Impact of physical exercises on immune function, bone mineral density, and quality of life in people living with HIV/AIDS: a systematic review with meta-analysis

Background Compromised immune function, associated with human immune deficiency virus(HIV) infection, is improved by antiretroviral therapy(ART) which also decreases bone mineral density(BMD), and possibly the quality of life(QoL). However, physical(aerobic/resistance) exercises, were reported to induce reverse effects in uninfected individuals and were appraised in the literature for evidence of similar benefits in people living with HIV/AIDS(PLWHA). The main study objective was to evaluate the impact of physical (aerobic and resistance) exercises on CD4+ count, BMD and QoL in PLWHA. Methods A systematic review was conducted using the Cochrane Collaboration protocol. Searching databases, up to June 2017, only randomized control trials investigating the effects of either aerobic, resistance or a combination of both exercise types with a control/other intervention(s) for a period of at least 4 weeks among adults living with HIV, were included. Two independent reviewers determined the eligibility of the studies. Data were extracted and risk of bias(ROB) was assessed with the Cochrane Collaboration ROB tool. Meta-analyses were conducted using random effect models using the Review Manager(RevMan) computer software. Results Nineteen studies met inclusion criteria(n = 491 participants at study completion) comprising male and female with age range 22–66 years. Two meta-analyses across 13 sub-group comparisons were performed. However, there were no RCTs on the impact of physical exercises on BMD in PLWHA. The result showed no significant change in CD4+ count unlike a significant effect of 5.04 point(95%CI:-8.49,-3.74,p = 0.00001) for role activity limitation due to physical health(QoL sub-domain). Overall, the GRADE evidence for this review was of moderate quality. Conclusions There was evidence that engaging in moderate intensity aerobic exercises (55–85% Maximum heart rate-MHR), for 30–60 min, two to five times/week for 6–24 weeks significantly improves role activity limitation due to physical health problems, otherwise physical(aerobic or/and resistance) exercises have no significant effects on CD4+ count and other domains of QoL. Also, there is lack of evidence on the impact of exercises on BMD in PLWHA due to the paucity of RCTs. The moderate grade evidence for this review suggests that further research may likely have an important impact on our confidence in the estimate of effects and may change the estimate. Electronic supplementary material The online version of this article (10.1186/s12879-019-3916-4) contains supplementary material, which is available to authorized users.


Background
Human immunodeficiency virus (HIV) infection persists as a global public health issue [1] and is presently regarded as a chronic condition [2] since the advent of Highly Active Antiretroviral therapy (HAART). This is a sequel to the significantly improved life expectancy in people living with HIV (PLWHA) who are HAART-experienced [3]. This view is given further credence by the decreasing number of HIV-related deaths post-HAART era, down from the extreme point of 2.3 million in 2005 to an estimated 1.0 million in 2016 [4]. However, high incidence of comorbidities has been recorded in PLWHA, and might be due to the high rate of metabolic abnormalities resulting from toxic side effects of antiretroviral therapy(ART) [5]. These include low bone mineral density (BMD), and consequent high risk of fracture [6], cardiovascular diseases [7] , and instability of fat metabolism [8], which may be amenable to physical exercises.
Several research studies have supported the role of physical exercise as a complementary alternative therapy in the management of chronic illnesses, and apparently both aerobic and resistance exercises are beneficial to PLWHA [9].
For instance, contemporary literature supports the conclusion that exercise has been the key strategy to improve lean body mass, cardiovascular fitness [10], improve strength [11], change mood state [5], increase BMD, reduce risk of fracture and invariably enhance quality of life (QoL) in PLWHA [12]. These developments may be attributed to the regulatory effects of structured exercise on immune function and probably as a result of expressive traction of muscles to the bones during training protocols.
Structured physical exercise at different intensities and duration has been shown to improve mental health, QoL, immune and physical function in PLWHA compared to an inactive lifestyle [3]. Aerobic exercise, when carried out for 16 weeks (30 min, 3 times/wk) at a moderate intensity, resulted in either an increased or a stable CD 4+ cell count which improves resistance to infection [5]. In fact, Maduagwu et al. [13], reported a significant improvement in the CD 4+ cell in a pre-test and post-test 12 week (40 min, 3 times/wk) experimental study. Similar findings were reported by Ezema et al. [14], who observed a significant increase in CD 4+ cell count and VO 2 max with exercise compared to the control in an 8-week (45-60 min) training program. However, Tiozzo et al. [15], reported a stable CD 4+ cell count with exercise during a 12-week intervention (45-60 min, 3 times/week) study, unlike the control who recorded a significant decrease. Similarly, structured resistance exercise also triggers a specific immune response in PLWHA as reported by Zanetti et al. [16] who demonstrated that resistance exercise for 12 weeks (3 times/week) is effective in boosting the CD 4+ and CD 8+ cell counts with consequent improvement in the integrity of the immune system. Furthermore, Anandh [17] also reported an increase in the CD 4+ cell count, marked improvement in functional capacity and positive changes in QoL in PLWHA after a 12-week progressive resistance exercise (60 min, 3times per week) intervention. Meanwhile, a significant increase in the BMD and maximum muscle strength was observed after strength training, 3 times per week, for 12 weeks [12]. However, it is by investigating the unique impact of physical exercises through relevant studies in PLWHA that its physiological and therapeutic effects on the immune system, BMD, and Qol, in this population may be determined.
Overall, there are mixed reports on whether there are therapeutic benefits associated with physical exercises on the immune system, BMD, and QoL in PLWHA. Therefore, a systematic review of the literature should be conducted to provide a synthetic knowledge that is required to guide practice. To the best of our knowledge, this has not been done and was the focus of this review. Thus, the aim of this systematic review was to determine the impact of physical exercises on immune function, BMD, and QOL in patients with HIV.

Methods
This systematic review was certified according to the International Prospective Register of Systematic Reviews (PROSPERO) on 27 June 2017 (registration number PROSPERO 2017: CRD42017069068).

Eligibility criteria
Eligibility criteria considered for selecting studies in the review include:  [20] to facilitate the management and selection of articles for inclusion. Eligibility questions and forms for the screening of the studies included in the review were then developed, piloted and subsequently refined. The title, abstract and full texts of selected studies were independently screened for eligibility by F.I.O and E.F.E based on the review eligibility criteria. Differences of opinions occurring at any stage regarding inclusion or exclusion were resolved by discussion and reflection, in consultation with S.C.I.

Quality appraisal and risk of bias
Adopting the Cochrane Collaboration Tool for Risk of Bias Assessment (Table 8.5a of the Cochrane Handbook for Systematic Reviews of Interventions) [21], risk of bias for each of the included studies were evaluated by two authors in six key domains: (i) selection bias (random sequence generation, allocation concealment); (ii) performance bias (blinding of personnel and participants); (iii) detection bias (blinding of outcome assessments); (iv) bias due to attrition (incomplete outcome data, including dropouts and withdrawals); (v) reporting bias (selective reporting) and (vi) other bias (other sources of bias not elsewhere addressed) (Additional file 4) [18]. The procedures undertaken to assess each domain for each study was explicitly described and rated as 'high risk' or 'low risk'. The risk of bias in a study was reported as unclear if there were insufficient details in the original study. In such instances, the study investigators were contacted to provide the required details. The judgments for the risk of bias was made independently by the first reviewer and the same with the second reviewer, based on the criteria for judging the risk of bias (Additional file 4) [18]. Both reviewers made judgments regarding the risk of bias independent of each other. Areas of differences were resolved by discussion and reflection, or in consultation with S. C. I.

Data item
Data were collected from variables including authors' references, participants' characteristics, inclusion and exclusion criteria, study sample size, components of the intervention, the intervention setting, who delivered the intervention, the duration of the intervention and follow-up (where available), attrition rate, aspects of outcome assessed, the outcome measurement, methods/ techniques, results, conclusions and funding sources.

Data synthesis and assessment of heterogeneity
The review question of the impact of physical exercises on immune function, BMD, and QOL, in patients with HIV/ AIDS, was answered. In doing this, all quantitative study outcomes which analyzed the effectiveness of these interventions were presented, considered and combined in a proof table. The Proper statistical method was used for different variables: for a continuous variable, weighted mean differences were applied when outcomes are uniform or standard mean difference when different outcomes are used with 95%: CI while for a dichotomous variable, the Risk ratio was applied with 95%: CI. This review also includes a meta-analysis to find pooled effect sizes across studies, using a random-effects model relying on the level of heterogeneity of intervention effects. Heterogeneity was assessed using the Cochrane's χ 2 test (10% significance level) and Higgins I 2 for which values of 25, 50, and 75% shows low, medium and high heterogeneity respectively as stipulated by the guidance in the Cochrane Handbook for Systemic Reviews of Interventions [19].

Data and sensitivity analysis
Investigation and presentation of outcomes were made using the primary outcome. Studies with homogenous characteristics in terms of design, intervention, and comparator(s) were pooled together for meta-analysis using a random-effects model [18]. Heterogeneous studies were interpreted by narrative synthesis following the guidelines of the Centre for Reviews and Dissemination to explore the relationship and findings between the included studies [19]. Sensitivity analysis was done to decide the impacts of studies with a high risk of bias on the general outcomes with and without these studies.

Rating quality of evidence and strength of recommendation
The quality of evidence of the studies was evaluated to determine the strength of recommendation in the systematic review. This was judged utilizing the Grading of Recommendations Assessment Development and Evaluation (GRADE) approach [22] which comprise consistency; design; directness; precision; publication bias and study limitations. The individual study was graded as high risk of bias or low risk of bias, and then again individual evidence statement for this review was graded from 'High Quality' to 'Very Low Quality' according to the criteria (Additional file 5).

Search result
Three different searches were carried out sequentially during the course of this study using the three-primary outcomes separately in the search strategy.

Bone mineral density
The result from all sources gave 18 citations, 6 duplicates were removed and after a full-text screening of 12 articles, none met the study inclusion criteria. (Additional file 7. PRISMA flow Diagram).

Reasons for exclusion
Three full-text articles retrieved (for immune function) were excluded because studies [38][39][40] did not assess immune function as an outcome. For the QoL, the 2 studies [41,42] excluded were non-randomized clinical trials. For bone mineral density, none of the studies were RCT and therefore were all excluded.

Quality of life
Studies that assessed QoL as an outcome were all RCTs.
Five studies had aerobic exercise as intervention [23,24,27,32,33,34], three studies had resistance exercise as an intervention [17,32,35] and two studies combined both interventions [36,37]. Aerobic exercise and combined studies had no exercise [23,33,37], maintain daily  activity [27,36], and Short-wave diathermy as a placebo [34], with counselling [24] groups as control while resistance exercise studies had no exercise [17], usual advice plus normal activities [35] as control groups. One of the three studies for resistance exercise assessed the effects of co-intervention of progressive resistance exercises (PRE)-and Whey protein [32]. This study also included a comparison group of whey protein only ( Table 1).

Participants of included studies Immune function
A total of 639 participants were included in this arm of the review (i.e. participant recorded at baseline). Participant were all HIV male and female adults with CD 4+ cell count ranging from < 150 cells/mm 3 to > 850 cells/mm 3 with age range 22-63 years and were on a HAART regimen. Five studies had participant located in Brazil [16,24,28,29,31], one in India [17], three in Nigeria [13,14,30], and four in U.S. A [23,[25][26][27] ( Table 1).

Risk of Bias in included studies
Overall, 571 participants withdrew from the included studies accounting for~54% withdrawal rate (571/1062 participants at baseline). Withdrawal rates within individual studies ranged from 3% [37] to 60% [35] ( Table 1). However, a high risk of bias due to attrition exists as 15 of the 19 included studies (78.9%) reported withdrawal rates of > 15% while 1 study [16] did not provide information on incomplete outcome reporting. The remaining 3 studies were judged the low risk of bias due to attrition (25%) with withdrawal rates of < 15% [14,24,30]. The withdrawal rate between comparison groups was similar in most groups. Almost all the included studies mentioned participant who was not complying with their exercise intervention or withdrew from the study. (Table 1 shows the proportion of participants who dropped from individual studies). All Authors but one [16] reported information on adherence to the exercise intervention. Adherence rate ranged from 40 to 93% [30,35].

Narrative synthesis
Immune function One study [31] which assessed immune function, and which was not included for meta-analysis combined aerobic and resistance exercise among 27 HAART treated HIV-infected patients (age 45 ± 2 years). This intervention was carried out in Brazil for 12 weeks whereby the exercise group (n = 19) were involved in aerobic training (cycle ergometer) for 30 min of moderate intensity, strengthening exercise (2 sets of 12 repetitions of 5 exercises at 60-80% 12 Repetition Maximum) for 50 min and flexibility exercise (2 sets of 30s at maximum range of motion of 8 exercises) while the control group (n = 8) received no treatment. Immune function (CD 4 and CD 4 %) were determined by specific monoclonal antibodies using fluorescein isothiocyanate and phycoerythrin monoclonal antibodies using a whole blood staining method. The study reported no significant change in the CD 4 T-cell count in either the exercise group or the control group (p = 0.19 for CD 4 T-cells and p = 0.22 for CD 4 %) [31].

Quality of life
Similarly, two studies [36,37] which investigated QoL also employed a combined exercise approach in PLWHA. In Rwanda, Mutimura [37] recruited 100 HIV individuals (21-50 years) with body fat redistribution (BFR) that were on HAART for greater than 6 months to find out the effect of exercise training on QoL. The participants (exercise group) were involved in a 6-month supervised exercise (EXS) programme which consists of warm-up exercise followed by 45-60 min of jogging, running, stair climbing, low-back & abdominal stabilization and strengthening exercises while the control group did not undergo any treatment. The outcome of interest (i.e. QoL) was measured at baseline and after   Comment: Probably appears to be free from other sources of bias.

Low risk High Quality
Terry 1999 Yes Comment: Missing data adequately accounted for.

Yes
Quote: "Before the program, at 6 weeks and the end of the program blood was collected…" Comment: Probably done.

Yes
Comment: Probably appears to be free from other sources of bias. 6 months using WHOQOL-BREF (short-form instrument) evaluating physical, psychological, independence and social relationship domains. Findings from this study state that over the 6-month training period, significant improvements were observed in BFR + EXS group for the psychological, independence, social relationships (p < 0.001) domains of QoL compared to BFR + noEXS group [37]. The other study also with a combined intervention on functional outcomes and QoL [36] was also carried out in Rwanda with 38 PLWHA (20-60 years) on HAART greater than/ 3 months. The EXS group received 8 weeks of exercises with the goal to foster strength, endurance, and cardiovascular exertion while the control group maintained normal daily activities. QoL outcome was assessed using the Medical Outcome Short Form (MOS-HIV) and report states that there is an improvement in QoL from participants in EXS group when compared to control group in overall health perception subscale (p = 0.04) [36]. The above studies [36,37] from inference have a common relationship in setting, type of intervention, study design and outcome of interest. Results from both studies are also similar suggesting a positive effect of combined aerobic and resistance exercise on PLWHA on HAART (> 3 months). The robustness of this synthesis is low and should be interpreted with caution because of the quality attributed to the studies (having a high risk of bias) judging by Cochrane risk of bias assessment tool [18].

Meta-analyseseffects of interventions
This review conducted two meta-analyses across thirteen sub-group comparisons which included meta-analyses for immune function (CD 4+ cell count) and QoL. The sub-group comparisons of the meta-analyses were: 1) Aerobic exercise compared to no exercise as a control 2) Aerobic exercise compared to normal routine activity as control 3) Moderate intensity aerobic exercise compared to high-intensity aerobic exercise 4) Aerobic exercise compared to other controls (conventional, counseling and soft stretching plus relaxation) 5) Resistance exercise compared to control (no exercise and maintain normal routine activities). 6) Physical activity limitation due to health 7) Role activity limitation due to physical health 8) Bodily pain 9) General health perception/overall 10) Vitality/Energy 11) Social limitation due to physical/emotional problems 12) Role activity limitation due to emotional problems 13) Mental health.
Three of the included studies compared aerobic exercise with no exercise control group [23,25,26]. Three studies compared aerobic exercise with normal routine activity [13,27,30]. Two studies compared moderate-intensity aerobic exercise to high-intensity exercise [27,29]. Ezema [14], Ogalha [24] and Terry [28], compared aerobic exercise with conventional therapy, Counselling, and soft stretching plus relaxation as control respectively (Table 1). Two studies compared resistance exercise to either no exercise or maintain normal activities [16,17]. For the QoL outcome, three studies included in the review [24,33,34] compared aerobic exercise with control. In all domains of QoL, only role activity limitation due to physical health showed a significant difference. Meta-analyses were limited for the QoL due to the outcome reported and outcome tool for the remaining studies.
Heterogeneity Heterogeneity (p < 0.1) was evident in the 2 meta-analyses which could be as a result of the differences in gender, location, and type of intervention. Sensitivity analysis was carried out with those greater than two studies since  -Fig. 1). Results demonstrated a non-significant trend towards an increase in CD 4+ cell for participants in the aerobic exercise compared to no exercise; aerobic exercise compared to normal routine activity; moderate intensity aerobic compared to high-intensity aerobic exercise; aerobic exercise compared to other controls and resistance exercise compared to control. The results of the metaanalyses indicated no difference in the CD 4+ cell count for aerobic exercise compared to no exercise; aerobic exercise compared to normal routine activity; moderate intensity aerobic compared to high intensity aerobic; aerobic exercise compared to other controls. Likewise, no difference in CD 4+ cell count was found for resistance exercise compared to control. II. Heterogeneity -CD 4+ count All the sub-group meta-analyses were statistically significant for heterogeneity (p < 0.1). Sensitivity analyses did not indicate a change in the overall effects of exercise on CD 4+ cell count. III. Quality of life Ten of the nineteen included studies assessed QoL. Meta-analyses were performed for three studies Fig. 1 Forest plot for CD 4 + cell count [24,33,34] on 8 sub-domains for QoL. The result of meta-analyses showed no statistically significant difference in all but one sub-domain (role activity limitation due to physical health) (forest plot- Fig. 2). This represents a clinically important improvement in role activity limitation due to physical health compared to control (no exercise, placebo, and counseling). IV. Heterogeneity -QoL All but one sub-group meta-analyses were statistically significant for heterogeneity (p < 0.1). Sensitivity analyses indicated the same overall effect of exercise on QoL.

V. Grade rating
The effect estimate demonstrated the overall significant effect of 5.04 points (95% CI: -8.49, − 3.74, p < 0.00001) for role activity limitation due to physical health (Qol sub-domain) when comparing aerobic exercise group to controls, which can be accepted as moderate evidence. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it may be substantially different. This outcome was downgraded from high to moderate GRADE quality of evidence, because of the inability of authors to conceal allocation in the assignment of participants to experimental and control groups.

Discussion
Meta-analyses for immune function (CD 4+ cell count) showed that neither aerobic exercise nor resistance exercise intervention carried out for a period of 8-24 weeks had an impact on CD 4+ cell count of PLWHA. The results also showed that neither aerobic exercise nor resistance exercise is safe for medically stable adults living with HIV/ AIDS. This finding is based on few available reports/literature on the impact of either aerobic or resistance exercise on CD 4+ cell count in HIV patients and also on participants that finished their exercise protocol prescription and for whom adequate follow-up data were provided.
Most of the studies that assessed CD 4+ cell count were included in a meta-analysis [13, 14, 16, 17, 23-30, ]. Generally, evaluating either aerobic exercise or resistance exercise with controls among PLWHA suggests minute to no impact on CD 4+ cell count. This agrees with the conclusion from previous reviews by Jaggers & Hand [5] , Nixon [10], and O'Brien [43,44]. Aside from these meta-analyses, individual studies that assessed CD 4+ cell count reported increase in CD 4+ cell count of effect size (Cohen's d) 0.9, 0.7, 0.8, 0.2, 0.9, 0.4, 1.2 [13,14,16,17,24,25,30] respectively, suggesting a mean/pooled (0.73) large effect of both aerobic and resistance exercise on HIV patient while the remaining studies reported minimal or no change in CD 4+ cell count [23,[26][27][28][29]31]. Among the Ten studies that assessed the QoL as an outcome, Meta-analyses conducted for three studies [24,33,34] suggest that aerobic exercise only demonstrated significant improvement in one sub-domain of QoL assessment (role activity limitation due to physical health). These findings give credence to an earlier review by O'Brien [44] that the QoL improved for physical function, role emotional, and mental health sub-domains among exercisers compared to non-exercisers.
Individual studies' results show that progressive resistance exercise group (PREG) tends to record improved physical activity score, general health perceptions, and vitality [32]. PREG group improved in QoL assessed by MOS scales [17,23] and another using health visual analog [35]. Aerobic exercise also significantly improved the QoL (physical and mental component; general health, vitality, and mental health) of the experimental group compared to the control [24,34] respectively. Likewise, Stringer [27] reported an improvement for both moderate-intensity aerobic exercise and heavy intensity aerobic exercise when compared to the control group while combined intervention (aerobic and resistance) improved several components of QoL [36,37]. In contrast, Gillespie et al. [33] reported no improvement between exercise and control group using MOS-HIV scale.

Bone mineral density
Evaluation of the impact of physical exercise on BMD was not included in this review due to the paucity of RCTs in PLWHA, assessing BMD as an outcome.

Quality of evidence
Since the interventions in this review were aerobic exercise and resistance exercise, the possibility of blinding participants and exercise supervisors, was not feasible. The exact impact of the absence of blinding on the extent and bearing of the treatment impact is unclear, and however, may constitute a high risk of performance bias [45][46][47][48][49]. On the other hand, some outcomes were self-reported, which also prompted a high innate risk of detection bias when blinding of members was unrealistic. The researchers chose not to downgrade studies for this 'risk of bias' item alone. However, reasons for allotting studies a high risk of bias were not due to lack of blinding. Furthermore, the GRADE evidence of this review was of moderate quality suggesting that further research may likely have an important impact on our confidence in the estimate of effect and may change the estimate.

Implication for practice
Engaging in moderate intensity aerobic exercises (55-85% Maximum heart rate-MHR), for 30-60 min, two to five times per week for 6-24 weeks can lead to significant improvement in role activity limitation due to physical health problems (a sub-domain of QoL), otherwise physical (aerobic or/and resistance) exercises have no significant effects on immune function (CD 4+ cell count) and other domains of QoL. However, lack of RCTs on the effects of physical exercises on the BMD in PLWHA makes it difficult to reach a scientific conclusion that will guide practice. The findings from this study are very important because a previous study [50] has shown that functional limitations of PLWHA affect their care needs and ability to perform social roles such as employment. According to Crystal et al. [50], the limitation in complex roles among PLWHAsuch as working at a job, working around the house, or going to school -was more prevalent than limitation in most specific physical tasks. Among physical tasks, limitation was more prevalent in energy-demanding activities including climbing stairs (43%) or walking > 1 block (26%) than in self-care tasks such as bathing and dressing (14%). Moreover, symptom intensity, pain, and fatigue were strongly associated with these limitations. Therefore, the evidence from this study recommends physical exercises (specifically moderate intensity aerobic exercises) as an effective clinical tool for addressing the wellbeing of PLWHA such that improved management of these disease symptoms using physical exercises might improve physical and social functioning at no extra financial cost, and with little or no side effects to the PLWHA.

Implication for research
The research found few studies on the impact of physical exercises (i.e. aerobic or resistance exercise or the combination of both exercise types) on immune function, and QoL for adults living with HIV, but none on the BMD. Moreover, the few studies identified are also of relatively low quality. All these reasons cumulate and emphasize the need for further research studies in this area. The lack of RCTs on the impact of physical exercises on BMD in PLWHA is an important gap that needs to be addressed in future studies. The increasing reports that bone demineralisation is amenable to physical exercises in non-HIV infected population [51] highlights the possibility of the translational relevance of such studies in HIV/HAART-experienced PLWHA considering the fact that HIV/HAART are associated with bone loss. The study population included both younger and older adults, as well as both male and female participants without considering the variation in the age of participants which could possibly influence the results. The small number of participants per study, that completed the exercise intervention (n = 6-52) suggest that the included studies may have lacked sufficient power to detect treatment effect or that the sample size is smaller than what would be required to detect a clinically important benefit. Therefore, further studies should address the limited number of RCTs and the weaknesses of the available studies as already mentioned above, otherwise it will be difficult to reach a scientific conclusion that will guide practice on the effects of exercises on such parameters as CD4 + cell count, BMD, and other domains of the QoL.