Fifteen studies met the inclusion criteria: seven non-prospective (n=1,466, range 28 to 588); and eight prospective (n=1,227, range 47 to 325). Blinding of both biologists and clinicians was reported in two studies, biologists only in two studies and clinicians only in three studies.
When the results of the prospective studies were pooled, aspirin non-responders had an increased risk of recurrent ischaemic events compared to aspirin responders (OR 2.1, 95% CI 1.4 to 3.4, p<0.001; eight studies); there was no statistically significant heterogeneity observed between studies when a random-effects model was used.
Of the seven non-prospective studies, five reported a significant association between a short PFA and the occurrence of an ischaemic event. Odds ratios ranged from 1.8 (95% CI 1.0 to 3.3) to 18 (95% CI 0.99 to 327); publication bias was observed for these studies.