Exercise on a GLP-1: Pilates, Walking, HIIT, and Resistance Training for Lean Mass Preservation

Exercise pairing on a GLP-1 is the area with the strongest mitigation evidence for the lean mass loss documented in STEP-1 and SURMOUNT-1. The S-LiTE trial^[1] — a 195-patient randomized comparison of liraglutide alone, exercise alone, and both combined — reported −9.5 kg in the combination arm vs −6.8 kg with liraglutide alone and −4.1 kg with exercise alone, a 40% bigger weight loss in the combination arm with roughly double the body-fat reduction. The single highest-leverage intervention is resistance training paired with adequate protein^[4][5]. Pilates and yoga produce minimal weight effects but are useful for stress and mobility^[6][10]. ACSM 2009 sets the cardio threshold at >250 min/week of moderate-intensity activity for clinically significant weight loss^[11], and the Saint-Maurice 2020 JAMA cohort^[8] identifies ~7,000-8,000 steps per day as the inflection point for mortality benefit. Here is the evidence map.

The lean mass problem GLP-1 trials surfaced

The STEP-1^[3] and SURMOUNT-1 body composition substudies showed that approximately 25-45% of total weight loss on semaglutide and tirzepatide is lean tissue when patients do not intervene with protein and resistance training. Neeland and colleagues^[3] reviewed the full GLP-1 lean mass literature in 2024 and concluded that the muscle changes appear adaptive (proportional to weight loss magnitude and improved insulin sensitivity), but that active mitigation is warranted in patients losing rapidly, in older adults, and in anyone planning sustained weight loss over 12-18 months. The recommended targets are 1.6-2.3 g/kg fat-free mass of protein per day plus structured resistance training at least three times per week.

For the protein side, see our GLP-1 protein & macro calculator, which targets these recommendations directly. This article covers the exercise side.

S-LiTE: the trial that proved exercise + GLP-1 is more than additive

Lundgren and colleagues^[1] randomized 195 adults with obesity, after an initial 8-week low-calorie diet that produced ~12% body weight loss, to one of four arms for one year:

• Exercise + placebo: −4.1 kg
• Liraglutide 3 mg + usual activity: −6.8 kg
• Liraglutide + exercise (combination): −9.5 kg
• Placebo + usual activity: reference

The combination arm did not just add the two effects; the body-fat percent reduction was approximately double either single arm (−3.9% in the combination vs −1.7% and −1.9% in the single arms), and the combination was the only arm that improved A1c, insulin sensitivity, AND cardiorespiratory fitness simultaneously. The structured exercise prescription was a 4-day-per-week program of moderate-to-vigorous aerobic activity totaling 150 minutes plus two strength sessions per week.

S-LiTE is the most direct trial answer to the question “does exercise still matter on a GLP-1?” The answer is yes, and the magnitude of the additional benefit is large enough that exercise should not be treated as optional in patients capable of doing it.

Resistance training: the canonical evidence base

Three studies anchor the resistance-training-during-deficit recommendation:

• Longland 2016^[4] randomized 40 young men in a 4-week intense caloric deficit (~40% below maintenance) plus resistance + HIIT 6 days/week to 1.2 g/kg/day vs 2.4 g/kg/day protein. The high-protein arm gained +1.2 ± 1.0 kg of lean body massand lost −4.8 ± 1.6 kg of fat. The standard-protein arm gained essentially no lean mass and lost less fat. High protein + resistance training preserved AND added muscle in a severe deficit.
• Mettler 2010^[5] randomized 20 resistance-trained athletes in a 2-week deficit to 1.0 g/kg/day vs 2.3 g/kg/day protein. Lean mass loss was −1.6 ± 0.3 kg (low protein) vs −0.3 ± 0.3 kg (high protein) — 1.3 kg of lean mass preserved by adequate protein in just two weeks.
• Cava 2017^[2] reviewed the published literature on muscle preservation during weight loss in Advances in Nutrition and concluded that resistance training plus adequate protein is the best-evidenced strategy and should be promoted in any weight-loss program where lean mass preservation matters.

The compound lifts that recruit the most muscle per session — squat, hinge (deadlift / Romanian deadlift), horizontal push (DB press / push-up), horizontal pull (row / pull-up), and carry (farmer carry / sled push) — are the highest-leverage exercises for the limited weekly training volume most patients can sustain.

Walking and the step count question

Walking is the most accessible exercise modality and has its own evidence base. Murphy and colleagues^[7] meta-analyzed 24 RCTs of walking interventions and reported statistically significant reductions in body weight, BMI, body fat percent, and resting diastolic blood pressure with walking programs of at least 150 minutes per week of brisk intensity. The magnitude is modest at the lower doses but compounds at higher doses.

Saint-Maurice and colleagues^[8] published the most-cited step count data in JAMA in 2020, using 4,840 US adults from NHANES 2003-2006 with objective accelerometer measurements followed through 2015. They reported:

• 8,000 steps/day: hazard ratio 0.49 vs 4,000 steps/day — a 51% reduction in all-cause mortality.
• 12,000 steps/day: HR 0.35 vs 4,000 — 65% reduction.
• Step intensity (faster vs slower walking) showed no independent benefit after adjusting for total daily steps. Total volume is what matters; speed is secondary.

The practical take-home: 8,000-10,000 steps per day is the ergonomic floor for general health benefit and weight loss support, achievable for most people via deliberate walks plus accumulated daily activity.

HIIT vs steady-state cardio

Wewege and colleagues^[9] meta-analyzed 13 RCTs directly comparing HIIT to moderate-intensity continuous training (MICT) for body composition outcomes in overweight and obese adults. The headline finding: no significant difference between HIIT and MICT for whole-body fat or waist circumference, but HIIT achieved equivalent results in roughly 40% less training time. For time-pressured patients, HIIT is a strong choice; for patients who prefer longer steady-state work or have joint limitations, MICT is equally effective. The Wewege analysis also flagged that running-based programs produced more fat loss than cycling-based programs across both training types.

Pilates and yoga: the honest position

Pilates and yoga are popular and have a small but real evidence base.

Aladro-Gonzalvo and colleagues^[6] systematically reviewed Pilates trials for body composition and reported only weak quantitative evidence of a positive effect. The underlying trials were methodologically limited (small samples, inconsistent measurement, inadequate dietary controls). Pilates produces minimal direct weight loss but improves core strength, postural control, and flexibility. For GLP-1 patients, Pilates is a reasonable adjunct for mobility maintenance, not a primary fat-loss modality.

Lauche and colleagues^[10] meta-analyzed 30 yoga trials covering 2,173 participants and reported that yoga does not produce significant weight, BMI, or body fat changes overall. Subgroup analysis showed modest waist-to- hip ratio improvement in healthy adults and a small BMI reduction in overweight/obese adults, but the effects were not robust against publication bias. The honest take: yoga is a stress-reduction and mind-body intervention with meaningful benefits for stress eating and quality of life, not a weight-loss intervention per se.

The ACSM frameworks

Two ACSM position stands set the practical floor:

Donnelly 2009^[11] — Appropriate Physical Activity Intervention Strategies for Weight Loss and Prevention of Weight Regain. Key thresholds:

• Weight gain prevention: 150-250 minutes/ week of moderate-intensity physical activity
• Modest weight loss: 150-250 minutes/week (yields modest results only)
• Clinically significant weight loss: >250 minutes/week
• Weight maintenance after loss: >250 minutes/week
• Resistance training was noted as “does not enhance weight loss but increases fat-free mass and fat loss” — the muscle preservation point.

Garber 2011^[12] — ACSM position stand on quantity and quality of exercise for cardiorespiratory, musculoskeletal, and neuromotor fitness. Key resistance training recommendations:

• Frequency: 2-3 days per week minimum
• Exercises: 8-10 exercises targeting major muscle groups
• Volume: 8-12 repetitions per exercise for strength and endurance
• Progression: progressive overload (gradual increase in resistance)

The deficit ceiling: why more is not better

Murphy and Koehler^[13] meta-analyzed studies of resistance training in caloric deficit and reported that lean mass gains were significantly impaired in deficits greater than 500 kcal/day below maintenance (effect size −0.57, p=0.02). Strength gains were preserved across the deficit range, but the muscle-building benefit of RT was attenuated in larger deficits.

For GLP-1 patients this finding is critically important. The drugs suppress appetite enough that many patients unintentionally drop into 800-1000 kcal/day eating ranges from a maintenance baseline of 2200-2500 kcal — a deficit far beyond the 500 kcal/day ceiling that the RT literature supports. The combination of intense training plus an aggressive GLP-1-induced deficit can produce more lean mass loss than either alone, exactly the opposite of the goal. Tracking calories and protein during the first 3-6 months of GLP-1 therapy is the simplest way to catch this.

The practical weekly template

Synthesizing the evidence into a starting weekly template for a typical GLP-1 patient (adjust to fitness and time):

• Resistance training: 2-3 days per week. 45-60 minutes per session. Compound lifts (squat, hinge, push, pull, carry) at 8-12 reps per set, 2-4 sets per exercise. Progressive overload week to week. Garber 2011 ACSM minimum^[12].
• Aerobic activity: total >250 minutes per week at moderate intensity, distributed across 5-6 days. Donnelly 2009 ACSM threshold for clinically significant weight loss^[11].
• Daily step target: 8,000-10,000 steps. Saint-Maurice 2020 JAMA inflection point^[8]. Walks count toward both the daily step target and the weekly aerobic minutes.
• HIIT (optional): 1-2 sessions per week for time-efficient cardio. 4-8 intervals of 30-60 seconds at near-maximum effort with equal recovery. Wewege 2017 shows equivalent fat loss to MICT in less time^[9].
• Mobility / flexibility / mind-body: 1-2 sessions per week (Pilates, yoga, foam rolling). Not for weight loss directly, but supports recovery and stress management.
• Recovery monitoring: track fatigue, subjective recovery, and training quality. If the RT sessions feel progressively harder week to week with no strength gains, the deficit is probably too steep (Murphy & Koehler 2022)^[13].
• Protein floor: 1.2-1.6 g/kg/day (or 1.6- 2.0 if on a GLP-1, per the Neeland 2024 review^[3]). See the protein calculator for personalized targets.

Common patterns to avoid

• Adding intense cardio without adding protein. The combination of GLP-1 appetite suppression + endurance training is the worst-case scenario for lean mass.
• Exclusive cardio with no resistance training. Cardio alone preserves less lean mass than resistance training alone. The lean mass goal requires structured loading.
• Tracking only minutes, not strength. Lifting the same dumbbells every week is not progressive overload. Logging your sets and reps is the cheapest way to verify your training is actually building (or preserving) capacity.
• Skipping resistance training because the gym is intimidating. Body weight progressions (push-up, squat, lunge, plank, band row) cover the major movement patterns and are a legitimate starting point. The progression matters more than the equipment.
• Ignoring fatigue signals. GLP-1 fatigue is real (see our fatigue article) and can mask exercise-induced overreaching. If sessions feel progressively harder without progress, drop volume for a week.

Bottom line

• The S-LiTE trial^[1] proves exercise + GLP-1 is more than additive: the combination arm produced 40% bigger weight loss and roughly double the body-fat reduction vs liraglutide alone.
• Resistance training plus adequate protein is the single-highest-leverage intervention for lean mass preservation on a GLP-1 (Longland 2016, Mettler 2010, Cava 2017, Neeland 2024).
• ACSM 2009 sets the cardio floor at >250 minutes/week for clinically significant weight loss. ACSM 2011 sets the resistance training floor at 2-3 days/week of compound lifts.
• Saint-Maurice 2020 JAMA identifies 8,000 steps/day as the inflection point for mortality benefit, with 12,000 steps adding additional protection.
• HIIT and MICT produce equivalent body composition outcomes; HIIT is time-efficient.
• Pilates and yoga are useful mobility and stress-reduction adjuncts, not primary weight-loss modalities.
• Murphy & Koehler 2022 warns that deficits >500 kcal/day impair lean mass gains in resistance-trained adults — directly relevant for GLP-1 patients whose appetite suppression can mask under-eating.

Related research and tools

• GLP-1 protein & macro calculator — the protein side of the muscle-preservation pair
• Loose skin after rapid GLP-1 weight loss — the downstream effect this article is designed to mitigate
• Semaglutide and muscle mass loss — the underlying lean tissue loss problem
• What to eat on a GLP-1 — food choices that meet the protein target
• GLP-1 fatigue and hair loss — relevant when titrating exercise volume
• Why am I not losing weight on a GLP-1? — the plateau picture exercise interacts with

Important disclaimer. This article is educational and does not constitute medical advice or an exercise prescription. Patients with cardiovascular disease, joint pathology, or other conditions limiting exertion should consult a clinician (and ideally a credentialed exercise physiologist or physical therapist) before starting any new exercise program. The S-LiTE trial used liraglutide 3 mg, not semaglutide or tirzepatide; the directional inference to the newer drugs is reasonable but not yet replicated in a head-to-head trial. Every primary source cited here was independently verified against PubMed on 2026-04-08 by a research subagent.