ABSTRACT
Context and objective In intensive care units (ICU), clinicians have little information to identify COVID-19 patients at high risk of poor prognosis requiring intubation. Considering the clinical and biological parameters of the patients during their admission to ICU, we determined the incidence of a pejorative evolution requiring intubation, and secondarily we searched among the starting parameters, which were predictors of the intubation during follow-up Methods We conducted a monocentric retrospective cohort study of adult patients admitted for moderate, severe or critical COVID-19/WHO clinical classification, during the first two waves of the pandemic in Kinshasa/DR Congo. Our primary end point was the incidence of intubation. Potential predictors of intubation were determined by the Cox regression analysis. The relative risk of death was assessed according to treatment with mechanical ventilation. (intubation) Results We included 219 patients (average age of 56.8 ±15.2 years; 75 % men), respectively 37 % in the 1st and 63 % in the 2nd wave of the pandemic.Cumulative incidence of intubation was 24% (1stwave: 26% vs 2nd wave: 24%). One-third of intubations were performed on the first 3 days versus two-thirds beyond the third day. The Cox's regression model showed that among data from the 1st day of ICU admission, those predicting intubation were: age (Hazard ratio: 1.025, CI 95%: 1.005-1.044), obesity (HR: 4.808; CI 95%: 2.660-8.696), corticosteroid therapy (HR: 0.313, CI 95%: 0.102-0.965), ROX index < 4.88 (HR: 2.024, CI 95 %: 1.003-4.080) and black race (HR: 0.502, CI 95%: 0.272-0.928). In total, 54 deaths (25 % of patients) were recorded with a higher relative risk (18.8) in intubated patients. Conclusion A quarter of COVID-19 patients admitted to ICU could worsen and be intubated. The majority of intubations were performed after the third day of admission and mortality was high. The predictors of intubation that have been identified can help anticipate management by being proactive