Intended for healthcare professionals


Which diabetic patients should be taking aspirin?

BMJ 1995; 311 doi: (Published 09 September 1995) Cite this as: BMJ 1995;311:641
  1. John S Yudkin
  1. Professor of medicine University College London Medical School, Whittington Hospital, London N19 3UA

    Those with vascular disease and those at greatly increased risk of vascular disease

    Low dose aspirin has been shown to reduce the risk of myocardial infarction and thrombotic stroke in patients at high risk of vascular death. Does aspirin work in diabetic patients? Is the diagnosis of diabetes sufficient to warrant starting a patient on aspirin? And does aspirin confer any specific advantages or risks on these patients?

    Firstly, does it help? Several large studies have been conducted in diabetic patients, some aimed at preventing cardiovascular disease and others at slowing progression of retinopathy or nephropathy. The Veterans Administration study looked at the efficacy of aspirin and dipyridamole in preventing progression of cardiovascular and peripheral vascular disease in 231 diabetic men with limb gangrene or recent amputation for ischaemia.1 There were no differences in the incidence of either end point. The early treatment diabetic retinopathy study randomised 3711 patients with insulin dependent and non-insulin dependent diabetes to aspirin 650 mg daily or placebo.2 Patients treated with aspirin had 13% fewer cardiovascular deaths—a non-significant difference.

    Inevitably, small numbers of patients and events in studies of this type result in outcomes with fairly wide confidence intervals. The Antiplatelet Trialists' Collaboration avoided the problem of small numbers by conducting a meta-analysis of the efficacy of antiplatelet agents in preventing myocardial infarction, stroke, and vascular death.3 Among nearly 47000 patients considered to be at high risk of such events, just under 10% had diabetes.

    Antiplatelet drugs reduced vascular events

    These diabetic patients had 17% fewer vascular events while taking antiplatelet drugs, compared with a 22% reduction in those without diabetes. Consequently, the benefit was similar in diabetic and non-diabetic patients—38 events prevented per 1000 patients treated, compared with 35. In the study as a whole, no significant differences in efficacy existed between aspirin and other antiplatelet drugs—although these comparisons were not possible in the smaller number of diabetic patients.

    The trialists subdivided the results into those from primary and secondary prevention studies, with very different conclusions. Overall, antiplatelet drugs produced about a 30% reduction in the risk of myocardial infarction and about a 20% reduction in the risk of cerebral infarction. These drugs, however, seemed to double the risk of cerebral haemorrhage. Consequently, the benefits of antiplatelet drugs in patients at high risk of thrombotic events—those with a history of transient ischaemic attack, stroke, angina, or myocardial infarction—outweighed the extra risks of cerebral haemorrhage.

    In primary prevention studies, however, in which patients are at substantially lower risks of thrombotic events, the additional risk of haemorrhagic stroke outweighed the 20-30% protection provided. Put another way, in primary prevention trials, in which subjects had a background risk of cardiovascular death of about 4 per 1000 patient years, the hazards exceeded the benefits; in the secondary prevention trials, in which untreated patients had a cardiovascular mortality of 45 per 1000 patient years, antiplatelet drugs were worth while.

    How can this be extrapolated to patients with diabetes? Diabetes is an independent risk factor for cardiovascular disease, increasing the risk twofold to threefold in men and postmenopausal women and around fivefold in premenopausal women.4 5 Moreover, among patients with non-insulin dependent diabetes microalbuminuria or proteinuria further doubles the cardiovascular risk.6

    From data from men screened for the multiple risk factor intervention trial7 it can be calculated that the 10 year cardiovascular mortality of a hypertensive middle aged man without diabetes (3.6 per 1000 man years of follow-up) would increase to 10.4 per 1000 if he had diabetes but normal urinary albumin excretion and to 18.7 per 1000 if he also had microalbuminuria. For this reason, having diabetes and microalbuminuria with a moderately raised blood pressure or serum cholesterol concentration would place anyone over 50 into a high risk category for cardiovascular death, even without a history of angina or myocardial infarction. If it is the level of risk and not the previous diagnosis of vascular disease that justifies treatment with aspirin then older microalbuminuric diabetic patients or younger patients with any additional risk factor might warrant antiplatelet treatment even without a history of vascular disease.

    The early treatment diabetic retinopathy study provides our only chance to extrapolate this hypothesis to a diabetic population. In that study, although none of the differences was significant, there was an 18% fall in the incidence of myocardial infarction, but the number of all vascular events fell by only 10% because of a 17% rise in strokes.2 This implies that in treating patients with diabetes, as in non-diabetic patients, the potential hazards of aspirin need to be set against its benefits. Yet the study does not really answer the question of whether aspirin works for primary prevention in diabetes, partly because around half of the patients had known cardiovascular disease but also because patients' microalbuminuria status was not documented.

    We must also consider whether there are any benefits or risks of antiplatelet agents specific to diabetes. Several studies have suggested that drugs acting on platelet function can reduce albumin excretion in patients with proteinuria8 9 but these effects may be the consequence of altered renal haemodynamics resulting from changes in glomerular production of eicosanoids and might not translate into long term benefit.

    Indeed, the early treatment diabetic retinopathy study found no difference in the incidence of end stage renal failure or death from kidney disease between aspirin and placebo groups.2 As to retinopathy, again there have been several reports from smaller studies of a slower rate of progression of microaneurysms with treatment with aspirin,8 but the early treatment diabetic retinopathy study found that aspirin at a dose of 650 mg daily neither slowed the progression of retinopathy nor reduced the risk of bleeding from new vessels.10

    Diabetic patients may need higher doses of aspirin

    There are some hints that, because of increased turnover of platelets, diabetic patients need larger doses of aspirin11 and that 300 mg enteric coated aspirin daily may be a better regimen than the more usual 75-100 mg daily used in non-diabetic patients. Indeed, the second international study of infarct survival did not show any benefit of aspirin at a dose of 165 mg daily after myocardial infarction in the diabetic subgroup.12

    Strong evidence therefore exists for treating with aspirin any diabetic patient with a history of thrombotic stroke, transient ischaemic attack, myocardial infarction, or angina, provided that there are no contraindications. It would also be valid to consider starting such treatment for any diabetic patient without known macrovascular disease, but who is at a similarly high level of risk (cardiovascular mortality exceeding 40-50 per 1000 person years) because of other reasons, such as microalbuminuria, age, or standard cardiovascular risk factors.

    A primary prevention trial of aspirin in diabetic microalbuminuric patients is urgently needed.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.
    View Abstract