Calprotectin serum levels on admission and during follow-up predict severity and outcome of patients with COVID-19: A prospective study.

Abstract

Calprotectin reflects neutrophil activation and is increased in various inflammatory conditions including severe COVID-19. However, serial serum calprotectin measurements in COVID-19 patients are limited. We assessed prospectively, calprotectin levels as biomarker of severity/outcome of the disease and a COVID-19 monitoring parameter in a large cohort of consecutive COVID-19 patients. Calprotectin serum levels were measured in 736 patients (58.2 % males; median age 63-years; moderate disease, n = 292; severe, n = 444, intubated and/or died, n = 50). Patients were treated with combined immunotherapies according to our published local algorithm. The endpoint was the composite event of intubation due to severe respiratory failure (SRF)/COVID-19-related mortality. Median (interquartile range) calprotectin levels were significantly higher in patients with severe disease [7(8.2) vs. 6.1(8.1)μg/mL, p = 0.015]. Calprotectin on admission was the only independent risk factor for intubation/death (HR=1.473, 95 %CI=1.003-2.165, p = 0.048) even after adjustment for age, sex, body mass index, comorbidities, neutrophils, lymphocytes, neutrophil to lymphocytes ratio, ferritin, and CRP. The area under the curve (AUC, 95 %CI) of calprotectin for prediction of intubation/death was 0.619 (0.531-0.708), with an optimal cut-off at 13 μg/mL (sensitivity: 44 %, specificity: 79 %, positive and negative predictive values: 13 % and 95 %, respectively). For intubated/died patients, paired comparisons from baseline to middle of hospitalization and subsequently to intubation/death showed significant increase of calprotectin (p = 0.009 and p < 0.001, respectively). Calprotectin alteration had the higher predictive ability for intubation/death [AUC (95 %CI):0.803 (0.664-0.943), p < 0.001]. Calprotectin levels on admission and their subsequent dynamic alterations could serve as indicator of COVID-19 severity and predict the occurrence of SRF and mortality.