No short-term link found between PPIs, heart attack

Proton pump inhibitor conferred no short-term increase in risk for myocardial infarction in a large retrospective insurance claims study. Over a median follow-up of 2-3 months, estimated weighted risks of first-ever MI were low and similar regardless of whether patients started PPIs or histamine-receptor antagonists (H2RAs).

“Contrary to prior literature, our analyses do not indicate increased risk of MI in PPI initiators compared to histamine2-receptor antagonist initiators. Epidemiologic studies have produced mixed findings on PPI use and MI risk.

Animal models and ex vivo studies of human tissue indicate that PPIs might harm coronary vessels by increasing plasma levels of asymmetrical dimethylarginine, the investigators noted.

To further assess PPIs and risk of MI while minimizing potential confounding, they studied new users of either prescription PPIs or an active comparator, prescription H2RAs.

The dataset included administrative claims for more than 5 million patients with no MI history who were enrolled in commercial insurance plans or Medicare Supplement Insurance plans.

The study data spanned from 2001 to 2014, and patients were followed from their initial antacid prescription until they either developed a first-ever MI, stopped their medication, or left their insurance plan.

After controlling for numerous measurable clinical and demographic confounders, the estimated 12-month risk of MI was about 2 cases per 1,000 commercially insured and about 8 cases per 1,000 Medicare Supplemental Insurance enrollees.

The estimated 12-month risk of MI did not significantly differ between users of PPIs and H2RAs, regardless of whether they were enrolled in commercial insurance plans (weighted risk difference per 1,000 users, -0.08) or Medicare Supplemental Insurance (weighted risk difference per 1,000 users, -0.45) plans.

Each antacid class also conferred a similar estimated risk of MI at 36 months, with weighted risk differences of 0.44 per 1,000 commercial plan enrollees and -0.33 per 1,000 Medicare Supplemental Insurance plan enrollees, the researchers reported.

Weighted estimated risk ratios also were similar between drug classers, ranging from 0.87 at 3 months among Medicare Supplementary Insurance enrollees to 1.08 at 36 months among commercial insurance plan members.

Results reassuring

In the late 2000s, several large epidemiologic studies suggested that proton pump inhibitors (PPIs) increase the risk for MI in users of clopidogrel. There was a proposed mechanism: PPIs competitively inhibit cytochrome P450 isoenzymes, which blocked clopidogrel activation and, ex vivo, increased platelet aggregation.

It sounded scary – but fortunately, some reassuring data quickly emerged. In 2007, the COGENT trial randomized patients with cardiovascular disease to a PPI/ clopidogrel versus a placebo/clopidogrel combination pill.

After three years of follow-up there was no difference in rates of death or cardiovascular events. In the glaring light of this randomized controlled trial data, earlier studies didn’t look so convincing.

So, why won’t the PPI/MI issue die? In part because COGENT was a relatively small study. It included 3,761 patients, but the main result depended on 109 cardiovascular events. Naysayers have argued that perhaps if COGENT had been a bigger study, the result would have been different.

In this context, the epidemiologic study provides further reassurance that PPIs do not cause MI. Two insurance cohorts comprising over 5 million patients were used to compare PPI users with histamine – receptor antagonist users after adjusting for baseline difference between the two groups. The large size of the dataset allowed the authors to make precise estimates: we can say with confidence that there was no clinically relevant PPI/MI risk in these data.

Can we forget about PPIs and MI? These days, patients worry more about dementia or chronic kidney disease. But the PPI/MI story is worth remembering. Large epidemiologic studies are sometimes contradicted and need to be evaluated in context.