The Strength They Measure, the Mass They Miss

The gap between what the public discourse on exercise assumes and what the trials on exercise actually test.

A social injunction with no matching evidence base

The public discourse on exercise and muscular training occupies a considerable place in health policy, preventive medicine, and institutional recommendations. It is repeated, correctly, that resistance training prevents sarcopenia, protects joints, and improves quality of life. What is nearly absurd is that this discursive omnipresence coexists with a scarcity of scientific work that has ever asked whether an obese individual with a muscle deficit can actually recover a normal muscle mass for their age and sex, rather than simply progressing somewhat above their own baseline. The gap between the volume of the recommendation and the paucity of the evidence for the recommendation is itself an object of legitimate astonishment. A treatment is prescribed with confidence to correct a condition when the question of whether it corrects the condition, as opposed to easing it a little, has almost never been posed in those terms.

What the literature measures, and what it declines to measure

The overwhelming majority of randomized controlled trials and meta-analyses on resistance training in overweight and obese populations report continuous variables, group means, and effect sizes expressed in kilograms or in percentages relative to baseline. The largest meta-analysis available on the question covers 116 articles describing 114 trials, with 4,184 participants who were overweight or obese, and reports a mean lean-mass gain of 0.7 kg across all training modalities [1]. Measured against a clinical deficit, that number is small.

More subtle is the reading that must be given to the combination of resistance training with caloric restriction, which is the protocol most commonly recommended to persons living with obesity. There the same meta-analysis reports not a gain but a preservation of lean mass (ES ≈ −0.3 kg, p = 0.550 to 0.727) [1]. The authors present the result as a protective effect: without training, caloric restriction erodes muscle; with training, muscle is maintained. For a subject who is already in deficit and who is seeking to normalize their muscle mass, the finding is not a silent failure. It is a correct answer to a question that is not theirs. The strategy most widely prescribed protects against loss. It does not restore.

This methodological architecture, centered on mean change rather than on the crossing of a clinical threshold, is not a neutral technical choice. Its structural consequence is to render invisible the question the present piece wishes to name: how many of the persons who follow a resistance training program in a context of obesity actually reach a normative muscle mass for their age and sex, and how many remain, in spite of all their documented effort, below the healthy threshold. A study can report a statistically significant result that is clinically insignificant at the individual level, and no one will ask about actual normalization. The omission is not anecdotal. It runs through nearly the whole of the scientific corpus consulted on the subject.

One exception, and a second nearby, in a field of hundreds

Across the randomized controlled trials and meta-analyses recensed on sarcopenic obesity and on muscle deficit in the overweight and obese, one recently published trial has explicitly posed the question under the angle of a binary diagnostic threshold, that is, has asked whether participants actually left the pathological category rather than merely approaching it. The trial by Polo-Ferrero and colleagues, published in Nutrients in 2025, is a 32-week randomized clinical trial in 40 community-dwelling older women meeting the EWGSOP2 and ESPEN criteria for sarcopenic obesity. It reports that 35.7% of the power-training group no longer met those diagnostic criteria at the end of the intervention [2]. The sample is modest, the duration is short of a year, and the endpoint is exactly the endpoint that matters.

The same absence extends beyond sarcopenic obesity and into sarcopenia without obesity. A single further trial has posed the binary question in that adjacent field. Flor-Rufino and colleagues, in Maturitas in 2023, followed 38 older sarcopenic women (20 in the high-intensity resistance training arm) for six months and combined MRI of the thigh with bioelectrical impedance to measure the change. They report a remission of sarcopenia in 50% of the training group [3]. That the same binary question, in the sarcopenia-only literature, has apparently been asked only once suggests that the omission is not a peculiarity of the sarcopenic obesity subfield. It is a shared feature of the wider muscle-mass literature.

These two trials are not isolated curiosities without significance. Their existence demonstrates that it is entirely possible to design a study that asks the actual question, the question of full normalization rather than of relative gain, and that when the question is asked it produces substantial and clinically meaningful results. Their extreme rarity in an ocean of literature devoted to the subject does not owe to methodological impossibility. It owes to a collective choice within the research community not to pose the question, or to pose it only very recently, and only marginally.

The omission as a form of institutional aberration

The situation must be named for what it is: a systemic failure of research to answer the question that matters most to the person concerned, the person who lives with a measurable muscle deficit and who is following an exercise program in the legitimate hope of correcting it entirely. The greater part of the available literature is content to measure whether exercise helps a little, never whether it corrects completely, even as the public, medical, and media discourse on exercise is framed in near-categorical terms, as if the expected result were already secured. A 2025 meta-analysis in older women with sarcopenic obesity reports significant gains in physical function, but no significant change in the skeletal muscle mass index (WMD −0.62, 95% CI −2.38 to 1.15, p = 0.49) [4]. Another 2026 meta-analysis, covering 518 older sarcopenic women across twelve controlled trials, reaches the same structural conclusion: improvement in grip strength and gait speed, absence of significant effect on muscle mass itself [5]. The 2023 umbrella review, which synthesizes the meta-analyses of randomized controlled trials then available on sarcopenic obesity, explicitly concludes that the certainty of the evidence remains low to very low, for lack of a sufficient number of primary studies that have even attempted to answer the question with rigor [6].

The pattern reproduces itself one level up. The umbrella review published in Frontiers in Nutrition in 2026 by Ma and colleagues synthesizes eight systematic reviews covering 33 unique primary trials on exercise-based interventions for sarcopenic obesity [7]. It reports the by-now-familiar signature: a modest and uneven muscle-mass response (mean difference of +0.40 kg in one review, directionally positive in some others but not uniformly replicated), a reduction in body-fat percentage in the range of 0.85 to 2.96 points, an improvement in grip strength between +1.30 and +3.85 kg, and an improvement in gait speed between +0.05 and +0.20 m/s. AMSTAR-2 confidence was low in seven of the eight included reviews and critically low in the eighth. Umbrella-level certainty was judged low or very low across every outcome. The authors themselves observe, without diagnosing it as an omission, that “strength and mobility improved more consistently than muscle-mass indices”. Across the 33 primary trials thus covered, none is cited as measuring the binary remission endpoint, and the umbrella review therefore neither extracts nor synthesizes remission proportions. The blind spot is not confined to the individual trials. It is stabilized at the level of the syntheses themselves.

Strength as a substitute target

Here is the point that turns the observation into a diagnosis. In the metabolic context of obesity, the improvement in strength that these trials all report is not nothing. It indexes real adaptations, neural and coordinative, and it protects against falls in the older subject. But it is the wrong criterion for the question the obese subject is posing. What contrasts a phenotype at metabolic risk from one that is protected is muscle mass, and not the strength that muscle can produce in a test. Resting energy expenditure depends on the volume of tissue available. Insulin-mediated glucose uptake depends on the volume of tissue that can store glucose. Long-term defense against weight regain, after a caloric-restriction phase, depends on the metabolic floor that muscle mass provides. In this context the improvement of strength functions as a substitute target. It gives the appearance of success to a protocol that has not reached the target that matters. The literature does not measure the wrong thing. It measures a real thing that does not answer the question posed.

The consequence, in the trial reporting that follows, is a peculiar kind of positive result. A protocol produces a statistically significant improvement in grip strength and a statistically significant improvement in gait speed. Its author reports the result and concludes that resistance training improves the state of the sarcopenic-obese subject. The reader who scans the abstract will understand that the intervention worked. The reader who lives with the deficit and reads the same abstract has been told, with the language of scientific success, that a protocol produced an outcome that leaves their metabolic risk intact. There is no falsification here, in the technical sense. There is a systematic mismatch between the endpoint the discipline reports and the endpoint the subject would recognize as an answer.

The weight of individual variability, not integrated into any design

This gap is compounded by a second phenomenon widely documented in the general hypertrophy literature but almost never crossed with the question of sarcopenic obesity: the inter-individual variability of the response to resistance training. Rates of non-response reported in the general hypertrophy literature range from 0% to 84% depending on the statistical definition used [8], a spread that reflects a real biological heterogeneity between individuals subjected to the same stimulus. An international conference dedicated to this question was held at the University of Jyväskylä in November 2025 [9], and its proceedings were summarized in the Journal of Applied Physiology in 2026 [10], evidence that the subject is only now beginning to receive the scientific attention it warrants. This reflection on individual variability, however, remains nearly absent from trials specific to sarcopenic obesity, where group means continue to be reported and continue to mask the diversity of individual trajectories. Two persons with obesity following exactly the same protocol may reach radically different outcomes, one arriving at full normalization and the other plateauing indefinitely below the healthy threshold, and no current study allows one to predict in advance who will belong to which group, or why.

An omission that is not neutral

The net result is that a person with obesity and a muscle deficit, following the most widely diffused exercise recommendations to the letter, is effectively advancing in a form of scientific darkness as to the real probability of reaching full normalization of their muscle mass. The rare spectacular transformations visible publicly, whether shared on social media or observed in personal acquaintances, are likely to be no more than the upper margin of a distribution of responses whose extent and determinants remain, to this day, largely uncharacterized by institutional research. This is not merely a lack of data. It is a lack of the question itself, posed with the rigor it warrants, in a domain where the public discourse acts as if it already had the answer.

The physiological right at issue, the right of the person concerned to have their muscle mass measured, named, and, where deficient, actually restored to a normative range, is not a technical claim about exercise physiology. It is the claim that the endpoint reported by the trial and the endpoint recognized by the subject should meet. As long as they do not, the recommendation carries the confidence of a medical injunction with none of the epistemic verification a medical injunction is normally supposed to require.

References

1. Lopez P, Taaffe DR, Galvão DA, Newton RU, et al. Resistance training effectiveness on body composition and body weight outcomes in individuals with overweight and obesity across the lifespan: A systematic review and meta-analysis. Obesity Reviews. 2022;23(5):e13428. PMC9285060

2. Polo-Ferrero L, Martin MJ, Puente-González AS, Barbero-Iglesias FJ, González-Manzano S, Méndez-Sánchez R. Efficacy of Power Training on Sarcopenic Obesity in Community-Dwelling Older Women: A 32-Week Randomized Clinical Trial. Nutrients. 2025. PubMed 40507091

3. Flor-Rufino C, Barrachina-Igual J, Pérez-Ros P, Pablos-Monzó A, Sanz-Requena R, Martínez-Arnau FM. Fat infiltration and muscle hydration improve after high-intensity resistance training in women with sarcopenia. A randomized clinical trial. Maturitas. 2023. PubMed 36423489

4. Guo C, Dai T, Zhang H, Luo M, Gao J, Feng X. Effect of resistance training on body composition and physical function in older females with sarcopenic obesity: a systematic review and meta-analysis of randomized controlled trials. Frontiers in Aging Neuroscience. 2025. PubMed 40370749

5. Zhou Y, Wen K, Zhang X, Sun Y. Effects of resistance training on muscle mass, strength, and physical function in older women with sarcopenia: a systematic review and meta-analysis. Frontiers in Public Health. 2026. PubMed 41668861

6. Reiter L, Bauer S, Traxler M, et al. Effects of Nutrition and Exercise Interventions on Persons with Sarcopenic Obesity: An Umbrella Review of Meta-Analyses of Randomised Controlled Trials. Current Obesity Reports. 2023;12(3). PubMed 37249818

7. Ma X, Sun G, Liu C, Yan X, Liang W, Ma Z. Exercise-based interventions for sarcopenic obesity in middle-aged and older adults: an umbrella review of systematic reviews with pairwise meta-analyses and network meta-analyses. Frontiers in Nutrition. 2026. doi:10.3389/fnut.2026.1859967

8. Hubal MJ, Gordish-Dressman H, Thompson PD, et al. Variability in muscle size and strength gain after unilateral resistance training. Medicine & Science in Sports & Exercise. 2005;37(6):964-972.

9. Inter-Individual Variation in Resistance Training Responses. 25th International Symposium, Biology of Physical Activity, University of Jyväskylä, Finland. November 19–21, 2025. Conference page

10. Research into resistance training response heterogeneity: a summary of the 2025 conference at the University of Jyväskylä. Journal of Applied Physiology. 2026. doi:10.1152/japplphysiol.00289.2026