Our study shows that pre-stroke treatment with A + D does not reduce acute stroke severity in comparison with aspirin only. However, at the end of the 12-months follow-up period, patients on A + D had lower all-cause mortality compared with patients on A only. There was no difference in functional outcome and/or non-fatal cardiovascular events between patients on A + D or A only during follow-up.
In some previous studies
[1–3], pre-stroke aspirin use has been associated with milder clinical deficits at stroke onset, but reduction in stroke severity may also differ by stroke mechanisms
Few and, partly inconclusive, data are available on the effect of pre-stroke treatment with A + D on stroke severity
. Theoretically, D is suggested to be neuroprotective
 and may reduce the inflammatory process. These effects in combination with the antiplatelet effects of aspirin might provide more efficient neuroprotection
However, our study, in agreement with recent data
 does not suggest that use of A + D before stroke onset would lessen severity of acute stroke. In spite of this, any conclusions must be drawn with caution since the number of patients using A + D before stroke was considerably smaller than the number of patients using aspirin only.
Our finding of reduced all-cause mortality in the A + D group at 12months follow-up is somewhat puzzling. Previously two large, randomised, studies
[8, 9] have shown A + D to be effective for secondary stroke prevention, but all-cause mortality was not reduced. Furthermore, in TIA patients The American-Canadian Co-Operative Study Group (21) has demonstrated that after a mean follow-up of 25 months there was no difference in mortality, stroke or retinal infarction between patients randomised to aspirin plus placebo or aspirin plus dipyridamole. There are several explanations for this discrepancy between our results and previous data. One explanation could be that the present study was non-randomised in contrast to previous randomised trials, which means that our results must be interpreted with extreme caution. In previous studies
[8, 9, 21] the mean follow-up was between 24 months and 3.5 years in comparison with 12-months follow-up in the present study. During longer follow-up, risk factors for cardiovascular mortality may be more impending and may counteract the early positive effects of A + D.
The difference in mortality rate between A + D vs. aspirin only users in the present study was solid in spite of a smaller population sample and was demonstrated early after stroke. It may be hypothesized, that time from onset of stroke to initiation of antiplatelet therapy may be important.
In the present study acute stroke patients were enrolled if they had experienced a stroke/TIA within preceding 7 days. In previous, randomised studies
[8, 9, 21] patients were eligible for inclusion in the non-acute setting i.e. the index stroke must have occurred within 3 and, respectively, 6 preceding months. Thus, in the present study treatment with antiplatelet was initiated very early and considering the suggested neuroprotective effects of dipyridamole and the antithrombotic properties of aspirin it may be hypothesized, that these early effects may be important for the significant reduction in all-cause mortality.
This association is further delineated by the clear difference in all-cause mortality between A + D vs. aspirin only users in the present study in comparison with previous studies
[8, 9, 21].
Our patients were older, than in previous studies
[8, 9, 21], which must be considered as a factor of potential importance. Increasing age may substantially augment mortality rate but in spite of this, the consistent effects on mortality persisted during the whole observational period. Furthermore, patients treated with Actilyse® (Boehringer-Ingelheim) were not included in the study. Only around 6.6% of all stroke patients were treated by Actilyse® during this period in Sweden so, tentatively, inclusion of Actilyse® treated patients would only have a marginal effect on the present results
There was no difference at discharge in systolic or diastolic blood pressure nor in LDL-C, total cholesterol and triglycerides. If anything, HDL-C was slightly lower in the A + D group which, conversely, may suggest an increased risk profile in A + D users, but the difference was quite small. Therefore, it is not possible to discern a higher risk profile in the aspirin only group.
In disparity with previous studies
[8, 9] there was no difference in secondary stroke prevention in A + D users vs. users of aspirin only. Conceivably, this finding may depend on the low numbers of recurrent strokes during follow-up.
There are several limitations of the present work. This study was not randomised, but a post-hoc analysis of data collected from a stroke registry study. Our findings, which are in contradiction with results from several randomized trials, should therefore be interpreted with great caution. Our sample size was also much smaller than in previous randomised studies
[8, 9, 21]. Numerous factors may influence use of A + D or aspirin only and these factors may be distributed dissimilarly among the two groups.
However, the influence of potential confounding factors such as age, stroke severity, previous stroke, previous myocardial infarction, and diabetes type 2, were taken into account and adjusted for in the survival analysis. We cannot, however, completely rule out that there are still some unknown factors that have substantially influenced our findings.
For pre-stroke use of A + D or aspirin only information was obtained from patients, next of kin or medical records, but there may be a bias even if information was meticulously collected. At discharge we recorded use of drugs but we cannot confirm that all patients remained on these drugs during follow-up.
The strength of the present study is that the population is large and that most patients have been followed for 12months.
Furthermore, all admitted stroke patients were included in the present study and there were very few patients lost to follow-up. However, our findings may not be valid for a general stroke population, but merely for a somewhat older stroke population, given the rather high mean age of our study population.