Same-day initiation of oral pre-exposure prophylaxis among gay … – The Lancet

Background

Although gay, bisexual, and other cisgender men who have sex with men (MSM) and transgender women have the highest HIV burden in Latin America, pre-exposure prophylaxis (PrEP) implementation is poor. We aimed to assess the feasibility of same-day oral PrEP delivery in Brazil, Mexico, and Peru.

Methods

Implementation PrEP (ImPrEP) was a prospective, single-arm, open-label, multicentre PrEP implementation study conducted in Brazil (14 sites), Mexico (four sites), and Peru (ten sites). MSM and transgender women were eligible to participate if they were aged 18 years or older, HIV-negative, and reported one or more prespecified criteria. Enrolled participants received same-day initiation of daily oral PrEP (tenofovir disoproxil fumarate [300 mg] coformulated with emtricitabine [200 mg]). Follow-up visits were scheduled at week 4 and quarterly thereafter. We used logistic regression models to identify factors associated with early loss to follow-up (not returning after enrolment), PrEP adherence (medication possession ratio ≥0·6), and long-term PrEP engagement (attending three or more visits within 52 weeks). This study is registered at the Brazilian Registry of Clinical Trials, U1111-1217-6021.

Findings

From Feb 6, 2018, to June 30, 2021, 9979 participants were screened and 9509 were enrolled (Brazil n=3928, Mexico n=3288, and Peru n=2293). 543 (5·7%) participants were transgender women, 8966 (94·3%) were cisgender men, and 2481 (26·1%) were aged 18–24 years. There were 12 185·25 person-years of follow-up. 795 (8·4%) of 9509 participants had early loss to follow-up, 6477 (68·1%) of 9509 were adherent to PrEP, and 5783 (70·3%) of 8225 had long-term PrEP engagement. Transgender women (adjusted odds ratio 1·60, 95% CI 1·20–2·14), participants aged 18–24 years (1·80, 1·49–2·18), and participants with primary education (2·18, 1·29–3·68) had increased odds of early loss to follow-up. Transgender women (0·56, 0·46–0·70), participants aged 18–24 years (0·52, 0·46–0·58), and those with primary education (0·60, 0·40–0·91) had lower odds of PrEP adherence. Transgender women (0·56, 0·45–0·71), participants aged 18–24 years (0·56, 0·49–0·64), and those with secondary education (0·74, 0·68–0·86) had lower odds of long-term PrEP engagement. HIV incidence was 0·85 per 100 person-years (95% CI 0·70–1·03) and was higher for transgender women, participants from Peru, those aged 18–24 years, Black and mixed-race participants, and participants who were non-adherent to PrEP.

Interpretation

Same-day oral PrEP is feasible for MSM and transgender women in Latin America. Social and structural determinants of HIV vulnerability need to be addressed to fully achieve the benefits of PrEP.

Funding

Unitaid, WHO, and Ministries of Health in Brazil, Mexico, and Peru.

Translations

For the Portuguese and Spanish translations of the abstract see Supplementary Materials section.

Tenofovir-based oral pre-exposure prophylaxis (PrEP) has dramatically reduced population-level HIV incidence in London (UK),

San Francisco (CA, USA),

New South Wales (Australia),

and in rural Kenya and Uganda.

Access to PrEP can reduce HIV incidence among groups disproportionately affected by HIV, including youth, gay, bisexual, and other cisgender men who have sex with men (MSM), cisgender women, and transgender women.

Initiation and continuation of PrEP during periods of higher susceptibility to HIV are required to realise its full benefits. Reasons for stopping or never initiating PrEP include, in addition to access, low self-perceived risk, concerns about side-effects or a desire not to take a daily pill, stigma, or competing life events.

^, 

Furthermore, linkage from HIV testing to a PrEP prescription appointment occurs only for 21–67% of individuals referred for PrEP care.

^, 

To address this gap, PrEP programmes focus on scaling up PrEP uptake, especially among the most clinically vulnerable

(eg, by implementing same-day PrEP, in which a prescription is provided the day that PrEP eligibility is confirmed).

In 2021, 2·2 million people were living with HIV in Latin America. New HIV infections in the region are estimated to have increased in the past decade, with most new cases among MSM and transgender women.

HIV prevalence in the region was estimated at 13·9% for MSM (2016–20) and 25·9% for transgender women (meta-analysis), with differences among and within countries.

^, 

^, 

Key populations in Latin America are at increased clinical vulnerability due to inadequate access to HIV services, stigma, discrimination, and other human rights violations.

Although daily oral PrEP was recommended for people at substantial HIV risk by WHO in 2014, PrEP availability is limited in Latin America. Only 10% of PrEP users across 77 countries are in Latin America, most of whom are in Brazil.

PrEP has been available within Brazil’s Public Health System (Sistema Único de Saúde) since December, 2017, has been recently implemented as a public health policy in Mexico, and is part of a new policy awaiting approval in Peru, where it remains accessible only via purchase or through a few research studies.

From Feb 6, 2018, to June 30, 2021, 9979 participants were screened and 9509 were enrolled (Brazil n=3928, Mexico n=3288, and Peru n=2293; appendix 3 p 2). Participants were followed up until June 30, 2021. 43 patients had acute HIV infection (HIV-negative rapid test or detectable HIV viral load) identified at enrolment and were subsequently referred to HIV care. Thus, the prevalence of acute infection at the screening visit was 0·5%. 915 (9·6%) participants returned more than 24 weeks after a previous visit and were thus re-enrolled in the study (appendix 3 p 3). Re-enrolments were more common in Peru (618; 26·9%) than in Brazil (232; 5·9%) and Mexico (65; 2·0%).

Among the 9509 enrolled participants, median age was 29 years (IQR 24–35); 2481 (26·1%) were aged 18–24 years and 5984 (62·9%) self-identified as mixed race (table 1). The majority had more than secondary education, had attended the clinic seeking PrEP, and reported no postexposure prophylaxis use in the previous 12 months (table 1). Sexual behaviour, binge drinking, and stimulant use are shown in table 1. Based on self-reported information, 5235 (55·1%) participants did not report missing any pill during the previous 30 days and were thus deemed adherent at week 4. 60 participants (0·63%, 95% CI 0·47–0·79) with creatinine clearance 60 mL/min or less at enrolment had repeat creatinine testing, and subsequent results were within the permitted range.

Overall, participants were followed up for 12 185·25 person-years, with more person-years of follow-up in Brazil than in Mexico or Peru (table 2). 104 HIV seroconversions occurred with an HIV incidence of 0·85 (95% CI 0·70–1·03) per 100 person-years. The incidence was greater in Peru than in Mexico or Brazil (table 2). HIV incidence was 2·5 times greater among transgender women than among MSM, 4·3 times greater among younger (aged 18–24 years) participants than older (aged >30 years) participants, and 2·2 times greater among participants who were Black, mixed race, Indigenous, or Asian than among participants who were White (table 2). HIV incidence among participants not adherent to PrEP (MPR table 2).

Among all enrolled participants, 795 (8·4%) never returned for any follow-up visit after enrolment and were considered as an early loss to follow-up (table 3). The proportion of participants recorded as early loss to follow-up was greater in Peru than in Mexico or Brazil (table 3). The odds of early loss to follow-up were greater among transgender women versus cisgender men, among participants aged 18–24 years and 25–30 years versus those older than 30 years, greater among participants with primary or secondary education versus those with more than secondary education, greater among participants reporting condomless anal sex with a partner or partners with unknown HIV status versus those reporting not taking part in condomless anal sex with a partner or partners with HIV, and greater among those reporting taking part in transactional sex versus those reporting not (table 3). The odds of early loss to follow-up were lower among participants who self-identified as White versus those of other races, lower among participants attending the service seeking PrEP versus those attending for other reasons, lower among participants reporting five to ten or more than ten sex partners versus those reporting fewer than five sex partners, and lower among participants reporting receptive condomless anal sex versus those reporting not taking part in this behaviour (table 3). Although similar results were observed in the models for each country, in Peru the following variables were not statistically significant: gender, race, and transactional sex. Non-significant variables were race in Mexico, and receptive condomless anal sex in Brazil (appendix 3 p 5).

Among all enrolled participants, 6477 (68·1%) were adherent to PrEP (MPR ≥0·6), and the proportion of participants adherent was greater in Brazil than in Mexico or Peru (table 4). The odds of adherence to PrEP were lower among transgender women versus cisgender men, lower among participants aged 18–24 years and 25–30 years versus those older than 30 years, lower among participants with primary or secondary education versus those with more than secondary education, lower among participants reporting taking part in condomless anal sex with a partner or partners with unknown HIV status versus those reporting not taking part in condomless anal sex with a partner or partners living with HIV, and lower among participants reporting taking part in transactional sex versus those reporting not taking part in this behaviour (table 4). The odds of adherence to PrEP were greater among participants who attended the service seeking PrEP versus those attending for other reasons, greater among participants reporting five to ten or more than ten sex partners versus those reporting fewer than five sex partners, greater among participants reporting receptive condomless anal sex versus those not reporting this behaviour, and greater among participants reporting condomless anal sex with a partner or partners living with HIV versus those reporting not taking part in this behaviour (table 4). Similar results were observed for models per country, with some exceptions such as gender for Peru and race for Peru and Mexico (appendix 3 p 7).

From the initial pool of enrolled participants, 1230 (12·9%) were not eligible for the long-term PrEP engagement outcome due to administrative censoring occurring before 52 weeks of follow-up. Of the 8279 eligible participants, 7556 (91·3%) attended at least one follow-up visit, while 2910 (35·1%) completed all five follow-up visits. We further excluded 54 participants who did not have the chance to complete a third visit due to adverse events (n=18), medical concerns (n=6), and HIV seroconversion (n=30). Therefore, the sample for this analysis was 8225 participants, with 5783 (70·3%) having long-term PrEP engagement (table 5). The proportion of participants with long-term PrEP engagement was greater in Brazil than in Mexico or Peru (table 5). The odds of long-term PrEP engagement were lower among transgender women versus cisgender men, lower among participants aged 18–24 years and 25–30 years versus participants older than 30 years, lower among participants with secondary education versus those with more than secondary education, and lower among participants reporting condomless anal sex with a partner or partners with unknown HIV status versus those reporting not taking part in condomless anal sex with a partner or partners living with HIV (table 5). Participants who self-reported as PrEP adherent at week 4 had three times greater odds of having long-term PrEP engagement than those who were not PrEP adherent at week 4 (table 5). The odds of long-term PrEP engagement were greater among participants who attended the service seeking PrEP versus those who attended for other reasons, greater among those reporting five to ten or more than ten sex partners versus those reporting fewer than five, greater among those reporting receptive condomless anal sex versus those reporting not taking part in this behaviour, and greater among those reporting condomless anal sex with a partner or partners living with HIV versus those reporting not taking part in this behaviour (table 5). Similar results were observed for models per country, with some exceptions such as gender for Peru and race for Peru and Mexico (appendix 3 p 9).

Prevalence of active syphilis at enrolment was 8·8% (95% CI 8·2–9·4), and overall incidence was 10·09 per 100 person-years (95% CI 9·40–10·82; appendix 3 p 11). Syphilis incidence was greater in Brazil than in Peru or Mexico, and among transgender women than cisgender men. Prevalence of rectal chlamydia decreased from enrolment to week 52 (appendix 3 p 12). Prevalence of gonorrhoea also decreased from enrolment to week 52 (appendix 3 p 12). Among participants aged 18–24 years and transgender women, prevalence of both chlamydia and gonorrhoea at enrolment were greater than in the overall population, although they did not increase over time (appendix 3 p 12). Prevalence of both chlamydia and gonorrhoea decreased over time among participants from Mexico and Peru but remained stable among participants from Brazil (appendix 3 p 12).

Prevalence of active hepatitis B remained stable from enrolment to week 52, whereas prevalence of hepatitis C increased over time (appendix 3 p 13). Prevalence of hepatitis C increased over time among participants from Mexico and Peru, cisgender men, and those aged 25–30 years.

Overall, 87 (0·91%) of 9509 participants interrupted PrEP due to adverse events, and 52 (0·55%) discontinued PrEP due to adverse events (appendix 3 p 14). Notably, very few participants interrupted PrEP (31 [0·33%]) or discontinued PrEP (14 [0·15%]) due to creatinine clearance of 60 mL/min or less.

In this large cohort of 9509 MSM and transgender women enrolled in Brazil, Mexico, and Peru, a same-day PrEP delivery strategy was shown to be feasible and effective, with only a small proportion of participants with early loss to follow-up. Most participants had PrEP adherence and long-term PrEP engagement, although country, sociodemographic, and behavioural disparities were observed. Although HIV incidence was low during the study compared with Latin American sites participating in the HPTN083 study (tenofovir disoproxil fumarate plus emtricitabine group)

and the placebo group in the iPrEX study for Brazil (5·0 per 100 person-years) and Peru (3·5 per 100 person-years),

it was greater than that observed in implementation studies conducted in France and Australia.

^, 

Participants from Peru, transgender women, younger participants (aged 18–24 years), and individuals who were not White had greater HIV incidence, as well as increased odds of early loss to follow-up and lower odds of PrEP adherence and long-term PrEP engagement. Our results corroborate the finding that early adherence, measured at week 4, is associated with higher likelihood of long-term PrEP engagement.

The first month after PrEP initiation can be a valuable window of opportunity to implement strategies aiming to improve PrEP outcomes, particularly among the most clinically vulnerable.

The moderately high prevalence (0·5%) of acute HIV infection, in agreement with that of the PrEP Brasil study

but not quite as high as in the iPrEx study,

suggests that MSM and transgender women interested in PrEP are highly vulnerable to HIV. The opportunity to identify individuals with acute HIV infection before PrEP initiation is particularly important with the introduction of novel long-acting PrEP technologies, such as injectable cabotegravir.

Initiation of PrEP in individuals with non-diagnosed acute HIV infection might lead to HIV drug resistance, potentially affecting treatment outcomes.

Our results underscore the importance of PrEP adherence to prevent new HIV infections, as shown by the lower HIV incidence among participants with an MPR of 0·6 or more. The association of younger age with higher HIV incidence is in line with trends observed in Latin America showing a higher number of new HIV cases among this subgroup.

A combination of lower adherence and lower long-term PrEP engagement might explain these worse outcomes. In a recent cross-sectional online survey conducted in Brazil, younger age was associated with decreased odds of high perceived HIV risk,

which can negatively affect PrEP outcomes. The higher HIV incidence among Peruvian participants might be explained by the inclusion of a younger population, more individuals enrolled who were not actively seeking PrEP, lower adherence, and lower long-term PrEP engagement. Moreover, lower PrEP awareness might explain the lower proportion of participants seeking PrEP in Peru than in Brazil and Mexico.

Low early loss to follow-up indicates the feasibility of same-day PrEP prescribing and delivery, which reduces the attrition between initial evaluation and receipt of a PrEP prescription, as previously seen in the USA.

Our results expand this body of evidence from the perspective of PrEP services in Latin America, and could also be useful for other resource-limited regions. Moreover, our results reinforce that same-day PrEP, as adopted by the Ministry of Health in Brazil since 2017 and in Mexico since 2021, should be maintained and serve as a model for other countries in the region.

ImPrEP participants with lower education compared with higher education in all countries, and of Black or mixed race compared with those of White race in Brazil had worse PrEP outcomes. Social determinants of health, such as race and education, contribute to widening health disparities and inequalities.

Furthermore, MSM and transgender women are also exposed to recurring marginalising experiences, such as homophobia and transphobia, which might affect mental health and wellbeing.

Black and less educated MSM and transgender women are additionally affected by intersectional discrimination related to racism and classism. Studies evaluating PrEP outcomes in Brazil found that Black MSM had lower PrEP awareness, PrEP uptake, and PrEP adherence than White MSM.

^, 

^, 

PrEP services in Latin America should focus on strategies to and explore and address stigma and to encourage PrEP reinitiation, adherence, and persistence among these vulnerable populations, with interventions addressing the gap in access and retention considering age, race, and education.

Young MSM and transgender women aged 18–24 years had worse PrEP outcomes compared with older peers. In Latin America, young MSM are more vulnerable to HIV infection due to poor access to comprehensive sex education and preventive strategies, condomless and PrEP-less sex, and low perceived HIV risk.

^, 

HIV prevalence has been increasing in the past decade in the region among young MSM,

who show lower PrEP awareness, willingness, uptake, and adherence compared with older MSM.

^, 

^, 

Alternative approaches to recruit and retain young populations in PrEP services should be explored, such as using new social media platforms (eg, TikTok), engaging them in development of digital campaigns, use of appropriate language, and having digital influencers to better communicate with these groups.

Transgender women had worse PrEP outcomes when compared with cisgender men, corroborating findings from previous studies.

^, 

Transgender women consistently face challenges engaging in prevention and treatment services, reflecting their underlying vulnerabilities and poor adaptation of services to their needs. A transgender-specific PrEP study in Brazil showed that although high engagement in prevention services was attainable, PrEP adherence decreased over time, with social determinants of health substantially affecting adherence during the study follow-up.

High PrEP retention was achieved in settings where staff were trained on transgender competency, and where gender-affirming care with feminising hormone therapy was provided as part of the study.

In a Peruvian study to support PrEP among transgender women, more than half of participants (49 [55%] of 89) were lost to follow-up after 3 months.

Stigma and previous discriminatory experiences in the health-care setting are crucial barriers to HIV prevention, including PrEP, among transgender women.

The country-level differences observed are likely to reflect distinct underlying characteristics of the populations enrolled, the effect of COVID-19 pandemic restrictive measures, and particularities of countries’ public health systems. Brazil, Mexico, and Peru together represent more than half of the Latin American population (approximately 368 million). These countries have vast socioeconomic disparities, which intensified during the COVID-19 pandemic. Key populations were particularly affected, including by reduced access to HIV prevention services.

A substantial part of the follow-up was affected by the restrictive measures of the pandemic, especially in Peru. Brazil was less affected due to available telehealth procedures, including HIV self-testing distribution for PrEP delivery.

The incidence of sexually transmitted infections was greater among younger participants and transgender women. Syphilis incidence during follow-up was similar in the PrEP Brasil study,

although prevalence at enrolment was lower than the 2010–20 pooled prevalence in Latin America and the Caribbean (8·8% vs 11·2%).

High prevalence of rectal chlamydia and gonorrhoea before PrEP initiation, and decreased prevalence at week 52 suggest that risk compensation did not play a role in our study.

A major strength of this study is its sample size, making it, to our knowledge, the largest PrEP cohort in Latin America. Notably, ImPrEP included the second largest absolute number of transgender women among all PrEP clinical trials and implementation studies to date (n=543). Furthermore, the study was implemented in Brazil at public health system sites, where most of the Brazilian population access PrEP, and contributed to PrEP implementation in Mexico. Some limitations should be acknowledged. We could not perform counterfactual calculations because there was no comparator group or recent incidence data in the region. Our results are not informative of PrEP uptake because participants who were screened and enrolled had previously expressed an interest in participating in the study. Data on PrEP refusal and qualitative data on individuals who had early loss to follow-up or low long-term PrEP engagement or adherence were not collected. We used different timeframes to evaluate number of sex partners across countries. PrEP adherence data were based on self-report and MPR and not on biomarkers of adherence. Although self-reported adherence could overestimate adherence, be context-dependent, and be limited by different biases, previous analyses have shown that MPR and self-reported adherence can discriminate against participants with and without protective tenofovir diphosphate levels in Brazil.

Study person-years, although longer than in most studies, is still short when compared with a person’s lifetime experience of HIV prevention. Studies with longer follow-up will be required to address the sustainability of long-term daily PrEP more comprehensively, and to measure the complexity of how MSM and transgender women start and stop PrEP in relation to HIV vulnerability.