Introduction

Human immunodeficiency virus (HIV) continues to be a significant global health challenge despite decades of research and intervention efforts. In 2023, there were 1.3 million new infections, with about half occurring in sub-Saharan Africa [1]. A substantial proportion (55%) of these infections were among key populations and their partners [2], including men who have sex with men, female sex workers, injecting drug users, and others. HIV prevention research, particularly on interventions like vaccines, focus on enrolling key populations at high risk of HIV acquisition to assess efficacy [3,4]. The success of these trials depends on factors such as HIV incidence, participant accrual, duration of participant follow-up, and maximising retention [5]. However, in the past, studies were prematurely terminated due to unexpectedly low HIV incidence leading to insufficient statistical power [6].

Global HIV incidence is declining, with fewer new infections reported in 2023 than at any point since the mid-1990s [1]. The most significant reductions have occurred in sub-Saharan Africa, where new infections among adults aged 15–49 dropped by 54%, from 1.1 million in 2010–510,000 in 2022 [2]. This decline is largely attributed to the expansion of prevention strategies such as oral pre-exposure prophylaxis (PrEP) and universal test-and-treat programs [7,8]. However, current incidence data remain limited. A 2021 systematic review of data from the reported HIV incidence rates ranging from 0.6 to 3.4 per 100 person-years (PYO) among sex workers, 1.03 to 15.4 per 100 PYO among men who have sex with men, and 1.59 to 6.04 per 100 PYO among fisher folk [9]. While declining incidence reflects progress in HIV prevention, it also poses challenges for HIV prevention trials by reducing observable new infections, lowering statistical power, and increasing trial complexity and costs [4,10–13]. Addressing these challenges requires larger sample sizes, longer follow-up periods, and the integration of proven prevention methods with new interventions [14]. To improve trial efficiency, researchers incorporate consideration of pre-existing infections, product adherence, and discontinuation in sample size calculations, while also developing biomarkers, and alternative trial designs [15]. Consequently, relative efficacy, rather than absolute efficacy, has becomes a key measure of intervention success [16]. Given the evolving HIV prevention landscape, updated incidence data are essential—particularly in the era of PrEP—to ensure effective trial planning.

To support the PrEPVacc HIV vaccine efficacy trial [17], we conducted the PrEPVacc registration cohort study across five research sites in Mozambique, South Africa, Tanzania, and Uganda. This study aimed to recruit high-risk, HIV-negative individuals, estimate HIV incidence, and inform sample size calculations. In Uganda, the study took place in Masaka, a hub for HIV research. Previous vaccine preparedness studies in Masaka reported HIV incidence rates ranging from 1% in rural communities to 4% among serodiscordant couples [18]. Additionally, earlier research estimated an incidence of 4.9 cases per 100 PYO among high-risk fishing communities [19]. Key risk factors for HIV acquisition included female sex, marital status, multiple sexual partners, a history of sexually transmitted infections (STIs), and lack of circumcision [20,21]. Further studies linked younger age, alcohol use, and genital discharge to increased risk among serodiscordant couples [22]. Given that these data are outdated, updating HIV incidence estimates and risk factors was essential for a more accurate and current assessment. This paper presents findings on HIV incidence and associated risk factors among participants in Masaka, offering critical insights to guide future HIV prevention research.

Materials and methods

Results

Baseline characteristics

A total of 1,736 adults were screened, of whom 1,422 (82%) were enrolled. The most common reason for screen failure (n = 314) was low risk of HIV acquisition, accounting for 94% (n = 295) of the failures. Among those enrolled, 1,115 (78%) attended at least one follow-up visit. of these, 691 (62%) exited the study early, mainly due to declining further follow-up (30%, n = 206) or being re-classified as low risk for HIV acquisition (21%, n = 142). Overall, 424 (38%) participants completed the study (Fig 1). Participants who did not attend any follow-up visits were more likely than those who did to be female (93% vs 69%, p-value<0.001), sex workers (69% vs 46%, p-value<0.001), and residents of non-fishing villages (94% vs 83%, p-value<0.001), (S1 Table).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Study profile of adults screened and enrolled into the vaccine preparedness cohort in Masaka, Uganda (July 2018-October 2022).

https://doi.org/10.1371/journal.pone.0348769.g001

Among participants who were enrolled and followed up, 69% (n = 769) were female and 55% (n = 612) were aged 24 years or younger. Most participants (61%) had primary school education or lower, 53% were single, and 46% were sex workers. The majority (83%) resided in non-fishing villages (Table 1).

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Baseline factors associated with HIV acquisition among adults enrolled in an HIV vaccine preparedness study in Masaka, Uganda.

https://doi.org/10.1371/journal.pone.0348769.t001

HIV incidence and risk factors during follow up

At the final follow-up assessment, incident HIV infection was significantly associated with reported engagement in sex after consuming alcohol in the past 3 months [aIRR = 2.65 PYO, 95% CI 1.11–6.31] and with STI diagnosis or treatment during the same period [aIRR = 2.52 PYO, 95% CI 1.09–5.80] (Table 2).

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Risk factors for HIV acquisition among adults in Masaka, Uganda: Analysis of final follow-up data from an HIV vaccine preparedness study.

https://doi.org/10.1371/journal.pone.0348769.t002

In the analyses accounting for repeated assessments of risk at individual level over time (S3 Table), reported engagement in sex after consuming alcohol in the past 3 months was significantly associated with HIV incidence [aIRR = 4.16 PYO, 95% CI 1.60–10.85]. A weaker non-statistically significant association was observed for STI diagnosis or treatment in the past 3 months [aIRR = 2.06 PYO, 95% CI 0.84–5.06].

Discussion

We observed a relatively high HIV incidence within this vaccine-preparedness cohort, indicating that HIV transmission continues to occur despite the availability of multiple prevention options. This ongoing transmission may partly reflect low uptake and suboptimal adherence to existing prevention strategies in some subpopulations [29,30]. Our findings are consistent with previous work in this population low uptake of PrEP, primarily due to pill burden, stigma, and the need for more time to make informed decisions about its use [31]. Although the observed incidence was high, it was lower than anticipated. This may reflect the increasing availability and uptake of a combination of effective interventions, including HCT, condom use, voluntary medical male circumcision, antiretroviral drugs for the prevention of mother-to-child transmission, PrEP, and treatment as prevention [32]. Previous studies in similar high-risk populations have reported HIV incidence rates ranging from 3.1 to 9.3 per 100PYs [33]. As noted above, declining HIV incidence in high-risk populations may present challenges for the conduct of HIV prevention trials, as lower incidence reduces the ability to measure the effectiveness of new interventions.

In our study, HIV incidence was significantly higher among women than among men. Women are at an increased risk of HIV acquisition due to a combination of biological, behavioral, socioeconomic, cultural and structural factors [34]. Structural barriers, such as limited access to HIV prevention services, lower education opportunities, limited skill development options, intimate partner violence, and migration driven by economic hardship, may further exacerbate their vulnerability to HIV infection [35]. Younger women, in particular, have become a major focus of HIV prevention interventions efforts [36].

We found that individuals residing in fishing communities were at higher risk of acquiring HIV than their counterparts in other communities. People living in fishing communities in sub-Saharan Africa are known to face increased HIV risk due to several factors, including high mobility, engagement in transactional sex, multiple sexual partnerships, heavy alcohol use, inadequate health infrastructure, and limited access to healthcare services [33,37,38]. Additionally, newer interventions such as oral PrEP remain underutilized in these communities, which may hinder efforts to reduce HIV incidence [39].

Illicit drug use can directly or indirectly increase the likelihood of engaging in risky sexual behaviors, thereby increasing the risk of HIV acquisition. This may explain our finding that use of reactional drugs and alcohol was associated with higher HIV incidence. Recreational drug use is relatively common among high-risk groups, including sex workers and individuals in fishing communities in SSA [40]. Both recreational drug use and alcohol consumption have been associated with increased likelihood of sexual encounters, inconsistent condom use, and other high-risk sexual practices [41].

STIs are known risk factors for HIV acquisition. They can cause genital inflammation and ulceration, which facilitate viral entry, particularly among women [42]. We found that both a diagnosis or treatment of an STI and a history of genital ulcers were significantly associated with HIV acquisition. Interestingly, abnormal genital discharge was associated with a reduced risk of HIV acquisition. This unexpected finding may reflect unmeasured confounding or possible misclassification. For example, some female participants may have misreported normal vaginal discharge associated with menstruation or the use of oestrogen-containing contraception as abnormal [43]. Alternatively, the reported discharge may have resulted from disturbances of the normal vaginal flora rather than from an STI. Although vaginal discharge itself does not directly protect against HIV, a healthy vaginal microbiome, dominated by Lactobacillus species, plays a key role in reducing HIV susceptibility by maintaining an acidic pH, strengthening the mucosal barrier, and preventing inflammation [44]. Nevertheless, other studies have similarly reported no clear association between abnormal vaginal discharge and an increased risk of HIV acquisition [45].

A major limitation of this study is that 22% of individuals did not attend any follow-up visits. These individuals were more likely to be young (<25years), female, recreational drug users, and persons who engage in transactional sex [46], subpopulations that are also associated with a higher risk of HIV acquisition. Their absence from follow-up likely introduced selection bias, potentially leading to an underestimation of HIV incidence. LTFU was also higher during the COVID-19 restriction period (2020–2021) compared with the pre-pandemic period (2018–2019). To support retention, several strategies were implemented, including collection of detailed locator data, short clinic visits (1–2 hours), flexible reimbursement of transport costs, immediate follow-up of missed visits (≤7 days) via phone and/or home visits, and community engagement meetings.

Additionally, data on HIV risk factors were self-reported. Despite being collected by well-trained and experienced staff, these may have been affected by recall bias, social desirability bias, or misreporting. To mitigate these biases, multiple questions were used to verify and crosscheck responses. The inclusion of objective measures, such as laboratory testing for STI testing, could have provided biological confirmation of HIV risk but was not feasible in this study.

HIV infection was diagnosed using antibody-based rapid diagnostic tests, which may not identify very recent infections, before seroconversion has occurred, potentially resulting in delayed diagnosis and misclassification of HIV status. Quarterly HIV testing substantially mitigated this risk by increasing the likelihood of detecting new HIV infections during follow-up, although some underestimation of HIV incidence may still have occurred.

Conclusions

HIV incidence remains high, especially among women and residents of fishing communities, despite the availability of effective prevention interventions such as oral PrEP. To address this issue, it is crucial to strengthen HIV prevention and care services specifically tailored to these vulnerable populations. Targeted HIV education and community outreach programs could enhance access to oral PrEP in fishing communities by addressing stigma, low awareness, and logistical barriers to service uptake. Additionally, counselling and support groups can play a crucial role in mitigating risk associated with substance use and encouraging PrEP uptake and adherence. Prevention efforts should prioritise targeted interventions that focus on reducing high-risk behaviours, including recreational drug and alcohol use, while also strengthening the prevention, screening, and treatment of STIs.

As the use of placebo-controlled trials becomes less feasible due to the increasing availability of highly effective HIV prevention tools, future HIV prevention trials must explore innovative methodologies to assess the efficacy of new interventions. Some of the proposed approaches include the use of registration cohorts, such as the one described in this paper, to estimate HIV incidence in real-world settings without the experimental intervention; recency assays to determine recent HIV infection at trial screening; external placebo arms using incidence data from previous placebo-controlled HIV prevention trials as a reference; biological markers of HIV incidence to assess HIV exposure and infer intervention efficacy; drug concentration measurements to evaluate pharmacokinetic effectiveness of prevention products; and immune biomarkers, such as antibody responses, to help predict the protective effectiveness of emerging HIV prevention strategies [47].

Supporting information

Acknowledgments

We thank all the participants and staff of the PrEPVacc registration cohort study at the MRC/UVRI and LSHTM Uganda Research Unit Masaka Clinical Research Centre.

This study is also made possible by the support of the American People through the U.S. President’s Emergency Plan for AIDS Relief (PEPFAR) through United States Agency for International Development (USAID). The contents of this publication are the sole responsibility of the PrEPVacc Team and do not necessarily reflect the views of PEPFAR, USAID, or the United States Government.