Background and aims: We hypothesized that mechanically ventilated cirrhotic patients not only have poor outcomes, but also that certain clinical variables are likely to be associated with mortality. We aimed to describe the predictors of mortality in these patients.
Methods: This observational study examined 113 mechanically ventilated cirrhotic patients cared for at our institution between July 1, 2014, and February 28, 2018. We performed bivariate and multivariate analyses to identify risk factors for mortality on mechanical ventilation and created an equation to calculate probability of mortality based on these variables.
Results: Seventy percent of patients had a history of a decompensating event. Altered mental status was the most frequently encountered indication for intubation (46%). 53% patients died on mechanical ventilation. After controlling for variables associated with increased mortality, multivariate analysis revealed that vasopressor use was the strongest predictor of mortality on mechanical ventilation (OR = 9.3) followed by sepsis (OR = 4.1). A formula with an area under the curve of 0.85 was obtained in order to predict the probability of mortality for cirrhotic patients on mechanical ventilation (available at https://medweb.musc.edu/mvcp/ ). This model (AUC = 0.85) outperformed the CLIF-SOFA score (AUC = 0.68) in predicting mortality in this cohort.
Conclusion: Cirrhotic patients requiring mechanical ventilation have an extremely poor prognosis, and in patients requiring vasopressors, having a history of decompensation, sepsis or low albumin, mortality is higher. Our data points to the clinical variables should be considered in the medical management of these patients and provide physicians with a formula to predict the probability of mortality.
Keywords: Critical care; Death; Hepatic decompensation; Intensive care; Intubation; Vasopressors.