Effects of Semaglutide on Muscle Structure and Function in the SLIM LIVER Study

Abstract

Background

Semaglutide is highly effective for decreasing weight. Concomitant loss of muscle mass often accompanies weight loss and may have consequences on muscle function.

Methods

This is a secondary analysis from the SLIM LIVER (Advancing Clinical Therapeutics Globally for HIV/AIDS and Other Infections, ACTG A5371) study, a single-arm study of semaglutide in people with human immunodeficiency virus (HIV, PWH) with metabolic dysfunction–associated steatotic liver diseases (MASLD). Participants received subcutaneous semaglutide for 24 weeks (titrated to 1 mg/week by week 4). Psoas volume and fat fraction were assessed from liver magnetic resonance imaging, and physical function was assessed by 10-time chair rise test and 4 m gait speed. Mean change from baseline to week 24 was estimated with linear regression modeling.

Results

Fifty-one PWH were enrolled (muscle measures n = 46). The mean age was 50 years (standard deviation, 11), body mass index was 35.5 kg/m² (5.6), 43% were women, 33% Black, and 39% Hispanic/Latino. Psoas muscle volume decreased by 9.3% (95% confidence interval [CI]: −13.4 to −5.2; P < .001) over 24 weeks, but psoas muscle fat did not significantly change (−0.42%; 95% CI: −1.00 to .17; P = .16). Chair rise and gait speed showed nonsignificant improvements of 1.27 seconds (95% CI: −2.7 to .10) and 0.05 m/sec (95% CI: −.01 to .10), respectively (both P > .07). The prevalence of slow gait speed (<1 m/sec) decreased from 63% to 46% (P = .029).

Conclusions

In PWH receiving semaglutide for MASLD, despite decreased psoas muscle volume, there was no significant change in physical function, suggesting function was maintained despite significant loss of muscle.

Clinical Trials Registration

NCT04216589.

People with human immunodeficiency virus taking 1 mg semaglutide for 24 weeks had a decline in muscle volume but no significant change in physical function. This suggests that function may be maintained despite loss of muscle concomitant with weight loss.

Metabolic dysfunction–associated steatotic liver disease (MASLD) is a prevalent comorbidity in people with human immunodeficiency virus (HIV, PWH) [1]. Increased fat, particularly visceral fat, can lead to increased intrahepatic triglycerides (IHTG), which contribute to the pathogenesis of MASLD [2]. Mortality from MASLD is primarily from cardiovascular complications associated with increased IHTG, which is associated with impaired insulin signaling, oxidative stress and inflammation, and impaired endothelial function [3–5]. HIV-specific factors, such as chronic systemic inflammation, persistent immune activation, and mitochondrial dysfunction, are thought to further complicate MASLD in PWH [1, 6].

MASLD and increased IHTG have implications on the musculoskeletal system. Accumulation of lipids in the muscle, termed myosteatosis, was reported in 93% of a patient cohort with chronic liver disease [7]. Although the mechanism that links liver and muscle fat is not completely understood, it has been proposed that liver dysfunction impairs lipid oxidation in the muscle due to an increase in skeletal muscle ammonia [8]. The inability to properly use lipids may then result in inappropriate accumulation of fat in the muscle tissue and decrease the mechanical ability of the muscle to contract [8]. Interventions to reduce IHTG are crucial to prevent the development of MASLD and its associated cardiovascular and musculoskeletal complications [9].

Weight loss is recommended as the first line of treatment for MASLD [10]. Semaglutide, a glucagon-like peptide-1 receptor agonist (GLP-1 RA), is a highly effective medication for decreasing weight and weight complications by suppressing appetite and improving insulin signaling in overweight and diabetic populations [11, 12]. Additionally, semaglutide has been shown to decrease liver fat [13]. Initial data from semaglutide trials among populations without HIV suggest that weight loss with semaglutide is accompanied by losses in both fat and lean mass, likely representing loss in skeletal muscle quantity [14]. Data from pooled studies show that loss of lean mass may contribute up to 40% of total weight loss [15].

Whether this loss of lean (skeletal) mass has implications on physical function is not well understood, as interventions to decrease body weight, particularly fat mass, also have the potential to improve physical function [16]. Limited evidence supports that semaglutide may improve physical function; however, current findings are mostly limited to participant self-reported measures [17, 18]. Our goal in this predefined secondary analysis was to examine the change in muscle volume, muscle fat, and objectively measured physical function among PWH receiving semaglutide as treatment for MASLD.

METHODS

Study Design

Advancing Clinical Therapeutics Globally for HIV/AIDS and Other Infections (ACTG) A5371 SLIM LIVER study was a phase 2b, single-arm, open-label, 24-week pilot study on the effects of semaglutide on IHTG in PWH and MASLD [19]. The following ACTG sites enrolled participants in the study: Johns Hopkins University, Massachusetts General Hospital, University of Washington, Ohio State University, University of Cincinnati, University of Colorado Hospital, Instituto de Pesquisa Clinica Evandro Chagas in Brazil, Houston AIDS Research Team, and University of Alabama. The primary outcome was a 24-week change in IHTG; exploratory outcomes included 24-week change in muscle and 24- and 48-week change in physical function.

Participants

PWH were enrolled from February 2021 to September 2022. Inclusion criteria included age ≥18 years, on antiretroviral therapy (ART) with undetectable HIV-1 RNA levels, ≥ 5% IHTG content by proton-derived fat fraction magnetic resonance imaging (PDFF-MRI), and central adiposity (defined by a minimum waist circumference [WC] of ≥95 cm for individuals assigned male sex at birth or ≥94 cm for individuals assigned female sex at birth). Participants also had to meet at least 1 of the following criteria for insulin resistance or prediabetes: fasting plasma glucose 100–125 mg/dL, hemoglobin A1c (HbA1c) between 5.7% and 6.4%, or homeostatic model assessment for insulin resistance (HOMA-IR) >3.0 [20]. Each study site obtained local institutional review board approval. All participants provided written informed consent prior to enrollment in the study.

Intervention

Participants self-administered semaglutide subcutaneously into the abdomen, thigh, or upper arm, starting at 0.25 mg/week and titrating up to 1.0 mg/week at week 4 for 24 weeks of therapy. Participants were then followed off drug, on study for an additional 24 weeks. Participants were instructed to remain on their same ART regimen over the course of the study.

Covariates

All covariates were selected a priori. Participants self-reported age, sex at birth, current gender identity, race, and ethnicity. The age stratifications were prespecified and selected to designate younger adults, middle-aged adults, and older adults. WC was measured during study screening and used as a surrogate for visceral adiposity because visceral adipose tissue quantification was delayed. Adherence was based on participant self-report in the treatment log. Any indication of a missed dose was put into the “<100%” group.

Study Outcomes

Participants underwent liver PDFF-MRI at screening and week 24. Psoas muscle volume and fat fraction were also measured during the same imaging session [19]. The psoas muscle is a key contributor to hip flexion and trunk stabilization during functional activities such as walking, rising from a chair, and balance [21]. Physical function was measured at baseline, week 24, and week 48. Physical function was assessed by the time to rise from a chair 5 and 10 times and 4-meter gait speed, where gait speed was calculated as the average of 2 measurements at usual pace. Slow gait speed was defined as walking <1 m/sec [22]. Physical activity was self-reported through the International Physical Activity Questionnaire as baseline, week 24, and week 48 [23].

Statistical Analyses

Baseline characteristics and change outcomes that were continuous were summarized using mean and standard deviation; categorical variables were summarized using frequency and percentage. Change from baseline to week 24 was estimated with a linear regression model without an intercept term. Relative risk from baseline to week 24 was estimated using a log link generalized estimating equation model without an intercept term. Subgroup analyses were performed for sex at birth, gender, race/ethnicity, age, and semaglutide adherence. Associations between IHTG, body mass index (BMI), weight, WC, glucose, HbA1c, HOMA-IR, total cholesterol, low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, triglycerides, and psoas muscle volume and fat were assessed with Spearman correlations. Significance was assessed with a 5% type 1 error. Statistical analyses were performed using SAS software (Version 9.4 for Linux, SAS Institute Inc, Cary, NC).

RESULTS

Participants

The study enrolled 51 participants. Participant demographics reflect those in the per-protocol population (n = 49). The per-protocol population included all participants on study treatment until within 4 weeks of the week 24 MRI and who did not start prohibited medications known to affect weight prior to week 24. Paired psoas muscle outcomes from the MRI and physical function assessments were available for 47 participants at week 24, and physical function measures were available for 42 participants at week 48. The mean age was 50 years (standard deviation 11). At baseline, participants weighed a mean of 103.1 kg (20.8) with a BMI of 35.5 kg/m² (5.6) and WC of 114.7 cm (11.8). Overall mean weight loss in the study was 7.8 kg (95% confidence interval [CI]: 6.1 to 9.5) over 24 weeks [19]. Baseline differences in psoas muscle measures, weight characteristics, and age by sex at birth and gender identity can be found in the Supplementary Material (Supplementary Table 1). Demographics are further detailed in Table 1.

Table 1.

Demographic Information

Characteristic	Semaglutide (N = 49)a
Age, y	50 (11)
Sex at birth
Female	18 (37%)
Male	31 (63%)
Gender identity
Cis female	18 (37%)
Cis male	28 (57%)
Transgender female	3 (6%)
Race and ethnicity
American Indian or Alaska Native	1 (2%)
Black or African Americanb	16 (33%)
Non-Hispanic White	13 (27%)
Hispanic	19 (39%)
Multiple	1 (2%)
Unknown	2 (4%)
Weight characteristics
Body mass index, kg/m2	35.5 (5.6)
Weight, kg	103.1 (20.8)
Waist circumference, cm	114.7 (11.8)
HIV characteristics
Median (interquartile range) CD4 T-lymphocyte count × 109 cells/L	0.701 (0.586–0.869)
HIV-1 RNA <50 copies/mL	49 (100%)
Antiretroviral therapy regimen
Protease inhibitor	2 (4%)
Nonnucleoside reverse transcriptase inhibitor	10 (22%)
Integrase strand inhibitor	40 (82%)

Presented as n (%) unless otherwise reported.

Abbreviation: HIV, human immunodeficiency virus.

^an = 3 participants discontinued treatment early and were not included in the per-protocol population [19].

^bIncludes 1 participant who reported Hispanic ethnicity and Black race.

Psoas Muscle Volume

Absolute psoas muscle volume decreased by 1.49 mL (95% CI: −2.15 to −.83), with a mean percent decrease of 9.3% (95% CI: −13.4 to −5.2) over 24 weeks (P < .001 for both; Figure 1). Decreases in muscle volume were greatest among PWH aged >60 years (n = 7), who saw a mean percent change of −22.8% (95% CI: −32.4 to −13.3) compared with a 7.9% decrease (95% CI: −12.3 to −3.4] in those aged 40–60 years (n = 32) and a 2.4% decrease (95% CI: −11.9 to 7.2) in those aged <40 years (n = 7; P = .008). We did not find statistically significant differences between other subgroups (sex at birth, gender, race/ethnicity, or semaglutide adherence) in percent or absolute change in psoas volume (P > .05; Figure 1). Reductions in psoas muscle volume (%) correlated with absolute decreases in BMI (r = +0.31, P = .038) and HbA1c (r = +0.39, P = .007) and a percent decrease in IHTG (r = +0.32, P = .028). Change in absolute psoas volume was associated with an absolute change in fasting triglycerides and percent change in IHTG (r = +0.33, P = .027 and r = +0.30, P = .040). No statistically significant correlations were found between changes in psoas volume and changes in weight (r = +0.22), WC (r = +0.12), fasting glucose (r = −0.06), HOMA-IR (r = 0.0), fasting total (r = +0.22), LDL cholesterol (r = +0.10), or HDL cholesterol (r = +0.02; all P > .10).

Figure 1.

Overall change and change among subgroups for psoas muscle measures. A, Absolute change in psoas muscle measures. B, Percent change in psoas muscle measures. ** indicates P < .05.

Psoas Muscle Fat

There was a small, absolute mean decrease (−0.42%; 95% CI: −1.00 to .17) and mean percent change increase (5.18%; 95% CI: −10.3 to 20.7) in psoas muscle fat that did not reach statistical significance (P = .16 and P = .50, respectively; Figure 1). Percent change in psoas muscle fat differed by age (P = .027); muscle fat decreased by 3.6% (95% CI: −21.2 to 13.9) among those aged 40–60 years and decreased by 2.9% (95% CI: −39.8 to 34.0) in those aged >60 years but increased by 52% among those aged <40 years (95% CI: 15.4 to 89.2). These percent change increases were influenced by a few participants with low levels at baseline in the <40 age group; the difference between age groups was no longer significant when these outliers were removed. In other subgroup analyses for either absolute or percent change of psoas fat, differences were not seen (Figure 1). The relative (percent) increase in psoas fat was associated with an absolute reduction in fasting triglycerides (r = −0.30, P = .049) but not IHTG, BMI, weight, WC, fasting glucose, HbA1c, HOMA-IR, and fasting total, LDL, or HDL cholesterol (P > .05). No statistically significant correlations were found between absolute changes in psoas fat and covariates.

Physical Function

Both the time to rise from a chair and gait speed tended to improve with semaglutide, but most changes did not reach statistical significance (Table 2). Time to rise from the chair decreased (improved) by 0.66 seconds (95% CI: −1.4 to .07; P = .077) and 1.27 seconds (95% CI: −2.7 to .10; P = .069) for the 5 time and 10 time chair rises, respectively. Gait speed improved by 0.05 m/sec (95% CI: −.01 to .10; P = .078), and the prevalence of slow gait speed (<1 m/sec) decreased from 63% to 46% (P = .029). These measures were not correlated with change in weight (10× chair rise: r = +0.14; gait speed: r = −0.14; all P > .35) or psoas muscle fat (5 time chair rise r = +0.07; 10 time chair rise: r = +0.15; or gait speed r = +0.01, all P > .33). There were moderate correlations between the 10× chair rise and change in psoas volume (r = +0.35, P = .018); change in gait speed was weakly correlated with change in psoas volume but did not reach statistical significance (r = −0.27, P = .073).

Table 2.

Change in Measures of Physical Function

Parameter	Baseline	Week 24	Change, Baseline to Week 24	P Value
5 time chair rise, seconds	12.5 (3.6)	11.9 (3.3)	−0.66 (2.5) 95% CI, −1.4 to .07	.077
10 time chair rise, seconds	26.2 (7.0)	25.0 (6.8)	−1.27 (4.7) 95% CI, −2.7 to .10	.069
Gait speed, m/sec	0.93 (0.23)	0.98 (0.24)	0.05 (0.19) 95% CI, −.01 to .10	.078
Presence of slow gait, <1 m/sec	No: 18 (37%) Yes: 31 (63%)	No: 26 (54%) Yes: 22 (46%)	Risk ratio, 0.73 (0.55, 0.97)	.029

Results for performance measures presented as mean (standard deviation). The presence of slow gait speed is presented as frequency (n) and percentage (%).

Abbreviation: CI, confidence interval.

In subgroup analyses, the effect varied by gender for the 5 time chair rise (1.4 seconds faster [95% CI: .46 to 2.3] among 27 cisgender males vs 0.32 seconds slower [95% CI: −.75 to −1.4] among 20 cis or transgender females; P = .019) and the 10 time chair rise (2.98 seconds faster [95% CI: 1.3 to 4.6] in cis males vs 1.03 seconds slower [95% CI: −.9 to 3.0] in cis or transgender females; P = .003). The effect also varied by sex at birth for the 10 time chair rise time: 30 male-at-birth participants improved by 2.8 seconds (95% CI: 1.2 to 4.4), while 17 females at birth were 1.4 seconds slower (95% CI: −.7 to 3.5; P = .002). No other significant differences were observed in the subgroup analyses, Figure 2.

Figure 2.

Subgroup analyses of change in physical function. A, Change in 5 time and 10 time chair rise (seconds). B, Change in gait speed (m/sec). ** indicates P < .05.

There were additional small incremental improvements from week 24 to week 48 in 5 time chair rise (−0.13 second), 10 time chair rise (−0.68 second), and gait speed (0.02 m/sec) that did not reach statistical significance (all P > .32).

Physical Activity

Self-reported physical activity did not significantly change from baseline to week 24 or week 48 (Supplementary Table 3). Median (interquartile range) daily walk time was 60 (20, 120) minutes at baseline, 40 (20, 90) minutes at week 24, and 40 (20, 120) minutes at week 48.

DISCUSSION

In this secondary analysis from the ACTG A5371 SLIM LIVER study, we found that low-dose semaglutide in the setting of MASLD resulted in a 9.3% decrease in psoas muscle volume, similar to what is typically observed with weight loss. We did not see significant changes in physical function, suggesting that function was maintained despite significant loss of overall weight and muscle.

Psoas muscle measures were selected as an outcome in the present study because they did not require additional imaging to the liver MRI, reducing participant burden and study cost. Muscle volume by MRI additionally provides a more accurate estimate of skeletal muscle than lean mass by dual-energy X-ray absorptiometry (DXA), as DXA lean mass also reflects a change in any non-bone, non-fat tissue [24]. Although psoas muscle measures have not been previously measured as an outcome of GLP-1 RA, the reduction observed in the present study is consistent with changes seen in other weight loss–focused interventions among populations of persons that are overweight or obese [14, 25]. A 15-week intensive weight loss intervention in a population of persons with obesity and diabetes resulted in a 6.8% reduction in psoas muscle area [25]. Furthermore, longitudinal studies that assessed loss of psoas volume with unintentional weight loss also reported similar changes in volume, with a 9.2% reduction in psoas volume seen in patients with cancer cachexia [26, 27]. Only 1 other study has reported the impact of semaglutide on lean body mass among PWH with lipodystrophy; that study found a 5.4% loss over 32 weeks among those receiving 1 mg weekly semaglutide compared with placebo [28, 29].

Importantly, we found that physical function was maintained despite the nearly 10% loss in muscle volume. Furthermore, the confidence interval bounds for gait speed approach the minimally clinically important difference (0.05 m/sec) [30], which suggests that a statistically and clinically significant effect may have been seen with a larger sample size. Function was maintained with no significant changes to self-reported physical activity. Previous studies of changes in muscle mass and physical function with semaglutide treatment have demonstrated similar findings, though they have been primarily limited to self-reported changes in function. Patients with obesity randomized to semaglutide (at a higher subcutaneous 2.4-mg weekly dose) had a significant loss of DXA-measured total lean mass over 68 weeks but reported subjective improvements in physical function [12]. In a cohort of patients with type 2 diabetes, a significant decline in total lean mass was observed in conjunction with trends toward improvement of handgrip strength after 26 and 52 weeks of 0.5–1.0 mg weekly subcutaneous semaglutide [31, 32]. Among patients with obesity and heart failure with preserved ejection fraction, 52 weeks of subcutaneous semaglutide improved both self-reported physical function and the 6-minute walk test, a measure of endurance, by 21.5 meters [33].

Previous literature suggests that no substantial strength improvements are seen in the setting of weight loss, without the addition of exercise training [16, 34]. Although participants in the present study received general physical activity recommendations, the lack of significant improvement in functional leg strength by chair rise likely reflects a lack of specific muscle-strengthening exercises combined with semaglutide treatment. Exercise implemented during and/or following semaglutide treatment may improve long-term health outcomes, as exercise may help facilitate long-term, sustained weight loss for those who discontinue semaglutide [35].

Last, we did not find a reduction in muscle fat content, as has been seen in diet-induced weight loss interventions [36–38]. HIV-specific factors may explain some of the differences noted in the present study compared with previous literature. Prior studies have shown that PWH tend to have greater psoas muscle fat compared with controls without HIV [39]. No prior studies have examined the impact of semaglutide on muscle fat among PWH; whether reduction of muscle fatty infiltrate in response to interventions (semaglutide or other) differs between PWH and control populations is not known.

Several limitations of the present analysis should be acknowledged. These secondary analyses were not powered to detect differences in muscle volume, muscle density, or physical function; subgroups were small and exploratory, with limited power to detect differences. Muscle volume and muscle density measures of the psoas muscle may not reflect muscle changes that would be observed with lower extremity or full-body musculature. Study strengths should also be acknowledged, including the proportion of women and minorities, the clinical trial design and rigor of the end points (central training for the function measures, use of MRI for muscle volume), and the inclusion of measures of both gait speed (function) and ability to rise from a chair (providing an estimate of strength or power), which are more reflective of daily function than grip strength [40].

CONCLUSIONS

Among PWH receiving semaglutide for MASLD, we found a nearly 10% decrease in muscle volume but with no significant change in muscle fat or physical function. Although findings need to be compared to placebo and confirmed in larger studies, our findings suggest that physical function is preserved while taking semaglutide. Further studies on interventions to preserve muscle mass and improve function while taking semaglutide are warranted for both people with and without HIV.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Author Contributions. All authors had full access to the data and were responsible for the decision to submit the manuscript for publication. K. M. E., J. E. L., and F. R. S. conceived the initial study idea. D. W. K., A. K., and C. M. verified the data and performed all statistical analyses. R. M. performed the magnetic resonance imaging readings and assisted with interpretation of muscle fat and volume findings. G. L. D. and K. M. E. wrote the original draft of the manuscript. All authors contributed to review and editing.

Disclaimer. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health (NIH). The funders of the study oversaw the development and monitoring of the study but had no role in the conduct, analyses, and conclusions of the study. P. F. B.-Z., co-author of this study and medical officer from NIH/National Institute of Allergy and Infectious Diseases (NIAID), participated in the design, data interpretation, manuscript revision, and intellectual contribution; however, his views are personal and do not represent the NIH/NIAID's views.

Financial support. This work was supported by the National Institute of Allergy and Infectious Diseases (under UM1 AI068634, UM1 AI068636, UM1 AI106701) with additional funding provided by the McGovern School of Medicine at UTHealth. Preparation of the manuscript was supported by the NIAID (T32 AI150547 to G. L. D. and K24 AI120834 to T. T. B.) and the National Institute on Aging (under K24 AG082527 to K. M. E.).