The Transformative Impact of the African Cohort Study (AFRICOS) Toward Reaching HIV 95-95-95 Goals in Sub-Saharan Africa

INTRODUCTION

When the HIV epidemic began, there were profound effects on countries’ social, economic, and healthcare landscapes. Unfortunately, Africa bore a disproportionate burden of the global HIV/AIDS pandemic, with nearly two-thirds of all people living with HIV residing on the continent. However, significant progress has been made in controlling the epidemic. According to the Joint UN Program on HIV/AIDS (UNAIDS), the annual number of new infections decreased from 1.6 million in 2010 to approximately 1.3 million in 2022. In addition, the number of AIDS-related deaths in Africa declined from 1.3 million in 2010 to approximately 630,000 in 2022.¹

The U.S. President’s Emergency Plan for AIDS Relief (PEPFAR) was established in 2003 in response to the escalating HIV crisis. Recognizing the need for a comprehensive and coordinated effort, PEPFAR works alongside partner governments, civil society organizations, private sector organizations, and external partners to control the HIV/AIDS pandemic. This collaborative effort has led to lifesaving antiretroviral therapy (ART) being available and provided to 20.5 million people living with HIV (PLWH).² In addition to ART, prevention efforts have intensified, with widespread educational campaigns, condom distribution, and increased awareness leading to a decline in new infections. However, challenges such as ensuring that the needs of the younger population living with HIV are met, ongoing stigma, and diagnosis and treatment of comorbidities persist.

Africa’s HIV landscape is complex, marked by diverse cultures, socioeconomic challenges, and variations in healthcare infrastructure. Understanding the nuances of the epidemic within this context is paramount for designing effective, culturally sensitive interventions. The African Cohort study (AFRICOS) was established in 2013 as a 15-year observational study to help guide PEPFAR programming and policy. Funded by PEPFAR, AFRICOS is a longitudinal observational cohort study of PLWH and a smaller group of people living without HIV (PLWoH) and runs in four countries within sub-Saharan Africa: Nigeria, Kenya, Tanzania, and Uganda. The African Cohort study is uniquely positioned to evaluate the longitudinal impact of clinical practices, biological factors, and socio-behavioral issues on disease progression and HIV infection in an African context. The robust data collected in the cohort allow for investigation of HIV disease progression and associations with a multitude of health outcomes to identify valuable insights that are described in later sections; further details of the cohort are described in prior publications and are available online at https://www.hivresearch.org/african-cohort-study-africos.^3,4

The African Cohort study has provided valuable insights into the natural course of diseases, the effectiveness of interventions, and the identification of potential risk factors. The goal of this paper is to describe the contributions of AFRICOS to the continual effort to fight against HIV/AIDS in Africa and to reach the UNAIDS’ goals of ending the epidemic by 2030. As a summary of the breadth of scientific insights generated over the last decade, the paper also provides examples of how PEPFAR’s sustained investment in science continues to inform HIV programming decisions in Africa and beyond. Going forward, as countries progress toward the 95-95-95 targets, the importance of following the science could not be more relevant/pertinent. In the following sections, key results across the major domains/topic areas (i.e., mental health and cognition, clinical profile, and adherence) and their contribution to research are discussed, as well as how AFRICOS is looking to sustain the gains in the future.

RESEARCH CONTRIBUTIONS

Over the past decade, the AFRICOS Study Group’s findings have contributed to our understanding of HIV epidemiology, clinical outcomes, and sociodemographic factors in the African context. These papers have informed how PEPFAR has shifted programmatic priorities with the evolving phases of the epidemic. Outlined below and in Table 1 are the key insights that have shaped PEPFAR’s response over the last 10 years.

1. 1. Clinical Profiles and Treatment Outcomes: AFRICOS has examined the clinical characteristics of PLWH, including CD4 cell counts, viral loads, and treatment adherence. These studies have contributed to our understanding of disease progression, the efficacy of ARTs, and the factors influencing treatment outcomes.

The findings from papers looking at viral suppression indicate that the use of second-line ART, low CD4 count, and missing any day of ART are important predictors of virologic failure.⁵ These results were further confirmed when we applied machine learning methods to better characterize participants at risk of viral failure.⁶ From a programmatic standpoint, we evaluated the role of low-level viremia on subsequent virologic failure. Any viral load test results >200 copies/mL were associated with an increased risk of subsequent virologic failure.^7,8

In 2019, the combination of tenofovir disoproxil fumarate, lamivudine, and dolutegravir (TLD) was rolled out as the preferred first-line regimen. The AFRICOS was able to monitor and evaluate this transition and its effect on HIV outcomes. We found that participants who switched to TLD had significantly lower rates of viral failure compared with those who did not switch.⁸ However, because of dolutegravir’s association with weight gain, we were concerned that the transition from tenofovir/lamivudine/efavirenz (TLE) to TLD would lead to a decrease in adherence and subsequently viral load suppression. The AFRICOS data showed that a substantial proportion of participants experienced weight gain with a body mass index (BMI) at 1.77 times the rate observed in those on other ART regimens.⁹ Despite this, there was no significant impact on adherence or viral load suppression.¹⁰ Nevertheless, despite the benefits of TLD, AFRICOS data identified that women were less likely to be switched to this lifesaving regimen than men were.¹¹ The importance of these results to PEPFAR clinical programs was substantial; AFRICOS data provided early, robust evidence that switching to TLD was safe and effective, helping to catalyze widespread scale-up. As of 2021, TLD accounted for at least of 80% of antivirals dispensed.¹²

Another key finding relates to predictors of all-cause mortality among PLWH. The causes of death varied, including malignancies, infections, and non–HIV-related conditions. Factors such as a lower CD4 count, increased viral load, recent fever, low BMI, clinical depression, and advanced WHO stage were identified as predictors of mortality. However, each additional year on ART was associated with a reduced hazard ratio for mortality.⁴ Although the findings underscored how mortality rates among PLWH were still three times higher than those among contemporaneous comparators with HIV, they provided compelling evidence from diverse African settings of the impact of ART on all-cause mortality.

Looking at data longitudinally across several indicators, health outcomes have improved since the inception of AFRICOS in 2013.¹³ Consistent with changing guidelines, the interval between diagnosis and ART initiation has decreased over time. Median time to ART initiation decreased from 22.0 months (interquartile range [IQR]: 4.0–77.3 months) among participants diagnosed prior to 2006 to 0.5 month (IQR: 0.2–1.8 months) among those diagnosed in 2016 and later.¹⁴ Upon more recent analysis, most participants started ART the same day as diagnosis. Consistent with these findings, from 2013 to 2021, the prevalence of CD4 <200 decreased from 10.5% to 3.1%, whereas the percentage on ART increased from 76.6% to 100%.^14,15 We also found accompanying decreases in mortality over time with the mortality rate for PLWH reduced from 19.74/1,000 person-years in 2013–2014 to 4.74/1,000 person-years in 2019–2020.⁴

1. 2. Socio-Behavioral Factors and Adherence: Socio-behavioral factors impacting HIV treatment adherence include social support systems, food security, satisfaction with care, and the presence of stigma. Understanding how these factors affect adherence is crucial, as they influence the effectiveness of HIV treatment, outcomes, and potentially transmission.

Psychosocial factors affecting individuals within the AFRICOS cohort have been a focal point of research, in alignment with PEPFAR’s strategic commitment to mainstreaming behavioral and social sciences into HIV programming.^16,17 One study found that although experiencing stigma was associated with decreased odds of adherence to ART and viral suppression, the prevalence of stigma had declined from 22% at study enrollment in 2013 to below 2% by 2018.¹⁸ We have not identified an association between clinic satisfaction or support group attendance with adherence to therapy.^19,20 However, we did identify that missed visits were more common among those <50 years old, those on ART <2 years, and among those with a history of alcohol use, incarceration, and depression. The most common reasons for missing visits were being busy or transportation concerns.²¹ During the COVID-19 pandemic, PEPFAR rapidly scaled up access to multi‐month dispensing to minimize risk to patients and the burden of having to travel to clinics monthly.

Behavioral factors play a crucial role in reducing the risk of transmission and enhancing adherence to HIV treatment as they foster overall health and minimize complicating factors for disease management. Our study looking at condom use found that 64% of participants reported using a condom during the last sexual encounter with their regular partner. Condom use was influenced by age, sex, HIV status, and HIV status of the partner and was less frequent among those who consumed alcohol.²² Regarding disclosure, we found that participants in sero-different relationships were more likely to disclose their status compared with participants in relationships with partners of the same HIV status.²³

The African Cohort study collects data on environmental, social, and economic factors that influence individual behavior. From these data, we have found that food insecurity was highly prevalent in our cohort though not significantly associated with HIV.²⁴ We have also noted differences in outcomes between our rural and urban sites, leading to variations in treatment adherence and overall health management. More analysis is forthcoming to better understand the relationship between environmental factors and health outcomes.

1. 3. Noncommunicable diseases: Several papers within the AFRICOS cohort studies have investigated the complex interplay between HIV and noncommunicable diseases (NCDs).²⁵ This emerging area of research recognizes that PLWH also face an increased risk of developing NCDs such as cardiovascular diseases, diabetes, and certain cancers.

Our analysis found that NCDs were geographically variable and largely consistent with metabolic complications of first-line ARTs.³ Building on this analysis and examining the role of age and NCDs, we found that compared with people without HIV aged <50 years, PLWH aged ≥50 years had increased adjusted odds of having diabetes, dysglycemia, and renal insufficiency.²⁶ We have also commenced analyses to understand how NCDs are managed and to understand potential unmet non–HIV-related health needs. From 2013 through 2023, approximately 20% of PLWH met our definition of hypertension, of whom about 60% did not receive any blood pressure–lowering medication.²⁷ This treatment gap was similar when comparing people with and without HIV and varied by site, likely reflecting differences in how ART clinics manage non–HIV-related conditions. We also found that among those receiving treatment, blood pressure control was suboptimal.

1. 4. Mental Health and Cognition: Cognitive impairments can affect an individual’s ability to adhere to complex medication regimens, impacting overall health outcomes. In addition, the psychological burden of living with HIV, including stigma and potential social isolation, can affect mental health, necessitating integrated care approaches that address both psychological and physical aspects of the disease.

The AFRICOS data revealed that depression is a common comorbidity among PLWH, with prevalence rates varying across regions and demographics. Decreasing age, female sex, worse ART adherence, a detectable viral load, and all-cause mortality were all associated with depression.^4,28 Among pregnant women, 21% reported depressive symptoms, which were associated with two times greater odds of viral nonsuppression.²⁹ The AFRICOS data also shed light on how ART adherence is influenced by mental health, underlining the importance of psychological support in HIV treatment programs.

As PLWH are living longer, understanding the role of HIV’s effect on brain function, leading to challenges in memory, attention, problem-solving abilities, and cognition is important. The AFRICOS data found that 38% of PLWH were cognitively impaired compared with 26% of PLWoH. Factors associated with an increased risk of cognitive impairment included literacy, country, WHO stage 4, and current CD4 cell count.³⁰

1. 5. HIV and Co-Infections: Co-infections remain a significant challenge for PLWH, and the AFRICOS cohort has been instrumental in unraveling the dynamics of co-infections. Papers within this realm have focused on the prevalence of co-infections, such as tuberculosis (TB), hepatitis, and sexually transmitted infections (STIs), and their impact on HIV progression and treatment outcomes.

The AFRICOS cohort had an STI prevalence of 7.7% which did not differ by HIV status. However, the odds of having an STI were higher among those with a lower education level, females, younger individuals, and those with depression.³¹ Risk factors associated with co-infection with syphilis specifically included lower education levels and alcohol consumption.³²

With respect to TB, between 2013 and 2021, the TB prevalence was 3% and the incidence rate was 600 per 100,000 person-years. Factors significantly associated with prevalent TB included higher educational attainment, being underweight, and having a CD4 count <200 cells/mm³.³³ Notably, individuals on TLD had a 78% lower risk of incident TB compared with those on TLE. To better understand the dynamics of subclinical disease, we looked at longitudinal samples of blood mononuclear cells from individuals with active and recurrent TB, latent infection, or incipient TB. The data revealed that Mycobacterium tuberculosis-specific CD4 T-cell activation can distinguish active TB from latent TB with a sensitivity and specificity of 86%. This activation decreases upon TB treatment initiation. Progression to active TB disease began 6–12 months before clinical diagnosis, suggesting that blood biomarkers could aid in early TB detection.³⁴

Co-infection with hepatitis and HIV complicates the clinical management of patients and requires careful coordination of treatment strategies to minimize liver damage. The AFRICOS data found that individuals co-infected with HIV and hepatitis B virus (HBV) exhibited higher levels of inflammation than those with either infection alone or uninfected individuals. This inflammation was linked to liver disease, with notable dysregulation in biomarkers matrix metalloproteinase-2, matrix metalloproteinase-9, and fibroblast growth factor-19. In addition, the co-infected group had a higher Fibrosis-4 score, suggesting undiagnosed advanced liver disease, indicating a more complex and severe disease progression in HIV/HBV co-infection.³⁵

Our work on malaria found that two single-nucleotide polymorphisms (SNPs) significantly influence tapasin Messenger RNA expression levels in Africans. These SNPs are associated with a lower prevalence and incidence of clinical malaria, particularly in individuals carrying tapasin-dependent human leukocyte antigen (HLA) class I molecules allotypes. This suggests that enhanced antigen presentation, influenced by genetic variations, can affect the course of malaria, providing important insights for vaccine efficacy evaluation.^36,37

1. 6. HIV and COVID: The emergence of the COVID-19 pandemic presented an unprecedented challenge to global health systems, including those already addressing the complexities of HIV. The AFRICOS cohort was able to illustrate the intricate relationship between COVID-19 and HIV, shedding light on how PLWH were affected by and responded to the pandemic.

Disruptions in healthcare services due to lockdowns and resource reallocations to the COVID-19 response affected HIV testing, treatment initiation, and adherence to ART. African Cohort Study activities were paused on March 19, 2020 to establish safety procedures for streamlined telephonic and subsequently in-person visits. The first telephonic visit was on May 7, 2020 with reopening dates varied by program. To document the effects of the pandemic on HIV, the AFRICOS protocol was amended to capture data related to COVID, including vaccination status, knowledge, attitude, and beliefs regarding vaccination and the ability to do nasal swab testing.

Initially, there was widespread concern that COVID-19 would have devastating effects in Africa. However, some countries in Africa demonstrated surprising resilience. There are many hypotheses regarding why many countries in Africa were able to evade widespread dissemination. Because AFRICOS participants provided samples for storage and retrospective testing, we were able to analyze retrospective samples to better understand circulating antibodies and assess infections that may have gone undetected. The prevalence of SARS–CoV-2 antibodies in western Kenya among stored samples was 3.3%. When matched with self-reported symptoms, only one person was symptomatic, suggesting asymptomatic infections may have been circulating prior to the first reported case.³⁸ When we compared AFRICOS samples with those from our Thailand studies, we found coronavirus antibodies were much higher among AFRICOS participants than among those from Thailand, especially for outbreak coronaviruses.³⁹ By using retrospective and prospective samples, we are able to better understand how COVD-19 affected diverse communities and the effectiveness of public health interventions and to track mutation patterns that can inform vaccine and treatment development. As such, AFRICOS offers a compelling model for how investment in health research infrastructure can be effectively and rapidly leveraged to advance a broader global health security agenda.

There was also concern that COVID-19’s disruption to public health services would affect HIV care and treatment outcomes.⁴⁰ The AFRICOS data revealed that clinic attendance, food security, and viral load suppression decreased early in the pandemic when control measures were most stringent.^41,42 As the pandemic progressed, PEPFAR implemented several changes to providing care and treatment, including community medication dispensaries, outdoor waiting rooms, and allowance for multiple months of medication to be dispensed at a time. As the pandemic progressed, data showed that adherence was unaffected, and food security stabilized and viral load increased later in the pandemic.^41,42

1. 7. Immunology and HIV Sequencing: These studies aimed to decipher the role of host-virus interactions, including the characteristics of viruses and immune responses that influence HIV acquisition, progression, and clinical outcomes.

The African Cohort study has characterized associations between inflammatory markers and disease outcomes. For example, we found that virally suppressed PLWH had higher levels of nonspecific inflammatory markers (CCL5, CXCL10, fatty acid binding protein (FABP), fas ligand (FASLG), matrix metalloproteinase (MMP) 1, MMP7, soluble (s)CD14, and sCD163) compared with PLWoH. Some of these biomarkers correlated with clinical scoring indices predictive of morbidity and mortality, such as the Veterans Aging Cohort Study index.⁴³ Inflammatory markers were even higher among viremic participants, with the magnitude and character of immune activation influenced by factors such as viral load, sex, age, CD4 count, and region.⁴⁴ Specific comorbid conditions—such as viral hepatitis, syphilis, and TB—contributed to phenotypically unique patterns of chronic inflammation.⁴⁵ Tuberculosis-specific T-cell activation also correlated with disease activity. The AFRICOS data suggest that CD38 expression on TB-specific CD4 T-cells may be a useful surrogate biomarker to detect incipient TB and predict TB recurrence after treatment.

We also looked at biomarkers to characterize HIV-related cognitive outcomes to help guide early detection and prevention efforts. We identified that sCD14 was inversely correlated to neuropsychological testing performance (NTP) for both PLWH and PLWoH. In PLWH, I-FABP was inversely associated with NTP, regardless of viral suppression. These results confirm that inflammatory sources contribute to cognitive performance regardless of HIV status.⁴⁶ More work is needed to determine how these biomarkers can be longitudinally used to guide targeted treatment and interventions.

The African Cohort study works with collaborators to understand the genetic diversity of HIV as it influences the virus’s transmissibility and its response to treatment, thereby informing the development of effective therapy and vaccine strategies. Among Kenya and Uganda AFRICOS participants, HIV-1 subtype A is the main circulating virus, with subtype D contributing a minority. In Tanzania, the majority are subtype C, with a minority being subtype A. In Nigeria, subtypes G and CRFC02_AG are the predominant circulating subtypes.⁴⁷ From 2013 to 2019, AFRICOS identified an almost 3-fold increase in pretreatment drug resistance in Uganda, from 8.0% to 24.3%, driven by increases in resistance to both Nucleoside reverse transcriptase inhibitors and non-nucleoside reverse transcriptase inhibitors. African Cohort Study data also showed that the vast majority of viremic ART-experienced participants had drug-resistant viruses that potentially limited the efficacy of standard first- and second-line regimens. These findings informed country-specific guidelines on standardized ART regimens and the urgency to consider switching from failing regimens.⁴⁸

More recently our preliminary drug resistance testing among individuals on TLD for more than 6 months who had viral nonsuppression (N = 133/2,004) and were successfully virally amplified (N = 65/133) found that 32 (49%) had a drug mutation. There was an integrase inhibitor mutation in seven (11%) specimens, two (3%) with a mutation-conferring high-level dolutegravir resistance.⁴⁹

1. 8. Women’s Health: Women face unique needs and challenges, including access to reproductive health services, mental health interventions, and tailored services perinatally.

Perinatal depression can affect a mother’s ability to adhere to HIV treatment regimens, potentially leading to viral nonsuppression and increased risk of vertical transmission of HIV. However, there are few data regarding antenatal, prenatal, and perinatal trajectories of depression and viral suppression. We found that throughout the 1-year preconception phase, 86% had elevated depressive symptoms, 92% during pregnancy and 94%postpartum.⁵⁰ In a multivariate analysis, elevated depressive symptoms were associated with viral nonsuppression.²⁹

With respect to HPV, we found the prevalence of HPV was 50.9% among women living with HIV (WLWH) compared with 38.1% in women living without HIV (WLWoH) in the AFRICOS cohort. The most common high-risk human papillomavirus (HR-HPV) types among WLWH were HPV-16 (13.5%), HPV-52 (9.5%), and HPV-35 (9.3%). The study also found that HR-HPV infection was more frequent among Tanzanian WLWH and among women with a CD4⁺ T cell count <200 cell/μL.⁵¹

LOOKING TO THE FUTURE

OTHER CONTRIBUTIONS AND CAPACITY BUILDING

Many AFRICOS analyses have been led by site investigators. In addition to building scientific capacity and promoting career growth, AFRICOS has supported sites with developing new technical and laboratory capabilities, including HIV sequencing and drug resistance testing, HPV sample collection, microbiome sample collections (stool, rectal/vaginal swabs), and Mpox sample collection. In alignment with PEPFAR’s commitment to transition toward sustainability, AFRICOS envisions a plan for increasing leadership of its science priorities, data analysis, and presentation of accomplishments by scientific leaders across the four partner countries.

CONCLUSION

In summary, the body of work emanating from the AFRICOS cohort has contributed to the global knowledge base, informing evidence-based interventions, policy decisions, and further research directions in the ongoing efforts to combat HIV and improve the lives of PLWH in the region. Moreover, AFRICOS serves as one of the best examples of PEPFAR’s commitment to investing in science and data to drive programming decisions. The AFRICOS findings have had a substantial impact on PEPFAR program priorities over the last decade. Looking to the future, the network is ideally placed to inform broader global health strategies in the ongoing fight against HIV. As the global community strives toward the UNAIDS 95-95-95 goals, the nuanced insights from cohort studies can be instrumental in refining and optimizing global strategies. The AFRICOS cohort serves as a catalyst for ongoing dialogue and collaboration, encouraging the exchange of best practices, innovative approaches, and culturally relevant interventions.