REF-1679 · Zheng, Gao & Song, 2026 · BMC Sports Science, Medicine and Rehabilitation, · Paper · PubMed · Free full text
Adding a heavy or explosive conditioning activity (squats, flywheel, isometric squats, drop jumps) to a general warm-up raised countermovement jump height by 2.41 cm (95% CI 1.20 to 3.61; 8 studies, moderate certainty) in trained and competitive athletes, against the general warm-up alone; change of direction also improved (SMD −0.75, 3 studies, low certainty), but sprint did not (SMD 0.29, −0.35 to 0.94; 3 studies, low certainty). A meta-analysis of randomised acute trials; 12 of 33 included studies could be pooled. Read in full
Main caveat: The authors': few studies in the sprint and change-of-direction analyses; evidence uneven across outcomes, so task-specific effects are tentative.
Zheng, Gao and Song asked a narrower question than earlier potentiation reviews: does a PAPE conditioning activity, put after a general warm-up, improve jump, sprint and change-of-direction performance beyond the same general warm-up alone, in trained and competitive athletes?
They searched five databases to March 2026 and kept only randomised acute trials with that comparison; 33 studies qualified and 12 had data that could be pooled (one effect per study, chosen by prespecified rules).
The conditioning activities were heavy or loaded squats, flywheel half-squats, isometric squats, drop or depth jumps, hops and sport-specific efforts, with testing typically 3-10 min afterwards.
The primary analysis, countermovement jump height in 8 studies, improved by 2.41 cm, a result that held after removing two influential studies (1.95 cm) and in a standardised analysis (SMD 0.52). All jump outcomes together improved by SMD 0.42. Change of direction improved (SMD −0.75) in only three studies, and sprint did not change significantly (SMD 0.29) in three studies.
GRADE rated jumps moderate and sprint and change of direction low. The authors conclude that the jump benefit is consistent, the change-of-direction benefit preliminary and the sprint benefit unsupported, and they advise finding each athlete's recovery window by testing.
Participants
Pooled studies (Table 1): 12 studies, 202 participants by our count (15 elite youth basketball players; 14 highly trained volleyball players; 11 elite female volleyball players; 17 collegiate female soccer players; 10 young professional soccer players; 27 young taekwondo athletes; 12 semi-professional basketball players; 14 semi-professional female volleyball players; 28 female collegiate basketball players; 10 top-level team-sport athletes; 32 national-level sprinters; 12 academy soccer players).Mean ages from 16 to 28 years.
Narrative studies: 21 further studies whose populations are listed in a supplement we did not read.
Protocol
Search: five databases from inception to 14 March 2026; 1,616 records, 681 screened after removing 935 duplicates, 182 full texts, 33 included (149 excluded, 101 of them for a comparator or warm-up that did not meet the criteria).
Inclusion: healthy athletes, trained or competitive, with no restriction by sex, sport or level; an acute PAPE conditioning activity after a general warm-up; comparator the general warm-up alone; randomised crossover, repeated-measures or parallel-group designs.
Conditioning activities in the pooled studies: flywheel half-squats (3 × 6), isometric half-back squats (1 or 3 sets of 3 s), back squats at 60-100% 1RM (single sets to 3 × 3), depth jumps (3-12 contacts), bilateral or unilateral drop jumps (3 × 5), stiff-legged hops (3 × 10), deadlift 5RM, tuck jumps, isometric knee extension, plyometric or kicking efforts; one study combined squats with caffeine or placebo.
Outcomes: countermovement jump (CMJ) height as the primary outcome; other jumps, sprints of 10-50 m and change-of-direction tests as secondary outcomes.
Analysis: mean difference in cm for the CMJ-only analysis; SMD for all-jump, sprint and change-of-direction analyses; fixed-effect when I² < 50% and χ² p > 0.10, otherwise random-effects; leave-two-out and SMD sensitivity analyses for CMJ, leave-one-out for all jumps; Egger's test for the all-jump analysis (k = 10); no subgroup analysis or meta-regression (too few studies).
Risk of bias and certainty: seven-domain Cochrane tool by two reviewers; GRADE.
Funding and interests: no specific funding; no competing interests declared.
Key Results
| Analysis (PAPE + warm-up vs warm-up alone) | Studies | Participants (exp/control) | Effect (95% CI) | p | I² | GRADE |
|---|---|---|---|---|---|---|
| CMJ height, primary (fixed-effect) | 8 | 110/86 | MD 2.41 cm (1.20 to 3.61) | < 0.0001 | 26% | moderate |
| CMJ, two influential studies removed | 6 | 85/68 | MD 1.95 cm (0.69 to 3.21) | 0.003 | 0% | — |
| CMJ, standardised (random-effects) | 8 | 110/86 | SMD 0.52 (0.23 to 0.82) | 0.0005 | 0% | — |
| All jump outcomes | 10 | 139/115 | SMD 0.42 (0.17 to 0.68) | 0.001 | 0% | moderate |
| Sprint | 3 | 41/35 | SMD 0.29 (−0.35 to 0.94), not significant | 0.37 | 49% | low |
| Change of direction (lower time = better) | 3 | 47/47 | SMD −0.75 (−1.17 to −0.33) | 0.0005 | 0% | low |
Sensitivity: leaving out any single study in the all-jump analysis gave SMD 0.36 to 0.49, all intervals above zero.
Small-study effects: all-jump funnel plot not markedly asymmetric; Egger intercept 3.31 (SE 1.51), p = 0.060, which the authors call borderline.
Risk of bias (no counts in the text): random sequence often adequately reported; outcome assessment mostly low risk (objective devices); missing data rarely a problem; allocation concealment usually unclear; participant and personnel blinding frequently high or unclear risk.
Timing: in the pooled studies the post-activity test came mostly about 3-10 min after the conditioning activity, in some up to 12 min; the authors did not test timing statistically.
Reporting slips (our reading): the two studies removed in the sensitivity analysis carry different reference numbers in the text, the figure legend and Table 1.
Practical Application
A heavy or explosive set added to the general warm-up raised countermovement jump height by about 2.4 cm in trained and competitive athletes (95% CI 1.20-3.61 cm; moderate certainty), tested mostly 3-10 min afterwards. The effect is acute and was measured on jumps.
It did not clearly speed up sprints (SMD 0.29, −0.35 to 0.94; three studies, low certainty), so a jump gain should not be assumed to carry over to a sprint start or a cycling sprint. Change-of-direction tests improved in three studies (low certainty).
Where this differs from the base: SYNTH-66's strength and gym row says a conditioning activity before explosive work gave a small boost to jumps and a moderate one to sprints (REF-688, an earlier meta-analysis read from its abstract, whose comparator the base does not record); this review, restricted to the activity added to a general warm-up against that warm-up alone, found no significant sprint benefit (three studies, low certainty, an interval that does not exclude a benefit), so the sprint half of that line is not confirmed for this comparison.The jump result agrees in direction with REF-688.
Inference, not tested here: for a cyclist or runner, the guide's below-threshold priming bout and sub-maximal openers (SYNTH-66) are a different tool; this review tested squats, flywheel work and plyometrics before jumps and short sprints, not before endurance efforts, and gives no reason to add heavy lifting to a race warm-up.
The authors' advice, untested as a rule: choose the conditioning activity by the athlete's strength and fatigue tolerance and find the recovery gap by repeated testing rather than a fixed number of minutes.
Limitations
The authors': few studies in the sprint and change-of-direction analyses; evidence uneven across outcomes, so task-specific effects are tentative; conditioning activities, athletes, devices, scoring and timing varied; one-effect-per-study rules left out other time points and arms; no moderator analysis; incomplete reporting of allocation concealment and protocols; publication bias not excluded for jumps.
21 of 33 included studies could not be pooled; the pooled 12 hold about 200 athletes, mostly from team and court sports.
The primary CMJ model was fixed-effect; one pooled study gave the conditioning activity with caffeine or placebo, a co-intervention.
Crossover and parallel designs were pooled as group means; participant numbers per arm in Table 2 appear to count crossover participants in both arms (our reading).
We did not read the supplements (search strings, GRADE profile, effect-selection rules, the list of narrative-only studies).
How this entry was checked and changed
Why this level
a meta-analysis of randomised trials whose risk of bias and GRADE certainty we read in the full text (moderate certainty for jumps, low for sprint and change of direction), held at A- because it rests on few small studies (8 for the primary jump result, 3 each for sprint and change of direction), allocation concealment was usually unclear and participants could not be blinded; jump height, sprint and change-of-direction times are performance outcomes, so no marker step