REF-1168 · Varovic et al., 2025 · International Journal of Sports Medicine · Paper · PubMed

Does training at long muscle lengths grow more muscle?

A Bayesian meta-analysis of 12 small trials found only trivial extra regional growth (SMD 0.05-0.09), slightly more distally. Mostly untrained people, modest length contrasts.

A-fairly strong evidenceMeasured on non-athletes and recreational people

Varovic, Wolf, Schoenfeld, Steele, Grgic and Mikulic asked whether training a muscle at a longer average length grows particular regions of it more.

They searched three databases to April 2024 for randomised trials of at least four weeks that compared training at different joint angles (isometric) or through different ranges or exercises (dynamic) and measured muscle size at two or more points along the muscle by ultrasound, CT or MRI.

Thirteen studies qualified and 12 were meta-analysed (184 effects, 22 arms); two of them were isometric knee-extension trials.

For a typical contrast of 21.8% of muscle length between "shorter" and "longer" conditions, the extra growth from the longer condition was trivial at the proximal (SMD 0.05), middle (0.07) and distal (0.09) sites, with most of each posterior inside the pre-set region of practical equivalence. The estimates rose from proximal to distal.

In a non-registered sub-model limited to studies that changed range of motion, the distal advantage was larger (SMD 0.23). The authors conclude that longer mean muscle lengths do not seem to produce greater regional hypertrophy, at least for the length contrasts studied so far.

Included studies

Search: PubMed/MEDLINE, Scopus and Web of Science from inception to April 2024, plus forward and backward citation searches and the authors' libraries; two reviewers, a third for disputes; pre-registered on OSF (7 April 2024), with an addendum adding two inclusion rules after registration.
Inclusion: healthy young adults; at least 4 weeks; two or more groups or conditions at different joint angles (isometric) or through different ranges or muscle lengths (isotonic) using the same muscle actions; randomised within- or between-subject design; direct imaging (ultrasound, CT, MRI) at two or more sites along the muscle.
Studies: 13 included, 12 analysed (one reused a previous dataset). 11 untrained, 1 resistance-trained; 3 men only, 4 women only, 5 mixed; 8 lower-body (mostly quadriceps; also hamstrings), 4 upper-body (elbow flexors, triceps long head); 9 changed range of motion, 3 changed exercise; 5 to 15 weeks.
Isometric studies included: Alegre 2014 and Noorkõiv 2014 (knee-extension holds at long vs short length), the same trials behind the long-length advantage in REF-1158.
Muscle length: estimated from joint angles as a percentage of maximum joint range (biarticular muscles averaged across both joints); mean contrast 21.8 ± 13.6% between conditions (about 32.4% vs 54.2%).
Quality: SMART-LD mean 11 ± 2 (8-15); 6 fair, 6 poor.

Key Results

Site along the muscle (published abstract)SMD for longer vs shorter length (95% QI)Percent change, exponentiated lnRR (95% QI)
Proximal (25% of length)0.05 (-0.07 to 0.16)0.57% (-1.92 to 3.24%)
Mid-belly (50%)0.07 (-0.02 to 0.15)1.22% (-0.77 to 3.22%)
Distal (75%)0.09 (-0.01 to 0.19)1.88% (-0.44 to 4.34%)
Preprint detail (version 3)Result
Posterior inside the ROPE, SMD (±0.1): proximal / mid / distal82.7% / 77.68% / 59.02%
Probability of a meaningful advantage (SMD > 0.1): proximal / mid / distal16.54% / 22.3% / 40.96%
Posterior inside the ROPE, percent scale (±3%)96.66% / 96.62% / 83.45%
Range-of-motion studies only, distal site (not pre-registered)SMD 0.23 (0.07 to 0.35); 4.6% (-0.08 to 8.02%)
Secondary models (upper vs lower body, muscle group, muscle action)inconclusive, wide intervals
Data184 effects, 22 arms, 12 studies

Mechanisms

Authors' interpretation: the length-tension relationship predicts more passive and total tension at long lengths, but over the contrasts tested the hypertrophy difference was trivial relative to the usual overall growth from resistance training (SMD about 0.34 in another meta-analysis they cite).
Authors' interpretation: a distal trend, and a clearer distal effect when range of motion (lengthened partials) was the lever, may reflect region-specific activation; untested here (they suggest T2 MRI studies).
Inference: the model is linear in muscle length, so a much larger contrast than 22% (for example a hold at near-maximal length) could still matter; the data cannot say.

Practical Application

Does training at long muscle lengths grow more muscle? A Bayesian meta-analysis of 12 small trials found only trivial extra regional growth (SMD 0.05-0.09), slightly more distally. Mostly untrained people, modest length contrasts.
Train through a full range and do not chase lengthened-only work for extra size. The base's earlier lean toward long lengths (REF-680) holds as "at least as good", not "clearly better"; that synthesis is the base's.
For isometric holds, the long-length advantage in [REF-1158] rests on the same two to three small knee trials this review pooled with dynamic studies; [REF-1162] found long-length holds similar to full-range lifting. Choose the joint angle for comfort, the sticking point or range-specific strength rather than for growth.
A possible distal-region effect from lengthened partials is a non-registered sub-analysis; treat it as a hypothesis.

Limitations

Twelve small trials, all rated poor or fair; 11 of 12 in untrained young adults; mostly quadriceps; only seven studies measured the most proximal site.
The average length contrast was only 21.8%; only one study trained at maximal muscle length, so conclusions do not extend to bigger contrasts.
Muscle length was estimated from joint angles under simplifying assumptions (equal, linear contributions of both joints for biarticular muscles).
The authors call their pre-registered analysis plan imprecise, with researcher degrees of freedom; sampling iterations and inclusion rules deviated from registration (disclosed). Secondary and range-of-motion sub-models are exploratory.
Full text read from the preprint; the published version's numbers differ slightly and other changes cannot be excluded.
Funding: none. Competing interests: BJS serves on the scientific advisory board of Tonal Corporation (fitness equipment); JS provides research and statistical consulting to health and fitness organisations. Some authors helped set the smallest effect size of interest for a pre-registered lengthened-partials trial.
How this entry was checked and changed
Why this level

meta-analysis of small randomised trials judged poor (6) or fair (6) on SMART-LD, 11 of 12 in untrained people, modest length contrasts; full text read from the SportRxiv preprint version 3 (SRXIV.464, updated 11 October 2024), because the journal version is closed. The published version may differ: its abstract gives slightly different estimates (proximal SMD 0.05 vs 0.04 in the preprint), and the Key Results use the published abstract's figures

Full text (checked 2026-10-16)
Preprint version: "The manuscript was last updated on October 11, 2024"; marked "PREPRINT - NOT PEER REVIEWED" (full text).
Search to April 2024; pre-registered on the Open Science Framework on April 7th, 2024 (full text).
13 studies included, 12 meta-analysed; 184 effects nested within 22 intervention arms (full text).
Participants 18.8 – 27.2 years of age; eleven studies untrained, one resistance-trained; duration 5 to 15 weeks (full text).
SMART-LD mean 11 ± 2 (range: 8-15 points); six fair, six poor (full text).
Mean contrast in muscle length between conditions 21.8 ± 13.6%, between 32.4% and 54.2% (full text).
SMD slopes 0.04 [-0.07, 0.15] at 25%, 0.07 [-0.02, 0.15] at 50%, 0.09 [-0.01, 0.19] at 75% (full text).
Within the ROPE: 82.7%, 77.68%, 59.02%; meaningful positive effect 16.54%, 22.3%, 40.96% (full text).
lnRR: 0.48%, 1.14%, 1.8%; within the ROPE 96.66%, 96.62%, 83.45% (full text).
ROM-only studies, distal site: SMD 0.23, 95%QI [0.07, 0.35]; 4.6% [-0.08%, 8.02%] (full text).
"No funding was received for this project"; BJS on the scientific advisory board of Tonal Corporation (full text).