In the realm of gymnastics, a well-structured warm-up regimen is a crucial component for optimizing overall performance aspects, including physical, technical, and cognitive performance (Guidetti et al., 2009). Combining physiological and biomechanical considerations reveals the multifaceted significance of warming up in addition to the conventional understanding of warming up merely as a preparatory phase (Behm et al., 2021). The elevated muscle temperature achieved through a systematic warm-up routine increases flexibility and joint mobility (Opplert and Babault, 2018), enabling gymnasts to perform complex movements better. Furthermore, researches have shown that increased temperature enhances central nervous system function and augments the transmission speed of nervous impulses, potentially benefiting overall athletic performance (Bishop, 2003).

In the context of gymnastics warm-up protocols, the choice between static stretching (SS) and dynamic stretching (DS) methods induces various benefits and disadvantages (Siatras et al., 2003). Several studies indicated that SS may provide benefits such as enhanced muscle flexibility and joint range of motion (Behm and Chaouachi, 2011). However, SS might also impair neuromuscular activation (Chaabene et al., 2019), potentially worsening the dynamic and explosive movements included into gymnastic routines (Siatras et al., 2003). Meanwhile, dynamic warm-ups, particularly when incorporating sport-specific exercises, are likely to induce a more comprehensive physiological and neural preparatory response (Ahmadabadi et al., 2015). Nevertheless, an accurate design is needed to ensure that the warm up routine adequately addresses the diverse demands of gymnastic performances (Takeuchi et al., 2019).

Post-activation potentiation enhancement (PAPEs) (Blazevich and Babault, 2019) that arise from intense maximal contractions introduce an additional stimulus to the preparatory phase (Sale, 2002). Including intense contractions in warm-ups may increase muscle temperature, and muscle fiber water content, as well as muscle activation (Blazevich and Babault, 2019). This effect could improve the early stages of gymnastic performances (Dallas et al., 2019). PAPEs may also contribute to heightened neural drive, potentially increasing the maximum voluntary rate of force development and maximal muscle force (Hodgson et al., 2005). These enhancements in muscle function may substantially improve the dynamic and explosive movements of gymnastic.

The literature lacks comprehensive systematic reviews and meta-analyses investigating the effects of various warm-up strategies for gymnasts (Siatras et al., 2003; Guidetti et al., 2009). The lack of a synthesized overview makes it challenging to draw definitive conclusions about the most effective warm-up protocols for gymnastics. A systematic review and meta-analysis would offer a valuable and comprehensive perspective on the effects of warm-up approaches, providing a broad overview of the evidence on this topic. Such a study could also highlight the specific interventions that consistently yield positive outcomes, offering gymnasts and coaches evidence-based insights into how they can optimize warm-up routines.

Given the considerations outlined above, the principal objective of this systematic review with meta-analysis is to systematically analyze and compare the impacts of various warm-up strategies, including SS and DS, as well as post-activation potentiation techniques, on the immediate performance of gymnasts.

Adhering to the rigorous standards delineated by the PRISMA 2020 guidelines, this systematic review was conducted. A comprehensive protocol outlining our review procedures has been formally documented on the Open Science Framework (OSF) platform, assigned the project number osf.io/fk6vz, and accompanied by a DOI:10.17605/OSF.IO/FK6VZ.

In our systematic review and meta-analysis, we have thoroughly incorporated all original articles published in peer-reviewed journals, including those in "ahead-of-print" status. Language restrictions were intentionally avoided to ensure a comprehensive range of articles.

Our establishment of eligibility criteria adhered to the PICOS framework, and a detailed breakdown of these criteria can be found in Table 1. These criteria were carefully formulated to include studies involving gymnasts from any speciality, competitive level, age, or sex, with a minimum requirement of being at least tier 2 (trained/developmental, associated with local-level representation, regular training ~3 times per week, and training with a purpose to compete) on the Participants Classification Framework, which classifies a spectrum of exercise backgrounds and athletic abilities (McKay et al., 2022).

In the context of warm-up interventions, the criteria were clearly outlined, requiring participants to engage in, at least, one of the following warm-up strategies: (1) SS; (2) DS; (3) PAPE; and (4) vibrating platforms. The control group consisted of gymnasts who performed warm-up types that did not include SS, DS, PAPE, and VP.

Regarding our primary outcomes, we focused on parameters related to acute gymnastic responses, including athletic physical performance (e.g., jumping performance, muscular power, range of motion), technical execution (e.g., executing specific gymnastics technical exercises), and cognitive/mental performance (e.g., increasing attentional focus, cognitive arousal). Eligible study types included experimental or observational studies, either cross-sectional or parallel.

To ensure a thorough evaluation, a comprehensive review of the full texts was conducted to ascertain their eligibility for inclusion in the review.

Our approach to identifying pertinent studies entailed a thorough search across multiple databases, namely: (i) PubMed, (ii) Scopus, and (iii) Web of Science, up to November 13, 2023. To enhance the comprehensiveness of our methodology and mitigate the risk of overlooking relevant materials, we additionally conducted manual searches within the reference lists of the studies included into our review.

The search was conducted employing Boolean operators AND/OR, with a decision to refrain from using filters or constraints concerning publication dates or language. This approach was implemented to enhance the probability of identifying relevant studies. All the terms included were searched within the title and abstract. It is noteworthy that in PubMed and Scopus databases, the keywords were also selected. In the Web of Science – Core Collection, the terms were chosen specifically as topics. The exact code line used for conducting these searches was: ((Gymnastic* OR gymnast*) AND ("Warm-Up*" OR "Warmup*" OR "Warming-Up*" OR "Warming Up*"OR "static* stretch*"OR "dynamic* stretch*" OR "stretch*" OR "proprioceptive neuromuscular facilitation" OR "PNF" OR "Post-activation potentiation" OR "post-activation" OR "PAP" OR "postactivation potentiation")).

The screening process was conducted by two of the authors. They independently reviewed the retrieved records, including both titles and abstracts. Subsequently, each author individually assessed the full texts of the selected records. In instances where discrepancies in the evaluation arose, a collaborative reevaluation process was initiated to reach a consensus. If a consensus could not be achieved, the final decision was deferred to a third author.

To streamline record management, we employed EndNote X9.3.3 software, developed by Clarivate Analytics in Philadelphia, Pennsylvania, USA.

The data collection process was independently carried out by two of the authors. In cases where disagreements emerged during this phase, a third author served as a mediator. To improve efficiency and uphold organization throughout this procedure, we employed a dedicated Microsoft® Excel datasheet. This datasheet included all relevant data and essential information, ensuring a structured and effective approach to data management.

The data collection process involved extracting a comprehensive range of participant details and contextual factors. These variables included details, as the publication date, primary research objectives, sample size, country of origin, age distribution, sex, study design specifics, and the competitive level of the participants.

Regarding intervention-related conditions, we documented information regarding the study duration, training context, and various aspects of the training regimen. This involved specific details on the duration, repetitions, rest intervals, intensity, frequency, and training density.

Moreover, we extracted information from both the experimental and control groups. This included details regarding the type of exercises, exercise intensity, and volume.

Our primary focus during the extraction of main outcomes was based on parameters associated with acute gymnastic responses. This included assessments of athletic physical performance, such as jumping, muscular power, and range of motion; technical execution, including the quality of specific gymnastics technical exercises; and cognitive/mental performance, which involved measures to evaluate attentional focus and cognitive arousal. These measures were collected both before (baseline) and after the intervention, guaranteeing a minimum of two time points for assessing performance.

In the case of non-controlled studies, we evaluated the methodological quality of the included studies by applying a set of 27 criteria outlined in the modified Downs and Black Checklist (Downs and Black, 1998; Simic et al., 2010). This checklist categorizes its 27 items into distinct domains, namely "reporting" (10 items), "external validity" (3 items), "internal validity - bias" (7 items), "internal validity - confounding (selection bias)" (6 items), and "power" (1 item) (Trac et al., 2016).

Each item was assigned a score of either 0 (indicating poor quality) or 1 (indicating good quality), except for question 5 ("clear description of principal confounders"), which had a scoring range from 0 (not satisfactory) to 2 (fully satisfactory). Consequently, each study had the potential to attain a maximum score of 28.

To assess study quality, we established the following thresholds: (i) poor quality (<14 points); (ii) fair quality (14-18 points); (iii) good quality (19-23 points); and (iv) excellent quality (24-28 points). This systematic approach ensured a comprehensive evaluation of the methodological rigor across the included studies.

Effect sizes, specifically Hedges' g, were calculated for variables related to gymnastic performance in both intervention and control groups. These measurements were obtained from the means and standard deviations before and after the intervention, and the standardization was conducted using the standard deviation values after the intervention. To account for potential differences between studies that could impact warm-up effects, the DerSimonian and Laird random-effects model was utilized (Deeks et al., 2008; Kontopantelis et al., 2013).

The representation of effect size (ES) values incorporated 95% confidence intervals (95% CIs), with interpretation based on a scale: <0.2 for trivial, 0.2-0.6 for small, >0.6-1.2 for moderate, >1.2-2.0 for large, >2.0-4.0 for very large, and >4.0 for extremely large effects (Hopkins et al., 2009). Subsequently, it was considered pertinent to exclude a study from a specific meta-analysis if its ES value was ≥2, as such a result is considered atypical in warm-up studies following most interventions and may be identified as an outlier. (Kadlec et al., 2023).

To assess the influence of heterogeneity, I² statistics were employed, categorizing values as <25% for low impact, 25-75% for moderate impact, and >75% for high impact. To investigate the potential publication bias in continuous variables, the extended Egger's test was applied. To mitigate this bias, a sensitivity analysis was carried out using the trim and fill method, with L0 as the default estimator for the number of missing studies. All statistical analyses were executed using SPSS (version 28, IBM, USA), and the significance level was set at p ≤ 0.05.

Figure 1 illustrates the outcomes of our initial investigation, revealing a total of 591 titles. Seventeen studies that adhered to our predetermined eligibility criteria were identified. In addition to our database screening, we conducted an extensive manual search within the references cited in the selected articles. This supplementary search revealed two additional articles meeting our inclusion criteria. Consequently, our systematic review included a total of 19 articles (Supplementary Material 1).

Figure 2 illustrates the assessment of the risk of bias using the Downs and Black Checklist. Of the included studies, all of them ranged between 14 and 17 points, placing them within the classification of fair quality. The lack of blinding for both participants and evaluators, as well as the absence of a sample size estimation were the most common issues hindering the generalization of findings. Detailed scores for each assessment item are provided in Supplementary Material 2.

Table 2 describes the principal characteristics of the individual studies incorporated in the present systematic review.

Table 3 describes the outcomes of individual studies that have specifically examined the impacts of various warm-up strategies on the jumping and technical execution performance of gymnasts. Table 4 showed the outcomes of individual studies that have specifically examined the impacts of various warm-up strategies on the muscle strength, range of motion and balance gymnastics elements.

Results showed a non-significant difference between the SS condition and control conditions in squat jump (SJ) performance (Figure 3: ES = -0.073, 95% CI = -0.500; 0.354, p = 0.539, I² = 0.00%, Egger test two-tailed = 0.585). Similarly, non-significant difference between the VP condition and SS conditions regarding SJ performance were found (Figure 4: ES = 0.159, 95% CI = -0.482; 0.799, p = 0.398, I² = 0.00%, Egger test two-tailed = 0.422).

Non-significant difference between the SS condition and control conditions in countermovement jump (CMJ) performance were found (Figure 5: ES = 0.012, 95% CI = -0.262; 0.286, p = 0.896, I² = 0.00%, Egger test two-tailed = 0.685) as well as non-significant difference between the VP condition and SS conditions regarding CMJ performance (Figure 6: ES = 0.298, 95% CI = -0.638; 1.234, p = 0.304, I² = 0.10%, Egger test two-tailed = 0.011).

Non-significant difference between the SS condition and control conditions in gymnastic technical performance were found (Figure 7: ES = 0.428, 95% CI = -0.427; 1.283, p = 0.164, I² = 19.8%, Egger test two-tailed = 0.345). Non-significant difference between the VP condition and SS conditions regarding gymnastic technical performance were observed (Figure 8: ES = -0.378, 95% CI = -1.983; 1.226, p = 0.417, I² = 52.7%, Egger test two-tailed = 0.643).

The present systematic review and meta-analysis examined the influence of various warm-up strategies on the immediate performance of gymnasts. In addition to summarize the studies’ outcomes that provided additional information and facilitated comparisons, the meta-analysis for cases with a sufficient number of studies showed no significant differences between SS, vibration platforms (VP), or control conditions in SJ, CMJ, or the technical execution of gymnastics elements.

Several studies in gymnastics, including those by Melocchi et al. (2021), Montalvo and Dorgo (2020), and Di Cagno et al. (2010), have explored the contrasting effects of SS versus traditional warm-ups, specifically focusing on physical and technical performance. Moreover, numerous studies have examined SS characterized by prolonged muscle elongation (McNeal and Sands, 2003; Melocchi et al., 2021). While some studies indicate potential benefits, such as increased range of motion (Melocchi et al., 2021), others suggest that SS may lead to a transient decrease in muscle power (McNeal and Sands, 2003). Interestingly, a meta-analysis exploring the effects of SS on SJ and CMJ revealed that impairment levels depend on the duration of the static stretch (Simic et al., 2013).

Conversely, our results suggest that SS enhances specific gymnastics elements to a moderate but non-significant extent compared to control conditions. Siatras, (2014) showed that the angle reached in legs-horizontal during the V-sit (an exercise in which the gymnast sits with their legs extended and torso lifted off the ground, forming a V shape with the body) was significantly better after SS than after the control condition. Additionally, Van Zyl et al. (2011) reported that the range of motion in a forward split was significantly improved after SS compared to control conditions.

Conversely, for ballistic elements such as split leaps with the leg stretched, the flight time was significantly greater in the control condition than in the SS condition (Di Cagno et al., 2010). Similarly, other research showed that the running speed on the handspring vault was significantly higher in the control condition than in the SS condition (Siatras et al., 2003). Scientific evidence supports the immediate benefits of SS for gymnasts while considering static elements (flexibility), enhancing range of motion and flexibility. Kataura et al. (2017) showed that engaging in SS at a high intensity augments one’s range of motion and reduces passive muscle-tendon stiffness. However, caution is warranted, as SS may induce temporary decreases in ballistic technical elements, which are crucial to gymnastics performance. SS also temporarily decreases muscle stiffness, affecting the muscles’ ability to rapidly generate force due to a decrease of voluntary activation and persistent inward current effects on motoneuron excitability (Behm et al., 2021).

The included studies also often compared SS with the same stretching exercises performed on VP (Kinser et al., 2008; McNeal et al., 2011; Dallas et al., 2014a). VP devices, also known as whole-body vibration (WBV) devices, are intended to complement SS by utilizing mechanical vibrations to stimulate muscle activity and enhance neuromuscular function (Luo et al., 2005). These devices are designed to induce rapid muscle contractions, thereby improving muscle activation and neuromuscular facilitation (Cochrane, 2011). The reflexive responses triggered by the vibrations may also increase strength and power (Alam et al., 2018), helping gymnasts achieve precise coordination and control during their routines (Zasada et al., 2016).

This meta-analysis comparing SS with VP revealed no significant differences in performance outcomes. Specifically, for SJ, both conditions exhibited similar effects in the studies of Dallas et al. (2014b) and Dallas et al. (2014a). Moreover, regarding the forward split, Sands et al. (2008) and Zyl et al. (2011) demonstrated significant benefits of VP. Conversely, Kinser et al. (2008) revealed that the use of VP reduced the range of motion in the forward split compared to SS by 9-12%.

Two studies have hinted at potential benefits on specific gymnastic elements such as the forward split after VP (Sands et al., 2008; Van Zyl et al., 2011). These observations align with a previous research demonstrating that acute exposure to vibration can enhance flexibility (Đorđević et al., 2022). This improvement can be attributed to a muscle extensibility enhancement and a reduced muscle stiffness (Fowler et al., 2019). The mechanism underlying this effect is associated with VP’s ability to induce muscle activity reflex (e.g., tonic vibration reflex, bone myoregulation reflex) (Cidem et al., 2017), thereby enhancing neuromuscular control and increasing the joint range of motion.

Though several studies included in this review lacked the necessary data for a meta-analysis, they examined the impact of DS versus SS on gymnastic performance. Melocchi et al. (2021) found that compared to SS, DS significantly increased SJ and CMJ performance by 15.4% and 10.8%, respectively. Similarly, Montalvo et al. (2020) demonstrated significant benefits of DS, showing an 8.86% improvement in CMJ performance compared to SS. The potential increases in muscle force and/or shortening velocity, possibly prompted by fluid shifts into the working muscles, could enhance muscle function in a fiber-type-specific manner. Concurrently, alterations in neural circuitry, such as changes in H reflex, motor-evoked potentials, and cortico-medullary evoked potential amplitudes, suggest that PAPE is a contributing factor to the observed positive effects (Blazevich and Babault, 2019). In the context of squat and counter-movement, DS primes the neuromuscular system, enhancing the recruitment of motor units and optimizing force generation.

Non-statistically significant values were reported by Melocchi et al. (2021), revealing that DS exhibited a 2.55% greater improvement in specific gymnastic jumps compared to SS. Similarly, the 1.96% increase of running speed in the handspring vault reported after DS reported by Siatras et al. (2003) did not reach statistical significance. However, Zaggelidou et al. (2023) showed a significant 8.48% enhancement in bicep range of motion after DS compared to SS.

Melocchi et al. (2021) evaluated the hip joint range of motion and found no significant difference between SS and DS. While DS involves movement through a range of motion and SS involves holding a position, both modalities have been shown to increase flexibility in comparison to control conditions. The theoretical reasons supporting their comparable effects are related to the commonality of their underlying mechanisms, such as alterations in muscle and tendon compliance, increased stretch tolerance, and neurophysiological adaptations (Behm and Chaouachi, 2011). Furthermore, both DS and SS can promote the viscoelastic properties of the muscles and tendons, thereby improving one’s range of motion (Kubo et al., 2002).

Current research on warm-up protocols for gymnasts reveals several limitations that hinder a comprehensive understanding of their impacts on performance. One limitation is related to the different methodologies employed in the studies. For instance, analyses of stretching warm-ups often overlook the effects of varying stretching durations and their potential impacts on impairments or enhancements, as well as the duration of these effects (i.e., the post-warm-up period). Hence, certain methodological aspects may constrain some of our results. Additionally, most research has concentrated on short-term performance outcomes while overlooking the potential long-term effects of warm-up routines on skill acquisition.

Methodological differences in measuring performance outcomes, such as objective biomechanical measures and surrogates (a variable measurable in lieu of one that cannot be assessed through gold-standard methods), further complicate data interpretation. Future research should include longitudinal studies that assess the cumulative effects of warm-up routines on skill development. Finally, utilizing advanced technologies, such as motion capture and wearable sensors, can provide more accurate and objective measurements of performance outcomes.

The current systematic review with meta-analysis revealed no significant differences in the extent to which SS, DS, VP, and control conditions enhanced jumping performance or technical execution in gymnastics. However, several studies suggest that SS slightly outperforms DS in improving technical executions related to static range of motion (e.g., forward split). However, DS appears to enhance jumping performance to a greater extent than SS. These findings suggest that DS is a preferred warm-up strategy in acute settings, especially when performed in close proximity to the gymnastic performance. Meanwhile, SS could be a more favorable warm-up strategy for events involving static movements and range of motion. Due to the heterogeneity of the findings, caution is advised when applying the results of this systematic review and meta-analysis; the warm-up process should be individualized.