Exercise Training Reduces Central Adiposity and Improves Metabolic Indices in HAART-Treated HIV-Positive Subjects in Rwanda: A Randomized Controlled Trial

Study population

Participants were screened for eligibility from August to December 2005 at the Centre Hospitalier Universitaire de Kigali, Treatment and Research AIDS Centre, and HIV/AIDS clinics of Kimironko, Kicukiro, Bilyogo-Nyiranuma, Kinyinya, and Kacyiru health centers in Rwanda. Eligible HIV-positive volunteers were between 21 and 50 years old and on a stable WHO-recommended HAART regimen for ≥6 months. Participants had moderate to severe BFR, determined by physical examination and subjects’ self reporting, and rating changes in fat content using a validated questionnaire.²⁸ The degree of body fat redistribution was rated as absent (score 0), mild (noticeable on close inspection, score 1), moderate (readily noticeable by the patient and the physician, score 2), or severe (readily noticeable to a casual observer, score 3). The overall score was the mean of the scores given by the participant and a score assigned to each participant by a consensus of three clinicians working in the field of HIV/AIDS. The presence and rating of BFR were confirmed in all participants by physical examination, in which 18% of HIV⁺ participants who self-reported moderate to severe BFR were excluded due to lack of confirmation from the clinicians. For the purposes of this study, a clinical diagnosis was given to HIV⁺ participants with moderate (score 2) to severe (score 3) BFR, and an overall mean score of ≥18 on a validated seven-item inventory for the face, neck, arms, breasts, abdomen, buttocks, and legs, with 21 as the highest score. All included HIV⁺ participants with BFR had a waist-to-hip ratio (WHR) ≥0.90 in men and 0.85 in women.

Subjects were excluded if they had emotional distress or psychosis, unstable angina, shortness of breath with exercise, a known medical history that would contraindicate exercise training, acute infections, or AIDS-defining opportunistic illnesses. Subjects were also excluded if they were obese [body mass index (BMI) ≥ 32 kg/m²], had musculoskeletal or neuromuscular disorders, participated in regular exercise (≥3 sessions/week, ≥30 min/session in the previous 4 weeks), were unwilling to exercise continuously for 6 months, or anticipated changes in HAART medications or were currently taking medications that affect lipid and glucose metabolism. Pregnant or breast-feeding women were excluded. Participants were receiving WHO-recommended first line HAART regimens, with 80% receiving stavudine, lamivudine, and nevirapine, a widely used first-line therapy in resource-limited regions of the world. Another 11% received zidovudine, lamivudine, and nevirapine and 5% received zidovudine, lamivudine, and efavirenz. A small percentage of subjects (3%) received stavudine, lamivudine, and efavirenz and only 1% received lamivudine, abacavir, and nevirapine. HAART regimen use was not different among the HIV⁺ participant groups.

One hundred and fifty-two HIV⁺ participants with moderate to severe BFR were screened and 52 were not eligible; 21 participants did not meet eligibility criteria for central obesity, 16 participants were determined unfit for the exercise program, 11 participants declined to participate, three anticipated changes in HAART medications, and one participant had epilepsy. One hundred HIV⁺ participants with moderate to severe BFR were stratified by gender and randomized to either an exercise group (n = 50; HIV⁺BFR⁺EXS) or a nonexercising control group (n = 50; HIV⁺BFR⁺noEXS). For comparative purposes, we also randomly recruited age, gender, and BMI-matched HIV-infected participants on stable HAART for ≥6 months, without BFR changes (as defined above) (n = 50; HIV⁺noBFR⁺ noEXS). Since there are few normative data on cardiorespiratory fitness and metabolic and body composition parameters in HIV-seronegative Rwandan adults, we also randomly recruited age, gender, and BMI-matched HIV-seronegative healthy adults as an additional comparative control group (n = 50; HIV-ve⁺noEXS). After obtaining ethical permission from the National Research Ethics Committee (Rwanda), written informed consent was obtained from all volunteers before participation.

Metabolic and anthropometrics

After an overnight fast, a venous blood sample was obtained and serum was isolated and analyzed for total cholesterol, high-density lipoprotein cholesterol (HDL cholesterol), triglycerides (TG), and glucose and insulin levels (Humalyzer 3000; Abbott Laboratories, Abbott Park, IL). The CD4⁺ count was quantified by flow cytometry using monoclonal antibodies (Simultest CD4, Beckon Dickinson). Low-density lipoprotein cholesterol (LDL cholesterol) was calculated from total cholesterol, HDL cholesterol, and TG using the Friedewald formula.²⁹

Height, weight, and hip and waist circumferences were measured while the participant was standing, wearing light clothing and no shoes. Height and weight were measured to the nearest 0.1 kg and 0.1 cm, respectively. Waist and hip circumferences were assessed by two well-trained research associates. For reproducibility and to minimize interobserver variability, associates were trained on proper anatomic placement of the tape measure, and their measurements were cross-validated on a number of participants until the variability between duplicate measures among associates was minimal. Waist circumference was measured using a cloth tape measure at the narrowest circumference, halfway between the lowest ribs and iliac crests. Hip circumference was measured at the level of the anterior superior iliac spine, where this could be palpated, or at the broadest circumference below the waist. Two measurements were taken; if they differed by >2 cm, a third measurement was taken and the mean of the closest two measurements was used to calculate waist-to-hip ratio (WHR). Mid-triceps, mid-biceps, suprailiac, and subscapular skinfolds were measured using Lange skinfold callipers. For reproducibility, an experienced trained investigator (E.M.) located and measured all skinfolds in triplicate. Measurements were read to the nearest 0.2 mm and averaged for each skinfold site. The sum of four total skinfold measures was used to calculate percentage body fat (% BFM) and lean body mass (% LBM) using equations and formulas^30,31 validated in a black population.³² All measurements were obtained at baseline and after 6 months of observation, placebo, or exercise intervention.

Exercise training protocol

HIV⁺BFR⁺EXS participated in a 6-month, supervised exercise training program (3 sessions/week, 1.5 h/session, alternating days) at Amahoro Fitness Club located in downtown Kigali. All participants were instructed on a progressive exercise program consisting of stretching exercises and 15 min of brisk walking followed by 45–60 min of jogging, running, stair climbing, low back and abdominal stabilization, and strengthening exercises. Participants were instructed to perform primarily jogging and running exercises with a goal of achieving at least 45% age-predicted maximal heart rate (HR) in the first 3 weeks, 60% age-predicted maximal HR in the next 6 weeks, and 75% age-predicted maximal HR by the end of the intervention. Heart rate was approximately assessed in one-third of participants in each training session with all participants age-predicted maximal HR assessed at least every week. This ensured that the exercise stimulus was adequate to elicit cardiorespiratory and musculoskeletal system adaptations. Each exercise training session ended with 15 min of cool down and stretching exercises. Participants’ attendance at each session was recorded and noncompliance was defined as 50% missed training sessions. Cardiorespiratory fitness was assessed at baseline and 6 months later using the 20-m multistage shuttle run test (20mMST)³³ to predict maximum oxygen consumption (VO₂ peak; ml/kg/min). During the 20mMST, participants ran continuously in an indoor fitness stadium on a flat hard synthetic surface between two points separated by 20 m at a pace that was incremented by 0.14 m/sec each minute or level. The required running pace was indicated by verbal cues and prerecorded audio signals that sounded when each 20-m “lap” should have been completed (much like a metronome). The inability to maintain the required pace for a particular increment was documented (level and number of 20-m laps completed) and represents the participants’ maximal effort, which is proportional to VO₂ peak. The 20mMST assessment tool for cardiorespiratory fitness and exercise capacity has been validated in health and fitness settings.³⁴ Each participant’s cardiorespiratory fitness was assessed individually during the 20mMST to minimize possible sources of error. Heart rate and blood pressure were taken immediately before and after the exercise test, and perceived exertion was evaluated with the Borg Rating of Perceived Exertion Scale (Borg RPE).

Three HIV⁺BFR⁺EXS participants did not complete the CET program: two dropped out, one died following surgery, and another one changed jobs to a distant location. One HIV⁺BFR⁺noEXS participant was lost to follow-up and three HIV⁺noBFR⁺noEXS subjects dropped out; one declined further participation and two started to obtain medications and social support from a rural nongovernmental organization. Seven HIV-ve⁺noEXS participants declined further participation due to lack of time, interest, family issues, or personal travel arrangements. Compliance with the exercise program was 82.2%. Figure 1 represents a flow diagram of study subjects.

Statistical analyses

Based on previous studies,^22–24 we estimated that a sample of 37 individuals in each group would have a power of 90% to detect a 10% difference in triglycerides and total cholesterols for α = 0.05. Assuming a drop-out of 20% per group, at least 44 participants would be required per group. Since all subject groups were ≥44 per group after the intervention, the primary efficacy analyses were based on data from participants who completed the study at a 6-month period.

Between-group comparisons of the changes in the outcome variables were performed using analyses of covariance (ANCOVA). Analyses of covariance with the 6 month change as the dependent variable (absolute change from baseline), the treatment group as the independent variable, and baseline values as the covariates, were used to determine whether there were between-group differences at 6 months after adjusting for the initial baseline values. As part of the ANCOVA, statistical contrasts were used to make pairwise comparisons across groups. All analyses were performed using SAS software, version 9.1.3 of the SAS System for Linux (SAS Institute Inc., Cary, NC). All statistical tests were two-tailed, and significance was accepted at a p-value = 0.05. Data are presented in tables as arithmetic mean ± SD, or median (interquartile range), and in figures as mean ± SE, unless stated otherwise.

Cardiovascular outcomes

At baseline, participants with BFR randomized to the exercise or nonexercise group were well matched for age, CD4 cell count, body fat redistribution score, gender, body composition, and metabolic variables (Table 1). HIV-seronegative controls had lower diastolic blood pressure at the start of the 20mMST than HIV⁺noBFR⁺noEXS subjects (p < 0.05). At 6 months, HIV⁺BFR⁺EXS achieved a significantly higher HR and RPE at the end of the 20mMST (p < 0.001) than other subject groups. Likewise, predicted VO₂ peak during the 20mMST increased more in the HIV⁺BFR⁺EXS group (4.7 ± 0.56 ml/kg/min, p < 0.0001) than in all other groups (Fig. 2). Following CET, the mean changes in CD4 cell count for HIV⁺BFR⁺EXS (19 ± 91 cells/μl) were not significantly different from that of HIV-infected subjects with BFR in the HIV⁺BFR⁺noEXS (34 ± 128 cells/μl) (p = 0.547) or HIV⁺noBFR⁺noEXS subjects (77 ± 160 cells/μl) (p = 0.079).

Body composition outcomes

After 6 months of CET, BMI, waist circumference (−7.13 ± 4.4) (Fig. 3) and WHR declined more in HIV⁺BFR⁺EXS, while it remained unchanged or increased in the other groups (Table 2). Hip circumference was unchanged after 6 months in all groups; therefore the decline in waist-to-hip ratio in HIV⁺BFR⁺EXS was attributed to the decline in waist circumference. In fact, waist circumference increased 1.84 ± 9.2 cm in HIV⁺noBFR⁺noEXS subjects. We also observed a decline in WHR in HIV⁺BFR⁺noEXS participants compared to HIV-ve⁺noEXS participants (0.00 ± 0.1 versus 0.01 ± 0.1; p < 0.05). After the CET, waist circumference in the exercise group was not different from that of HIV-seronegative subjects (Table 2). The BFR mean score significantly declined more in HIV⁺BFR⁺EXS than HIV⁺BFR⁺noEXS, while the BFR mean score significantly increased in HIV⁺noBFR⁺noEXS (Table 2). Subscapular (−1.9 ± 3.2 mm), suprailiac (−2.13 ± 3.5 mm), and sum skinfold thicknesses (−6.15 ± 8.2 mm) (Fig. 3) declined more in HIV⁺BFR⁺EXS than in all other groups (p < 0.0001). HIV⁺BFR⁺EXS experienced a significant decrease in % BFM (−1.5 ± 3.3, p < 0.001) and a significant increase in % LBM (1.5 ± 3.1, p < 0.0001) compared to all other groups (Table 2).

Metabolic outcomes

At baseline, total cholesterol was significantly greater in HIV⁺BFR⁺EXS and HIV⁺BFR⁺noEXS than HIV⁺noBFR⁺noEXS and HIV-ve⁺noEXS (Table 3). Following CET, total cholesterol levels declined more in the HIV⁺BFR⁺EXS group than the HIV⁺BFR⁺noEXS group (p < 0.05). HIV-ve⁺noEXS showed significantly lower mean changes in total cholesterol than HIV⁺BFR⁺EXS (p < 0.0003). There were no significant changes in HDL cholesterol, LDL cholesterol, and HOMA values, but insulin levels declined more in the HIV⁺BFR⁺EXS than in the HIV⁺BFR⁺noEXS group (Table 3). Subjects in the exercise group demonstrated a significant decrease in fasting plasma glucose than HIV⁺ subjects with BFR in the nonexercise group [−0.21 (0.71) versus 0.39 (1.25), p < 0.0001] and HIV⁺noBFR⁺noEXS subjects (p < 0.002) (Fig. 3). Interestingly, HIV-infected subjects with BFR- and HIV-infected subjects with no BFR in the nonexercise groups had a significant increase in fasting plasma glucose, whereas HIV-seronegative subjects did not. Also the serum TG level declined more in HIV⁺BFR⁺EXS than in HIV⁺BFR⁺noEXS and was not changed in other subject groups (Time 3, Fig. 3).