Physical activity interventions for cardiometabolic and psychosocial outcomes in people living with HIV in sub Saharan Africa: A systematic review Smart Z Mabweazara, Jennifer Manne-Goehler, Mark Hamer, Jacqueline Cellini, Takshita Sookan, Lisa K Micklesfield, Mark J Siedner, Miguel Reina Ortiz Abstract Antiretroviral therapy (ART) has extended the lifespan of people living with HIV (PWH), resulting in a growing burden of cardiometabolic disease (CMD) associated with chronic HIV infection and the effects of ART. Physical activity (PA) may reduce CMD risk while improving psychosocial health among PWH. We systematically reviewed the efficacy of PA interventions for improving cardiometabolic and psychosocial outcomes among PWH in sub-Saharan Africa (sSA). Web of Science, PubMed, African Index Medicus, Embase, and CINAHL were searched from inception to 10 August 2023 for randomised controlled trials and quasi-experimental studies published from 2000 onwards involving adults with HIV in sSA. Methodological quality was assessed using the PEDro scale and Downs and Black checklist, and findings were synthesised narratively. Seventeen studies involving 1,130 participants met the inclusion criteria and generally demonstrated good methodological quality. Most participants were female (51%), and all studies included PWH receiving ART. Interventions included aerobic exercise (n = 8), resistance exercise (n = 3), combined aerobic and resistance exercise (n = 5), while one study did not report exercise modality. Intervention durations ranged from 6 weeks to 12 months. Nearly all studies reported improvements in cardiometabolic outcomes, including blood pressure, lipids, body mass index, body fat percentage, and waist circumference, and/or psychosocial outcomes, including depression, anxiety, self-efficacy and quality of life. Two studies assessing PA outcomes reported significant improvements in step count and metabolic equivalents per week. Limitations included small sample sizes and generally short post-intervention follow-up, limiting assessment of sustained benefits and cardiovascular events. PA interventions may therefore be a valuable addition to standard HIV care in sSA. Future research should identify optimal exercise modalities, duration, and intensity and develop feasible, context-sensitive, scalable interventions, particularly for resource-constrained and underserved populations. Introduction Of the 37 million people living with HIV (PWH), approximately 68% live in sub-Saharan Africa (sSA) []. Health priorities for PWH in sSA have significantly transformed in the region, with the worldwide dissemination of antiretroviral therapy (ART) []. Widespread availability of ART has transformed HIV from a frequently fatal condition into a manageable chronic disease, extending life expectancy to nearly that of uninfected individuals when started early and maintained [,]. As life expectancy has increased, there has been a higher incidence of non-communicable diseases within this population []. Cardiovascular and metabolic diseases (CMDs) have been identified as the most common and most severe non-communicable diseases among PWH [] and multimorbidity is already high in sSA populations []. For example, the Randomized Trial to Prevent Vascular Events in HIV (REPRIEVE) study recently reported a median BMI of 25.9 kg/m2 and that 36% of the 7382 participants had hypertension []. These conditions are of particular concern given the use of newer ART regimens, such as dolutegravir, which are strongly linked to obesity, particularly in sSA []. Physical activity (PA) has proven to be effective for reducing the likelihood of CMD in PWH []. Resistance exercise, for example, is linked to significant reductions in high-density lipoprotein, total cholesterol, low-density lipoprotein, and triglycerides []. Resistance exercise is also associated with significant reductions in haemoglobin A1c (HbA1c), fasting blood glucose, body fat percentage [], and blood pressure [], and increased lean body mass percentage in PWH [,]. Physical activity has also been found to increase cardiorespiratory fitness []. There is also overwhelming evidence reported in systematic reviews for the efficacy of PA to improve body composition [,], fitness and functional capacity [,], and some psychosocial conditions, including quality of life and mental health [,] among PWH. These benefits of PA underscore the necessity of promoting its implementation for PWH, particularly in regions where public health systems encounter significant resource limitations []. As CMDs increasingly contribute to illness and death among PWH and given the established link between PA and enhanced CMD outcomes in various populations, comprehensively grasping methods to promote PA among individuals with HIV in sSA will be an essential initial stride []. This will aid in addressing an unfulfilled health requirement for this population []. Given the differences in HIV/AIDS burden, access to healthcare, and socioeconomic factors between regions, particularly the Global North and South, coupled with variations in PA interventions, a systematic review on the effects of PA among PWH sSA is warranted to provide evidence on its impact in this context. We include studies published after 2000 because the early 2000s marked the beginning of massive ART roll-out, with the number of PWH on treatment in sSA rising from 11,000 in 2000 to over 11 million by 2015 []. The development of low-cost generic ART and the launch of major international funding initiatives, specifically the President’s Emergency Plan for AIDS Relief, launched in 2003, caused a massive, rapid expansion in HIV treatment access []. The rapid scale-up of ART significantly decreased AIDS-related deaths and improved life expectancy, shifting HIV from a fatal disease to a chronic, manageable condition []. As the population of PWH ages, the focus of care has shifted to include managing long-term comorbidities such as cardiovascular disease, diabetes, cancer, and neurocognitive disorders alongside HIV, requiring a strong emphasis on lifestyle interventions [–]. We further restricted inclusion to studies conducted in sSA to ensure that findings are contextually relevant and applicable to populations most affected by HIV. Sub-Saharan Africa bears a disproportionate share of the global HIV burden, accounting for nearly two-thirds of all PWH, making it the region where interventions are most urgently needed []. In addition, sSA, particularly South Africa, is experiencing a dual burden of disease, characterised by high HIV prevalence alongside rapidly increasing rates of overweight and obesity. In South Africa, more than half of adults are overweight or obese, with prevalence disproportionately higher among women, including those living with HIV []. Furthermore, sSA settings are shaped by unique socioeconomic and structural constraints, including poverty, food insecurity, limited access to recreational facilities, and unsafe environments for PA [,]. These contextual realities are likely to influence both the design and implementation of PA interventions, meaning that studies conducted in the region may differ in important ways from those undertaken elsewhere, even if similar health-enhancing effects are ultimately observed. To our knowledge, however, no systematic review has specifically examined the efficacy of PA interventions for improving cardiometabolic and psychosocial outcomes among PWH in sSA. As a result, our interpretations and comparisons will necessarily draw largely on evidence from systematic reviews conducted outside the region or not specifically focused on sSA. In this study, we examined the features and effects of PA interventions aimed at enhancing CMD and psychosocial outcomes in PWH in sSA []. A secondary aim was to describe the interventions and identify the behaviour change techniques (BCTs) that were used in these interventions to help inform best practices for elements of such interventions to be used in practice. Methods We adhered to the guidelines of the Preferred Reporting Items for Systematic review and Meta-analysis Protocols (PRISMA-P) []. The populated PRISMA checklist for the review is shown in . A protocol for the systematic review exists and was registered in PROSPERO (I.D: CRD42021271937) []. Initially, in our protocol we had planned to use the revised tool to assess the risk of bias in randomised trials (RoB 2) []. However, due to a lack of expertise in using RoB 2 within our team, we opted to use the PEDro scale [] in this review instead. We also intended to assess the certainty of evidence using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) guidelines []. Unfortunately, this was not done because a meta-analysis was not conducted due to the extensive heterogeneity of the interventions included in our review. Inclusion criteria We developed a library search to identify published manuscripts with the following criteria []: - Publication date in 2000 or after - Studies including adults (18 years or older) living with HIV in sSA. - Randomized clinical trials or quasi-experimental study designs that evaluated PA-based interventions. No specific comparator or control group was required, but could include passive (e.g., usual care) or active (e.g., alternative behavioural approach) control group. - Primary outcome must include at least one cardiometabolic health indicator, such as body composition including BMI, waist circumference, body weight, waist-to-hip-ratio, lipids, measures of diabetes risk, blood pressure and/or a psychosocial outcome/s, or health indicator, such as HRQOL, anxiety, depression, self-efficacy, and social support. Secondary outcomes are PA outcomes and BCTs used in effective PA promotion interventions. We excluded manuscripts that did not include original data about the effects of PA on cardiometabolic and psychosocial outcomes as well as observational studies, case series, and unpublished pre-prints. Sources of information We conducted systematic searches of five electronic databases: Embase, PubMed, Web of Science, African Index Medicus, and the Cumulative Index to Nursing and Allied Health Literature [], from 15 May 2021 to 10 August 2023. Articles that were published in any language were included, being translated in Google translator when the language was not native to the authors. Reference lists of previous reviews and included papers were searched to check for any further studies. We did not search for any grey literature (e.g., reports, theses, conference abstracts, and non–peer-reviewed documents), as we focused exclusively on published randomised controlled trials and quasi-experimental study designs. We excluded grey literature to ensure the inclusion of methodologically rigorous, high-quality evidence. Peer-reviewed publications undergo formal evaluation processes that enhance the validity, reliability, and transparency of study methods and findings, which is critical for synthesising evidence on PA interventions and their cardiometabolic and psychosocial outcomes []. Grey literature often lacks detailed reporting of study design, intervention components, outcome measurement, and statistical analyses, limiting the ability to critically appraise methodological quality and risk of bias using established tools []. This is particularly important in intervention research, where insufficient reporting can compromise reproducibility and interpretation of effectiveness. While excluding grey literature may increase the risk of publication bias, this limitation is mitigated by conducting comprehensive searches across multiple databases, screening reference lists, and applying transparent inclusion criteria, as recommended by PRISMA guidelines []. The search strategy is available in . Summary of manuscript identification Screening. The full search output was imported into EndNote (Version 20) first, then into Covidence software (Veritas Health Innovation 2020) for screening. The Covidence software assisted in handling the data and study records throughout the review. Two authors (J.M-G and S.Z.M) screened the first 250 abstracts in duplicate, in sets of 50 []. In instances where agreement between the 2 authors exceeded Cohen κ > 0.90 for 2 consecutive sets of 50 abstracts, a single reviewer (S.Z.M.) screened the remainder of the abstracts []. The same strategy was applied for full text screening. Extraction of data Each included study was analysed using a standardized consensual form to obtain the following data: study year, country, study design, intervention setting, outcomes, and participant characteristics such as age and CD4 count []. Intervention features also extracted included: description of the intervention, frequency of the exercise sessions and duration of the intervention, exercise modality, description of the control or comparator intervention, description of primary and secondary outcome measures, and results, by treatment arm []. Missing elements of these features were also captured. We used differences in mean from pre-to post-intervention to estimate the effect of the intervention on cardiometabolic outcomes. These were either calculated or extracted from the published manuscripts. Missing or incomplete data were addressed in a stepwise manner. When key outcomes, sample characteristics, or effect estimates were not reported, we contacted corresponding authors via email. If no response was received after two attempts spaced two weeks apart, the study was retained in the review, and the missing data were marked as “NR” (Not Reported). Ascertainment of quality of the included studies We used the PEDro scale [] to assess the methodological quality of the included randomised controlled trials (RCTs). The PEDro scale scores 11 items: random allocation, concealed allocation, similarity at baseline, subject blinding, therapist blinding, assessor blinding, > 85% follow up for at least one key outcome, intention-to treat analysis, between-group statistical comparison for at least one key outcome and point and variability measures for at least one key outcome. Items are scored as either present [] or absent (0) and a score out of 11 is obtained by summation. An inter-rater reliability generalised kappa statistic of between 0.40 and 0.75 has been reported for the PEDro scale []. Studies that did not reach a cut-off of five points [] were excluded from subsequent analysis. For non-randomised studies, methodological quality was assessed using the Downs and Black [] checklist. Methodological quality was assessed using the Modified Downs and Black checklist, which simplifies the original scoring system, particularly the power item, to improve feasibility and inter-rater reliability in systematic reviews including randomised and non-randomised studies [–]. In particular, the Modified Downs & Black checklist refines the original 27-item tool to better suit non-randomised study designs, such as cohort and case–control studies, by modifying or de-emphasising items specific to RCTs, particularly those related to randomisation concealment and blinding []. Ultimately, this results in a more versatile and balanced instrument for assessing and comparing the methodological quality of both randomised and non-randomised studies in systematic reviews. The item that pertains to the power of the study is modified from a five point score to a one point score which instead assesses whether or not the study performed a power calculation []. Thus, the maximum score of the checklist becomes 28 instead of 32. Downs and Black score ranges were given corresponding quality levels as previously reported []: excellent [–]; good [–]; fair [–]; and poor (≤ []. Because some questions (items 14, 15, 21–25, 27) are not relevant for uncontrolled studies, the maximum score for uncontrolled studies was 20 points. Thus, for uncontrolled studies we considered: excellent [–]; good [–]; fair [–]; and poor (≤ []. Two authors (SZM and JM-G) were tasked with individually assessing the quality of the studies. In cases of disagreement between the reviewers, a third reviewer (MS) was consulted []. The ultimate decision regarding the study’s rating was reached through collaborative discussion and consensus among all reviewers. Synthesis of results Results are presented as a narrative synthesis of data on the efficacy and components/characteristics of successful interventions to enhance cardiometabolic and psychosocial outcomes among PWH in sSA []. Where possible, the effects of the interventions on cardiometabolic and psychosocial health outcomes were assessed using the differences in pre- and post-test means. When not reported, we calculated these differences by subtracting the pre-test mean from the post-test mean. We also synthesised our results using the reported post-test results and the conclusions stated in the publication. In addition, the characteristics of the studies, including (a) name of author and year of publication (b) study design (c) intervention setting (d) outcomes assessed (e) characteristics of participants (f) description of the PA interventions and their efficacy are also presented in tables. Results Study selection In total, 6 406 abstracts were retrieved by the search strategy (). Nine hundred and ten duplicates were removed, and we were left with 5496 unique studies for abstract screening, of which 5478 were found to be irrelevant to the main subject of our study, i.e., 4,320 studies did not assess the efficacy of PA interventions in improving cardiometabolic or psychosocial outcomes but instead focused on other interventions, while 1,158 studies did not address our outcomes of interest. Eighteen studies were then assessed for eligibility through full manuscript review. One study [] was excluded because the study did not assess any of our pre-specified outcomes of interest, leaving 17 studies that were included in the final systematic review. Study characteristics reports the characteristics of the studies included in this systematic review. Sixteen RCTs, and 1 quasi-experimental study were included. The included studies were conducted in Nigeria (n = 8), South Africa (n = 4), Zimbabwe (n = 2), Rwanda (n = 2), and Mozambique (n = 1). Studies were generally small, with sample sizes ranging from 21 to 100. Two studies focused on PWH with concurrent comorbidities, including one with HIV-related poly-neuropathy [] and another on people with antiretroviral therapy-associated body fat redistribution syndrome []. Six of the 17 studies had cardiometabolic outcomes alone [,–], eight had psychosocial outcomes alone [,–] and three included both cardiometabolic outcomes and at least one psychosocial outcome [–]. In addition, two studies had PA as a behavioural outcome [,]. Cardiometabolic outcomes included blood pressure, body weight, waist circumference, hip circumference, waist-hip ratio, body fat, visceral fat, body mass index, skinfolds, cholesterol, LDLs, HDLs, triglycerides, blood glucose, HOMA-IR, and insulin. Psychosocial outcomes included depression, anxiety, perceived stress, body self-image, quality of life and self-efficacy. In the interventions that also targeted PA as a behavioural outcome, PA was assessed using pedometer step count in both studies [,], and in addition, Mabweazara [] also assessed PA using the Global Physical Activity Questionnaire (GPAQ). Description of the interventions Full description of the interventions and their efficacy on improving cardiometabolic and psychosocial health outcomes are presented in . Physical activity intervention trials with cardiometabolic health outcomes Most (7/8) of the interventions included structured interventions with supervised training plans, whereas one was self-led []. The trials with cardiometabolic outcomes evaluated PA interventions with resistance exercise (n = 2) [,], aerobic exercise (n = 4) [,,,] or a combination of aerobic and resistance exercise (n = 2) [,]. Most interventions set goals for moderate exercise intensity (i.e., 40–79% of maximum heart rate) with gradual increases in intensity as the intervention progressed. The interventions set exercise goal ranges from 45-90 minutes per session, with sessions occurring 3 times/week (7/8), and total intervention durations of 6–12 weeks (n = 6), and 6 months (n = 2). Neither the locations of exercise nor the supervisor was reported in most (5/8) studies. Two studies evaluated gym-based programs [,] and another focused on a home-based exercise strategy []. Finally, only one study included education on information on cardiometabolic health risk factors [] in the intervention and no study included diet or lifestyle modification interventions in addition to PA. Physical activity intervention trials with psychosocial health outcomes Most (7/11) of the interventions included structured interventions with supervised training plans. The trials with psychosocial outcomes evaluated PA interventions with resistance exercise (n = 1) [], aerobic exercise (n = 6) [,,–] or a combination of aerobic and resistance exercise (n = 3) [,,].One study did not report the exercise mode used in the intervention []. Most interventions set goals for moderate exercise intensity (i.e., 50–75% of maximum heart rate) with gradual increases in intensity as the intervention progressed. The interventions set exercise goal ranges from 30-90 minutes per session, with sessions occurring 3 times/week (8/11), and total intervention durations in the 6–12 week range(6/11). Five studies did not report exercise location, but 3 studies had a home exercise component [,,]; one study [] also had a community-centre based component and the last [] evaluated a gym-based program. Exercise supervision was not reported in 4 studies, but the intervention was self-led in 2 studies [,], a component led by an exercise scientist in one study [], and led by physiotherapists in two studies [,]. Finally, 3 studies included educational information on cardiometabolic health [] and PA [,]. Only one study incorporated incentivisation as a strategy to promote and maintain participant adherence []. This involved providing fruit packs and food parcels to participants allocated to the exercise intervention arm, which were distributed contingent upon attendance at each scheduled contact session during the intervention period. Exercise monitoring was reported in 3 studies using log books [,,], and pedometers [,]. Physical activity intervention trials with physical activity participation outcomes Two interventions [,] had PA outcomes in addition to cardiometabolic and psychosocial outcomes and are also listed in . The intervention in one [], was delivered in the indigenous Xhosa language and included a community centre-based program and a home-based exercise program. At the community centre, participants were taught simple home-based exercise activities, strategies for increasing walking step count, increasing PA through activities of daily living, role of PA for NCD prevention in PWH, health benefits of PA for PWH, safety when exercising, PA self-efficacy and social support, identification and strategies for overcoming barriers to PA. Participants also engaged in a 30-minute group aerobic exercise session. Participants then received a PA manual with all the taught material and were then instructed to regularly do the home-based exercise activities and walk more frequently. Exercise monitoring was via a pedometer and a PA log book. In another study that measured PA [], participants attended 5 monthly contact sessions where they were taught about ischaemic heart disease (IHD), risk factors for IHD, benefits of PA, strategies for increasing PA, monitoring PA intensity when walking, and safety suggestions when walking in the community. Participants then received a home-based walking program which was monitored via a pedometer and an activity diary. Effects of physical activity interventions on health outcomes The numerical results of selected studies are presented in for cardiometabolic outcomes and for psychosocial outcomes, while presents the efficacy of the PA interventions to improve cardiometabolic and psychosocial outcomes. Effects of physical activity interventions on cardiometabolic health outcomes The most commonly evaluated cardiometabolic health indicators were blood pressure (n = 5) [,,,,], body weight (n = 3) [,,], BMI (n = 4) [,,,], body and visceral fat (n = 5) [,,,,], waist circumference (n = 4) [,,–], cholesterol (n = 4) [,–], HDLs (n = 4) [,,,], triglycerides (n = 3) [,,], glucose (n = 4) [,,,], and waist-hip-ratio (n = 4) [,,,]. By and large, PA interventions evaluated in these studies had positive impacts on cardiometabolic health indicators ( and ). For example, all 8 studies that evaluated the efficacy of PA interventions to improve cardiometabolic health outcomes reported at least one positive impact. The most common improvements reported were on reducing blood glucose in 3 of 4 studies [,,], waist circumference in 3 of 4 studies [,,], body fat in 3 of 3 studies [,,], cholesterol in 3 of 4 studies [,,], waist hip ratio in 3 of 3 studies [,,], triglycerides in 2 of 3 studies [,], and blood pressure in 2 of 3 studies [,]. Two studies reported no impact on some cardiometabolic outcomes: one [] reported no impact of a formal exercise intervention and a playful exercise intervention on blood pressure, HDLs, triglycerides, and blood glucose, whilst another [] reported that a cardiorespiratory training exercise intervention had no impact on hip-circumference, HDLs, LDLs, and HOMA-IR. Using an education and home-based pedometer walking program, one study [] reported a negative impact on weight, waist circumference, and BMI, whilst another study [] also reported negative impacts on weight, BMI, and cholesterol. Effects of physical activity interventions on psychosocial health outcomes The most commonly evaluated psychosocial health indicators were quality of life (n = 6) [,,,,,], depression (n = 2) [,], and self-efficacy (n = 2) [,]. While a single study evaluated perceived stress [], and another evaluated body self-image []. All 11 studies that focused on psychosocial health reported a positive impact on the outcomes of interest ( and ). However, in one study [], the positive impact of the intervention was restricted to quality of life as measured using the state of health visual analogue scale, and only certain aspects of body self-image in the one study which incorporated that measure []. Effects of physical activity interventions on physical activity outcomes also lists the studies that evaluated the effects of PA interventions on PA outcomes. Only two studies assessed PA outcomes [,]. A positive impact on PA was reported for pedometer step count in both studies. Additionally, one of the studies [] reported a positive impact on PA (metabolic equivalent/week) as measured by the GPAQ. Behavioural change techniques identified in physical activity interventions promoting physical activity shows a detailed description of the BCTS identified in the studies that had an effect on PA outcomes. The following BCTs were identified in one study []: information about health consequences, feedback on the behaviour, feedback on outcomes of the behaviour, goal-setting outcome, self-monitoring of the behaviour, material incentive (behaviour), demonstration of the behaviour, problem solving, instruction on how to perform the behaviour, social support (unspecified), social support (practical), goal-setting behaviour, review the behavioural goals, and action planning. Whereas in the other study evaluating PA promotion []: information about health consequences, feedback on the behaviour, goal-setting outcome, self-monitoring of the behaviour, problem solving, and instruction on how to perform the behaviour were all reported. Methodological quality The included RCTs generally scored high on the PEDro scale, ranging from 7 [] to 10 [,,]. Risks identified included absence of blinding of participants, therapists, and assessors who measured key outcomes. Only 3 studies reported blinding of therapists and assessors [,,]. Baseline comparability was challenging in three studies [,,]. Randomisation, concealed allocation, adequate follow-up, intention-to-treat analysis, between-group comparisons, and reporting point estimates and variability were present in all studies. In summary, PEDro scores were as follows: 10/11, n = 3 [,,]; 8/11, n = 9 [,,,,,,,,]; 7/11, n = 2 [,]; and 6/11, n = 1 [] (). The study by Odunaiya [] was a non-randomized quasi-experimental study, assessed using the Downs and Black checklist, and scored 12 out of 20, which was considered a good quality score (). The study scored well in reporting on items 1–10 (7/10), did not score any points on external validity on items 11–13 (0/3), and scored all relevant items (items 14–20) for internal validity. Discussion We retrieved 16 randomized controlled trials and one quasi-experimental trial to examine effects of PA on cardiometabolic and psychosocial health among PWH. Improvements in cardiometabolic health outcomes were reported across a range of markers, including blood pressure, lipids, BMI, body weight, waist circumference, and body fat. Likewise, improvements in psychosocial outcomes included augmented quality of life, reduction in depression, improved self-efficacy, decreased perceived stress, and improved body self-image. Notably, the largely positive impact of PA interventions on health was seen despite the heterogeneity of the PA interventions evaluated, which ranged from structured and supervised programs to individual and community-based models, and from relatively modest interventions, such as 30-to-40-minutes of aerobic workouts 3 times/week for six weeks, to much more substantial interventions, such as 1.5 hours of combined aerobic and resistance training, 3 times/week for 6-months. A key limitation of most trials was the relatively short duration (e.g., 6–12 weeks), making it impossible to detect major clinical events. Notwithstanding, this body of work offers strong preliminary support for the use of PA interventions to improve cardiometabolic and psychosocial health among PWH, and argues for additional research to evaluate these interventions at larger scale. Physical activity interventions have been studied for their effectiveness in improving cardiometabolic and psychosocial outcomes among PWH. A systematic review by O’Brien et al. [] analysed 24 randomized controlled trials (RCTs) involving 936 participants, predominantly male and on antiretroviral therapy. The interventions, which included aerobic exercise alone or combined with resistance training, ranged from 5 to 52 weeks in duration. The findings indicated significant improvements in cardiorespiratory fitness, strength, body composition, and quality of life among participants who engaged in these PA interventions compared to non-exercisers. Notably, no significant changes were observed in CD4 counts or viral load, suggesting that PA interventions are safe and beneficial for medically stable adults living with HIV. A systematic review and meta-analysis by Kalatzi et al. [] examined the effects of supervised aerobic exercise on clinical, physiological, and mental health parameters in PWH. Findings revealed that supervised aerobic exercise significantly reduced depressive symptoms. However, the meta-analysis found no significant changes in CD4 T-cell count or blood pressure, indicating that aerobic exercise did not have a substantial impact on these immunological and hemodynamic markers. Another systematic review and meta-analysis by Aminde et al. [] evaluated the effectiveness of minimally supervised home and community exercise (HCE) interventions in PWH. The analysis encompassed 13 studies involving 857 participants aged between 29 and 56 years. The interventions primarily consisted of aerobic and strength exercises performed with minimal supervision in home or community settings. The results showed that HCE interventions led to significant improvements in PA levels, with an increase of approximately 1,097 steps per day compared to control groups. Additionally, there was a notable reduction in body fat percentage by 3.36%, although no significant change was observed in body mass index (BMI). Quality of life (QoL) also improved, particularly in the physical health domain, with an increase of 13 points relative to controls. Notably, interventions conducted at moderate or higher intensities were associated with greater benefits, and walking-only programs were found to be at least as effective as other forms of exercise. Importantly, no adverse events were reported across the included studies. Lastly, an umbrella review [] prepared for the Guideline Development Group for the WHO Guidelines on PA and sedentary behaviour for children and adolescents, adults and older adults highlights the significant advantages of regular PA for PWH. Results revealed that engaging in appropriately graded exercise can help preserve lean body mass, enhance cardiopulmonary function, bolster immune status, and improve overall quality of life. Additionally, PA was shown to improve quality of life, cardiometabolic health, and cardiorespiratory fitness in this population. Although not specifically focused for PWH in sub-Saharan Africa, the evidence for the benefits of PA shown in these systematic reviews shows that incorporating regular PA is crucial for optimising the health and well-being of PWH. In our study, a greater proportion of participants were female. Similarly, a review examining the health benefits of PA for Black women, including African Black women [], reported that PA improved cardiometabolic and psychosocial outcomes, as well as benefits for pregnancy- and menopause-related conditions. African women encounter distinct challenges and bear a higher risk of obesity compared to men and women in the Global North usually shaped by a complex interplay of individual, cultural, social, economic, and environmental factors []. In South Africa, research shows that 60% of African women are either overweight or obese, with obesity rates among women being five times higher than those observed in men []. Based on our findings, PA appears to be an effective therapeutic approach for improving health outcomes, although the considerable heterogeneity among studies warrants cautious interpretation. In our study, the included studies discussed in the discussion section scored well on the PEDro scale (≥9/11 points). However, most fell short in blinding subjects, therapists, and assessors [,,,,–], as well as in external validity []. Perhaps the most compelling feature of this body of evidence is the consistency in improvements seen across multiple CMD and psychosocial endpoints as shown in . Essentially every study showed a benefit of PA, and the benefits were wide ranging although of relatively modest effect size. All studies examining blood pressure demonstrated potentially significant improvements. The study by John et al. [] reported the most substantial improvements, with reductions in systolic blood pressure by 46 mmHg and diastolic blood pressure by 6 mmHg. Notably, reviews of randomized controlled trials of PA interventions have also demonstrated relatively modest changes in blood pressure (4 and 2 mm Hgb, respectively) []. Thus, these data are consistent with data outside the HIV literature. Similarly, although one study by Maduagwu et al [] showed large improvements in lipids (i.e., a 22 mg/dl increase in HDL and a 55mg/dl decrease in LDL), others, such as that by Roos et al [], Mbayo [], and Mutimura [] were associated with much smaller changes. Smaller differences were also observed in a prior systematic review with nine randomized controlled trials which, although not focusing in PWH in sSA, found that various exercise types had modest but meaningful benefits in terms of lipid profiles and blood glucose levels in PWH []. Finally, improvements in body composition outcomes were also observed, ranging from large reductions in BMI (i.e., a decrease of 4 kg/m2) to a 4% reduction in body fat in the study by John et al. [], > 5% body weight reduction in Mutimura et al. [,], a 7.13 cm reduction in waist circumference in both studies by Mutimura [,], and 0.2 kg/m2 reduction in BMI, 0.5 kg reduction in body weight, 2 cm reduction in waist-circumference, and 1% reduction in body fat in the study by Mbayo []. These magnitudes of change are consistent with general population reviews that demonstrate that regular exercise significantly improves body composition parameters [–]. The impact of PA on body composition are particularly important in the era of modern ART, which has been associated with significant increases in weight in PWH []. It is important to note that some studies reported null effects on certain cardiometabolic outcomes despite improvements in others. For example, Mutimura et al. [] found improvements in central adiposity, cardiorespiratory fitness, triglycerides, and fasting glucose, but no significant changes in HDL-C, LDL-C, or HOMA-IR following the PA intervention. Across methodological scrutiny and comparison with adjacent evidence, these null findings are likely explained by several factors, including: limited baseline risk (“room to improve”), as HDL and LDL levels were not markedly abnormal; study design and power limitations, with analyses not powered for HDL, LDL, or HOMA-IR and reliance on per-protocol rather than intention-to-treat approaches; measurement constraints, including the use of calculated LDL using the Friedewald equation, which is known to possess significant limitations in terms of accuracy [] and the limited sensitivity of HOMA-IR []; and biological influences of HIV and ART, which may attenuate improvements in lipid profiles and insulin resistance []. Importantly, similar trials often report no change in HDL, suggesting this finding is not unusual. Improvements in LDL and insulin resistance appear more likely when baseline abnormalities are greater, exercise dose and intensity are tightly controlled, or more sensitive measures of insulin sensitivity are used []. Mangona et al. [] reported “no exercise effect” on somatic (weight/BMI/body fat/most anthropometrics) and most cardiovascular risk variables (blood pressure, HDL, triglycerides) improvements in cardiorespiratory fitness (VO2peak). The null findings may be best explained by a convergence of (i) limited baseline impairment and ceiling effects (BP, glucose, lipids, and CD4 were generally in relatively normal/stable ranges), (ii) an intervention that likely prioritized fitness over energy deficit/body-composition change (only ~20 min structured aerobic work per session in the formal group; resistance emphasized endurance with low volume), (iii) statistical design/analysis choices (small per-group number of participants, no reported power calculation or primary outcome hierarchy; reliance on within-group paired tests rather than a formal (interaction model) testing whether changes over time differ between groups (e.g., intervention vs control). Importantly, the benefit of PA interventions extended beyond CMD outcomes. Moreover, the benefits were consistent across all studies that considered psychosocial outcomes, even if the outcomes differed. For example, in Odunaiya et al [], an aerobic exercise programme led to improvements in depression, anxiety, and self-efficacy. Although the outcomes differed, the studies by Mutimura et al [] and Shittu et al [] reported improvements in quality of life. These data are consistent with prior reviews in this area. For example, another systematic review [] not focused on PWH in sSA on the efficacy of aerobic exercise for PWH reported statistically significant improvements in selected outcomes of depression and quality of life in the intervention groups than controls. These results are consistent with an extensive meta-review (97 reviews, 1039 trials, and 128,119 participants) in the general population, which found PA to be highly beneficial for improving symptoms of depression, anxiety, and distress across various adult populations, including those with mental health disorders and chronic diseases []. In that review, some of the greatest benefits in terms of psychosocial outcomes were seen among PWH. Interestingly, in that review briefer exercise programs provided greater benefit than extended regimens, and the benefits diminished with longer-duration programs. In our review, interventions to improve mental health conditions and quality of life also had relatively short durations ranging from 30 minutes in Odunaiya et al [] and Shittu et al [] to 1.5 hours in Mutimura et al []. On balance, these data suggest that a maximum therapeutic benefit of PA on mental health conditions may be drawn from shorter duration PA activities. Only two studies in our review incorporated interventions to specifically promote PA among PWH [,]. Both interventions were effective in promoting PA, measured as step-count captured via pedometers [,] and as self-reported in the Global Physical Activity Questionnaire []. Emphasizing increased step-count through walking [,] and incorporating self-led home exercise activities and activities of daily living [], these interventions align with the positive outcomes of a meta-analysis showing the efficacy of walking interventions, especially in group settings, for increasing PA in adults []. Both interventions evaluated in our review implemented activity diaries for self-monitoring and included an educational component aimed at increasing understanding of the health benefits of PA for PWH. Common BCTs incorporated in the interventions included providing information about health consequences, feedback on behaviour, goal setting, self-monitoring, problem solving, and instruction on how to perform the behaviour (). Notably, these interventions had an unsupervised home-based component, an educational component, and were not overly rigorous, which may enhance their potential for feasibility and future scalability. A prominent limitation in many PA intervention studies is the lack of robust long-term follow-up and limited evidence of sustained behavioural change beyond the active intervention period. Although acute increases in PA are commonly achieved during intervention delivery, maintaining these gains and translating them into sustained health benefits remains challenging. Beyond identifying effective BCTs during intervention delivery, greater attention is needed on the role of maintenance strategies to support long-term effects and sustainability. Evidence from long-term reviews suggests that interventions incorporating tailored exercise prescriptions and systematic follow-up, such as booster sessions (e.g., phone calls, text messages, in-person follow-up sessions, or tailored reminders), ongoing contact (e.g., follow-up prompts and self-monitoring), behavioural counselling, or tailored self-management support, are more effective in reinforcing PA behaviour beyond the intervention period and demonstrate better maintenance of PA behaviour compared with interventions that do not include such strategies [,]. Limitations of the current review The studies evaluated were of modest quality. Nonetheless, the interventions in our review varied widely in terms of target populations, types of interventions, delivery personnel, measurement tools, and duration, which prevented the ability to consider pooling the data. Limitations of the study included short follow-up periods, absence of measurements for maintenance after the initial intervention period, absence of active controls, and lack of blinding of study staff when assessing outcomes. This lack of blinding is particularly concerning for interventions targeting psychosocial outcomes. Additionally, small sample sizes and observation period prevented estimation of intervention effects on clinical outcomes such as cardiovascular events. Such data may ultimately be needed to help ensure PA interventions are ultimately adopted at scale and in public health guidelines. We also acknowledge that while the PEDro scale can be helpful in identifying potential methodological issues, it cannot definitively confirm the presence of publication bias without further analysis considering factors such as the overall publication landscape. Future research implications Future research should focus on well-powered, longitudinal studies to determine whether the short-term improvements in cardiometabolic and psychosocial outcomes observed among PWH in sSA are sustained over time and translate into reductions in hard clinical endpoints such as cardiovascular disease, diabetes, and HIV-related multimorbidity. Larger, multi-site trials across diverse and resource-limited settings are needed to enhance generalisability, alongside the use of objective PA measures to improve assessment accuracy. Further studies should identify the key intervention components and BCTs responsible for observed benefits and apply implementation science approaches to evaluate feasibility, scalability, cost-effectiveness, and integration of PA interventions into routine HIV care, while also addressing equity-related factors such as sex, age, treatment regimen, comorbidities, and socio-environmental determinants of PA participation. Policy and practice implications The consistent evidence of cardiometabolic and psychosocial benefits highlights the need to integrate structured PA promotion into HIV care policies and clinical guidelines in sSA, recognising PA as a core component of comprehensive HIV management alongside ART. Health systems should support task-shifting approaches that enable trained lay health workers and community-based providers to deliver low-cost, culturally appropriate PA interventions within existing HIV programmes, strengthening holistic and person-centred care. At the policy level, sustained investment and cross-sectoral collaboration are essential to support scalable, community-based PA initiatives and create supportive environments, particularly in underserved settings, with the potential to improve long-term health outcomes and quality of life among PWH. Conclusions In conclusion, amidst rising concerns about CMD consequences of lifelong HIV infection in the current ART era, this review supports further study of PA to improve CMD risk and psychosocial health among PWH. While preliminary, our results suggest that PA interventions could be a valuable addition to standard HIV care packages in sSA, offering an alternative approach to enhance both cardiometabolic and psychosocial well-being in this population, particularly at high risk for obesity and metabolic conditions []. Future research should explore the most effective exercise modalities, intervention durations and intensities to elicit positive improvements in both cardiometabolic and psychosocial aspects among PWH. Importantly, focus should be on the use of effective BCTs to develop feasible, context-sensitive interventions that can be scaled up to promote PA among PWH, especially in low socioeconomic settings or amongst underserved populations. Funding Statement This research was funded, in part, by the Africa Health Research Institute and Wellcome [Wellcome Strategic Core Award: 227167/A/23/Z], awarded to SZM and MJS. For the purpose of open access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission. This publication resulted from research partly supported by the Harvard University Center for AIDS Research (CFAR), an NIH-funded program (P30 AI060354), which is supported by the following NIH Co-Funding and Participating Institutes and Centers: NHLBI, NIDCR, NIA, NIAID, NIMHD, NIDA, NIDDK, NICHD, NINR, and NCI. Awarded to SZM. Research reported in this publication was also supported by the Fogarty International Center of the National Institutes of Health under Award Number K43TW012862, awarded to SZM. The contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health.