A Nomogram for Predicting Hospital Mortality in Patients with COVID-19 Admitted to the Intensive Care Unit

The objective: to develop a predictive model for assessing the risk of death in patients with COVID-19 admitted to the intensive care unit (ICU). Subjects and Methods. This was a single-center retrospective cohort study of hospital mortality in patients admitted to ICU of V.P. Demikhov City Clinical Hospital from March 6 to June 3, 2020. Results. Case histories of 403 patients were analyzed. In-hospital 30-day mortality among patients treated in ICU was 44.9% (181/403 patients). A multivariate analysis showed that significant predictors of death in patients with COVID-19 were the age of 60 years or more (adj. OR 3.84; 95% CI 1.56‒9.44, p = 0.003), COPD (adj. OR 2.35; 95% CI 1.12‒4.95, p = 0.024), peripheral artery diseases (adj. OR: 5.08; 95% CI 1.87‒13.76, p = 0.001) and CKD stage 3b and higher (adj. OR 4.58; 95% CI 2.36‒8.90, p < 0.001), LDH 300 IU/l or more (adj. OR 3.05; 95% CI 1.23‒7.58, p = 0.016), and CRP 200 mg/l or more (adj. OR 3.65; 95% CI 1.95‒6.85, p < 0.001). Predictive model quality: AUC = 0.811 [0.733‒0.874], p < 0.001. Conclusions. A nomogram to assess the risk of death in patients with COVID-19 has been developed. It includes the assessment of age, history of COPD, CKD stage 3b and higher, peripheral artery diseases, and monitoring of LDH and CRP.