Although many demographic and clinical predictors of mortality have been studied in relation to COVID-19, little has been reported about the prognostic utility of inflammatory biomarkers.
Retrospective cohort study. All patients with laboratory-confirmed COVID-19 treated in a hospital emergency department were included consecutively if baseline measurements of the following biomarkers were on record: lymphocyte counts, neutrophil-to-lymphocyte ratio NRL, and C-reactive protein (CRP) and procalcitonin (PCT) levels. We analyzed associations between the biomarkers and all-cause 30-day mortality using Cox regression models and dose-response curves.
We included 896 patients, 151 (17%) of whom died within 30 days. The median (interquartile range) age was 63 (51-78) years, and 494 (55%) were men. NLR, CRP and PCT levels at ED presentation were higher, while lymphocyte counts were lower, in patients who died compared to those who survived (P .001). The areas under the receiver operating characteristic curves revealed the PCT concentration (0.79; 95% CI, 0.75-0.83) to be a better predictor of 30-day mortality than the lymphocyte count (0.70; 95% CI, 0.65-0.74; P .001), the NLR (0.74; 95% CI, 0.69-0.78; P = .03), or the CRP level (0.72; 95% CI, 0.68-0.76; P .001). The proposed PCT concentration decision points for use in emergency department case management were 0.06 ng/L (negative) and 0.72 ng/L (positive). These cutoffs helped classify risk in 357 patients (40%). Multivariable analysis demonstrated that the PCT concentration had the strongest association with mortality.
PCT concentration in the emergency department predicts all-cause 30-day mortality in patients with COVID-19 better than other inflammatory biomarkers.