Reply to Althuwaybi et al.: Hospitalization Outcomes for COVID-19 in Patients with Interstitial Lung Disease: A Potential Role for Aerodigestive Pathophysiology?

From the Authors:

Althuwaybi and colleagues propose an interesting hypothesis to explain the potential mechanism for increased risk of mortality following hospitalization for coronavirus disease (COVID-19) in patients with interstitial lung diseases (ILDs) (1). Unfortunately, in our study these data were not collected, and therefore, we cannot not make comparisons with the control population. Gastroesophageal reflux is strongly linked to hiatus hernia, which is twice as prevalent in patients with idiopathic pulmonary fibrosis (IPF) as the general population (2, 3). It is therefore plausible that proton pump inhibitors (PPIs) may have been prescribed in excess in patients with ILD.

However, the cited evidence for a putative role of PPIs in pathophysiology is inconclusive and likely does not overcome the residual effect of significant comorbidities in our cohort. In an unpowered pilot randomized controlled trial into the effect of omeprazole therapy on cough in IPF, Dutta and colleagues reported safety events including three respiratory tract infections in the placebo group compared with six using omeprazole, which could be explained by chance (4). In a large population-based study of patients with IPF in a real-world clinical practice setting, PPI use was not associated with a difference in survival or the incidence of respiratory-related hospitalization compared with those not using PPIs (5). Althuwaybi and colleagues describe a dose response of PPIs with risk of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection reported in a large North American survey, as well as a retrospective health insurance database study in Korea that observed an association of PPI use with COVID-19 severity (6, 7). In contrast, the same Korean study found no association of PPIs with risk of infection; further study and meta-analyses will be required to build certainty on an effect. Lee and colleagues did observe a significant association between PPI use and severe outcomes in an adjusted model, particularly for a subgroup with <30 days of PPI records (7). Although the study included 132,316 individuals, a more limited 267 were currently using PPIs and were positive for SARS-CoV-2. A 79% greater risk of severe intervention was observed from 24/267 people with current PPI use (18/175 with <30 d of records), in comparison with 14/267 propensity-matched controls. The propensity-matched group had fewer significant comorbidities, and analyses did not adjust for lung function or body mass index, which we identify as important predictors of hospitalization outcomes. PPI use was not available for the individuals included in the study by Drake and colleagues, although the cited studies suggested little effect in past users or those with >30 days of treatment. It is likely that disease status, enhanced respiratory support, and significant comorbidities had a more substantial impact than PPI on outcomes in our cohort.

Furthermore, the role of PPIs in IPF pathogenesis remains uncertain. Post hoc analysis from international trials in IPF reported that pulmonary infections were higher in patients with advanced IPF (i.e., FVC <70%) who were receiving antacids compared with those not treated with antacids (8). However, PPIs have pleiotropic activity including antiinflammatory and antiproliferative effects (9). It is also worth noting, however, that although hiatus hernia has been independently linked to mortality in patients with IPF (2) and almost 5,000 participants in the AGES-Reykjavic birth cohort (3), in these uncontrolled studies, no association was found between PPI prescription and outcomes. The potential of antacid drugs to contribute to viral infections including COVID-19 has been the subject of some discussion (10). Thus, until further data are available, the potential adverse or favorable effects of PPIs in ILDs during the current pandemic remain to be determined, but understanding the balance of benefits and potential harms from antacids is imperative.

There are also a number of other potential mechanistic explanations why patients with ILD may have had poor outcomes following COVID-19, including an increase in SARS-CoV-2 entry genes, such as ACE2 (angiotensin-converting enzyme 2), as well as baseline alterations in IL-6 and type 1 IFN response genes in cells from patients with ILD (11). Furthermore, patients with ILD, but especially IPF, have high levels of the αvβ6 integrin in their alveolar epithelium, which increases mortality in response to a host of viral and bacterial pathogens in animal models (12), is associated with a worse prognosis in patients with IPF (13), and contains a binding site for SARS-CoV-2 virus (14). Finally, SARS-CoV-2 is known to act on the systemic and pulmonary vasculature and patients with IPF are at higher risk of cardiovascular disease and pulmonary thromboembolism (15–17).

Therefore, although at the current time it is not possible to determine whether host or iatrogenic factors are responsible for our observation, this is clearly an area in need of further investigation. We would like to reassure Althuwaybi and colleagues that the audit has been expanded and future waves now record gastroesophageal reflux disease to support such lines of investigation; however, the current evidence for a role of PPIs in the severity of COVID-19 hospitalizations in ILD is insufficient.