Meta-analysis: The Effects of Proton Pump Inhibitors on Cardiovascular Events and Mortality in Patients Receiving Clopidogrel

C. S. Kwok; Y. K. Loke

Disclosures

Aliment Pharmacol Ther. 2010;31(8):810-823. 

In This Article

Results

The search results yielded 23 relevant studies with 93 278 patients and the process of selection is shown in Figure 1. These comprised of 20 retrospective studies, two post hoc analyses of randomized trial participants and one prospective RCT.[12–34] Main characteristics of the included studies are described in Table 1. The outcomes, interventions and quality assessments of the included studies are shown in Table 2.

Figure 1.

Flow diagram of the process of article selection for meta-analysis.

As many of the included studies were presented in abstract form, we found it difficult to assess methodological quality fully. A majority of studies were based on registries of patients undergoing PCI, or were part of a health insurance database/reimbursement scheme. Database studies relied on prescription claims to ascertain exposure, while outcomes were commonly assessed based on ICD-9 codes. Misclassification or inconsistent recording of exposures and outcomes is a possibility within such databases. In contrast, the few studies that relied on data from trial participants generally had more rigorous follow-up and predefined ascertainment of CV events that included adjudication by an independent committee.[13,15,22] Four studies used a propensity scoring method to try to match the participants and reduce potential confounding.[20,22,25,30]

Twelve studies reported either myocardial infarction or acute coronary syndrome events as an outcome (Figure 2). There was a high degree of heterogeneity (I 2 = 77%) in the overall analysis. Unadjusted data from observational studies gave the highest RR of 1.82 (95% CI: 0.90–3.70) that fell to an RR of 1.54 (95% CI: 1.23–1.92) with data adjusted for confounders and an RR of 1.15 (95% CI: 0.89–1.48) based on propensity-matched or trial participants.

Figure 2.

Meta-analysis of myocardial infarction or acute coronary syndrome with clopidogrel and proton pump inhibitor use.

Thirteen studies were included in the analysis of all-cause mortality (Figure 3). The overall pooled result was nonsignificant (RR 1.09, 95% CI: 0.94–1.26) with some degree of heterogeneity of I 2 = 60%. Similar to the analysis of MI above, the pooled RR was highest with the unadjusted observational data and lowest (RR 1.00, 95% CI: 0.94–1.26) with the propensity-matched or trial participants.

Figure 3.

Meta-analysis of overall mortality with clopidogrel and proton pump inhibitor use.

Nineteen studies were included in the analysis of MACE (Figure 4). Definitions of the composite endpoints within MACE varied considerably between studies and this may have accounted for the substantial heterogeneity in the overall analysis, as illustrated by the diverse pattern in the Forest plots (I 2 = 79%). Again, data from propensity matched or trial participants gave the lowest pooled RR of 1.07 (95% CI: 0.90–1.28), whereas the adjusted observational data yielded a pooled RR of 1.44 (95% CI: 1.24–1.67).

Figure 4.

Meta-analysis of major adverse cardiac events with clopidogrel and proton pump inhibitor use.

processing....