In Reply: Mr Osnabrugge and colleagues note the presence of heterogeneity for 1 end point evaluated in the meta-analysis. As described in the article, the presence of heterogeneity of risk was assessed using the Q statistic and the degree using the I2 measure, which reflects the percentage of the total variability due to between-study heterogeneity vs within-study variability. If heterogeneity was found, then sensitivity analyses were performed, serially excluding studies to determine the source of heterogeneity. Therefore, rather than remove trials based on subjective assessment of the baseline characteristics of patients as suggested, we serially removed studies on objective grounds based on their numerical contribution to heterogeneity. In these sensitivity analyses, exclusion of 2 trials yielded a result that no longer displayed heterogeneity (P = .59 for heterogeneity, I2 = 0%). The resultant point estimate for cardiovascular death, myocardial infarction, or stroke among carriers of 1 reduced-function CYP2C19 allele vs noncarriers was similar to that calculated with all 9 studies, demonstrating that even after restricting the analyses to a statistically homogenous group of studies, the results did not fundamentally change. Additionally, as clinicians are well aware, P2Y12 blockade is particularly efficacious in reducing the risk of stent thrombosis. Correspondingly, we found that both patients who carried 1 reduced-function CYP2C19 allele, as well as those who carried 2, had a large and highly statistically significant risk of stent thrombosis, with no evidence of heterogeneity.
Mega JL, Sabatine MS. Carriage of Reduced-Function CYP2C19 Allele Among Patients Treated With Clopidogrel—Reply. JAMA. 2011;305(5):467-468. doi:10.1001/jama.2011.78