Prognostic predictors for mortality of patients with COVID-19 in an intensive care unit

Abstract

Introduction: Fatality due to COVID-19 continues to be a challenge. Timely identification of critical COVID-19 patients is crucial for their close clinical follow-up and treatment. We aimed to identify the mortality predictors of critical COVID-19 patients.

Methodology: We analyzed medical records of 232 out of 300 patients with COVID-19 hospitalized in the intensive care unit (ICU) whose medical records were available in the hospital database. Non-survivors and survivors were compared for parameters. Medical records of demographics, comorbidities, radiological signs, respiratory support, and laboratory tests on the first day of ICU admission were included. The durations of ICU stay and hospitalization were also evaluated.

Results: The patients with Acute Physiology and Chronic Health Evaluation II (APACHE-II) score above 28.5 and the patients with blood urea nitrogen (BUN) above 45.5 mg/dL were significantly more mortal (95% CI: 0.701, p = 0.0001; 95% CI: 0.599, p = 0.022; respectively). Partial oxygen pressure/fraction of inspired oxygen (P/F) ratio below 110.5 mmHg was a predictor for mortality (95% CI: 0.397, p = 0.018). Older age, smoking, crazy paving pattern on computed tomography (CT), and short duration of hospitalization were also predictors of mortality. The patients requiring invasive mechanical ventilation were significantly more mortal whereas the patients requiring high flow oxygen and non-invasive ventilation were significantly more likely to survive.

Conclusions: We recommend evaluating APACHE-II score, BUN value, P/F ratio, age, smoking status, radiological signs on CT, length of hospitalization and modality of respiratory support upon ICU admission to identify critical patients with poor prognoses.