Fig. 2

Cox proportional hazard model. CEA is the only laboratory indicator with significant results in all univariate and multivariate analyses. To identify the optimal cutoff point of CEA, the cyclic log-rank test was performed. And the results showed that CEA = 7.3 ng/ml was the optimal cutoff point with the most significant P value in log-rank test (A, B). Then, 12 potential significant indicators (showing prognostic values in Kaplan–Meier analysis) and two demographic information (age and gender) were incorporated into the initial Cox proportional hazard models, and the final multivariate models were constructed to confirm the effects of significant covariates in the initial models to the OS of COVID-19 patients (C). The variable of nasal catheter is integrated into a new variable named "Mode of ventilation," a variable with five levels (no oxygen; oxygen with nasal canula (OWNC); oxygen through high flow nasal canula (HFNC); noninvasive ventilation (NIV); and invasive ventilation (IV)). The results suggested that patients with lower CEA had better OS (HR 0.57; 95% CI 0.354 to 0.920; P = 0.020) in multivariate model, which suggested that CEA independently prognostic indicator for COVID-19 patients