REF-1476 · Currier et al., 2026 · Medicine & Science in Sports & Exercise · Paper · PubMed · Free full text

What does ACSM's 2026 stand say on how to progress a lift?

Almost nothing direct. Across 137 reviews, load, sets and 2+ sessions a week moved strength; failure and periodisation did not clearly. No rule for when to add or how much.

Main caveat: Variable-level verdicts, no pooling. The authors say the method prevents statistical insight into comparative effectiveness or dose-response, yet Table 6…
strong evidenceMeasured on non-athletes to trained athletes

This stand updates the 2009 ACSM "Progression models" stand with a registered overview of systematic reviews of randomised trials in healthy adults (at least 6 weeks of training, at least 12 exposures), searched to October 2024.

It takes 137 reviews, scores each with AMSTAR, and turns them into a verdict per prescription variable: impacts, does not impact, cannot determine, with a quality-of-evidence percentage. Against no exercise, resistance training improved strength, size, power, endurance, contraction velocity, gait speed, balance, chair stand and timed up-and-go.

Between programmes, few variables mattered: for strength, load (at least 80% 1RM, dose-response), sets (2-3 per session), frequency (at least 2 sessions a week), starting the session with the lift, full range and flywheel eccentrics; for size, sets (at least 10 per muscle a week) and eccentric overload; for power, 30-70% 1RM, Olympic-style lifting and fast concentrics.

Failure, equipment, set structure, time under tension and periodisation did not consistently matter. On progression itself the stand is thin. Its three "Progression" rows are intrasession autoregulation (cannot determine), varied exercise selection (cannot determine) and periodisation (cannot determine for strength, no impact on size).

Progressive overload appears only in the discussion, as a concept: the authors write that progression is "not necessary to achieve beneficial outcomes" and that increasing the stimulus is "likely a requirement only for those seeking continued longer term progress", while the same paragraph says "progression is a requirement to achieve a meaningful benefit" for some inexperienced people whose loads are necessarily low, a later one says untrained people benefit "provided that progression" is among the core principles, and the abstract's conclusion says healthy adults should perform progressive training.

No increment, trigger or hold rule is offered. The authors' primary recommendation is high effort at least twice weekly with all major muscle groups, and they rank individualising for participation above specific prescription criteria. Our reading: a strong map of which levers matter in general adults, and honest silence on how to move them from session to session.

Protocol

Design: overview of systematic reviews; registered (INPLASY202360071), reported to PRIOR. Updates the ACSM 2009 stand.
Search: Ovid MEDLINE, Ovid Emcare, Ovid Embase, Cochrane Database of Systematic Reviews, SPORTDiscus, Web of Science, October 2024; 5751 records after duplicates, 137 reviews included.
Eligibility: reviews of randomised trials in healthy adults (≥18 yr), resistance training ≥6 wk (range 6-52 wk) with at least 12 exposures, against no exercise or a distinct programme. No separation by age. "Standard" resistance training is whatever a review did not specify.
Who: n = 30000 or more participants across reviews (">30,000" as printed); the authors state much of the evidence is from inexperienced trainees. No sport-specific analysis.
Appraisal: AMSTAR (1 to 11); included reviews scored 1 to 9. Screening and extraction by four reviewers independently, in duplicate.
Certainty: an outcome-level quality-of-evidence score based on GRADE, from review design and AMSTAR, averaged and expressed as a percentage. The authors state no standardised method exists for overviews and that cohesive GRADE pooling was not possible.
Overlap: corrected covered area for strength only ("moderate", 6-10%); not computed for other outcomes.
Funding and interests (as printed): "No specific source of funding was used for this work." B. S. C. was supported by an Alexander Graham Bell Canada Graduate Scholarship-Doctoral. S. M. P. thanks the Canada Research Chairs program and acknowledges grant support from NSERC and CIHR. B. J.S. formerly served on the scientific advisory board for Tonal Corporation, a manufacturer of fitness equipment. "No other authors reported conflicts relevant to this work." Published as an official ACSM pronouncement, reviewed by members-at-large and the Pronouncements Committee.

Key Results

Between programmes (Table 4). Each cell gives the verdict, then the reviews behind it. ↑ impact, ? cannot determine, ↓ no impact is our reading of the table's arrows against the three directions named in the Methods. QoE = the authors' quality-of-evidence percentage.

VariableStrengthHypertrophy
LoadImpacts: ↑ 6 reviews, n = 6,574, QoE 79%; ? 1, n = 151; ↓ 2, n = 734No impact: ↓ 8 reviews, n = 5,340, QoE 66%; ↑ 1, n = 231; ? 1, n = 108
SetsImpacts: ↑ 7 reviews, n = 5,633, QoE 71%Impacts: ↑ 5 reviews, n = 2,267, QoE 50%; ? 1, n = 181; ↓ 1, n = 555
FrequencyImpacts: ↑ 4 reviews, n = 3,531, QoE 69%; ↓ 3, n = 608, QoE 75%No impact: ↓ 4 reviews, n = 3,177, QoE 63%; ↑ 1, n = 200; ? 2, n = 166
Failure or fatigueNo impact: ↓ 3 reviews, n = 1,371, QoE 83%; ↑ 1, n = 199No impact: ↓ 4 reviews, n = 800, QoE 69%; ↑ 1, n = 189
Exercise orderImpacts: ↑ 4 reviews, n = 941, QoE 88%No impact: ↓ 4 reviews, n = 759, QoE 81%; ? 1, n = 71
Range of motionImpacts: ↑ 2 reviews, n = 1,262, QoE 50%Cannot determine
Inter-set restNo impact: ↓ 2 reviews, n = 982, QoE 63%Cannot determine: ? 4 reviews, n = 265, QoE 44%
Progression: intrasession autoregulationCannot determine: ↑ 1 review, n = 308, QoE 75%; ↓ 1, n = 356, QoE 25%Cannot determine: ? 1 review, n = 243
Progression: varied exercise selectionCannot determine: ? 1 review, n = 198, QoE 50%Cannot determine: ? 1, n = 95; ↓ 1, n = 218
Progression: periodisationCannot determine: ↑ 1 review, n = 616; ? 1, n not reported; ↓ 2, n = 1,075No impact: ↓ 4 reviews, n = 1,414, QoE 63%
Concurrent trainingCannot determine: ↑ 2, n = 725; ? 1, n = 119; ↓ 3, n = 354Cannot determine
Against no exercise (Table 3): strength ↑ 26 reviews, n = 23,204, QoE 73%; hypertrophy ↑ 12 reviews, n = 14,924, QoE 79%; power ↑ 4 reviews, n = 1,001, QoE 63%. Circuit, elastic-band and home-based training also raised strength.
Recommended to enhance (Table 6): strength: at least 2 sessions a week, at least 80% 1RM (dose-response), flywheel eccentrics, full range, 2-3 sets per session, start of the session. Hypertrophy: eccentric overload, at least 10 sets a week (dose-response). Power: 30-70% 1RM, flywheel, Olympic-style and power training, low-to-moderate volume (reps times sets ≤24).
Sets, in the discussion: two sets beat one, but "the exact number of sets required to optimize adaptations cannot be ascertained". Reading REF-1477, the authors describe a plateau with diminishing returns beyond about 2-3 sets per exercise for strength and about 18-20 weekly sets for hypertrophy, and advise at least two sets per exercise.
Failure, in the discussion: not necessary for benefit. "Near-failure or a target of 2-3 repetitions in reserve" is offered as sufficient effort, then: "there is insufficient evidence to quantify exact RIR and perceived exertion targets".
Frequency, in the discussion: no dose-response could be concluded and the effect on strength is diminished when volume is equated.
Could not establish: the number of sets that optimises; the upper limit of frequency; exact RIR or RPE targets; whether periodisation, intrasession autoregulation or varied exercises change strength; concurrent training effects; a prescription for muscular endurance; minimal-dose programmes; any comparison between body regions; any comparison of progressing against not progressing.
Safety: the stand calls resistance training safe for healthy adults of all ages, citing a pooled analysis in which exercise did not raise serious adverse events. Lifting to failure "may" be inadvisable for some, such as older people, because of vascular risk and poor form (authors' suggestion, no effect size).

Mechanisms

Measured (review level): load moved strength more than size; sets moved both; failure moved neither consistently. These are variable-by-variable verdicts, and the authors state an overview cannot compare interventions against each other or show dose-response.
Authors' interpretation: physiological systems adapt to reduce the stress of a stimulus, so continued gains need the stimulus to grow "in some manner" (load, volume, frequency, exercise selection or duration, even if slightly). The same relative load can also be kept by raising absolute load as strength grows, by testing or perceived exertion. Neither statement is a tested comparison in this overview.
Inference: a programme that holds reps at a target and raises kg as the athlete gets stronger keeps relative load about constant rather than raising it. That is one reading of the authors' sentence, and is what a log can check.

Practical Application

What does ACSM's 2026 stand say on how to progress a lift? Almost nothing direct. Across 137 reviews, load, sets and 2+ sessions a week moved strength; failure and periodisation did not clearly. No rule for when to add or how much.
Levers the stand names for strength (it does not rank them, and says an overview cannot compare them): heavier load, more sets up to about 2-3 per exercise, the main lift early in the session, at least 2 sessions a week if the week allows. For most endurance athletes this supports a short session of heavy lifts, not more of everything. The dose as a target for an endurance athlete is not from this paper.
Effort: stopping about 2-3 reps short is called sufficient by the authors and is not a measured optimum. Agrees with the short-of-failure line in SYNTH-43 and SYNTH-44, which already label the reps-in-reserve number as practice.
%1RM: the app has none, and the stand itself notes that load can be stated as a repetition maximum and that conversions vary by exercise and between people. Not from this paper; the base's own practice: read a heavy load off low reps at a logged RPE.
Autoregulation and periodisation: the stand does not support calling either proven for strength. Against the base: REF-683 (A, 18 studies) reports periodised beating non-periodised on 1RM, and REF-685 (B+, published after this search closed) ranks autoregulated methods above fixed percentages.
Against the base, on progression: PHYS-39 rests "further gains need a progressing stimulus" on the 2009 stand.This stand updates 2009 and says increasing the stimulus is "likely a requirement only for those seeking continued longer term progress", yet also calls progression a requirement for some inexperienced people on low loads and a core principle for the untrained, and its abstract says adults should do progressive training.No tested comparison sits behind any of these sentences (REF-1481 is one, outside its search).

Limitations

Variable-level verdicts, no pooling. The authors say the method prevents statistical insight into comparative effectiveness or dose-response, yet Table 6 labels load and sets "dose-response".
Who. Much of the evidence is from inexperienced trainees; the last paragraph nevertheless speaks of what can be changed "in experienced trainees". Reviews spanning young and old were not separated.
Quality. AMSTAR tops out at 9 of 11; the authors call randomised trials in the field often poor quality, with small samples, poor randomisation and little preregistration. The at-least-10-sets threshold for size rests on a QoE of 50%, among the lowest in the table.
A null can reflect sparse data. The authors warn this for the Short Physical Performance Battery; it applies equally to other "cannot determine" and thin "no impact" cells.
Things that do not add up. (1) "As the results in Table 4 highlight, progression is not necessary", but Table 4 has no row comparing progression with none, and the abstract concludes healthy adults should perform progressive training; the discussion itself later calls progression a core principle for untrained people.(2) Table 6 gives 2-3 sets per session as enhancing strength; the discussion says the optimum cannot be ascertained. (3) Reference 198, cited for the insufficient evidence on RIR and RPE targets, is by its title a meta-analysis of multimodal training in frail older people.
Search date. October 2024; later trials such as REF-1481 are outside it, though the discussion cites REF-1477.
Interests. As printed above; one author formerly on an equipment maker's advisory board.
How this entry was checked and changed
Why this level

read from the full text (open access, CC BY-NC-ND). The rubric allows A+ for a governing-body statement built on systematic review; held one step below because the review-level AMSTAR scores reach only 9 of 11, the certainty score is the authors' own method ("no standardized method" exists), overlap between reviews was checked for strength only, much of the trial evidence is from inexperienced trainees, and the primary trials are described by the authors as often poor quality

Full text (checked 2026-10-30)
We synthesized data from 137 systematic reviews (>30,000 participants); search current to October 2024; training ≥6 wk, range 6-52 wk (full text)
AMSTAR scores for included reviews ranged from 1 (lowest) to 9 (highest) out of a possible 11 (full text)
Strength enhanced by ≥80% one-repetition maximum, 2-3 sets, beginning of training sessions and ≥2 sessions/wk; hypertrophy by ≥10 sets/wk and eccentric overload (full text)
Table 4 periodization, strength: 1 review n = 616, 1 review n = NR, 2 reviews n = 1,075; hypertrophy: 4 reviews n = 1,414, QoE = 63% (full text)
Table 4 fatigue/failure, strength: 1 review n = 199, 3 reviews n = 1,371, QoE = 83%; hypertrophy: 1 review n = 189, 4 reviews n = 800 (full text)
Table 4 intrasession autoregulation, strength: 1 review n = 308, QoE = 75%, and 1 review n = 356, QoE = 25% (full text)
Progression is not necessary to achieve beneficial outcomes, and overload is likely a requirement only for those seeking continued longer term progress (full text)
A plateau and diminishing returns beyond ~2-3 sets/exercise for strength and ~18-20 weekly sets for hypertrophy, per the meta-regression at reference 187 (full text)
Near-failure or a target of 2-3 repetitions in reserve; insufficient evidence to quantify exact RIR and perceived exertion targets (full text)
No specific source of funding was used for this work; B. J. S. formerly served on the scientific advisory board for Tonal Corporation (full text)