Lactate to Albumin Ratio and Mortality in Patients with Severe Coronavirus Disease-2019 Admitted to an Intensive Care Unit

<b>Aim:</b> This study sought to evaluate the effectiveness of lactate/albumin ratio for ICU mortality prediction in a large cohort of patients with severe Coronavirus Disease-2019 (COVID-19) admitted to an intensive care unit (ICU). <b>Methods:</b> This is a single-center retrospective cohort study of prospectively collected data derived from the COVID-19 dataset for all critically ill patients admitted to an academic ICU. Data were used to determine the relation between lactate/albumin ratio and other laboratory parameters measured on the first day of the ICU stay and to evaluate the prognostic performance for ICU mortality prediction. <b>Results:</b> A total of 805 ICU patients were included, and the median age (IQR) was 67 (57-76) years, with 68% being male. ICU mortality was 48%, and the median lactate/albumin ratio was 0.53 (0.39-0.59). A survival analysis showed that patients with higher lactate/albumin ratio values had significantly lower survival rates (Log Rank <i>p</i> < 0.001). A multivariable analysis revealed that the lactate/albumin ratio was an independent risk factor for ICU mortality with a hazard ratio of 1.39 (CI: 1.27-1.52). The lactate/albumin ratio showed a receiver operating characteristics area under the curve (ROC-AUC) value to predict ICU mortality significantly higher than that of lactate alone (0.71 vs. 0.68, DeLong test <i>p</i> < 0.001). The optimal lactate/albumin ratio cut-off for predicting ICU mortality was 0.57, with 63% sensitivity and 73% specificity. A subgroup analysis revealed that the lactate/albumin ratio was significantly associated with mortality across different patient groups, including age and sex categories, and those with or without hypertension and coronary heart disease. <b>Conclusions</b>: Lactate/albumin ratio is a reliable prognostic marker in critically ill COVID-19 patients and could predict ICU mortality more accurately than lactate alone.