Importance: A male predominance is reported in hospitalised patients with COVID-19 alongside a higher mortality rate in men compared to women.
Objective: To assess if the reported sex bias in the COVID-19 pandemic is validated by analysis of a subset of patients with severe disease.
Design: A nationwide retrospective cohort study was performed using the Austrian National COVID Database. We performed a sex-specific Lasso regression to select the covariates best explaining the outcomes of mechanical ventilation and death using variables known before ICU admission. We use logistic regression to construct a sex-specific "risk score" for the outcomes using these variables.
Setting: We studied the characteristics and outcomes of patients admitted to intensive care units (ICUs) in Austria.
Participants: 5118 patients admitted to the ICU in Austria with a COVID-19 diagnosis in 03/2020-03/2021.
Exposures: Demographic and clinical characteristics, vital signs and laboratory tests, comorbidities, and management of patients admitted to ICUs were analysed for possible sex differences.
Main outcomes and measures: The aim was to define risk scores for mechanical ventilation and mortality for each sex to provide better sex-sensitive management and outcomes in the future.
Results: We found balanced accuracies between 55% and 65% to predict the outcomes. Regarding outcome death, we found that the risk score for pre-ICU variables increases with age, renal insufficiency (f: OR 1.7(2), m: 1.9(2)) and decreases with observance as admission cause (f: OR 0.33(5), m: 0.36(5)). Additionally, the risk score for females also includes respiratory insufficiency (OR 2.4(4)) while heart failure for males only (OR 1.5(1)).
Conclusions and relevance: Better knowledge of how sex influences COVID-19 outcomes at ICUs will have important implications for the ongoing pandemic's clinical care and management strategies. Identifying sex-specific features in individuals with COVID-19 and fatal consequences might inform preventive strategies and public health services.
Keywords: COVID-19; ICU; diabetes; gender; mortality; sex.