REF-1156 · Bohm et al., 2021 · Proceedings of the Royal Society B: Biological Sciences · Paper · PubMed · Free full text

Heavy isometric calf training lowered the energy cost of running about 4% in recreational runners

14 weeks of 90% MVC isometric plantarflexion (5 × 4 × 3 s, 3-4×/week) raised plantarflexor strength 10% and Achilles stiffness 31%, and the soleus then shortened at more efficient speeds; the fall in cost was a within-group effect, not a significant difference from controls

Blimited evidenceMeasured on recreational people

Bohm and colleagues repeated their group's 14-week isometric calf programme (REF-1155) in recreational runners to find out why it improves economy. Thirty-six runners were randomised; 13 in the intervention group and 10 controls finished.

Plantarflexion strength rose 10% (3.12 to 3.44 Nm/kg) and Achilles tendon stiffness 31% (85 to 111 kN/strain), both with significant group × time interactions, while controls did not change. The energy cost of running at 2.5 m/s fell from 10.6 to 10.2 W/kg (about 4%, p = 0.028 within group); controls went from 11.2 to 11.1 W/kg (p = 0.688).

For energy cost, the table marks a significant main effect of time but not a group × time interaction. Foot strike, stance time, cadence and joint kinematics did not change.

In the trained runners, soleus fascicles tended to shorten faster early in stance (p = 0.073) and shortened significantly slower late in stance (p = 0.007), closer to the velocity of peak efficiency, so their estimated enthalpy efficiency rose 15% during the lengthening phase of the muscle-tendon unit and 7% over the whole stance. Soleus EMG over stance fell 12%.

The authors conclude that the economy gain comes from the soleus working at a more efficient shortening velocity and recommend this muscle-tendon training to improve running economy.

Protocol

Participants: 36 recreational runners randomised (19 intervention, 17 control); inclusion: age 20-40, at least two running sessions a week, no muscle or tendon injury in the past year, habitual rearfoot strike (strike index < 0.3). Completers: intervention 13 (29 ± 5 y, 73 ± 8 kg, 4 women), control 10 (31 ± 3 y, 70 ± 11 kg, 7 women). Sample size: a priori power analysis required 12 in the intervention group.
Training: supervised, biofeedback-based isometric ankle plantarflexion (overcoming, against a dynamometer), 90% of MVC, adjusted every two weeks; 5 sets of 4 contractions, 3 s loading and 3 s relaxation; 3-4 times a week for 14 weeks, added to the runners' usual training. "The same 14-week muscle–tendon training" as REF-1155. The training joint angle is not given in the main text (the supplementary material was not read).
Control: continued usual running.
Tests: plantarflexion moment by inverse dynamics; Achilles stiffness from ramp MVCs with ultrasound (50-100% of maximal tendon force); energy cost from gas exchange in the last 3 min of an 8-min treadmill run at 2.5 m/s (RER below 1.0); kinematics, soleus fascicle length and velocity by ultrasound, EMG; soleus force-length and force-velocity relationships; enthalpy efficiency from fascicle velocity (intervention group only).

Key Results

OutcomeIntervention pre → post (n = 13)gControl pre → post (n = 10)gStatistics
Plantarflexion moment, Nm/kg3.12 ± 0.48 → 3.44 ± 0.370.773.10 ± 0.46 → 2.99 ± 0.320.32group × time significant; within p = 0.024 vs p = 0.296
Achilles stiffness, kN/strain85 ± 36 → 111 ± 590.6773 ± 29 → 71 ± 280.10group × time significant; within p = 0.048 vs p = 0.745
Energy cost at 2.5 m/s, W/kg10.6 ± 0.6 → 10.2 ± 0.70.7411.2 ± 1.0 → 11.1 ± 1.00.12main effect of time only; within p = 0.028 vs p = 0.688
Cadence, steps/min160 ± 11 → 159 ± 90.39162 ± 9 → 161 ± 90.26no change
Stance time, ms310 ± 23 → 316 ± 230.29327 ± 17 → 324 ± 230.34no change
Soleus (intervention group only): enthalpy efficiency during MTU lengthening up 15% (p = 0.006, g = 0.85), over stance up 7% (p = 0.025, g = 0.66), no change during MTU shortening (p = 0.640). Fascicle velocity during MTU shortening fell (p = 0.007, g = 0.83). Stance-phase EMG fell 12% (p = 0.017). Maximal Achilles strain during MVC unchanged (6.2 ± 1.6% → 6.0 ± 1.2%, p = 0.501).
Baselines: no significant group differences in strength (p = 0.894), stiffness (p = 0.421) or energy cost (p = 0.143), although controls started at 11.2 vs 10.6 W/kg.

Mechanisms

Measured: after training, soleus fascicles shortened at velocities closer to the peak of the efficiency-velocity curve (about 0.18 Vmax) during the first part of stance, so estimated enthalpy efficiency rose; whole-stance shortening, force-length and force-velocity potentials did not change.
Measured: stance EMG fell while maximal strength rose without a change in maximal EMG, which the authors read as hypertrophy rather than neural adaptation; estimated average soleus force did not change.
Authors' interpretation: the lower cost of running comes from the soleus doing its work at a more efficient shortening velocity, not from less active muscle volume; the first, lengthening half of stance matters for overall cost.
Authors' interpretation: gaining both strength and tendon stiffness keeps tendon strain balanced (maximal strain unchanged), which they see as protective; strength without stiffness could raise strain.
Inference: efficiency is modelled from frog and mouse muscle data and a fixed Vmax; the authors' sensitivity analysis held the effect down to Vmax -30%.

Practical Application

Heavy isometric calf holds (90% MVC, 14 weeks) cut running cost ~4% in 13 recreational runners, with a 31% stiffer Achilles, but the change was not significant against controls and 13 of 36 dropped out. Easy pace only. Strength rose 10%.
The programme, copied: isometric calf push at 90% of maximal voluntary force, re-tested every two weeks; 3 s on, 3 s off, 4 per set, 5 sets; 3-4 sessions a week for 14 weeks, on top of usual running. It was supervised with force biofeedback.
Why it might work (authors): the soleus works at a more efficient speed in early stance. Stride, cadence and foot strike did not change, so the athlete will not feel a different stride.
Not from this paper; the base's practice: two trials from one laboratory point the same way (REF-1155 and this one), but a meta-analysis that pooled isometric studies found no significant effect (REF-801) and did not include this trial. Use this as an optional add-on to heavy lifting (SYNTH-44), not as the main strength work.

Limitations

13 of 36 randomised participants did not finish (6 intervention, 7 control); analysis of completers only, reasons not given in the main text.
The energy-cost change is a within-group result; the table flags a main effect of time, not a group × time interaction, so the trial does not show the intervention beat control for economy.
Groups differed in sex mix (4 of 13 vs 7 of 10 women) and in baseline energy cost (10.6 vs 11.2 W/kg, p = 0.143).
The fascicle and efficiency analysis was done in the intervention group only; the authors argue controls would not have changed.
One easy speed (2.5 m/s) on a treadmill; rearfoot strikers only; no race or time-trial outcome; no blinding reported.
Enthalpy efficiency is modelled, not measured.
Funding: German Federal Institute of Sport Science (grant ZMVI14-070604/17-18). The authors declare no competing interests.
How this entry was checked and changed
Why this level

randomised parallel groups, but n=23 completers of 36 randomised (13 of 36 lost, 36%); energy cost fell within the intervention group while the table flags only a main effect of time, not a group × time interaction; the mechanism outcomes have no control data; read from the full text

Full text (checked 2026-10-16)
Recruited 36 participants, randomly assigned to intervention (n = 19) or control (n = 17); twenty-three completed: 13 intervention (four females), 10 control (seven females) (full text).
Training: five sets of four isometric ankle plantar flexion contractions (3 s loading and 3 s relaxation) at 90% of MVC, adjusted every two weeks, three to four times a week for 14 weeks (full text).
Plantar flexion moment 3.12 ± 0.48 to 3.44 ± 0.37 Nm kg−1 (g 0.77) vs control 3.10 ± 0.46 to 2.99 ± 0.32 (full text).
AT stiffness 85 ± 36 to 111 ± 59 kN strain−1 (g 0.67) vs control 73 ± 29 to 71 ± 28 (full text).
Energy cost 10.6 ± 0.6 to 10.2 ± 0.7 W kg−1 (g 0.74; p = 0.028) vs control 11.2 ± 1.0 to 11.1 ± 1.0 (p = 0.688); Table 1 footnote b, significant main effect of time (full text).
Soleus enthalpy efficiency higher in MTU lengthening (p = 0.006, g = 0.85) and over the entire stance phase (p = 0.025, g = 0.66) (full text).
Maximum AT strain during MVCs pre 6.2 ± 1.6%, post 6.0 ± 1.2%, p = 0.501 (full text).
Running trial 8 min at 2.5 m s−1, average of the last 3 min (full text).
Funding: German Federal Institute of Sport Science (grant no. ZMVI14-070604/17-18); no competing interests (full text).