Intended for healthcare professionals


Collaborative overview of randomised trials of antiplatelet therapy - II: Maintenance of vascular graft or arterial patency by antiplatelet therapy

BMJ 1994; 308 doi: (Published 15 January 1994) Cite this as: BMJ 1994;308:159
  1. Antiplatelet Trialists' Collaboration
  1. Correspondence to : APT Statistical Secretariat, ICRF/BHF/MRC Clinical Trial Service Unit, Nuffield Department of Clinical Medicine, Radcliffe Infirmary, Oxford OX2 6HE, or APT Clinical Secretariat, Department of Clinical Neurosciences, Western General Hospital,Edinburgh EH4 2XU.
  • Accepted 29 November 1993


Objective : To determine the efficacy of antiplatelet therapy in maintaining vascular patency in various categories of patients.

Design : Overviews of 46 randomised trials of antiplatelet therapy versus control and 14 randomised trials comparing one antiplatelet regimen with another.

Setting : Randomised trials that could have been available by March 1990 and in which vascular graft or arterial patency was to be studied systematically.

Subjects : About 8000 patients at varying degrees of risk of vascular occlusion (by virtue of disease or of having some vascular procedure) were in trials of antiplatelet therapy versus control and 4000 such patients were in trials directly comparing different antiplatelet regimens.

Results : Overall, antiplatelet therapy produced a highly significant (2P<0.00001) reduction in vascular occlusion, with similar proportional reductions in several different types of patient. Hence the absolute reductions tended to be largest among patients at highest risk of occlusion, with smaller but still significant absolute reductions among lower risk patients. The proportions of patients with confirmed occlusion among those allocated antiplatelet therapy versus appropriately adjusted control proportions (and mean scheduled treatment durations and net absolute benefits) were: (a) among about 4000 patients with coronary artery grafts, 21% antiplatelet therapy v 30% control (seven month benefit about 90 patients protected per 1000 allocated antiplatelet therapy (2P<0.00001); (b) among about 800 patients after coronary angioplasty, 4% antiplatelet therapy upsilion 8% control (six month benefit about 40/1000 (2P = 0.02)); (c) among about 3000 patients with peripheral artery procedures or disease, 16% antiplatelet therapy v 25% control (19 month benefit about 90/1000 (2P<0.00001); (d) among about 400 renal patients with a shunt or fistula placed for haemodialysis access, 17% antiplatelet therapy v 39% control (two month benefit about 200/1000 (2P<0.00001)).

Conclusion : Antiplatelet therapy (chiefly aspirin alone or aspirin plus dipyridamole) greatly reduces the risk of vascular occlusion in a wide range of patients at high risk of this complication. Further studies are required to determine exactly when treatment should start (to limit any perioperative bleeding while still preventing most early occlusion) and for how long it should be continued.

Clinical implications

  • Clinical implications

  • Patients who have undergone coronary artery bypass grafting or angioplasty or leg artery bypass grafting or angioplasty are at high risk of both major clinical “vascular events” (myocardial infarction, stroke, or vascular death) and subclinical vascular occlusions

  • As well as preventing about one quarter of clinical vascular events in patients undergoing such vascular procedures, antiplatelet therapy also reduced the odds of vascular graft or arterial occlusion by about 40% while treatment continued

  • Typically, among 1000 patients who have coronary artery revascularisation procedures, about six months of antiplatelet therapy should prevent several dozen patients from having coronary artery occlusion as well as a few dozen major clinical vascular events

  • Patients who have peripheral artery revascularisation procedures tend to be at lower risk of vascular events, but even so about one or two years of antiplatelet therapy in 1000 such patients should not only prevent several dozen from having peripheral artery occlusion but also avoid several major vascular events

  • Antiplatelet therapy also substantially reduced the rate of occlusion of arteriovenous fistulas or shunts established for haemodialysis access (and should also reduce the high incidence of major vascular events in such patients)


After coronary artery revascularisation, whether by coronary artery bypass grafting or by percutaneous transluminal coronary angioplasty,1 angiographic studies show substantial rates of reocclusion. For example, about one fifth of coronary artery bypass grafts occlude during the first postoperative year2 and a few per cent per year occlude thereafter.3,4 These occlusions are often subclinical, though some may produce clinical signs of myocardial infarction.5 Occlusion or reocclusion is also seen after peripheral artery revascularisation, though many such occlusions also are subclinical. Finally in patients with renal disease who have arteriovenous fistulas or shunts established to allow haemodialysis, occlusion of these is common (and such patients are also at particularly high risk of other serious thrombotic complications6). Experimental7,8 and clinical*RF 5,9-11* evidence suggests that antiplatelet therapy may help prevent vascular graft or arterial occlusions, particularly during the period soon after vascular procedures, before any intimal damage has healed.

Randomised evidence of the effects of antiplatelet therapy on myocardial infarction, stroke, and death was reviewed in part I.5 To assess the effects of this treatment on vascular occlusion this paper provides a systematic overview12,13 of all randomised trials of antiplatelet therapy in which vascular graft or arterial patency was to be studied systematically. (This includes arteriovenous shunts but excludes those trials in which the chief aim was to prevent deep venous thrombosis or pulmonary embolism, which are reviewed separately in part III.9) Wherever possible information on any serious bleeding complications was also reviewed.

Materials and methods data acquisition Identification of all unconfounded randomised trials

The aim was to include all unconfounded randomised trials of antiplatelet therapy versus no antiplatelet therapy, or of one antiplatelet regimen versus another, that could have been available for review by March 1990 and in which vascular graft or arterial patency was monitored systematically (see appendices 1 and 2). A fuller description of the methods used for seeking trials and data was given in part I.5 Relevant randomised trials were identified in three main categories: (a) coronary artery trials - patients having coronary artery bypass grafting*RF 14-43* or percutaneous transluminal coronary angioplasty*RF 44-46*; (b) peripheral artery trials - patients with symptomatic peripheral vascular disease in whom patency of native peripheral arteries*RF 47-50* was studied, patients having any non-coronary arterial grafting procedures (usually in the legs),*RF 51-65* or patients having transluminal angioplasty in the legs*RF 66-69*; (c) haemodialysis access trials - patients with arteriovenous fistulas or shunts placed for haemodialysis.*RF 70-82*

As in part I, trials were to be excluded if allocation was believed not to have been randomised in a manner that precluded prior knowledge of the next treatment (for example, where allocation was alternate or based on odd or even dates, or where the comparison was to be with some historical controls) or if the treatment comparisons were considered to be confounded (that is, where the scheduled treatment in one group differed from that in the other in some aspect other than antiplatelet therapy). Nor were randomised trials to be included that did not seek vascular graft or arterial occlusion systematically (for example, those designed to assess only partial vessel occlusion, only clinical events, or only short term haematological changes (see appendices 1 and 2 in part I). Trials that were not available for review by March 1990, generally because they were still in progress (see references in part I), do not contribute to the main analyses, though the larger such trials that were reported subsequently are included in appendices 1 and 2 and the discussion.

Definition of outcome measures

Methods used to assess vascular patency varied widely between trials: angiography in all coronary artery trials; clinical examination, Doppler ultrasonography, limb plethysmography (often with confirmatory radionuclide scanning or angiography), or systematic angiography in peripheral artery trials; and clinical examination in all haemodialysis access trials (see appendices 1 and 2). But within any single trial the same approach was to be used for assessing both treatment and control patients, so variation between different trials in the methods used does not invalidate an appropriately constructed overview of their separate results. In some surgical settings, particularly in the coronary circulation, one patient might receive several grafts, but data were sought only on the number of patients in each treatment group who suffered at least one main proximal graft occlusion within the scheduled follow up period. Similarly, for renal patients, in whom multiple occlusions might occur sequentially in the same arteriovenous fistula or shunt, the information sought was whether any occlusions had occurred.

Many trials also attempted to assess “partial occlusion,” but there are substantial difficulties with this. (For example, among patients who have had percutaneous transluminal coronary angioplasty the proportion suffering it varies from 2% to 20%,83 depending on the definition used.) This report therefore excludes partial occlusions. The denominator was to be the number of patients originally randomised rather than the number who received follow up angiography or some other form of assessment of vascular patency. Inevitably, when the method of assessing arterial or graft occlusion was invasive some patients were not subjected to it (see appendices 1 and 2). In principle this might bias outcome assessment, particularly if treatment influenced the rate of symptomatic occlusion, but in practice within any one trial the proportions assessed were generally similar in the treatment and control groups. Analyses confined to placebo controlled studies, which may be less subject to treatment dependent biases in the assessment of vascular occlusion, were also considered separately when appropriate.

Information on non-fatal clinical events (such as myocardial infarction and stroke) and on death from vascular and other causes was also sought, though these events were often recorded only as secondary outcome measures. In the trials where treatment was to last at least one month such results already contribute to part I. Information was sought on any haemorrhage that was either fatal or severe enough to require transfusion (here defined as a major bleed), and was generally available. Other measures of perioperative haemorrhagic complications (such as bleeds resulting in reoperation, wound haematomas, or infection) were also sought, but the definitions used were not uniform and such measures were not available from most trials.

When the data collected did not include information about outcome among all patients initially randomly assigned on all of the outcomes of interest extra details were sought from the principal investigators (see part I). In trials in which some patients had been excluded after randomisation from the published report it was often possible to obtain by correspondence follow up information on these outcome measures among most or all of the missing patients, so that appropriately unbiased intention to treat analyses of such events could be conducted. When this was not possible the available data have still been included in the overview unless the number of exclusions was so extensive that the trial could no longer be considered properly randomised.

Statistical methods: proportional and absolute reductions

Statistical methods used to obtain an overview of the results from the trials were detailed in part I. Proportional reductions in vascular occlusion may be more widely generalisable to somewhat different medical settings, whereas absolute reductions may be more directly relevant to deciding whether to use therapy in particular medical settings. Standard methods for combining trial results5,10,84,85 that provide an unbiased test of whether treatment has any effect at all can also provide an estimate of the “typical” proportional reduction in the odds of adverse events observed in the trials. (These methods are completely robust and assume not that the real proportional effects of treatment are all exactly the same in the different trials but merely that any real effects probably point in the same direction. Fuller discussion has been given elsewhere.85) When the overall difference between treatment and control patients is very highly significant and all the trials are approximately evenly randomised, a simple description of the absolute reduction can be provided just by adding together all the treatment groups, adding together all the control groups, and comparing these two grand totals. (If any of the trials involve deliberately uneven treatment allocation - for example, two thirds treatment versus one third control - then it can first be “adjusted” to an evenly randomised comparison by counting the control group more than once (see part I).)

Even when there is a highly significant overall effect, separate analyses of the effects in small subgroups of patients may be statistically unstable. Consequently, unless there are good prior reasons for expecting large difference between the effects of treatment in different circumstances, the approximate benefit of antiplatelet therapy in some particular subgroup may best be assessed indirectly, not from an analysis that is restricted just to that one subgroup but, instead, by approximate extrapolation from the proportional effect that is observed in a much wider class of patients86 (see discussion in part I).


Effects of antiplatelet therapy on vascular graft or arterial occlusion

Information on vascular graft or arterial occlusion was available from 46 trials of antiplatelet therapy. The proportional odds reductions are described in figure 1 and (of more direct relevance to medical practice) the absolute reductions in risk are given in figure 2. The numbers of patients suffering a vascular occlusion among those allocated antiplatelet therapy totalled 864/4889 (17.7%), while the corresponding adjusted total among controls was 1327/4911 (27.0%). These crude risks of 18% and 27% indicate a risk reduction of about one third, which corresponds to an odds reduction of somewhat more than one third (see legend to figure 1). More formal statistical methods for combining the results from these 46 trials, which entail no unjustified assumptions,5 indicate that the typical reduction in the odds of a vascular occlusion occurring was about 44% (SD 4%), which is very highly significantly favourable (2P<0.00001; fig 1).

FIG 1 - Proportional effects of antiplatelet therapy on numbers of patients found to have occlusion after coronary artery procedures, peripheral artery procedures or disease, and after establishing haemodialysis shunts or fistulas. O-E=Observed minus expected.

In most trials patients were allocated roughly evenly to antiplatelet therapy or control, but in some more were deliberately allocated to active treatment. To allow direct comparisons between proportions observed to have an event in each treatment group, adjusted totals have been calculated after converting any unevenly randomised trials to even ones by counting control groups more than once. Statistical calculations are, however, based on actual numbers from individual trials and crude, unadjusted control totals are given. Stratified ratio of odds of occlusion in treatment group to that in control group is plotted for each group of trials (black square: area proportional to amount of information contributed) along with its 99% confidence interval (horizontal line). Black square to left of solid vertical line suggests benefit (significant at 2P<0.01 only where entire confidence interval is to left of solid vertical line). Overviews of results for certain subtotals of trials (and 95% confidence intervals) are represented by diamonds. Observed reductions in odds of occlusion in particular groups of trials are given to right of solid vertical line)


Risk reduction and odds reduction: Crude totals occluded (18% and 27%) suggest risk reduction of about one third but odds reduction of more than one third. These crude risks are equivalent to odds of 18/82 (0.22) and 27/73 (0.37) respectively (indicating odds ratio of about 0.22/0.37 (=0.59) - that is, approximate odds reduction of about 41%, which is similar to result (44% (SD 4%) of more formal statistical analysis)

When the trials were subdivided into the three main patient categories (coronary artery, peripheral artery, and haemodialysis access trials) the odds reductions were highly significant in each separate category (each 2P<0.00001; fig 1). The 46 separate trials did not all indicate exactly the same odds reduction, but the odds reductions in the two largest patient categories were similar to each other and only slightly smaller than that observed among dialysis patients. The consistency of the proportional effects observed in these very different clinical settings is important because it suggests a general finding of wide applicability. It is the absolute effects, however, that determine how worth while antiplatelet therapy is, and for these reasons both proportional (fig 1) and absolute (fig 2) reductions are described in the more detailed results that follow.


Absolute effects of antiplatelet therapy on occlusion. Adjusted totals calculated after converting any unevenly randomised trials to even ones by counting control groups more than once, for calculating adjusted percentages and numbers of patients prevented from having occlusion per 1000 allocated antiplatelet therapy. (Significance (2P) based on stratified analyses of original, unadjusted numbers in each trial: see statistical methods in part I). (dagger)Months of A=Means of scheduled antiplatelet durations

Effects in different categories of patients Coronary artery trials

Thirty trials of antiplatelet therapy versus control were identified among patients who had had coronary artery bypass grafting or percutaneous transluminal coronary angioplasty (see part I), but vascular occlusion was monitored systematically in only 23 of these trials. In these, allocation to antiplatelet therapy was associated with a proportional reduction of 41% (SD 6%) in vascular occlusion, which was highly significant (2P<0.00001; fig 1). Twenty of the 23 trials that could provide information on vascular occlusion concerned patients with coronary artery bypass graft surgery rather than angioplasty. The proportional reductions in the odds of suffering at least one occlusion appeared similar after coronary artery bypass grafting and after percutaneous transluminal coronary angioplasty. But although the reductions were significant in each setting (2P<0.00001 and 2P=0.02 respectively), the former is much more accurately known. Among patients who had had coronary artery bypass grafting, allocation to a mean scheduled duration of seven months of antiplatelet therapy was associated with prevention of occlusion in 92 (SD 15) patients per 1000 (21.1% of antiplatelet allocated patients versus 30.3% of corresponding controls; fig 2). In the coronary angioplasty studies occlusion was less common (4.1% antiplatelet therapy v 8.0% control) but allocation to an average of six months of antiplatelet therapy nevertheless achieved an absolute reduction of 39 (SD 16) per 1000 patients.

Peripheral artery trials

Thirty nine trials of antiplatelet therapy versus control were identified among patients having peripheral vascular procedures or with peripheral vascular disease (see part I) but vascular occlusion was monitored systematically in only 14 of them. Allocation to antiplatelet therapy in these 14 trials was associated with a proportional reduction of 43% (SD 8%) in vascular occlusion, which was highly significant (2P<0.00001; fig 1). Studies of patients with saphenous vein grafts or prosthetic implants for lower limb disease contributed most of the data; of the three other studies, one assessed the patency of native vessels in patients with intermittent claudication and two concerned patients who had had peripheral angioplasty. As before, however, the proportional reductions in occlusion appeared similar in each of these three different settings, although only the first of the three is accurately known (2P = 0.00001, 2P<0.0005, and 2P = 0.06 respectively; fig 1). Overall, allocation to a mean scheduled duration of 19 months of antiplatelet therapy produced a substantial absolute reduction of 92 (SD 15) per 1000 in the risk of peripheral artery occlusion (15.7% of antiplatelet allocated patients versus 24.9% of corresponding controls; fig 2).

Haemodialysis access trials

Ten trials of antiplatelet therapy versus control were identified among renal patients having surgery to establish haemodialysis access, and vascular occlusion was monitored systematically in nine of these trials (all evenly balanced and placebo controlled) among a total of only 418 patients. Allocation to antiplatelet therapy in these nine trials was associated with a proportional reduction of 70% (SD 14%) in vascular occlusion, which was highly significant (2P<0.00001; fig 1). Similar sized proportional reductions in occlusion of fistulas and of shunts were observed, both being highly significant (2P = 0.005 and 2P<0.0002 respectively). Thrombotic occlusion of a fistula or shunt may occur rapidly, so that, although the mean duration of these trials was only two months, 38.7% of the controls suffered an occlusion as against only 16.5% of the antiplatelet allocated patients - an average absolute reduction of 222 (SD 38) per 1000 (fig 2). The risk of occlusion was greatest among patients in whom arteriovenous shunts were established, so that the absolute benefits observed with antiplatelet therapy among such patients appeared to be even greater than among those with fistulas.

Avoiding bias when assessing efficacy

Systematic assessment of patency was planned in all patients randomised in these trials, but the methods to be used differed widely (appendices 1 and 2). Variation of the method used to detect occlusion should not bias assessment of whether or not antiplatelet therapy prevents occlusion because within any particular trial the same technique was to be used for both the treatment and control groups. But as the methods used were sometimes rather subjective, ascertainment of occlusion might, in principle, have been influenced by knowledge of the allocated treatment group. In practice, however, the results (44% (SD 4%) proportional reduction overall, with 41% (SD 6%), 43% (8%), and 70% (14%) reductions in coronary artery, peripheral artery, and haemodialysis access studies respectively) were not materially altered when the analyses were confined to placebo controlled trials (43% (SD 4%) proportional reduction overall, and 40% (6%), 41% (8%), and 70% (14%) reductions in these three main categories).

As noted in part I, data dependent emphasis on analyses among specific categories of patients, or of trials, may lead to biased conclusions. Hence, though there was marginally significant heterogeneity (X2 on 2 df=7.6; P=0.02) between the effects of antiplatelet therapy on vascular occlusion in the three different categories of patients, the effects in any particular category may best be assessed indirectly by approximate extrapolation from the 44% (SD 4%) reduction in occlusion observed in the overview of all trials rather than directly from the effects observed only in that one category (fig 1). Such arguments would not materially alter the size of the estimated effects on coronary or peripheral artery patency but would suggest that the effect among haemodialysis patients, though certainly real, may not be quite as large as these small trials indicate.

Comparisons of different antiplatelet regimens Indirect comparisons of when to start therapy

During the period covered by this overview there were no direct randomised comparisons of starting treatment before the invasive vascular procedure versus starting treatment after it. Indirect comparisons between the average effects observed in trials of treatment started before the procedure and the average effects in trials where it was started after the procedure indicated that similar reductions in vascular occlusion could be achieved by starting before the procedure or within the first 24 hours afterwards (fig 3). But even in trials of treatment started somewhat more than 24 hours postoperatively, a highly significant reduction in occlusion was observed that was not significantly different from that seen in the trials of treatment started earlier. Nevertheless, the right hand ends of the confidence intervals for the sizes of the effects in figure 3 (which indicate the minimum benefit that is readily compatible with the trial results) show that the size of benefit that is statistically guaranteed to exist by these trials is substantially greater if treatment is started reasonably promptly.


Indirect comparisons of proportional effects of antiplatelet therapy started before or after vascular procedures on occlusion. Symbols and conventions as in figure 1

Direct and indirect comparisons of different regimens

The effects of different antiplatelet drug regimens (high dose aspirin, medium dose aspirin, aspirin plus dipyridamole, dipyridamole, sulphinpyrazone, ticlopidine, suloctidil) on “vascular events” (non-fatal myocardial infarction, non-fatal stroke, or vascular death) were reviewed in part I, with no clear evidence that one regimen differed much from another in its ability to prevent such vascular events. Likewise, in the present trials (figs 4,5) of the prevention of vascular occlusion neither direct comparisons nor indirect comparisons of the effects of different regimens on vascular patency provided good evidence that any one antiplatelet drug regimen studied was more effective than any other.

Some direct randomised comparisons of the effects on occlusion were available (fig 4), but these were generally too small to be reliable. The only apparent trend was towards a slightly better effect with somewhat lower doses of aspirin, but that trend was not statistically significant (and parts I and III of this overview5,9 suggest that higher and lower doses of aspirin are of about similar efficacy). Nevertheless, if aspirin is to be given to maintain arterial patency, then these trials provide no directly randomised evidence that higher (and hence more gastrotoxic87,88) aspirin doses are any more effective than 75-325 mg daily. Indirect comparisons of the effects observed on occlusion in separate overviews of trials comparing various different regimens with control (fig 5) likewise provide no clear evidence that any one antiplatelet regimen was more or less effective than another.


Direct comparisons of proportional effects on occlusion of different antiplatelet regimens. Symbols and conventions as in figure 1. Asp=Aspirin. Dip=Dipyridamole. Sulph=Sulphinpyrazone


Indirect comparisons of proportional effects on occlusion of different antiplatelet regimens. Symbols and conventions as in figure 1

Effects of antiplatelet therapy on bleeding

The table summarises the bleeding complications reported in all trials comparing antiplatelet therapy with control in patients who were to undergo, or had just undergone, a vascular procedure, irrespective of whether vascular patency was to be monitored systematically.

Among the trials in which some months of antiplatelet therapy was to be started just before the procedure there was an apparent excess of about one fatal bleed per 1000 patients allocated antiplatelet therapy (95% confidence interval 0 to 3 per 1000), but it was not quite statistically significant. There was a significant excess of non-fatal major bleeds reported (2.2% v 0.9%; 1P=0.002) but the absolute excess was small (about 13 (SD 4) per 1000). Information on other complications (reoperations, wound haematomas, or infections due to bleeding) was available from only a few trials. No large excess was apparent, but the incompleteness of the data precluded reliable estimation of the exact size of any excess.

Among the trials in which some months of antiplatelet therapy was to be started after the vascular procedure, however - in many of which the treatment was to be started <24 hours afterwards - there was no apparent excess of fatal or non-fatal major bleeds (nine among patients allocated antiplatelet therapy versus an adjusted control total of five) or of other haemorrhagic side effects (one versus one).

Reported bleeding complications in all trials of antiplatelet therapy among patients having vascular procedures (irrespective of whether patency was monitored systematically)

View this table:


Similar proportional reductions in occlusion

Among patients having coronary artery bypass operations, coronary artery angioplasty, leg artery bypass operations, or leg artery angioplasty, this overview of randomised trials provides unequivocal evidence that antiplatelet therapy (beginning either just before or, possibly slightly more safely, just after the procedure) reduces the odds of graft or arterial occlusion by about 40% while treatment continues. The numbers studied were largest, and hence the evidence was strongest, for patients having coronary and peripheral artery grafts, but there was also significant benefit after angioplasty. In absolute terms the benefits were substantial, occlusions being averted in about 40 to 90 patients per 1000 allocated antiplatelet therapy for about seven months after coronary artery procedures and about 19 months after peripheral artery procedures (fig 2).

When considering the effect of antiplatelet therapy it is important to remember that owing to non-compliance in the trials the trial results (and hence the present overview results) will systematically under-estimate the size of the effect produced by actually taking treatment. Moreover, patients who have had vascular interventions are at high risk not just of subclinical occlusion but also of major vascular events such as myocardial infarction, stroke, or vascular death. Part I5 showed that among many types of patients prolonged antiplatelet therapy prevents about a quarter of such vascular events, the effects being about as great (22% (SD 7%) odds reduction; 2P=0.002) among patients who had undergone a vascular procedure as among other types of patient.

Treatment for only six months of 1000 patients who have had coronary revascularisation procedures might therefore not only prevent several dozen from having coronary artery occlusions but, even more important, might avoid a few dozen serious or fatal clinical vascular events.5 Even among patients with peripheral arterial disease alone (who are at lower risk of myocardial infarction, stroke, or vascular death) treatment of 1000 patients with a year or two of antiplatelet therapy might not only prevent several dozen from having peripheral artery occlusions but also avert several more serious vascular events.5

Among haemodialysis patients antiplatelet therapy reduced the risk of occlusion of fistulas and shunts by about two thirds. The monthly rate of occlusion was high in these studies, and antiplatelet therapy reduced it from about 20% per month to about 10% per month (2P<0.00001; fig 2). It is relevant that these patients also suffer a particularly high incidence of serious vascular events6 that could be reduced by prolonged antiplatelet treatment.

Risk of bleeding Many patients at very high risk of subclinical occlusion or vascular events (for example, those with unstable angina or with suspected acute myocardial infarction who are having emergency coronary artery bypass surgery or angiography) have been denied antiplatelet therapy for fear of perioperative bleeding.89,90 This overview of trials of antiplatelet therapy during and after surgery suggests that the substantial benefits of such therapy (fig 2) generally outweigh any risks of bleeding (table). Detailed analysis of bleeding complications in a trial of aspirin started preoperatively in patients having coronary artery bypass surgery found a small but significant increase in drainage from the chest tube, in perioperative transfusion requirements, and in the reoperation rate, but there was no excess mortality due to bleeding complications.30

It has been suggested that starting antiplatelet therapy after invasive vascular procedures may reduce the risk of bleeding while still preventing occlusion.11 No direct randomised comparisons of starting before or just after vascular procedures were available for this overview, but indirect comparisons suggested that similar reductions in occlusion were produced by starting treatment before and just after these procedures. More recently, the results of a trial have been reported among 351 patients randomly allocated to receive aspirin starting preoperatively or six hours after coronary artery bypass graft surgery.91 Aspirin (325 mg daily) started preoperatively was associated with more blood being transfused (900 ml v 725 ml; 2P<0.05) and with more reoperations for bleeding (6.3% v 2.4%; 2P=0.07). But there were no statistically significant differences between the treatment groups in graft occlusion at an average of eight days after surgery (30 of 176 patients allocated preoperative antiplatelet therapy versus 27 of 175 allocated postoperative therapy), reinforcing - if on smaller numbers - the conclusion suggested by figure 3 that treatment started just after the vascular procedure may be about as effective as treatment started preoperatively.

Choice of antiplatelet regimen

Comparisons between different antiplatelet regimens may be based on several factors, including efficacy, ease of use, severity of side effects, and cost. This overview provides some direct and indirect randomised comparisons of the effects of different drug regimens on the prevention of occlusion but finds no evidence of any differences in efficacy. The numbers of patients studied and the numbers of events that occurred were, however, not large enough (even with the inclusion of recent trials*RF 92-94*; appendices 1 and 2) to exclude some small but real differences in efficacy between different drug regimens. Consequently, the choice of which regimen to use must be determined by other factors. At present medium dose aspirin is the most widely tested, most convenient, and least expensive antiplatelet treatment, with direct evidence of substantial reductions in occlusion and in important vascular events5 that would generally far outweigh any small risks of bleeding. The choice of which antiplatelet regimen to use is, however, of secondary importance: what chiefly matters is that some such treatment should routinely be considered for the vast majority of patients having vascular procedures.

Duration of antiplatelet treatment

It remains uncertain how long antiplatelet therapy should be continued after vascular procedures. As yet there are no large directly randomised comparisons of different durations of treatment (though a recent small trial comparing two weeks versus six months of aspirin after coronary angioplasty suggested some additional effect on restenosis with continued treatment95). The highest absolute risk of occlusion generally occurs during the first few months, while any traumatic damage is healing, but even after those first few months a small percentage of patients per year may suffer late occlusions, as well as other more serious vascular events. The average duration of treatment in these trials was only about one year, but indirect evidence from studies in other settings suggests that more prolonged treatment may be more effective (see part I). It may therefore be prudent to consider continuing antiplatelet therapy after vascular procedures for as long as the risk of occlusive vascular events remains high.

Appendix 1

Individual results of unconfounded randomised comparisons of antiplatelet therapy with control (A upisilon C)for maintaining vascular patency

View this table:
Appendix 2

Individual results of unconfounded randomised comparisons of one antiplatelet regimen with another (A1 upsilion A2) for maintaining vascular patency

View this table: