Monocyte absolute count as a preliminary tool to distinguish between SARS-CoV-2 and influenza A/B infections in patients requiring hospitalization.

Curtolo, Ambrogio; Oliva, Alessandra; Volpicelli, Lorenzo; Ceccarelli, Giancarlo; D'Ettorre, Gabriella; Borrazzo, Cristian; Mastroianni, Claudio Maria; Venditti, Mario

Since the most frequent symptoms of novel coronavirus 2019 disease (COVID-19) are common in influenza A/B (FLU), predictive models to distinguish between COVID-19 and FLU using standardized non-specific laboratory indicators are needed. The aim of our study was to evaluate whether a recently dynamic nomogram, established in the Chinese population and based on age, lymphocyte percentage and monocyte absolute count, might apply to a different context. We collected data from 299 patients (243 with COVID-19 and 56 with FLU) at Policlinico Umberto I, Sapienza University of Rome. The nomogram included age, lymphocyte percentage and monocyte absolute count to differentiate COVID-19 from FLU. Odds ratios (ORs) and 95% confidence intervals (95% CIs) were calculated for all associations. Multivariate logistic regression models were used to adjust for potential confounding. A p-value of less than 0.05 was considered statistically significant. Patients with COVID-19 had higher age, lymphocyte percentage and monocyte absolute count than patients with FLU. Although univariate analysis confirmed that age, lymphocyte percentage and monocyte absolute count were associated with COVID-19, only at multivariate analysis was monocyte count statistically significant as a predictive factor of COVID-19. Using receiver operating characteristic (ROC) curves, we found that a monocyte count >0.35x1000/mL showed an AUC of 0.680 (sensitivity 0.992, specificity 0.368). A dynamic nomogram including age, lymphocyte percentage and monocyte absolute count cannot be applied to our context, probably due to differences in demographic characteristics between Italian and Chinese populations. However, our data showed that monocyte absolute count is highly predictive of COVID-19, suggesting its potential role above all in settings where prompt PCR nasopharyngeal testing is lacking.