Once the anti-aggregant effect of aspirin on platelets had been described in the mid 1960s, calls for trials of aspirin in myocardial infarction were made by John O’Brien in the UK (19) and by Harvey Weiss and others in the USA.21
In fact, these calls had been anticipated. Laurence Craven, a family practitioner in Glendale, California, published a series of three papers on aspirin in the early 1950’s.22-24 Each of these focused on some aspect of bleeding.
In the first, Craven pointed out that women take aspirin rather often for their pains and aches, while men tend to scorn such an ‘effeminate’ remedy and he went on to wonder if this difference might explain the sex difference in the incidence of heart attacks. (22) In his second paper,
Craven described bleeding in tonsillectomy patients given Aspergum, a mint flavoured gum impregnated with aspirin. (23) His third paper reports uncritically a remarkable benefit of aspirin on heart attacks and strokes (see panel). Craven’s ideas were however so unacceptable at that time, and his experimental work so flawed that he had difficulty reporting his findings and had to submit his papers to a rather obscure journal. (24)
The calls of O’Brien and Weiss were however heard in the MRC Epidemiology Unit in Cardiff, and in 1969 a double-blind placebo-controlled randomised trial of low-dose aspirin was commenced. (25) Men who had been recently discharged from local hospitals following an MI were invited to cooperate. Those who agreed were randomised to receive either 300 mg aspirin in a gelatine capsule or a placebo capsule.
Two points about the published report on this first trial are worth noting (see panel below). First, the journal published it under the headline: ‘For debate’. This was totally appropriate. Clinical practice should never be based on the results of a single trial, however convincing these are.
Replication is essential and if a drug or a preventive measure truly works then it will show in different patient groups, in different situations and in trials run by different trialists. Secondly, the results of the trial were evaluated only in terms of a reduction in total deaths. The inclusion of non-fatal events gives opportunity for bias due to a possible differential reporting of symptoms. Cardiovascular disease makes a substantial contribution to all-cause deaths, and if a measure does reduce it, an effect on total mortality should be seen.
A Randomised Controlled Trial of Acetyl Salicylic Acid in the Secondary Prevention of Mortality from Myocardial Infarction.
P.C.Elwood, A.L.Cochrane, M.L.Burr, P.M.Sweetnam, G.Williams, E Welsby, S.J.Hughes, R.Renton
British Medical Journal 1974,1, 436-440
The results of a randomised controlled trial of a single daily dose of acteyl salicylic acid (aspirin) in the prevention of re-infarction in 1,239 men who had had a recent myocardial infarct were statistically inconclusive. Nevertheless, they showed a reduction in total mortality of 12% at six months and 12% at twelve months after admission to the trial. Further trials are urgently needed to establish whether or not the effect is real.
Naturally, clinicians at that time were somewhat more than sceptical of the whole situation, and there was no detectable change in clinical practice. However research groups around the world took note, a number of trials were set up and by 1980 six trials had been reported. (25-30)
Cardiff 1 (1974)25 1239 26% n.s.
CDP (1976) 26 1529 30% n.s.
Cardiff 2(1979)27 1725 30% n.s.
German (1979)28 626 18% n.s.
AMIS (1980)29 4524 10% n.s.
PARIS (1980)30 1216 18% n.s.
23% reduction by aspirin (P < 0.0001)
Each of these early trials suggested a beneficial effect of aspirin, but the results of none achieved an acceptable level of statistical significance. Overall, however, there is clearly convincing evidence of benefit.
The reporting of overviews of the results from all relevant published trials relating to a particular clinical intervention is becoming increasingly common throughout clinical practice and the first such overview, or meta-analysis, based on these six trials was presented by Richard Peto and his colleagues of Oxford to the inaugural meeting of the Society for Clinical Trials in Philadelphia in 1980.(31)
This overview was one strand in the thinking that led eventually to the setting up of the Cochrane Collaboration, the worldwide initiative that aims to conduct overviews within every area of clinical activity. (32)
Since that first report, the Oxford group have produced several monumental overviews of aspirin and cardiovascular disease: in 1988 (33) ,1994 (34). and 1997 (35).
The following is based on the overview published in 1994, which combines results from 145 RCTs, with a total of 102,459 patients and 10,943 outcome events. A remarkably consistent reduction in vascular events is demonstrated (22 to 32%).
An overview of RCT's of aspirin and cardiovascular disease
Prior MI 11 25%
Acute MI 9 29%
Prior stroke/TIA 18 22%
Other high risk 104 32%
All trials 25%
Non-fatal MI 122 34%
Non-fatal stroke 124 25%
Vascular death 144 17%
All-cause death 144 16%
Also considering the following factors
male/female; under/over 65yrs; hypertensive / normotensive; diabetic / non-diabetic
gave no evidence of any significant differences in benefit.
An overview of studies with such a large number of clinical outcomes enables the drawing of conclusions from sub-group analyses with a fair degree of confidence. The benefit of aspirin is similar in all groups of patients whatever the prior indication for prophylaxis. Furthermore, there is no evidence of differences in the proportionate reduction by aspirin in different patient groups: males and females, older and younger subjects, diabetic and non-diabetic, hypertensive and normotensive subjects etc.
Together with this overview a Press Release by the group in Oxford was widely reported by the media. (36) This gave the estimate that around 100,000 premature deaths from cardiovascular disease could be prevented world-wide by the appropriate use of low-dose aspirin together with at least this number of non-fatal heart attacks and strokes.