The effects of high-intensity interval training versus continuous moderate-intensity exercise on body composition among older adults with HIV

Abstract

We compared two exercise intensities on lean and fat mass among older people with HIV (PWH). The High Intensity Exercise to Attenuate Limitations and Train Habits (HEALTH) randomized sedentary PWH ≥50 years to 16 weeks of high-intensity interval training (HIIT) or continuous moderate-intensity exercise (CME), both with resistance exercise. Body composition was measured using dual-energy x-ray absorptiometry (DXA) at baseline and week 16. The primary outcome was the percent change in body fat percentage (ratio of fat mass to total mass). Secondary outcomes included percent change in lean, fat, appendicular lean (ALM), and total mass. Linear regression models examined between- and within-group changes from baseline to week 16. Of the 95 participants with pre- and post-DXA scans, the median (interquartile range [IQR]) age was 58 [54-61] years, 14% female, 14% Black, and 14% Hispanic. Fat percentage decreased by −3.3% (95% confidence interval [CI]: −5.2, −1.6) in HIIT and −3.3% (95% CI: −5.3, −1.6) in CME. Fat mass decreased by −3.5% (95% CI: −5.9, −1.0) in HIIT and −3.8% (95% CI: −6.3, −1.3) in CME. Lean mass increased by 1.7% (95% CI: 0.7, 2.7) and ALM by 3.3% (95% CI: 1.3, 5.3) in HIIT and 1.2% (95% CI: 0.2, 2.2) and ALM by 2.6% (95% CI: 0.6, 4.6) in CME. Although both arms had significant improvements compared to baseline, no between-arm comparisons were statistically significant. Supervised HIIT or CME combined with resistance training for 16 weeks is an effective strategy for improving body composition in previously sedentary older PWH.

Clinicaltrials.gov: NCT04550676

Introduction

People living with HIV (PWH) experience an accentuation or acceleration in the aging process,¹ with earlier occurrence and greater prevalence of many comorbid conditions such as metabolic dysregulation and cardiovascular disease.^2,3 Available evidence shows that obesity prevalence in PWH is increasing in parallel with, and in some subgroups surpassing, that of people without HIV, further contributing to the development of comorbid conditions.³ Although modern antiretroviral therapy (ART) tends to be better tolerated with fewer side effects, contemporary regimens are thought to contribute to the rising prevalence of obesity.^4–6 This weight gain, combined with an increased risk for loss of muscle mass among PWH, is associated with an increased risk for physical function decline, frailty, greater chronic comorbidity burden, and all-cause mortality both in PWH and the general population.^7–12

Although there are effective pharmacologic treatments to reduce excessive fat mass, such as glucagon-like peptide-1 receptor agonists (GLP-1 RAs), rapid weight loss in the setting of these therapies is often accompanied by loss of muscle mass, including among PWH.¹³ Additionally, loss of muscle mass that occurs in the setting of weight loss is not regained during a weight rebound following cessation of the treatment.¹⁴ Few pharmaceutical treatments, and none currently approved for use in clinical practice,^15,16 successfully target both fat mass reduction and muscle mass preservation.

Exercise interventions are effective at reducing fat mass while being the most effective way to increase muscle mass, thereby improving overall body composition.¹⁷ Among PWH, we previously found that 6 months of combined aerobic and resistance exercise improved body composition, with loss of fat and gain of lean (a proxy of muscle) mass.¹⁸ High-intensity interval training (HIIT) is a popular form of aerobic exercise that produces superior cardiometabolic benefits compared to continuous moderate-intensity aerobic exercise (CME) in some populations.¹⁹ HIIT has been shown to elicit greater fat mass loss than CME alone, and when combined with resistance training to improve muscle mass may be superior for improving overall body composition.²⁰ Extensive prior literature on aerobic or resistance exercise among PWH is primarily among younger PWH, earlier in the HIV epidemic, when body composition was focused on HIV wasting or ART-related changes.^21,22 Evidence comparing different types of exercise is limited. To the best of our knowledge, only one study has compared HIIT and CME on body composition and was limited to 19 participants (with only 1 female).²³ Determining whether HIIT provides superior benefits to CME has important implications for optimizing exercise prescriptions in this population. Furthermore, current use of contemporary ART, exposure to older thymidine analogue-based ART regimens, and use of statins may influence fat and muscle metabolism as well as the effects of exercise on body composition in older PWH.^24,25

The purpose of this analysis was to compare the efficacy of HIIT versus CME on the body composition of older PWH. The secondary objective was to compare the changes in body composition in clinically relevant subgroups, specified a priori (eg, age, sex, and prior medication exposures).

Methods

Study design and participants

This is a pre-specified secondary analysis of the High Intensity Exercise to Attenuate Limitations and Train Habits among Older Adults with HIV (HEALTH) trial (NCT04550676). This study enrolled sedentary older adults with HIV at the University of Colorado (Aurora, CO, USA) and the University of Washington (Seattle, WA, USA). Participants were enrolled from March 2021 to August 2024, and were included if they were: (1) aged 50 years or more; (2) virologically suppressed (HIV-1 RNA level <200 copies/mL) for at least 12 months; (3) sedentary (defined as self-report of <3 days a week of physical activity that breaks a sweat over the 3 months preceding study enrollment); and fatigued (≥2.0 on either of the first two screening items on the HIV-Related Fatigue Scale). Additional details regarding inclusion and exclusion criteria have been previously published.²⁶

After all baseline assessments were completed, participants were randomized 1:1 to 16 weeks of either HIIT or CME training, using permuted blocks balanced by site, biological sex, and age. The duration of the exercise training was selected in the parent HEALTH trial based on feasibility considerations and prior evidence demonstrating meaningful outcome improvements from exercise interventions of similar length.²⁶ Although the participants and trainers were not blinded to the intervention due to the observable differences in HIIT and CME, the site’s primary investigators were anonymized to participant randomization assignment.

Exercise intervention

The intervention consisted of aerobic and resistance exercise performed on 3 days per week for 16 weeks. Each session started with the randomized aerobic exercise (walking or jogging). The aerobic exercise intensity was prescribed using heart rate reserve (HRR), calculated from the resting heart rate and maximum heart rate obtained during a graded exercise treadmill test performed at baseline. For the first 8 weeks, participants in each arm were familiarized and progressed according to their respective aerobic protocols. For example, participants in the HIIT arm increased the number and intensity of the high-intensity bouts, and participants in the CME arm increased the duration of time walking at their heart rate goal. By week 8, participants in the HIIT arm performed five 4-minute bouts at 90% of HRR, alternating with four 3-minute bouts at 50% HRR for a total of 42 minutes/session in the remaining 8 weeks. In the CME arm, after the 8-week ramp-up period, participants walked/jogged continuously for 50 minutes at 60% of HRR for the remaining 8 weeks. Both arms performed the same resistance training protocol, which consisted of 3 sets of 8-10 repetitions at 70%-80% of their 1-repetition maximum (1-RM) of leg press, chest press, and lateral pulldown using machine-loaded weights. Further details on the exercise intervention have been previously published.²⁶

Covariates

Demographics, social and behavioral factors, comorbidities, and medications were self-reported at baseline, with comorbidities and medications additionally confirmed in the medical records.

Measurement of body composition

Body composition parameters were measured at baseline and week 16 using dual-energy x-ray absorptiometry (DXA) on the Hologic Horizon W instrument (Hologic Inc., Bedford, MA, USA) at the University of Colorado and GE Lunar Prodigy (GE Healthcare Lunar, Madison, MI, USA) at the Prevention Center at Fred Hutchinson. DXA is a valid and reliable measurement of fat and lean mass and is the most common imaging technique for lean mass among people with HIV.^27,28 Body fat percentage was calculated as the ratio of fat mass to total body mass. Appendicular lean mass (ALM) was calculated as the sum of the lean mass of the upper and lower extremity limbs (kg). Body mass index was calculated using the total weight at the time of the DXA scan (kilograms/meters²). Visceral fat from the Hologic software was measured in the abdominal region (from L4/5 to 1 cm above the iliac crest); the software then estimates visceral fat mass by subtracting subcutaneous abdominal fat from the total abdominal fat mass measured.²⁹ Visceral fat mass from the GE Healthcare software was measured in the android region, defined as the distance between the top of the iliac crest to 20% from the base of the skull.³⁰

Statistical analyses

The primary body composition outcome for this secondary analysis was the percent change in body fat percentage from baseline. Secondary outcomes include percent change in total lean mass, total fat mass, appendicular lean mass, total mass, body mass index (BMI), and upper and lower extremity lean mass. Visceral fat was included as an exploratory outcome. Data were combined across sites for the analysis, with the exception of visceral fat mass due to proprietary differences in machine algorithms.³¹ Linear regression modeling was conducted with and without adjustment for randomization stratification variables (age, birth sex, and site). Effect sizes for unadjusted models were calculated using Cohen’s F². Analysis of visceral fat mass was stratified by site due to differences in the proprietary machine algorithms,³¹ reported as absolute change in kilograms (kg), and adjusted for age, sex at birth, and baseline value to account for baseline heterogeneity. For the primary and secondary outcomes, stratified analyses were completed to examine overall change (pooled exercise intensities) among subgroup populations that were determined a priori (ie, age categories, sex, BMI, prior exposure to thymidine analogues [zidovudine, didanosine, or stavudine], and statin use). Participants who reported being “unsure” of their prior exposure to thymidine analogues were excluded from that subgroup analysis. The decision to pool exercise intensities was determined a priori if no significant differences between HIIT and CME were found in the primary analysis. Linear mixed effects models were used to further assess for a potential moderation effect of exercise intensity on subgroup change using a 3-way interaction (study week × subgroup × exercise intensity). Patterns of body composition response heterogeneity were classified into 4 groups: gained fat and lean mass, gained fat and lost lean mass, lost fat and lean mass, and lost fat and gained lean mass. Baseline demographic characteristics were summarized using descriptive statistics across these groups. Subgroup analyses were considered exploratory. Two-sided tests are reported assuming statistical significance at p < .05. Study data were collected and managed using REDCap electronic data capture tools hosted at the University of Colorado and supported by Vanderbilt University, Nashville, Tennessee, USA.³² All analyses were performed using RStudio (v2024.09.1 + 394; Boston, Massachusetts).

Results

Among 118 enrolled PWH, 95 completed pre-and post-training DXA scans and are included in the present analysis (Figure 1). The median (interquartile range [IQR]) age was 58 (54, 61) years. Approximately 14% were female, 14% Black or African American, and 81% were classified as overweight or obese (BMI ≥25 kg/m²). Further demographics are detailed in Table 1.

Figure 1.

CONSORT diagram of study flow.

Table 1.

Baseline demographic characteristics of participants.

	Overall	High-intensity interval training (HIIT)	Continuous moderate-intensity exercise (CME)
n	95	49	46
Age (median [IQR])	58.0 [54.0, 61.0]	57.0 [54.0, 61.0]	58.0 [54.0, 62.0]
Birth sex = female (%)	14 (14.7)	8 (16.3)	6 (13.0)
Race/ethnicity (%)
African American/Black, non-Hispanic	14 (14.7)	8 (16.3)	6 (13.0)
Hispanic or Latino/a, regardless of race	14 (14.7)	4 (8.2)	10 (21.7)
Other or unreported	5 (5.3)	3 (6.1)	2 (4.3)
White/Anglo, non-Hispanic	62 (65.3)	34 (69.4)	28 (60.9)
Education (%)
11th grade or less	3 (3.2)	1 (2.0)	2 (4.3)
High school or GED	13 (13.7)	2 (4.1)	11 (23.9)
Some college or technical school	36 (37.9)	24 (49.0)	12 (26.1)
College (BA or BS)	25 (26.3)	13 (26.5)	12 (26.1)
Master’s degree or higher	18 (18.9)	9 (18.4)	9 (19.6)
Employment (%)
Disabled (permanently or temporarily)	20 (21.1)	10 (20.4)	10 (21.7)
Retired	22 (23.2)	11 (22.4)	11 (23.9)
Working (full or part-time)	42 (44.2)	22 (44.9)	20 (43.5)
Missing/prefer not to answer	4 (4.2)	2 (4.1)	2 (4.3)
Other	7 (7.4)	4 (8.2)	3 (6.5)
Smoking status (%)
Current	8 (8.4)	4 (8.2)	4 (8.7)
Former	42 (44.2)	22 (44.9)	20 (43.5)
Never	45 (47.4)	23 (46.9)	22 (47.8)
Marijuana use in the last year (%)	47 (49.5)	25 (51.0)	22 (47.8)
Illicit drug use in the last 2 years (%)	15 (15.8)	7 (14.3)	8 (17.4)
Categorical BMI (%)
18.5 to <25 kg/m2	18 (18.9)	9 (18.4)	9 (19.6)
25.0 to <30.0 kg/m2	36 (37.9)	20 (40.8)	16 (34.8)
≥30.0 kg/m2	41 (43.2)	20 (40.8)	21 (45.7)
BMI (mean (SD))	30.0 (6.1)	30.0 (6.4)	29.9 (5.8)
Hypertension (%)	50 (52.6)	26 (53.1)	24 (52.2)
Depression, anxiety, or bipolar disorder (%)	49 (51.6)	26 (53.1)	23 (50.0)
Diabetes (%)	15 (15.8)	9 (18.4)	6 (13.0)
Years since HIV diagnosis (median [IQR])	23.0 [16.5, 30.0]	23.0 [17.0, 31.0]	23.0 [16.0, 29.5]
Years of ART (median [IQR])	20.0 [13.0, 26.0]	20.0 [13.0, 26.0]	20.0 [11.2, 26.0]
Prior thymidine analogue exposurea (%)
Exposure to 1 or more	31 (32.6)	17 (34.7)	14 (30.4)
No exposure	44 (46.3)	23 (46.9)	21 (45.7)
Unsure	20 (21.1)	9 (18.4)	11 (23.9)
Baseline CD4 (median [IQR])	679 [482, 835]	720 [477, 854]	646 [496, 82]
current statin use (%)	44 (46.3)	22 (44.9)	22 (47.8)
current testosterone use (%)	7 (7.4)	3 (6.1)	4 (8.7)
Current GLP-1RAb	3 (3.2%)	2 (2.1%)	1 (1.1%)

Abbreviations: ART, antiretroviral therapy; BA, Bachelor of Arts; BMI, body mass index (kg/m²); BS, Bachelor of Science; GED, general education degree; IQR, interquartile range.

^aThymidine analogue exposure included any use of zidovudine, didanosine, or stavudine.

^bGlucagon-like peptide-1 receptor agonist, participants were on a stable dose for ≥3 months for diabetes.

Change in body composition parameters and total mass

Although we observed statistically significant changes from baseline to week 16 for body fat percentage, fat mass, lean mass, and ALM in both arms, there were no statistical differences between the HIIT and CME arms in adjusted analyses (Figure 2). Body fat percentage decreased by an average of 3.3% in both HIIT and CME (95% confidence interval [CI]: −5.0, −1.6 in HIIT; −5.1, −1.6 in CME). Total fat mass decreased by −3.5% (95% CI: −5.8, −1.2) in the HIIT arm and −3.7% (95% CI: −6.1, −1.3) in the CME arm. Both intervention arms had significant increases from baseline in total lean mass (1.7% [95% CI: 0.7, 2.7] in HIIT; 1.1% [95% CI: 0.1, 2.2] in CME) and ALM (3.4% [95% CI: 1.5, 5.3] in HIIT; 2.2% [95% CI: 0.2, 4.1] in CME). There was no significant change in BMI or total body mass within arms (−0.27% [95% CI: −1.3, 0.8] in HIIT; −0.45% [95% CI: −1.5, 0.6] in CME) or between arms (Figure 2). Changes in upper and lower extremity lean mass are shown in the supplementary material and were not statistically different between HIIT and CME (Table S1). All results were similar in unadjusted analyses and in analyses using absolute change (Tables S2 and S3).

Figure 2.

Model estimates of percent change from baseline of body composition parameters. Adjusted for age, sex at birth, and study site. * denotes statistically significant within-group change (p < .05). HIIT, high-intensity interval training; CME, continuous moderate-intensity exercise; ALM, appendicular lean mass.

Change in visceral fat mass

At the University of Colorado site, there were no significant changes in visceral fat mass within or between arms (−0.06 kg, 95% CI: −0.11, −0.01 in HIIT vs. −0.03 kg, 95% CI: −0.08, 0.02 in CME). At the University of Washington, changes in the HIIT group were minimal (−0.01 kg, 95% CI: −0.15, 0.12) while those in CME had a significant decrease in visceral fat by −0.18 kg (95% CI: −0.31, −0.04).

Change in body composition by subgroup population

Changes in any body composition parameter by age categories, sex at birth, or BMI category were not significantly different across groups (Table 2). However, compared to participants not on statins, those on statins had significantly greater decreases in body fat % (−4.9% [95% CI: −3.5, −0.3] vs. 1.9% [95% CI: −6.7, −3.2]) and total fat mass (−6.0% [95% CI: −8.3, −3.7] vs. −1.5% [95% CI: −3.6, 0.7]) (both p < .05). Changes in total lean mass, ALM, or total body mass were not significantly different between statin subgroups. Similarly, compared to participants with no prior exposure to thymidine analogues, those with prior exposure had a significantly greater percent decrease in body fat % (−4.9% [95% CI: −7.0, −2.9] vs. −1.4% [95% CI: −3.1, 0.3]) and total fat mass (−5.3% [95% CI: −8.1, −2.5] vs. −1.3% [95% CI: −3.7, 1.0]). Changes in total lean mass, ALM, or total body mass were not significantly different between thymidine exposure subgroups. Baseline demographic characteristics and anthropometrics by statin subgroups and thymidine exposure subgroups are shown in Tables S4 and S5. There was no differential effect of exercise intensity on the changes seen among any subgroup population (Table S6).

Table 2.

Subgroup analyses of the changes in body composition parameters.

	n	Body fat %	Total fat mass	Total lean mass	Appendicular lean mass	Total mass
Age
<55 years	31	−2.63 (−4.76, −0.50)	−2.84 (−5.72, 0.04)	1.09 (−0.15, 2.32)	1.99 (−0.39, 4.38)	−0.30 (−1.62, 1.01)
56-60 years	33	−2.99 (−5.05, −0.93)	−3.00 (−5.79, −0.21)	1.73 (0.53, 2.93)	4.18 (1.87, 6.50)	−0.14 (−1.41, 1.14)
>60 years	31	−4.35 (−6.48, −2.23)	−4.91 (−7.79, −2.03)	1.43 (0.20, 2.93)	2.16 (−0.25, 4.54)	−0.65 (−1.96, 0.66)
Sex at birth
Male	81	−3.61 (−4.92, −2.30)	−3.82 (−5.60, −2.04)	1.47 (0.71, 2.24)	2.76 (1.28, 4.25)	−0.32 (−1.13, 0.49)
Female	14	−1.60 (−4.75, 1.55)	−2.14 (−6.42, 2.14)	1.13 (−0.71, 2.96)	3.08 (−0.49, 6.65)	−0.59 (−2.53, 1.36)
BMI category
Normal	18	−3.99 (−6.80, −1.18)	−3.85 (−7.65, −0.05)	1.41 (−0.21, 3.03)	2.70 (−0.47, 5.86)	0.02 (−1.69, 1.73)
Overweight	36	−3.16 (−5.14, −1.17)	−2.94 (−5.63, −0.26)	1.82 (0.67, 2.96)	2.94 (0.70, 5.18)	0.15 (−1.06, 1.36)
Obese	41	−3.17 (−5.03, −1.30)	−4.00 (−6.52, −1.49)	1.08 (0.00, 2.15)	2.74 (0.64, 4.84)	−0.97 (−2.11, 0.16)
Medication exposure
Prior thymidine  analogue	31	−4.91 (−6.99, −2.82)a	−5.31 (−8.19, −2.42)a	2.00 (0.68, 3.33)	3.79 (1.46, 6.12)	−0.53 (−1.92, 0.87)
No prior thymidine  analogue	44	−1.43 (−3.18, 0.32)a	−1.34 (−3.76, −1.09)a	0.92 (−0.20, 2.03)	1.94 (−0.01, 3.90)	−0.02 (−1.15, 1.19)
Current statin use	44	−4.94 (−6.67, −3.21)a	−6.00 (−8.32, −3.68)a	1.47 (0.43, 2.50)	2.98 (0.96, 4.99)	−1.17 (−2.24, −0.09)a
No statin use	51	−1.92 (−3.53, −0.31)a	−1.47 (−3.63, 0.68)a	1.38 (0.42, 2.35)	2.66 (0.79, 4.53)	0.34 (−0.66, 1.33)a

Thymidine analogue exposure included any use of zidovudine, didanosine, or stavudine.

Abbreviations: ALM, appendicular lean mass; BMI, body mass index.

^aDenotes statistical significance (p < .05) between subgroups.

Heterogeneity of body composition response

Over the course of the intervention, there were distinct patterns of body composition responses. The majority of participants (n = 40) lost fat and gained lean mass (−2.3 [SD: 1.6] kg fat; +1.5 [SD: 1.3] kg lean), 20 lost both fat and lean mass (−2.8 [SD: 1.8] kg fat; −1.3 [SD: 0.8] kg lean), 25 gained both fat and lean mass (+1.3 [SD: 0.9] kg fat; +2.0 [SD: 1.5] kg lean), and 10 gained fat but lost lean mass (+0.91 [SD: 0.6] kg fat; −1.2 [SD: 0.9] kg lean) (Figure 3). Participants who lost both fat and lean mass had the observed highest percentage of participants with hypertension (85%) and participants using statins (70%). Baseline demographics and body composition characteristics of each group can be found in Table S7.

Figure 3.

Heterogeneity of body composition response stratified by exercise intensity. Quadrant I represents those who lost lean mass and gained fat mass; II represents those who gained lean and fat mass; III represents those who lost lean and fat mass; IV represents those who gained lean and lost fat mass. The gray box represents the coefficient of variation of DXA measurements (0.7% for lean mass and 1.7% for fat mass); data points within the box represent changes that may reflect measurement error.

Discussion

In a cohort of nearly 100 older PWH completing 16 weeks of HIIT versus CME, combined with resistance training, we found significant overall improvements in all body composition parameters, with a gain in lean mass in the setting of fat mass loss, regardless of randomized intervention arm. We did not find that the magnitude of changes differed significantly by intervention arm. In stratified analyses, we found that PWH using statins and PWH with prior thymidine analogue exposure both had significantly greater decreases in total fat mass and body fat percentage (compared to no statin use or no thymidine analogue exposure) in response to training. Finally, we characterized the heterogeneity of body composition response by grouping participants according to whether they lost or gained fat and lean mass.

To our knowledge, this is the first study comparing the effects of HIIT versus CME with resistance exercise on body composition changes in older PWH. A prior systematic review found that aerobic exercise reduced fat mass and progressive resistance exercise increased lean or muscle mass among adults (range from 18 to 70 years) PWH.³³ Of note, the studies included in the review were exclusively resistance training or aerobic training, and no combined exercise studies were included.³³ In our prior study of older PWH (≥50 years) undergoing 24 weeks of combined higher intensity continuous aerobic and resistance exercise, we found a significant reduction in fat mass¹⁸ and increased lean mass of similar magnitude to that of the present study (+0.6 kg vs. +0.6-0.9 kg, respectively).

Our lack of significant differences between HIIT versus CME in PWH is consistent with studies comparing the effects of HIIT and CME on body composition in adults without HIV.^34,35 A distinction, however, is that other studies did not include resistance training. A meta-analysis of studies restricted to overweight or obese participants found that fat mass decreased by approximately 6% from baseline in both HIIT and CME,³⁴ and lean mass did not increase significantly (+0.06-0.07 kg), with no significant between-arm differences.³⁴ In another meta-analysis of studies in older adults, changes in body composition tended to favor HIIT over CME, but the estimates similarly did not reach statistical significance.³⁵ Authors reported that fat mass was significantly reduced only within the HIIT condition, and lean mass did not change significantly in either HIIT or CME.³⁵

Although no studies to date have established standardized, clinically relevant thresholds for changes in fat and lean mass, reductions in body fat percentage of at least ∼3% have been associated with greater improvements in obesity-related health outcomes.³⁶ Furthermore, the lean mass increases and fat mass decreases seen in this study exceeded the DXA coefficients of variation for these measurements (0.7% and 1.7%, respectively), suggesting that the changes reflect true physiologic adaptation rather than measurement variability.

We undertook exploratory subgroup analyses that included demographic and clinical characteristics of the participants. An interesting finding was that PWH using statins lost significantly more fat mass than statin nonusers, whereas lean mass increased regardless of statin use. In contrast, Mikus et al. found that sedentary, overweight or obese adults without HIV on statin therapy lost less fat mass after 12 weeks of aerobic exercise (45 minutes of treadmill walking or jogging at 60%-75% HRR), compared to participants who were not on statin therapy.³⁷ However, statin therapy began concurrently with exercise training in the Mikus study, whereas participants in HEALTH were on stable statin therapy prior to study enrollment. Our findings could be explained in that statin users had greater total fat mass at baseline compared to nonusers. Encouragingly, participants on statin therapy significantly improved lean mass in the study, despite mixed literature describing the effects of short-term statin use on skeletal muscle size or strength adaptations in response to resistance training in prospective randomized trials.^37–40 To our knowledge, no other studies to date have assessed body composition changes following exercise among those on long-term statin therapy.

Another notable finding from our subgroup analyses was that those with prior exposure to thymidine analogues had significantly greater reductions in fat mass and body fat percentage following exercise compared to those without prior exposure. Thymidine analogue-based ART is known to induce lipoatrophy, or the excessive loss of subcutaneous fat commonly seen in the face and upper and lower extremities.⁴¹ Although these regimens have been replaced by newer ART regimens, several longitudinal cohort studies have demonstrated that alterations to adipose tissue quality, density, and distribution have persisted more than 10 years after treatment discontinuation.^42–44 The mechanisms by which prior thymidine analogue exposure may modify adipose tissue responses to exercise have yet to be elucidated. To our knowledge, no other studies to date have examined body composition response after exercise stratified by prior exposure to thymidine analogues.

Encouragingly, more participants experienced favorable body composition changes (fat loss with lean mass gain) than any other response pattern in the setting of exercise. Identifying interventions that facilitate fat loss while maintaining or increasing lean mass is particularly important with the rapid emergence of GLP-1RA use, which induces reductions in fat mass that are often accompanied by losses in lean mass.¹³ However, similar to concerns from GLP-1RA use, a subset of participants in the present study also lost both fat and lean mass. These participants had the highest baseline BMI and a greater prevalence of those with obesity-related comorbidities, such as hypertension and statin use. Although weight loss may have been clinically indicated in this group, preserving lean mass remains a challenge. Further studies should evaluate adjunctive strategies, such as protein intake, to support lean mass preservation during weight loss.

A few limitations are worth noting. First, all subgroup analyses were considered exploratory and should be interpreted with caution. The current statin use and prior thymidine analogue exposure subgroup analyses were limited by the non-randomized assignment of these therapies. Second, diet composition and caloric intake can significantly influence body composition, but we were not able to control for dietary changes. As participants did not receive nutritional counseling during the study, it is reasonable to assume that any dietary changes were balanced by randomization to the exercise arm. The DXA instruments at each study site differed by manufacturer, and thus, absolute body composition measures could not be compared between sites.⁴⁵ However, all participants had their pre- and post-training measurements on the same DXA instrument, and we reported the changes from baseline, which are comparable. Given that the sample size for HEALTH was driven by the primary outcomes of physical function and fatigue, the study may have been underpowered to detect statistically significant between-group differences in body composition; however, the magnitude of the changes was consistent across exercise arms. Additionally, female participants were under-represented in the study, which limits the generalizability to a sex-specific interpretation of the findings. Finally, given the nature of the superiority trial design without the presence of a non-exercise control group, we cannot completely attribute the changes in body composition to the intervention alone. However, the assumed trajectory of body composition without an intervention would be declining lean mass with increasing fat mass based on prior literature among PWH.¹²

Conclusion

Sixteen weeks of HIIT or CME, combined with resistance training, is an effective strategy for reducing fat mass while preserving or improving lean mass in previously sedentary older adults with HIV. As both intensities were effective for improving body composition, PWH should be encouraged to perform the type of aerobic exercise according to individual preference. Furthermore, with the increasing use of incretin-based weight loss medications, the role of aerobic and resistance training in complementing fat loss while minimizing lean mass loss, particularly in older adults, is of increasing relevance. Given the growing prevalence and subsequent consequences of obesity among PWH, optimizing exercise interventions to improve body composition with or without concomitant pharmacologic therapy will be paramount.

Supplementary material

Supplementary material is available at The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences online.

Funding

This work was supported by the National Institute of Aging (NIA) of the National Institutes of Health (NIH) through R01AG066562 and the Prevention Center Shared Resource, RRID: SCR026631, of the Fred Hutch/University of Washington/Seattle Children’s Cancer Consortium (P30 CA015704). This work is also supported by NIH/NCATS Colorado CTSA Grant Number UM1 TR004399, NIA K24AG082527 (to K.M.E.), NIDDK P30 DK048520, and the National Institute of Allergy and Infectious Diseases T32 AI150547 (to G.L.K.). This manuscript is the result of funding in whole or in part by the NIH and is subject to the NIH Public Access Policy. Through acceptance of this federal funding, NIH has been given the right to make this manuscript publicly available in PubMed Central upon the Official Date of Publication, as defined by NIH.

Conflicts of interest

K.M.E. has received grant funding from Gilead Sciences and has served as a consultant for ViiV, Merck, and Gilead (all paid to the University of Colorado). S.M. serves as a consultant to ViiV. The other authors declare no conflicts of interest.

Data availability

De-identified participant data with a data dictionary will be available upon reasonable request to the corresponding author (K.M.E.), upon approval by the institutional review board, and with additional institutional agreement. Data sharing will be dependent on the approval of a proposal with a complete data analysis plan.

Author contributions

All authors had full access to the data and were responsible for the decision to submit the manuscript for publication. Grace L. Kulik wrote the original draft of the manuscript and performed all statistical analyses. Kristine M. Erlandson and Allison R. Webel conceived the initial study idea. Kristine M. Erlandson, Allison R. Webel, and Samantha MaWhinney was responsible for the study design, randomization, and data analysis plans. Debashis Ghosh and Melissa P. Wilson verified the data and supervised the statistical analyses. Vincent Khuu administered the intervention at UCD, and Vitor H.F. Oliveira oversaw the administration of the intervention at UW. Stephanie Dillon, Paul Cook, and Catherine M. Jankowski were involved with study oversight and administration. All authors contributed to the review and editing.