Abstract
Objectives
In HIV-infected drinkers, alcohol types more likely to cause inflammation could plausibly increase the risk of HIV disease progression. We therefore assessed the association between alcohol type and plasma HIV RNA level (HIV viral load) among HIV-infected drinkers not on antiretroviral therapy (ART) in Russia and Uganda.
Methods
We analyzed the data of participants from cohorts in Russia and Uganda and assessed their HIV viral load at enrollment by the alcohol type predominantly consumed. We defined predominant alcohol type as the alcohol type contributing >50% of total alcohol consumption in the 1 month (Russia) or 3 months (Uganda) prior to enrollment. Using multiple linear regression, we compared log10 HIV viral load by predominant alcohol type, controlling for age, gender, socioeconomic status, total number of standard drinks, frequency of drinking ≥6 drinks/occasion, and in Russia, history of injection drug use.
Results
Most participants (99.2% of 261 in Russia and 98.9% of 352 in Uganda) predominantly drank one alcohol type. In Russia, we did not find evidence for differences in viral load levels between drinkers of fortified wine (n=5) or hard liquor (n=49), compared to drinkers of beer/low-ethanol-content cocktails (n=163); however, wine/high-ethanol-content cocktail drinkers (n=42) had higher mean log10 viral load than beer/low-ethanol-content cocktail drinkers (β=0.38, 95%CI: 0.07 to 0.69; p=0.02). In Uganda, we did not find evidence for differences in viral load levels between drinkers of locally-brewed beer (n=41), commercially-distilled spirits (n=38), or locally-distilled spirits (n=43), compared to drinkers of commercially-made beer (n=218); however, wine drinkers (n=8) had lower mean log10 HIV viral load (β=−0.65, 95% CI −1.36 to 0.07, p = 0.08), although this did not reach statistical significance.
Conclusions
Among HIV-infected drinkers not yet on ART in Russia and Uganda, we observed an association between the alcohol type predominantly consumed and the HIV viral load level in the Russia sample. These exploratory results suggest that, in addition to total number of drinks and drinking patterns, alcohol type might be a dimension of alcohol use that merits examination in studies of HIV and alcohol related outcomes.
Keywords: Alcohol types, HIV-infected patients, HIV-disease progression, Uganda, Russia, HIV viral load
Background
Heavy alcohol consumption can increase risk of HIV transmission and acquisition [1, 2]. In HIV-infected individuals, heavy drinking may also affect disease progression [3] and adherence and response to antiretroviral therapy (ART) [4]. While previous studies assessing the negative effects of alcohol consumption have tended to quantify alcohol use via either the volume of alcohol consumed or the drinking pattern [5, 6], the type of alcohol consumed may also be important. Alcohol types vary widely in ethanol concentration, processing methods, and ingredients [7]. Such differences may explain why certain alcohol types have been reported to be more strongly associated with negative clinical outcomes in general population studies [8] and in studies of HIV-infected patients [9].
Associations between alcohol type and clinical outcomes might be explained by variations in either the total volume of pure ethanol consumed or the drinking patterns by alcohol type. For example, those drinking certain types might consume larger amounts of pure ethanol or be more likely to binge-drink [10, 11]. However, different alcohol types could also exert different biological effects independent of both volume of ethanol consumed and the drinking pattern. For example, fermented alcohols, especially wines, are suggested to have polyphenols that can reduce inflammation [12, 13]. Alcoholic beverages that are made through distillation (i.e., “purification”) processes, such as liquor, might be deficient of such “beneficial” chemicals, and the ethanol in such drinks could in turn be more inflammatory [14]. Alternatively, imperfectly processed “locally-made” drinks may retain excessive levels of harmful non-ethanol chemicals like methanol [15], which could promote inflammation. Such differences in the propensity to cause inflammation may be important in HIV-infected drinkers, since inflammation plays a key role in HIV disease progression [16, 17].
Previous studies have examined the association between heavy alcohol use and HIV disease progression, but findings remain inconclusive. In the systematic review by Azar et al., 2010, HIV-infected patients with alcohol use disorders were more likely to experience decreased adherence to ART and poor treatment outcomes [18]. However, in a narrative review by Hahn et al., 2010, studies in the pre-ART era generally found no associations between heavy drinking and either viral load or CD4, while studies in the ART era found some associations between heavy drinking and CD4+ cell count declines and viral suppression [19]. Such associations could be mediated by behavioral factors like adherence to ART or biological factors like immune activation, microbial translocation or overlapping metabolic pathways between alcohol and ART [20].
If alcohol types are differently associated with inflammation, then alcohol type is a possible biologic mediator of observed associations between heavy drinking and HIV disease progression. In one study in the United States involving 165 HIV-infected drinkers on ART, those drinking only liquor (N = 55) were less likely to achieve viral suppression after 6 months of ART than those drinking only beer or wine [9]. The liquor-only drinkers were also less likely to increase their thymic volumes, further suggesting diminished responses to ART [21]. These associations were adjusted for total volume of alcohol consumed, but not drinking patterns. We are not aware of any studies reporting on the association of “locally-made” alcohols with health outcomes; these “locally-made” alcohols are commonly consumed in Uganda [22].
In Uganda a large proportion of adults (58.7% in 2010) report abstaining from alcohol, but the per capita alcohol consumption is high (9.8 liters of pure ethanol in 2010). In Russia smaller proportion (32.2%) report abstention from alcohol, and the per capita alcohol consumption is even higher (15.1 liters of pure ethanol in 2010) [23]. HIV prevalence is high (7.4% in 2013) in Uganda and ~1% in Russia [24, 25]. In addition to allowing the exploration of pathways through which heavy drinking might be associated with HIV disease progression, studying the association of alcohol type with health outcomes in these settings could potentially expand the range of available interventions for unhealthy drinking, e.g., through beverage substitution interventions or through beverage specific volume reduction messaging.
We thus took advantage of existing cohort studies in Russia and Uganda to describe the alcohol type preferences of HIV-infected drinkers and to assess associations between the alcohol type predominantly consumed prior to enrollment and the HIV viral load at enrollment. Based on the hypothesis that alcohols with higher ethanol content are more likely to promote inflammation and, through this, increase HIV viral load, we evaluated whether drinkers of higher ethanol content drinks had higher viral loads, compared to drinkers of commercially brewed beers.
METHODS
Study design
We analyzed the baseline data of HIV-infected adults enrolling into 3 cohort studies between 2011 and 2015 in Russia (1 cohort of 360 participants) and Uganda (2 cohorts with a total of 697 participants). These studies were separate studies with differing eligibility criteria, but all recruited HIV-positive participants not yet on ART. For this analysis, we selected participants who reported consuming any alcohol prior to enrollment (past 30 days in Russia, past 3 months in Uganda, per the individual study enrollment criteria). The Ugandan studies used a longer time frame of alcohol consumption prior to enrollment to allow for identification of drinkers, since drinking in this setting is often underreported and a longer time frame was considered more socially acceptable to mitigate the risk of under-reporting [26]. Participants in both countries completed interviewer-administered structured questionnaires and provided blood samples at the time of the interview, and received HIV care independently of study activities. We restricted our analysis to only the self-reported drinkers since only those would have provided alcohol type information. Given the distinct alcohol types consumed in the two countries, the analysis was stratified by country.
Ethics statement
All participants provided written informed consent to participate in their respective cohort study; those who had a cognitive impairment resulting in inability to provide informed consent were excluded. The Russian study was approved by Institutional Review Boards (IRBs) at Boston University School of Medicine/Boston Medical Center, and The First St. Petersburg Pavlov State Medical University. The Ugandan studies were approved by IRBs at The University of California San Francisco, Boston Medical Center, and Mbarara University of Science and Technology, and by Uganda’s National Council for Science and Technology.
Setting and study population
URBAN ARCH Russia cohort
The Russia sample included participants from the Uganda Russia Boston Alcohol Network for Alcohol Research Collaboration on HIV/AIDS (URBAN ARCH) consortium’s Russia cohort, a prospective observational cohort study of HIV-infected individuals from St. Petersburg, Russia, to assess the longitudinal association between alcohol consumption and biomarkers of microbial translocation and inflammation/altered coagulation. Eligibility criteria were: documented HIV infection, 18–70 years of age, not yet on ART, living within 100km of St. Petersburg, providing contacts of at least 2 relatives or close friends who could assist with follow-up, having a telephone and being fluent in Russian. Participants were enrolled irrespective of their CD4+ T cell count levels and alcohol consumption. URBAN ARCH’s Russia cohort started enrollment in 2012. For this analysis, we included participants enrolled from 2012 to 2015.
The ADEPT and BREATH cohorts (Uganda Cohorts)
The Uganda sample included participants from the URBAN ARCH consortium’s Uganda cohort, known as the Alcohol Drinking Effects on Progression prior to Treatment (ADEPT) study, and from the Biomarker Research of Ethanol among Those with HIV (BREATH) study. Both studies were prospective observational cohorts of HIV-infected adults at the Immune Suppression Syndrome (ISS) clinic in Mbarara, Uganda. ADEPT’s aim was to determine the effect of alcohol consumption on pre-ART HIV disease progression; BREATH aimed to describe changes in alcohol consumption during the first year of HIV care [22, 27, 28]. Eligibility criteria for both cohorts included: HIV-infected adult (age ≥ 18), enrolled into care at the ISS clinic and not yet on ART, fluent in English or Runyankole (the local language), living within 60km of the clinic, and for BREATH only, new to the ISS Clinic and HIV care, and reporting past-year alcohol use. ADEPT initially (August 2011–February 2014) recruited only patients with CD4+ T cell count >350 cells/mm3. After February 2014, Uganda’s national guidelines for ART initiation were changed such that patients with CD4+ T cell counts ≤500 became eligible for ART; we thus then recruited only those patients with CD4+ T cell counts >500. By this time, however, all but 68 patients had already been enrolled. Participation in ADEPT ended and participants were exited from the study once the clinic booked them for ART initiation. BREATH enrolled patients irrespective of CD4+ T cell count.
Measurements
Alcohol type
The primary independent variable was the alcohol type predominantly consumed. We defined the predominant alcohol type per participant as the type contributing >50% of the participants’ overall absolute alcohol consumption. The total number of standard drinks, and the frequency of drinking 6 or more drinks per occasion, were also measured and used as covariates in the analysis.
In Russia, ethanol concentration estimates were based on manufacturer labels since all the alcohol types assessed in this setting were commercially produced, and survey questions included 4 alcohol types as (from lowest to highest alcohol concentration): beer or low-ethanol-content cocktails; wine or high-ethanol-content cocktails; fortified wine; and hard liquor (e.g., vodka). “Cocktails” refers to canned or bottled mixed drinks sold in commercial alcohol stores. They were classified as either “high-ethanol-content cocktails” (~9% ethanol by volume), or “low-ethanol-content cocktails” (~5.5% ethanol by volume). On the study survey, a question was asked about drinking low-ethanol-content cocktails and commercial beer together since these were thought to have similar ethanol concentrations. The two are thus grouped into one alcohol type: “beer or low-ethanol-content cocktails”. Similarly, a question was asked about high-ethanol-content cocktails and wine together leading to the “wine or high-ethanol-content cocktails” category.
In Uganda, ethanol concentration estimates were based on a previous study, which measured alcohol content in different alcohol types using an alcohol analyzer at a brewing company [29]. The classification of alcohol types on the study surveys in Uganda depended on the methods of production (commercial vs. local/traditional), as well as whether a drink was a beer, a spirit, or a wine. Data on 5 alcohol types were collected (from lowest to highest estimated alcohol concentration): commercially-brewed beer, locally-brewed beer, wine, commercially-distilled spirits, and locally-distilled spirits.
Total alcohol volume consumed
In both countries, measurements of alcohol volumes were beverage-specific, and illustrations of common containers in which different drinks are sold were used to aid recall (Figure 1). In Russia, a beverage-specific timeline follow-back method with 30-day recall was used [30]. With the aid of a calendar, participants were asked to report the volume of each alcohol type that they drank yesterday, the day before yesterday, etc., for the past 30 days [31]. The daily amounts were added to obtain monthly beverage-specific total volumes of alcohol. In Uganda, a beverage-specific quantity frequency method was used. Participants were asked to report volumes consumed for each alcohol type on a “typical drinking day” in the past 3 months and their frequency of drinking in the past 3 months. These two quantities were used to calculate the total (beverage-specific) alcohol volumes consumed in the past 3 months as previously described [22].
Figure 1. Illustrations of containers that were used to aid patient recall during interviews for both the number of standard drinks and the total alcohol volume consumed.
In Russia, a complex chart was used, as shown. The captions in the top left and top right corners are Russian text translating to: “Examples of standard drink types (alcohol beverages)” and “Picture 1”, respectively. The pictures on the chart show actual beverages: the first two (cans) are beers and are labelled “beer”, the next three bottles are wines, labelled as “table wine”, the red bottles are “fortified wine” and are labelled as such. The two cans at the end of the chart are cocktail drinks and are each labelled with Russian text for “cocktail”. The illustrations of glasses represent how the associated drinks are commonly consumed in this setting, with the small glass on the right side representing a “shot of liquor” (1 standard drink), the one on the left representing a “glass of wine” (one standard drink), while the one in the middle represents a “glass of fortified wine” (1.5 standard drinks). The numbers at the bottom of the chart represent the total number of standard alcoholic drinks in the associated container. For example, the 0.3L beer-can is 1.0 standard drink, while the 0.5L can is 1.5 standard drinks, etc.
In Uganda, simple illustrations showing a beer bottle, a wine glass, and a shot of liquor were used without any associated text as shown. Volumes and standard drink quantities were explained to participants by the interviewer.
Total number of standard alcoholic drinks
To obtain the total number of standard alcoholic drinks consumed by a participant, we first estimated the beverage-specific grams of alcohol by multiplying the volume of each drink type consumed by estimated ethanol content per drink as previously described [29]. The following estimates were used for each drink type: in Russia, beer/low-ethanol-content cocktails (3.92%), wine/high-ethanol-content-cocktails (9.52%), fortified wine (13.44%), and hard liquor (31.50%); in Uganda, beers (3.95%), wines (9.87%), and spirits (31.57%). Beverage-specific grams of alcohol were then summed into a total for the reference period and divided by 14, the US National Institutes on Alcohol Abuse and Alcoholism (NIAAA) standard number of grams for one alcoholic drink.
Defining a participant’s predominant alcohol type
To determine each participant’s predominant alcohol type, we assessed fractional contributions to the total number of standard drinks by each alcohol type. The type contributing more than half of the participant’s total standard drinks was their predominant type. For example, if a participant reported consuming 8 standard drinks of alcohol from wine in the reference period out of a total consumption (from all drink types) of 10 standard drinks, the fraction of absolute alcohol due to wine was 0.8, and wine was their predominant alcohol type. If no single type accounted for >50% of reported alcohol consumption, the participant was considered as “having no predominant alcohol type”.
Drinking patterns
In both countries, we assessed the number of days when a participant drank 6 or more drinks on one occasion as a proxy measure of drinking patterns [32]; we categorized responses into three groups: never, less than monthly or once to thrice a month, and weekly or more often. For this question, we defined a drink for the participant as a 140ml glass of 12%-alcohol wine, a 40ml container of hard liquor, or a 360ml bottle or can of beer, also using illustrations of relevant containers (Figure 1).
Other covariates
We also obtained the participants’ age and socioeconomic status (SES). In Russia, we used individual-level monthly income to measure SES. In Uganda we created a household asset index based on household ownership of durable goods, housing quality, and available energy sources as a proxy measure of SES (in this setting, the asset index is suggested to be a better measure of SES) [33]. We also asked about history of injection drug use in both countries (although this was not reported by any participants in the Uganda cohorts). Participants reported the date of first HIV-positive diagnosis; from this, we calculated the years since diagnosis, at enrollment. As underreporting of alcohol use is common in HIV-infected patients in Uganda, we measured levels of the alcohol biomarker phosphatidylethanol (PEth) as previously described [22], and controlled for this variable in a sensitivity analysis. In both countries, we also measured participants’ CD4+ T-cell count levels (APC-H7, BD Biosciences, for the Russia sample, and Beckman Coulter, Brea, California, for the Uganda sample).
Plasma HIV RNA level
The outcome was the plasma HIV RNA level measured on frozen and batched samples by the RIBO-sorb AmpliSens HIV-Monitor-FRT, for the Russia samples (Federal Budget Institute of Science, Central Research Institute for Epidemiology, Moscow), and the Versavt HIV-1 RNA 3.0 Assay, for the Uganda samples (Bayer system 340 bDNA analyzer, Bayer HealthCare Corporation, Whippany, NJ). Log10-transformed values of the HIV viral load were used in the regression analyses.
Analysis
We described the participants’ characteristics and assessed the association between predominant alcohol type and the log10 viral load stratified by country. In linear regression models comparing the log10 HIV viral load by predominant alcohol type, beer/low-ethanol-content cocktails (in Russia) and commercially-made beer (in Uganda) were used as the reference categories. Participants without a predominant alcohol type were excluded from the regression analyses.
In the adjusted analysis, we controlled for covariates which we believed a priori to be correlates of both alcohol type preference and HIV disease progression (i.e., potential confounders) based on the literature and clinical knowledge. These included age, gender, SES (income/asset index), total number of standard alcoholic drinks, frequency of drinking ≥6 drinks per occasion, and injection drug use (Russia only). Numeric covariates (age, SES, and number of standard drinks) were all modelled as restricted cubic splines. Since HIV-infected patients in Uganda may underreport volumes of alcohol consumed [34], we repeated the analysis in the Uganda sample further adjusting for PEth concentrations, also modelled as a restricted cubic spline. Since some descriptive differences in years since diagnosis were observed by alcohol type, we performed a second sensitivity analysis post-hoc, adding years since diagnosis to the adjusted model, also modelled as a restricted cubic spline. Analyses were performed in Stata 13 (College Station, Texas), and, for all analyses, p-values <0.05 were considered statistically significant.
Three participants in Russia and 12 in Uganda were missing income information; one participant in Russia and two in Uganda were missing date of HIV diagnosis; and one participant in Uganda lacked PEth measurements. We substituted the missing values with sample’s median for these variables so as to retain these observations in the adjusted analyses [35]. One participant in Russia and 6 in Uganda lacked HIV viral load (outcome) measurements. For these participants, we chose not to impute their viral load values, given that the viral load was the main outcome of interest, and we did not have other appropriate biological data to rely on during the imputations. The 7 participants were thus excluded from the analysis.
RESULTS
Participants’ characteristics
Russia
In Russia, a total of 360 participants were enrolled between 2012 and 2015. We excluded 99 individuals (50 who were HIV antibody negative or had undetectable viral loads at enrollment, suggesting either HIV negativity or ART positivity; 49 who did not report any alcohol consumption in the past month). A total of 261 individuals were thus analyzed. Median age was 33 years (interquartile range (IQR) 30 to 37), and 69.4% were male. Median CD4+ T-Cell count was 465 cells/mm3 (IQR 299 to 683).
Uganda
ADEPT enrolled a total of 484 participants; we excluded 255 from these analyses (37 were ineligible because they were either ART-positive or HIV-antibody negative/indeterminate, and 218 did not report any alcohol consumption in the past 3 months). BREATH enrolled a total of 213 participants; we excluded 90 from these analyses (8 for not meeting study eligibility criteria, 42 for being co-enrolled in ADEPT, and 40 for not reporting any alcohol consumption in the past 3 months). Consequently, the Uganda sample comprises 352 self-reported ART-naïve HIV-infected drinkers assessed between 2011 and 2014 (229 from ADEPT and 123 from BREATH). Their median age was 31 years (IQR 25 to 38); 45.5% were male (Table 1). Median CD4+ T-Cell count was 486 cells/mm3 (IQR 332 to 626).
Table 1.
Characteristics of participants at cohort entry by country. The table shows the characteristics at cohort entry of HIV-infected participants enrolling in 3 cohort studies in Russia and Uganda. The reference period for drinking reports was past month in Russia and past 3 months in Uganda.
| Russia (n=261) | Uganda (n = 352) | |
|---|---|---|
| Year of enrollment | 2012 to 2015 | 2011 to 2014 |
| Age | 33 (30 to 37)* | 31 (25 to 38) |
| Male sex | 181 (69.4%) | 160 (45.5%) |
| Monthly income (USD) | 309.6 (123.8 to 464.4) | 30 (17 to 60) |
| Asset index score | - | 0.0 (−1.7 to 1.7) |
| History of injection drug use | ||
| No | 154 (59.0%) | 352 (100.0%) |
| Yes | 107 (41.0%) | 0 (0.0%) |
| Years since first HIV-positive diagnosis | 6.8 (2.7 to 11.6) | 0.2 (0.0 to 2.9) |
| Total number of standard drinks | 62.3 (36.8 to 116.4) | 32.1 (8.5 to 126.9) |
| Frequency of 6+ drinks | ||
| Never | 22 (8.4%) | 228 (64.8%) |
| Less than monthly or once to thrice a month | 121 (46.4%) | 78 (22.2%) |
| Weekly or more often | 118 (45.2%) | 46 (13.1%) |
| Alcohol type predominantly drank† | ||
| Russia | ||
| Beer or low-ethanol-content cocktails‡ | 163 (62.5%) | - |
| Wine or high-ethanol-content cocktails | 42 (16.1%) | - |
| Fortified wine | 5 (1.9%) | |
| Liquor such as vodka ∞ | 49 (18.8%) | |
| No specific predominant type | 2 (0.8%) | |
| Uganda | ||
| Commercially-brewed beer‡ | - | 218 (61.9%) |
| Locally-brewed beer | - | 41 (11.7%) |
| Wine | - | 8 (2.3%) |
| Commercially-distilled spirits∞ | - | 38 (10.8%) |
| Locally-distilled spirits∞ | - | 43 (12.2%) |
| No specific predominant type | - | 4 (1.1%) |
| Phosphatidylethanol (PEth) (ng/mL) | - | 73.0 (14.4 to 265.5) |
| CD4+ T cell count (cells/mm3) | 465.0 (298.9 to 683.4) | 485.5 (332.0 to 625.5) |
| HIV viral load (log10 copies/ml) | 4.6 (4.0 to 5.2) | 3.9 (3.2 to 4.5) |
Median (Interquartile range) unless otherwise specified
Contributing >50% of total alcohol volume in the reference period; listed in order of lowest to highest estimated alcohol content
Lowest ethanol concentration
Highest ethanol concentration (in Uganda, unclear whether locally distilled spirits considered to have the same concentration as commercially distilled spirits, but the former’s ethanol concentrations may be more variable).
Alcohol type preferences
In both countries, nearly all participants predominantly drank one alcohol type; only 2 participants (0.8%) in Russia and 4 participants (1.1%) in Uganda did not have a predominant alcohol type. In Russia, beer/low-ethanol-content cocktails were the most common type consumed (163/261; 62.5%). In Uganda, most participants predominantly drank commercially-brewed beer (218/352; 61.9%) (Table 1).
Alcohol type and participant characteristics
Distributions of participant characteristics, by predominant alcohol type, are presented in Table 2 among participants reporting a predominant alcohol type. There appeared to be some differences by alcohol type. For example, in Russia, 52.4% of wine/high-ethanol-content-cocktail drinkers reported injection drug use, compared to 42.3% of beer/low ethanol content cocktail drinkers, 40.0% of fortified wine drinkers, and 28.6% of hard liquor drinkers. In Uganda, 76.3% of all commercial spirit drinkers were male, compared to 37.6% of commercial beer drinkers. The median monthly income in Uganda of wine and commercially distilled spirit drinkers was 65 USD (IQR 32 to 135) and 55 USD (IQR 30 to 90), respectively, compared to 30 USD (IQR 18 to 60) among commercial beer drinkers and 24 USD (IQR 14 to 54) among locally-brewed beer drinkers. In Uganda, liquor drinkers appeared to have higher PEth levels (median 167 ng/ml (IQR 26 to 730) for commercial spirits, and 181 ng/ml (IQR 59 to 510) for locally distilled spirits), compared to beer drinkers (median 56 ng/ml (IQR 10 to 148) for commercial beer, and 88 ng/ml (IQR 23 to 334) for locally-brewed beer).
Table 2.
Descriptive table of study participant characteristics at cohort entry by predominant drink type and stratified by country*. The reference period for drinking reports was past month in Russia and past 3 months in Uganda.
| Russia‡ | ||||||
|---|---|---|---|---|---|---|
|
| ||||||
| Overall | Beer/low-ethanol cocktails | Wine/high- ethanol cocktails | Fortified wine | Hard liquor (e.g., vodka) | ||
| Characteristic. | n = 259 | n = 163 | n = 42 | n = 5 | n = 49 | |
| Male sex | 180 (69.5%) | 116 (71.2%) | 24 (57.1%) | 5 (100.0%) | 35 (71.4%) | |
| Age | 33 (30 to 37) ∞ | 33 (30 to 36) | 32 (29 to 35) | 31 (29 to 39) | 36 (32 to 39) | |
| Monthly income (USD) | 310 (124 to 464) | 232 (77 to 464) | 310 (124 to 464) | 310 (155 to 929) | 310 (155 to 464) | |
| History of injection drug use | ||||||
| No | 152 (58.7%) | 94 (57.7%) | 20 (47.6%) | 3 (60.0%) | 35 (71.4%) | |
| Yes | 107 (41.3%) | 69 (42.3%) | 22 (52.4%) | 2 (40.0%) | 14 (28.6%) | |
| Years since first HIV-positive diagnosis | 6.8 (2.8 to 11.4) | 7.1 (3.8 to 12.0) | 7.2 (2.6 to 9.9) | 3.9 (1.2 to 5.9) | 4.4 (1.6 to 9.7) | |
| Total number of standard drinks | 62 (37 to 116) | 59 (34 to 100) | 74 (48 to 167) | 115 (53 to 220) | 72 (37 to 126) | |
| Frequency of 6+ drinks | ||||||
| Never | 22 (8.5%) | 10 (6.1%) | 4 (9.5%) | 1 (20.0%) | 7 (14.3%) | |
| Less than monthly or once to thrice a month | 119 (46.0%) | 82 (50.3%) | 19 (45.2%) | 2 (40.0%) | 16 (32.7%) | |
| Weekly or more often | 118 (45.6%) | 71 (43.6%) | 19 (45.2%) | 2 (40.0%) | 26 (53.1%) | |
| HIV viral load (log10 copies/ml) | 4.6 (4.0 to 5.2) | 4.5 (3.9 to 5.0) | 4.8 (4.2 to 5.7) | 4.1 (4.0 to 4.7) | 4.5 (3.6 to 5.3) | |
| CD4+ T cell count (cells/mm3) | 462 (299 to 683) | 448 (292 to 668) | 486 (354 to 642) | 632 (471 to 839) | 478 (278 to 724) | |
|
| ||||||
| Uganda† | ||||||
|
| ||||||
| Overall | Commercially- brewed beer | Locally-brewed beer | Wine | Commercially- distilled spirits | Locally-distilled spirits | |
| Characteristic. | n = 348 | n = 218 | n = 41 | n = 8 | n = 38 | n = 43 |
| Male sex | 157 (45.1%) | 82 (37.6%) | 24 (58.5%) | 3 (37.5%) | 29 (76.3%) | 19 (44.2%) |
| Age | 30 (25 to 38) | 30 (25 to 36) | 34 (29 to 43) | 29 (26 to 34) | 34 (27 to 44) | 32 (27 to 39) |
| Monthly income (USD) | 30 (18 to 60) | 30 (18 to 60) | 24 (14 to 54) | 65 (32 to 135) | 55 (30 to 90) | 27 (12 to 45) |
| Asset index score | 0.0 (−1.7 to 1.7) | 0.3 (−1.6 to 1.7) | −1.2 (−2.1 to 0.0) | 1.2 (−0.2 to 4.2) | 1.1 (−0.5 to 3.0) | −1.7 (−2.9 to 0.3) |
| Years since first HIV-positive diagnosis | 0.2 (0.0 to 2.8) | 0.2 (0.0 to 2.5) | 0.6 (0.1 to 3.6) | 3.9 (0.6 to 6.2) | 0.2 (0.0 to 3.3) | 0.1 (0.0 to 2.0) |
| Total number of standard drinks | 32 (8 to 127) | 21 (7 to 80) | 40 (17 to 190) | 17 (2 to 27) | 172 (71 to 358) | 45 (12 to 176) |
| Frequency of 6+ drinks | ||||||
| Never | 227 (65.2%) | 146 (67.0%) | 29 (70.7%) | 8 (100.0%) | 16 (42.1%) | 28 (65.1%) |
| Less than monthly or once to thrice a month | 76 (21.8%) | 51 (23.4%) | 7 (17.1%) | 0 (0.0%) | 11 (29.0%) | 7 (16.3%) |
| Weekly or more often | 45 (12.9%) | 21 (9.6%) | 5 (12.2%) | 0 (0.0%) | 11 (29.0%) | 8 (18.6%) |
| HIV viral load (log10 copies/ml) | 3.9 (3.2 to 4.5) | 3.9 (3.3 to 4.5) | 4.1 (3.4 to 4.9) | 3.2 (2.9 to 3.7) | 3.9 (3.1 to 4.9) | 3.9 (3.4 to 4.7) |
| CD4+ T cell count (cells/mm3) | 486 (332 to 625) | 488 (354 to 618) | 436 (307 to 634) | 495 (434 to 554) | 493 (322 to 672) | 431 (326 to 588) |
| Phosphatidylethanol (PEth) (ng/mL) | 73 (14 to 257) | 56 (10 to 148) | 88 (23 to 334) | 25 (6 to 51) | 167 (26 to 730) | 181 (59 to 510) |
Limited to participants with a predominant alcohol type
Median (Interquartile range) unless otherwise specified
Excludes 2 Russia participants without a predominant alcohol type
Excludes 4 Uganda participants without a predominant alcohol type
Alcohol type and plasma HIV viral load
In the unadjusted analysis in both countries, the HIV viral load varied according to the alcohol type predominantly consumed. Drinkers of wine/high-ethanol-content cocktails and liquor (Russia), and commercially- and locally-distilled spirits and locally-made beer (Uganda) tended to have higher levels than drinkers of beer/low-ethanol-content cocktails (Russia) or commercially-made beer (Uganda). Drinkers of fortified wine in Russia (n = 5) and wine in Uganda (n = 8) had substantially lower HIV viral load levels (Table 3).
Table 3.
Mean HIV viral load by alcohol type predominantly consumed among HIV infected drinkers in Russia and Uganda
| Predominant alcohol type | N | Mean HIV viral load (SD) (copies/ml) |
|---|---|---|
| Russia* | ||
| Beer/Low-ethanol content cocktails‡ | 162 | 161,178 (413,293) |
| Wine/High-ethanol content cocktails | 42 | 505,434 (1,183,791) |
| Fortified wine | 5 | 66,980 (113,228) |
| Liquor∞ | 49 | 286,239 (749,554) |
| No predominant alcohol type | 2 | 14,031 (4,242) |
|
| ||
| Uganda† | ||
| Commercially-made beer‡ | 215 | 42,712 (102,845) |
| Locally-made beer | 41 | 52,611 (95,054) |
| Wine | 8 | 4,549 (6,458) |
| Commercially-distilled spirit∞ | 35 | 112,245 (314,840) |
| Locally-distilled spirit∞ | 43 | 46,232 (79,833) |
| No predominant alcohol type | 4 | 2,595 (3,193) |
excludes n = 1 Russia participants missing viral load.
excludes n = 6 Uganda participants missing viral load.
Lowest ethanol concentration
Highest ethanol concentration (in Uganda, unclear whether locally distilled spirits considered to have the same concentration as commercially distilled spirits, but the former’s ethanol concentrations may be more variable).
In the adjusted analysis in Russia, those drinking wine/high-ethanol-content cocktails had higher log10 HIV viral load levels than those drinking beer/low-ethanol-content cocktails (β = 0.38, 95% CI 0.07 to 0.69, p = 0.02) (Table 4). There was no significant difference between those drinking beer/low–ethanol-cocktails and those drinking liquor or fortified wine. In the adjusted analysis in Uganda, compared to commercially-brewed beer drinkers, wine drinkers appeared to have lower HIV viral load levels (β = −0.65, 95% CI −1.36 to 0.07, p = 0.08) (Table 4). This result did not reach statistical significance, likely due to the small number of wine drinkers (n=8), but was consistent even after adding PEth concentrations to the model (β = −0.65 95% CI −1.36 to 0.07, p = 0.08). We did not find evidence of differences between those drinking commercially-brewed beer and those drinking spirits or the locally-brewed beers in the adjusted analysis in Uganda.
Table 4.
Unadjusted and adjusted mean differences in log10 HIV viral load level (coefficients of linear regression) and 95% confidence intervals by alcohol type among HIV-infected patients in Russia and Uganda.
| Variable | Unadjusted β (95% CI) | P | Adjusted* β 95% CI | P |
|---|---|---|---|---|
| Predominant alcohol type | ||||
| Russia (n=258)† | ||||
| Beer or low-ethanol cocktails‡ | Ref | Ref | Ref | Ref |
| Wine or high-ethanol cocktails | 0.42 (0.11 to 0.72) | 0.008 | 0.38 (0.07 to 0.69) | 0.018 |
| Fortified wine | −0.23 (−1.03 to 0.57) | 0.571 | −0.12 (−0.95 to 0.72) | 0.782 |
| Hard liquor (e.g., vodka)∞ | 0.04 (−0.25 to 0.33) | 0.783 | 0.00 (−0.30 to 0.30) | 0.989 |
|
| ||||
| Uganda (n=342)ε | ||||
| Commercially-brewed beer‡ | Ref | Ref | ||
| Locally-brewed beer | 0.17 (−0.17 to 0.51) | 0.327 | 0.07 (−0.27 to 0.42) | 0.672 |
| Wine | −0.58 (−1.30 to 0.14) | 0.115 | −0.65 (−1.36 to 0.07) | 0.076 |
| Commercially-distilled spirits∞ | 0.04 (−0.32 to 0.41) | 0.818 | −0.08 (−0.47 to 0.32) | 0.704 |
| Locally-distilled spirits∞ | 0.05 (−0.29 to 0.38) | 0.779 | 0.01 (−0.33 to 0.35) | 0.951 |
Adjusted for age, sex, asset index (or monthly individual-level income in Russia), number of standard drinks, frequency of drinking 6 or more drinks per occasion, and in Russia only, history of injecting drugs.
Excludes 1 participant without HIV viral load measurement and 2 participants without a predominant alcohol type.
Lowest ethanol concentration
Highest ethanol concentration (in Uganda, unclear whether locally distilled spirits considered to have the same concentration as commercially distilled spirits, but the former’s ethanol concentrations may be more variable).
Excludes 6 participants without HIV viral load measurements and 4 participants without a predominant alcohol type.
In a sensitivity analysis adding years since HIV diagnosis to the adjusted models, the results from models with and without years since HIV diagnosis were similar for all drink types in Russia. For example comparing drinkers of wine or high ethanol content to drinkers of commercial beer, β was 0.38 (95% CI 0.07 to 0.69) in the model without years since HIV diagnosis, and 0.38 (95% CI 0.07 to 0.70) in the model with years since HIV diagnosis. In Uganda, there was some attenuation in the observed association among wine drinkers (β = −0.44, 95% CI −1.15 to 0.27, p = 0.22) in the model with years since HIV diagnosis.
DISCUSSION
Alcohol consumers tend to prefer specific alcohol types [36]; their preferred type may expose them to varying health risks [12]. In particular, among HIV-infected individuals, alcohol types that are more likely to promote inflammation [37] could impact HIV disease progression [38]. In this analysis, we assessed the alcohol types used by ART-naïve HIV-infected patients enrolling into three separate cohort studies at HIV treatment clinics in Russia and Uganda. We evaluated the association between the alcohol type predominantly consumed in the one month (Russia) or three months (Uganda) prior to enrollment and the HIV viral load at enrollment. Notably, nearly all participants predominantly drank one alcohol type, supporting the presumption that alcohol consumers, including those with HIV infection, have preferred alcohol types.
In our analysis, we assessed whether those drinking higher-ethanol-content drinks would have higher viral loads. Alcohol types with high ethanol content might be more likely to cause inflammation irrespective of the overall amount of alcohol consumed. Some experimental studies suggest that ethanol concentrations above 30%, a threshold that is lower than the ethanol concentrations of most liquors, are more likely to damage biological membranes than lower concentrations [39]. In theory, such damage could promote increased microbial translocation, which in HIV-infected patients is associated with HIV disease progression [40]. In Russia, those consuming wine/higher-ethanol-content cocktails had significantly higher HIV viral load levels than those consuming beer/lower-ethanol-content cocktails; however, those consuming liquor, contrary to our expectation, did not have significantly higher viral loads. Similarly, we did not find a significant association between high ethanol content and HIV viral load in the Uganda sample.
Although wine may have considerably high ethanol levels, participants consuming wine might have lower HIV viral loads, since wine is believed to have some anti-inflammatory properties [12]. In Uganda, a small number (n=8) of participants who reported predominantly drinking wine had lower HIV viral loads than those drinking commercially-brewed beer. In Russia, a small number of participants (n=5) predominantly drank fortified wine, and had lower viral loads than those drinking beer/lower-ethanol-content cocktails. However, neither of these results were statistically significant (p = 0.08 and p = 0.78, respectively). Due to the small number of participants in both wine groups, further investigation is warranted. If confirmed, such an observation would be consistent with suggestions that fermented drinks like wine may be less likely to promote inflammation than other types of alcohol [12, 13] and thus may be least associated with risk of HIV disease progression. In Russia, drinkers of wine/higher-ethanol-content cocktails had significantly higher viral loads than those drinking beer/lower-ethanol-content cocktails. Unfortunately, wine and high-ethanol-content cocktails were combined during patient interviews in Russia and thus their associations with HIV viral load cannot be assessed separately. Consequently, this observation in Russia may be due to either the high ethanol content inherent in both high-ethanol-content cocktails and wine (perhaps less likely given the lack of higher viral load observed among liquor drinkers), or, it could be due primarily to unknown constituents in the cocktail drinks. Further studies assessing the cocktail drinks separately from wine and beer are thus recommended.
In the Uganda, sample, we expected that those consuming locally-made alcohols would have higher viral loads, possibly due to differences in methods of production or methods of distribution. We did not find evidence of differences in HIV viral load level by production method (i.e. whether the alcohol type was locally- or commercially-made).
We suggest that our results be interpreted with consideration given to the limitations in our data and study design. First, we were only able to assess associations with HIV viral load, which may be an imperfect marker of both HIV disease progression and the overall effects of inflammation in HIV-infected patients. Future studies should assess whether differences by alcohol type consumed exist in aspects or markers of HIV disease progression other than the viral load. For example, in Uganda, we recently found increased levels of monocyte activation among ART-naïve persons consuming unhealthy amounts of alcohol compared to those consuming lower amounts [41]. Whether such associations are driven by the consumption of specific alcohol types remains unknown.
Secondly, the number of participants consuming some alcohol types in our study was small (e.g., fortified wine in Russia and wine in Uganda). Also, as noted earlier, wine and high-ethanol-content cocktails were assessed together during interviews in Russia, as were beer and low-ethanol-content cocktails. As such, associations with HIV viral load could not be assessed separately for these types.
Thirdly, the estimated ethanol concentrations used in both countries may not be accurate. This may affect how the total number of standard ethanol drinks is estimated, leading to insufficient adjustment in our analyses. In Uganda, for example, locally-brewed beers are estimated to have similar ethanol concentrations as commercially-brewed beers, but the former has been reported to have higher ethanol concentrations [29].
We excluded participants who were suspected to be either HIV-negative or to have been already exposed to ART (based on either undetectable viral load results, or testing HIV-antibody negative), since this would have otherwise classified them as ineligible for these studies. This excluded a substantial number of participants (50 in Russia and 37 in Uganda); for those testing HIV-negative, their earlier results could possibly be due to false positive rapid tests [42]. While this affected our sample size, excluding these participants allowed us to better reflect those who were truly HIV-positive and not yet on treatment in both locations.
Finally, as our measurements of alcohol type, total number of standard drinks, and frequency of binge drinking were self-reported, social-desirability could lead to misreporting on any of the variables, and this could bias our estimates. In our previous study in this setting, we found only moderate correlation between self-reported measures of drinking and PEth [22].
The overall impact of these limitations could be that we missed differences in viral load levels by alcohol type consumed, especially in the adjusted analyses, even where such differences might actually exist. We suggest that future alcohol surveys in similar settings consider separate categories for locally- versus commercially-made alcohols, as well as different drink types (i.e. wine versus cocktails), as well as give due consideration to inherent ethanol concentrations in these drink types.
Despite these limitations, we adjusted for both the overall amount of ethanol consumed and the drinking pattern, suggesting that any observed associations are unlikely to be due to these two dimensions of alcohol-related risk. Assessing volumes in a beverage-specific fashion allowed us to identify drinkers of each alcohol type. In Uganda, our findings were consistent even when adjusting for PEth, a direct metabolite of alcohol, which may suggest that observed associations are unlikely to be mediated by alcohol per se but quite possibly something else that may be inherent in specific alcohol types.
Conclusions
Among HIV-infected drinkers not on ART enrolling into cohort studies in Russia and Uganda, we observed an association between the alcohol type consumed and the HIV viral load in the Russia sample. In Russia, those consuming wine/higher-ethanol-content cocktails had higher HIV viral load levels than those consuming beer/lower-ethanol-content cocktails. In Uganda, compared to those drinking commercially-brewed beer, those drinking wine appeared to have somewhat lower HIV viral load levels, but this was not statistically significant. Our observations suggest that in addition to assessing total number of standard drinks and drinking patterns, it may also be important to assess alcohol type when investigating alcohol use and HIV disease progression. We recommend further study to evaluate the possible link between alcohol type and clinical outcomes among HIV-infected patients, and what relevance this might have to interventions to reduce alcohol-related harm in this population.
Acknowledgments
Funding
Funding was provided by the NIH to: Judith Hahn (R01 AA018631, U01AA020776, K24 AA022586), Jeffrey Samet (U01AA020780, U01AA021989, U24AA020778), and Debbie Cheng (U24AA020779)
Footnotes
Compliance with Ethical Standards
Conflicts of interest: All authors declare that they have no conflicts of interest.
Ethical approval: This study involved human subjects. All procedures involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Informed consent: Informed consent was obtained from all individual participants included in the study.
Author contributions
SBA conceptualized the study, performed the analyses, wrote and edited the manuscript, and led the study. JH is the PI for ADEPT and BREATH cohort studies in Uganda; JS is the PI for the URBAN ARCH Russia cohort study. Both JH and JS provided mentorship to SBA, conceptualized the study, and wrote and edited the manuscript. RF, CLT, and GP, conceptualized the study, prepared and cleaned the data, assisted with data analysis, and wrote and edited the manuscript. NG, WM, NA, EK, DC, and AK, conceptualized the study, prepared the data, and wrote and edited the manuscript. All authors read and approved the final version of the manuscript.
References
- 1.Kalichman SC, Simbayi L, Jooste S, Vermaak R, Cain D. Sensation seeking and alcohol use predict HIV transmission risks: prospective study of sexually transmitted infection clinic patients, Cape Town, South Africa. Addictive behaviors. 2008;33(12):1630–1633. doi: 10.1016/j.addbeh.2008.07.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hahn JA, Woolf-King SE, Muyindike W. Adding fuel to the fire: alcohol’s effect on the HIV epidemic in Sub-Saharan Africa. Curr HIV/AIDS Rep. 2011;8(3):172–180. doi: 10.1007/s11904-011-0088-2. [DOI] [PubMed] [Google Scholar]
- 3.Baum MK, Rafie C, Lai S, Sales S, Page JB, Campa A. Alcohol use accelerates HIV disease progression. AIDS research and human retroviruses. 2010;26(5):511–518. doi: 10.1089/aid.2009.0211. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hendershot CS, Stoner SA, Pantalone DW, Simoni JM. Alcohol use and antiretroviral adherence: review and meta-analysis. Journal of acquired immune deficiency syndromes. 2009;52(2):180–202. doi: 10.1097/QAI.0b013e3181b18b6e. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Saitz R. Clinical practice. Unhealthy alcohol use The New England journal of medicine. 2005;352(6):596–607. doi: 10.1056/NEJMcp042262. [DOI] [PubMed] [Google Scholar]
- 6.Finucane MM, Samet JH, Horton NJ. Translational methods in biostatistics: linear mixed effect regression models of alcohol consumption and HIV disease progression over time. Epidemiologic perspectives & innovations : EP+I. 2007;4:8. doi: 10.1186/1742-5573-4-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Greenfield TK, Nayak MB, Bond J, Patel V, Trocki K, Pillai A. Validating alcohol use measures among male drinkers in Goa: implications for research on alcohol, sexual risk, and HIV in India. AIDS and behavior. 2010;14(Suppl 1):S84–93. doi: 10.1007/s10461-010-9734-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Razvodovsky YE. The effect of beverage type on alcoholic psychoses rate in Russia. Alcohol and alcoholism. 2015;50(2):200–205. doi: 10.1093/alcalc/agu104. [DOI] [PubMed] [Google Scholar]
- 9.Miguez-Burbano MJ, Lewis JE, Fishman J, Asthana D, Malow RM. The influence of different types of alcoholic beverages on disrupting highly active antiretroviral treatment (HAART) outcome. Alcohol and alcoholism. 2009;44(4):366–371. doi: 10.1093/alcalc/agp024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Shacham E, Agbebi A, Stamm K, Overton ET. Alcohol consumption is associated with poor health in HIV clinic patient population: a behavioral surveillance study. AIDS and behavior. 2011;15(1):209–213. doi: 10.1007/s10461-009-9652-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Dey M, Gmel G, Studer J, Dermota P, Mohler-Kuo M. Beverage preferences and associated drinking patterns, consequences and other substance use behaviours. European journal of public health. 2014;24(3):496–501. doi: 10.1093/eurpub/ckt109. [DOI] [PubMed] [Google Scholar]
- 12.Arranz S, Chiva-Blanch G, Valderas-Martinez P, Medina-Remon A, Lamuela-Raventos RM, Estruch R. Wine, beer, alcohol and polyphenols on cardiovascular disease and cancer. Nutrients. 2012;4(7):759–781. doi: 10.3390/nu4070759. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Palmieri D, Pane B, Barisione C, Spinella G, Garibaldi S, Ghigliotti G, Brunelli C, Fulcheri E, Palombo D. Resveratrol counteracts systemic and local inflammation involved in early abdominal aortic aneurysm development. The Journal of surgical research. 2011;171(2):e237–246. doi: 10.1016/j.jss.2011.07.041. [DOI] [PubMed] [Google Scholar]
- 14.Waldschmidt TJ, Cook RT, Kovacs EJ. Alcohol and inflammation and immune responses: summary of the 2006 Alcohol and Immunology Research Interest Group (AIRIG) meeting. Alcohol. 2008;42(2):137–142. doi: 10.1016/j.alcohol.2007.11.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Paine A, Davan AD. Defining a tolerable concentration of methanol in alcoholic drinks. Human & experimental toxicology. 2001;20(11):563–568. doi: 10.1191/096032701718620864. [DOI] [PubMed] [Google Scholar]
- 16.Ipp H, Zemlin AE, Erasmus RT, Glashoff RH. Role of inflammation in HIV-1 disease progression and prognosis. Critical reviews in clinical laboratory sciences. 2014;51(2):98–111. doi: 10.3109/10408363.2013.865702. [DOI] [PubMed] [Google Scholar]
- 17.Hunt PW. HIV and inflammation: mechanisms and consequences. Current HIV/AIDS reports. 2012;9(2):139–147. doi: 10.1007/s11904-012-0118-8. [DOI] [PubMed] [Google Scholar]
- 18.Azar MM, Springer SA, Meyer JP, Altice FL. A systematic review of the impact of alcohol use disorders on HIV treatment outcomes, adherence to antiretroviral therapy and health care utilization. Drug and alcohol dependence. 2010;112(3):178–193. doi: 10.1016/j.drugalcdep.2010.06.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Hahn JA, Samet JH. Alcohol and HIV disease progression: weighing the evidence. Current HIV/AIDS reports. 2010;7(4):226–233. doi: 10.1007/s11904-010-0060-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Williams EC, Hahn JA, Saitz R, Bryant K, Lira MC, Samet JH. Alcohol Use and Human Immunodeficiency Virus (HIV) Infection: Current Knowledge, Implications, and Future Directions. Alcoholism-Clinical and Experimental Research. 2016;40(10):2056–2072. doi: 10.1111/acer.13204. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Rubio A, Martinez-Moya M, Leal M, Franco JM, Ruiz-Mateos E, Merchante E, Sanchez-Quijano A, Lissen E. Changes in thymus volume in adult HIV-infected patients under HAART: correlation with the T-cell repopulation. Clinical and experimental immunology. 2002;130(1):121–126. doi: 10.1046/j.1365-2249.2002.01950.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Asiimwe SB, Fatch R, Emenyonu NI, Muyindike WR, Kekibiina A, Santos GM, Greenfield TK, Hahn JA. Comparison of Traditional and Novel Self-Report Measures to an Alcohol Biomarker for Quantifying Alcohol Consumption Among HIV-Infected Adults in Sub-Saharan Africa. Alcohol Clin Exp Res. 2015;39(8):1518–1527. doi: 10.1111/acer.12781. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.The World Health Organization. Global Status Report on Alcohol and Health 2014. Geneva, Switzerland: The World Health Organization; 2014. [Google Scholar]
- 24.The world Health Organisation. Summary country profile for HIV/AIDS treatment scale-up. Russian Federation; 2005. [Google Scholar]
- 25.THE HIV AND AIDS UGANDA COUNTRY PROGRESS REPORT. 2014 http://www.unaids.org/sites/default/files/country/documents/UGA_narrative_report_2015.pdf.
- 26.Muyindike WR, Lloyd-Travaglini C, Fatch R, Emenyonu NI, Adong J, Ngabirano C, Cheng DM, Winter MR, Samet JH, Hahn JA. Phosphatidylethanol confirmed alcohol use among ART-naive HIV-infected persons who denied consumption in rural Uganda. AIDS care. 2017:1–6. doi: 10.1080/09540121.2017.1290209. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Sundararajan R, Wyatt MA, Woolf-King S, Pisarski EE, Emenyonu N, Muyindike WR, Hahn JA, Ware NC. Qualitative study of changes in alcohol use among HIV-infected adults entering care and treatment for HIV/AIDS in rural southwest Uganda. AIDS Behav. 2015;19(4):732–741. doi: 10.1007/s10461-014-0918-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Hahn JA, Emenyonu NI, Fatch R, Muyindike WR, Kekiibina A, Carrico AW, Woolf-King S, Shiboski S. Declining and rebounding unhealthy alcohol consumption during the first year of HIV care in rural Uganda, using phosphatidylethanol to augment self-report. Addiction. 2016;111(2):272–279. doi: 10.1111/add.13173. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Hahn JA, Dobkin LM, Mayanja B, Emenyonu NI, Kigozi IM, Shiboski S, Bangsberg DR, Gnann H, Weinmann W, Wurst FM. Phosphatidylethanol (PEth) as a biomarker of alcohol consumption in HIV-positive patients in sub-Saharan Africa. Alcoholism, clinical and experimental research. 2012;36(5):854–862. doi: 10.1111/j.1530-0277.2011.01669.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Sobell LC, Sobell MB. Timeline Follow-back: A technique for assessing self-reported alcohol consumption. In: Litten RZ, Allen I, editors. Measuring Alcohol Consumption: Psychosocial and Biological Methods. New Jersey: Humana Press; 1992. pp. 41–72. [Google Scholar]
- 31.Sobell LC, Agrawal S, Annis H, Ayala-Velazquez H, Echeverria L, Leo GI, Rybakowski JK, Sandahl C, Saunders B, Thomas S, et al. Cross-cultural evaluation of two drinking assessment instruments: alcohol timeline followback and inventory of drinking situations. Substance use & misuse. 2001;36(3):313–331. doi: 10.1081/ja-100102628. [DOI] [PubMed] [Google Scholar]
- 32.Greenfield TK. Ways of measuring drinking patterns and the difference they make: experience with graduated frequencies. Journal of substance abuse. 2000;12(1–2):33–49. doi: 10.1016/s0899-3289(00)00039-0. [DOI] [PubMed] [Google Scholar]
- 33.Filmer D, Pritchett LH. Estimating wealth effects without expenditure data--or tears: an application to educational enrollments in states of India. Demography. 2001;38(1):115–132. doi: 10.1353/dem.2001.0003. [DOI] [PubMed] [Google Scholar]
- 34.Hahn JA, Fatch R, Kabami J, Mayanja B, Emenyonu NI, Martin J, Bangsberg DR. Self-Report of Alcohol Use Increases When Specimens for Alcohol Biomarkers Are Collected in Persons With HIV in Uganda. Journal of acquired immune deficiency syndromes. 2012;61(4):e63–64. doi: 10.1097/QAI.0b013e318267c0f1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Acuna E, Rodriguez C. Classification, Clustering, and Data Mining Applications. Springer; 2004. The treatment of missing values and its effect on classifier accuracy; pp. 639–647. [Google Scholar]
- 36.Lindgren KP, Westgate EC, Kilmer JR, Kaysen D, Teachman BA. Pick your poison: stimuli selection in alcohol-related implicit measures. Addictive behaviors. 2012;37(8):990–993. doi: 10.1016/j.addbeh.2012.03.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Imhof A, Woodward M, Doering A, Helbecque N, Loewel H, Amouyel P, Lowe GD, Koenig W. Overall alcohol intake, beer, wine, and systemic markers of inflammation in western Europe: results from three MONICA samples (Augsburg, Glasgow, Lille) European heart journal. 2004;25(23):2092–2100. doi: 10.1016/j.ehj.2004.09.032. [DOI] [PubMed] [Google Scholar]
- 38.Lane HC. Pathogenesis of HIV infection: total CD4+ T-cell pool, immune activation, and inflammation. Topics in HIV medicine : a publication of the International AIDS Society, USA. 2010;18(1):2–6. [PubMed] [Google Scholar]
- 39.Gurtovenko AA, Anwar J. Interaction of ethanol with biological membranes: the formation of non-bilayer structures within the membrane interior and their significance. The journal of physical chemistry B. 2009;113(7):1983–1992. doi: 10.1021/jp808041z. [DOI] [PubMed] [Google Scholar]
- 40.Sandler NG, Douek DC. Microbial translocation in HIV infection: causes, consequences and treatment opportunities. Nature reviews Microbiology. 2012;10(9):655–666. doi: 10.1038/nrmicro2848. [DOI] [PubMed] [Google Scholar]
- 41.Carrico AW, Hunt PW, Emenyonu NI, Muyindike W, Ngabirano C, Cheng DM, Winter MR, Samet JH, Hahn JA. Unhealthy Alcohol Use is Associated with Monocyte Activation Prior to Starting Antiretroviral Therapy. Alcoholism, clinical and experimental research. 2015;39(12):2422–2426. doi: 10.1111/acer.12908. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Baveewo S, Kamya MR, Mayanja-Kizza H, Fatch R, Bangsberg DR, Coates T, Hahn JA, Wanyenze RK. Potential for false positive HIV test results with the serial rapid HIV testing algorithm. BMC research notes. 2012;5:154. doi: 10.1186/1756-0500-5-154. [DOI] [PMC free article] [PubMed] [Google Scholar]