Associations between clinical AUDIT-C screens and HDL cholesterol are observed across primary care patient subgroups.

backgroundThe Alcohol Use Disorders Identification Test-Consumption (AUDIT-C) is a validated, scaled marker of past-year alcohol consumption that is increasingly used in population-based screening and research. Like other screening questionnaires, AUDIT-C scores are influenced by patient and system factors affecting self-report. High-density lipoprotein (HDL) cholesterol increases with alcohol consumption and is routinely measured in primary care. Researchers using AUDIT-C scores as an outcome could potentially use HDL as a population-level check on the performance of alcohol screening, for example, to assess the extent to which changes in AUDIT-C scores after an intervention reflect changes in drinking or changes in self-report. However, the association between AUDIT-C scores and HDL has only been evaluated in limited populations.

methodsCross-sectional associations between AUDIT-C scores and HDL were examined in 290,091 Kaiser Permanente Washington primary care patients who had HDL measured as part of clinical care in the 365 days before or 14 days after routine screening with the AUDIT-C. Linear regression models examined the association between AUDIT-C scores and HDL and explored effect modification by sociodemographic and clinical characteristics.

resultsAUDIT-C scores were positively associated with HDL, including for subgroups defined by age, sex, race, ethnicity, geographically estimated socioeconomic status, presence of cardiovascular disease, history of alcohol or drug treatment, tobacco use, receipt of lipid-lowering medications, and, for female patients, receipt of oral estrogen or progestin medications. Effect modification analyses showed that most sociodemographic and clinical characteristics modified the association between AUDIT-C and HDL.

conclusionsThe association between AUDIT-C and HDL is present in a range of sociodemographic and clinical subgroups. However, effect modification by sociodemographic and clinical characteristics may limit the use of that association in assessing the validity of alcohol screening scores across heterogeneous populations.