Validation of blood phosphatidylethanol as an alcohol consumption biomarker in patients with alcohol use disorder and liver disease at a liver transplant center

Background

Phosphatidylethanol (PEth), ethyl glucuronide (EtG), and ethyl sulphate (EtS) are highly sensitive and specific biomarkers of alcohol intake. This study investigated their application and relationship to traditional self-report measures in a mixed cohort of liver disease patients to guide decision making in liver transplant populations.

Methods

We recruited 183 participants (mean age 49.2 years, 62% male), with N = 99 liver disease (88% alcohol-associated liver disease [ALD]), N = 35 alcohol use disorder (AUD), and N = 49 healthy volunteers. Patient-reported alcohol intake and AUDIT score served as references and were compared to traditional biomarkers, PEth and serum EtG/EtS. Receiver operating characteristic (ROC) analysis and a range of biomarker cutoffs were examined to determine optimal test characteristics. A subset of blood samples modified to a standardized hematocrit analyzed the relationship between hematocrit and PEth.

Results

Compared to traditional biomarkers, both PEth and EtG were sensitive and specific for alcohol intake. At the limit of detection (LOD), PEth was 95% sensitive at detecting any drinking. PEth cutoff of 300 μg/L was 86% sensitive and 92% specific for “heavy drinking,” and 600 μg/L was 88% sensitive and specific for “very heavy drinking.” PEth displayed superior test characteristics (sensitivity, specificity, PPV, NPV, and AUC) to all measured traditional biomarkers over two-day and one-month time frames. A subset of participants suspected of drinking but reporting abstinence had positive PEth tests (35%), suggestive of unreported drinking. PEth was positively correlated with hematocrit (r² = 0.83, p < 0.01) and correction to a standardized median resulted in increases in PEth concentration in most cases.

Conclusions

PEth is clinically useful as an alcohol biomarker in patients with liver disease and is superior to traditional biomarkers, providing good test characteristics for “heavy” and “very heavy” drinking using stepwise cutoffs. PEth detected a subset of patients underreporting their alcohol use, with implications for the management of patients in liver transplant clinics.