Abstract
AIMS: To evaluate the association between self-reported alcohol consumption and phosphatidylethanol (PEth) concentrations in blood in a large general population study, and discuss optimal cut-off PEth concentrations for defined levels of alcohol consumption. METHODS: Population based, longitudinal cohort study including 24,574 adults from The Trøndelag Health Study 4 (HUNT4) conducted in Trøndelag County, Norway. Data included PEth concentration, self-reported alcohol consumption and CAGE score. RESULTS: PEth levels in whole blood increased with the number of alcohol units consumed, the frequency of alcohol consumption, the frequency of binge drinking and the CAGE score (lifetime, i.e. 'have you ever'). The cut-off concentrations with highest combined sensitivity and specificity were 0.057 μmol/l (40 ng/ml) for identification of those consuming >1 alcohol unit per day (sensitivity 86%, specificity 76%), 0.087 μmol/l (61 ng/ml) for consuming >2 units per day (sensitivity 87%, specificity 81%) and 0.122 μmol/l (86 ng/ml) for consuming >3 alcohol units per day (sensitivity 80%, specificity 86%). By defining a CAGE score ≥ 2 as potentially harmful consumption, a cut-off of 0.100 μmol/l (70 ng/ml) identified 52% of all those subjects. CONCLUSIONS: Cut-off limits of PEth concentrations should take into account the indication for sampling. Using cut-offs for the PEth concentrations of about 0.05 μmol/l (35 ng/ml) and 0.08 μmol/l (56 ng/ml) would identify about 90% of the subjects consuming more than 1 and 2 alcohol units per day, respectively. Concentrations above these cut-offs should lead to a more detailed interview related to alcohol use.