Correspondence: Dr Seth C. Kalichman, Department of Psychology, 406 Babbidge Road, University of Connecticut, Storrs, CT 06269, USA. Tel: 860 208 3706; fax: 960 486 8706; e-mail: email@example.com
The success of antiretroviral therapy (ART) for treating HIV infection is now being turned towards HIV prevention. The Swiss Federal Commission for HIV/AIDS has declared that HIV-positive persons who are treated with ART, have an undetectable viral load, and are free of co-occurring sexually transmitted infections (STIs) should be considered noninfectious for sexual transmission of HIV. This study examined the implications of these assumptions in a sample of HIV-positive individuals who drink alcohol.
People living with HIV/AIDS (n = 228) were recruited through community sampling. They completed confidential computerized interviews and underwent monthly unannounced pill counts for ART adherence. HIV viral loads were obtained from medical records.
One hundred and eighty-five HIV-positive drinkers were currently receiving ART and 43 were untreated. Among those receiving ART, one in three were not virally suppressed and one in five had recently been diagnosed with an STI. Adherence was generally suboptimal, including among those assumed to be less infectious. As many as one in four participants reported engaging in unprotected intercourse with an HIV-uninfected partner in the past 4 months. There were few associations between assumed infectiousness and sexual practices.
Less than half of people who drank alcohol and took ART met the Swiss criteria for noninfectiousness. Poor adherence and prevalent STI threaten the long-term potential of using ART for prevention. In the absence of behavioral interventions, the realities of substance use and other barriers call into question the use of ART as prevention among alcohol drinkers.
Research in untreated HIV-serodiscordant couples has shown that sexual transmission of HIV is significantly less likely to occur when HIV is undetectable in peripheral blood plasma [1, 2]. Antiretroviral therapies (ARTs) effectively suppress HIV replication and have improved the health of people living with HIV/AIDS. Most ART regimens penetrate the genital compartment of the immune system and potentially reduce sexual infectiousness [3-6]. Accumulating evidence for the HIV prevention benefits of ART led the Swiss Federal Commission for HIV/AIDS to declare that HIV-infected persons should be considered noninfectious when peripheral blood plasma viral load is undetectable for at least 6 months and there are no co-occurring sexually transmitted infections (STIs). The so-called ‘Swiss Statement’ draws directly from studies of men showing that viral load in semen is suppressed and concordant with viral load in blood plasma when assumptions of treatment adherence, viral load and co-occurring STIs are met .
The Swiss Statement is now supported by the results of HIV Prevention Trials Network (HPTN) randomized clinical trial 052, which demonstrated a 96% reduction in HIV transmission following early initiation of ART compared with delayed ART . Early treatment also resulted in improved health for the HIV-infected partner. The World Health Organization (WHO) now recommends ART for HIV-positive sex partners in HIV-serodiscordant couples with CD4 counts of < 350 cells/μL to reduce HIV transmission. The WHO's position on HIV treatment as prevention is part of a more general movement calling for all HIV-infected partners in serodiscordant couples with CD4 cell counts between 350 and 500 cells/μL to qualify for ART as a means of curbing HIV transmission.
There are, however, several reasons to question the potential effectiveness of ART for HIV prevention in real-world settings. First, evidence garnered from prospective treatment cohort studies, such as the Swiss Cohort Study, and randomized trials, including HPTN 052, is contaminated by the intensive attention and clinical care provided to research participants which are rarely achieved in practice. Participants in these studies are carefully monitored for adherence, counselled to use condoms, and routinely tested and treated for co-occurring STIs. Selecting serodiscordant couples for research may also introduce biases towards more stable and monogamous relationships which can reduce the risks for co-occurring STIs. It is noteworthy that 11 of the 39 HIV infections observed in HPTN 052 were not genetically linked to the sex partner enrolled in the trial.
The assumptions underlying the Swiss Statement are often not realized in the lives of people with HIV/AIDS. Particularly important is the inference that sexual infectiousness mirrors blood plasma infectiousness. A review of 19 studies that investigated the concordance between blood plasma viral load and semen viral load found an average correlation of 0.44 . Thus, as little as 16% of the variance in sexual infectiousness can be discerned from knowing one's blood plasma viral load. One likely factor in the discordance between blood plasma and semen viral load is co-occurring STIs, because local inflammation spikes viral shedding in the genital tract. A review of 37 studies found a 12% median point-prevalence of confirmed STIs in people living with HIV/AIDS, and the most common STIs in people living with HIV were those that cause HIV shedding, specifically syphilis (median 9.5% prevalence), gonorrhoea (9.5%), chlamydia (5%), and trichamoniasis (18.8%) . Factors that impede ART adherence and increase risks for STIs will therefore undermine the use of ART as HIV prevention. Alcohol is among the most reliable impediments to ART adherence [11, 12] and among the most robust predictors of STIs in people living with HIV/AIDS . Alcohol use may therefore pose a particularly significant challenge to employing ART to prevent the spread of HIV.
In the current study, we tested the association between HIV treatment status and HIV transmission risk behaviours in a sample of HIV-positive drinkers. Based on previous research of risk compensation in response to HIV treatment [14, 15], we hypothesized that individuals receiving ART would demonstrate higher rates of HIV transmission risk behaviours compared with their untreated counterparts. In a second series of analyses focused on individuals receiving ART, we examined risk behaviours for persons meeting the Swiss Statement criteria for noninfectiousness. We hypothesized that persons who met the Swiss Statement assumptions for noninfectiousness would engage in more unprotected sex behaviours with HIV-uninfected partners, potentially offsetting the protective benefits of reduced infectiousness.
Participants and recruitment
People living with HIV/AIDS (n = 228) were recruited through community-based strategies that included both targeted clinic recruitment and snowball sampling techniques. Recruitment relied on responses to brochures placed in waiting rooms of AIDS service providers and infectious disease clinics throughout Atlanta, GA as well as an explicit systematic approach to word-of-mouth chain recruitment to extend our sampling. The study entry criteria were (1) 18 years of age or older, (2) HIV positive, and (3) drank any alcohol in the past week.
Data were collected through three sources; audio-computer-assisted structured interviews (ACASIs) [16, 17], monthly unannounced pill counts to monitor ART adherence, and chart-abstracted viral loads and CD4 counts.
Demographic and health characteristics
Participants were asked to state their gender, sexual orientation, age, years of education, income, ethnicity and employment status. We also asked participants about their histories of incarceration and substance abuse treatment. Participants self-reported whether they knew their most recent viral load test results and if so whether the result was detectable or undetectable.
To assess global alcohol use, we administered the Alcohol Use Disorders Identification Test (AUDIT), a 10-item scale designed to measure alcohol consumption and identify risks for alcohol abuse and dependence . The first three items of the AUDIT represent the quantity and frequency of alcohol use and the remaining seven items concern problems incurred from drinking alcohol. Scores of > 8 indicate a high risk for alcohol use disorders and problem drinking, with demonstrated specificities between 0.80 and 0.90 . In the current sample, the AUDIT was internally consistent (alpha = 0.90).
Participants reported whether they had been diagnosed with gonorrhoea, chlamydia, syphilis, genital herpes [herpes simplex virus (HSV) infection] or trichomoniasis. Participants who had been diagnosed with an STI were asked to give the approximate date of their last diagnosis. We used these dates to determine if the diagnosis had occurred within the previous 12 months of the assessment session. We used the same format to assess the occurrence of unexplained genital discharge to detect potentially undiagnosed STI symptoms. We did not, however, include nonspecific symptoms in our definition of having contracted an STI.
Sexual risk and protective behaviours
Participants reported vaginal and anal intercourse with and without condoms for HIV-seroconcordant and -serodiscordant partnerships in the previous 4 months . Participants were instructed to think back over the past 4 months and estimate the number of sex partners and number of sexual occasions on which they practiced each behaviour. Data were analysed within HIV-serodiscordant relationships, with individual behaviours examined as well as behaviours collapsed across unprotected and protected aggregates. In addition, we calculated the percentage of intercourse occasions on which condoms were used by taking the ratio (condom-protected vaginal + condom-protected anal intercourse)/(total vaginal + total anal intercourse).
Unannounced pill count medication adherence
Participants consented to monthly unannounced telephone-based pill counts for the duration of the study, constituting a prospective objective measure of adherence. Unannounced pill counts are reliable and valid in assessing medication adherence when conducted in participants' homes  and on the telephone [22, 23]. In this study we conducted unannounced cell-phone-based pill counts. Participants were provided with a free cell phone that restricted service to project contacts and emergency use (e.g., 911). Following in-office training in the pill counting procedure, participants were called at unscheduled times by a phone assessor. Pill counts occurred at 21- to 35-day intervals and were conducted for each of the antiretroviral medications participants were taking. Pharmacy information from pill bottles was also collected to verify the number of pills dispensed between calls. Adherence was calculated as the ratio of pills counted relative to pills prescribed, taking into account the number of pills dispensed.
Chart-abstracted viral loads and CD4 cell counts
We used a participant-assisted method for collecting chart-abstracted viral loads and CD4 cell counts from medical records. Participants were given a form that requested their doctor's office to provide results and dates of their most recent viral load and CD4 cell count measurements. These data were therefore obtained directly by the participants from their primary HIV care providers. The form included a place for the provider's office stamp or signature to assure data authenticity. Participants collected their chart-abstracted data prior to the initial assessment and again at the end of the study, proximal to the final assessment.
After written informed consent had been obtained, participants completed the baseline ACASI assessment, which required approximately 45 min. Participants were then provided with a study cell phone and instructed in its use. We trained participants in the steps required to compete the monthly unannounced pill counts. Viral loads and CD4 cell counts were collected at the start and end of the 1-year study. The ACASI was repeated at the 12-month follow-up. Participants were provided with US$20 cash reimbursements for each study activity completed. Data were collected between 30 November 2009 and 29 June 2011. The study was approved by the university Institutional Review Board.
Two series of analyses were performed to examine the associations between (a) receiving ART and sexual behaviour and (b) meeting the Swiss Statement criteria for noninfectiousness and sexual behaviour. To address the first question, participants who were being treated with ART were compared with those who were currently not treated. The second set of analyses focused only on those who were receiving ART. We partitioned participants into more and less infectiousness groups on the basis of the Swiss Statement criteria. Specifically, participants who either had been diagnosed with an STI during the 1-year follow-up period or were not virally suppressed at either time-point, as indicated in their medical records, were defined as not meeting the Swiss criteria, or assumed to be more highly infectious. Participants who were both not diagnosed with an STI and were virally suppressed at baseline and follow-up were defined as meeting Swiss criteria for noninfectiousness. We used 200 HIV RNA copies per mL of blood plasma to define undetectable viral loads. All analyses were conducted using logistic regression models to obtain odds ratios with 95% confidence intervals.
The total sample included 178 men and 50 women living with HIV/AIDS who were currently drinking alcohol; 55% reported drinking less than once a week, 16% drank two to three times a week, and 6% drank at least four times a week. Forty-three per cent of participants reported drinking one to two drinks during typical drinking days, 22% three to four drinks, and 12% five or more drinks. A total of 32% of participants had received substance use treatment in their lifetime, with 14% treated for substance use in the past year.
Receiving ART and sexual behaviour
One hundred and eighty-five participants were currently receiving ART and 43 were currently untreated. Table 1 shows the demographic and health characteristics of untreated and treated participants. Receiving ART was only associated with a lower likelihood of having been incarcerated in the previous year. As expected, treatment was associated with having an undetectable viral load at the start and completion of the study. Participants receiving ART were more likely to have been diagnosed with an STI in the past year, although this association only approached significance. In addition, receiving ART was associated with a greater likelihood of having HIV-negative and unknown HIV status partners (Table 2). This association was significant at time 1 and was a statistical trend at time 2. Importantly, treatment status was not significantly associated with engaging in unprotected intercourse with HIV-negative partners.
Table 1. Demographic and health characteristics of HIV-positive drinkers who were and were not receiving antiretroviral therapy (ART)
Not receiving ART (n = 43)
Receiving ART (n = 185)
AUDIT, Alcohol Use Disorders Identification Test; CI, confidence interval; OR, odds ratio; SD, standard deviation; STI, sexually transmitted infection.
Figure 1 shows the sample breakdown with respect to ART status, viral suppression and STI diagnoses. Among the 228 participants enrolled in the cohort, 43 (18%) were not receiving HIV treatment. The 12-month retention for the remaining 185 ART-treated participants was 91% (n = 170). Among those retained over the year, 67% (n = 114) had undetectable chart-abstracted viral loads at both baseline and 12-month follow-up, whereas 33% (n = 56) had detectable virus at either time 1 or time 2. In addition, 26 participants who were virally suppressed had been diagnosed with an STI over the 12-month follow-up period. Thus, 88 participants (51%) were virally suppressed and STI-free, meeting the Swiss Statement criteria for noninfectiousness, and were defined as being ‘less infectious’. The remaining 82 participants (49%) did not meet the Swiss Statement criteria for noninfectiousness and were defined as being ‘more infectious’; 42 had an STI during the 12-month follow-up period, 56 were not virally suppressed and 16 were not virally suppressed and had been diagnosed with an STI. The most common STI diagnoses over the 12-month follow-up period were genital herpes, syphilis and gonorrhoea. Among those diagnosed with an STI, 16 had multiple STI diagnoses during the follow-up period (Table 3).
Table 3. Demographic and health characteristics of participants defined as more and less infectious
With respect to demographics, participants defined as less infectious were significantly more likely to be married. As expected, less infectiousness was associated with a greater likelihood of a higher CD4 cell count at both time-points. There were no other differences between groups for demographic, health or substance use characteristics. With respect to viral load awareness, participants in both infectiousness groups were not always aware of their viral load and it was common for self-reported viral load to be discrepant from the viral load abstracted from medical records. More than one in four participants at time 1 reported having a viral load that was discrepant with their most recent medical records viral load. However, at time 2 significantly more less infectious participants reported having an undetectable viral load than did the more infectiousness group and the number of discrepant self-reported and chart-abstracted viral loads was less than observed at time 1.
Infectiousness and ART adherence
Adherence monitored by unannounced pill counts for the two infectiousness groups is shown in Table 4. Overall, participants defined as more infectious demonstrated poorer ART adherence than those in the less infectious group. Nearly half of more infectious participants were < 85% adherent. During the typical month, the average adherence for more infectious participants was > 80% of pills taken. However, adherence among the less infectious group was uneven over the year of observation, with more than one in four participants failing to achieve 85% of pills taken each month.
Table 4. Monthly unannounced pill count antiretroviral therapy (ART) adherence among participants defined as more and less infectious
More infectious (n = 82)
Less infectious (n = 88)
< 85% adherent
< 85% adherent
CI, confidence interval; OR, odds ratio; SD, standard deviation.
The majority of participants were currently sexually active at times 1 and 2. As shown in Table 5, it was common for participants to report more than one sex partner in the previous 4 months at both time-points. In addition, nearly half of participants reported HIV-negative or unknown HIV status sex partners during those time periods, with as many as half of those with uninfected partners indicating unprotected anal or vaginal intercourse. Between 30 and 60% of intercourse occasions with HIV-uninfected sex partners were not protected by condoms. The only significant difference between assumed infectiousness groups was for condom use with uninfected partners at time 1; less infectious participants reported less condom use with uninfected partners.
Table 5. Sexual behaviours at time 1 and time 2 among participants defined as more and less infectious
More infectious (n = 82)
Less infectious (n = 88)
CI, confidence interval; OR, odds ratio; SD, standard deviation.
*P < 0.05.
Sex partners [n (%)]
HIV-negative partners [n (%)]
Unprotected sex with HIV-negative partners [n (%)]
Unprotected sex with HIV-negative partners [n (%)]
% condoms with HIV-negative partners [mean (SD)]
The current study of HIV-positive alcohol drinkers found few differences between those receiving and not receiving ART. Overall, individuals receiving ART were more likely to report sexual partners who were not HIV positive. However, there were no differences in their likelihood to engage in unprotected intercourse with those partners. Among persons receiving ART, one in three were not virally suppressed and one in five had been diagnosed with an STI during the study. In addition, 22% of participants who were virally suppressed during the study had been diagnosed with an STI, increasing their likelihood of sexual infectiousness regardless of having an undetectable viral load in their blood plasma. Poor adherence was observed in nearly one-third of those defined as less infectious, calling into question their assumed infectiousness over time.
Findings from the current study indicate that as many as one in three people living with HIV who drink alcohol and are taking ART engage in unprotected anal or vaginal intercourse with HIV-uninfected sex partners. Rates of serodiscordant unprotected sex were similar for persons defined as less and more infectious. Thus, suboptimal adherence, multiple sex partners, and inconsistent condom use were common in this cohort of HIV-positive alcohol drinkers, suggesting a high risk for increased infectiousness over time. Alcohol-related problems did not differentiate groups, suggesting that lighter and heavier drinkers may not differ in their meeting the assumptions of infectiousness. Research with larger samples of heavy drinkers is needed to confirm these associations. These findings highlight the limitations of applying the Swiss Statement for noninfectiousness to HIV-positive drinkers, and perhaps others, receiving ART. Because alcohol use is common among people living with HIV/AIDS  and alcohol use is prevalent in developing countries , the cautions raised by the current findings may have broad implications for the use of HIV treatment with ART as prevention.
The Swiss Statement unambiguously specifies the circumstances under which a person with HIV should be considered noninfectious. Unfortunately, these conditions do not reflect the realities of many individuals living with HIV/AIDS. The Swiss Statement advises providers to inform their patents that it is no longer necessary to worry about infecting others with HIV when the conditions of noninfectiousness are met. The Statement emphasizes the potential ‘liberating effect’ on patients who are told they are no longer infectious, and the importance of sensitizing patients to the symptoms of STIs. However, the majority of empirical support for the Swiss Statement comes from research with heterosexual couples in southern Africa [25, 26], which may not apply equally to men who have sex with men who practice anal intercourse. The HIV transmission dynamics of anal sex differ from those of vaginal sex in relation to infectivity. Unfortunately, there is also a wide gap between the applicability of these recommendations and what we know about HIV-positive persons who drink. Drinkers may be less likely to follow through with medical recommendations and have less diligence in their health care. Further complicating matters is the issue that individuals who believe they are less infectious because they are receiving ART may relax their use of condoms and actually be at greater risk for co-occurring STIs. Research from the Swiss Cohort Study supports these concerns, finding that after the publication of the Swiss Statement HIV-positive persons were more likely to report unprotected sex with stable sex partners if they were treated with ART and their HIV was suppressed . The conditions of the Swiss Statement should therefore be considered an ideal that will remain difficult to achieve in the context of substance use, mental health problems, poverty, and other significant barriers to adherence and facilitators of contracting STIs.
These findings should be interpreted in light of the study limitations. First, the sample was one of convenience and cannot be considered representative of people living with HIV infection. We recruited from multiple clinical services as well as through participant referrals. We cannot therefore know the range of care and clinical services that participants were receiving. In addition, participants were taking a variety of ART combinations and for variable lengths of time. The study also relied on self-report instruments to assess sexual behaviours, substance use, and other participant characteristics. While each of these measures was collected using state of the science procedures, they may still be subject to biases. Self-reported viral load results were often discrepant with medical records, offering a sense of the limitations of self-reported data in general. Socially sensitive behaviours such as sex and substance use assessed by self-report may be underreported, suggesting that rates of risk practices in this study should be considered lower-bound estimates, with actual risks possibly even higher. Another limitation of our study was our definition of nonadherence applied to all medication regimens, which differ in their demand for optimal adherence [28-30]. With these limitations in mind, we believe that the current study results have important implications for efforts to use ART to reduce HIV transmission risks.
The remarkable success of HIV treatments in improving the health and well-being of people living with HIV/AIDS is being extended to preventing new HIV infections. Using ART as prevention is successful in the context of perinatal transmission and stands to reason when applied to sexual transmission risks. However, poor concordance between blood plasma and genital tract viral loads calls into question the optimism underlying the use of ART as prevention. These doubts are amplified for alcohol drinkers. Alcohol use is a double threat to the efficacy of ART as prevention because it impedes treatment adherence and increases risks for co-occurring STIs. Scaling up ART as prevention requires that the same amount of attention be given to behaviour in the general HIV-infected population as in participants in research cohorts and trials that have deemed the strategy effective. Under-resourcing of substance use treatment, routine medical care, and behavioural counselling will undermine the potential for HIV treatment with ART to prevent HIV infections.
This research was supported by an America Reinvestment and Recovery Act (ARRA) Challenge Grant from the National Institute of Alcohol Abuse and Alcoholism (NIAAA) RC1AA018983.