Association between hypertension and risk of respiratory failure and mortality in patients with COVID-19: a retrospective cohort study with predictive model development

Background: Hypertension is highly prevalent among patients with Coronavirus Disease 2019 (COVID-19), yet its impact on respiratory failure and mortality remains controversial. Our objective was to evaluate the association between hypertension and adverse outcomes in COVID-19 patients and to develop predictive models for risk stratification. Objective: To investigate the association between hypertension and respiratory failure or mortality in patients with COVID-19: mediation analysis of laboratory biomarkers and development of a risk prediction model for adverse outcomes in hypertensive individuals with COVID-19. Methods: We retrospectively analyzed clinical data from hospitalized COVID-19 patients. After 1:1 propensity score matching (PSM) of hypertensive and non-hypertensive patients, multivariable logistic regression assessed hypertension's independent association with respiratory failure and in-hospital mortality. Mediation and subgroup analyses explored biomarker mediators and effect modifiers. Least Absolute Shrinkage and Selection Operator (A LASSO)-selected risk prediction model was validated by Area Under the ROC Curve (ROC-AUC) analysis. Results: = 0.004). Mediation analysis revealed that eosinophil count and Estimated Glomerular Filtration Rate (eGFR) partially mediated these associations: for respiratory failure, they accounted for 44.5 and 15.6% of the total effect, respectively; for mortality, 50.0 and 16.7%. LASSO-based prediction models demonstrated an AUC of 0.788 for respiratory failure (using neutrophil count, C-Reactive Protein (CRP), Blood Urea Nitrogen (BUN), and albumin) and 0.694 for mortality (using neutrophil count and BUN). Conclusion: Compared with non-hypertensive COVID-19 patients, those with hypertension had significantly higher risks of respiratory failure and in-hospital mortality. Eosinophil count and eGFR significantly mediated these associations. A multi-marker model-combining neutrophil count, CRP, BUN, and albumin-showed good predictive performance for respiratory failure, supporting early risk identification and precision clinical decision-making.