Concurrent Training Interference: 2026 Meta-Analysis

Concurrent training is the simultaneous pursuit of resistance and endurance adaptations within the same training week. Hickson opened the field in 1980 with a study showing that adding endurance work to a strength program blunted lower-body strength gains by approximately 30 percent over ten weeks.^[7] The phenomenon was named "interference" and the question of how, when, and how badly two opposing stimuli interfere has driven hundreds of trials since.

The 2026 picture is meaningfully different from the Wilson 2012 meta-analysis that anchored a generation of strength coaches. Newer studies, better dose control, and a clearer mechanistic picture have shrunk the average interference effect by roughly two-thirds for hypertrophy and made the interaction more dose-and-mode dependent. This article walks through Wilson 2012, the Schumann 2022 update, the Markov 2023 meta-regression, and the practical splits that fall out.

The Wilson 2012 baseline

Wilson and colleagues pooled 21 concurrent-training trials in 2012 and computed weighted effect sizes for hypertrophy, lower- and upper-body strength, and power.^[1] Headline numbers:

• Hypertrophy: pooled effect -0.35 (95% CI -0.55 to -0.15). Roughly a third of a standard deviation slower mass gain in concurrent groups vs strength-only.
• Lower-body strength: pooled effect -0.18. Borderline meaningful.
• Upper-body strength: pooled effect 0.00. No interference detected.
• Power output: pooled effect -0.78. Large negative effect, the strongest signal in the analysis.
• Mode interaction: running showed larger interference than cycling on every outcome.
• Frequency interaction: more than 3 endurance sessions per week worsened the strength outcome.

The Wilson model became the canonical reference. It was easy to summarise in a single sentence: running hurts hypertrophy more than cycling, more endurance is worse, power is the most fragile outcome.

The Schumann 2022 update

Schumann and colleagues re-ran the meta-analysis in 2022 with 27 additional trials published since Wilson 2012, plus stricter inclusion criteria for resistance-program quality.^[2] Several findings reversed or shrank.

• Hypertrophy interference: pooled effect -0.11 (95% CI -0.21 to -0.01). Roughly a third of the Wilson estimate. Statistical significance held; clinical significance is closer to noise for moderate-volume lifters.
• Lower-body 1RM: no detectable interference (pooled effect -0.02). Even with high endurance frequency, 1RM gains were equivalent to strength-only controls.
• Power and rate-of-force-development: still negative (pooled effect -0.30 for jump performance). The power outcome remains the most fragile.
• Type II fibre cross-sectional area: showed selective interference. Type I fibre area was unaffected by endurance addition.

The shift between Wilson and Schumann reflects a real change in the studied populations: newer trials use more carefully controlled resistance programs, with adequate volume and recovery, while older trials often pitted a moderate strength block against a heavy endurance load with insufficient recovery. The interference effect was partly an artefact of poor program design.

The Markov 2023 meta-regression

Markov, Hauser, and Chaabene 2023 went further and ran a meta-regression on the moderators of the interference effect.^[3] Their dataset (45 trials, 1,370 participants) generated dose-response curves rather than a single pooled effect size.

• Endurance frequency: each additional weekly endurance session beyond 2 reduced hypertrophy effect size by about 0.05. At 5 sessions per week the cumulative interference matched the original Wilson estimate.
• Endurance mode: running reduced hypertrophy effect size by an additional -0.15 vs cycling, holding total volume constant.
• Endurance intensity: high-intensity intervals (above lactate threshold) interfered less than steady-state work matched for total kilojoules. Counter-intuitive but consistent with the AMPK activation literature.
• Recovery interval: less than 6 hours between sessions doubled the interference signal vs more than 24 hours.
• Order of work: resistance-before-endurance produced slightly better hypertrophy outcomes (effect +0.08) than the reverse, but the effect was small and inconsistent across studies.

The meta-regression makes the interference effect a function of dose, mode, and timing rather than a fixed property of concurrent training. A lifter doing two cycling sessions per week, separated from lifting by 24 hours, sits near the no-interference point on the curve. A lifter doing five running sessions, half of them within 4 hours of lifting, sits at the steepest part of the slope.

The mechanism: AMPK-mTOR cross-talk

Atherton 2005 proposed the molecular mechanism that anchors the modern interference framework.^[6] Endurance work activates AMP-activated protein kinase (AMPK), a metabolic sensor that responds to depleted ATP and rising AMP. Resistance work activates the mTOR pathway, the master regulator of muscle protein synthesis. AMPK and mTOR have inhibitory cross-talk: when AMPK is high, mTOR signalling is suppressed.

The clinical consequence is that endurance work performed shortly before a resistance session attenuates the muscle protein synthesis response to that session. The duration of the AMPK suppression is not infinite; AMPK signalling decays over 3 to 6 hours after moderate-intensity endurance work, longer after exhaustive work. This is the molecular basis for the "6-hour rule": separate endurance and resistance by at least 6 hours and the AMPK signal has cleared before the lifting stimulus is presented.

Robineau and colleagues 2016 tested this directly.^[10] Three groups trained concurrently with 0, 6, or 24 hours between endurance and resistance work for 8 weeks. Hypertrophy and 1RM gains scaled with separation: the 24-hour group matched strength-only controls; the 0-hour group showed roughly half the gains. The 6-hour group landed in between, closer to the 24-hour outcome.

Order of operations

If both modalities have to happen on the same day, which order is better? Eddens 2018 ran a controlled comparison and reported small advantages for resistance-first, but the effect was inconsistent and modulated by which adaptation was being optimised.^[8] Enright 2017, working with soccer players, found no order effect on soccer-specific outcomes.^[9]

Practical synthesis: when same-day work is unavoidable, do the priority modality first while the system is fresh. For a hypertrophy-priority block, lift first. For a VO2-max-priority block, run intervals first.

Endurance-mode and intensity interactions

Gergley 2009 compared cycling vs running endurance modes added to a 9-week strength program.^[11] Cycling produced no interference with squat 1RM; running produced a measurable deficit. The mechanical eccentric loading of running adds direct interference to lower-body recovery that cycling avoids.

Blagrove 2018 reviewed the strength-training literature for distance runners and found the inverse: running performance improves with added strength work, especially explosive lifting (heavy loads at high velocity, jump training).^[5] The interference is asymmetric. Strength work improves running economy without harming aerobic capacity; running, especially high volumes, blunts hypertrophy and power.

Practical splits by goal

Synthesising the 2022 to 2023 evidence into programmable splits:

Hypertrophy priority (lifter who runs)
  Resistance:           4 sessions/week, full programmed volume
  Endurance:            2 sessions/week max
  Mode:                 cycling preferred over running
  Intensity:            zone 2 base + 1 short interval session
  Separation:           24 hours from any lifting session
  Expected interference: minimal (Schumann pooled -0.05 to -0.10)

Endurance priority (runner who lifts)
  Resistance:           2 sessions/week, heavy + explosive
  Endurance:            5–6 sessions/week, full volume
  Mode:                 running (the priority specificity)
  Lifting focus:        compound, low-rep, high-intent
  Separation:           lift on quality run days, after the run
                        AM run + PM lift, or vice versa
  Expected interference: small for running performance,
                         hypertrophy is gated to 60-70% of strength-only

Hybrid athlete (true concurrent)
  Resistance:           3–4 sessions/week
  Endurance:            3–4 sessions/week
  Mode:                 mix cycling + running, with running for race specificity
  Separation:           same-day if needed, prefer 6+ hours
  Order:                priority adaptation first
  Expected outcome:     ~10-15% interference per modality vs single-focus
                         training; total work capacity exceeds either
                         single-focus program

Where the meta-analyses still disagree

• Magnitude of upper-body interference. Wilson 2012 reported zero; Schumann 2022 reported a small but non-zero negative effect for bench press. The effect sizes are within the noise floor.
• Concurrent training in older adults. Doma 2017 reviewed acute interference and noted larger effects in untrained older populations.^[4] Schumann 2022 includes both age groups and reports no significant moderation by age, but the older-adult sample is small.
• Female-specific data. The included studies are predominantly male. The Schumann update flags this as a limitation; the few female-only trials suggest similar but possibly slightly smaller interference effects.
• Long-duration trials. Most studies run 8 to 12 weeks. The interference effect compounded over a year of concurrent training is largely extrapolated, not measured.

Cross-link tools

• 1RM Calculator tracks the strength outcome that interference is most likely to compromise on power-priority blocks.
• VO2 Max Estimator tracks the endurance outcome on the other side.
• Workout Volume Calculator manages the resistance volume that determines whether the interference window matters.

• Wilson 2012 set the field with moderate hypertrophy interference, large power interference, and no upper-body strength interference.
• Schumann 2022 shrank the hypertrophy effect to roughly a third and showed no detectable lower-body 1RM interference.
• Markov 2023 made the interference effect dose- and mode-dependent: frequency, mode (running vs cycling), and recovery interval each independently moderate the size.
• The molecular mechanism is AMPK-mTOR cross-talk; separating sessions by 6 to 24 hours largely resolves the conflict.
• Practical splits cap endurance at 2 to 3 sessions for hypertrophy priority, prefer cycling, and put separation between modalities when possible.