Article Text
PDF
Statistics from Altmetric.com
We read with interest the study by Lee et al.1 The authors conducted a propensity score (PS)-matched analysis of a national South Korean cohort evaluating the association between proton pump inhibitor (PPI) use and SARS-CoV-2 susceptibility (primary outcome) and COVID-19 clinical severity (secondary outcome). Between January and May 2020, 4785 patients tested positive for SARS-CoV-2 (3.6% positivity); 267 current PPI users and 148 former PPI users were 1:1 PS-matched to non-users for the secondary outcomes. The authors reported current PPI use versus non-use was associated with a statistically significant increased risk of the composite endpoints: (1) oxygen therapy, intensive care unit (ICU) admission, mechanical ventilation use or death (composite OR 1.63; 95% CI, 1.03–2.53); and (2) ICU admission, mechanical ventilation or death (composite OR 1.79; 95% CI, 1.30 to 3.10).
We assembled a national retrospective cohort of US veterans who tested positive for SARS-CoV-2 (index date). Current outpatient PPI use up to and including the index date (primary exposure) was compared with non-use, defined as no PPI prescription fill in the 365 days prior to the index date (online supplemental figure 1). The primary composite outcome was mechanical ventilation use or death within 60 days; the secondary composite outcome also included hospital or ICU admission. In contrast to PS matching, PS weighting allowed inclusion of all patients. Weighted logistic regression models evaluated severe COVID-19 outcomes between current PPI users versus non-users.
Supplemental material
Our analytic cohort included 97 674 veterans with SARS-CoV-2 testing, of whom 14 958 (15.3%) tested positive (6262 (41.9%) current PPI users, 8696 (58.1%) non-users). In the unweighted cohort, current PPI users were older, more often current or former smokers, and had more comorbidities than non-users. After weighting, all covariates were balanced (table 1, online supplemental figure 2). In the unweighted cohort, we observed higher odds of the primary (9.3% vs 7.5%; OR 1.27; 95% CI, 1.13-1.43) and secondary (25.8% vs 21.4%; OR 1.27; 95% CI, 1.18-1.37) composite outcomes among PPI users versus non-users (figure 1, online supplemental table 1). After PS weighting, PPI use versus non-use was not associated with the primary (8.2% vs 8.0%; OR 1.03; 95% CI, 0.911.16) or secondary (23.4% vs 22.9%; OR 1.03; 95% CI, 0.95-1.12) composite outcomes. There were no significant interactions between age and PPI use on composite or individual outcomes.
Forest plot of primary and secondary COVID-19 outcomes within 60 days of the index date, weighted and unweighted cohorts (semi-log scale, range 0.1 to 10). In the unweighted cohort, current outpatient PPI use compared with PPI non-use was associated with increased odds of severe COVID-19 outcomes, defined based on composite (primary: death or mechanical ventilation; secondary: death, mechanical ventilation, ICU admission or hospitalisation) and individual component outcomes. Each of these associations were statistically non-significant after more fully accounting for covariates in the propensity-weighted cohort, including date of SARS-CoV-2 testing and VHA facility location. Of note, there was no significant interaction between age group and PPI use on these outcomes. ICU, intensive care unit; PPI, proton pump inhibitor.
Characteristics of veterans with positive SARS-CoV-2 testing, stratified by current PPI user versus PPI non-user
Disparate results are reported in studies analysing COVID-19-related outcomes among PPI users versus non-users2–6 due to varied PPI exposure definitions; COVID-19 severity outcomes; covariate assessment and adjustment; study design and populations; contemporaneous treatments; and healthcare infrastructure. In our unweighted analysis, we also observed an association between PPI use and severe COVID-19 outcomes (separately and as composites) which was not demonstrated in the PS-weighted cohort, suggesting that the associations in previous studies might reflect incomplete covariate adjustment.7 Indeed, the low E-values (all <2.0) for the weak associations between PPI exposure and COVID-19 severity outcomes (although variably defined) that are demonstrated in previous studies suggest incomplete covariate adjustment and residual confounding (see online supplemental material).8 Similar to the Lee et al study, prior studies also include data from the first months of the COVID-19 pandemic, when management and available treatments were rapidly evolving. Lee et al’s outcome definition also included oxygen therapy. Oxygen administration may not correlate with COVID-19 severity and may be considered routine protocol, especially early in the pandemic. Similarly, ICU admission may be influenced by health system factors, such as bed availability. Our study was designed to avoid immortal time, lag time and protopathic biases, which have been present in some PPI studies (see online supplemental material).9 We further accounted for the pandemic timeframe and clinical management evolution by considering COVID-19 prevalence and US geography.
In conclusion, with respect to COVID-19, our robust PS-weighted analysis provides patients and providers with further evidence for PPI safety.
Ethics statements
Patient consent for publication
Acknowledgments
Jason Denton (in memoriam): We would like to acknowledge and remember Jason Denton, a friend, colleague and team member, who contributed to database development and data stewardship.
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors SS helped in study concept, study design, dataset verification, interpretation of data and statistical analysis, drafting of initial manuscript and critical revision of the manuscript. AH and RG were responsible for study design, primary statistical analysis, dataset verification, interpretation of data and statistical analysis, critical revision of the manuscript and methodological oversight. CD and OW helped in dataset creation and stewardship. BM, ST, KC, AS, CH, ES, MM and AH performed manuscript revision. CR helped in study design, interpretation of data and statistical analysis, methodological and study oversight and critical revision of manuscript. All authors approved the final version of the manuscript.
Funding American Gastroenterological Association 2019 Research Scholar Award (SS); US Dept of Veterans Affairs ICX002027A01 (SS), Million Veteran Programme Core MVP000 (KC); National Institute of Health K23HL143161A01 (ST).
Competing interests The authors report no conflicts of interest that are relevant to this article. Dr. Shah is an ad hoc consultant for Phathom Pharmaceuticals.
Provenance and peer review Not commissioned; internally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.