Parâmetros hematológicos como fatores de predição precoce do desfecho da COVID-19

Abstract

Novel antiviral COVID-19 medications, such as Paxlovid and Molnupiravir, are most effective if administered within the first five days of symptoms, helping patients with a high risk of developing the severe form of COVID-19 to avoid hospitalization. Hence, the sooner patients who may benefit from these drugs are identified, the better their chances of survival. Although the SARS-CoV-2 infection already has some established risk groups, such as men, elderly people, and people with comorbidities, it is still necessary to find biomarkers of disease outcomes to stratify patients' risk and improve clinical management. The complete blood count (CBC) can be a viable, efficient, and affordable option to find biomarkers able to predict the prognosis of the disease, since this viral infection can alter various blood parameters. Therefore, the goal of the present study was to identify a possible association between hematological parameters and different clinical forms of COVID-19 and use them as predictors of disease outcomes. We performed a CBC in blood samples from 297 individuals treated in Belo Horizonte, Brazil, and observed that in the first 4 days of symptoms the classic hematological COVID-19 alterations, such as lymphopenia, are not yet perceptible. However, the percentage of monocytes (MON%) and the granulocyte-to-lymphocyte ratio (GLR) were already reduced and increased, respectively, during this period in patients who presented mild symptoms and who later progressed to the severe form of the disease and required hospitalization. We also observed that during this period, the platelet-to-lymphocyte ratio (PLR) was able to differentiate COVID-19 patients from individuals with flu-like symptoms or negative controls. We performed ROC and TG-ROC Curves, as well as Decision Tree algorithms, to evaluate the accuracy of these hematological predictors and to establish their cutoff values. Thus, our findings demonstrated that patients with COVID-19 with MON% values lower than 7.7% and GLR values greater than 8.75 are assigned to the hospitalized group with a precision of 86%. This suggests that these variables can be important biomarkers in predicting disease outcomes and could be used to discriminate patients at hospital admission and manage therapeutic approaches, such as antivirals, in the beginning of infection.