C-reactive protein on admission as an early predictor of covid-19 severity in a reference health center in Peru

Abstract

Objective: To determine a cut-off point and the operational characteristics of the C-reactive protein in the severity of COVID-19 infection in adult patients treated at the Corazón de Lurín Medical Center during the years 2020 to 2021.
Materials and methods: Observational, descriptive, retrospective, cross-sectional, diagnostic test type design, we worked with a sample size of 502 patients with SARS COV-2. A binary logistic regression model was used to obtain the operational characteristics of the C-reactive protein for the diagnosis of Severe SARS COV-2, as well as a ROC curve to obtain the area under the curve (AUC) with their respective intervals of confidence and optimal cut-off point. Likewise, the Fagan normogram was used to represent the pre-test and post-test probability.
Results: An AUC of 79.5%, an overall precision of 77.49%, a sensitivity of 80.66%, a specificity of 75.10%, a positive predictive value of 65.18%, and a negative predictive value of 87.41% were obtained. Values of 36.06%, 65% and 13% were obtained for the pre-test probability, negative post-test probability and positive post-test, respectively. In addition to this, the maximum positive (3.32) and negative (0.26) likelihood ratios were calculated.
Conclusions: It can be stated that C-reactive protein on admission is an adequate early marker for the prognosis of severity due to COVID-19 disease.
DeCS / MeSH: COVID-19; Clinical Study; C-Reactive Protein; ROC curve; Prognosis; Severity of Illness Index