Introduction

HIV remains a complex public health problem worldwide1. In 2022, approximately 40 million individuals worldwide were identified as people living with HIV (PLHIV)2. In Latin America, 120,000 people were newly infected with HIV in 2019, a 21% increase since 20103. Conversely, some high-income countries have recently reported significant reduction of HIV4,5. This success has been attributed to their concerted efforts in developing effective prevention tools, including early diagnosis4,5.

Early diagnosis of HIV infection facilitates the timely initiation of antiretroviral therapy (ART). This approach significantly decreases the circulating virus concentration, resulting in viral suppression and reduced transmission6. In 2021, the Joint United Nations Program on HIV/AIDS, the World Health Organization (WHO), and several signatory countries reviewed previously established objectives and proposed 95-95-95 targets2. These targets stipulate that 95% of PLHIV will be diagnosed, 95% of those diagnosed will receive ART, and 95% of those receiving ART will achieve a suppressed viral load (SVL) by 20302. The HIV test-and-treat cascade7 enables the monitoring of these targets. It is a strategy for following up with PLHIV in health services, encompassing five stages: diagnosis of HIV infection, linkage to care, retention in care, use of ART, and viral suppression8,9,10.

These targets present challenges, especially among key populations disproportionately affected by the HIV epidemic, such as men who have sex with men (MSM). These challenges are exacerbated in Latin America, where significant inequalities create social and structural barriers that affect the effectiveness of health services11. These countries reported 110,000 individuals newly diagnosed with HIV in 2022, accounting for 10% of all new HIV infections worldwide2, and an HIV prevalence of 9.5% in MSM in 2018–2022, which is nearly 20-fold higher than that in adults in the general population aged 15–49 years12. In Brazil, the HIV infection rate increased by 17.2% from 2020 to 202213. Furthermore, new HIV infections have been detected principally among MSM (52.6%) over the last two decades13.

The ongoing violation of the social rights of MSM, partly due to stigma and discrimination14, hinders access to HIV testing and treatment. Additional barriers include lack of trust in partners, fear of breach of confidentiality, and lack of information15,16.

Thus, this study aimed to analyze the HIV test‒and‒treat cascade in the Brazilian MSM population in 12 state capitals and to identify the factors associated with the stages of this cascade.

Methods

Study site and type

A survey among MSM from 12 capitals in the five administrative regions of Brazil was conducted from June to December 2016. The surveyed cities were Manaus and Belém (northern region); Fortaleza, Recife, and Salvador (northeast); Brasília and Campo Grande (central-western); Belo Horizonte, Rio de Janeiro, and São Paulo (southeast); and Curitiba and Porto Alegre (south). The study was conducted in public health system clinics, except in Belo Horizonte and Salvador, where rented rooms were used. Each location included reception and waiting areas, interview spaces, and private rooms for testing and counseling.

Study population and sample size

The sample comprised MSM aged ≥ 18 years who reported engaging in oral or anal sex with another man in the past 12 months and who lived, studied, or worked in the selected capital. The participants signed separate informed consent forms (ICFs) for the questionnaire and biological tests, with the option of decreasing participation in the tests. Individuals under the influence of drugs or alcohol or who were identified as transwomen were excluded.

The sample size calculation for RDS is the same as that for other studies, and a design effect (DE) is recommended. DE refers to the ratio between the actual and expected variances in comparison to simple random sampling. On the basis of recommendations from Wejnert et al.17 and using the HIV prevalence rates estimated in a previous study18, we calculated sample sizes for the 12 cities ranging from 204 to 474. Ultimately, the donors determined a maximum sample size of 350 per city. Most of the sites achieved convergence on important outcome variables between the fifth and sixth waves before reaching 350 participants.

Formative research

Formative research (FR) was conducted before data collection, as recommended by Johnston et al.19, following the Strengthening the Reporting of Observational Studies in Epidemiology guidelines for studies using the RDS method. Individual interviews and focus groups (FGs) were used. The interviews covered topics such as sexual identity, the social and geographical organization of MSM, and perceived acceptance by the community. Questions pertaining to study coordination, including details about the study location and operating hours; inquiries about social network size and willingness to participate and test; and the potential to act as seeds, were explored until saturation was achieved on each topic. A total of 184 MSM participated in the FR, of whom 58 participated in individual interviews; 17 FGs were conducted.

Data collection

Five to six participants were identified as seeds in each study location. Each seed received three nonreproducible coupons to invite three acquaintances. All eligible participants who completed the survey, including the seeds, received USD 10. The participants received an additional 10 USD for every recruit who completed the survey.

After providing consent, the interviewer administered a questionnaire to assess the size of the participant’s social network through a computer-assisted self-interview. Before the interview, the participant was offered a tablet if they preferred to answer the questionnaire independently. The questionnaire was divided into several sections to assess various domains: (1) identification and eligibility; (2) socioeconomic and demographic information; (3) access to general and HIV-specific health services; (4) history of HIV, syphilis, and hepatitis B and C testing; (5) knowledge about HIV and AIDS, preexposure prophylaxis (PrEP), postexposure prophylaxis (PEP), and treatment; (6) gay and homosexual visibility; (7) sexual behavior; (8) drug and alcohol use; and (9) social inclusion and participation.

The participants’ social network was investigated via a four-question cascade: “How many men have you met or talked to in the last 2 months whom you know by name and who know you by name; who had sex with other men (oral or anal) in the last 12 months; who live, study, and work in [city]; and who are 18 years old or over? How many of them would you invite to participate in this study?” After completing the questionnaire, the respondent was referred for blood tests.

The participants initially underwent the RT1 Anti-HIV rapid test (Alere, Bioeasy). If the result was positive, RT2 HIV testing was performed (Abon). Participants with two positive test results were considered to be living with HIV. HIV-positive participants were counseled and immediately sent to a referral center or health unit. Following the rapid tests, the blood samples were centrifuged and stored at − 20 °C. The samples were sent to the national reference laboratory (Instituto Adolfo Lutz, São Paulo) for confirmation. The transportation of samples was performed according to standards set by the Brazilian National Health Surveillance Agency. Plasma sample HIV RNA was quantified via the Abbott real-time HIV-1 assay (M2000), following the manufacturer’s instructions, in an accredited laboratory of the National HIV Viral Load Network (IAL, Centro de Virologia, São Paulo). Reverse transcription polymerase chain reaction (RT‒PCR) was employed to generate RNA genome amplicons. The amplification cycle used to detect the fluorescent signal was proportional to the logarithm of the HIV concentration.

Data analysis

The Gile sequential sampling (SS) estimator was used to calculate the RDS required weights for every participant in each city (RDS Analyst 1.7–16). This estimator assumes a finite population and requires a population size estimation for each sample. The estimation we used was based on the proportion of men who self-reported having at least one same-sex relationship in the 2013 National Research of Knowledge, Attitudes and Practices in the Brazilian population (aged 18–64 years)20. In our study, we applied those regional estimates to the total male population aged 18–64 years in each city within that region on the basis of data from the Brazilian Institute of Geography and Statistics21. The seeds were included in the sample for analysis.

After applying participant weights, the data from the 12 cities were merged and analyzed via Stata 17.0™ software. Each city was treated as a stratum. A univariate analysis was initially conducted to explore the study variables. Social class was defined according to standard socioeconomic strata (A–E) developed by the Brazilian Association of Research Organizations22. Alcohol use was assessed via the Alcohol Use Disorders Identification Test (AUDIT)23, a 10-item questionnaire recommended by the WHO. AUDIT scores range from 0 to 40 points, with a score of ≤ 7 indicating low or no risk and a score of ≥ 8 indicating moderate to high-risk consumption. Consistent condom use was defined as the use of condoms in all insertive and receptive sexual acts, with any partner, in the past 6 months. Access to supplies was defined as receiving free lubricant and condoms.

The HIV treatment cascade was analyzed in two complementary ways. In the first approach, the percentage of participants at each stage was calculated based on the total number of MSM living with HIV identified during the study (Fig. 1A). In the second approach, the percentage at each stage was calculated relative to the number of participants in the immediately preceding stage (Fig. 1B). The cascade included four stages: previous HIV diagnosis, use of ART, suppressed viral load (SVL), and undetectable viral load (UVL).

A previous HIV diagnosis was defined as a report of a positive result of HIV testing prior to the study. Moreover, ART use was defined as self-reported continuous use of ART and used as a proxy for treatment retention in this study. This is justified by the fact that the overwhelming majority of HIV treatment in Brazil is provided free of charge by the public health system, even for people with private health insurance. To access the medication, one must be registered and officially monitored by the public or private health service. Only those who reported current treatment were analyzed for viral load. These tests were performed in 2017 using biological samples collected in 2016, which had been properly stored.

Following recent WHO24 and current Brazilian Ministry of Health25 recommendations, SVL was defined as ≤ 1,000 copies/ml, whereas UVL was defined as < 200 copies/ml.

Bivariate analyses of the outcomes, such as a previous HIV diagnosis, ART use, SVL and UVL, were conducted via the chi-square test and prevalence ratio (PR). Variables associated with outcomes at a significance level of 20% were initially included in the multivariate model. Only the variables that maintained a significance level of 5% were retained in the final logistic model.

Ethical considerations

The study protocol was approved in 2015 by the Committee on Research Ethics of the Federal University of Ceará (UFC), accredited by the Brazil National Commission on Research (#1.024.053-23/06/2015), and all methods were performed in accordance with the ethical standards as laid down in the Declaration of Helsinki.

Results

A total of 4,176 MSM were included in the study. Of these, 585 (18.3%, CI95%: 15.4–21.7) were PLHIV. Most MSM diagnosed with HIV were aged ≥ 25 years (69.1%, CI95%: 59.6–77.3), reported previous STIs (69.8%, CI95%: 60.7–77.5), and reported inconsistent condom use (50.5%, CI95%: 40.9–60.0) (Table 1).

Table 1 Sociodemographic and health characteristics of MSM living with HIV in Brazil, 2016.
Full size table

Among all participants who were diagnosed with HIV in the study, only 55.5% (95% CI: 46.2–64.4) were aware of their serologic status, 54.7% (95% CI: 45.5–63.7) received ART, 43.8% (95% CI: 43.9–53.2) had SVL, and 43.5% (95% CI: 34.6–52.8) had UVL (Fig. 1A). With respect to continuous HIV care, 98.4% (95% CI: 96.3–99.3) of the participants living with HIV who were aware of their HIV status received ART. Among those treated with ART, 80.0% (95% CI: 64.8–89.7) had a suppressed viral load, whereas 79.4% (95% CI: 64.4–89.2) had an undetectable viral load (Fig. 1B).

Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

HIV test‒and‒treat cascade for MSM. Brazil, 2016.

1 A. The percentage of participants at each stage was based on the total number of MSM with HIV identified during the study.

1B. The percentage of participants at each stage was based on the total number of MSM in the previous stage.

Bivariate analysis revealed positive associations between a previous diagnosis of HIV infection and parameters such as older age, higher education level, having a religion, not being bisexual, disclosed sexual identity, a previous diagnosis of STI, access to supplies, participation in an NGO, participation in public events, inconsistent condom use, low-risk alcohol consumption, and no use of illicit drugs at least once a week (except marijuana) (Table 2).

Table 2 Factors associated with a previous HIV diagnosis, ART use, SVL and UVL in MSM living with HIV in Brazil.
Full size table

ART use was also positively associated with older age, higher education level, religion, not being bisexual, disclosed sexual identity, a previous diagnosis of STI, access to supplies, participation in NGO, participation in public events, irregular condom use, low-risk alcohol consumption, and no use of illicit drugs at least once a week (except marijuana) (Table 2). SVL was positively associated with older age, higher education level, having a religion, not being bisexual, disclosing sexual identity, a previous diagnosis of STI, access to supplies, participation in NGO, participation in public events, exchanging sex for money, having sex for money in the past 6 months, inconsistent condom use, low-risk alcohol consumption, and no use of illicit drugs at least once a week (except marijuana) (Table 2). UVL was positively associated with older age, higher education level, religion, not being bisexual, disclosed sexual identity, a previous diagnosis of STI, access to supplies, participation in an NGO, exchange of sex for money, inconsistent condom use, low-risk alcohol consumption, and no use of illicit drugs at least once a week (except marijuana) (Table 2).

Multivariate analysis revealed that awareness of HIV serostatus was positively associated with age ≥ 25 years (PR: 1.5, CI95%: 1.1–2.0), completion of high school or higher (PR: 1.5, CI95%: 1.1–2.2), a previous diagnosis of STI (PR: 4.3, CI95%: 1.9–9.4), not being bisexual (PR: 1.5, CI95%: 1.0–2.1), consistent condom use (PR: 1.2, CI95%: 1.0–1.5), and low risk of alcohol consumption (PR: 1.2, CI95%: 1.0–1.5) (Table 3).

Table 3 Multivariate model of factors associated with previous HIV diagnosis, ART, and UVL in MSM living with HIV in Brazil.
Full size table

After adjustment, use of ART remained positively associated with age ≥ 25 years (PR: 1.5, CI95%: 1.1–2.1), completion of high school education or higher (PR: 1.7, CI95%: 1.1–2.6), a previous diagnosis of STI (PR: 4.0, CI95%: 1.8–8.8), not being bisexual (PR: 1.6, CI95%: 1.1–2.3), attending public events (PR: 1.2, CI95%: 1.0–1.4), and a low risk of alcohol consumption (PR: 1.2, CI95%: 1.0–1.5) (Table 3).

After adjustment, SVL remained associated with high school education or higher (PR: 2.0, CI95%: 1.2–3.4), a previous diagnosis of STI (PR: 3.7, CI95%: 1.4–9.9), participation in NGOs (PR: 1.4, CI95%: 1.1–1.9), consistent condom use (PR: 1.5, CI95%: 1.1–2.1), and a low risk of alcohol consumption (PR: 1.4, CI95%: 1.1–1.9) (Table 3).

After adjustment, UVL remained associated with age ≥ 25 years (PR: 1.5, CI95%: 1.1–2.2), completion of high school education or higher (PR: 1.6, CI95%: 1.1–2.5), a previous diagnosis of STI (PR: 3.4, CI95%: 1.4–8.2), not being bisexual (PR: 1.6, CI95%: 1.1–2.4), and participation in public events (PR: 1.2, CI95%: 1.0–1.4) (Table 3).

Discussion

Brazil has significantly reduced its public policies for HIV prevention and control in general, especially those targeting younger LGBTQIAPN + populations. This setback has notably impacted MSM26, probably contributing to the significantly increased HIV prevalence between 2009 and 2016 in this population, from 12%18 to 17%27. A 320% increase was observed among Young MSM of low Social Economic Status, who, not coincidentally, are those who suffered the most from having entered adolescence without benefiting from the public prevention policies that were interrupted during this period. Therefore, presenting these data, even if older, can contribute to current and future assessments of the HIV prevalence situation among MSM and the different stages of the HIV care cascade.

Although the global effort to reduce the transmission of HIV has progressed in recent years28, that progress is not well reflected in key populations, especially MSM. Our study revealed that almost half of the MSM living with HIV were unaware of their HIV status. This means that even though HIV treatment rates are high, only 44% of the total population of MSM with HIV have achieved SVL.

Brazil has made significant strides toward achieving the proposed global targets, reaching 88–83–95 in 20232. These results suggest that 88% of PLHIV are aware of their HIV status, 83% of people who know their HIV status are receiving ART, and 95% of people on ART have a suppressed viral load.

This partly reflects changing policies surrounding HIV. In the first decades of the epidemic, key populations at higher risk of infection were targeted, and campaigns and policies were directed specifically to these populations; for a number of reasons, efforts were made to “normalize” the epidemic and develop strategies for the general population. This also entailed severely cutting resources for HIV prevention NGOs focused on key populations.  . While testing is always freely available at clinics, our study demonstrates how simply providing a service does not guarantee its use.

Since 1967, general health testing has been mandated for young men conscripted into the Brazilian armed forces29. Testing for HIV began in 199230. Currently, this is conducted on a sample of men eligible for the draft. Another public policy aimed at the general population encourages professionals at STI clinics to offer HIV testing to their users31. MSM, however, are less likely to seek treatment for reasons of fear of stigma and discrimination32 or to avoid anticipated depressing consequences of a positive diagnosis. Although PrEP and PEP users, patients with a clinical suspicion of AIDS, and people who had contact with PLHIV or people suspected of being infected are encouraged to test33, the results of this policy for the populations in our study are that only half the population of potential beneficiaries take advantage of the program. MSM are among the populations with the highest HIV rates in Brazil, and our study demonstrates the large number of MSM who are not being tested.

Brazil faces significant challenges concerning access to HIV prevention, particularly due to stigma and cultural barriers as well as socioeconomic inequalities that hinder vulnerable people and groups from gaining access to HIV/AIDS prevention and treatment34. The results of this study show that older MSM with higher education levels and a previous diagnosis of STIs had greater knowledge of their serologic status, were more adherent to ART, and had higher UVL, corroborating other studies conducted in Brazil and other countries34,35. Higher education levels are often associated with better living and health conditions and lower HIV infection rates in the Brazilian population36. Undergoing STI/HIV testing provides an opportunity to establish a link with health services and disseminate information among healthcare users34.

The literature presents different findings regarding the association between age and awareness of serologic status among MSM. Some studies suggest that younger MSM tend to have better awareness of their HIV status34,35,37; in other studies, age does not significantly correlate with knowledge of HIV-positive status, with lower awareness observed at both ends of the spectrum38,39.

Older individuals, generally with longer sexually active lives, are more likely to have already been tested and received a diagnosis. Furthermore, they may have accumulated greater experience using the healthcare system and accessing HIV-related services, which favors engagement with ART and virological suppression. In contrast, younger individuals may face barriers such as stigma, less knowledge, or low risk perception, which reduces their engagement with care40,41. Sexual orientation also influenced the study results. Participants who do not identify as bisexual have greater knowledge about HIV and better risk perception and integration into healthcare services, which may facilitate access to testing, engagement with care, and sustained adherence to ART37.

A prior history of STIs emerged as a strong marker of both increased exposure and increased contact with health services. Individuals who have had an STI frequently receive counseling and are tested for HIV, which increases the chances of early diagnosis and treatment initiation. Similarly, participation in NGOs and public events was positively associated with HIV-related outcomes, reflecting the crucial role of these community organizations and collective spaces in disseminating information, distributing prevention supplies, and facilitating access to health services40. Behavioral factors also influenced outcomes. Inconsistent condom use can increase risk perception after exposure, leading to increased demand for testing and health services, while also reflecting greater vulnerability to HIV15.

High-risk alcohol consumption was another important factor associated with negative outcomes at different stages of the HIV care cascade in our study. Individuals classified as low-risk may have more organized lifestyles, which favor ART adherence, while those at higher risk face barriers such as missed doses, social instability, or prioritizing substance use over health care15,42.

Alternatives for controlling problematic alcohol consumption, such as behavioral and pharmacological interventions43, should be considered with a view to expanding positive results in HIV testing and treatment.

The widespread dissemination of information on transmission routes is crucial for promoting safer behaviors, particularly among key populations44. PrEP, PEP, and HIV self-testing complement existing strategies to support decision making in cases of potential risk exposure45. Some cities, such as London, Amsterdam, and Sydney, have already achieved global targets in response to AIDS by reducing stigma and discrimination, testing, and increasing access to PrEP4,5.

Given the success of Brazil’s early model AIDS program, which was specifically targeted at MSM and other key populations, our findings suggest that the movement to normalize HIV and AIDS has not had the expected consequences, at least for prevalence and new infections. Our findings underscore the consequences of inadequate investment in the MSM community and in targeted programs to address the epidemic among MSM and other key populations. In the past, meaningful support for MSM facing and MSM community organizations was provided, actively encouraging community members to help develop and deliver solutions. Current efforts seem limited to printing and distributing educational materials, which limits opportunities for other persuasive communication methods46. It is vital to increase investments in HIV control programs, including NGO support. In this study, MSM who participated in NGO activities had a 40% higher prevalence of SVL. One study showed that the HIV incidence would be 50% higher in a scenario without NGOs promoting prevention47.

Some limitations of this study include the sample size and weights used to adjust estimates, as previously described26. Recent modifications to RDS Analyst use a new method, such as visibility imputation, to calculate survey participation probabilities48. However, we opted to use our prior analysis strategy for imputation, as the differences are small and the methods and results would differ from those previously published by our group.

This study used data collected in 2016, a period prior to the broader introduction of PrEP in Brazil, the expansion of the use of new, more effective and less toxic ART regimens, and the impact of the COVID-19 pandemic on the organization of health services. It is therefore possible that the current pattern of testing, access, and adherence to treatment among Brazilian MSM differs from that observed in this survey, which may limit the generalizability of the results. The availability of PrEP may have increased demand for health services and testing frequency, while the new ART regimens possibly favored greater adherence and viral suppression. In turn, the COVID-19 pandemic imposed restrictions and interruptions in access to in-person services, which may have exacerbated existing inequalities. Nevertheless, the findings presented here remain relevant, as they reflect structural and behavioral determinants that continue to influence engagement in HIV care, offering important insights for understanding persistent vulnerabilities and for planning prevention and health care strategies.

Conclusions

The response against the HIV epidemic has several dimensions, including increasing access to testing and education about the infection, reducing the viral load in PLHIV, and implementing treatment as prevention to curb transmission if the infection is not prevented at an earlier stage. Brazil was the first country to provide universal ART13 in 1997, with the enactment of Law No. 9,313/1996, which guaranteed universal and free distribution through the Unified Health System (SUS) and marked a turning point in access to treatment. Between 1997 and 2003, the number of individuals receiving ART in the country nearly tripled49. After 2010, Brazil already reported ART adherence rates of around 90% in the general population50, reinforcing the effectiveness of this universal access policy. Subsequently, the country consolidated its position as an international reference by continuously expanding treatment coverage, incorporating more effective and less toxic therapeutic regimens, and adopting guidelines increasingly aligned with global recommendations, such as the “test and treat” strategy following the results of the START trial in 201551. Building upon this pioneering policy framework, our study demonstrates that the public health system has effectively supported MSM living with HIV by ensuring access to free health services, facilitating ART provision, and assisting individuals in achieving SVL/UVL. However, Brazil still faces challenges in HIV diagnosis, which remains a significant barrier to achieving SVL/UVL. Although most patients who knew their serologic status completed all stages of the care cascade, adhering to treatment and achieving SVL/UVL, these achievements were below the recommended levels of target population coverage. Brazil’s current approach is insufficient in safeguarding this key population from HIV infection and adequately preventing and diagnosing those who are already infected. Consequently, this contributes to challenges in maintaining MSM in the care cascade and the ultimate goal of eliminating HIV.