Reply to "A limitation regarding the association between intranasal corticosteroid use and better COVID-19 outcomes: Nasal symptoms matter".

To the Editor:

We thank the editors for the opportunity to respond to the comments by Maeda et al regarding our article, and appreciate their positive comments. We acknowledge that our article was limited by its retrospective design and the information available in the Cleveland Clinic COVID-19 Research Registry (CCCRR). Despite this, the significant associations between intranasal corticosteroid (INCS) and improved COVID-19 outcomes highlight the need for a randomized controlled trial (RCT) to determine whether INCS are effective in treating or preventing COVID-19.

Maeda et al hypothesized that the association of early nasal symptoms with improved COVID-19 outcomes was confounded by INCS therapy and recommended repeat analysis adjusting for nasal symptoms. Although we appreciate their recommendation, the CCCRR does not include data on nasal symptoms. Nonetheless, adding nasal symptoms as a covariate to our models, although helpful, does not preclude the need for a prospective cohort study or an RCT to corroborate our findings.

We note that observational studies of nasal symptoms in COVID-19 have been limited by small sample sizes and missing data. Kim et al demonstrated that both nasal congestion (34.3%) and rhinorrhea (26.2%) were relatively common in a cohort of 172 symptomatic patients with COVID-19. However, larger studies, including a systematic review of 1,770 patients with a positive COVID-19 test, showed that nasal congestion (4%) and rhinorrhea (2%) were relatively uncommon in COVID-19. Rhinorrhea (1% to 6.8%) and nasal congestion (3% to 4.8%) were also uncommon in two additional systematic reviews. , These conflicting results and the low prevalence of nasal symptoms support our concern that data on nasal symptoms, extracted from electronic health records, are frequently incomplete and misleading. Hence, a prospective observational study or an RCT is needed to test the hypothesis raised by Maeda et al.

Our analysis sought to exclude acute INCS therapy during COVID-19, as noted by a median (interquartile) time between INCS prescription and COVID-19 testing of 379 (147-679) days. To address this further, we repeated the analysis after excluding patients (n = 837) who had an INCS prescription within 14 days of COVID-19 test date and found similar results, including lower risk for COVID-19 related hospitalization (adjusted odds ratio [OR] [95% confidence interval [(CI)]: 0.77 [0.71-0.83]), admission to the intensive care unit (0.74 [0.63-0.87]), and in-hospital mortality (0.72 [0.57-0.89]).

Compared with non-users, INCS users reported more age-adjusted comorbidities and an increased use of medications at baseline. The INCS users were more likely to have asthma, and reported more respiratory symptoms during acute COVID-19, consistent with united airway disease when upper and lower airways are considered a united morphologic and functional unit. Despite this, our adjusted analysis associated INCS independently with improved COVID-19 outcomes. The association we demonstrate is founded on plausible biologic mechanisms, sound statistical methods, and a large sample size. Our findings are also supported by a recent study that associated fluticasone propionate with ACE2 gene suppression in human nasal epithelial cells. Given these results, we believe it is time for an RCT to determine the role of INCS in the management of COVID-19.