Revisiting the stretch-induced force deficit: A systematic review with multilevel meta-analysis of acute effects

Highlights

• •In literature and practice, static stretching is described to harm subsequent maximal strength and speed performance parameters. As previous systematic reviews were performed more than 10 years ago and partially included studies with uncontrolled designs and/or lacking pre-post comparisons, this systematic review with meta-analysis was performed to update the evidence.
• •While the results confirmed that stretching performed >60 s acutely impairs maximal strength when tested in isolation, neither complex speed performance tasks such as jumping or sprinting, nor explosive strength/rate of force development showed significant reductions.
• •Therefore, it was summarized that researchers and practitioners should avoid extensive stretching before testing or performing isolated maximal strength tests, while a rigorous prohibition of including stretching to (dynamic) warm-up routines seems without evidence.

Abstract

Background

When recommending avoidance of static stretching prior to athletic performance, authors and practitioners commonly refer to available systematic reviews. However, effect sizes (ES) in previous reviews were extracted in major part from studies lacking control conditions and/or pre–post testing designs. Also, currently available reviews conducted calculations without accounting for multiple study outcomes, with ES: –0.03 to 0.10, which would commonly be classified as trivial.

Methods

Since new meta-analytical software and controlled research articles have appeared since 2013, we revisited the available literatures and performed a multilevel meta-analysis using robust variance estimation of controlled pre–post trials to provide updated evidence. Furthermore, previous research described reduced electromyography activity—also attributable to fatiguing training routines—as being responsible for decreased subsequent performance. The second part of this study opposed stretching and alternative interventions sufficient to induce general fatigue to examine whether static stretching induces higher performance losses compared to other exercise routines.

Results

Including 83 studies with more than 400 ES from 2012 participants, our results indicate a significant, small ES for a static stretch-induced maximal strength loss (ES = –0.21, p = 0.003), with high magnitude ES (ES = –0.84, p = 0.004) for stretching durations ≥60 s per bout when compared to passive controls. When opposed to active controls, the maximal strength loss ranges between ES: –0.17 to –0.28, p < 0.001 and 0.040 with mostly no to small heterogeneity. However, stretching did not negatively influence athletic performance in general (when compared to both passive and active controls); in fact, a positive effect on subsequent jumping performance (ES = 0.15, p = 0.006) was found in adults.

Conclusion

Regarding strength testing of isolated muscles (e.g., leg extensions or calf raises), our results confirm previous findings. Nevertheless, since no (or even positive) effects could be found for athletic performance, our results do not support previous recommendations to exclude static stretching from warm-up routines prior to, for example, jumping or sprinting.

Graphical abstract

1. Introduction

For centuries, the use of stretching has been recommended as an essential component of numerous warm-up programs to acutely enhance range of motion (ROM).1, 2, 3 However, around the new millennium, it was found that stretching may induce harmful acute effects on strength-dominant performance capacities.4, 5, 6, 7, 8 Since then, we have seen a plethora of studies discussing deteriorations in maximal strength and speed performance in response to static stretching9, 10, 11, 12, 13, 14 as well as others that have been unable to confirm these reductions.15, 16, 17, 18, 19 Consequently, systematic reviews with meta-analyses have been performed—some including more than 120 acute stretching studies—with the aim of exploring the “stretch-induced force deficit”. While in 2004, Shrier²⁰ found no acute performance changes, with an effect size (ES) of –0.1¹³ and strength losses of –4.2% to 61%,¹⁴ the review articles that followed concluded that static stretching should be avoided in movement preparation, especially for periods >60 s^9,10,14,21 and/or prior to high-intensity workouts.²² In contrast to Simic et al.,¹³ who found non-significant or trivial explosive strength deteriorations post stretch (ES up to –0.04 with a standard error up to 0.06), Li et al.²³ included 35 studies and showed static stretching to significantly diminish explosive strength performance compared to other stretching types, foam rolling, or combined alternatives. High awareness of these results is reflected by a questionnaire study showing that 85% of practitioners who were asked responded that they believe static stretching would negatively affect their subsequent maximal and speed strength performance.²⁴

Underlying mechanisms of the stretch-induced force deficit were summarized by several review articles. The most popular explanatory approaches were extensively discussed by neuromuscular experts^{2,5,7,10,11,21,25} or specialized biomechanics labs from Australia.^14,26 Accordingly, stretching seems to acutely reduce electromyography (EMG)-activity (muscle action potential wave (M-Wave) amplitude), indicating diminished central nervous innervation patterns. Takeuchi et al.²⁷ showed acute muscle–tendon unit stiffness reduction, which could harm tendon energy storage and, in turn, reduce stretch-shortening cycle output by decreasing energy storage and releasing capability.^28,29 In contrast, more frequently suggested stretch-specific effects, such as H-reflexes, exteroceptive (E-) reflexes, muscle spindle- or Golgi-tendon reflexes, remain significantly unchanged or were shown to dissipate within seconds and so were classified as unlikely to affect subsequent performance.²¹

However, it is noteworthy that the currently available systematic reviews with meta-analyses^13,14 pooled study effects extracted from uncontrolled (e.g., Refs.^25,30, 31, 32, 33, 34, 35, 36, 37) studies and from those lacking pre–post comparisons (e.g., Refs.^15,16,38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52). Moreover, while Simic et al.¹³ highlighted how previous literature reviews did not use an appropriate statistical tool to quantify the effects, the authors neither accounted for multiple study outcomes by using multilevel calculation models nor considered unknown origins of variance by using robust variance estimation calculation models.⁵³ Of course, drawing conclusions based on uncontrolled study results seems to be of questionable validity; thus, a revisitation of the most recent evidence using controlled study designs is called for.

Even when assuming diminished acute stretch-induced subsequent performance, the stated underlying mechanisms do not appear to be stretch-specific. First, strength and performance reductions^32,54 as well as diminished EMG (i.e., M-wave amplitude and duration^21,55,56) do not appear to be exclusively attributable to static stretching, as these are also evident in response to several other activities with the potential to induce (among other effects) fatigue.57, 58, 59 Second, stiffness reductions can also be induced via several activities that are sufficient to enhance body and muscle temperature and thus are not stretch-specific.²⁴ Consequently, by revisiting the current acute stretching literatures, we hypothesized a static stretch-induced force and performance deficit (>60 s), while one could speculate that alternative routines sufficient to decrease EMG-activity and stiffness could produce similar strength and speed performance losses.

To counteract the methodological issues listed above, we used the robust variance estimation meta-analysis model to account for multiple study outcomes with unknown origin of variance. First, we limited study inclusion to controlled pre–post comparison studies to explore the current state of the stretch-induced force deficit literatures with a reduced bias. Second, ES for direct comparison studies were pooled to oppose stretching with alternative activities and determine whether or not possible acute reductions in performance are indeed stretch-specific.

2. Methods

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines⁶⁰ were considered to perform this systematic review with meta-analysis. The search term was designed under consideration of the Patient/population, Intervention, Comparison, Outcomes (PICO) guidelines⁶¹ and applied to MEDLINE/PubMed, Web of Science, and Scopus (inception to December 2023), which was supplemented by a manual search of the first 500 Google Scholar results. To counteract the aforementioned limitations of previous reviews, the following eligibility criteria were applied for study inclusion: (a) healthy participants; (b) any type of stretching, including any type of static, dynamic/ballistic, and proprioceptive neuromuscular facilitation stretching; (c) acute effects on either maximal strength, jumping, sprinting, or throwing performance; (d) non-intervened controlled studies or direct comparisons to other stretching or non-stretching alternatives, such as strength training or cycling; and (e) pre–post study design. Studies that (a) were uncontrolled, (b) lacked pre–post testing, (c) combined stretching with other interventions, or (d) did not use stretching interventions were excluded.

The search terms were created based on the requirements of each database. The literature search in PubMed (search terms for Web of Science and Scopus in the Supplementary Material) was conducted using the following term:

(stretch* AND acute) AND (strength OR force OR torque OR jump OR throw OR performance OR explosive OR speed OR “force deficit”).

The systematic, term-guided literature search was supplemented by carefully reviewing the list of previous systematic^13,14,23 and narrative works^9,10,21,26 as well as the reference lists of the found literatures.

To distinguish between different types of stretching, the following definitions were used. Static stretching involves lengthening a muscle until reaching a stretch sensation/the point of discomfort and holding the muscle in a lengthened position; this can be performed passively by a tool, partner, or external weight, or actively with functional movements.² Proprioceptive neuromuscular facilitation stretching incorporates a maximal voluntary contraction into a static stretching bout with or without antagonist contraction.² In accordance with Warneke et al.,²⁴ dynamic stretching was defined as controlled back and forth movements in the end ROM, while ballistic stretching was assumed to be a subcategory of dynamic stretching including less controlled, bouncing movements in the end ROM. Other articles used a less restrictive definition and considered full ROM movements to be dynamic stretching; as a consequence, full ROM resistance training would be considered dynamic stretching. We exclusively included studies using the Warneke et al.²⁴ definition as well as those studies that described their intervention as dynamic stretching, including movements such as backward or forward lunges,62, 63, 64, 65 front kicks,⁶² frontal/backward and side leg swings,^63,64 butt kicks,^65,66 or kicking actions.⁶⁷

2.1. Methodological study quality and risk of bias

Risk of bias assessment was performed by both investigators (KW and LHL) using the Physiotherapy Evidence Database (PEDro) scale.^68,69 Furthermore, study quality assessment was extended by exploring the publication bias via visual inspection of modified funnel plots that account for dependency of multiple study outcomes,^70,71 with result quantification via the Egger's test. To classify the certainty of evidence, the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) working group criteria⁷² were applied, categorizing effects as “very low” (effect estimate very uncertain), “low” (further research is very likely to change the effect estimate), “moderate” (further research is likely to change the effect estimate), or “high” (further research is very unlikely to change the effect estimate). Certainty would be downgraded due to risk of bias, inconsistency, uncertainty of directness, imprecise data, or reporting bias and upgraded if studies showed strong evidence of association, evidence of a dose–response gradient, and plausible confounders.

2.2. Data processing and statistics

Both authors (KW and LHL) performed the data extraction. If no means and standard deviations (SDs) were provided in the full text, the corresponding author of the respective study was contacted via e-mail or ResearchGate to request the original data. If available, data were imputed from graphical illustrations. If the corresponding author did not respond and there was no other possibility of determining exact mean and SD, the study was excluded. We calculated ES based on mean difference from pre- to post-test by applying:

$$\text{Mea}{\mathrm{n}}_{\text{post}-\text{test}}-\text{Mea}{\mathrm{n}}_{\text{pre}-\text{test}}$$

pooled SDs were determined by

$$\mathrm{S}{\mathrm{D}}_{\text{pooled}}=\sqrt{\frac{\left(n_1-1\right)\times {\text{SD}}_1^2+\left(n_2-1\right)\times {\text{SD}}_2^2}{\left(n_1-1\right)+\left(n_2-1\right)}}.$$

As recommended by Tipton,⁷³ the robust variance estimation meta-analysis calculation model was used to pool the standardized mean differences and 95% confidence intervals (95%CIs) of acute stretching effects on maximal strength outcomes, speed strength (explosive strength, rate of force development (RFD)), and athletic performance parameters such as jump, sprint, or throwing performance. Furthermore, a separate analysis was performed that only included studies which directly compared the effects of stretching to those of alternative training routines on strength-dominant performance or athletic performance. While between-studies-within-cluster variance was reported as intracluster heterogeneity via Ω², between-study (intercluster) heterogeneity (τ²) was minimized by separately analyzing subgroups for outcome parameters as well as type of stretching. Stretching times per bout and overall volume were also analyzed separately for above and below a 60-s cut-off, which is frequently reported as a border.^9,14,21 Searching for an evidence-based subgroup cut-off for a practically relevant stretching duration, we were not able to pinpoint a generally accepted guideline. However, some literatures74, 75, 76 have suggested a stretching volume of 480 s to be relevant in practice. Furthermore, as different effects might occur in children (<18 years) and adults (≥18 years), age-related subgroup analyses were performed. If 2 studies or fewer were available for the respective analysis, we performed a sensitivity analysis for each of the investigated moderators (stretching time per bout, stretching volume per session with 60-s and 480-s cut-offs, as well as age groups). While for maximal strength we considered isometric and dynamic force measurements, speed performance included athletic performance such as sprinting, jumping, or throwing as well as explosive strength/RFD. Pooled ES were interpreted as follows: trivial ES: <0.2; small ES: 0.2–0.49; moderate ES: >0.49–0.79; and large ES: >0.79.⁷⁷ All calculations were performed using R with the robumeta package^53,73 under special consideration of the study design (parallel and cross-over design).

3. Results

The overall literature search resulted in the inclusion of 83 studies (Fig. 1).4, 5, 6^,8,19,54,62, 63, 64, 65, 66, 67^,74, 75, 76^,78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145 Controlled pre–post comparisons from 37 studies showed acute stretching effects on maximal strength performance,4, 5, 6^,8,54,67,74, 75, 76^,78,81,87, 88, 90, 91, 92, 93^{, 103,104,}106, 107, 108^{,111,112,118,119,121,123,125,127,}132, 133, 134, 135, 136^,141,142 while 22 studies investigated stretching effects on speed parameters such as RFD, power, or maximal jumping/sprinting/throwing performance as compared to a non-intervened control condition.^66,75,76,78, 79, 80, 81^{,84,86,96,104,111,115,117,}120, 121, 122^,126,138, 139, 140^,144 Additionally, 31 studies compared acute static stretching effects to dynamic stretching, proprioceptive neuromuscular facilitation,^{54,62,65,66,78,80,85,89,94,95,}97, 98, 99, 100, 101^,105,107,110, 112, 113^,116, 118, 119, 120^{,124,125,127,130,131,137,140} or alternative training routines, including resistance training,⁵⁴ foam rolling with and without vibration,^110,113,124 and aerobic activity^95,116 (see study characteristics in Supplementary Table 1).

Fig. 1.

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart of the systematic literature search.

Of the 83 studies, 11 included females only,^66,74,87,90, 91, 92^{,95,108,121,132,133} 40 males only,^{4,5,65,67,75, 79,80,82,88,94,}96, 97, 98, 99, 100^{,102,103,105,}110, 111, 112, 113, 114^,116, 117, 118, 119^,123, 124, 125, 126, 127^,129, 130, 131^{,134, 135,140,143,145} and the remaining 32 studies^6,8,19,54,62, 63, 64^{,76,78, 81,}83, 84, 85, 86^{,89,93,101,104,106,107,109,115,120,122,128,}136, 137, 138, 139^,141,142,144 included a mixed population (Supplementary Table 1). Most of the research (56 out of 83 studies) recruited untrained and/or recreationally active participants,^4,5,8,19,54,62, 63, 64^,66,74, 75, 76^,80, 81, 82, 83, 84^,86, 89, 90, 91, 92^,94,95,98,100, 101, 102, 103, 104, 105, 106, 107^{,109,111,113,114,116,}119, 120, 121, 122, 123, 124^,126,127, 128, 129^,134,135, 138, 139, 140, 141, 142, 143 while only 18 studies investigated acute stretching effects in trained participants.^{65,67,78,79,85,87,88,96,97,99,112,118,}130, 131, 132, 133^{, 137,145} The remaining studies either did not state any information on subjects’ physical activity backgrounds,^{93,108,110,136} investigated school children from physical education classes,¹¹⁵ or included subjects with both recreational and trained^6,116 or sedentary and trained sport backgrounds.^125,144 Six studies^{65,79,87,115,118,122} included participants younger than 18 years of age (range: 9¹¹⁵ to 15⁶⁵) while the remaining articles investigated effects in healthy adults, with most including participants between 18 and 30 years. Only Cesar et al.,⁸⁸ Connolly et al.,⁸⁹ and Fonta et al.⁹⁸ reported mean age ≥30 years, while Shelton and Kumar⁶⁴ provided a range up to 35 years and Kopec et al.¹⁰⁹ included participants up to 39 years of age.

Note that during the systematic search alone, 120 studies were excluded based on their lack of pre–post comparisons (n = 71) (e.g., Refs.^38,146, 147, 148) or for having uncontrolled research designs (n = 49) (e.g., Refs.^149,150).

3.1. Methodological quality, risk of bias, and certainty of evidence

With a PEDro score of 3.95 ± 0.67 (mean ± SD) (score ranging from 2 to 6), risk of bias was rated as poor. Sixty-seven of 83 included studies randomly allocated participants into groups or the respective intervention sequence,^{4,6,8,19,54,62,63, 65,}76, 78, 79, 80, 81, 82, 83, 84, 85, 87, 88, 89^,94, 95, 96, 97, 98, 99, 100, 101^,103,105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125^,127, 128, 129, 130, 131, 132, 133, 134, 135, 136^,139, 140, 141, 142, 143, 144, 145 81 studies had equal pretest values,4, 5, 6^,8,19,54,62, 63, 64, 65, 66, 67^,74, 75, 76^,78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100^,102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140^,142, 143, 144, 145 1 study⁹⁵ blinded all investigators; however, allocation concealment as well as blinding of the participants and therapists was never reported (Supplementary Table 2).

Considering the GRADE criteria, the certainty of evidence was initially rated as high due to the inclusion of (randomized) controlled trials. The level of evidence was finally graded as low (downgraded 2 levels due to risk of bias considering the poor PEDro score for this sub-group) for controlled studies addressing static stretching on maximum strength parameters, as moderate (downgraded 1 level due to risk of bias considering the fair PEDro score for this sub-group) for static stretching on speed strength parameters when compared to passive control, and as moderate (downgraded 1 level due to risk of bias considering the fair PEDro score for this sub-group) when compared to alternative interventions.

3.2. Quantitative analysis of stretch-induced force deficit with passive control comparisons

3.2.1. Strength and performance effects

We performed overall statistics including all participants and, additionally, separate analyses for age-dependent effects (adults vs. underage). However, a minimum of 3 eligible studies were considered the minimum to further calculate subgroup analyses. (If 3 or more studies were included to the overall sample size to improve methods homogeneity). If this was not possible, sensitivity analyses were used excluding studies with underaged participants. The results of all analyses are provided in Supplementary Table 3.

Overall, including all types of stretching, 220 ES extracted from 53 studies4, 5, 6^{,8,54,66,67,74,75,}78, 79, 80, 81^,84,86, 87, 88^,90, 91, 92, 93^,96,103,104,106, 107, 108^{,111,112,115,}117, 118, 119, 120, 121, 122, 123^{,125,126,127,129,}132, 133, 134, 135, 136^,138, 139, 140, 141, 142, 143, 144 yielded an ES of –0.06 (p = 0.29, τ² = 0.16), showing no significant impairment of subsequent maximal strength and speed performance capacity when considered together. Static stretching (51 studies,4, 5, 6^{,8,54,66,67,74,75,}79, 80, 81^,84,86, 87, 88^{,90,92,93,96,103,104,}106, 107, 108^{, 111,112,115,}117, 118, 119, 120, 121, 122, 123^,125, 126, 127^,129,132, 133, 134, 135, 136^,138, 139, 140, 141, 142, 143, 144 163 ES) resulted in ES = –0.11 (p = 0.12, τ² = 0.13); durations ≥60 s per bout (10 studies,^{5,8,54,103,108,115,123,134,136,139} 17 ES) increased the ES to –0.48 (p = 0.17, τ² = 0.61); and stretching volumes >480 s per intervention (9 studies,^5,6,8,54,78, 79, 80^,90,103 30 ES) showed reductions in ES = –0.28 (p = 0.04, τ² = 0.14). No effects are reported for lower volumes (≤480 s) (p = 0.37) (Table 1). Separate analyses including only underaged participants (5 studies^{79,87,115,118,122} with passive control) did not reach the level of significance (p: 0.10–0.56) and had high heterogeneity (τ²: 0.36–2.64). The adult-only analyses revealed a small effect for static stretching using overall volumes <60 s (8 studies,^{106,117,119,121,123,125,134,135} 20 ES) on maximal strength and speed performance combined, with ES = –0.3 (p = 0.02) and without interstudy heterogeneity (τ² = 0). The remaining subgroup or sensitivity analyses did not reach the level of significance (p: 0.053–0.49) (Supplementary Table 3).

Table 1.

Meta-analytic results for acute effects of stretching on strength and performance capacity compared to a passive control condition or group.

Comparison	Effect size (95%CI)	p	τ2	Ω2	Number of clusters/outcomes
Stretch-induced force and performance deficit	–0.06 (–0.19 to 0.06)	0.29	0.16	0	53/220
Dynamic stretch-induced force and performance change	0.20 (–0.05 to 0.45)	0.11	0.29	0	19/46
Static stretch-induced force and performance deficit	–0.11 (–0.25 to 0.03)	0.12	0.13	0	51/163
Static stretch-induced force and performance deficit (volume ≥60 s)	–0.07 (–0.21 to 0.07)	0.34	0.13	0	49/149
Static stretch-induced force and performance deficit (volume <60 s)	–0.30* (–0.52 to –0.07)	0.02	0	0	8/20
Static stretch-induced force and performance deficit (bout ≥60 s)	–0.48 (–1.21 to –0.24)	0.17	0.61	0	10/17
Static stretch-induced force and performance deficit (bout <60 s)	–0.07 (–0.21 to 0.06)	0.29	0.08	0	43/146
Static stretch-induced force and performance deficit (volume >480 s)	–0.28* (–0.54 to –0.02)	0.04	0.14	0	9/30
Static stretch-induced force and performance deficit (volume ≤480 s)	–0.07 (–0.22 to 0.08)	0.37	0.11	0	43/139
Stretch-induced force deficit	–0.15 (–0.310 to 0.005)	0.06	0.16	0.01	37/150
Dynamic stretch-induced force change	0.09 (–0.25 to 0.44)	0.55	0.05	0	9/26
Static stretch-induced force deficit	–0.21* (–0.39 to –0.02)	0.003	0.21	0	36/124
Static stretch-induced force deficit (volume ≥60 s)	–0.18 (–0.39 to 0.03)	0.08	0.23	0	34/107
Static stretch-induced force deficit (volume <60 s)	–0.47* (–0.66 to –0.28)	<0.001	0	0.003	7/22
Static stretch-induced force deficit (volume >480 s)	–0.46* (–0.85 to –0.06)	0.03	0.14	0	6/23
Static stretch-induced force deficit (volume ≤480 s)	–0.14 (–0.39 to 0.10)	0.24	0.24	0.06	27/95
Static stretch-induced force deficit (bout ≥60 s)	–0.84* (–1.32 to –0.37)	0.004	0.16	0	8/15
Static stretch-induced force deficit (bout <60 s)	–0.13 (–0.32 to 0.07)	0.20	0.17	0	30/109
Stretch-induced performance change	0.08 (–0.06 to 0.21)	0.26	0.10	0	22/68
Dynamic stretch-induced performance change	0.34 (–0.19 to 0.88)	0.19	0.54	0	13/20
Dynamic stretch-induced explosive strength/RFD change	0.30 (–0.58 to 1.18)	0.39	0.42	0	4/7
Dynamic stretch-induced performance change (athletic performance)	0.38 (–0.47 to 1.23)	0.33	0.65	0	9/13
Dynamic stretch-induced performance change (jump performance)	0.87 (–0.89 to 2.63)	0.27	1.1	0	7/8
Static stretch-induced performance change	0.07 (–0.07 to 0.22)	0.29	0	0	21/49
Static stretch-induced explosive strength/RFD deficit	–0.03 (–0.39 to 0.34)	0.86	0	0	6/12
Static stretch-induced performance change (athletic performance)	0.13 (–0.08 to 0.35)	0.20	0.20	0	15/37
Static stretch-induced performance change (jump performance)	0.21 (–0.04 to 0.45)	0.09	0.23	0	12/32

^⁎Indicates a significant difference.

Abbreviations: 95%CI = 95% confidence interval; RFD = rate of force development.

3.2.2. Static stretch-induced force deficit

Focusing solely on static stretching and its effects on force production (36 studies,4, 5, 6^{,8,54,67,74,75,78,81,87,88,90,92,93,103,104,}106, 107, 108^{,111,112,118,119,121,123,125,127,129,}132, 133, 134, 135, 136^,141,142 124 ES) confirmed a significant decrease, with ES = –0.21 (p = 0.003, τ² = 0.21). Stretching durations >60 s per bout (8 studies,^{5,8,54,103,108,123,134,136} 15 ES) showed significant effects, with ES = –0.84 (p = 0.004, τ² = 0.16), as did stretching volume >480 s (6 studies,^{5,8,54,78,90,103} 23 ES), with ES = –0.46 (p = 0.03, τ² = 0.14). Shorter stretching durations did not reach the level of significance (p: 0.08–0.24) (Fig. 2 and Table 1).

Fig. 2.

Forest plots for static stretch-induced force deficit for volumes of ≥60 s per bout. CG = control group; CMD = calf muscle device; ES = effect size; LP = leg press; MVC = maximal voluntary contraction; PT = peak torque; SS = static stretching.

Analyses including underaged participants only (2 studies^87,118) did not reach the level of significance (p = 0.29) and had high heterogeneity (τ² ≤ 0.65). The adults-only analyses revealed significant results for the stretch and the static stretch-induced force deficit (ES = –0.21, p = 0.005, τ² = 0.09, from 35 studies4, 5, 6^,8,54,67,74, 75, 76^,78,81,88,90, 91, 92, 93^,103,104,106, 107, 108^{,111,112,119,121,123,125,127,}132, 133, 134, 135, 136^,141,142 with 140 ES; and ES = –0.26, p = 0.002, τ² = 0.11, from 34 studies4, 5, 6^,8,54,67, 74, 75, 76^{,78,81,88,90,92,93,103,104,}106, 107, 108^{,111,112,119,121,123,125,127,}132, 133, 134, 135, 136^,141,142 with 118 ES, respectively). Also, volumes of ≥60 s (ES = –0.25, p = 0.007, τ² = 0.15, from 32 studies4, 5, 6^,8,54,67,74, 75, 76^{,78,81,88,90,92,93,103,104,}106, 107, 108^{,111,112,121, 123,125,127,}132, 133, 134^,136,141,142 with 101 ES) and <60 s (ES = –0.47, p < 0.001, τ² = 0, from 7 studies^{106,119,121,123,134,135,141} with 22 ES) as well as volumes of ≤480 s (ES = –0.22, p = 0.04, τ² = 0.11, from 25 studies^{4,6,67,76,81,88,93,104,}106, 107, 108^{,111,112,119,121,123,125,127,}132, 133, 134, 135, 136^,141,142 with 89 ES) showed small but significant strength reductions. While stretching bouts <60 s caused trivial reductions (ES = –0.18, p = 0.03, τ² = 0.07, from 28 studies^{4,6,67,74,75,76,78,81,88,90,92,93,104,106,107,111,112,119,121,123,125,127,}132, 133, 134, 135^,141,142 with 103 ES) in adults only, studies with stretching bouts ≥60 s were only performed in adults and revealed large magnitude effects (ES = –0.84, p = 0.004, τ² = 0.16, from 8 studies^{5,8,54,103,108,123,134,136} with 15 ES) (Supplementary Table 3). While in 3 cases (volume ≥60 s, volume ≤480 s, and bout <60 s) the overall analysis did not reach the level of significance, subgroup/sensitivity analyses showed significant strength reductions, with ES: –0.18 to –0.25, p: 0.007–0.040, and τ²: 0.07–0.15 in adults only (Supplementary Table 3).

3.2.3. Static stretch-induced performance deficit

Excluding maximal strength measures, acute static stretching showed significant positive but trivial effects on speed performance (RFD, explosive strength, jumping, sprinting, and throwing), with ES = 0.07 (p = 0.29, τ² = 0, from 21 studies^66,75,76,78, 79, 80, 81^{,84,86,96,104,111,115,117,}120, 121, 122^,126,138, 139, 140 with 49 ES). Athletic performance parameters (jumping, sprinting, and throwing) could be extracted from 15 studies78, 79, 80, 81^{,84,86,96,115,117,120,122,126,}138, 139, 140 with 37 effects and showed significant but trivial positive effects (ES = 0.13, p = 0.20, τ² = 0.20), while jump-only performance even tended to be positively influenced (ES = 0.21, p = 0.09, τ² = 0.23) (from 12 studies^{78,80,81,84,86,96,115,117,120,122,126,138} with 32 ES) (Table 1). Explosive strength/RFD was not affected by static stretching in any of the subgroups (p: 0.39–0.99, τ²: 0–0.42). Even though Behm et al.^12,151 showed that the intra-individual control leg might be a considerable passive condition, Zhou et al.¹⁵² indicated contralateral and global effects when performing training. Thus, the same analyses were performed, and the contralateral conditions were left aside, without changing the significance or magnitude of the ES. Separate analyses including underaged participants only did not reach the level of significance (p: 0.65–0.90) and had small-to-high heterogeneity (τ²: 0.38–2.04) in 2 analyses. The adults-only analyses revealed significant results for static stretching on athletic movements, including jumping, sprinting, and throwing (ES = 0.12, p = 0.03, τ² = 0, from 12 studies^{78,80,81,84,86,96,117,120,126,138,139,140} with 33 ES), and on jumping performance when considered alone (ES = 0.15, p = 0.006, τ² = 0, 10 studies^{78,80,81,84,86,96,117,120,126,138} with 30 ES), which was only significant in the adult group.

3.3. Quantitative analysis of dynamic stretching effects on maximal strength and speed performance

Dynamic stretching subgroup analysis against passive controls (19 studies,^{66,67,78,79,87,91,95,107,109,111,115,}117, 118, 119^{,122,133,139,143,144} 46 ES) showed ES = 0.20 (p = 0.11, τ² = 0.29). While no subgroup or sensitivity analysis showed significant stretch-induced changes (p: 0.18–0.95), the adults-only analysis of jumping performance did reveal small magnitude dynamic stretch-induced performance benefits (ES = 0.28, p = 0.02, τ² = 0, from 5 studies^{78,95,109,117,144} with 6 ES). Similar to previous results, we saw large heterogeneity among the few studies done in children, and the results did not reach significance (up to p = 0.51, τ² = 6.78).

3.4. Quantitative analysis of stretch-induced force deficit opposed by alternative interventions

The second meta-analysis was performed for studies directly comparing stretching effects with concurrent exercise routines (active controls). Incorporating both maximal strength and speed performance, static stretching showed significant decreases overall (ES = –0.18, p = 0.03, τ² = 0.004, 31 studies^{54,62,65,66,78,80,85,89,94,95,}97, 98, 99, 100, 101^{,105,107,110,112,113,116,}118, 119, 120^{, 124,125,127,130,131,137,140} with 90 ES) and with bouts <60 s (ES = –0.17, p = 0.04, τ² = 0.008, 27 studies^{62,65,66,78,80,85,94,95,}97, 98, 99, 100, 101^{,107,110,112,113,116,}118, 119, 120^{,125,127,130,131,137,140} with 84 ES), while other comparisons failed to reach the level of significance (p: 0.051–0.07). When focusing exclusively on maximal strength parameters, static stretching exhibited decreased outcomes (ES = –0.25, p < 0.001, τ² = 0, 16 studies^{54,62,78,89,98,101,105,107,110,112,113,118,119,125,127,131} with 40 ES), as did volumes ≥60 s (ES = –0.25, p < 0.001, τ² = 0, 14 studies^{54,62,78,89,98,101,105,107,110,112,113,118,127,131} with 36 ES) and stretching bouts <60 s (ES = –0.28, p < 0.001, τ² = 0, 13 studies^{62,78,98,101,107,110,112,113,118,119,125,127,131} with 35 ES) (Table 2).

Table 2.

Meta-analytic results for acute effects of stretching compared with alternative interventions assumed to induce muscle fatigue.

Comparison	Effect size (95%CI)	p	τ2	Ω2	Number of clusters/outcomes
Static stretch-induced force and performance deficit	–0.18* (–0.33 to –0.02)	0.03	0.004	0	31/90
Static stretch-induced force and performance deficit (bout <60 s)	–0.17* (–0.330 to –0.008)	0.04	0.008	0	27/84
Static stretch-induced force and performance deficit (volume ≤480 s)	–0.16 (–0.320 to 0.001)	0.05	0.007	0	28/78
Static stretch-induced force and performance deficit (volume >480 s)	–0.32 (–0.71 to 0.07)	0.07	0	0	3/12
Static stretch-induced force deficit	–0.25* (–0.36 to –0.13)	<0.001	0	0	16/40
Static stretch-induced force deficit (volume >60 s)	–0.25* (–0.38 to –0.13)	<0.001	0	0	14/36
Static stretch-induced force deficit (bout <60 s)	–0.28* (–0.41 to –0.15)	<0.001	0	0	13/35
Static stretch-induced performance deficit	–0.12 (–0.35 to 0.11)	0.30	0.04	0	15/45
Static stretch-induced performance deficit (volume ≤480 s)	–0.09 (–0.32 to 0.14)	0.41	0.04	0	13/39
Static stretch-induced performance deficit jump	–0.05 (–0.24 to 0.13)	0.55	0	0	13/40

^⁎Indicates a significant difference.

Abbreviation: 95%CI = 95% confidence interval.

In contrast, but similar to passively controlled comparisons, Table 2 shows that no difference in speed performance was observed (ES: –0.12 to –0.043, p: 0.30–0.62, τ²: 0–0.04). Since only 1 study⁶⁵ opposed stretching to active controls in underaged participants, there were no separate subgroups, but sensitivity analyses were performed, and the results were not different from those of the overall analysis (Supplementary Table 4).

Publication bias was assessed via visual funnel plot inspection (Supplementary Fig. 1), and no publication bias was indicated. With p > 0.05, the Egger's regression confirmed this impression (p = 0.76 and 0.56).

4. Discussion

Since previous systematic review results arose from studies that were uncontrolled and/or lacked pre–post comparisons,^13,14 this meta-analysis exclusively included controlled, pre–post studies investigating acute stretch-induced effects on maximal strength and speed performance parameters. In accordance with previous literatures, our analysis revealed significant maximal strength decreases, especially when using longer stretching durations (>60 s per bout).^9,14,21 However, in contrast to previous suggestions, we were unable to confirm negative effects of acute static stretching on speed and/or athletic performance, including jumping, sprinting, or throwing. Moreover, some subgroup analyses found significant trivial effects on athletic performance and even exhibited beneficial stretching effects on performance when compared to a passive control. In accordance with common recommendations, dynamic stretching did not affect strength performance (p: 0.11–0.95); and similar to static conditions, a positive effect on jumping performance was measured in adults (ES = 0.28, p = 0.02). These results seem to be partially in conflict with the previous analysis performed by Li et al.,²³ who found a positive influence of dynamic stretching. However, this discrepancy might originate from differences in the comparison conditions, as Li et al.²³ did not include a passive control condition but compared dynamic stretching with dynamic + static and static stretching only.

As acute maximal strength and speed performance losses have been attributed to diminished M-wave EMG amplitude as well as compliance changes in the muscle–tendon unit²¹ (which is, in fact, not stretch specific²⁴), we pooled ES from studies directly comparing static stretching effects with alternative interventions. We found significantly superior maximal strength decreases from static stretching independent of the applied stretching volume, while speed performance (including sprint, jump, and throwing performance) did not show a significant difference from other interventions.

In all comparisons (passive and active), the subgroups for underaged populations showed high heterogeneity and 95%CIs, making the interpretation of the results not meaningful. However, decreasing outcome and methods heterogeneity by performing separate analyses for adult participants only furthered the significance of stretch effects, indicating that especially high volumes of static stretching may harm strength outcomes in adults. No conclusive statements can be made for children due to study scarcity.

4.1. Previous research

In the latest systematic review with meta-analysis performed in 2013, Simic et al.¹³ concluded: “Our results clearly show that SS (static stretching) before exercise has significant and practically relevant negative acute effects on maximal muscle strength and explosive muscular performance, while the corresponding acute effects on muscle power remain unclear”. This statement, however, was unjustified in several ways. First, the authors reported a pooled maximal strength ES of –0.10. Even though statistically significant, it is commonly classified as trivial. While in elite athletes a force loss of just a few percent might have critical impact, in the broad recreational and sedentary population the effects seem negligible. Furthermore, several included studies showing negative influence reported isolated muscle force losses after impractically long stretching durations^5,8,54 without checking the transferability to sports practice. Second, the classification of unclear power effects seems to be untrue because since their meta-analysis included 104 studies with 57 data points for explosive strength, a sufficient study number was included. Obtaining non-significant effects is, in fact, not unclear. It seems questionable to interpret the trivial (ES = –0.10) and non-significant effects (ES = –0.03) of stretching on performance when recommending the avoidance of static stretching in warm-up routines. The analysis by Li et al.²³ showed static stretching variants to reduce explosive strength; however, all studies were not compared to passive controls but to alternative stretching variants.

We updated the literature search and included more than 80 controlled pre–post studies to improve interpretability. Our findings confirm a significant stretch-induced force deficit with small to large effects (ES up to –0.84) with a confirmed dose–response relationship (as suggested by Kay and Blazevich,¹⁴ Behm et al.,⁹ and Behm et al.²¹) and a generally small heterogeneity in overall and adults-only analyses, with 2 outliers (bouts ≥60 s overall and adults with τ²: 0.43–0.61). In their systematic review, Kay and Blazevich¹⁴ reported no meaningful strength decreases for stretching durations <30 s or those from 30–45 s. However, referring to dose–response evaluations, Kay and Blazevich¹⁴ stated a rapid enhancement of the percentage reduction to 61% when stretching durations exceed 1 min. No meaningful further strength decrease was reported when enhancing the stretching duration. This meta-analysis, however, found a significant decrease for short stretching durations when using a static stretch cut-off volume of 60 s, while higher volumes did not reach the level of significance. This controversial finding might be attributable to the biased sample size of the subgroups, as there was a comparatively large number of higher dosage studies as opposed to those using accumulated stretching volumes of <60 s for, for example, the static stretch-induced maximal strength and speed performance deficit (49 vs. 8 studies, respectively).

Nevertheless, most of the studies tested isolated muscle strength or power production, which might not be highly applicable in most sport and therapeutical settings in which complex movements are performed. Since acute effects seem to have little, if any, influence on complex movement tasks with speed performance characteristics, previous recommendations to avoid static stretching with practically relevant stretching durations^{74,76,92,129,153} seem questionable as well, especially since previous reports highlighted counteracting the stretch-induced force deficit by adding dynamic and activating warm-up activities after stretching.^154,155

4.2. Suggested underlying mechanisms

4.2.1. Reflex mechanisms

Behm et al.,^9,21 Trajano et al.,²⁶ and Chaabene et al.¹⁰ reviewed the available literatures regarding the central and peripheral mechanisms, including reflex mechanisms such as H-,156, 157, 158, 159 E-,¹⁶⁰ muscle spindle-, or Golgi-tendon reflexes;²⁶ EMG activity;^10,26,151 as well as structural changes (e.g., stiffness in the muscle–tendon unit). However, even longer stretching durations showed conflicting results,^158,161, 162, 163 and potential effects were described as dissipating within seconds.^164,165 Consequently, Behm et al.²¹ classified stretch-specific presynaptic H-reflexes, postsynaptic exteroceptive (E-) reflexes^21,166 (which were measured after long stretching durations¹⁶⁷), as well as tendon (T-)¹⁵⁹ and muscle spindle reflexes¹⁶³ as factors unlikely to explain the force loss.

4.2.2. Central nervous innervation

In contrast to highly specific stretching effects, a reduced EMG activity (i.e., diminished M-wave) was reported as the exclusive evidence-based factor that could be clearly related to the stretch-induced force deficit.²¹ Literature provides several explanations for the diminished EMG activity, especially after higher training dosages, including stretch-induced restricted blood flow¹⁶⁸ or muscle architecture changes, such as fascicle length, pennation angle, passive fascicle stiffness, or changes in passive torque-joint angles.169, 170, 171 These alterations might affect EMG results, as measurement conditions (e.g., the electrodes’ relative position to the muscle fiber) could have changed.^172,173

While reduced central nervous innervation was the main contributor for reduced strength parameters, especially after higher static stretching dosage (≥60 s per bout and >480 s volume), determining whether the resulting diminished EMG activity was stretching-specific is still required. Trajano et al.²⁶ noted that several of the reported contributors for changed EMG activity could be attributed to general fatigue instead of to stretch-specific responses. As such, it is possible that general muscular fatigue or damage could cause the reduction in subsequent strength and performance.

4.2.3. Exercise-induced fatigue

Based on shared underlying mechanisms in general (e.g., fatigue and stretch-induced deficits), our second analysis aimed to investigate whether stretching reduces the ability to maximally contract the muscles more than alternative training routines. Even though we hypothesized that general fatigue could be responsible for strength and performance deficits after concurrent interventions, most authors exclusively investigated performance outcomes, such as maximal or explosive strength, jumping or sprinting performance. On the other hand, underlying mechanisms, such as EMG activity, reflex mechanisms, or stiffness parameters, were only assessed in a handful of the active comparison studies. However, even though researchers^8,12,19,21 showed significant EMG changes, impaired reflex mechanisms, or decreased stiffness accompanying the stretch-induced force/performance deficit, these effects might not be exclusively attributable to stretching. It has been known for decades that muscular fatigue or acutely damaged tissue are functional and structural signs following exhausting exercises (especially if eccentric muscle actions are included) that diminish the contractile ability of muscles via mechanical stress, which can be induced by several muscular activities.¹⁷⁴ For instance, fatiguing muscular exercises, including evoked contractions for the plantar flexors¹⁷⁵ or electrical stimulation training after isometric contractions,¹⁷⁶ decreased EMG up to 57%. However, it is not necessary to perform any type of (sub)maximal contractions: downhill running,¹⁷⁷ 100 repetitions of squats (10 × 10 with 70% bodyweight),¹⁷⁸ 100 repetitions of eccentric squats with 80% of the one-repetition maximum,¹⁷⁹ and 100 eccentric knee extensor contractions¹⁸⁰ were sufficient to induce fatigue and were designed to induce muscle damage. In all studies, the authors reported a significantly reduced force production capacity (18%–35%) accompanied by an elevation of creatine kinase activity and reduced jumping performance.¹⁸¹ Indeed, over 20 years ago, Allen¹⁸² described how cytoskeletal proteins stabilize the sarcomere structure, which might be injured by overextension; such overextension could, in turn, be caused by stretching but also by mechanical overload or exhausting activities.

Even though we were not able to find studies separating stretch-induced changes from general activity-induced fatigue, deteriorations from static stretching were superior to alternatives, suggesting there are specific effects that cannot be exclusively explained by general fatigue independent of the performed exercise. Nevertheless, many aspects of stretching and its effects on muscular and non-muscular tissue remain unknown. Consequently, our results call for conducting future research to investigate the origin of EMG reductions that can be exclusively attributed to static stretching. The current literatures do not allow us to draw final conclusions.

4.2.4. Maximal strength vs. complex athletic performance activities

While the potential mechanisms behind strength loss were discussed extensively above, the diminished ability to innervate muscle could be reasonably assumed to negatively affect athletic performance as well.^96,183,184 However, in accordance with Simic et al.,¹³ who found negligible ES, our analysis revealed no impairments in athletic performance such as sprinting, jumping, and throwing. On the contrary, effects indicated improvements in athletic movements (ES = 0.12, p = 0.030), and particularly in jumping (ES = 0.15, p = 0.006). This controversy might be attributable to the fact that stretching was frequently performed for single muscles only, such as the plantar flexors,^{54,101,103,107} hamstrings,^{90,94,102,124} or quadriceps.^74,90,92,145 The ability to exert high forces within a short time frame might influence jumping performance; however, the static stretch-induced force deficit with small magnitude effects (overall static ES = –0.21, p = 0.003) might only play a negligible role in multijointed, complex movements that are determined by a complex interaction of central nervous innervation patterns (innervation frequency, synchronization, and fiber recruitment²⁸) in complete muscle chains. While maximal strength in complex movements such as the squat or deadlift showed high correlations with jumping and sprinting,185, 186, 187 the ability to exert large muscle forces in individual and small muscle groups such as the calf muscles explained less than 30% of the variance in jumping performance.¹⁸⁸ Even though long stretching durations showed higher ES, stretching effects on athletic performance were rarely investigated with stretching durations exceeding 90 s per bout.¹³⁹ However, to improve the general understanding of these effects, further studies exploring the specific underlying mechanisms of the stretch-induced force deficit as well as the role of maximal strength of individual muscle groups in athletic performance (jumping, sprinting, throwing) are necessary.

Since no differences between the effects of static stretching and concurrent warm-up activities on jumping, sprinting, and throwing performance were observed, there seems to be no evidence to recommend avoiding specific stretching interventions before maximal athletic performance. It seems that it is more harmful to induce prolonged and/or fatiguing muscular activities in general because they might cause metabolic or mechanical fatigue, which could then result in reduced contractability of muscular tissue. Since our results at least partially oppose those reported by Li et al.,²³ further research is necessary.

4.3. Limitations

Even though we were able to pool outcome parameters (e.g., maximal speed or explosive strength) and separate those from alternative interventions, specific underlying mechanisms were not pooled based on methodological issues (EMG-recording variability and individuality); thus, specific stretch mechanisms need further investigations.

Additionally, apart from high outcome homogeneity, which can be achieved by subgroup analyses, meta-analyses assume a high degree of method homogeneity to compare pre–post effects. Even though we aimed to reduce methodological study heterogeneity by applying restrictive eligibility criteria and performing several subgroup and sensitivity analyses, the comparability of different study methods in sport and exercise science remains a serious limitation when aiming to pool ES, which also biases our study results. Especially when it comes to interpreting the results of our second round of analyses, which compare stretching with alternative interventions, it must be noted that meta-analytical software compares numbers but does not interpret their origin. Even though there are valid concerns over biased interpretability, it is common in sport science to include incomparable interventions;^24,189 this is also the case when it comes to pooling intervention effects vs. passive or active control conditions.190, 191, 192 To ensure maximal homogeneity within groups and avoid comparing static stretching with non-comparable interventions, subgroup analyses and sensitivity analyses were performed.

It must be noted that the “practical durations” classification used for specific subgroup analyses was set at 8 min because this was the norm in the existing literatures.74, 75, 76 However, since a total stretching time of 8 min seems to be somewhat long in terms of practical application, the cut-off appears to be quite arbitrary. The lack of official guidelines classifying any stretching duration as practically relevant calls for clearer terminology and classification guidelines to improve consensus for future research.

Another aspect limiting the interpretability of our results is attributable to the sample size per se. Hypothetically, stretching could induce different effects in different age groups,^193,194 with a potential influence from maturation. While we aimed to account for potential moderators by performing separate analyses by age, the literature lacks sufficient studies investigating these effects in underaged participants, or in females only. Even among the few studies included, the results showed high interstudy outcome heterogeneity (τ² up to 2.64) and large 95%CIs, which made the results uninterpretable. Thus, we are not able to provide a final statement for age or sex-specific effects, and we call for further well-designed and controlled research to improve future analyses and enable authors to perform a higher number of valid subgroup analyses.

4.4. Practical applications

Our review sharpens previous analytical methods to improve interpretability by enhancing the quality requirements of included studies; in doing so, it confirms the static stretch-induced force deficit. However, while present when strength was tested under isolated laboratory conditions, these detrimental effects were not observed during athletic performance, such as jumping, sprinting, or throwing. Moreover, it should be noted that the subgroup analysis investigating static stretch-induced effects on jumping and sprinting showed trivial magnitude but significant performance enhancements, which is in contrast to common beliefs.¹⁹⁵ Also, when differentiating static stretching from concurrent warm-up alternatives, there was a significantly superior harmful effect on maximal strength, without transfer to speed and athletic performance. The practical relevance of the results must be considered in light of actual warm-up research, which has shown active warm-up routines can counteract the stretch-induced force deficit when they are performed subsequently.^154,155 Not only is static stretching often not performed in isolation in sports practice, our results showed no harmful effects on athletic performance measured via jumping, sprinting, and throwing performance. Thus, the actual analysis challenges current recommendations of categorial avoidance of static stretching in warm-up routines, especially when integrated with other (dynamic) conditions,^154,155 as both static and dynamic stretching indicated a small but significant benefit for subsequent jumping performance in adults.

5. Conclusion

Current evidence underlines the need to pay attention in specific laboratory settings when applying static stretching prior to isolated maximum strength testing, while non-significant to significant but trivial effects on sports-relevant performance parameters challenge current beliefs and reflect negligible relevance for sports practice. Even though we are not able to recommend stretching before performance based on the current evidence, rigorous avoidance of any type of stretching before performance seems to be without evidence, especially since the literature shows that potential strength impairments can be counteracted by subsequent dynamic activities.

Authors’ contributions

LHL performed the literature search, data extraction/processing, and quality assessment and revised the manuscript; KW performed the literature search, data extraction/processing, and quality assessment, took the lead in writing, revised the manuscript, and performed the statistical calculations. Both authors have read and approved the final version of the manuscript, and agreed with the order of presentation of the authors.

Competing interests

Both authors declare that they have no competing interests.