Alcohol risks are higher in HIV positive people and linked to increased mortality

Gareth Hardy, HIV i-Base

Alcohol may affect HIV positive people differently and be an independent factor related to all cause mortality, according to a paper from one of the largest US cohort studies. [1]

The paper, published in January 2016 in Drug and Alcohol Dependence, looked the relationship between alcohol use and physiological harm or mortality in the most comprehensive HIV study to date.

Amy Justice from the Yale School of Medicine worked with colleagues looking at medical records from the Veterans Aging Cohort Study (VACS), a large cohort of HIV positive and negative patients receiving care at the Veterans Health Administration (VHA) from 1997 to 2014. The study used medical records of mortality as well as two index-based assessments of alcohol consumption and clinical events.

The VACS Risk Index predicts hospitalisation, frailty and mortality among HIV positive and negative patients and includes measures of HIV-specific and general organ injury. [2] The index is a composite of weighted point values assigned to age, CD4 cell count, HIV RNA, haemoglobin, kidney function, HCV infection, and liver cirrhosis. A higher index for an individual indicates worse clinical markers.

The researchers also used the Alcohol Use Disorders Identification Test – Consumption (AUDIT-C), which is a tool to screen for alcohol use in large clinical populations.

AUDIT-C consists of 3 questions:

  • How often do you have a drink containing alcohol?
  • How many standard drinks containing alcohol do you have on a typical day?
  • How often do you have six or more drinks on one occasion?

Answers to these questions were scored 0-4 based on severity and summed for a total score ranging from 0-12. Heavy episodic drinking (HED) was assessed using a cut-off based on any positive response to AUDIT-C item 3, and its frequency was considered separately.

The primary aim of the study was to determine whether alcohol use was associated with all-cause mortality and physiologic injury as measured by the VACS Index from the date of the first AUDIT-C to death or July 31, 2014. The second aim was to determine whether physiologic injury was associated with different AUDIT-C scores according to HIV status. Women were excluded from the analysis because they made up only 3% of the cohort and have a markedly different sensitivity to alcohol.

The analysis included 18,145 HIV positive patients and 42,228 HIV negative individuals for whom the mean age was 52.5 years and 54.0 years respectively. HCV infection was more common in HIV positive (31%) than HIV negative individuals (16%, p<.001). Of the HIV positive people, 76% were on ART with undetectable viral load.

The median follow-up was 4.8 years (IQR: 3.6 to 5.3) with a mortality rate of 2.7 per 100 person-years among HIV positive and 1.8 among HIV negative individuals (p<0.001). Mortality rates increased together with AUDIT-C scores for both HIV positive and negative individuals and this difference became wider with higher AUDIT-C scores (interaction term from Cox PH model: p=0.02). A similar pattern was found between mortality and alcohol exposure when measured as the estimated total number of drinks per month or HED.

In a combined multivariate model of hazard ratios (HR), HIV was significantly associated with mortality (HR 1.35; 95%CI: 1.26-1.45), with a significant interaction for HIV and AUDIT-C (p=0.02). Age, smoking, and HCV infection were also significant predictors of mortality in this model.

Of note, the number of drinks per month and HED were important predictors of mortality, with a greater effect in HIV positive people. Mortality HRs were significant for HIV positive people who consumed an estimated 30 or more drinks per month compared to HIV negative people where this level was more than 70 drinks per month. Furthermore, in HIV negative individuals there was a slight decline in mortality risk for those who consumed 3-7 drinks per month compared to 1-2 drinks per month, suggesting a protective effect from light alcohol consumption. This effect was not seen in HIV positive people. In combined models both drinks per month and HED, over the previous 12 months, were independently associated with mortality, with a significant interaction term for HIV and drinks per month (p<0.001).

Physiologic injury, as measured by the VACS Index, increased together with AUDIT-C scores for HIV positive people, where AUDIT-C scores of 5 to 7 and 8 to 12 were significantly associated with injury compared to those with AUDIT-C scores of 1 to 3 (beta 0.47; 95%CI: 0.22 to 0.73; and beta 1.05; 95%CI 0.77 to 1.34, respectively). In contrast, only a score of 8 to 12 was associated with injury in HIV negative people (beta 0.29: 95% CI 0.16, 0.42). Similar relationships were observed between VACS Index and the estimated total drinks per month and HED for HIV positive people, with a protective effect seen for those consuming between 3 to 29 drinks per month in HIV negative individuals, which was not seen in HIV positive people.

These results show that mortality and physiologic injury are associated with moderate to low levels of alcohol consumption in HIV positive patients. Harm caused by alcohol occurred at lower levels of consumption in HIV positive compared with HIV negative people. The authors suggest that HIV positive patients consuming more than 30 alcoholic drinks per month (or one per day) are at increased risk of all-cause mortality and physiologic frailty. National recommended limits for alcohol consumption in the US are 14 drinks per week (two per day).

The association between harm and lower levels of alcohol use in HIV positive patients may be explained by two factors. Firstly, previous studies have reported that HIV positive people have higher blood alcohol levels given a unit exposure. [3] A second factor is that even modest alcohol use has been associated with poor adherence to antiretroviral therapy. [4]


  1. Justice AC et al. Risk of mortality and physiologic injury evident with lower alcohol exposure among HIV infected compared with uninfected men. Drug and Alcohol Dependence. Jan 2016. (16) 00045-4. (PDF)
  2. Akgun KM et al. Risk factors for hospitalization and medical intensive care unit (MICU) admission among HIV-infected Veterans. J. Acquir. Immune. Defic. Syndr. 2013. (62) 52-59 (PDF)
  3. McGinnis KA. Number of drinks to “feel a buzz” varies by HIV Status and viral load in men. AIDS Behav. 2015. 1-8.
  4. Braithwaite RS and Bryant KJ. Influence of alcohol consumption on adherence to and toxicity of antiretroviral therapy and survival. Alcohol Res. Health. 2010. (33) 280-287. (PDF)

Links to other websites are current at date of posting but not maintained.