Famotidine and Coronavirus Disease 2019

Dear Editors:

Yeramaneni et al¹ reported results from a retrospective study testing associations between the use of famotidine and outcomes among patients with coronavirus disease 2019 (COVID-19). Like our recent retrospective study on the same topic,² they classified the use of famotidine based on exposure within 24 hours after hospital admission and followed patients with COVID-19 for death for up to 30 days. Interestingly, although our study found a nearly 2-fold protective association between the use of famotidine and death or intubation (adjusted hazard ratio, 0.42; 95% confidence interval [CI], 0.21–0.85), Yeramaneni et al found no association between famotidine and death (adjusted odds ratio, 1.59; 95% CI, 0.94–2.71). Why might the 2 studies, so similar in design, have such different results?

First, it is possible that differences related to institutional patterns of use of famotidine underlie the discrepancy in study findings. For example, if famotidine was often used for stress ulcer prophylaxis in critically patients at Yeramaneni et al’s institution, then patients who received famotidine may have been sicker at baseline than those who did not. Sixteen percent of patients used famotidine in Yeramaneni et al’s study compared with 5% in ours, implying a fundamental difference related to institutional patterns of use. Before matching, patients who used famotidine at Yeramaneni et al’s institution were sicker in almost every way (higher oxygen requirements, more comorbidities, etc), whereas this was not true in our cohort. After matching, differences within Yeramaneni et al’s cohort are likely to persist in the unmatched categories. Given the significant baseline differences between those who used famotidine and those who did not, these residual confounders would likely bias results toward showing harm associated with famotidine.

Second, home use of famotidine may help to explain the differences between studies. An assumption of our study was that use of famotidine in the hospital represented a continuation of home use of famotidine. Intriguingly, home use of famotidine in Yeramaneni et al’s study seemed to have the opposite relationship with death compared with use in the hospital (adjusted odds ratio, 0.49 [95% CI, 0.16–1.52] for home use of famotidine vs 1.59 [95% CI, 0.94–2.71] for use of famotidine in the hospital). This hint of an interaction between home and hospital use of famotidine is puzzling and suggests that hospital use of famotidine does not represent a continuation of home use in Yeramaneni et al’s study. An analysis of home use of famotidine in Yeramaneni et al’s prematched cohort, excluding those who used famotidine in the hospital, would be interesting. One possibility is that early, but not late, use of famotidine may be beneficial in COVID-19.³

Examining the totality of evidence, what do we have? Our study and other retrospective studies of famotidine suggest there may be an association between the use of famotidine and improved outcomes among hospitalized patients with COVID-19^{4 , 5}; this was also suggested by a case series of famotidine with quantitative symptom tracking in nonhospitalized patients.³ The data from Yeramaneni et al and other retrospective studies^{6 , 7} show no association. We agree with Yeramaneni et al that famotidine should only be used as COVID-19 therapy in the context of a clinical trial. Such trials are ongoing, and the results of these trials will be the crucial next step in answering the question of whether there is a role for famotidine in the treatment of COVID-19.^{8 , 9}