known both to increase the risk of an adverse cardiovascular event, and also of bleeding144. A subsequent analysis of OASIS-5 indicated that the benefit of fondaparinux over enoxaparin was actually greatest in those with the most renal impairment (glomerular filtration rates (GFR) of <58 mls/min)145. It would therefore be illogical to use dose-adjusted enoxaparin146 as an alternative to fondaparinux for those with greater degrees of renal impairment (who were excluded from OASIS-5), especially as it is known that such dose adjustment is often not undertaken appropriately in practice147. Unfractionated heparin, with dosage guided by monitoring of blood clotting would be a more logical alternative to fondaparinux where there is particular clinical concern regarding bleeding risk.

The GDG concluded that:

  • The use of enoxaparin in patients with UA or NSTEMI is a cost-effective treatment when compared to UFH, and is easier to administer.
  • Fondaparinux has been shown to be superior in clinical outcome to enoxaparin, particularly with respect to its lower bleeding risk.
  • Clinicians should carefully consider factors (such as renal impairment) which increase bleeding risk. Unfractionated heparin, with dose adjustment guided by monitoring of clotting function, is an alternative to fondaparinux for those with renal impairment excluded from OASIS-5 (creatinine >265 μmol/l).
  • People on fondaparinux undergoing PCI should receive unfractionated heparin, and not additional fondaparinux, at the time of the procedure.
  • Fondaparinux is likely to be dominant (cost saving and improved health outcomes) compared to enoxaparin.

4.3. Bivalirudin

4.3.1. Clinical introduction

Hirudin is a naturally occurring substance secreted by leeches which has a powerful anticoagulant effect. It is a natural inhibitor of thrombin and has some potential advantages over heparin; it does not interact with other serum proteins, and it has the ability to lyse existing thrombus, unlike heparin which acts only on soluble thrombin. As it is difficult to extract large amounts of hirudin from natural sources, and hirudin was shown to be associated with a risk of increased bleeding, synthetic analogues were developed. The only one of these analogues that is licensed in the UK for use in acute coronary syndromes is bivalirudin.

Bivalirudin is a direct inhibitor of soluble and clot-bound thrombin. It has a rapid onset of action and has a half-life of 25 minutes, so is given as an intravenous infusion. It is cleared principally by proteolytic cleavage, but a significant component is also cleared by renal excretion.

In current UK practice bivalirudin is used as an anticoagulant during percutaneous coronary intervention (PCI), as an alternative to the combination of heparin and a GPI, (initiated at the time of PCI). It is also approved for use in UA/NSTEMI patients planned for urgent or early invasive intervention (coronary angiography with PCI/CABG/medical management as indicated), and is initiated prior to angiography in combination with aspirin and clopidogrel and continued through PCI in those who undergo this procedure.

When initiated at PCI, recommended dosing is an initial bolus of 0.75mg/kg and an infusion of 1.75mg/kg/hr during the PCI. Following PCI, an infusion of 0.25mg/kg/hr can be optionally continued if clinically appropriate.

When initiated pre-angiography, recommended dosing is a bolus of 0.1mg/kg and infusion of 0.25mg/kg/hr. If the patient continues to PCI following angiography an additional bolus of 0.5mg/kg is administered and an infusion of 1.75mg/kg/hr is used during the PCI. Following PCI, an infusion of 0.25mg/kg/hr can be optionally continued but is generally not required. For patients who are managed medically or go on to CABG following angiography, the infusion can also be optionally continued.

The GDG aimed to assess the clinical and cost-effectiveness of bivalirudin and therefore asked the following questions around which the literature was searched:

“What is the efficacy and safety of adding a thrombin inhibitor (bivalirudin) to aspirin, with or without a GPIIb/IIIa inhibitor, in the management of patients with UA or NSTEMI compared to the combination of LMWH/UFH, and aspirin, with or without a GPIIb/IIIa inhibitor?”

“What is the efficacy and safety of adding a thrombin inhibitor to aspirin with or without a GPIIb/IIIa inhibitor as adjunct therapy to patients with UA/NSTEMI undergoing PCI compared to the combination of LMWH/UFH, aspirin, and a GPIIb/IIIa inhibitor?”

4.3.2. Methodological introduction

The literature was searched from 1999 to 2009 for RCTs and systematic reviews comparing direct thrombin inhibitors in combination with aspirin, with or without a GPI compared with the combination of LMWH/UFH, aspirin, with or without a GPI in people with non ST-segment elevation ACS. The rationale for searching from January 1999 onwards was to reflect current practice,, and that bivalirudin is the only licensed direct thrombin inhibitor and was not available before then.

Outcomes of interest were thirty day survival, re-infarction, LV function, revascularisation, quality of life, and serious complications. For a study to be included at least 60% of patients enrolled needed to have a diagnosis of non ST-segment elevation ACS or the study had to report outcomes in a non ST-segment elevation ACS subgroup.

Three systematic reviews 148 149 150 and four RCTs 151 44 152 110 were identified which compared a direct thrombin inhibitor to heparin in ACS patients. However, one RCT 44 152 and two meta-analyses 149 150 assessing hirudin were rejected as this drug does not have a license for ACS. The third systematic review 148 was rejected because study quality was not appraised, and three of the five RCTs included in the meta-analysis (BAT153, REPLACE-1 154 CACHET 155) had populations containing < 60% unstable angina or NSTEMI.

Two RCTs, ACUITY 110,156and an ACS subgroup of REPLACE-2 151 112, were identified that addressed the use of bivalirudin in people with non ST-segment elevation ACS.

Bivalirudin initiated before angiography: ACUITY RCT

The ACUITY open-label RCT 110 recruited patients (N=13819) with UA (41%) or NSTEMI (59%) who were scheduled for an early invasive strategy (angiography within 72 hours). Following angiography patients were triaged to PCI, CABG or continued medical management alone. In the first randomization, people were randomised to one of three arms:

In the second randomization and only within the heparin plus GPI and the bivalirudin plus GPI arms, patients were randomised to either upstream GPI use (where all patients received early GPI- either tirofiban or eptifibatide) or deferred GPI use (where only patients who went on to PCI received GPI and only during the PCI; patients received abciximab or eptifibatide).

The length of time from antithrombotic study drug to angiography was 4.0 h (median) and to PCI was 4.1 h (median).

All trial participants were given aspirin (daily dose 300-325 mg orally or 250 to 500 mg iv).. Clopidogrel (dose and timing) was left to investigator discretion; although a 300 mg or greater loading dose was required in all people undergoing PCI no later than two hours following their procedure.

Bivalirudin was given as an initial bolus of 0.10 mg/kg, then 0.25 mg/kg per hour continued through angiography. Dosing of bivalirudin beyond angiography depended on the type of management strategy: PCI, CABG or medical management. If PCI followed, an additional bolus of bivalirudin (0.5 mg/kg) was given and the infusion was increased to 1.75 mg/kg per hour and no post-PCI infusion dose was recommended, although 0.25 mg/kg per hour for 4 to 12 hours could be used (in the absence of a GPI) at operator discretion. Full details of the study design and doses of antithrombotic agents were reported in a prior publication 157

For upstream GPIs, either tirofiban (0.4 microgram/kg/minute infusion for 30 minutes followed by 0.1 microgram/kg/minute infusion) or eptifibatide (180 microgram/kg bolus plus 2.0 microgram/kg/minute infusion) were started immediately. If PCI was to follow, the same GPI was to be used during PCI and discontinued 12-18 hours later. The infusion was typically discontinued in people triaged to CABG or those to medical management, although the infusion could be maintained if clinically indicated.

For those randomised to downstream GPIs, either abciximab (0.25 microgram/kg bolus plus 0.125 microgram/kg/minute infusion, with a maximum of 10 microgram/min) or eptifibatide (180 microgram/kg bolus plus 2.0 microgram/kg/minute infusion, with a second bolus given in ten minutes) were administered only to those people getting PCI, begun 5 to 10 minutes prior to the balloon inflation, and continued for 12 hours (abciximab) or 12 to 18 hours (eptifibatide) thereafter.

A pre-specified subgroup analysis of people undergoing PCI in the ACUITY trial, 156 (N=5170) compared:

And also:

The current SPC for bivalirudin states that bivalirudin is indicated “for the treatment of adult patients with acute coronary syndromes (unstable angina/non-ST segment elevation myocardial infarction (UA/NSTEMI)) planned for urgent or early intervention. Bivalirudin should be administered with aspirin and clopidogrel”. After stakeholder consultation on the draft guideline, it was deemed appropriate to present unpublished data from the Medicines Company in a subgroup of people who received clopidogrel either before angiography or before PCI (N=8677). The rationale for this was that data from the clopidogrel subgroup was reviewed by the EMEA when it undertook its licensing review of bivalirudin. The EMEA had concerns regarding a numerical increase in ischemic events in the bivalirudin alone arm in certain patient groups in the ACUITY trial and so requested identification of a target group where the benefit/risk profile was clearly positive158. A subgroup using aspirin and clopidogrel was identified, with clopidogrel given pre-angiography or pre-PCI. A published subgroup analysis of the ACUITY trial 159assessed outcomes in people in a different clopidogrel subgroup - people who received clopidogrel any time before angiography or peri-PCI were compared with people who received clopidogrel more than thirty minutes after PCI or not at all. This subgroup analysis was rejected because it was felt that the clopidogrel subgroup on which the EMEA made its licensing decision was the subgroup that was more appropriate and more clinically relevant.

The evidence statements below for the ACUITY trial refer to the group of people who received clopidogrel either before angiography or before PCI. Tables summarise the relative risks in this clopidogrel subgroup as well as in the entire published ACUITY trial population. Three primary 30-day end points were prespecified; a composite ischaemia endpoint (death from any cause, myocardial infarction, or unplanned revascularisation for ischaemia), major bleeding (not related to CABG), and a net clinical outcome endpoint (defined as the occurrence of the composite ischaemia end point or major bleeding).

Bivalirudin initiated after angiography and before PCI: REPLACE-2

In people undergoing PCI, the double blind REPLACE-2 RCT 151 (N=6,010) compared:

The population in REPLACE-2 had a low proportion of people with ACS (defined as unstable angina within preceding 48 h or MI within previous 7 days; N=1351), but was included because results in the ACS subgroup were reported separately 112,151. The primary outcome was a quadruple composite outcome of death, MI, urgent revascularization or major bleeding by 30 days, but the ACS subgroup is too small (underpowered) to reliably detect a difference in most outcomes.

Bivalirudin dosing in the REPLACE-2 RCT was different from the ACUITY RCT. In REPLACE-2, bivalirudin was given as a bolus of 0.75 mg/kg prior to the start of PCI, followed by infusion of 1.75 mg/kg/hour for the duration of the procedure. The median duration of bivalirudin infusion was 0.73 hours (IQR 0.43 – 1.33 hours).

In the heparin plus planned GPI arm, a heparin bolus (65 U/kg, maximum 7000 U) was given prior to PCI, with either abciximab (0.25 mg/kg bolus, 0.125 microgram/kg/minute infusion for 12 hours) or eptifibatide (two boluses of 180 microgram/kg boluses given ten minutes apart, followed by 2.0 microgram/kg/minute infusion for 18 hours). The median duration of eptifibatide infusion was 18.0 hours (IQR 16.5-18.1 hours) and the median duration of abciximab infusion was 12.0 hours (IQR 11.9-12.2 hours). Aspirin was given to all; pre-treatment with clopidogrel (300 mg loading dose) was encouraged two to twelve hours before the interventional procedure.

4.3.3. Clinical evidence statements

Bivalirudin infusion initated before angiography: ACUITY trial

Heparin + GPI versus Bivalirudin + GPI in people who received clopidogrel before angiography or before PCI: Outcomes at 30 days (refer to Table 4-2)
Table 4-2. Outcomes at 30 days in the ACUITY RCT for people with NSTEMI or UA randomised to bivalirudin + GPI or heparin + GPI in the entire trial population, and also in the subgroup of people who received clopidogrel before angiography or before PCI .

Table 4-2

Outcomes at 30 days in the ACUITY RCT for people with NSTEMI or UA randomised to bivalirudin + GPI or heparin + GPI in the entire trial population, and also in the subgroup of people who received clopidogrel before angiography or before PCI .

In those who received clopidogrel (before angiography or before PCI) 109 there was no significant difference between people randomised to heparin + GPI versus Bivalirudin + GPI for:

  • death/MI/unplanned revascularisation
  • death/MI/unplanned revascularisation/major bleeding
  • death or MI
  • death
  • Unplanned revascularisation
  • All major bleeding
  • major bleeding not related to CABG
  • minor bleeding not related to CABG
  • major TIMI bleeding
  • minor TIMI bleeding

Evidence Level 1+

Heparin + GPI versus Bivalirudin alone in people who received clopidogrel before angiography or PCI 109: Outcomes at 30 days (refer to Table 4-3)
Table 4-3. Outcomes at 30 days in the ACUITY trial for people with NSTEMI or UA randomised to bivalirudin alone or heparin + GPI in the entire trial population, and also in the subgroup who received clopidogrel before angiography or before PCI .

Table 4-3

Outcomes at 30 days in the ACUITY trial for people with NSTEMI or UA randomised to bivalirudin alone or heparin + GPI in the entire trial population, and also in the subgroup who received clopidogrel before angiography or before PCI .

In the clopidogrel subgroup and compared with people randomised to heparin + GPI, people randomised to bivalirudin alone had a significantly:

  • decreased risk of death/MI/unplanned revascularisation/major bleeding
  • decreased risk of all major bleeding
  • decreased risk of major bleeding not related to CABG
  • decreased risk of minor bleeding not related to CABG
  • decreased risk of TIMI major bleeding
  • decreased risk of TIMI minor bleeding

Evidence Level 1+

In the clopidogrel subgroup 109 there was no significant difference between people randomised to heparin + GPI versus Bivalirudin alone for:

  • death/MI/unplanned revascularisation
  • death/MI
  • death
  • unplanned revascularisation

Evidence Level 1+

However, in subgroup analysis of people who did not receive a thienopyridine anti–platelet agent (such as clopidogrel) before angiography (N=3304), the bivalirudin alone group had a significantly increased risk of death/MI/ unplanned revascularisation compared with the heparin + GPI group (RR 1.29 [1.03, 1.63]) 110

Evidence Level 1+

Subgroup analysis of people undergoing PCI in the ACUITY trial

(refer to Table 4-4)

Table 4-4. Subgroup analysis of the ACUITY trial: Outcomes at 30 days for people undergoing PCI randomised to bivalirudin alone, or bivalirudin + GPI, or heparin + GPI, as well as for the PCI subgroup who received clopidogrel (before angiography or before PCI) .

Table 4-4

Subgroup analysis of the ACUITY trial: Outcomes at 30 days for people undergoing PCI randomised to bivalirudin alone, or bivalirudin + GPI, or heparin + GPI, as well as for the PCI subgroup who received clopidogrel (before angiography or before PCI) . (more...)

In people undergoing PCI who received clopidogrel (before angiography or before PCI) 109, there was no significant difference between the bivalirudin + GPI and the heparin + GPI inhibitor groups for:

  • death/MI/unplanned revascularisation
  • death/MI/unplanned revascularisation/major bleeding
  • death
  • Unplanned revascularisation
  • major bleeding not related to CABG
  • minor bleeding not related to CABG
  • major TIMI bleeding
  • minor TIMI bleeding

Evidence Level 1+

In people undergoing PCI who received clopidogrel (before angiography or before PCI) 109, there was no significant difference between the bivalirudin group and the heparin + GPI group for:

  • death/MI/unplanned revascularisation
  • death/MI
  • death
  • Unplanned revascularisation

Evidence Level 1+

In people undergoing PCI who received clopidogrel 109 those randomized to bivalirudin alone (compared with heparin plus GPI) had a significantly decreased risk of:

  • death/MI/unplanned revascularisation/major bleeding
  • major bleeding not related to CABG
  • minor bleeding not related to CABG
  • all major bleeding
  • major TIMI bleeding
  • minor TIMI bleeding

Evidence Level 1+

Bivalirudin initiated after angiography and before PCI: REPLACE-2

(ACS subgroup results only; refer to Table 4-5)

Table 4-5. Outcomes at 30 days for people with ACS (defined as unstable angina within preceding 48 h or MI within previous 7 days) undergoing PCI randomised to bivalirudin + provisional GPI or heparin + planned GPI in the REPLACE -2 trial .

Table 4-5

Outcomes at 30 days for people with ACS (defined as unstable angina within preceding 48 h or MI within previous 7 days) undergoing PCI randomised to bivalirudin + provisional GPI or heparin + planned GPI in the REPLACE -2 trial .

In the ACS subgroup undergoing PCI in the REPLACE-2 RCT, there was no significant difference between patients randomised to heparin plus planned GPI versus those assigned bivalirudin plus provisional GPI for:

  • death/MI/urgent revascularization/major bleeding
  • death/MI/urgent revascularisation
  • death/MI
  • death
  • Urgent revascularization
  • Major bleeding

Evidence Level 1+

In the REPLACE-2 RCT, people with ACS undergoing PCI who were randomised to bivalirudin plus provisional GPI had a significantly reduced risk of:

  • minor bleeding

Evidence Level 1+

4.3.4. Health economic methodological introduction

Two economic evaluations were identified from the literature that addressed the use of bivalirudin in people with UA/NSTEMI. One based on the ACUITY study160 and one based on the ACS subgroup of the REPLACE-2 study112 – the ACUITY study and REPLACE-2 study design and outcomes are described in detail in the clinical evidence section above. One economic evaluation submitted during consultation was also included (this was unpublished but had been presented at conference and submitted for publication); this was based on the ACUITY study with a UK perspective161. In addition five studies were identified, taking non-UK perspectives, that examined the clinical question but the population of the clinical studies used to inform effectiveness did not meet the population cut off of >60% UA/NSTEMI and so were not reviewed162-166. One study was not reviewed as it was judged to have serious methodological limitations in terms of the outcome measure used and the method for calculating relative treatment effects167.

Bivalirudin initiated before angiography: ACUITY-based analyses

Two studies were identified relating to use of bivalirudin initiated before angiography in UA/NSTEMI, based on the ACUITY study160,161.

Pinto et al. 160 reported an economic evaluation based on resource use and outcomes from a US subgroup of the ACUITY study (n=7,851). A US healthcare system perspective was taken. In-hospital and 30-day costs were estimated based on resource use from the trial and US unit costs. The ACUITY study was an early angiography population (where all patients received angiography and were then triaged to PCI, CABG or continued medical management alone). It had two stages of randomisation: first to bivalirudin monotherapy (provisional GPI use allowed), heparin plus GPI, or bivalirudin plus GPI. In addition, within the heparin plus GPI and the bivalirudin plus GPI arms, patients were further randomised to either upstream GPI use (where all patients received early GPI) or deferred GPI use (where only patients who went on to PCI received GPI and only during the PCI). Resource use and costs were presented for the upstream and deferred PCI patients groups separately in this analysis. Disaggregated costs and events were presented (i.e. there was no cost–effectiveness ratio reported).

The Pinto et al. study is judged partially applicable to the UK with potentially serious limitations. There is uncertainty regarding the applicability of US resource use and costs to the UK. The ACUITY study has short times to intervention that may not represent UK practice. Resource use may also be impacted by the trial setting. The study used a short time-horizon, does not use QALYs and does not estimate a cost-effectiveness ratio.

Schwenkglenks et al.161 reported an economic evaluation based on a decision analytic model using data from a clopidogrel subgroup of the ACUITY study (defined by clopidogrel use at any point during the index hospitalisation). A UK NHS perspective was taken and lifetime costs and QALYs were estimated. Two analyses were presented: 1) bivalirudin (bailout GPI use allowed) versus heparin plus planned GPI (50% upstream use, 50% deferred selective during PCI); 2) bivalirudin (bailout GPI use allowed) versus heparin plus GPI use during PCI. For each analysis two scenarios were examined a) the whole population and b) a subgroup at high bleeding risk (defined as having at least two of the following risk factors for bleeding: age ≥65 years, female gender, renal impairment, baseline haemoglobin <12mg/dL [women] or <13mg/dL [men], weight <60kg, diabetes).

The Schwenkglenks et al. study is judged to be partially applicable to the UK setting. The perspective taken is inline with the NICE base case and as such is more relevant than the Pinto et al. study above. However, there is uncertainty regarding the applicability of the outcomes based on the ACUITY study to the wider UK setting due to short times to intervention in the ACUITY study. The study is generally judged to be of good methodological quality; however there are some potentially serious limitations that may impact results. Differences in QALYs between treatments are only impacted by the difference in the relative risk of mortality between bivalirudin and heparin+GPI; new non-fatal MI events do not impact QALYs. The relative risk for mortality (which drives the QALY difference) in the clopidogrel subgroup used in this model (clopidogrel at any time during index hospitalisation) is slightly more favourable to bivalirudin and with a narrower confidence interval than that from the pre-angiography/pre-PCI clopidogrel subgroup selected as the most appropriate by the GDG.

Bivalirudin initiated after angiography and before PCI: REPLACE-2 ACS subgroup-based analyses

One study was identified assessing bivalirudin initiated after angiography and before PCI in UA/NSTEMI 112. Rajagopal et al. reported an economic evaluation based on resource use and outcomes from an ACS subgroup (n=1351) from the REPLACE-2 trial (63% UA, 37% unspecified MI). All patients in the study underwent PCI. A US hospital perspective was taken in terms of costs. The study compared bivalirudin (with provisional GPI) to heparin with planned GPI in patients undergoing PCI with ACS. 30-day costs and 30-day, six-month and one-year outcomes in terms of events (death, MI, revascularisation, major and minor bleeding) were reported. Disaggregated costs and events were presented (i.e. no cost-effectiveness ratio was reported).

The study is judged partially applicable to the UK. There is uncertainty regarding the applicability of international resource use and US costs to the UK. Resource use may also be impacted by the trial setting. The study used a short time-horizon, did not use QALYs and did not estimate a cost–effectiveness ratio. The unit costs used were not reported.

4.3.5. Health economic evidence statements

Direct thrombin inhibitors in medical management

Pinto et al. 160 presented disaggregated costs and outcomes. Costs are summarised in Table 4-6 below. Total costs were lowest in the bivalirudin monotherapy arm. A significant difference is reported across all five arms of the trial (this includes the bivalirudin + GPI arm which had the highest costs); pair-wise significance tests were not reported. Only in-hospital health outcomes were reported in this subgroup analysis. These are reported as consistent with the full trial analysis where significant differences are seen in terms of bleeding endpoints with bivalirudin monotherapy but no significant difference is seen across the groups in terms of ischemic endpoints. 30-day results for the US subgroup are reported as ‘similar’ but not presented. No cost-effectiveness ratio is presented but the authors interpret the evidence to suggest that bivalirudin offers similar ischemic protection with lower bleeding and lower costs.

Table 4-6. ACUITY US subgroup analysis costs.

Table 4-6

ACUITY US subgroup analysis costs.

Schwenkglenks et al.161 reported an incremental cost-effectiveness ratio for bivalirudin (bailout GPI use allowed) versus heparin plus planned GPI (50% upstream use, 50% deferred selective during PCI) in patients with UA/NSTEMI undergoing an early invasive strategy of £10,009 per QALY gained. This was reduced to £3750 per QALY gained in a high bleeding risk subgroup. Bivalirudin was cost-effective in 72% and 89% of simulations respectively, at a threshold of £20,000 per QALY gained. The conclusion that bivalirudin is cost effective compared to use of heparin plus GPI was also robust to a range of deterministic sensitivity analyses. A sensitivity analysis was not carried out to address the possibility that non-fatal MI events may have a short-term or long-term impact on QALYs. The parameters with the strongest influence on results were reported as the relative risk of death and the index hospitalisation length of stay.

When bivalirudin (bailout GPI use allowed) was compared to heparin plus GPI use during PCI in an early invasive UA/NSTEMI population only, the incremental cost-effectiveness ratio was lower (that is cost effectiveness was improved) at £4514 per QALY gained in the overall population and £3,416 per QALY gained in the high bleeding risk subgroup.

Bivalirudin initiated after angiography and before PCI: REPLACE-2 ACS subgroup

Rajagopal et al.112 presented disaggregated costs and outcomes. Taking a 30-day perspective bivalirudin (plus provisional GPI) compared with heparin plus planned GPI reduced costs by £245 and reduced the rate of the composite of death, MI, urgent revascularisation and major bleeding, although not significantly. However, disaggregated results show that MI and urgent revascularisation were numerically, but non-significantly more frequent with bivalirudin, there was no difference in death, major bleeding was non–significantly less frequent and minor bleeding was significantly less frequent. Without extrapolation of these events to overall outcomes (e.g. life years or QALYs) it is difficult to interpret these results. The US setting limits its UK applicability.

Health economic modelling

Cost effectiveness modelling was undertaken for the guideline to look at the use of GPIs taking into account contemporary management In particular it addressed the use of GPIs in combination with clopidogrel, bivalirudin was included as a possible alternative to heparin plus a GPI, and fondaparinux as an alternative to heparin was incorporated.

For the full analysis methods, detailed results and discussion see the report in Appendix CAppendix C. A summary is provided below.

Methods

A cost–utility analysis was undertaken with costs and quality-adjusted life-years (QALYs) considered over patients' lifetimes from a UK NHS perspective. The analysis is primarily relevant to patients undergoing an early invasive management approach – that is coronary angiography with revascularisation if indicated – because trial results for GPIs and bivalirudin used in the analysis were only relevant to a population undergoing angiography. This is discussed in more detail in the full report in Appendix C.

This analysis compared the following treatment strategies in the acute management of UA/NSTEMI (heparin baseline):

In addition the analysis was run as above but with fondaparinux substituted for heparin in the first three arms (fondaparinux baseline). Fondaparinux was not incorporated in the bivalirudin arms in this analysis as there is no experience with these agents combined.

Cost effectiveness was analysed by six risk subgroups, as summarised in Table 4-7 below. The creation and interpretation of these risk groups is discussed in more detail in the Risk chapter of the guideline (section 2) and the report of the analysis of MINAP data for the cost effectiveness analysis (Appendix B).

Table 4-7. Risk groups.

Table 4-7

Risk groups.

The general approach taken was to obtain contemporary UK estimates of events for the aspirin, clopidogrel and heparin arm of the model from recent MINAP (the national audit of ACS management) data. These were stratified by acute management strategy: PCI, CABG, angiography only. Where inputs were not available from MINAP, data were sourced from the literature or discussion with the GDG. One-year death, MI and revascularisation, and in-hospital bleeding were incorporated. The effects of different treatment combinations are then modelled by applying relative risks from randomised controlled trials identified by the systematic review of the clinical literature for the guideline – one-year relative risks were used where available except for bleeding. Relative risks were applied to the appropriate part of the population; for example, only PCI patients, if only relevant to these patients.

Lifetime QALYs were estimated based on one-year status: dead, alive having had a new MI, alive without new MI. At one-year patients were attributed a number of life-years based on this status. Those alive at one year with new MI were attributed a lower estimate than those alive without new MI. Life-years were adjusted by a quality of life weight for people with ACS to estimate QALYs. As the rates of death and MI will vary with treatment strategy, so will the QALYs.

Lifetime costs were estimated taking into account initial drug treatment costs, the cost of MI, bleeding and revascularisation events up to one year and average disease-related costs incurred if alive post one-year.

Treatment effects were based on studies identified in the clinical review. Only studies with at least 50% clopidogrel use were used. Relative treatment effects were based on the following studies:

The Early ACS trial also compares upstream GPI vs PCI GPI use in an early angiography UA/NSTEMI population100. It was published late in the guideline development process and only reports 30-day outcomes, whereas the model was developed with one-year event rates and effectiveness data. Sensitivity analyses examined the possible impact of this study.

The model was built probabilistically in order to take account of the uncertainty around input parameter point estimates. Probability distributions in the analysis were based on error estimates from data sources, for example confidence intervals around relative risk estimates. Various one-way and scenario sensitivity analyses, where one or more inputs were varied, were undertaken to test the robustness of model assumptions and data sources.

Two analyses were run:

3.

Trial aligned analysis (costing based on bivalirudin vial usage in the ACUITY trial with a pre-angiography treatment period median 4hrs/mean 10hrs; ACUITY management split)

  • Costing based on trial vial usage; ACUITY management split
  • The ACUITY trial includes 3 of the 5 comparators and had a median treatment period pre-angiography of 4hrs (mean 10hrs)
  • This analysis is most aligned with the available trial data
4.

Adjusted analysis (costing based on 72hr pre-angiography treatment period; MINAP management split)

  • Costing based on a simulation assuming 72hr pre-angio treatment duration and a 1hr PCI treatment duration; MINAP management split
  • This analysis makes some adjustments to costing and management split that may be more typical for the UK
  • Note that this analysis potentially biases against upstream treatments as costs are increased but efficacy is not adjusted and so the analysis should be interpreted carefully.
Results
Fondaparinux baseline analysis

The analysis incorporating a fondaparinux baseline (that is fondaparinux replaces heparin in the aspirin+clopidogrel+heparin, aspirin+clopidogrel+heparin+GPI during PCI, aspirin+clopidogrel+heparin+GPIupstream arms of the model), was considered most relevant to clinical decision making in the majority of cases. Fondaparinux has been found to be cost-effective compared to heparin as shown in the published literature114. Fondaparinux is cheaper than enoxaparin and is associated with clinical benefits. In the model Aspirin+clopidogrel+fondaparinux dominated Aspirin+clopidogrel+heparin in all of our analyses (although this comparison was a secondary objective of the analysis).

In the trial aligned analysis (when trial vial usage was used for costings and the ACUITY management split employed) routine addition of upstream GPIs seems to be most cost-effective for patients in risk groups 2 and 3, with selective PCI GPI use the most cost-effective in risk group 4. This is based on these options having the highest mean INB at a £20,000 per QALY threshold. In the adjusted analysis (with treatment costs estimated using a 72hr pre-angiography treatment duration and the MINAP management split employed) selective use of GPIs at PCI was found to be most cost-effective strategy; however, this analysis was considered likely to bias against upstream use of GPIs as treatment costs are increased but efficacy is not adjusted.

There was considerable uncertainty in the results. This is evidenced by differences between the deterministic optimal strategy and probabilistic optimal strategy especially in Groups 1a and 4. Also, there is a wide spread of the probability of cost-effectiveness across different strategies. In places the optimal strategy as based on mean INB is not the one with the highest probability of being cost-effective as based on the highest proportion of simulations. In addition there is uncertainty regarding applicability as the trial aligned analysis may not represent typical treatment durations in the UK; whereas the longer term analysis is limited by the lack of effectiveness data. It was also noted that from a clincal perspective, the longer the wait for angiography the more likely a patient would need a GPI prior to angiography and deferring use until PCI is undertaken may not be a clinically acceptable option.

Interpretation of results is complicated by the uncertainty in the analysis. Additional clinical considerations should be employed in interpretation and it was considered that it may be reasonable to recommend more than one option to reflect this uncertainty. Risk group 1 is considered least likely to benefit from additional treatment over and above aspirin+clopidogrel+fondaparinux. Dependent on appropriate clinical interpretation, due to the uncertainty it was considered that either GPI use upstream of angiography or selective GPI use in PCI might be considered likely to be cost-effective in higher risk groups. This is due to the fact that different options were found to be most cost-effective in the trial aligned and adjusted analysis but limitations in the analysis mean that a definitive conclusion is not possible based on these model results alone.

Note that the fondaparinux baseline analysis is dependent on the assumption that the relative effect of GPIs will not be impacted by whether heparin or fondaparinux is used as the baseline antithrombin – there were no studies that assessed GPIs against no GPIs in a population using fondaparinux. The OASIS-5 trial addresses this issue somewhat by examining 30-day outcomes for fondaparinux versus enoxaparin in subgroups of patients receiving clopidogrel and GPIs115. This analysis suggested that the benefits of fondaparinux are maintained in patients receiving clopidogrel or GPIs.

Heparin baseline analysis

If fondaparinux is not an appropriate option, then the analysis with a heparin baseline is most appropriate to review. In this analysis, risk group one is least likely to benefit from additional treatment over and above aspirin+clopidogrel+heparin. Heparin use with selective bivalirudin during PCI seems to be most cost-effective in risk groups 2-4. This is based on the mean INB from the heparin baseline analyses in both the trial aligned analysis (reflective of a short time to angiography) and the adjusted analysis (with treatment costs estimated using a 72hr pre-angiography treatment duration and the MINAP management split employed). Bivalirudin use pre-angiography was associated with more QALYs than the selective bivalirudin use but also additional costs and based on the mean INB this use was not cost effective at a £20,000 per QALY threshold.

As in the fondaparinux baseline analysis there was considerable uncertainty in the heparin-baseline analysis. In the trial aligned analysis (reflective of a short time to angiography) bivalirudin PCI was considered the most cost-effective treatment based on mean INB, bivalirudin use upstream of angiography, and upstream GPI use generally also had a high level of simulations where they were optimal. As risk increased the likelihood of bivalirudin initiated upstream of angiography being cost effective increased. It was also raised that there will sometime be a clinical need to give additional treatment upstream of angiography, for example if the patient is actively unstable. Interpretation of results is complicated by the uncertainty in the analysis. Additional clinical rationale should be employed in interpretation and it was considered that it may be reasonable to recommend more than one option to reflect this uncertainty. Dependent on appropriate clinical interpretation, due to the uncertainty it was considered that use of the following might be considered likely to be cost effective: bivalirudin used selectively during PCI; upstream bivalirudin; heparin plus upstream GPIs.

In the adjusted analysis (where costing was based on a 72hr pre-angiography treatment duration) PCI bivalirudin was also most cost effective, as would be expected as the upstream treatments will have higher costs in the model but the effectiveness was not adjusted. In addition, this analysis was considered the least clinically relevant because if patients were not going for angiography relatively quickly they would be most likely to be considered suitable for fondaparinux.

4.3.6. Evidence summary

Bivalirudin infusion initated before angiography: ACUITY trial

Bivalirudin + GPI versus heparin + GPI

For the entire ACUITY trial population 110 156, there were no significant differences between people randomised to bivalirudin + GPI versus heparin + GPI for any of the outcomes of interest. In the subgroup that received clopidogrel 109, results were also non-significantly different. Thus, the GDG focused on the comparison of bivalirudin monotherapy versus heparin + GPI.

Bivalirudin vs heparin + GPI (with background aspirin)

In the entire ACUITY trial population, 110 bivalirudin monotherapy was associated with decreased rates of the net clinical outcome (death, MI, unplanned revascularisation, or bleeding) compared with heparin + GPI. The bivalirudin monotherapy group also had significantly decreased major or minor bleeding.. These results were also similar in the subgroup of people who received clopidogrel before angiography or PCI 109. Angiography was performed in all patients within 72 hours of randomisation. The effect of bivalirudin appeared to be dependent on the use of upstream thienopyridine (such as clopidogrel) use; pre-treatment with clopidogrel was particularly important in the group given bivalirudin alone; without clopidogrel the absolute rate of composite ischaemic endpoints (death, MI, unplanned revascularisation) was 2% higher (9.1% bivalirudin group vs 7.1% heparin + GPI group)110.. In keeping with SPC for bivalirudin, the analysis was focussed on those people who received clopidogrel before angiography or PCI.

PCI subgroup analysis

In the ACUITY PCI subgroup who received clopidogrel 109, ischaemic complications were non-significantly different and major bleeding was significantly decreased in the bivalirudin monotherapy arm.

ACUITY showed a reduced bleeding risk for bivalirudin compared to heparin plus GPI, but only when bivalirudin was used alone (without a GPI).

Bivalirudin initiated after angiography and before PCI: REPLACE-2 ACS subgroup

Bivalirudin vs heparin + GPI (with background aspirin) in patients undergoing PCI

In REPLACE-2, a GPI was mandated in the heparin arm but allowed, if clinically indicated, in the bivalirudin arm. At 30 day follow-up there was no difference in ischaemic endpoints. REPLACE-2 showed bivalirudin to reduce significantly the rate of minor bleeding compared to heparin.

4.3.7. Evidence to recommendations

Trials comparing bivalirudin with heparin + GPI suggest that bivalirudin may offer equivalent ischemic protection with reduced bleeding. However, interpretation of bivalirudin trial data is complicated by differences in dosages, duration of therapy, adjunctive therapies (such as clopidogrel and GPIs), trial design, and study populations. Hence, making recommendations regarding the place of bivalirudin in the management of patients in the UK admitted with UA/NSTEMI is difficult.

The ACUITY trial recruited 13,819 patients with UA or NSTEMI who were described as having “moderate or high risk acute coronary syndromes”. As commented upon elsewhere in this guideline (see RISK chapter) patients recruited to trials described as being moderate/high risk may nevertheless be lower risk than many patients in unselected registry populations, such as in the MINAP database. For instance, those considered ‘high risk’ in ACUITY (TIMI risk score 5-7) had 1 year mortalities of 6.1% in the bivalirudin, and 6.7% in the heparin arms of the trial, which puts them into our intermediate (group 2b) or lower risk category (predicted 6-month mortality 3-6%). The overall ACUITY trial concluded that in patients undergoing early angiography (median time from admission to angiography around 19.5 hours, median duration of treatment from randomisation to angiography only 4 hours), the use of bivalirudin alone (but with bail-out GPI if clinically indicated) was associated with rates of ischaemia that were similar to those of patients receiving either heparin+GPI or bivalirudin+GPI, but that the rate of bleeding complications was significantly reduced. Bivalirudin was also found to reduce bleeding complications in the REPLACE-2 trial.

However, achieving this desirable outcome without compromising the risk of ischaemic events, is influenced by the background therapy used, When used in comparison to a combination of heparin given together with a GPI, bivalirudin given alone and without prior treatment with a thienopyridine (most usually clopidogrel) can increase the risk of ischaemic events (ACUITY trial). For this reason the European Medicines Agency (EMEA) now licenses the use of bivalirudin for patients with acute coronary syndromes but states that “bivalirudin should be administered with aspirin and clopidogrel”.

The ACUITY trial, on which much of the evidence for the benefit of bivalirudin rests, recruited patients with “moderate and high risk acute coronary syndromes”o who were scheduled to undergo angiography. It did not address the use of bivalirudin for those where an early invasive approach was not considered appropriate. In the ACUITY trial the time from hospital admission to angiography (around 19.5 hours), and the duration of treatment prior to angiography (median four hours), was short compare to UK practice. Historically, delays to angiography in the UK have been long (often many days), and although waits for angiography in the UK are declining, they are still longer, than in ACUITY. Extrapolation of the results of ACUITY to an unselected UK population of patients with UA or NSTEMI is therefore difficult.

Subsequent publications showed bivalirudin to be of potential benefit for the subset in ACUITY who underwent PCI (ACUITY-PCI156) and that at 1-year the rates of composite ischaemia and mortality were similar to patients treated with a heparin together with a GPI168. However, ACUITY-PCI has been criticised because randomisation was not stratified by the treatment assigned, different GPIs and heparins were used in the control arm, and the analysis was not powered for non-inferiority testing169.

What is generally accepted is that any benefit of bivalirudin is predicated on its ability to reduce bleeding in patients who undergo angiography. It has not been shown to reduce ischaemic risk and therefore its potential value is in providing ‘net benefit’ – reducing bleeding events and being equivalent with regards ischaemic events). This amalgamation of efficacy (ischaemic events) and safety (bleeding events) has also been criticised on principle 169) because drugs that are ineffective but safe can appear to be better than effective drugs in a non-inferiority trial. Nevertheless, reducing bleeding risk is accepted as important because bleeding is known to be associated with adverse outcomes (see bleeding risk in risk chapter) and this association was confirmed in the ACUITY trial 170, particularly in patients of advanced age46. However, whilst this association between bleeding and adverse outcomes is generally accepted and has been noted in trials of other antithrombotic and anti-platelet agents such as OASIS-5, the reduction in bleeding associated with bivalirudin did not significantly reduce mortality.

In the REPLACE-2 study bivalirudin (with provisional use of a GPI if clinically indicated) was compared with heparin and GPI in patients undergoing urgent or elective PCI. The trial recruited a mixed population of acute and stable coronary syndromes; however a subgroup was available reporting outcomes in ACS patients where >60% were UA/NSTEMI. Patients were randomised in the catheter laboratory and therefore durations of treatment were shorter than in ACUITY. Again bivalirudin reduced bleeding but had no significant impact on ischaemic outcomes.

No trial has investigated the use of bivalirudin versus fondaparinux. However, as detailed elsewhere in this guideline, fondaparinux is more clinically and cost effective than heparin. Fondaparinux is associated with reduced bleeding risk, and therefore it is possible that the difference in bleeding complications between bivalirudin and fondaparinux+GPI may be less than that seen between bivalirudin and heparin+GPI. Also, whilst switching patients from upstream use of heparin to the use of bivalirudin prior to angiography may be safe171, there are no data concerning a switch of patients from fondaparinux to bivalirudin. Since this guideline recommends fondaparinux as the baseline antithrombin most patients with UA/NSTEMI admitted in the UK are likely to be on fondaparinux prior to angiography and not heparin.

The difference in bleeding complications seen in ACUITY was largely due to reduced bleeding from the site of arterial access for the angiogram procedure. Other factors may also reduce bleeding risk, such as greater use of a radial rather than femoral arterial access172, a move towards smaller diameter catheters173, selective rather than more routine use of GPIs, and lower heparin doses. Differences in the frequency of these between practice in the USA and UK also add to the need for caution in extrapolating the results of ACUITY to the UK.

Thus, in the light of trial evidence, if bivalirudin were to be considered it would be in patients:

  • with acute coronary syndromes who are pre-treated with clopidogrel,
  • and who will undergo very early angiography (< 24 hours from admission),
  • and who would otherwise be considered appropriate for a GPI
  • and who have not already been started on fondaparinux

Health Economics

We undertook a health economic analysis modelled both on short term (<24 hours) upstream use of a GPI or bivalirudin prior to angiography, to reflect trial (ACUITY) data, and their use when initiated in the catheter laboratory and given selectively to patients undergoing PCI (REPLACE-2 data). We also modelled longer term (72 hours) use of a GPI or bivalirudin (in recognition of the longer average times to angiography in the UK compared to those reported in the ACUITY trial). The cost-effectiveness analysis undertaken for the guideline and its results and limitations are summarised above and described in detail in Appendix C.

Short term upstream bivalirudin

Cost-effectiveness evidence from our analysis of shorter term use of bivalirudin suggested that it may be cost effective when heparin is the antithrombin used in the alternative strategies (heparin + GPI), in patients at intermediate and higher risk (predicted 6-month mortality >3.0). However, when fondaparinux was incorporated into the analysis instead of heparin, results suggested that bivalirudin may no longer be the most cost-effective option. Given the statistical uncertainty in the cost-effectiveness analysis, and limitations such as that associated with the indirect fondaparinux–bivalirudin comparison, it was nevertheless concluded that bivalirudin should be considered as a possible treatment option in patients at intermediate and higher risk of an adverse cardiovascular event It was considered that either using bivalirudin as an adjunct to PCI (as in REPLACE-2), or as a short term infusion prior to angiography (as in ACUITY) were reasonable. While selective use of bivalirudin during PCI was found to be most cost-effective, there was considerable uncertainty in the analysis.

Longer term upstream bivalirudin

Longer term (72 hours) use of bivalirudin is technically within its license but there is no trial evidence to support its use for this duration upstream of angiography. In addition in the longer term model, the overall cost difference between bivalirudin used upstream of angiography, and other treatment options, increases. This makes bivalirudin less cost effective, although it is acknowledged that the model does not incorporate any additional potential treatment benefit of longer upstream treatment. Given the lack of clinical evidence for longer use and the additional costs it was concluded that bivalirudin use should be restricted to shorter term (<24 hours) scenarios.

Also, clinically it would be unacceptable to defer the addition of potentially beneficial therapy for 72 hours pending angiography, because if it is judged that a patient requires more than heparin the clinician should be free to offer additional medication ahead of the angiography procedure and not withhold it simply on the basis of this cost-effectiveness modelling.

It is noted that population average risks of events are used in the economic model, based on best available data, but it is possible that clinical assessment may refine risk such that the net benefit (ischaemic risk vs bleeding risk) may be improved for an individual patient over that assumed for the average within a trial population. For this to be fully elucidated improved measures of bleeding risk will need to be incorporated into future clinical trials. In addition it is assumed in the economic model that relative risks of benefit (reduction in mortality or ischaemic events) and of harm (such as bleeding events) are constant across the various patient risk groups (low, intermediate, high, highest). Based on available data this was considered a reasonable assumption, but it is possible that this may vary.

Conclusions

Whilst sensitivity analyses were undertaken to model different assumptions, observations such as the uncertainties listed above, caused the GDG to be cautious about mandating the use of bivalirudin or GPIs, concluding it was more appropriate to recommend that they be “considered” by clinicians for certain patients at intermediate or higher risk (predicted 6-month mortality >3.0%). When such patients are scheduled for very early angiography (<24 hours from admission) and are pre-treated with aspirin and clopidogrel, bivalirudin either used selectively at the time of PCI or for a few hours upstream of angiography, is a reasonable alternative to the combination of heparin+GPI. With potential benefit of earlier angiography reported30, more patients in future may be considered appropriate for one of these pharmacological strategies.

The lack of data comparing fondaparinux (with or without a GPI) with bivalirudin, and the need for unfractionated heparin to be given at the time of PCI to patients receiving fondaparinux, led the GDG to conclude that if very early angiography was scheduled, upstream unfractionated heparin (with or without a GPI) should be the alternative to bivalirudin. Whilst switching patients from initial treatment with heparin to starting bivalirudin has been shown to be safe 171 there have been no studies of patients being switched from fondaparinux to bivalirudin.

For those patients waiting >24 hours for angiography, fondaparinux is the preferred upstream antithrombin (with or without a GPI), as has been concluded elsewhere in this guideline.

Bleeding risk

All anticoagulants are necessarily associated with a risk of bleeding complications and weighing this risk against the potential benefits of such agents requires an understanding of the factors associated with bleeding risk, measures by which the magnitude of this risk can be estimated, and the potential for benefit from these agents in reducing the rate of ischaemic events. Close attention to appropriate dosing of these agents is particularly important174. This topic is covered in detail in the RISK section of this guideline, to which readers are encouraged to refer (cross reference the bleeding section of the RISK chapter).

4.3.8. Recommendations

R17.

Offer fondaparinux to patients who do not have a high bleeding risk, unless coronary angiography is planned within 24 hours of admission.

R18.

Offer unfractionated heparin as an alternative to fondaparinux to patients who are likely to undergo coronary angiography within 24 hours of admission.

R19.

Carefully consider the choice and dose of antithrombin in patients who have a high risk of bleeding associated with any of the following:

  • advancing age
  • known bleeding complications
  • renal impairment
  • low body weight.
R20.

Consider unfractionated heparin, with dose adjustment guided by monitoring of clotting function, as an alternative to fondaparinux for patients with significant renal impairment (creatinine above 265 micromoles per litre).

R21.

Offer systemic unfractionated heparin (50–100 units/kg) in the cardiac catheter laboratory to patients receiving fondaparinux who are undergoing PCIp.

R22.

As an alternative to the combination of a heparin plus a GPI, consider bivalirudin for patients who:

  • are at intermediate or higher risk of adverse cardiovascular events (predicted 6-month mortality above 3%), and
  • are not already receiving a GPI or fondaparinux, and
  • are scheduled to undergo angiography (with follow-on PCI if indicated) within 24 hours of admission.
R23.

As an alternative to the combination of a heparin plus a GPI, consider bivalirudin for patients undergoing PCI who:

  • are at intermediate or higher risk of adverse cardiovascular events, and
  • are not already receiving a GPI or fondaparinux.

Footnotes

n

Costs are converted from 2006 US dollars using Purchasing Power Parities138

o

Moderate and high risk was defined as being one or more of the following: new ST-segment depression or transient ST elevation of at least 1mm; elevations in troponin-I, troponin-T, or creatine kinase MB levels; known coronary artery disease; or all four other variables for predicting thrombolysis in Myocardial Infarction (TIMI) risk scores for unstable angina (JAMA 2000;284:835-842)

p

Unfractionated heparin is not licensed for use during angiography and PCI. Such use is an off-label use. Informed consent should be obtained and documented before it is used during angiography and PCI.

Copyright © 2010, National Clinical Guidelines Centre.

Apart from any fair dealing for the purposes of research or private study, criticism or review, as permitted under the Copyright, Designs and Patents Act, 1988, no part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior written permission of the publisher or, in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK. Enquiries concerning reproduction outside the terms stated here should be sent to the publisher at the UK address printed on this page.

The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore for general use.

The rights of the National Clinical Guidelines Centre to be identified as Author of this work have been asserted by them in accordance with the Copyright, Designs and Patents Act, 1988.

Bookshelf ID: NBK62744

Views

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...