View Abstract

A6047 - Phophatidylethanol (PEth) Accurately Predicts Presence of an Alcohol Use Disorder in Critically Ill Patients
Author Block: A. Majid-Moosa1, S. E. Jolley2, B. P. deBoisblanc3, L. Simon4, P. Molina4; 1Pulmonary Critical care, New Orleans, LA, United States, 2Pulmonary/Critical Care Medicine, Louisiana State University, New Orleans, LA, United States, 3Pulmonary/Critical Care Medicine, Louisiana State Univ Medical Ctr, New Orleans, LA, United States, 4Department of Physiology, New Orleans, LA, United States.
Introduction:Alcohol Use Disorders (AUD) are associated with increased risk of critical illness, greater complications, poorer outcomes, and increased healthcare utilization. AUDs are poorly recognized in critically ill patients. AUD recognition is limited by lack of an objective test. Phosphatidylethanol (PEth) is a novel serum marker validated in outpatient alcohol assessment. Our goal was to determine the diagnostic utility of PEth in critically ill patients. Methods:Intensive care unit (ICU) patients with admission diagnoses of respiratory failure, sepsis and alcohol withdrawal were included. Whole blood was obtained for PEth level measurement on admission. An Alcohol Use Disorders Identification Test (AUDIT) score was obtained from patient or a proxy for non-responsive patients. AUDIT scores ≥ 8 were considered to represent potentially hazardous drinking in men with scores ≥14 representing AUD. For Women, scores ≥5 and ≥13 represented hazardous drinking and AUD, respectively. We used logistic regression to determine whether PEth values ≥ 8 ng/mL were associated with hazardous drinking or AUD. Receiver operator curves (ROC) were generated to assess the diagnostic utility of PEth. Sensitivity and specificity were obtained at varying cut-points based on the regression model to determine the ideal predictive value of PEth. A linear regression model was generated to determine the change in quantitative PEth value for each point increase in AUDIT score. Results:A total of 13 patients were enrolled with an average age of 53 years (SD 16) and 69% were male. Median ICU length of stay (LOS) was 4 days [IQR 3-9] with a median hospital LOS of 13 days [2-17]. Most patients were admitted for diagnosis alcohol withdrawal (38%), 31% with acute respiratory failure and 31% with sepsis. Median AUDIT scores for the cohort were 21 with median serum PEth values of 84. PEth values ≥8 ng/mL perfectly predicted potentially hazardous drinking (ROC AUC 1.00) and had good predictive value for presence of an AUD (ROC AUC 0.91, 95% CI 0.64-1.00). PEth levels ≥21 ng/mL had a sensitivity of 100% for AUD with a specificity of 83.3%. Each unit increase in AUDIT score was associated with a 33.3 ng/mL increase in PEth level (95% CI , p=0.02).Conclusions:PEth accurately identified presence of hazardous drinking and AUDs in ICU patients. PEth may be a useful biomarker for alcohol assessment in the ICU and may help identify patients at high risk for worse outcomes.