ABSTRACT
Background: Mechanical ventilation of patients with COVID-19 is associated with high mortality. Understanding risk factors for developing mechanical ventilation may allow for more targeted monitoring and therapeutics that may improve outcomes. Methods: We performed a retrospective case series of all patients admitted within thirty days of a positive Sars CoV-2 test to an integrated health system near Chicago, Illinois between March 12 and May 31, 2020. Covariates evaluated included demographics, symptoms on admission, vital signs, medications, comorbidities, census tract data and social history. Univariable analysis was performed and variables with an alpha of 0.05 or less were included in multivariable regression modeling to identify factors associated with mechanical ventilation. The area under the curve (AUC) was used to assess performance of the model. Results: Of 990 patients admitted with COVID-19 12.6% (125) were mechanically ventilated. The median age was 68 (interquartile range 55-82), 48.4% (479) were female and 49.6% (491) were Caucasian. Independent factors associated with mechanical ventilation included female sex (Adjusted OR [AOR] 0.621, Confidence Interval [CI] 0.427-0.903;p=0.0363), body mass index (BMI) (AOR 1.035, CI 1.011-1.060;p=0.0175), percent of english speaking population within patient's census tract (AOR 0.989, CI 0.979-0.998;p=0.0454), respiratory rate (AOR 1.054, CI 1.027-1.083;p=0.0011), oxygen saturation (AOR 0.922, CI 0.901-0.943;p< 0.0001), cerebrovascular accident (CVA) (AOR 0.176, CI 0.051-0.605;p=0.0207) and hematologic malignancy (AOR 3.668, CI 1.403-9.590;p=0.0261). AUC of the model was 0.8 (0.75-0.84). Conclusion: Risk factors associated with mechanical ventilation included male gender, elevated BMI, census tract with lower percentage of english speakers, increased respiratory rate, low oxygen saturation, hematologic malignancy and not having a CVA. We suspect that history of CVA may have been associated with overall patient debility in which aggressive measures such as intubation were not deemed appropriate. Identifying patients with risk factors associated with mechanical ventilation may allow for early and targeted interventions to improve outcomes. (Table Presented).