Comorbidities and Procedures Associated with Mortality in COVID-19 Hospitalizations

Abstract

COVID-19 is associated with multisystem illness, but national evidence on mortality correlates and selected inpatient procedures remains limited. We examined associations of documented COVID-19, comorbidities, bronchoscopy, and left heart catheterization with in-hospital outcomes. This retrospective cross-sectional analysis used adult discharges from the Healthcare Cost and Utilization Project National Inpatient Sample from 1 April 2020 through 31 December 2021. Survey weights, strata, and hospital clusters generated national estimates. Survey logistic regression estimated associations with in-hospital mortality, and survey-weighted general linear models examined ln(LOS+1). Analyses were discharge-level and associative. The analysis included 9,763,989 discharges, representing approximately 48.82 million weighted discharges. COVID-19 was documented in 8.2% of weighted discharges, and in-hospital mortality was 12.9% among COVID-19 discharges compared with 2.6% among non-COVID-19 discharges. In the fully adjusted model, COVID-19 was associated with higher mortality (adjusted odds ratio [aOR], 5.456; 95% confidence interval [CI], 5.379–5.535). The association remained in sensitivity analyses restricted to principal-diagnosis COVID-19 (aOR, 3.805; 95% CI, 3.737–3.873) and in a reduced model with ordinal APR-DRG severity (aOR, 1.484; 95% CI, 1.462–1.506). Among COVID-19 admissions, bronchoscopy was associated with mortality (aOR, 6.625; 95% CI, 6.135–7.155) and longer transformed length of stay; the mortality association remained after severity adjustment (aOR, 4.362; 95% CI, 4.022–4.731). Among COVID-19 admissions with principal-diagnosis acute coronary syndrome, left heart catheterization was not significantly associated with mortality in the primary model (aOR, 0.890; 95% CI, 0.721–1.099), although a severity-adjusted sensitivity model yielded an inverse association (aOR, 0.743; 95% CI, 0.592–0.934). Documented COVID-19 was associated with substantially higher in-hospital mortality. Bronchoscopy identified a particularly high-risk, resource-intensive group. The left heart catheterization results were sensitive to model specification. These discharge-level findings do not establish procedural treatment effects and remain susceptible to residual confounding and uncertain temporal ordering.