Lifting and intervals mixed in one circuit session or split across separate days gave similar gains, with no significant group difference, in squat and bench 1RM, VO2peak and maximal aerobic power over 7 weeks at matched volume, in a preregistered RCT of 79 recreationally active adults (group × time p 0.38-0.76); the one difference, a slightly larger drop in body-fat percentage with separate days (−0.63 points, 95% CI −1.21 to −0.05, p = 0.035), was trivial in size and does not survive the authors' correction for 16 outcomes
Main caveat: The authors' own: diet, alcohol and energy expenditure not controlled; outside exercise only self-reported.
Sharp and colleagues asked whether doing aerobic and resistance exercise inside the same session changes adaptation compared with doing them on separate days when the total work is matched.
Seventy-nine recreationally active adults were randomised: one group did four 45-min circuits a week that alternated a resistance exercise (Romanian deadlift, row, goblet squat, chest press) with an aerobic station (bike, step, rower, shuttle run); the other did the same exercises as two 45-min resistance sessions and two 45-min aerobic sessions a week.
Intensity was set at RPE 7 or more, resistance sets at 10 reps, and session RPE did not differ between groups. After 7 weeks both groups gained small amounts of relative squat and bench 1RM, VO2peak and peak power in a ramp test, with no group × time difference on any of these. Both lost a trivial amount of fat and gained a trivial amount of lean mass.
Body-fat percentage fell slightly more with separate days (group × time b = −0.628 percentage points, p = 0.035, d = −0.082), which the authors call trivial and which falls short of their Bonferroni threshold (p < 0.003). The authors conclude that both formats work and that outcomes tracked total training exposure rather than format.
Participants
79 randomised (HIMT 41, ISCT 38); 75 analysed in the abstract's count (38 and 37); 31 men, 48 women; age 36.3 ± 11.9 years (HIMT 35.8 ± 11.7, ISCT 37.0 ± 12.5); BMI 25.0 ± 3.7.
Recreationally active by the criteria above; some had not done the specific exercises or intensity before (authors).
A priori sample size 47 (24 per group), raised to 58 for 20% dropout; the anticipated dropout did not occur.
12 were excluded from the squat 1RM analysis (8 switched to a 3RM for safety, 4 stopped the test), leaving 67.
Protocol
HIMT: 4 × 45-min sessions a week, resistance and aerobic exercises strictly alternated in a circuit.
ISCT: 2 × 45-min resistance and 2 × 45-min aerobic sessions a week, same exercises, intensity, sets, reps and work-to-rest ratios.
Both: short intervals of 30-60 s work and 15-30 s rest (authors' description); RPE ≥ 7 (CR-10), resistance sets of 10 reps at ≤ 3 reps in reserve (about 75% 1RM, by the authors), exercise selection and work-to-rest changed every two weeks; small groups of 3-8 per instructor; outside activity checked weekly by questionnaire.
Tests (fasted, 6-9 a.m., before and after): DEXA body composition; back-squat and bench-press 1RM; cycle ramp test from 50 W, +25 W per minute, to exhaustion, for VO2peak and peak power (Workmax).
Analysis: intention to treat, linear mixed models with age, sex, baseline and location as fixed effects; p < 0.05 with a Bonferroni threshold of p < 0.003 for 16 outcomes also reported; Cohen's d; a per-protocol sensitivity analysis (≥ 80% attendance) gave the same results.
Funding: Government Research Training Programme, university faculty funding and an in-kind donation; one author is affiliated with a non-university organisation, Performance Everyday (Sydney), whose role is not described, and the last author declares a perceived conflict as a former colleague of the research partner.
Key Results
| Outcome | HIMT before → after | ISCT before → after | Group × time (ISCT − HIMT), 95% CI, p |
|---|---|---|---|
| Squat 1RM, kg/kg | 0.99 → 1.11 (d 0.40) | 0.94 → 1.07 (d 0.45) | 0.015 (−0.027 to 0.056), p = 0.478 |
| Bench 1RM, kg/kg | 0.72 → 0.78 (d 0.20) | 0.61 → 0.64 (d 0.14) | −0.014 (−0.046 to 0.017), p = 0.375 |
| VO2peak, mL/kg/min | 34.13 → 36.77 (+2.5) | 33.84 → 36.48 (+2.7) | 0.207 (−1.139 to 1.554), p = 0.761 |
| Workmax, W | 218.8 → 242.8 (+21.0) | 214.7 → 237.1 (+23.0) | 2.03 (−4.90 to 9.00), p = 0.563 |
| Fat mass, kg | 20.99 → 20.73 (−0.41) | 22.78 → 22.05 (−0.85) | −0.447 (−0.968 to 0.074), p = 0.092 |
| Lean mass, kg | 47.62 → 48.98 (+0.75) | 47.5 → 48.18 (+0.74) | −0.008 (−0.5 to 0.484), p = 0.974 |
| Body fat, % | 29.49 → 28.78 (−0.57) | 30.82 → 29.75 (−1.20) | −0.628 (−1.21 to −0.046), p = 0.035, d −0.082 |
Within-group changes in the table are model estimates (in brackets) or raw means; all listed within-group changes were significant at p < 0.05, squat, VO2peak and Workmax effects small (d 0.30-0.45), bench effects trivial by the authors (d 0.14-0.20), the body-composition effects trivial (d under 0.16).
No within- or between-group change in maximal heart rate, body mass, BMI, hip circumference or bone mass.
Waist circumference and waist-to-hip ratio fell in both groups with no group difference.
Practical Application
For a recreationally active adult, a combined lifting-and-intervals circuit four times a week built strength and aerobic fitness over 7 weeks with no significant difference from the same work split into separate lifting and cardio days. No group × time difference reached significance for squat or bench 1RM, VO2peak or peak power; the intervals were short (30-60 s), sets were 10 reps and power was not tested, so this does not test SYNTH-21's point that the explosive end is what same-session training costs.
The body-fat edge for separate days is not a reason to choose them. The difference was 0.63 percentage points over 7 weeks, trivial by the authors (d −0.08), with an interval reaching near zero (−0.05), not significant after their correction for 16 outcomes, and diet was not controlled.
Not settled by this design: whether the same holds for trained athletes, heavier lifting, longer aerobic sessions, longer programmes, or for jump and sprint power. The authors note that some participants were new to the exercises, which may have hidden a small interference effect.
Inference, not tested here: for an endurance athlete short of time, a mixed circuit can stand in for a separate gym session in a general-fitness phase without a detectable strength cost; it does not replace the athlete's aerobic sessions (SYNTH-107: a circuit is added to the running, not swapped for it).
Limitations
The authors' own: diet, alcohol and energy expenditure not controlled; outside exercise only self-reported; 1RM testing is specific and training was not near 1RM; one strict circuit format, so findings may not transfer to other multimodal styles or to how people usually do concurrent training; 7 weeks in recreationally active adults; with 16 outcomes, the fat-percentage interaction and several within-group body-composition changes miss the Bonferroni threshold.
Ours: not blinded; 12 of 79 missing from the squat analysis; the abstract's analysed numbers (38 and 37) differ from the 41 and 38 randomised and the flow diagram was not available as text; Table 4 marks the body-fat interaction with a footnote "p < 0.001" that contradicts the p = 0.035 printed beside it.
No power or jump outcome, which is where pooled interference shows (REF-330); the endurance part was short intervals, not long steady work.
Interests: an in-kind donation whose donor is not named, a co-author affiliated with a non-university organisation (Performance Everyday), and a declared perceived conflict of the last author as a former colleague of the research partner, who states no longer being associated with that organisation and holding no financial or non-financial interests; the partner is not named and its role is not described.
Attendance and delivered training are reported in a separate paper not read here.
How this entry was checked and changed
Why this level
an RCT read in full with about 38-40 per arm (79 randomised, 41 and 38), the rubric's ≥20 per arm; not A because under 50 per arm; outcomes are strength, capacity and body composition, not markers