RESUMO
OBJECTIVE: Early detection of patients with COVID-19 who will need mechanical invasive ventilation (MIV) may aid in delivering proper care and optimizing the use of limited resources. METHODS: In this single-center retrospective observational study, we aimed to identify simple laboratory parameters that in combination with ferritin (a surrogate marker of severe inflammation) may help predict early (first 48 hours) MIV. A total of 160 patients with COVID-19 in whom serum ferritin, absolute lymphocyte count (ALC), platelet count, C-reactive protein (CRP), and lactate dehydrogenase (LDH) had been analyzed at admission were included. RESULTS: We found that ferritin, LDH, ALC, and CRP predicted with 88% accuracy the probability of early MIV. Results indicated that LDH showed the greater area under the curve (AUC), with a value of 89.1%. Using the AUC, we established cutoff values for clinical application. Finally, we developed a classification tree based on LDH for its clinical use. CONCLUSION: Ferritin, LDH, ALC, and CRP predict with 88% accuracy the probability of early MIV.