REF-1523 · Held et al., 2026 · Sports Medicine · Paper · PubMed

Does cardio cancel my strength gains?

In an umbrella of 17 meta-analyses strength, power and size came out comparable to lifting alone; concluded for recreationally trained people. Abstract only; comparable is not zero interference.

Main caveat: Abstract only. Overlap, certainty, review quality, heterogeneity and all intervals are unread; nothing here can be reported on them.
fairly strong evidenceMeasured on non-athletes to trained athletes

Held and colleagues searched for meta-analyses of concurrent training, the combination of resistance and endurance training, and summarised 17 of them (144 individual studies, 1492 healthy participants) for aerobic capacity, muscle strength, power and hypertrophy. Aerobic capacity came out comparable with endurance training alone, and strength better (SMD 0.59).

Against resistance training alone, aerobic capacity was better (SMD 0.77), and strength, power and hypertrophy came out comparable. On sequence the authors report no significant effects and a trend favouring resistance first.

They call concurrent training valuable for comprehensive fitness in recreationally trained individuals and say data in highly trained to elite athletes remain scarce. Our reading: a useful high-level summary for general adults.

It is not a new data source: it cites REF-330, REF-512 and REF-588 among its references, so it should not count as independent confirmation of them.

Its "comparable" on power sits beside the explosive-strength loss in REF-330 (SMD −0.28) without cancelling it, because a summary of reviews that may overlap, reporting no difference, cannot erase a specific pooled effect, and the abstract gives no interval to show how large a loss it could exclude.

Protocol

Design: umbrella review of meta-analyses; PRISMA guidelines, four-eyes principle; PROSPERO CRD42025646460.
Search: PubMed, Web of Science, SPORTDiscus, PsycNet, to 28 February 2025.
Who: n = 1492 healthy participants, across 144 individual studies in 17 meta-analyses (as printed; see Limitations).
Outcomes: aerobic capacity, muscle strength, power, hypertrophy.
Planned subgroups: training modality, performance level, age, resistance-training load, endurance intensity. The abstract gives no result for any of these except its overall statement on performance level.
Not read: which 17 meta-analyses, their quality scores, any overlap index, any certainty rating, heterogeneity, intervals, and the supplementary XLSX on the publisher's page. PubMed lists 188 references; among those we resolved by PMID are REF-330, REF-512, REF-588, Petré 2021, Lundberg 2022, Huiberts 2024 and the AMSTAR paper (Shea 2009). Being cited is not the same as being included.
Funding and interests: the publisher's page states that no funding sources were used; the PubMed declarations state that all eight authors have no conflicts of interest relevant to the review.

Key Results

ComparisonOutcomeAs reported in the abstract
Concurrent vs endurance onlyAerobic capacityComparable (no number)
Concurrent vs endurance onlyStrengthSMD 0.59; p < 0.001, favours concurrent
Concurrent vs resistance onlyAerobic capacitySMD 0.77; p = 0.02, favours concurrent
Concurrent vs resistance onlyStrength, power, hypertrophyComparable (no numbers)
Resistance before endurance vs the reverseStrengthSMD 1.69; p < 0.001, called a trend
Resistance before endurance vs the reverseHypertrophySMD 0.83; p = 0.36, called a trend
A statement that does not add up: the abstract says no significant effects of sequence were found, then gives p < 0.001 for strength, and its conclusion says no relevant differences. Without the full text we cannot tell which holds.
A count that looks odd: 1492 participants over 144 studies is about 10 per study (our arithmetic). REF-330 alone reports 1090 participants in 43 studies. If it is among the 17, the printed total cannot be a simple sum over all included trials; how the authors counted is unread.

Mechanisms

Measured: nothing mechanistic in what was read; the outcomes are training adaptations pooled from earlier meta-analyses.
Authors' interpretation: concurrent training suits comprehensive fitness in recreationally trained people, and resistance first may help strength and hypertrophy.
Inference: the 17 meta-analyses probably share many primary trials (several cited reviews cover the same question and years), which is why an umbrella's agreement is weaker than it looks unless overlap is quantified. That quantification, if present, is in the unread text.

Practical Application

Does cardio cancel my strength gains? In an umbrella of 17 meta-analyses strength, power and size came out comparable to lifting alone; concluded for recreationally trained people. Abstract only; comparable is not zero interference.
For a general adult who wants both: the direction is reassuring. Aerobic capacity comparable with endurance-only and strength, power and size comparable with strength-only, by the abstract.
For an endurance athlete: the authors state that highly trained to elite data are scarce, so this entry neither shows nor rules out a cost at a large aerobic volume. Do not cite it as "no interference"; REF-330's explosive-strength loss and same-session finding stay the working guide for power, and the reasons to lift rest on REF-071, not on this.
Order: do not quote SMD 1.69 for strength first. The abstract's own text is inconsistent; the base's number for order stays REF-588 (+6.91%, lower-body dynamic strength). Not from this paper; the base's own practice: strength first when the two must share a session (SYNTH-21).
Not an endurance performance result: aerobic capacity is what was summarised. Time-trial and economy evidence is REF-071.

Limitations

Abstract only. Overlap, certainty, review quality, heterogeneity and all intervals are unread; nothing here can be reported on them. The numbers above are the only ones available.
A null, mostly without numbers. The comparable results for strength, power and hypertrophy carry no effect size or interval in the abstract, so the size of a loss they can exclude is unknown.
Populations. Healthy; the conclusion is for the recreationally trained and the sample's training status is unread; the highly trained are scarce by the authors' own statement. Age and modality subgroups are unreported in the abstract.
Probably not independent. Its reference list contains meta-analyses already in the base; cited is not proven included, and how much trial evidence it shares with them is unread.
Internal inconsistency. Sequence wording vs p < 0.001; participant count vs REF-330's own (see Key Results).
Interests. No funding sources and no relevant conflicts stated.
How this entry was checked and changed
Why this level

the rubric allows A+ for an umbrella review; held at A- because only the abstract was read, so review quality, overlap between the 17 meta-analyses, certainty of evidence and every confidence interval are unseen; the printed 1492 participants over 144 studies would mean small trials, though that count is itself unexplained; the abstract's sequence wording contradicts its own number