Do supersets save gym time without losing gains?
Time, yes: about 37% shorter, no clear volume loss, but harder. Gains: no clear difference, on 3 small trials at 50-80% 1RM, wide CIs. No heavy chronic trial, no endurance athletes.
Main caveat: Chronic evidence is thin. Three trials, three different superset types, different loads and rest. CIs include a clinically real difference either way.
Zhang and colleagues compared supersets (two exercises in sequence with little or no rest, then a rest) with traditional sets, in 19 studies and 313 participants; 16 were acute, 3 chronic.
Pooled, supersets gave similar repetitions and volume load, higher training efficiency, higher blood lactate, energy cost and RPE, with similar creatine kinase, surface EMG, acute swelling, blood pressure and perceived recovery. Chronic outcomes showed no clear difference from traditional sets.
By type, agonist-antagonist pairs allowed more repetitions, and similar-muscle pairs produced less volume load. The authors conclude that supersets are a time-efficient alternative for people with limited time, and that the higher internal load may need longer recovery.
Our reading: the acute numbers support "no clear difference in volume, in less time, and it feels harder"; the intervals (about ±0.5 SMD, prediction intervals about ±1.7) and the volume lost in same-muscle pairs do not show equal volume. Those are acute responses, not training adaptation.
The adaptation claim rests on three small trials of different superset types, with very wide intervals. That justifies supersets as a reasonable tool for a short-of-time lifter, but does not show they are equal for strength.
Protocol
Search: PubMed, Web of Science, Embase, EBSCO, inception to 10 February 2024; 5881 records, 108 assessed, 19 included.
Inclusion: healthy individuals; superset structure vs traditional sets; plyometric-plus-resistance complexes excluded; at least two sets of each exercise.
Studies and participants: n = 313 (232 men, 66 women, 15 sex not reported). Three studies excluded from the pooling for unbalanced training.
Types: agonist-antagonist (bench and row); similar biomechanical (bench and dumbbell press); alternate peripheral (squat and bench).
Chronic trials (three): Garcia-Orea (17 men, alternate peripheral, 3 sets to 15% or 20% velocity loss, 55-70% 1RM, squat and bench, 45 s within a pair and 2 min between pairs vs 3 min); Fink (23 athletes from a university gymnastics club, agonist-antagonist, 3 sets to failure, 50-60% 1RM, arm exercises, 8 weeks, 1 min rests); Merrigan (32 recreationally active women, similar biomechanical, 3-4 sets of 8-12, 70-80% 1RM, squat and leg press, 2.3-2.5 min vs 1 min between sets).
Analysis: Hedges' g SMD, 95% CI and prediction intervals; a pre-post correlation of 0.5 assumed; I² ≥ 75% high; GRADE.
Funding and interests: no external funding; open-access publishing by Universidad de Granada/CBUA; no conflicts of interest.
Key Results
| Outcome | k | SMD (95% CI) | p | I² | GRADE |
|---|---|---|---|---|---|
| Total repetitions | 6 | −0.03 (−0.56 to 0.51) | 0.92 | 66.9% | High |
| Volume load | 8 | 0.05 (−0.48 to 0.57) | 0.86 | 78.4% | Moderate |
| Training efficiency | 2 | 1.74 (0.46 to 3.01) | 0.01 | 73.0% | Moderate |
| Lactate during | 4 | 0.94 (0.08 to 1.81) | 0.03 | 72.3% | Moderate |
| Lactate after | 7 | 1.13 (0.42 to 1.84) | <0.01 | 79.6% | Moderate |
| Energy cost | 2 | 1.93 (0.08 to 3.78) | 0.04 | 83.3% | Very low |
| Creatine kinase after | 2 | 0.22 (−0.25 to 0.68) | 0.36 | 0% | Moderate |
| RPE | 6 | 0.77 (0.15 to 1.40) | 0.02 | 74.8% | High |
| Perceived recovery | 2 | 0.32 (−0.32 to 0.95) | 0.33 | 11.6% | Moderate |
| Chronic: 1RM | 3 | 0.10 (−0.40 to 0.60), PI −3.14 to 3.35 | 0.36 | 0% | Moderate |
| Chronic: maximal repetitions | 2 | 0.07 (−0.51 to 0.68) | 0.81 | 0% | Moderate |
| Chronic: muscle area | 2 | −0.05 (−0.63 to 0.54) | 0.87 | 0% | Moderate |
| Agonist-antagonist: repetitions | 2 | 0.68 (0.20 to 1.17) | 0.01 | 0% | Low |
| Similar biomechanical: volume load | 2 | −1.08 (−1.72 to −0.44) | <0.01 | 0% | Moderate |
Time: about 37% shorter sessions, from the discussion; it is given as a figure, not as a pooled result with an interval.
Rest in the chronic trials was not always shorter. In Merrigan the superset arm rested longer between sets than the traditional arm (see Protocol), so that trial tests structure more than time saved (our reading).
Session-RPE load (one study, as the authors report it): Weakley found supersets rated harder per session yet lower in session perceived load, because the session was shorter; a duration-times-RPE measure may hide how hard they were.
Mechanisms
Measured: higher lactate and energy cost with less rest, and similar EMG and creatine kinase on pooled data.
Authors' interpretation: shorter recovery raises internal load, so later sessions may need more recovery; antagonist preloading may help repetitions; the 16 studies with different muscle groups preserved volume by avoiding local fatigue. They call the recovery advice sensible and did not test it.
Inference: a logged "same load, same reps" in a superset is harder work than the same numbers done with full rest, so a superset session's numbers should not be compared directly with an earlier straight-set session's.
Practical Application
Do supersets save gym time without losing gains? Time, yes: about 37% shorter, no clear volume loss, but harder. Gains: no clear difference, on 3 small trials at 50-80% 1RM, wide CIs. No heavy chronic trial, no endurance athletes.
Which pairs: agonist-antagonist pairs allowed more repetitions (SMD 0.68, k = 2, low certainty); upper-with-lower pairs showed no significant difference in repetitions (SMD −0.46, 95% CI −1.08 to 0.15, k = 3); same-muscle pairs lost volume load (SMD −1.08, k = 2).
Acute is not adaptation: the lactate, energy-cost and RPE results describe a session. Only 1RM, maximal repetitions and muscle area speak of adaptation; the review states 8 weeks for one chronic trial (Fink) and does not tabulate the length of the others.
Before a key session: Not from this paper; the base's own practice: treat it as a hard gym day when spacing it from a quality session (SYNTH-31), since the review found higher internal load and did not test what follows.
Reading the log: compare a superset session with the last superset session, and watch RPE (inference from the acute findings).
Does not license short rests on heavy lifts: rests between supersets ran 1 to 3 min and the chronic loads 50-80% 1RM; the one heavy protocol (Gahreman: squat and deadlift triples at 90% 1RM, 2.5 min rests, 15 adolescents) was acute. The authors say load still needs care when the aim is maximal strength, and suggest about 55% 1RM may explain Fink's lack of a significant 1RM gain. It tests no progression rule.
Limitations
Chronic evidence is thin. Three trials, three different superset types, different loads and rest. CIs include a clinically real difference either way; prediction interval for 1RM −3.14 to 3.35. Their "moderate" GRADE here is generous.
Internal conflict. The conclusion and Key Points state "more severe muscle damage" with supersets, while the pooled results show no significant difference in creatine kinase (SMD 0.22, p = 0.36) or swelling (SMD −0.28, p = 0.36); only one study (Weakley) found higher creatine kinase.
Numbers that do not fit. The chronic 1RM entry gives SMD 0.10 (−0.40 to 0.60) with p = 0.36; that interval implies p of about 0.7 (our arithmetic). The conclusion is the same but one number is probably mistaken.
Study pool. Mostly young trained men; 66 women; order effect unclear in most studies, no allocation concealment or outcome blinding reported, no protocol preregistered by the primary studies; publication bias not assessed.
Assumed correlation. A 0.5 pre-post correlation was assumed for chronic effects.
Interests. No external funding, none declared.
How this entry was checked and changed
Why this level
a meta-analysis of small studies, often with high heterogeneity; the chronic result, the one that matters for a time-short athlete, rests on 3 trials and about 72 participants (our sum) with intervals of about ±0.5 SMD, so it carries less than the level suggests. Read from the full text (open access, CC BY)
Full text (checked 2026-10-30)
Nineteen studies, 313 participants; 16 acute only, three chronic (full text)
Training time reduced by approximately 37% (full text)
Chronic: 1 RM SMD = 0.10 (95% CI − 0.40 to 0.60), PI − 3.14 to 3.35; MNR 0.07 (− 0.51 to 0.68); muscle CSA − 0.05 (− 0.63 to 0.54) (full text)
Total repetitions − 0.03 (− 0.56 to 0.51), volume load 0.05 (− 0.48 to 0.57); training efficiency 1.74 (0.46 to 3.01) (full text)
Lactate after 1.13 (0.42 to 1.84); RPE 0.77 (0.15 to 1.40); energy cost 1.93 (0.08 to 3.78) (full text)
Agonist-antagonist repetitions 0.68 (0.20 to 1.17); similar biomechanical volume load − 1.08 (− 1.72 to − 0.44) (full text)
Creatine kinase 0.22 (− 0.25 to 0.68), p = 0.36, though the authors write of more severe muscle damage (full text)
No external sources of funding; no conflict of interest (full text)