Background
Alcohol abuse has been recognized as an important public health problem, common and frequently undiagnosed among people living with HIV/AIDS (PLWHA) and is amenable of intervention [
1,
2]. Several studies have examined the prevalence of alcohol abuse and its associated factors among PLWHA. The data available are the result of research conducted within very different populations and regions, using different methodologies. Alcohol use disorders among PLWHA seems to be 2–4 times higher than those disorders in the general population [
3‐
5], and it is estimated that 40 to 50% of these individuals have a history of alcohol abuse or dependence [
1,
6].
This association can generate a higher disease burden for this group of individuals [
7‐
12]. Among the negative effects related to high alcohol consumption is its ability to interfere with the immune system [
13‐
15], predisposing the individual to the increased occurrence of bacterial infections, particularly tuberculosis [
16,
17], enhancing the liver damage caused by co-infection with Hepatitis C Virus [
18] and altering the metabolism of antiretroviral drugs [
19,
20]. Alcohol consumption is also associated with harmful behavior such as smoking, illicit drug use and the adoption of unsafe sexual practices [
21]. Although not always conclusive, several studies have suggested that alcohol consumption is also related to a lower adherence to highly active antiretroviral therapy (HAART) with consequent inadequate viral suppression, the potential emergence of viral resistance and treatment failure [
22‐
24].
Certain factors affecting alcohol abuse and the consequences that such behaviors can have on the evolution of the disease are amenable to intervention. The identification of these factors may be useful in the implementation of actions aimed at controlling alcohol use and providing relevant information for detecting individuals at a higher risk of adverse clinical outcomes.
Another aspect of interest is the effect alcohol can have on certain clinical and laboratory parameters, which may be important in the evaluation of PLWHA and disease progression. Studies suggest that in these individuals, alcohol consumption may be associated with lower CD4 cell counts; however, this finding is not consistent in the literature [
23,
25‐
33].
Therefore, the aim of this study was to analyze, in a sample of PLWHA in southern Brazil, the prevalence of alcohol abuse using the Alcohol Use Disorders Identification Test (AUDIT), to determine the risk factors associated with this alcohol abuse, and finally to study the independent effect of alcohol use on CD4 cell counts and HIV VLs.
Discussion
This study found a high prevalence of alcohol abuse among PWLHA that exceeded the values found in the general population. The most important factors associated with alcohol abuse were male gender, mixed or black skin and the use of illicit drugs (inhaled and/or intravenous). In contrast, patients in the 4th income quartile and with more education had a reduced risk. The hypothesis that alcohol can alter laboratory parameters was confirmed by observing that the number of CD4 cells decreased linearly as the AUDIT risk zone increased. Supporting this hypothesis, the linear regression revealed that every increase of one unit on the AUDIT score reduced the CD4 cell count by more than five cells/mm3.
Studies show that alcohol consumption among PWLHA is frequent and may be higher than in the general population. In our study, the prevalence of alcohol abuse was more than three times higher than that previously observed in the same region for the general population [
41]. In 2014, the United States of America (USA) reported a prevalence of alcohol use disorder of approximately 7% in the population aged 18 or older [
42]; furthermore, several USA studies show high alcohol consumption among PLWHA, where 8–30% of subjects were classified as “heavy drinkers” [
3,
27,
31,
43]. In other countries, the prevalence of alcohol abuse among PLWHA evaluated with the application of the AUDIT questionnaire showed similar or even much higher rates than those found here. These rates are affected by the habits and customs of the region in which each study is carried out [
28,
37,
44]. The consistently higher prevalence of alcohol abuse among PLWHA that is observed in different geographic areas, cultures and socioeconomic conditions provides evidence that this behavior can be contextualized to the condition of seropositivity and possibly to behavioral aspects.
The probability of the outcome was nearly three times higher among males. This finding is consistent with biological, behavioral and social aspects determining that men drink more alcohol and are at a greater risk of abusing the substance [
45]. Socioeconomic factors were also identified as important risk/protection factors for abuse; this outcome was more prevalent among individuals with mixed and black skin and less prevalent among individuals with a higher income and/or education. Other studies conducted both in Brazil and in other countries have also found that although the frequency of alcohol use is higher among more favored economic classes, alcohol abuse occurs more often among individuals of the less privileged social classes [
46‐
49]. Regarding skin color, a gradient was observed in which the risk of alcohol abuse increased as the skin darkened. This fact can be explained by socio-economic aspects because there is a clear association in the population between skin color, economic class and education level [
50]. This consistency of findings points to inequalities in health. Individuals with higher socioeconomic status and more education are more protected from alcohol abuse. This may be related to these individuals’ ability to better cope with problems and stressful situations due to their easier access to information and health services, which takes them to adopting a healthier lifestyle [
51‐
53].
The factors associated with alcohol abuse in this study group are similar to those found in the general population [
47], indicating that socioeconomic and demographic factors really are macro determinants of alcohol abuse and that the condition of seropositivity is part of a context that already predisposes an individual to this behavior.
We found that the AUDIT score can predict CD4 cell count, and therefore, individuals with high scores, may have very low cell counts. As the R2 of the model was low, other explanatory variables are influencing the variation of CD4 cell count. Regarding to VL, no association was found.
Studies evaluating the effects of alcohol consumption on these HIV markers have not shown consistent results. Among PLWHA not using HAART, some researchers have found an inverse association between alcohol consumption and CD4 cell counts [
30,
31], whereas others have failed to show such results [
26,
29,
33] or have only found it for heavy drinkers [
27]. The same effect occurred when individuals receiving HAART were evaluated. Although some researchers have shown lower CD4 cell counts [
25,
30,
31,
33,
54], others have found no significant differences [
26,
27,
29,
55‐
58]. This inconsistency in findings therefore underscores the need for further studies on the subject. If we consider as valid the hypothesis that there is an inverse relationship between alcohol consumption and CD4 cell counts, the reasons for this phenomenon occurring appear to be multifactorial and related to immunological, toxic, nutritional and behavioral effects [
59].
Studies evaluating the effect of alcohol consumption on HIV VL among individuals who were not receiving HAART found no association [
27,
31,
33]; however, among individuals receiving HAART, higher VL values or even viral suppression failure have been demonstrated [
2,
30,
31,
33,
56,
60]. This finding appears to be, at least in part, a reflection of lower adherence to HAART [
2,
30,
33]. In the present study, no association was established between alcohol consumption and HIV VL, even when individuals who had undergone HAART for more than 6 months were selected.
This lack of consistency in the relationship between alcohol and HIV markers should be addressed with the implementation of new prospective studies. Using better definitions of the pattern and volume of alcohol consumed with control for cofactors, such as the length of HIV infection, the use of HAART and treatment adherence, type and amount of illicit drugs used, could facilitate a better understanding of this relationship. However, the findings of this study provide evidence that the use of the AUDIT score in clinical practice could be useful in predicting, at least in part, changes to the CD4 cell count in these individuals. Furthermore, the identification of alcohol abuse should prompt a better evaluation and counseling for abandoning or reducing this habit and consequently, prevent or minimize its adverse effects on the health of the individual.
Study limitations
The limitations of this study include its cross-sectional design because this does not allow establishing a causal relationship, particularly between alcohol use and HIV markers. In the case of associated factors, in which the temporality between certain factors (such as age, gender and skin color) and the outcome appears to be maintained, reverse causality is less likely, despite the use of a cross-sectional design. Another limitation is the possible lack of statistical power for some of the studied associations. The AUDIT score was the alcohol exposure measure used to evaluate the effect on HIV markers, however the amount and frequency of consumption were not quantified. The positive result regarding CD4 cells cannot therefore be interpreted exclusively as a biological effect, but it should be considered that other mechanisms, such as behavioral mechanisms, may somehow also affect this marker. However, the effect remained even after adjusting for variables clearly related to the CD4 cell level, such as the disease duration, the use of HAART and the use of illicit drugs, which reinforces the association found. Finally, individuals who did not participate in the study could have affected the results because non-participation was related to non-attendance at the service and not to the refusal to participate. These individuals may have had lower adherence to HAART and may be major consumers of alcohol and illicit drugs, in which case the prevalence of alcohol abuse would have been underestimated and likewise the effect of risk factors. However, the percentage of losses did not exceed 5%, which minimizes the possibility of this selection bias.
Acknowledgments
The authors wish to acknowledge the cooperation and support provided by the staff of the HIV/AIDS Service of FURG University Hospital during the course of this research. We thank Dr. Lulie Rosane Odeh Susin that critically reviewed this manuscript.