REF-1163 · Lum et al., 2026 · Journal of Strength and Conditioning Research · Paper · PubMed

Push or hold for isometrics?

In 10 adults over 6 weeks, pushing at 70% MVC raised concentric and eccentric strength; holding a load did not, but grew the quadriceps more. Mostly small effects; a first signal only.

Blimited evidenceMeasured on recreational people

Lum, Oranchuk, Chen and Kong compared the two ways of doing an isometric contraction. In pushing isometric muscle action (PIMA, "overcoming") the limb pushes against a fixed lever; in holding isometric muscle action (HIMA, "yielding") the limb keeps a set angle against a movable load that would otherwise move it.

Ten adults trained one leg each way: 4-6 contractions of 20 s at 70% of maximal voluntary contraction (MVC) at an 80-degree knee angle, 12 sessions over 6 weeks. Both styles raised isometric torque (g 0.30-0.36). Only pushing raised concentric (g 0.31) and eccentric (g 0.50) torque.

Both thickened the muscle (g 0.27-0.88) and lengthened vastus lateralis fascicles (g 0.20-0.31), but holding grew total quadriceps (g 0.46) and especially rectus femoris (g 1.44) thickness more, and only holding increased patellar-tendon thickness (g 0.12, trivial). The authors suggest pushing may suit strength and holding morphology, and call for larger, more realistic trials.

Protocol

Participants: 10 adults (5 men, 5 women), age 31.4 ± 6.5 years; 20 limbs randomised to PIMA or HIMA.
Pushing (PIMA, overcoming): force exerted against a fixed lever.
Holding (HIMA, yielding): the limb held a set joint angle while resisting an isotonic (movable) lever.
Dose (both): 70% MVC, 20 s per contraction, 4-6 repetitions, knee at 80 degrees; 12 sessions over 6 weeks. Rest intervals and progression not given in the abstract.
Tests: isometric, concentric and eccentric knee-extension torque per limb; ultrasound of quadriceps architecture (thickness, fascicle length) and patellar-tendon thickness, before and after.

Key Results

OutcomePushing (PIMA)Holding (HIMA)Statistics (within each condition unless marked)
Isometric torqueincreasedincreasedboth p < 0.05, g = 0.30-0.36
Concentric torqueincreased (p < 0.05, g = 0.31)not increased
Eccentric torqueincreased (p < 0.01, g = 0.50)not increased
Muscle thickness (sites)increasedincreasedboth p < 0.05, g = 0.27-0.88
Vastus lateralis fascicle lengthincreasedincreasedboth p < 0.05, g = 0.20-0.31
Total quadriceps thickness gainsmallerlargerp < 0.05, g = 0.46
Rectus femoris thickness gainsmallerlargerp < 0.05, g = 1.44
Patellar-tendon thicknessnot increasedincreased (p < 0.05, g = 0.12)
n = 10 people, 20 limbs.

Mechanisms

Authors' interpretation: the two isometric actions differ in motor control, so they deliver different stimuli; pushing may transfer better to dynamic strength, holding may favour morphological change.
Inference: holding contractions fatigue sooner at the same relative force (REF-1164), which could mean more motor-unit recruitment and metabolic stress per contraction; not measured here.
Inference: a patellar-tendon thickness effect of g = 0.12 is trivial in size and, at 70% MVC, sits right at the threshold REF-1158 proposes for tendon adaptation.

Practical Application

Push or hold for isometrics? In 10 adults over 6 weeks, pushing at 70% MVC raised concentric and eccentric strength; holding a load did not, but grew the quadriceps more. Mostly small effects; a first signal only.
Rectus femoris thickness was the largest difference (g 1.44 in favour of holding); tendon thickness rose only with holding, by a trivial g 0.12.
If the goal is strength that shows up in movement, push against something fixed; if the goal is muscle or tendon tissue, holding a load may do more. That split is the authors' suggestion from one small trial.
Holding may be harder to sustain at the same intensity (REF-1164, elbow extensors at 80% MVC); in this trial both styles completed 20-s holds at 70%, so shorter yielding holds at higher loads is the base's inference.
Base's practice, not the paper's: for endurance athletes neither style replaces dynamic heavy lifting (SYNTH-43, SYNTH-44).

Limitations

Ten people, 6 weeks, 12 sessions, one exercise; the authors call for larger samples and more ecologically valid programmes.
Within-subject unilateral design: cross-education between limbs could blur differences.
From the abstract: training status, rest, progression, how force was matched between conditions and the exact statistics behind "greater in HIMA" are not given.
Many outcomes and no stated correction for multiple comparisons in the abstract; the tendon effect (g 0.12) is trivial.
Funding and conflicts of interest: not stated in the abstract.
How this entry was checked and changed
Why this level

within-subject design with n=10 (counts as a crossover under 15); first training study of this question, small and short, effects mostly small; from the abstract