REF-1715 · Nuuttila et al., 2026 · Scandinavian Journal of Medicine & Science in Sports, · Paper · PubMed · Free full text
Easy endurance training alone builds fitness in untrained adults: training kept below the first threshold (or at most 60% of VO2max or heart-rate reserve, 75% of HRmax) for 3 weeks or more raised VO2max by 4.5 mL/kg/min over controls (Hedges' g 0.94, 95% CI 0.74 to 1.13; GRADE high) and peak cycling power by 35 W, in a meta-analysis of 50 RCTs with 824 people in the training groups (healthy, mostly sedentary, aged 18-65); between studies (not in head-to-head trials), higher intensity within that easy range went with larger VO2max gains, no minimum effective intensity was found down to about 45% of VO2max, and session length, weekly hours and frequency showed no dose-response; no trained athletes were included. Read in full
Main caveat: The authors': only one study set intensity from LT1/VT1, so physiological strain varied and some people may have trained in the heavy domain.
Nuuttila and colleagues (UKK Institute, Tampere, and the Universities of Jyväskylä, Turku and Tampere) pooled randomised trials in which healthy working-age adults trained only at low intensity, defined as below the first lactate or ventilatory threshold or, failing that, at most 60% of VO2max, VO2 reserve or heart-rate reserve, or 75% of HRmax, and asked what this does to aerobic fitness and cardiometabolic risk factors, and whether a lowest effective intensity exists.
Fifty studies qualified, with 824 people in the training groups, mostly sedentary or untrained, training on average 3.8 times a week for 41 min over about 12 weeks, mostly by cycling or walking and running.
Against controls, low-intensity training raised relative VO2max by a large amount (g 0.94, 4.5 mL/kg/min), absolute VO2max (0.84) and peak cycling power (1.09, 35 W), and VT1 moderately (0.74, 5 studies). Total and LDL cholesterol, triglycerides and blood pressure fell and HDL rose by small amounts; glucose did not change significantly.
Of the training characteristics, only intensity showed a dose-response with the VO2max gain: higher intensity within the low-intensity range went with larger gains, in subgroups and in a meta-regression, while session length, weekly volume, frequency, total dose and programme length showed no dose-response (some subgroup splits differed significantly, without a consistent pattern); of the participant characteristics, age moderated the VO2max gain, larger at 45 or under.
No intensity was ineffective down to the lowest tested, about 45% of VO2max. Heterogeneity was high and prediction intervals wide. The authors conclude that easy training improves fitness and, less, cardiometabolic health in untrained adults, that intensity within the easy range still matters for VO2max, and that gains come early and further gains need progression.
Participants
Pooled sample: 824 participants in 54 intervention groups and 708 in 50 control groups; intervention groups 441 men and 316 women, control groups 387 men and 244 women; 4 studies did not give the sex split of finishers.
Age: mean 35.5 ± 12.5 years (intervention) and 33.2 ± 11.8 (control); eligibility 18-65.
Fitness: mainly sedentary or untrained; mean baseline VO2max 32.5 mL/kg/min in both arms; only 3 studies had groups with a mean above 40 mL/kg/min; no study of trained or competitive athletes.
Health: healthy, BMI up to 35, no diagnosed disease or cardiometabolic medication, not pregnant; mean baseline blood glucose within the reference range in every study.
Protocol
Search: PubMed and SPORTDiscus, English, human studies, published before 31 October 2023, plus reference lists of related reviews and studies known to the authors; 9702 records screened, 50 studies included; 4 excluded as duplicate datasets.
Inclusion: RCTs (parallel, or crossover with control periods) of structured endurance-only training for at least 3 weeks, intensity exclusively below LT1/VT1 or within the fixed limits above (chosen because the authors' laboratory found them below LT1 in most people); excluded: unclear programmes, other training modes, nutritional, ergogenic or drug co-interventions, environmental changes.
Training in the trials: intensity set by % of HRmax in 21 studies (66.5 ± 5.7%), % of VO2max in 20 (53.2 ± 6.9%), % of heart-rate reserve in 8 (54.0 ± 7.0%) and by LT1/VT1 in 1; from a measured maximum in 36 studies, an estimated one in 13; 3.8 ± 1.1 sessions a week of 41 ± 13 min for 12.2 ± 5.7 weeks, 0.8-7 h a week; cycling 19 studies, walking and/or running 18, swimming 1, a choice of modes 11, unclear 1; frequency, intensity or session length progressed in 8, 10 and 10 studies.
Outcomes: VO2max (relative and absolute), peak power in a cycle test, VO2 at VT1, systolic and diastolic blood pressure, fasting total, HDL and LDL cholesterol, triglycerides and glucose. Peak power, total cholesterol and LDL were added after registration; economy was dropped (1 study).
Analysis: Hedges' g with 95% CI and 95% prediction intervals (SPSS 29); random effects; I² and Q; leave-one-out; exclusion of outliers (mean ± 2.68 SD), of high-risk studies and of studies with estimated maxima; subgroups by baseline VO2max, intensity (<50, 50-55, 55.1-60% of VO2max, HRmax converted with Howley's equation and heart-rate reserve taken as % of VO2max), session length (0-30, 31-50, >50 min), weekly volume (≤2.5, 2.6-4, >4 h), frequency, total dose (TRIMP), programme length (3-6, 7-11, ≥12 weeks), age, BMI and sex; studies whose programme progressed across subgroup bands left out of those splits; a meta-regression on intensity (added after registration) using studies whose intensity progressed by 5 points or less.
Risk of bias and certainty: RoB 2 by two reviewers; GRADE by two reviewers, with Egger's test for publication bias.
Funding and interests: the authors report nothing on funding and declare no conflicts of interest.
Key Results
| Outcome (low-intensity training vs control) | Groups | Change vs control (weighted mean ± SD) | Hedges' g (95% CI) | 95% prediction interval | GRADE |
|---|---|---|---|---|---|
| VO2max, mL/kg/min | 37 | +4.5 ± 2.4 | 0.94 (0.74 to 1.13), I² 0.73 | −0.07 to 1.94 | high |
| VO2max, mL/min | 7 | +390 ± 190 | 0.84 (0.36 to 1.33) | −0.66 to 2.35 | high |
| Peak cycle power, W | 9 | +35 ± 11 | 1.09 (0.86 to 1.31) | 0.82 to 1.36 | high |
| VT1 (VO2), pooled | 5 | +2.3 ± 0.9 mL/kg/min (4 groups) | 0.74 (0.26 to 1.22), I² 0.57 | −0.77 to 2.25 | moderate |
| Total cholesterol, mmol/L | 20 | −0.19 ± 0.29 | −0.29 (−0.50 to −0.09) | −1.03 to 0.45 | moderate |
| LDL cholesterol, mmol/L | 15 | −0.32 ± 0.31 | −0.42 (−0.62 to −0.22) | −1.00 to 0.16 | moderate |
| HDL cholesterol, mmol/L | 21 | +0.09 ± 0.10 | 0.32 (0.16 to 0.48) | −0.18 to 0.83 | low |
| Triglycerides, mmol/L | 19 | −0.09 ± 0.18 | −0.38 (−0.60 to −0.16) | −1.22 to 0.47 | low |
| Systolic blood pressure, mmHg | 17 | −4.0 ± 5.8 | −0.41 (−0.76 to −0.06) | −1.83 to 1.01 | low |
| Diastolic blood pressure, mmHg | 17 | −3.6 ± 4.0 | −0.43 (−0.68 to −0.18) | −1.34 to 0.48 | moderate |
| Glucose, mmol/L | 12 | −0.05 ± 0.29 | −0.18 (−0.66 to 0.30), not significant | −1.91 to 1.55 | low |
VO2max change: Table 1 gives +4.5 mL/kg/min against control; the Discussion gives a mean change of 3.9 mL/kg/min in the low-intensity groups (the two figures are not reconciled in the text).
Intensity: associated with the gain in relative VO2max (subgroup P = 0.021; meta-regression P = 0.01) and absolute VO2max (subgroup P = 0.03), and with no other outcome; the lowest intensity in the meta-regression was 45% of VO2max, and every intensity included gave positive effects against control.
Other training characteristics: several subgroup differences were significant, but no dose-response with session length, weekly volume, frequency, TRIMP or programme length; gains were comparable in programmes under 6 weeks and over 12 weeks (per-band numbers in supplementary tables we did not read).
Participant characteristics: age moderated the VO2max gain (P < 0.008): g 1.04 (0.79 to 1.30) at 45 years or under, 0.63 (0.45 to 0.80) at 45.1-65; baseline VO2max moderated the systolic pressure fall (P < 0.001): −1.43 (−2.19 to −0.67) below 30 mL/kg/min against −0.17 (−0.56 to 0.22) at 30-40; sex and BMI moderated nothing.
Sensitivity: removing one outlier each, HDL 0.27 (0.14 to 0.40) and LDL −0.36 (−0.54 to −0.19); excluding the 7 high-risk studies or the 13 with estimated maxima changed no outcome's size or significance (data not shown); leave-one-out changed nothing.
Risk of bias: 86% of studies some concerns, 14% high; reasons: randomisation not described, adherence unreported in 24 studies and below 80% in 3, no pre-registration, no blinding possible. Where reported, adherence was 66-100% (mean 93%).
Publication bias: Egger's test P = 0.01 for HDL and 0.04 for triglycerides.
Practical Application
Easy training alone builds an untrained adult's fitness: riding, walking or running kept below the first threshold raised VO2max by about 4.5 mL/kg/min over controls (g 0.94, high certainty) and peak cycling power by 35 W, over programmes averaging about 12 weeks (programmes under 6 weeks gained about as much). The 50 trials were in healthy, mostly sedentary adults training about 4 times a week for about 40 min; the spread between studies was wide (the prediction interval for VO2max runs from just below zero to large), so not every programme or person gains.
What it adds to [SYNTH-51]'s easy-volume section: the untrained "builds" cells there rested, before this review was added on 2026-11-10, on one trial, an older abstract-only analysis and two meta-analyses not limited to easy work — REF-1328 (B+, one trial at 50% of VO2peak in sedentary women), REF-1327 (C+, no floor in groups below about 40 mL/kg/min), REF-749 and REF-1330 (A-, moderate work and brisk walking).This review pools 50 RCTs restricted to the easy range and confirms the direction: for a beginner, easy minutes are training, and no floor was found down to about 45% of VO2max, in line with REF-1327 for less fit groups.It cannot speak to groups starting above 40 mL/kg/min (3 studies), so SYNTH-51's recreational and trained cells, its % of FTP translation and its recovery cut-offs are unchanged.
Where it differs from [SYNTH-51]: (1) the guide's default keeps easy work comfortably below the threshold because, in trained runners, sitting at the threshold bought nothing extra (REF-705, B+); in these untrained adults higher intensity within the easy range went with larger VO2max gains (meta-regression P = 0.01).That is a between-study association with intensities set as % of maxima, not a trial of placement, and in a population the guide does not use REF-705 for, so it does not overturn the default for trained athletes; for a beginner it supports the upper part of the easy range rather than the lower, and since 2026-11-10 the guide carries that as a beginner clause (still below the first threshold; a between-study association, not a tested placement).(2) The guide treats weekly minutes as the dose (REF-1328), since 2026-11-10 as a softer rule for an untrained beginner, and has the VO2max gain rising with session length up to about an hour (REF-749), now beside this review's finding of no dose-response; across these 50 trials session length, weekly hours (≤2.5 to >4 h), frequency and total dose showed no dose-response, and short and long programmes gained alike.On session length this disagrees with REF-749; both are comparisons between studies, and REF-749's continuous arm was not limited to work below the first threshold.Between-study subgroups with small cells are weaker on weekly minutes than REF-1328's randomised doses, so the minutes rule stands; what this review adds is that the first gains come fast and need progression to continue, and since 2026-11-10 the guide softens the weekly-minutes and session-length rules for an untrained beginner on this review's lack of a dose-response between studies: start with regular easy sessions at whatever length is sustainable and grow weekly time after the first weeks (practice).
Health markers: small falls in LDL (−0.32 mmol/L), total cholesterol, triglycerides and blood pressure (−4.0/−3.6 mmHg) and a small HDL rise (+0.09 mmol/L), at low to moderate certainty; glucose did not change significantly (g −0.18, −0.66 to 0.30). This is not treatment advice: blood pressure, lipids and glucose are managed with a physician.
Inference, not tested here: the base's practice of setting easy work from a measured threshold, heart rate at it or the talk test (SYNTH-51) is untouched — only one of these 50 trials set intensity from a threshold, so some participants may have trained above their own VT1.
Limitations
The authors': only one study set intensity from LT1/VT1, so physiological strain varied and some people may have trained in the heavy domain; high heterogeneity; adherence unreported in 24 studies, mean achieved intensity reported in 12 and the intensity distribution in none; VT1 reported by only 5 studies; VO2max mostly relative to body mass; no familiarisation tests reported, so learning effects may inflate the VO2max gains; day-to-day variation in lipids and glucose is large; subgroup cells were small, and some subgroup results looked contradictory (largest cholesterol change with the shortest sessions).
Certainty: GRADE high for VO2max and peak power, moderate for VT1, total and LDL cholesterol and diastolic pressure, low for HDL, triglycerides, glucose and systolic pressure; risk of bias some concerns in 86% of studies and high in 14%; possible publication bias for HDL and triglycerides.
Design (our reading): two databases only, English only, search ended October 2023; three outcomes and the meta-regression were added after registration; the intensity finding is between studies, after converting % of HRmax with a population equation.
Reporting: VO2max change given as 4.5 mL/kg/min against control in Table 1 and 3.9 in the Discussion; per-band subgroup estimates and GRADE reasons are in supplementary tables we did not read.
Transfer: untrained, healthy adults aged 18-65 (mean about 35); no trained athletes, so nothing here tests how easy training works for a fit rider or runner.
Revision 2026-11-10 (decision): the lines on SYNTH-51 now say what the guide carries from this review — a beginner clause for intensity and softer weekly-minutes and session-length rules for an untrained beginner — and that its untrained cells now cite it.
Revision 2026-11-10 (headline): the headline now says that the intensity, minimum-intensity and dose-response findings are comparisons between studies (subgroups and a meta-regression), not head-to-head trials. The body already said so (Methods, Limitations: Design, and the lines on SYNTH-51); the headline is the card the assistant is handed, and without these words an association could be passed on as a tested effect.
How this entry was checked and changed
Why this level
a meta-analysis of RCTs whose risk of bias and GRADE certainty we read in the full text (86% of studies some concerns, 14% high risk; GRADE high for VO2max and peak power); held at A- by serious heterogeneity (I² 0.73 for VO2max, prediction interval −0.07 to 1.94) and small groups (824 participants over 54 intervention groups, about 15 per group, our arithmetic); the VO2max, peak-power and VT1 results are capacity outcomes, so no marker step; the blood-lipid, glucose and blood-pressure results are markers and on their own would sit one step lower