Twenty-five studies (n=3,688) were included: 2 RCTs (n=482), 3 case-control studies (n=233) and 20 prospective cohort studies (n=2,973).
Prevalence of clopidogrel responsiveness (22 studies, n=2,574).
The pooled prevalence of clopidogrel responsiveness was 21% (95% CI: 17, 25).
Statistically significant heterogeneity was detected (I-squared 90.5%; p<0.0001).
A higher prevalence of nonresponsiveness was found when it was measured within 24 hours (36%, 95% CI: 28, 44) of the loading dose compared with 24 to 48 hours (13%, 95% CI: 5, 21), 2 to 7 days (10%, 95% CI: 2, 18) or later (0%, 95% CI: 0, 7). There was no difference in prevalence across different time period when only studies that used a 600-mg loading dose of clopidogrel were analysed. A lower adjusted prevalence of nonresponsiveness was found when the loading dose of clopidogrel was 600 mg (7%, 95% CI: 0, 15) compared with 300 mg (22%, 95% CI: 15% 29). The prevalence of nonresponsiveness was not associated with the measurement method or concomitant dose of aspirin.
Clinical outcomes (11 studies, n=2,319).
The pooled OR showed a significant increase in the risk of cardiovascular outcomes for patients with clopidogrel unresponsiveness (8.0, 95% CI: 3.4, 19.1), based on 8 studies with data available for pooling. Statistically significant heterogeneity was detected (I-squared 63%; p=0.009). In their discussion, the authors stated that almost all studies showed a positive relationship between unresponsiveness and the risk of cardiovascular events. The pooled PPV for nonresponsiveness was 34% and the NPV was 92%. There was no significant difference in the pooled OR for clopidogrel nonresponsive patients in the occurrence of stent thrombosis (3 studies, n=218), but the OR was significantly increased for a composite end point of cardiovascular events (OR 12.0, 95%: 5.9, 24.4; 4 studies, n=837); the PPV was 33% and the NPV 96%.