|Home | About | Journals | Submit | Contact Us | Français|
We sought to determine the prevalence of any alcohol use and hazardous alcohol consumption among HIV-infected individuals engaged in care and to identify factors associated with hazardous alcohol use.
During 2003, 951 patients were interviewed at 14 HIV primary care sites in the USA. Hazardous drinking was defined as > 14 drinks/week or ≥ 5 drinks/occasion for men and > 7 drinks/week or ≥ 4 drinks/occasion for women. Moderate alcohol use was consumption at less than hazardous levels. We used logistic regression to identify factors associated with any alcohol use and hazardous alcohol use.
Forty per cent of the sample reported any alcohol use in the 4 weeks prior to the interview; 11% reported hazardous use. In multivariate regression, male sex [adjusted odds ratio (AOR) 1.52 (95% confidence interval, CI, 1.07–2.16)], a college education (compared to < high school) [AOR 1.87 (1.10–3.18)] and illicit drug use [AOR 2.69 (1.82–3.95)] were associated positively with any alcohol use, while CD4 nadir ≥ 500 cells/μL [AOR 2.65 (1.23–5.69)] and illicit drug use [AOR 2.67 (1.48–4.82)] were associated with increased odds of hazardous alcohol use (compared to moderate and none).
Alcohol use is prevalent among HIV-infected individuals and is associated with a variety of socioeconomic and demographic characteristics. Screening for alcohol use should be routine practice in HIV primary care settings.
Alcohol use is prevalent among HIV-infected individuals, and is associated with adverse health effects [1–3]. It is associated with decreased antiretroviral adherence [4,5] and virological suppression , worsens HIV-associated neuropathy , hastens the progression of liver disease related to hepatitis C (HCV)  and is associated with liver-related mortality among HIV/HCV co-infected persons . Alcohol use is also associated with high-risk sexual and injection-related behaviours that increase the likelihood of HIV transmission [10–12].
The World Health Organization defines hazardous alcohol use as ‘alcohol consumption which confers the risk of physical or psychological harm’ [13,14]. While hazardous drinkers are not alcohol-dependent, they are at risk for the adverse consequences of alcohol use. Despite the myriad complications of alcohol use, alcohol use disorders are frequently not detected in health-care settings . Among a Veterans Administration cohort, healthcare providers often missed alcohol disorders in individuals with higher CD4 cell counts and without HCV .
With the advent of highly active antiretroviral therapy (HAART) in 1996, morbidity and mortality among HIV-infected individuals decreased dramatically . HIV has become a chronic illness, thus increasing the importance of identifying alcohol use among HIV-infected persons. We sought to (1) determine the prevalence of any alcohol use and hazardous drinking in HIV-infected individuals engaged in primary care and (2) identify factors associated with any alcohol use and hazardous drinking.
The HIV Research Network (HIVRN) is a consortium of 21 sites that provide primary and subspecialty care to HIV-infected adult and paediatric patients [18–21]. The 14 participating sites treating adults are located in the eastern (6), mid-western (3), southern (2) and western (3) regions of the USA. Seven of the sites have academic affiliations; seven are community-based.
Between December 2002 and December 2003, face-to-face patient interviews were conducted at 14 adult HIVRN sites. Interviews were administered to a convenience sample of 951 adult (≥ 18 years old) HIV-infected patients who volunteered when asked to participate in an interview, as described previously . The median sample size per site was 59 patients (range 38–172 patients). Gender, race/ethnicity and HIV transmission distributions were similar in the larger population of patients at these sites and in the interviewed sample [sex: 70% vs. 68% male; race/ethnicity: 29% vs. 31% White, 48% vs. 52% Black, 20% vs. 14% Hispanic; HIV transmission: 16% vs. 16% injection drug use (IDU), 38% vs. 34% men who have sex with men (MSM), 3% vs. 3% MSM/IDU, 32% vs. 30% heterosexual (HET), 6% vs. 8% HET/IDU].
Interviews were conducted by professional interviewers trained and supervised by Battelle Corporation (Columbus, OH, USA). The interviews assessed a wide range of demographic, HIV- and substance abuse-related topics. For comparability, interview questions were taken from the interview developed for the HIV Cost and Services Utilization Study (HCSUS) .
HIPAA waivers and IRB approval/exemption of the project, including the interview, were obtained by the data coordinating centre and each site. Informed consent was obtained from each participant before the start of the interview. Participants were reimbursed $30 for the interview, which lasted approximately 1 h.
Interview data included age as of 1 July 2003, calculated from self-reported month and year of birth (categorized as 18–34, 35–44 and 45 or older), racial/ethnic group (non-Hispanic White, non-Hispanic Black, Hispanic and other), gender (male or female) and illicit drug use. We defined current drug use as using illicit drugs within 6 months of the interview, and former drug use as using illicit drugs more than 6 months prior to the interview. Additional interview data included HAART use, highest educational level completed (less than a high school degree, a high school degree or some college and a 4-year college degree or greater), insurance status, number of primary care visits in the past 6 months (categorized as ≤ 3 visits, 4–5 visits, 6–7 visits and ≥ 8 visits), outpatient psychiatric care (yes or no), self-reported CD4 nadir (categorized as < 50, 50–199, 200–499 and > 500 cells/μL) and self-reported anti-retroviral adherence (≥ 95% in the past 2 weeks). HIV risk factor – classified as IDU, MSM, MSM + IDU, heterosexual contact (HET) or HET + IDU – was extracted from medical records.
Alcohol use was ascertained, as in HCSUS , from a series of questions asking: (1) how many days in the past 4 weeks the respondent drank alcohol; (2) how many drinks the person consumed on a typical day when drinking; and (3) the number of days the person consumed more than five drinks. As per US guidelines put forward by the National Institute on Alcohol Abuse and Alcoholism (NIAAA), we defined hazardous drinking as > 14 drinks per week (196 g of alcohol) or ≥ 5 drinks (70 g) per any single occasion for men and > 7 drinks (98 g) per week or ≥ 4 drinks (56 g) per any single occasion for women . Moderate alcohol use was any alcohol use at less than hazardous levels. The NIAAA definition of hazardous alcohol use varies slightly from the UK’s definition of hazardous use, defined as > 21 U (168 g) per week or > 8 U (64 g) per occasion for men and > 14 U (112 g) per week or > 6 U (48 g) per occasion for women . We categorized alcohol use as none, moderate or hazardous.
Analyses were conducted on the 922 patients with complete data. We used logistic regression to examine bivariate associations between any alcohol use and demographic and clinical variables. Insurance status and education were included as markers of socioeconomic status; primary care visits were included as a measure of engagement in medical care. We performed multivariate logistic regression of any alcohol use, including in the model all variables with a P < 0.20 on bivariate analysis. We then examined factors associated with hazardous alcohol use, compared to moderate drinking or abstaining, using the same analytic plan. Finally, we limited our analysis to individuals on HAART (n = 641) and examined factors associated with any alcohol use and hazardous drinking, including adherence as a covariate.
Analyses were conducted using STATA 9.0 (Statacorp LP, College Station, TX, USA). In all regressions, adjustment was made for site of care in order to account for variations in practice patterns and demographic differences across sites. All reported P-values are two-tailed.
Characteristics of this sample are presented in Table 1. Forty per cent reported any alcohol use in the 4 weeks prior to the interview; 11% drank at hazardous levels; 29% drank at moderate levels.
The results of bivariate and multivariate analyses are presented in Table 2. In multivariate regression, current illicit drug use [adjusted odds ratio (AOR) 2.69 (95% confidence interval, CI) 1.82–3.95], male sex [AOR 1.52 (1.07–2.16)] and college education (compared to < high school) [AOR 1.87 (1.10–3.18)] were associated with increased odds of any alcohol use compared to none. Meanwhile, > 7 primary care visits compared to < 4 in the past 6 months [AOR 0.51 (0.35–0.78)] and having Medicaid insurance compared to private insurance [AOR 0.58 (0.36–0.96)] were associated with decreased odds of alcohol use. HAART use was associated positively with any alcohol use [AOR 1.62 (1.15–2.29)]. When the sample was limited to individuals on HAART (results not shown), adherence was not associated significantly with any alcohol use [AOR 0.72 (0.49–1.05); P = 0.09]; other factors associated with alcohol use did not change.
In multivariate logistic regression, factors associated with hazardous alcohol use, compared to moderate or none (Table 3), included current illicit drug use [AOR 2.67 (1.48–4.82)] and a CD4 nadir ≥ 500 cells/μL [AOR 2.65 (1.23–5.69)] compared to those with CD4 < 50 cells/μL. Conversely, > 7 primary care visits compared to < 4 in the past 6 months was associated with decreased odds of hazardous alcohol use [AOR 0.45 (0.23–0.86)]. When the sample was limited to those on HAART (Table 4), high adherence was associated with lower odds of hazardous alcohol use [AOR 0.49 (0.28–0.85); P = 0.01].
Among this sample of HIV-infected individuals in primary care in the USA, the overall 4-week prevalence of any alcohol use was 40%, with 11% drinking at hazardous levels and 29% drinking at moderate levels. Active illicit drug use, decreased antiretroviral adherence and a nadir CD4 count > 500 cells/μL were associated with increased odds of hazardous alcohol use, while > 7 primary care visits was associated with lower odds of hazardous alcohol use. Male sex, higher educational attainment, private insurance (compared to Medicaid) and HAART use were associated with any alcohol use, but not with hazardous use.
The prevalence of any alcohol use was lower in our sample compared to the HIV Cost and Services Utilizations Survey (HCSUS), a USA-based national survey of HIV-infected adults in care, where the 1-month prevalence of alcohol use was 53% with 8% classified as heavy drinkers . This difference may be explained partially by the increased proportion of women and Hispanic people in this HIVRN interview sample – two groups in the USA that have a lower prevalence of alcohol use (compared to men or non-Hispanic people) . In addition, this lower prevalence of alcohol use may reflect sampling bias. Although the interview participants were similar to the overall cohort, the sample was largely one of convenience and non-drinkers may have been more likely to participate in the interview.
The 11% prevalence of hazardous use in our population was greater than the prevalence of heavy drinking in HCSUS (8%) and may in part be explained by a broader definition of hazardous use in this study compared to ‘heavy’ alcohol use in HCSUS (defined as five or more drinks on ≥ 4 days in the previous 4 weeks). Despite the differences in prevalence, factors associated with alcohol use did not differ significantly between our sample (in the current HAART era) and HCSUS (in the early HAART era). Both studies demonstrate that men, individuals with a college education and illicit drug users are more likely to consume alcohol, and that persons with higher CD4 cell counts and illicit drug users are more likely to drink at hazardous levels. The consistency between our findings and those of HCSUS suggest that the increased longevity and clinical benefits of HAART have not changed the correlates of alcohol consumption.
That illicit drug use was associated with both any use and hazardous levels of alcohol use was not surprising, because alcohol and illicit drug use frequently co-occur . The association of > 7 primary care visits with decreased hazardous alcohol use may reflect a population with advanced disease who may be too sick to drink alcohol. Consistent with this finding, CD4 nadir > 500 cells/μL compared to < 50 cells/μL was associated with hazardous alcohol use, suggesting that those with more advanced immunosuppression may be too ill to drink at this level .
Male sex was associated with any alcohol use, which is consistent with multinational studies examining the epidemiology of alcohol use . Other factors associated with any alcohol use included higher educational attainment, private insurance and HAART use. These variables may be capturing aspects of socioeconomic status, suggesting an association between higher socioeconomic status and any alcohol use [30–32].
In applying the results of this study to clinical settings, it is important to note that the definition of hazardous alcohol use varies among countries, as does the grams of alcohol comprising a standard drink. Because this is a USA-based study, we defined hazardous use by the NIAAA definition of hazardous drinking, and we used standard drinks as our unit of measurement. Fourteen grams of alcohol is equivalent to one US standard drink. As a comparison, 8 g of alcohol is equivalent to one British unit of alcohol. Notably, the cut-off for hazardous alcohol use in the USA is lower than in many other countries such as Australia, which defines hazardous use as > 280 g per week or > 60 g per occasion in men, and > 140 g per week or > 40 g per occasion in women . What is considered hazardous use in the USA may be considered moderate use in other countries, where the threshold for hazardous alcohol use is higher. Thus in countries where the cut-off for hazardous drinking is higher than in the USA and the UK, the adverse effects of alcohol use may occur at what they consider to be moderate levels of alcohol consumption.
Our study has potential limitations. Firstly, this is a cross-sectional study, so we are unable to make any causal inferences based on these data. Secondly, alcohol use is based on self-report, which may be subject to bias – individuals frequently underestimate how much alcohol they consume. Thirdly, the interview sample was largely one of convenience; as a result, there may have been a selection bias. However, our sample did reflect the demographics of the larger HIVRN population. Also, the sample is not representative and does not generalize to all HIV patients. However, the sites from which patients were sampled do encompass a broad geographic distribution, and multi-site studies afford greater generalizability than single-site studies.
In conclusion, any alcohol use and hazardous drinking are prevalent among HIV-infected individuals participating in primary care, and are associated with a variety of socioeconomic and demographic characteristics, and markers of disease severity. Efforts should be made to identify individuals with alcohol use because alcohol use is a modifiable risk factor for adverse HIV-associated outcomes. Screening for hazardous alcohol consumption among all HIV-infected individuals is an important aspect of comprehensive HIV care.
Supported by the Agency for Healthcare Research and Quality (290-01-0012), the National Institute on Drug Abuse (K23-DA00523) and the National Institute of Alcohol Abuse and Alcoholism (K23 AA015313). K.G. was also sponsored by the Johns Hopkins Clinician Scientist Award.
*The views expressed in this paper are those of the authors. No official endorsement by DHHS, the National Institutes of Health or the Agency for Healthcare Research and Quality is intended or should be inferred.