Monday, July 20, 2020

PPI use and COVID-19 infection: a meaningful correlation?

- Jennifer Middleton, MD, MPH

In its current issue, the American Journal of Gastroenterology preliminarily published the results of a health survey correlating proton pump inhibitor (PPI) use with COVID-19 infection. The journal placed a disclaimer at the top of the document noting that "[t]his version will undergo additional copyediting, typesetting and review before it is published in its final form." Its findings, widely reported in the lay press, may result in patient questions regarding PPI safety during the pandemic, but discontinuing PPIs on the basis of this study is likely premature.

The authors enlisted an online research firm to conduct a health survey throughout the United States. The survey asked participants about their use of PPIs or H2-receptor antagonists (H2RA) and also asked if they had been diagnosed with COVID-19. The manuscript does not specify how the 250,000+ adults were chosen to receive a survey invitation, but they report that over 86,000 took the survey. Of those persons who responded between May 3 and June 24, 2020, just over 53,000 reported gastroesophageal reflux disease (GERD) symptoms. This cohort of 53,160 individuals comprised the study population. 6.4% of this group reported a positive COVID-19 test. The researchers then compared PPI and H2RA use among participants who did and did not report a positive COVID-19 test. The odds ratio (OR) for a positive COVID-19 test among participants reporting use of a PPI once a day, compared to those not taking any antacid therapy, was 2.15 (95% confidence interval [CI] 1.90-2.44); the odds ratio among those reporting use of a PPI twice daily for a positive COVID-19 test was 3.67 (95% CI 2.93-4.60). The results for H2RA use were not deemed statistically significant (though the CI for once daily H2RA, compared to no therapy, fell just below 1.0 [OR 0.85, 95% CI 0.74-0.99]).

Methodological concerns with this study include an inadequately-described participant recruitment strategy and the decision to narrow the study cohort to only those participants who self-reported GERD symptoms. The demographic information in Table 1 also brings up some odd discrepancies; while it's not unexpected that a larger percentage of participants with COVID-19 used tobacco, identified as Latinx, and had not completed high school, it's harder to explain why a significantly higher percentage of participants with COVID-19 were married compared to those without a COVID-19 diagnosis (74.5% vs 27.1%) or reported an annual income of greater than $200,000 (63.5% vs 7.4%).

PPI use has previously been associated with an increased risk of enteric infections, pneumonia, and fractures, though a 2019 meta-analysis of PPI safety largely refuted those claims. Hopefully, future research will further elucidate any possible PPI-COVID connection; in the meantime, it's certainly advisable to continue to discourage twice-daily PPI use for most patients, since once-daily PPI use results in similar symptom improvement