We read with great interest the recent large population-based study by Nolde et al1 reporting that proton pump inhibitors (PPIs) did not increase the risk of myocardial infarction (MI) in the first decade after treatment initiation in comparison with H2-receptor antagonists (HR 0.96; 95% CI 0.80-1.16). Although this study has strengths, we have some concerns.

First, the study only described the overall safety of PPIs and did not present data about individual PPIs. An analysis of FDA data found that, among the PPIs, esomeprazole appeared to have the highest risk of MI, whereas lansoprazole was associated with a lower rate.2 In addition, Nolde et al included patients taking anti-platelet agents (clopidogrel and aspirin). The safety of PPIs for patients taking anti-platelet agents, especially clopidogrel, has been controversial. A meta-analysis showed that the administration of a PPI with clopidogrel was associated with a 29% increased risk of combined major cardiovascular events, and a 31% increased risk of MI.3 The interaction between clopidogrel and PPIs is agent-specific.4 Previous studies have shown that omeprazole and esomeprazole were more likely to inhibit the activation of clopidogrel than other PPIs such as pantoprazole.5-7 The interaction between clopidogrel and certain PPIs is based on competitive inhibition of cytochrome P450 2C19 (CYP2C19). Clopidogrel is converted to its active metabolite by CYP2C19, and PPIs are also metabolised by this enzyme to various degrees. Therefore, whether different PPIs may be associated with different risks of MI remains unclear.

Second, Nolde et al did not provide data about recurrent MI in the 10 years follow-up with PPI use. A previous study showed that concomitant use of clopidogrel and PPIs was still associated with increased recurrent adverse cardiovascular outcomes, including MI following coronary angioplasty.8 Even in clopidogrel non-users, current PPI use could also be associated with an increased risk of recurrent MI (OR 1.38, 95% CI 1.18-1.61) when compared with no PPI use.9

We congratulate Nolde et al for their novel study that reduces the concern about increased MI risk with PPIs suggested by previous research.10 However, the safety of PPIs in certain populations still remains to be further elucidated.

Declaration of personal interests: None.