Nikolaj Torp, Stine Johansen, Ellen Lyngbeck Jensen, Markus Maagaard, Georg Semmler, Mette Lehmann Andersen, K. Bech, Helle Lindholm Schnefeld, Katrine Prier Lindvig, Katrine Holtz Thorhauge, Ellen Elise Petersen, Johanne Kragh Hansen, Ida Falk Villesen, Peter Andersen, Marianne Lerbæk Bergmann, Diana Julie Leeming, M.A. Karsdal, C. Rønn Hansen, Maja Thiele, Mads Israelsen, Aleksander Krag
BACKGROUND & AIMS: Alcohol is a key driver of liver-related mortality, and phosphatidylethanol (PEth) is a direct biomarker of alcohol intake. We investigated the prognostic performance of PEth with noninvasive liver fibrosis tests (NITs) for predicting hepatic decompensation in an alcohol-related liver disease (ALD) at-risk population. METHODS: Prospective cohort study with 411 at risk of ALD without prior known chronic liver disease of whom 162 had a follow-up PEth. PEth was measured from whole blood by liquid chromatography-mass spectrometry. PEth, self-reported alcohol intake, and 3 NITs: enhanced liver fibrosis (ELF) test, an algorithm incorporating PRO-C3 (ADAPT), and transient elastography (TE) were assessed at baseline. By review of medical records, participants were followed for up to 5 years for hepatic decompensation. RESULTS: Baseline PEth was 338 ng/mL (interquartile range [IQR]: 32-921 ng/mL) and median time to follow-up PEth was 26 months (IQR: 17-33). Baseline PEth was associated with decompensation (subdistribution hazard ratio [sHR] per 100 ng/mL, 1.04; 95% confidence interval [CI], 1.01-1.06), independently of NITs. Liver fibrosis NITs were the strongest individual predictors, but PEth added incremental 6-month prognostic value. Discriminatory accuracy of PEth declined with an area under the curve of 0.77 at 6 months to 0.62 at 2 years, whereas fibrosis-based NITs maintained an area under the curve of ∼0.90. Of those with a follow-up PEth, 79 (49%) had increased PEth levels compared with baseline, and 83 (51%) had stable or lowered PEth. An increased follow-up PEth was significantly associated with a higher risk of subsequent hepatic decompensation (sHR, 4.92; 95% CI, 1.09-22.34; P = .039). CONCLUSIONS: PEth predicts hepatic decompensation independently of fibrosis-based NITs in individuals at risk of ALD. Although its prognostic performance declines rapidly, repeated PEth measurements maintain prognostic discrimination.