Q 42What is the evidence for the use of antiplatelet agents in the prevention of arterial disease in adults with Type 1 diabetes?

Author/Title/Reference/YrNHS Centre for Reviews and Dissemination 1998, Aspirin for the secondary prophylaxis of vascular disease in primary care., University of Newcastle upon Tyne, Centre for Health Services Research; York: University of York, Centre for Health Economics..
N=83 trials (Diabetes 8 trials).
Research DesignSystematic review and Meta analysis
AimTo determine the efficacy of aspirin to reduce cardiovascular morbidity and mortality in the primary care setting
PopulationAll primary care some studies with people with diabetes (type not stated)
InterventionAspirin at various doses
ComparisonPlacebo or other antiplatelet agents
OutcomeThe effectiveness of antiplatelet therapy overall; dose of aspirin and duration of treatment; side effects and cost; the effectiveness of aspirin to the secondary prophylaxis of cardiovascular disease and stroke in specific groups of patients.
Vascular events or death
CharacteristicsVaried between studies
ResultsResults relevant to Diabetes

Aspirin as an antiplatelet agent in patient at raised vascular risk
A pooled risk ratio by clinical subgroup and overall is calculated by combining The meta-analysis of the Antiplatelet Collaborative Group (containing properly randomised trials that compared prolonged antiplatelet therapy (≥1 month) with control, and trials that compared different antiplatelet agents) with trials published after 1990; to establish the impact of antiplatelet therapy on subsequent MI, stroke and vascular death. Overall pooled risk ratio = 0.79 (95% CI: 0.76–0.82)
There is good homogeneity between trials (Q=74.81, degrees of freedom=0.70)
Thus there is strong evidence for a general protective effect of antiplatelet therapy in patients at raised vascular risk
Too few studies make a direct comparison between aspirin and alternative antiplatelet agents to enable relative effectiveness to be defined. However, when the available data is combined a resulting analysis provides no evidence of systematic differences in effect between aspirin and other antiplatelet agents.
Substantially different reductions in mortality were found for different clinical conditions.

Aspirin therapy in diabetes
8 Trials were identified in people with diabetes.
The results contribute to an overall estimate of risk difference of 1.2%, which is of uncertain statistical significance (95% CI: – 0.9 to 3.3).
There is no evidence of heterogeneity of treatment effects between these trials (Q=1.66, df =0.36)
The pooled incidence rate difference, by random effects model, estimate that this translates to a 0.3% reduction in the risk of MI, stroke or vascular death from antiplatelet therapy for 1 year (NNT=360), which is not statistically significant
Two major trials ETDRS and DAMAD studies indicate that on its own, a diagnosis of diabetes does not represent a risk factor of sufficient severity to predict worthwhile benefits from aspirin therapy. The similar relative for MI, stroke and vascular death found in diabetes trials and other trials of patients at raised vascular risk, indicates that patients with diabetes alongside other indications of vascular risk are likely to benefit from routine aspirin therapy.
Hierarchy of Evidence GradingIa
CommentsGuideline development group containing individuals from appropriate sectors (GPs, secondary care physicians, medical and pharmaceutical advisers and pharmacists) discussed the evidence and developed recommendations.
Literature search: MEDLINE and controlled trials register in the Cochrane library to locate systematic reviews and meta- analyses, RCTs, quality of life studies and economic studies. Use was also made of recent, high quality review articles and bibliographies and contact with subject area specialists.
Study quality was assessed by internal validity, external validity and construct validity. Categorisation of quality was performed by a single reviewer
Where appropriate results of randomised studies were combined using established meat-analytic techniques.
Guideline includes a systematic appraisal of evidence, which is graded and used to form recommendations graded according to the level of evidence on which they are based.
Fixed effects models were used in studies estimating a single underlying effect, random effects models where a study looked at a distribution of effects.
Statistical effects of heterogeneity used to assess the likelihood that the variation between the results are not due to change or systemic differences between effects measured, using the heterogeneity statistic Q test.
Risk differences used to describe the practical importance of an intervention in practice
Guideline states that recommendations will cease to apply on 31st December 1999, by which time new, relevant results that may affect its recommendations are likely to be available
Trials includedDiabetes: Pollock & Wright 1979 (Sanofi internal report 105062/0019), Belgian Ticlopidine Retinopathy Study Group (BTRS) 1992, DAMAD Study Group 1989, Pannebakker et al 1980 (Abstract), EDTRS Investigators 1992, Oakley et al 1983 (Sanofi internal report 105062/0019), Nyberg et al 1984, Mirouze (on behalf of the TIMAD Study Group) 1984.
Author/Title/Reference/YrETDRS Investigators 1992, “Aspirin effects on mortality and morbidity in patients with diabetes mellitus. Early Treatment Diabetic Retinopathy Study report 14.”, JAMA, vol. 268, pp. 1292–1300.
Aspirin =1856, placebo= 1855
2 sites
Research DesignRandomised controlled trial
AimTo evaluate the effect of early aspirin treatment in people with diabetes to prevent mortality
PopulationMixed diabetes population
InterventionAspirin (two 325 mg tablets once a day)
OutcomePrimary outcome: mortality from all causes.
Additional outcomes: cause-specific mortality, cardiovascular events (measured by fatal and nonfatal MI and stroke), amputations, and kidney disease or failure (measured by dialysis, transplantation, or death due to kidney failure).
CharacteristicsMale (%): aspirin = 55.5, placebo = 57.4; Age <30: aspirin = 17.5%, placebo = 16.3%; Age 30–49: aspirin = 30.8%, placebo = 32.1%; Age ≥50: aspirin = 51.7%, placebo = 51.6%; Type 1 diabetes: aspirin = 30.1%, placebo = 30.8%; type 2 diabetes: aspirin = 31.6%, placebo = 30.5
ResultsTotal and cause specific mortality
Total mortality
Aspirin: n = 340 (18.3%), placebo: n= 366 (19.7%); z= –1.10
5-year life table rates for total mortality: aspirin=12.1%, placebo=14.9%
Relative risk based on total mortality for the entire follow-up period=0.91 (99% CI: 0.75–1.11, p=0.24)
Cardiovascular deaths
Aspirin: n=244 (13.1%) (71.8% of all deaths), placebo: N=275 (14.8%) (75.1% of all deaths), z=–1.47
25 deaths in each group were attributable to cerebrovascular disease (1.3% of all patients)
5-year cumulative rates for cardiovascular mortality: aspirin=9.3%, placebo=11.2%
Relative risk of cardiovascular mortality estimated for the entire follow-up period=0.87 (99% CI: 0.70–1.10), p=0.12)

Morbidity and combined mortality/morbidity end points
Fatal and non-fatal MI: aspirin: n=241 (13.0%), placebo: n=283 (15.3%), z=–1.99
No significant difference was seen between participants randomised to aspirin or placebo for fatal or nonfatal stroke (5.0% and 4.2% respectively), hypertension (75.1% and 73.1% respectively), renal transplants (7.8 and 7.9%), patient identified as a candidate for dialysis (5.4 and 5.4%) or reported dialysis (1.9 and 1.8%).
Occurrence of ≥ 1 of these conditions or death due to kidney disease occurred in 12.9% of aspirin and 12.2% placebo assigned patients.

Fatal or nonfatal Myocardial infarction
5-year event rates for fatal and nonfatal event rates for MI: aspirin=9.1%, placebo=12.3%
Relative risk of MI estimated for the entire follow up period = 0.83 (99%CI: 0.66–1.04, p=0.04)

Fatal or nonfatal Stroke
5-year event rates for fatal and nonfatal event rates for stroke: aspirin=4.5%, placebo=3.8%
Relative risk of stroke estimated for the entire follow up period = 1.17 (99%CI: 0.79–1.73, p=0.32)

Combined end points
Cumulative 5-year event rates for cardiovascular mortality, non-fatal MI, or stroke: aspirin=14.0%, placebo=16.6%
Treatment difference for a combined end point was similar to that for cardiovascular death alone.
The relative risk adjusted for age, gender, type of diabetes and clinical centre estimated using Cox regression, was very similar to the unadjusted estimates.

Side effects
When patients presented at routine visits with an indication of possible toxicity laboratory tests were performed for investigation. The percentage of patients with ≥1 laboratory tests scheduled for this reason at any time during follow-up was 10.9% for patients receiving aspirin and 9.6% for those in the placebo group. The detection of abnormalities occurred in 53% of these tests and results were equal between the two groups.
Few patients in both groups (2%) had some indication of bleeding (Hb<100g/L, haematocrit <0.30, haematuria, or blood in the stool)

Examined by Electrocardiogram Reading Centre staff using the Minnesota Code without knowledge of treatment assignment, these showed 60/942 (6%) of patients assigned to aspirin and 63/916 (7%) assigned to placebo who were examined at follow- up had one or more major abnormalities identified on central reading.
Only 6 and 8 (10 and 13%) of these patients in the aspirin and placebo groups respectively had previously reported any nonfatal myocardial infarctions before the ECG was obtained. These patients were considered to have had a ‘silent’ myocardial infarction.
The majority of the total ‘silent’ MIs detected throughout the study were not included in the number of fatal and nonfatal MIs at analysis.
Addition of these events to the other data for MI events increases the number to 289 (16%) and 336 (18%) in the aspirin and placebo groups, respectively (p=0.038)
Hierarchy of Evidence GradingIb
CommentsMortality classified as cardiovascular by the Mortality and Morbidity Committee
Study designed to evaluate the effects of photocoagulation and aspirin on ocular events. At the same time information was collected on each death and cardiovascular hospitalisation and on the results of regular medical examinations.
Patients enrolled between April 1980 and July 1985

Aspirin dose
Initial dose based on a previous study showing it to be sufficient to ensure complete inhibition of platelet aggregation by blocking cyclooxygenase but not enough to completely inhibit prostacyclin production by endothelial tissue.
During the course of the study, evidence pointed to lower doses was uncovered, suggesting that 80 mg every other day may be sufficient to inhibit platelet aggregation.
Review of all information including ETDRS data on side effects, the EDTRS investigators decided to continue with the dose of 650 mg/d.

Patient follow up
≥ 5 years, with scheduled visits every 4 months
Examination during initial 3.5 years: physical examination, medical history assessment of drug adherence based on pill count and patient interviews, and biochemical and haematologic determinations at annual intervals.
Additionally serum thromboxane B2 levels and urine salicylate levels were used to assess compliance.
For the remainder of the study, physical examinations were performed at only the third and fifth annual visit and biochemical compliance tests were discontinued.
Author/Title/Reference/YrRoffi, M., Chew, D. P., Mukherjee, D., Bhatt, D. L., White, J. A., Heeschen, C., Hamm, C. W., Moliterno, D. J., Califf, R. M., White, H. D., Kleiman, N. S., spacing, a., & Topol, E. J. 2001, “Platelet glycoprotein IIb/IIIa inhibitors reduce mortality in diabetic patients with non-ST-segment-elevation acute coronary syndromes”, Circulation, vol. 23, pp. 2767–2771.
N=6 Randomised, double blind, placebo controlled trials evaluating the medical management of acute coronary syndromes (ACS) in the absence of ST-segment elevation
Total 6485 people with diabetes and non-ST-segment elevation ACS
Research DesignSystematic review and Meta analysis
AimTo evaluate the efficacy of platelet inhibitors to reduce mortality in a subgroup of people with diabetes and non-ST-Segment elevation
PopulationMixed diabetes population
InterventionSecondary prevention with intravenous platelet GP IIb/IIIa antagonists
4 different agents were included in the trials (tirofiban, lamifiban, eptifibatide or abciximab)
OutcomeDeath and death or nonfatal MI at 30 days
MI defined by creatine kinase (CK) or CK-MB greater than the upper limit or normal.
CharacteristicsVaried between trials
Platelet GP IIb/IIIA inhibition was associated with a significant reduction of the aggregate 30-day mortality in people with diabetes.
Breslow-Day statistic demonstrated a lack of heterogeneity between the trials for mortality
Pooled results from the trials showed a reduction from 6.2% to 4.6% (OR:0.74, 95% CI:0.59–0.92, p=0.007)
Logistic regression analysis demonstrated a significant interaction between diabetic status and treatment (OR:0.75, p=0.036)
Patients undergoing PCI
The most marked benefit from active therapy was seen in patients with diabetes undergoing PCI:
Mortality in patients randomised to active treatment = 1.2% vs. 4.0% in placebo groups (OR: 0.30, 95%CI:0.14–0.69, p=0.002)

Composite death or MI
Treatment was associated with a reduction in the composite death or MI at 30 days in people with diabetes
A statistically significant reduction was seen in the PRISM (OR: 0.50, p=0.038) and PRISM-PLUS (OR: 0.27, p=0.001).
Pooling of the ORs also showed a significant reduction but Breslow-Day statistic demonstrated heterogeneity (p=0.024) between the trials for composite death or MI, so pooled results are not reported
Patients undergoing PCI
Active therapy was associated with significant reduction of death or MI at 30 days (15.8% –9.9%, OR:0.58, 95%CI:0.41–0.82, p=0.002) in the absence of heterogeneity.

GP IIa/IIIb receptor inhibition was associated with similar proportionate reduction in mortality for patients treated with insulin and those controlled by diet or oral therapy. Mortality was reduced from 6.9% to 5.2% (OR: 0.74, p=0.14) in 1799 people on insulin and 5.8% to 4.3% (OR:0.74, p=0.025) in 4551 people controlling their diabetes with diet or oral hypoglycaemic drugs
Hierarchy of Evidence GradingIa
CommentsIncluded trials differed in design, inclusion criteria, therapeutic agents, regimens and access to percutaneous revascularisation.
Enzyme definition of MI was not uniform across the trials, which may have influenced the incidence of events and thus the extent of therapeutic benefit.
Included trials were inconsistent with their prescription of heparin.
No details given of drugs other than intervention and aspirin administered throughout the trial
Primary endpoint is consistent throughout the trials, and assessed at 30 days, irrespective of length of follow-up.
Mantel-Haenszel statistic used to test significance of treatment effect within each study
Heterogeneity of ORs across the trails examined with Breslow-Day statistic—if p value was nonsignificant, trial results were weighted and pooled
Pearson chi-squared test applied to poled event rates to assess overall significance of treatment effects
P<0.05 was considered statistically significant
Logistic regression modelling assessed interaction of diabetic status with treatment.
No details given of type of diabetes in included patients.
Details are not provided of adverse effects of drugs used in the trials.
Revascularisation strategy not randomised thus survival advantage in patients undergoing PCI may be influenced by selection bias.
Limitations of meta-analysis discussed by authors, concluding that use of the Breslow-Day test allows for combination of the trail results with respect to mortality reduction, and suggest that due to the limitations of this being a subgroup analysis and the lack of randomisation of PCI results should be seen as exploratory and for PCI, hypothesis generating, requiring further validation.
Trials includedPlatelet Receptor Inhibition in Ischemic syndrome management trial (PRISM); Platelet Receptor Inhibition in Ischemic Syndrome Management in Patients Limited by Unstable Signs and Symptoms (PRISM-PLUS); Platelet IIb/IIIa Antagonism got the Reduction of Acute Coronary Syndrome Events in a Global Organization Network trial A (PARAGON A); PARAGON B trial; Platelet Glycoprotein IIb/IIIa in Unstable Angina: Receptor Suppression Using Integrilin Therapy trial (PURSUIT); Global Utilisation of Strategies to Open Occluded Coronary Arteries IV trial (GUSTO)

From: Appendix D, Evidence tables

Cover of Type 1 Diabetes in Adults
Type 1 Diabetes in Adults: National Clinical Guideline for Diagnosis and Management in Primary and Secondary Care.
NICE Clinical Guidelines, No. 15.1.
National Collaborating Centre for Chronic Conditions (UK).
Copyright © 2004, Royal College of Physicians of London.

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