Predictors of viral suppression and rebound among HIV-positive men who have sex with men in a large multi-site Canadian cohort

Gay, bisexual and other men who have sex with men (MSM) have the highest prevalence of HIV in Canada [1]. Between 1985 and 2011, 54.7 % of diagnosed HIV cases with known exposure status in Canada were attributable to MSM (n = 69,856), even though self-identified MSM comprise only an estimated 2.1 % of the Canadian population [1, 2]. In the first two decades of the epidemic, this disproportionate burden was characterized by premature mortality across MSM communities [3], with the estimated life expectancy of gay men in some urban environments being 8 to 20 years less than that of the general male population [4]. Since the implementation of combination antiretroviral therapy (cART), people living with HIV/AIDS (PHAs) have experienced significant improvements in health outcomes and can now achieve life expectancy near that of the general population [5–7]. High levels of adherence to cART, usually defined as taking >95 % of prescribed medication [8], usually results in full suppression of HIV-1-RNA levels in plasma, markedly improving health outcomes and similarly reducing the risk of HIV transmission [9]. Adherent patients generally achieve viral suppression between 8 and 24 weeks after initiating treatment [10].

Despite the proven clinical benefits of cART, Canadian MSM continue to experience a sustained rate of new HIV infections compared with the general population [1, 11]. A myriad of factors, including the heightened risk of infection via anal sex compared with vaginal sex [12]; behavioural factors, such as condomless anal intercourse and substance use [13]; and social factors, including homophobia, stigma, and social exclusion [11, 12, 14] have been advanced to explain these high rates of new infections. Additionally, structural barriers such as criminalization of HIV exposure [15], insufficient access to culturally appropriate health services and distrust of available health care providers [16–18], may deter some MSM from seeking HIV testing and treatment, as well as negatively impact retention rates. Continuation of treatment is crucial for long-term clinical success and prevention of viral rebound [19]. Failure to remain virally suppressed increases the risk of poor health outcomes, such as HIV drug resistance, progression to AIDS, premature death, and HIV transmission [20, 21].

While risk factors for HIV seroconversion among MSM have been largely explored, there is a sizable research gap regarding modern-day cART treatment responses. To date, there has not been an analysis of the clinical and social circumstances associated with virologic outcomes among MSM in Canada. The Public Health Agency of Canada has specifically identified this knowledge gap [1]. Furthermore, viral suppression is a principal component of the new UN 90-90-90 Target (i.e., to achieve 90 % of PHA diagnosed globally, 90 % of them on treatment, 90 % of them on cART virally suppressed by 2020) [22]. This ambitious plan to end AIDS as a global pandemic provides a timely and pertinent framework for assessing where Canadian MSM on cART stand with regard to this target. The purpose of this study was to identify socio-demographic and clinical correlates of treatment response among MSM in Canada, as measured by viral suppression and subsequent virologic rebound. This will help inform cART retention strategies for MSM living with HIV.

In bivariate analysis, time to viral suppression was significantly associated with IDU history (p < 0.001), calendar year of cART initiation (p < 0.001), initial composition of ARV regimen (p < 0.001), baseline viral load (p < 0.001), baseline CD4 count (p = 0.001), HCV co-infection (p = 0.001), age (p = 0.002), ethnicity (p = 0.002), having an ADI at baseline (p = 0.006), and province of residence (p = 0.011).

Close to 90 % of HIV-positive MSM receiving cART in CANOC achieved viral suppression within a median of 5 months of initiating treatment [27]. However, nearly one-in-eight HIV-positive MSM who achieved viral suppression on cART experienced viral rebound at some point following suppression during follow-up. This is the largest longitudinal, multi-provincial analysis of HIV treatment responses among MSM in Canada and demonstrates that MSM are close to UNAIDS’ proposed targets that 90 % of individuals on cART achieve viral suppression. This high rate suppression will ensure the long-term health of these men and limit new infections in this population.

The proportion of MSM achieving suppression in CANOC is similar to other studies from several international settings where the HIV burden is also predominantly concentrated among MSM. A large North American cohort study found that between 2001 and 2009, the cumulative incidence of 1-year viral suppression was 84 %, observing that MSM participants had a higher likelihood of achieving suppression compared with other male participants [25]. Between 2010 and 2012 in Australia, it was estimated that the proportion of HIV-positive MSM on cART with undetectable viral loads increased from 85 to 90 % [28]. Surveillance data from the United Kingdom (UK) suggests that 89 % of MSM who initiated cART in 2009 were virally suppressed by 2010 [29]. Unfortunately these international viral suppression estimates for large HIV-positive MSM populations do not provide follow-up regarding episodes of subsequent rebound. But, in a large UK-based cohort study of previously antiretroviral naïve PHA on cART, of which approximately half were MSM, 11 % experienced rebound [30], which is a similar proportion to MSM who experienced rebound in our analysis.

Consistent with previous research [31], our analysis found more recent calendar year of cART initiation to predict better treatment response within a large HIV-positive cohort. This may partially be explained by the simple fact that ARV regimens in more recent years have enhanced drug efficacy and reduced side effects, promoting more optimal treatment response. As part of our analysis, we also stratified antiretroviral regimens by era of cART initiation and found that in earlier calendar years, a significantly (p < 0.001) higher proportion of participants initiated treatment comprised of a Zidovudine and Lamivudine NRTI fixed combination (51, 23, and 1 % of participants from 2000 to 2003, 2004 to 2007, and 2009 to 2012, respectively). In contrast, the better-tolerated and more effective [32] Tenofovir and Emtricitabine fixed-dose regimen was the predominant baseline NRTI combination in later calendar years (0, 25, and 82 % of participants initiating cART from 2000 to 2003, 2004 to 2007, and 2009 to 2012, respectively). Furthermore, in terms of ARV drug class, regimens comprised of NNRTIs were more effective than either boosted or unboosted PI-based for the treatment of patients with no previous exposure to antiretroviral therapy. This is congruous with a meta-analysis examining ARV regimens among HIV-infected patients with limited or no previous exposure to antiretroviral therapy [33]. NNRTIs were also most prevalent in more recent calendar years, with 49, 33, and 51 % of participants initiating a regimen comprised of NNRTIs from 2000 to 2003, 2004 to 2007, and 2009 to 2012, respectively. More frequent rebound episodes, especially in British Columbia and in earlier calendar years, may also reflect physician-recommended structured treatment interruptions, which were more common during earlier cART eras and in British Columbia [34].

The low median baseline CD4 cell count among participants (230 cells/mm³) is concerning, although guidelines for the timing of cART initiation have changed significantly over the period of study. The median baseline CD4 cell count increased throughout the study timeline, likely in response to these updated guidelines. In 2000, the median baseline CD4 cell count among all participants was 183 cells/mm³, whereas in 2011, this measure was 358 cells/mm³. Although an improvement, other studies have demonstrated that PHA who initiate cART with CD4 cell measures below 350 cells/mm³ are at higher risk of AIDS and death than individuals who begin treatment at higher CD4 thresholds [35–37]. Recurrent viral load testing among PHA who started cART with lower CD4 counts is crucial for informing changes in ARV regimens that are necessary for preventing treatment failure and adverse health outcomes.

Our finding that higher baseline CD4 cell count was a significant predictor of viral rebound was surprising [38, 39]. It is possible that cART adherent MSM with high baseline CD4 cell counts retained comparatively better health throughout the study timeline, which may have influenced them to “take a break” from their medication [40, 41]. However, recent research set in the Canadian context has found no association between high CD4 counts (≥500 cells/mm³) and lower rates of adherence [42]. Going forward, it is important to examine this issue more closely. Recommendations for initiating cART immediately, regardless of CD4 cell count, emerged after our period of study, first in 2013 for select populations [10], and in 2015 for all PHA [37, 43]. As such, the proportion of PHA initiating cART at higher CD4 counts will increase in the future, providing an opportunity to obtain a better understanding of its predictive value regarding cART adherence.

This analysis discerned a few populations at risk of poorer virologic outcomes, including younger MSM and those with a history of injection drug use. Adverse treatment response was particularly evident in this latter group, who had 0.71 times the likelihood of achieving suppression, and over 2.5 times the chance of experiencing rebound. Their reduced probability of achieving and sustaining undetectable viral loads hinders UNAIDS suppression targets and has potential negative implications for future health outcomes, and on going HIV transmission. Younger MSM can face a wide array of challenges, including sexual identity issues, substance abuse, precarious employment, and housing instability [44]. Understandably, these conditions may complicate balancing a lifelong cART regimen and coming to terms with having a chronic and stigmatized disease. Health care providers with younger HIV-positive MSM patients should become more knowledgeable about local LGBT-focused hotlines, agencies, and media so that they can better connect them to these services, which may help improve HIV treatment self-efficacy and facilitate improved cART adherence [45]. Furthermore, HIV case management services should accommodate more flexible hours to promote continued engagement in care, as younger HIV-positive MSM are more likely to miss scheduled appointment times, and provide more flexible scheduling as this been has previously demonstrated to improve clinical care attendance and HIV treatment response among this population [46]. For improving treatment responses among MSM who also use injection drugs, more intensive case management and psychosocial support services have a strong record of improving retention in care among more vulnerable HIV-positive populations [47, 48]. For example, incorporating harm reduction services such as supervised injection facilities into HIV care settings and increasing the availability of methadone maintenance therapy may also lead to more consistent viral load suppression among MSM-IDU [49, 50].

Our results show that over 80 % of HIV-positive MSM with at least one year of follow-up achieved and maintained suppressed HIV viral levels. However, 12 % of MSM who achieved viral suppression went on to experience viral rebound within the study period. To better realize the UNAIDS 90 % suppression target, further strategies are required to optimize cART outcomes in MSM in Canada, specifically targeting younger MSM and those with a history of IDU.