Validation of hair ethyl glucuronide using transdermal monitoring and self‐reported alcohol use in women of childbearing potential

Abstract Aim The present study aimed to evaluate the validity of hair ethyl glucuronide concentrations compared with transdermal alcohol concentration and self‐reported alcohol use. Methods This trial included 25 adolescent and young adult females, aged 16‐24, who reported at least one heavy drinking episode (≥4 drinks) in the two weeks prior to baseline. All participants were asked to wear an alcohol biosensor over a one‐month prospective study. Detailed self‐report of drinking behaviors was assessed weekly. Estimates of blood alcohol concentration were computed from self‐report data using the National Highway and Transportation Safety Administration equation. Transdermal alcohol concentration and estimated blood alcohol concentration data were categorized into at‐risk (>0.05 g/dL alcohol) and high‐risk (>0.08 g/dL alcohol) drinking events. Hair ethyl glucuronide concentration, total number of drinking events, moderate (>0.05 g/dL) and high level (>0.08 g/dL) of transdermal alcohol concentration, and estimated blood alcohol concentration drinking events were analyzed with Spearman's rank correlation test for validity comparisons. Results No significant correlations were found between hair ethyl glucuronide values and total number, and moderate or high levels of detected drinking events by estimated blood alcohol concentration or transdermal alcohol concentration. Total number of drinking events detected and number of drinking events >0.08 g/dL using estimated blood alcohol concentration and transdermal alcohol concentration methods were significantly correlated with each other (respectively, R = .33, P < .05; R = .42, P < .05). Conclusion Our findings indicate that, due to the number of false negatives, hair ethyl glucuronide concentrations should be used with caution for monitoring abstinence from alcohol use.


| INTRODUC TI ON
Alcohol is the leading global lifestyle cause of death among 15-to 49-year-olds, 1 contributing to one in twenty deaths globally in 2016. 2 Alcohol use is particularly concerning among adolescents and young adults, who represent a significant economic investment to society and who suffer disproportionately from the consequences of alcohol use. 3 More than 4000 adolescents and young adults (under the age of 21) die each year in the United States from alcohol-related causes. 4 Many US adolescents engage in risky alcohol use, despite multiyear trends of decline. Nationally, in the past 30 days, nearly one in three (30%) reported drinking alcohol and nearly one in five (19%) reported engaging in heavy episodic drinking (4+ drinks for females or 5+ drinks for males in one setting). 3 Biological sex is also a critical consideration of risk associated with alcohol use. Rates of use have historically been highest among males; however, the gap in alcohol use between sexes is waning, particularly among young adults. 5 This trend is largely attributable to societal gender equality. 6 In the United States, females (62.6%) are more likely to engage in underage drinking than males (58.1%). 3 Rates of heavy and binge drinking among US women have increased by more than 50% over the past decade, 7 with peak episodic consumption occurring during childbearing years (age . 5 The childbearing potential of underage and young adult female drinkers further amplifies the risks among this age and gender group because these behaviors continue into early pregnancy and pregnancy is known later among women who drink. 8 Accurate measurement of intoxication from alcohol, particularly in early stages of use, is critically important to better understand the potential for intergenerational risk associated with unintended and alcohol-exposed pregnancy.
Alcohol use behaviors can be monitored by self-report, direct measurement of recent drinking events, or through specialized biological sampling. Each methodology may allow examination for specific windows of use. Participant self-reported use is a standard in the field and allows examination across multiple time frames.
Direct measurement of alcohol using blood, breath, or transdermal approaches can provide data on a current or very recent drinking event, due to the rapid metabolism of alcohol in the body. Emerging techniques in the field include sampling of metabolites of alcohol in hair or nails which can provide a retrospective assessment of consumption within a several month time frame. 9,10 Participant self-reports for measurement of alcohol use have been used for research and in clinical settings. 11 Self-reported alcohol use can be used to calculate estimated blood alcohol concentration (eBAC) when in vivo blood alcohol concentrations are inconvenient or unattainable. 12 There is some controversy associated with the use of eBAC compared with direct measurements of alcohol intoxication; however, eBAC has been shown to be highly correlated with actual blood alcohol concentration. 13 Self-report of drinking events and transdermal alcohol sensor data are also highly correlated. [14][15][16] When participants are asked to recall drinking events after an extended period of time, self-reports underestimate drinking 17 ; however, newer methods using shorter recall time spans have increased the accuracy and decreased recall bias. 18,19 The Wrist Transdermal Alcohol Sensor (WrisTAS) is an electrochemical sensor device resembling a wristwatch. 20 Ethanol detected by transdermal devices generally reflects blood alcohol levels, BAC. 21 These devices are often used for monitoring individuals in substance abuse treatment programs and intervention research among adult populations. 15,[22][23][24][25] Transdermal alcohol sensors can be used to approximate, but not directly derive, blood alcohol concentration and may be inconsistent in detection of low-intensity drinking events. 14,16,26,27 Emerging techniques allow for the indirect measurement of previous drinking events through direct metabolites including ethyl glucuronide (EtG) 28 and ethyl sulfate (EtS). EtG is a minor metabolite of alcohol and can be measured in urine, 29 hair, 30 and nails. 31 EtG accumulates in the hair shaft through perspiration. 30,[32][33][34][35] Currently, hair sampling is employed to monitor alcohol use via EtG in alcohol and drug treatment programs. 35 Hair EtG has been shown in a number of studies to be a reliable measure of chronic alcohol use. 30,32,33,[36][37][38][39] Studies have shown that hair EtG has an extremely low rate of false positives; therefore, a positive result is strongly associated with chronic consumption of alcohol. 30,37,40 Results of studies have shown the impact of cosmetic treatments, 10 hair color, 41 gender, 10,28,42,43 and level of drinking, such as heavy episodic drinking, occasional use, or chronic use. 38,39,42,44 However, the 2019 consensus from the Society of Hair Testing (SoHT) states that natural hair color and ethanol-containing hair care products do not influence EtG concentrations, 45 nor does hair washing. 46 Despite several studies which have examined validity of hair EtG in chronic heavy drinkers 32,34,36,47 and factors, such as hair color, that may affect EtG values, [37][38][39]48 there have been no studies to date that have looked at hair EtG as a method of monitoring alcohol use in adolescent and young adults with intermittent heavy drinking events. The purpose of this manuscript was to evaluate and identify valid measures of alcohol consumption among adolescent and young adult women. This manuscript presents a comparison of hair EtG, transdermal alcohol concentration (TAC), and self-reported alcohol (eBAC) use in participants, with inconsistent levels of alcohol consumption.

| Participants and criteria
Participants were recruited as part of a larger study examining alcohol effects on nutriture. Recruitment of participants largely occurred K E Y W O R D S alcohol, biosensor, ethyl glucuronide, hair EtG, transdermal through respondent-driven sampling. Potential participants were directed to the study website for initial screening. The study website contains a video explaining the study, frequently asked questions, and a link to a screening survey.
Individuals qualified for the study if they met the following criteria: female, between the ages of 14 and 24, live within a 45-minute radius of the research office, and reported at least one heavy drinking episode (≥4 standard alcoholic drinks) on one occasion within the past 2 weeks. Only female participants were recruited as this is part of a larger trial examining periconceptional behaviors in women of childbearing potential. When potential participants met the screening criteria, a member of the research team called to rescreen the potential participant and schedule a consent and baseline appointment.
Parents were required to accompany any potential participant under 18 years of age; her parents were required to consent and she to assent to her participation in the study. Anthropometric data collected at baseline included height, weight, and waist circumference. Selfreports were collected from participants on a weekly basis; self-report included day of drinking events, amounts and types of alcohol use, and length of drinking events. All participants were asked to wear a Giner WrisTAS sensor over a one-month prospective study.
In a subsample of the study population (n = 25), a hair sample was taken at the end of the four-week longitudinal study and sent for analysis. This subsample was based on participants who had at least one heavy drinking episode within the past 14 days (≥4 standard alcoholic drinks) and ability to obtain a sufficient quantity of hair sample. The National Highway and Transportation Safety Administration (NHTSA) equation was used to compute eBAC from self-report data.

| Transdermal alcohol sensor (TAS)
The Giner WrisTAS generation 7 (21 participants) and generation 9 (4 participants) were utilized to detect ethanol vapor and as a data acquisition-recording device. This device consists of three components: a sensor to detect ethanol vapor from the skin, thermistors which produce temperature signals, and a data-recording device to store data. 14 Participants wore the WrisTAS sensors for at least 21 hours per day for the study period of 28 days. Each device continuously sampled ethanol vapor detected near the skin, as well as temperature, humidity, and skin conductivity at 5-minute intervals. Members of the research staff informed the participants how to adjust the straps on the sensor and to remove the sensor while exercising or showering to prevent water exposure. Both WrisTAS generation 7 and generation 9 are water resistant, but not waterproof. Participants were instructed to wear the sensor on their ankle. While these sensors are marketed to be worn on the wrist the manufacturer's website states "you can wear it discretely on your wrist, or on just about any part of your body where blood vessels are close to the skin surface". 49 The participants were instructed they could move the sensor from one ankle to another depending on comfort. Finally, participants were instructed to avoid placing lotion or perfumes on the ankle on which they wore the sensor. Data were downloaded from the sensor weekly throughout the monthlong study. The data are stored in the sensor unit until connected via a serial port to a computer for download. Sensors were maintained weekly, including battery changes in both generation 7 and generation 9, and addition of deionized water to the generation 7 sensors.
The timing of maintenance, and subsequent quality control, resulted in the participants being fitted with a new monitor each week. Sensor data are reported as transdermal alcohol concentration (TAC), with values that correlate with blood alcohol concentration (BAC).

| Estimated blood alcohol concentrations (eBACs)
Blood alcohol concentration was estimated using the National Highway Traffic Safety Administration (NHTSA) equation with data collected as part of the study. Weights of participants were measured during the baseline visit. Each week, participants completed a retrospective report of alcohol use following a timeline follow-back (TLFB) approach. The research staff asked participants to report on which days they drank alcohol. Additionally, participants were asked to report how long each drinking event lasted as well as the number of standard drinks of beer, wine, liquor/spirits, and mixed drinks consumed (the following standard drink equivalents were used: 12-ounce beer or wine cooler, 5-ounce glass of wine, and 1.5 ounce of liquor). Participants were grouped into total number of drinking events, total number of moderate and binge drinking events >0.05 g/dL, and total number of binge drinking events >0.08 g/dL.

| Hair ethyl glucuronide (EtG)
Hair samples were collected from each participant by cutting the hair with scissors close to the scalp near the posterior vertex (crown) region of the head. The proximal end of each sample was marked. The samples were analyzed for EtG at US Drug Test Centers Analysis.
All analyses were conducted according to validated and published methods. 50

| RE SULTS
Of the 25 participants, the data included 700 days of continuous alcohol monitoring. Demographics for study population and comparison of alcohol consumption among risk groups are presented in  15). Based on self-reported alcohol use, participants fell into two risk groups: low-risk drinkers, ≤20 g alcohol/day (n = 18, 72%), and at-risk drinkers, >20 g alcohol/day (n = 7, 28%). Hair EtG results indicated individuals in every risk stratum: low-risk drinkers (n = 21, 84%), at-risk drinkers (n = 2, 8%), and high-risk drinkers (n = 2, 8%). There were no significant differences in race, ethnicity, BMI, mother's education level, or employment status across EtG and self-report risk groups (Table 1). Significant differences by alcohol use risk were found by age groups based on results from self-report, and younger participants (<20 years of age) were less likely to self-report at-risk consumption than older participants (P < .001). Based upon the risk groupings for hair EtG, this could indicate that younger participants under-report alcohol consumption and older participants may over-report alcohol use.
The total number of drinking events detected using eBAC was significantly correlated with the total number of drinking events .4239 * .7062*** .8987*** -Note: All drinking events were recorded over a 28-day period. * P < .05, **P < .01, ***P < .001 two-tailed. N = 25. Nonparametric Spearman's rank correlation test. detected using TAC methods (R = .33, P < .05). The number of drinking events >0.08 g/dL detected using eBAC was significantly correlated with the number of drinking events >0.08 g/dL detected using TAC methods (R = .42, P < .05). However, no statistically significant correlations were found between hair EtG values and the number of self-reported drinking events by eBAC or detected drinking events by TAC (Table 2). Further, no significant correlation was identified between the number of events with eBAC from low-risk (>0.05 g/dL) and at-risk (>0.08 g/dL) drinking, and the number of events with TAC from low-risk (>0.05 g/dL) and at-risk (>0.08 g/dL) and hair EtG values (Table 2).
Sensitivity, specificity, positive predictive values, and negative predictive values of hair EtG for detecting alcohol consumption are displayed in Table 3. Sensitivity of hair EtG was consistently low

| D ISCUSS I ON
This study found hair EtG tests to be a very low sensitivity assessment. Total self-reported drinking events and TAC-detected drinking events were moderately and significantly correlated; however, hair EtG did not correlate to these other alcohol detection methods.
Hair EtG testing among women has the ability to indicate alcohol exposure during pregnancy, with sampling taken in the second trimester indicative of first-trimester exposures. Hair EtG could potentially impact both clinical and legal decisions and is increasingly being implemented as standard practice in courts and treatment programs as a tool for ensuring participant compliance 40 ; moreover, among women tested during pregnancy, this may impact rates of child removal from the home. Hair EtG has previously been used for confirmatory testing of alcohol use or abstinence among heavy alcohol-using 30,32,33,[36][37][38][39] and light alcohol-using participants. 42,43 In

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

AUTH O R CO NTR I B UTI O N S
AB, AC, EC, KT, and JC have made substantial contribution to the conception, study design, data collection, data analysis, and/or interpretation of data and a contribution to manuscript writing and intellectual content of the article; and acknowledge that they have exercised due care in ensuring the integrity of the work.

A PPROVA L O F TH E R E S E A RCH PROTO CO L BY A N I N S TITUTI O N A L R E V I E WER B OA R D
The protocol for this research project has been approved by the Oklahoma State University Institutional Review Board, and a certificate of confidentiality was obtained from NICHD.

I N FO R M E D CO N S E NT
Written informed consent was obtained from all participants and/ or guardian(s) after a complete explanation of the procedures by researchers at the study facility. Participants were able to refuse study participation.

R EG I S TRY A N D TH E R EG I S TR ATI O N N U M B E R O F TH E S TU DY/ TR I A L
N/A.

A N I M A L S TU D I E S
N/A.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy restrictions.