Phosphatidylethanol vs Transdermal Alcohol Monitoring for Detecting Alcohol Consumption Among Adults

IMPORTANCE: Alcohol biomarkers can improve detection of heavy alcohol use in clinical care, yet cutoffs for phosphatidylethanol (PEth), a blood biomarker, have not been established. OBJECTIVE: To determine the optimal cutoff for PEth for heavy alcohol consumption in a study of middle-age and older adults. DESIGN, SETTING, AND PARTICIPANTS: This was a 4-week diagnostic study of adults with paroxysmal atrial fibrillation (AF) and current alcohol consumption, recruited from general cardiology and cardiac electrophysiology outpatient clinics from September 2014 to September 2019. Data were analyzed from October 2021 to March 2022. MAIN OUTCOMES AND MEASURES: The main aim was to determine the optimal PEth cutoff for heavy alcohol consumption, using the Secure Continuous Remote Alcohol Monitor (SCRAM) to measure transdermal alcohol. Area under the curve (AUC) for PEth-detected compared with SCRAM-detected heavy alcohol consumption in any week over the prior 4 weeks (ie, ≥3 [women] and ≥4 [men] episodes) or any estimated breath alcohol of 0.08% or greater in any week, and the PEth cutoff was calculated using the Youden J statistic. Similar analyses were conducted comparing PEth with individual drinks reported by pressing an event monitor, retrospective self-report via the Alcohol Use Disorders Identification Test–Consumption (AUDIT-C), and using 2-week look-backs. RESULTS: In this diagnostic study of 64 patients with both PEth and SCRAM measures over 4 weeks (54 [84.4%] men; mean age, 65.5 [95% CI, 62.6-68.5] years; 51 [79.7%] White), 31 (48.4%) had any SCRAM-detected heavy alcohol consumption over the 4 weeks, and the median (IQR) PEth at 4 weeks was 23 ng/mL (<limit of quantification to 60 ng/mL). The AUC for PEth vs any SCRAM-detected heavy alcohol consumption was 0.83 (95% CI, 0.72-0.93). The optimal PEth cutoff was 18.5 ng/mL (AUC, 0.83; 95% CI, 0.72-0.93), with sensitivity of 83.9% (95% CI, 66.3%-94.5%) and specificity of 72.7% (95% CI, 54.5%-86.7%). The PEth test characteristics by individual drink reporting using the event monitors and by the AUDIT-C, and by these measures and by SCRAM collected for 2-week intervals, were similar to those compared with the 4-week SCRAM. CONCLUSIONS AND RELEVANCE: In a predominately middle-age and older White male population, PEth compared well with SCRAM. A PEth cutoff of 18.5 ng/mL (or rounded to 20 ng/mL, a recommended PEth cutoff for significant alcohol consumption) can be used in clinical care to detect heavy alcohol consumption in middle-age and older men.