Do Subjective Alcohol Screening Tools Correlate with Biomarkers among High-Risk Transgender Women and Men who have Sex with Men in Lima, Peru?

Abstract

Alcohol abuse can influence sexual risk behavior; however, its measurement is not straightforward.

This study compared self-reported alcohol use, via the AUDIT and CAGE, with levels of phosphatidylethanol (Peth), a phospholipid biomarker that forms with chronic, heavy drinking, among high-risk MSM and TW in Lima, Peru. Chi-square, Fisher’s exact, Wilcoxon ranksum tests compared the instruments. Receiver operating curves determined sensitivity and specificity of the self-reported measures.

Among 69 MSM and 17 TW, PEth was positive for 86% (95% CI 77% – 93%) of participants, while 67% reported binge-drinking in the last 2 weeks. The AUDIT classified 25% as hazardous drinkers while CAGE identified 6% as problem drinkers. Self-reported binge drinking was more sensitive than the AUDIT for PEth positivity (71% vs. 27%, p=0.022).

Among high-risk MSM and TW in Lima, validated, self-report measures of alcohol abuse underestimated biological measures. Further research correlating bio-markers and self-reported alcohol abuse measures is needed.

Introduction

The association between unhealthy alcohol use and HIV/STIs has been demonstrated internationally among diverse populations.^1,2,3,4 Studies in sub-Saharan Africa have demonstrated associations between heavy episodic alcohol use with unsafe sex and HIV transmission.⁵ In Peru, key populations of men who have sex with men (MSM) and transgender women (TW) face the concurrent problems of heavy alcohol use and a concentrated HIV epidemic.⁶ The prevalence of alcohol use disorders among MSM and TW in Peru has been reported to be as high as 55–63%.^{7,8, 9} In addition to potentially impacting HIV acquisition through an influence on high-risk behaviors,^{10, 11, 12, 13} alcohol use disorders (AUDs) can threaten the ability of the World Health Organization’s “treatment as prevention” strategy¹⁴ to effectively curb the HIV epidemic, most commonly by decreasing adherence to antiretroviral therapy (ART).¹⁵ In Peru, AUDs have been found to significantly reduce anti-retroviral therapy adherence among HIV-infected MSM ¹⁶, and globally, alcohol use is associated with worse outcomes for HIV-positive patients.¹⁷ At the virologic level, alcohol’s effect on the human immune system may exacerbate HIV progression.^{18, 19, 20, 21}

While it is necessary to capture accurate alcohol use data for individual and population-level interventions, the measurement of alcohol abuse remains fraught with challenges. Computerized alcohol screening tools have already shown higher detection of at-risk drinking when compared to direct questioning in 1-on-1 interviews since interviewer-administered surveys frequently elicit social desirability bias. ²² In Peru for instance, there may be an incentive among already stigmatized sexual and gender minority groups to minimize alcohol use due to a perceived risk that ART, which is offered universally by the Peruvian government, may be denied if a substance use disorder is suspected. The ability to trust subjective reports of alcohol use is of particular concern in the Peruvian context, where qualitative work has uncovered the tendency to conceptualize problem alcohol use as merely social alcohol use.²³ Also, use of the Alcohol Use Disorders Identification Tool (AUDIT) is a challenge in Peru where the sharing of common containers is a common practice, which may complicate the quantification of standard drink sizes. This could pose a problem if standard drink sizes are used to derive the level of alcohol consumption in the AUDIT. Karno et al has suggested that how one quantifies their alcohol use can cause substantial variance in AUDIT scores which contrasts with the use of the AUDIT as a one-factor screening tool with a single cutoff score.²⁴ Additionally, there remains a dearth of research on the psychometric properties of the AUDIT outside English-speaking countries.²⁵

For both reasons, it is imperative that we corroborate subjective reports of alcohol use with objective measures, such as the phosphatidylethanol (PEth) biomarker, an abnormal cellular membrane phospholipid formed only in the presence of alcohol (see Figure 1). This direct biomarker can be used to detect chronic, excessive exposure to alcohol because human red blood cells lack the ability to efficiently degrade this byproduct which remains detectable in whole blood for up to 4 weeks after sobriety.²⁶ The assay itself is measurable as a concentration in venous blood and can be analyzed as a continuous variable. Globally, PEth has been used in HIV-infected populations to demonstrate the reality of the under-reporting of alcohol use.^27,28 In a meta-analysis of clinical trials that used PEth quantification in blood, cut-off levels for PEth yielded diagnostic sensitivities ranging from 98–100%.²⁹ When evaluating clinical specificity of the biomarker in its ability to distinguish alcohol-dependent subjects from social drinkers and/or abstainers, the available controlled trials have all obtained a 100% value.³⁰ To our knowledge PEth testing has not been utilized in Peru to date.

Figure 1.

Direct Alcohol Biomarker Testing: Phosphatidylethanol (PEth)

Image credit: United States Drug Testing Laboratories, Inc.

The aim of this pilot study was to use PEth as the gold standard to calculate the sensitivity and specificity of the AUDIT, AUDIT-Consumption (AUDIT-C: first three questions of full AUDIT) and CAGE (four-question screening tool). We therefore conducted this study to make such extensive comparisons across alcohol screening tools and then attempt to examine the correlates of those measures among a high-risk population.

Materials and Methods

Study Design and Population

Participants in this study were part of the longitudinal Picasso cohort of 401 high-risk MSM and TW from 35 of Lima’s 49 districts.³¹ The cohort was recruited from 2013 – 2014 with the intention of targeting high-risk individuals who could benefit from HIV prevention services. The study consists of 8 follow-up visits occurring every 3 months from 2014–2016 wherein participants respond to an interviewer-administered survey in Spanish and provide biologic specimens for syphilis, HIV, and rectal gonorrhea/chlamydia testing. Confidentiality of participation and interview responses was assured and emphasized to subjects at each visit. Those who were at least 18 years of age and fulfilled at least 3 of the following inclusion criteria were eligible for enrollment in the Picasso study: (i) sexually active for more than 5 years, (ii) a positive syphilis test in the last 2 years, (iii) a positive HIV test, (iv) more than 5 sexual partners in last 3 months, (v) STI diagnosis in last 6 months, (vi) current STI symptoms, or (vii) more than 5 episodes of condomless anal intercourse in the last 6 months. While the aforementioned criteria qualified a participant for their baseline visit, all the results in this manuscript are for the sub-study population of 86 participants who had PEth levels measured at one of their various follow-up visits after (and therefore excluding) their baseline visit. Baseline data is published separately.³²

Alcohol Consumption Measures: PEth as the Gold Standard

As part of the routine survey administered at each follow-up visit, participants completed the AUDIT,³³ AUDIT-C³⁴ and CAGE³⁵ questionnaires. An AUDIT score ≥8 was used to determine the presence of an alcohol use disorder,³⁶ while the AUDIT-C and CAGE questionnaire served as ancillary screens for hazardous alcohol use. Participants were also asked specifically about the frequency of binges in the last two weeks, which was defined as consuming 5 or more drinks on the same occasion in accordance with the definition put forth by the National Institute on Alcohol Abuse and Alcoholism (NIAAA).³⁷ In addition, participants were asked to characterize substance use as it related to their last sexual encounter, i.e. if any alcohol or drug was consumed at last sex and to what extent (was the participant subjectively intoxicated at the last sexual encounter?).

During one of the Picasso study follow-up visits, we assessed for the presence of the PEth biomarker by collecting dried blood spots from a sub-group of 86 participants who were offered participation in the PEth pilot study. We also collected two samples to be used as negative controls from known alcohol abstainers who worked in the university system at Cayetano Heredia. Knowing that no false positive results have been observed in clinical trials to date using the high-performance liquid chromatography method of processing,³⁸ we collected samples from negative controls to ensure our methodology was equally reliable. Healthcare workers collecting the samples were explicitly instructed not to use alcohol wipes for skin sterilization prior to drawing blood. Samples were stored in −80 C° freezers prior to being shipped for processing by Agilent 6460 liquid chromatography-tandem mass spectrometry (LC-MS/MS) to the United States Drug Testing Laboratory (USTDL) in Des Plaines, Illinois. The current manufacturer-recommended limit of quantitation (8 ng/mL) was used to determine PEth-positivity. Additional analyses were performed at a cutoff of 20 ng/mL as several studies have proposed a higher cutoff to better approximate heavy alcohol use.^{39, 40,41,42}

Independent Variables

Last, we administered a comprehensive survey assessing domains such as age, gender identity, education level, and socioeconomic status. The rationale for including these data in this analysis was to assess for correlates of unhealthy alcohol use. Given the confluence of factors that underlie both alcohol consumption and high-risk sexual behavior, such as psychiatric conditions,⁴³ depression was also assessed using a 5-item version of the Center for Epidemiologic Studies of Depression Scale (CESD). We chose to employ the version which has been validated in Spanish among a clinical population in Lima, Peru.⁴⁴ In this study a score of ≥6 was found to be reliably indicative of a positive screen for major depression with a sensitivity of 96% and a specificity of 93%. Participants were also asked to identify their preferred sex role and answer questions about high-risk sexual behaviors such as number of sex partners in the last 3 months, types of sex partners (casual, “friends with benefits,” stable, anonymous, etc), types of sex (anal, oral), and whether or not it was condomless (insertive or receptive). For the analysis, sex role was operationalized as a dichotomized variable looking at insertive only vs. receptive/versatile participants, given the significantly higher likelihood for HIV/STI prevalence shared by the receptive and versatile groups.⁴⁵ Participants were also asked if in the last 3 months they had anal sex in any of the following venues types: discos, saunas, hostels, hair salons, or public places. No assessment of frequency was obtained for this measure. The specific venue types which comprised the answer choices were selected from an ethnographic mapping study with Peruvian MSM and TW populations.⁴⁶ The decision to use this variable was based on preliminary results revealing that the risk of incident HIV infection was increased among participants reporting anal sex in high-risk venues such as saunas and discos (adjusted hazard ratio 3.89, 95% CI 1.03–14.62).⁴⁷ Outside of Peru, sexual partner meeting venue can be important risk factor for HIV/STI acquisition,^48,49,50 especially among alcohol-serving establishments if alcohol is considered to be a facilitator of sexual encounters.

Statistical Analysis

For univariate analysis of categorical variables, we calculated proportions. For univariate analysis of continuous variables, we calculated medians. We then calculated the sensitivities of our various self-report alcohol use measures compared to PEth using the limit of quantitation (≥8 ng/mL) and a higher cut-off of ≥20 ng/ml to avoid misclassifying social alcohol users. In bivariate analysis, we used chi-square tests to determine factors associated with PEth positivity for all variables. However, to determine if the continuous variables of age and median partner number were independently associated with PEth positivity, we used non-parametric Mann-Whitney tests. A p-value threshold of 0.05 was used to determine statistical significance. Using PEth as the gold standard, we calculated the sensitivity and specificity of the various subjective measures (AUDIT, AUDIT-C and CAGE). Lastly to assess the performance of the subjective alcohol measures as binary classifier systems, receiver operating characteristic (ROC) curves were constructed to assess their agreement with PEth positivity using an area under the ROC curve (AUROC) with a value of 1.0 to indicate high agreement. All analyses were conducted using STATA 12 (StataCorp, College Station, TX).

Human subjects and Ethical Review

The institutional review boards of the Universidad Peruana Cayetano Heredia and the University of California, Los Angeles approved the study. Written informed consent was obtained from 86 of the enrolled Picasso participants to participate in the PEth sub-study.

Results

Participant characteristics

Of the 86-person sample, 20% identified as TW. The median age of these participants was 30 years (interquartile range 24 – 37). For the participants analyzed, 63% reported achieving university/post-graduate/technical school level training. Just over 10% of the sample endorsed a history of sexual coercion (meaning they were forced to do something sexual that they did not want to do) (Table I). Of the participants in the sample, 30% had at least 1 recent STI diagnosis at the time of their PEth sample collection. With respect to individual STIs, the prevalence of HIV for this sample was 40% while that of syphilis, chlamydia, and gonorrhea were far fewer in comparison (3%, 9% and 5% respectively). STI status did not differ significantly by PEth positivity. Based on the CESD screening, 7% of the participants were at risk of clinical depression; no significant differences were seen between PEth-positive and PEth-negative individuals for risk of depression.

Table I.

PETH positivity and associations among high-risk men who have sex with men and transwomen in Lima, May 2013 - May 2014^*

	All (n=86)	Peth <8 (n =12)	Peth ≥8 (n =74)		Peth <20 (n =33)	Peth ≥20 (n =53)
Median Age (IQR)	30 (24–37)	30 (24–31)	31 (24–38)		30 (24–38)	30 (25–37)
Unable to meet basic needs in last 12 months	5%	0%	6%		3%	6%
University/Tech/Post- grad Education	63%	67%	62%		61%	64%
Transwomen	20%	33%	18%		27%	15%
History of sexual coercion**	11%	0%	13%		7%	14%
HIV+	40%	33%	41%		42%	38%
Syphilis	3%	8%	3%		6%	2%
Chlamydia	9%	8%	9%		9%	9%
Gonorrhea	5%	0%	5%		0%	8%
Any STI	30%	42%	28%		49%	45%
CESD +	7%	8%	7%		9%	6%
CAI at last sex	16%	17%	16%		21%	13%
CAI in last 3 months	22% (18/77)	17% (2/10)	23% (16/67)		21% (6/28)	25% (12/49)
Median # partners (IQR)	4 (1–10)	3 (1–6)	4 (1–10)		3 (1–9)	4 (1–10)
Report anal intercourse in a high-risk venue	70%	50%	73%		63%	74%
Receptive/versatile anal sex role	74%	67%	76%		73%	76%
AUDIT ≥8**	25%	9%	27%		19%	29%
Median AUDIT score (IQR)	6 (3–8)	2 (0–6)	6 (3–8)		4 (1–7)	6 (3–8)
AUDIT-C ≥4**	60%	36%	63%		47%	67%
Endorse a binge in last 2 weeks	67%	36%	71%*		59%	71%
EtOH use last sex**	3%	0%	4%		0%	6%
Subjectively report being intoxicated at last sex	3%	0%	4%		0%	6%
Baseline AUDIT ≥8	12%	0	14%		9%	13%
CAGE questionnaire positive	6%	0%	7%		0%	9%
Recent drug use (last 3 months)**	7%	0%	9%		3%	10%

^*Bold text indicates significant p-value of 0.022

^**Missing data for 5 participants.

PEth Levels and Alcohol Use

At the currently defined limit of quantitation (≥8 ng/mL), PEth positivity was 86%. Using a PEth cut-off of ≥20 ng/mL yielded a PEth-positive prevalence of 62%. AUDIT scores at the time of PEth sample collection ranged from 0 to 9, while for the same group of participants, prior baseline AUDIT scores ranged from 0 to 21 as published separately.⁵¹ The AUDIT classified 25% of the sample as hazardous alcohol users while the AUDIT-C identified 60% of the sample to be hazardous alcohol users. The prevalence of CAGE positivity, however, was 6% for the sample. A higher proportion of PEth-positive as opposed to PEth-negative participants were AUDIT and AUDIT-C positive (27% vs. 9%, p=0.191 and 63% vs. 36%, p=0.093), though these comparisons did not achieve statistical significance. We also evaluated AUDIT score and PEth as continuous variables and no difference was seen from that presented here. The same was true for CAGE screening wherein a higher proportion of PEth-positive vs. PEth-negative participants screened CAGE positive (7% vs. 0%, p=0.463). However, significantly more participants who reported an episode of binge drinking in the last 2 weeks were PEth-positive compared to PEth-negative (71% vs. 36%, p=0.022).

As shown in Table II, at both PEth cut-offs, asking participants about binges was a more sensitive measure than the AUDIT. Respective AUROCs are shown in Table II. Again, an AUROC of 1.0 indicates high agreement between a subjective measure and PEth positivity. The AUROCs in Table II are quite low (especially when confidence intervals are considered) suggesting poor prediction of high PEth. At a PEth cutoff of 8, the binge question did have the highest AUROC, albeit not considerably so.

Table II.

Sensitivity for self reported alcohol use measures vs. blood PEth level

Subjective Alcohol Us Measures	PEth Results (>8 ng/mL)		AUROCa (95% CI)	Peth Results (>20 ng/mL)		AUROCa (95% CI)
	Positive N=74 (86%)	Negative N=12 (14%)		Positive N=53 (62%)	Negative N=33 (38%)
AUDIT Score
≥8	20 (27%)b	1 (9%)	0.56 (0.49–0.62)	15 (29%)b	6 (19%)	0.56 (0.45–0.68)
<8	51 (73%)	9 (91%)c		37 (71%)	26 (81%)c
AUDIT-C Score
≥4	46 (63%)b	4 (36%)	0.56 (0.48–0.64)	35 (67%)b	15 (47%)	0.60 (0.49–0.71)
<4	27 (37%)	7 (63%)c		17 (32%)	17 (53%)c
Binge in last two weeks
Yes	52 (71%)b	4 (36%)	0.60 (0.50–0.68)	37 (71%)b	19 (59%)	0.56 (0.45–0.68)
No	21 (29%)	7 (64%)c		15 (29%)	13 (41%)c
CAGE Screen
Positive	1 (1%)b	0 (0%)	0.57 (0.50–0.80)	1 (2%)b	0 (0%)	0.69 (0.46–0.63)
Negative	73 (99%)	12 (100%)c		52 (98%)	33 (100%)c

^aArea under the receiver operating characteristic

^bSensitivity

^cSpecificity

PEth as a Correlate for Sexual Risk Behaviors

The median number of sex partners in the last 3 months was 4 (IQR 1–10) for the sample. The median partner number for those who were PEth-positive was higher than that of the PEth-negative participants (4 vs. 3) although this result did not achieve statistical significance. A total of 70% of the sample had anal intercourse in one of the high-risk venues in the last 3 months and a higher proportion of participants who were PEth positive had engaged in a sex act in a public space (73% vs. 50%, p=0.105). To examine a potential dose-response relationship between a higher PEth score and our variables of interest, bivariate analyses were conducted with the higher PEth cutoff of 20 ng/mL. However, with limited power, we found no new significant results.

Discussion

In our study of Peruvian MSM and TW, we found high prevalence of both PEth-positivity and recent binge drinking, consistent with prior reports of high levels of problem drinking in this population.⁵² While other studies have compared PEth levels to self-reported alcohol consumption measures,^{53, 54} to our knowledge this is the first study comparing the CAGE, AUDIT, and AUDIT-C with the PEth biomarker. It is also the first study to examine these relationships among MSM and TW along with the relationship between the PEth biomarker and sexual risk behaviors. Ultimately, we found that instruments to screen for alcohol use disorders and problem drinking were poorly correlated with the PEth biomarker, however direct questioning of binge-drinking was a more sensitive measure.

These results are noteworthy for several reasons. Using AUDIT cutoffs, study participants did not approach the category of alcohol dependence, which is typically indicated by a score of 20 or more.⁵⁵ Similarly, using the CAGE questionnaire, 6% of our participants endorsed what are typically thought of as the behavioral, cognitive, and physiologic manifestations of dependence that develop with repeated alcohol use. The utility of the CAGE questionnaire to detect those with hazardous or harmful alcohol use who have not yet progressed to overt alcohol dependence has been questioned,⁵⁶ while the AUDIT is known to be subject to limitations depending on the cultural context in which it is administered. We should note that the AUDIT and CAGE scores reported in this study are significantly lower than we have previously reported in this and similar cohorts, which may have resulted from social desirability bias as participants in our longitudinal study grew accustomed to self-reported measures.^57,58 As binge drinking is more culturally acceptable in Lima,⁵⁹ it is not surprising that this measure would correspond more highly with a physiologic biomarker.

Unlike other indirect biomarkers of alcohol typically used to diagnose longstanding, heavy alcohol use, PEth concentrations are not influenced by age, gender, other substances, or other comorbidities such as hypertension, kidney, or liver disease.⁶⁰ The superiority of PEth compared with other biological markers of alcohol use has been demonstrated in multiple studies.⁶¹ However, PEth concentrations are yet to be mapped to precise drinking levels in ways that are optimal for clinical use. For example, a set level of how much alcohol one needs to consume and for how long, depending on his/her body composition, to obtain a given PEth result has not been defined. A recent study from Sweden demonstrated that PEth can be used to detect violations of abstinence, as it correlates with low/moderate levels of habitual alcohol consumption.⁶² Discerning high-risk drinking, however, can be more difficult, as quantities consumed and time since abstinence will all impact PEth measurement. We attempted to overcome any effect of low alcohol consumption on our study results by performing a sensitivity analysis using a higher cutoff of 20 ng/mL. However, the performance of either cutoff when compared with other screening instruments did not considerably change. The higher cutoff other merely decreased the number of individuals who were PEth-positive (from 86% to 62%). Controlled, longitudinal studies are necessary to refine PEth measurements for use in reliably distinguishing between safe, social alcohol drinking and heavier at-risk consumption.

Despite the decreased prevalence of alcohol abuse detected by the AUDIT and CAGE instruments compared with prior studies, our findings demonstrate increased sensitivity with screening instruments that are shorter to administer. Self-reported binge drinking correlated most highly with PEth positivity, yielding the highest sensitivity (71%). In addition, the AUDIT-C out-performed the sensitivity of the AUDIT for PEth positivity at both the limit of quantitation (8 ng/mL) and the higher cutoff (20 ng/mL). Using a cut-off of ≥ 4, our calculated sensitivity and specificity of the AUDIT-C (63% for each), were somewhat comparable to originally validated performance characteristics (79% and 56%, respectively);⁶³ differences likely reflected the use of a different gold standard in a different cultural context. Overall, the improved performance of the abbreviated AUDIT which we observed, when compared to the full version, may reflect the fact that questions are focused on quantity and frequency of intake, both of which would be linearly proportional with an individual’s PEth level.

The poor correlation between PEth and alcohol screening instruments does not necessarily mean these measures are not useful. If anything, such poor inter-measure agreement highlights the importance of using multiple screening modalities, both objective and subjective, to better ascertain whether one’s alcohol use is posing a danger to one’s health. A downside to using multiple screening modalities include the time, cost, and burden to the participant. Yet the argument for screening with multiple modalities is only further bolstered by the variation in AUDIT scores observed in this select group over time, highlighted in the results section. Test and retest reliability remains understudied in research on the AUDIT,⁶⁴ and future longitudinal studies that utilize biomarkers would allow for a better understanding of the longevity of these scores.

With respect to sexual risk behavior, we found a high prevalence of anal intercourse in high-risk venues (70%), with a trend toward higher prevalence among individuals who were PEth-positive. Our decision to include this venue variable in the analysis was twofold, partly due to evidence from prior literature, and to plan future interventions. Prior work in Peru has demonstrated increased risk of incident HIV infection among participants reporting anal sex in high-risk venues such as saunas and discos (adjusted hazard ratio 3.89, 95% CI 1.03–14.62).⁶⁵ Similarly, in a prior study of baseline data from this cohort, we found a significant independent association between hazardous drinking (via AUDIT screen) and reporting anal sex in two or more distinct high-risk venues.⁶⁶

These findings suggest there might be potential for behavioral interventions in non-clinical venues in Peru, which has become an increasing practice in various settings internationally such as barber shops, beerhalls, and bathhouses.^{67, 68, 69, 70} Our data further show that these non-clinical spaces could be conceptualized as ideal settings for alcohol-related interventions or interventions that couple HIV-risk reduction with alcohol-related harm reduction. Such work has already been conducted in Cape Town, South Africa where a brief HIV risk-reduction intervention reduced sexual-risk behaviors among drinkers at alcohol serving establishments.⁷¹ By infiltrating the alcohol-serving venues frequented by the community for socialization to bring attention to one’s alcohol use, it may be possible to affect positive change on one’s beliefs, attitudes, and norms related to alcohol use along with the perceived behavioral control one has in these environments.⁷²

Our study has several limitations. The present analysis represents a cross-sectional sub-study from a longitudinal cohort, and thus, causal association between alcohol consumption and risk behaviors cannot be inferred. The small size of our sample limits the ability to generalize our findings but also, due to restricted power, makes the detection of any differences between PEth-positive and negative participants difficult unless they were substantial. Additionally, we are confined in our ability to discern whether PEth levels are indicative of recent versus longstanding heavy drinking, as the literature is yet to offer a consensus on this issue.⁷³ It was also not possible to determine with precision the timing of certain behavioral acts (e.g. last episode of condomless anal intercourse) and their temporal proximity to alcohol use/binge episode. Last, our assurances of confidentiality, aimed at minimizing the potential for social desirability bias, may have been less relevant than otherwise expected, given the unique context of alcohol use in Peru described above.

In conclusion, we found that (i) self-reported alcohol consumption poorly correlated with the PEth biomarker in this modest sample, (ii) asking specifically about binge-drinking simply using the NIAAA definition may be a better screen than the AUDIT to detect episodic heavy alcohol use among high-risk Peruvian MSM and TW, and (iii) alcohol-related interventions are urgently needed in specific situational contexts. The ability to reliably identify alcohol use in populations exhibiting high-risk behaviors that may predispose one to HIV infection remains imperative to guarantee comprehensive care allowing for early intervention and prevention. Given the discrepancy between the various self-report measures we found for this sample, the utility of biologic testing as an additional screening measure is supported. Future studies should establish more clinically meaningful thresholds if we hope to develop interventions based on biomarker levels; particularly if interventions are to be launched in the very sites where alcohol is being consumed.