Efficacy of exercise interventions on prevention of sport-related concussion and related outcomes: a systematic review and meta-analysis Authors: Branimir Ivanic, Anna Cronström, Kajsa Johansson, Eva Ageberg Published: 2024 ABSTRACT AbstractObjectiveTo review the efficacy of exercise interventions on sport-related concussion (SRC) incidence, as well as on linear and rotational head accelerations, and isometric neck strength and to assess reporting completeness of exercise interventions using the Consensus on Exercise Reporting Template (CERT).DesignSystematic review and meta-analysis, according to the Prisma in Exercise, Rehabilitation, Sport medicine and SporTs science guidelines.Data sourcesSix databases (MEDLINE, Embase, CINAHL, Scopus, Web of Science CC and SPORTDiscus) were searched up to 26 June 2023.Eligibility criteria for selecting studiesRandomised controlled trials (RCTs), cluster RCTs or quasi-experimental studies, evaluating exercise interventions on SRC incidence, linear and rotational head accelerations, and/or isometric neck strength in male and/or female athletes of any age, and/or in a healthy general population.ResultsA total of 26 articles were included. A large effect size was observed for resistance training (RT) on isometric neck strength (standardised mean difference (SMD) 0.85; 95% CI 0.57 to 1.13; high-quality evidence). Non-significant effect sizes were observed for neuromuscular warm-up programmes on SRC incidence (risk ratio 0.69; 95% CI 0.39 to 1.23; low-quality evidence), or for RT on linear head acceleration (SMD −0.43; 95% CI −1.26 to 0.40; very low-quality evidence) or rotational head acceleration (SMD 0.08; 95% CI −0.61 to 0.77; low-quality evidence). No studies assessed the impact of RT on SRC incidence. CERT scores ranged from 4 to 16 (out of 19) with median score of 11.5 (IQR 9–13).ConclusionRT increases isometric neck strength, but the effect on SRC incidence is unknown. More adequately powered and rigorous trials are needed to evaluate the effect of exercise interventions on SRC incidence, and on linear and rotational head accelerations. Future studies should follow CERT guidelines, as the included interventions were generally not reported in sufficient detail for accurate replication.PROSPERO registration numberCRD42023435033. WHAT IS ALREADY KNOWN ON THIS TOPIC Given the reported rise in the incidence of sport-related concussion (SRC), it is important to explore primary prevention strategies. Given the reported rise in the incidence of sport-related concussion (SRC), it is important to explore primary prevention strategies. Potential primary prevention strategies include policy and rule changes, individual protective equipment and exercise interventions. Research on the efficacy of exercise interventions in the prevention of SRC is scarce. Potential primary prevention strategies include policy and rule changes, individual protective equipment and exercise interventions. Research on the efficacy of exercise interventions in the prevention of SRC is scarce. Higher values for linear and rotational head accelerations and lower values for isometric neck muscle strength may be potential indicators of SRC risk. Higher values for linear and rotational head accelerations and lower values for isometric neck muscle strength may be potential indicators of SRC risk. WHAT THIS STUDY ADDS This systematic review and meta-analysis on the efficacy of exercise interventions demonstrates that there is no significant effect size observed following neuromuscular warm-up programs on SRC incidence. This systematic review and meta-analysis on the efficacy of exercise interventions demonstrates that there is no significant effect size observed following neuromuscular warm-up programs on SRC incidence. A large effect size is observed following resistance training (RT) on isometric neck strength, whereas non-significant effect sizes are observed for RT on linear and rotational head accelerations. A large effect size is observed following resistance training (RT) on isometric neck strength, whereas non-significant effect sizes are observed for RT on linear and rotational head accelerations. Dynamic and multimodal resistance programmes are more efficacious in increasing isometric neck strength than static resistance programmes, with higher effect sizes for males than females and adults than adolescents. Dynamic and multimodal resistance programmes are more efficacious in increasing isometric neck strength than static resistance programmes, with higher effect sizes for males than females and adults than adolescents. No study evaluated the efficacy of neck RT on SRC incidence. No study evaluated the efficacy of neck RT on SRC incidence. Exercise interventions are generally not reported in sufficient detail, which can hinder accurate replication. Exercise interventions are generally not reported in sufficient detail, which can hinder accurate replication. Introduction Sport-related concussion (SRC) is defined as ‘a traumatic brain injury caused by a direct blow to the head, neck or body resulting in an impulsive force being transmitted to the brain that occurs in sports and exercise-related activities’.1 When the head is impacted, the skull rapidly accelerates while the brain’s movement is delayed due to its own inertia.2 This delay can lead to strain and pressure gradients within the brain tissue, which can result in injury if the tolerable limits are exceeded.2 SRC can result in a range of symptoms and signs, such as headache, nausea, confusion and sleep disturbance. These symptoms can appear immediately or develop over time, and while they usually resolve within days, some cases may take longer to recover.3 Approximately, 10%–25% of participants may experience prolonged symptoms beyond the 4-week recovery period.4 5 Given the reported rise in the incidence of SRC in recent years,68 and its potential for severe short-term and long-term health consequences, there is a growing focus on exploring the effects of primary prevention strategies.1 A recent systematic review on prevention strategies and modifiable risk factors for SRC and head impacts demonstrated that prevention strategies such as policy/rule changes (eg, limiting contact practice in American football) and individual protective equipment (eg, mouthguards in ice hockey) reduced the rate of SRC.9 The review also suggested that future research should focus on exercise components targeting SRC prevention.9 Since neck muscle strength has been proposed as a potentially modifiable factor in SRC prevention,10 alongside policy/rule changes and individual protective equipment, neck muscle training could become another effective SRC prevention strategy.11 It is hypothesised that targeted training to strengthen the neck musculature may improve the initial resistance of the head to external forces.11 This, in turn, could reduce the postimpact kinematic response of the head and lower the risk of SRC.11 Although there is no definitive evidence confirming a direct association between neck muscle strength and SRC incidence,9 several studies have suggested that individuals with lower isometric neck muscle strength experience greater head acceleration following head impacts.1217 Furthermore, some studies have reported that such individuals may be at a greater risk of sustaining SRC.1820 Research has also suggested that youth and females exhibit increased head accelerations following head impacts,122123 likely due to their significantly lower mean isometric neck muscle strength when compared with adults12 and males.12 14 20 In the interests of reducing the risk of SRC, it has been recommended that future research investigates the inclusion of neck strength components within already existing injury prevention programmes.9 Therefore, by evaluating the efficacy of different neck exercise modalities on isometric neck muscle strength, valuable insights could be gained into the implementation of appropriate neck strength components in such injury prevention programmes. It is believed that linear and rotational head accelerations accurately reflect the inertial response of the brain24 and that both types of head accelerations contribute to concussive injury, however, with two distinctly different injury mechanisms.25 Linear acceleration can cause a transient increase in intracranial pressure while rotational head acceleration can lead to shear strain due to relative motion.25 It is important to recognise that the severity and outcome of concussive injuries are often determined by the combined impact of both types of head accelerations. This highlights the complex interrelationship between linear and rotational forces in head trauma.2 26 For these reasons, further exploration of exercise interventions and their efficacy on kinematic parameters (linear and rotational head accelerations), and a potential intrinsic modifiable factor (isometric neck muscle strength) is warranted. Therefore, the primary aim of this study was to systematically review the efficacy of exercise interventions on SRC incidence, as well as on kinematic parameters (linear and rotational head accelerations), and a potentially modifiable factor related to the prevention of SRC (isometric neck strength). The secondary aim was to assess reporting completeness of the exercise interventions using the Consensus on Exercise Reporting Template (CERT). Methods Guidelines and review process The review was preregistered in PROSPERO (CRD42023435033) and conducted in accordance with the guidelines for implementing Prisma in Exercise, Rehabilitation, Sport medicine and SporTs science.27 A completed Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist is available in online supplemental file 1. Two authors (BI and KJ) independently evaluated articles for inclusion, performed data extraction and carried out the risk of bias, reporting completeness and certainty of evidence assessments. All disagreements were resolved by a consensus discussion between the two authors and remaining authors (AC and EA) if consensus could not be reached. Information sources and search strategy The literature search strategy was developed with the assistance of two experienced librarians. The search was performed between 24 June 2023 and 26 June 2023 in the following databases: MEDLINE, Embase, CINAHL, Scopus, Web of Science CC and SPORTDiscus (details available in online supplemental file 2). No language or year restrictions were applied. A citation search was performed in the reference lists of included articles, and the systematic and narrative reviews identified in the literature search. Selection process and eligibility criteria All identified articles were imported into the systematic review software Covidence28 for duplicate removal, selection process and data extraction. The inclusion criteria were (1) full-text, original, peer-reviewed randomised controlled trials (RCTs), cluster RCTs, crossover-RCTs and quasi-experimental studies; (2) articles including: (a) male and/or female athletes of any age, on all competitive levels and/or healthy general population; (b) exercise interventions of any modality and delivery method; (c) control groups that received interventions which did not include cervical exercises, and/or received no intervention and (d) outcome measures of SRC incidence and/or primary kinematic parameters (linear and rotational head accelerations) and/or a potential modifiable factor related to the prevention of SRC (isometric neck muscle strength). The exclusion criteria were (1) publications in languages other than English; (2) abstract-only publications, review articles, cohort studies, case–control studies, cross-sectional studies and/or case reports; (3) articles including (a) athletes and/or general population with symptomatic acute or chronic neck or head injuries at the time of enrolment and (b) cognitive training, protective equipment and/or policy/rule changes aimed at outcomes related to prevention of SRC. Data collection process Data extraction included (1) general study information; (2) number and characteristics of study participants; (3) description of the intervention arm; (4) description of the control arm and (5) change from baseline values of measured outcomes (details available in online supplemental file 3). Any missing data were sought out from the authors via email. The participants were classified as either athletes or healthy individuals from the general population based on their enrolment status in the respective studies. Those enrolled in a sport-related capacity at any competitive level (eg, elite, amateur and recreational) were classified as athletes. Conversely, individuals enrolled in non-athletic capacities (eg, military personnel, university students) were categorised as part of the healthy general population. If articles did not provide SD of change from baseline, these values were estimated using the formula: SDchange=√SD2baseline+SD2final–(2r×SDbaseline× SDfinal) as per guidelines in the Cochrane Handbook for Systematic Reviews of Interventions.29 If values of isometric neck strength were presented separately for multiple directions of movement (eg, flexion, extension and lateral flexion), a composite score was calculated (sum of change from baseline values for all reported directions divided by the number of reported directions). Composite scores were presented alongside mean SDs which were calculated using the formula: SDmean=√(s12+s22+…+sk2)/k (sk–SD for kth group; k–total number of groups). If articles with two or more intervention arms met the inclusion criteria, the change from baseline values for the intervention arm that was most applicable in real-world sport settings and/or the intervention group that best matched the control group were extracted. The first author (BI) categorised resistance training (RT) into various modalities (eg, dynamic, static and multimodal resistance programmes) based on neck muscle action.30 Adolescent participants were defined as younger than 19 years old while adult participants were 19 years and older. This age threshold was chosen based on standard organisational practices within the sports included in this review, that is, American football, soccer and rugby. In these sports, the under-19 age group typically represents the last youth category before athletes’ transition to senior competition. Risk of bias assessment The risk of bias assessment was performed using the Cochrane Risk-of-Bias tool for randomised trials (RoB 2)31 for every outcome across all articles in five domains: (1) bias arising from the randomisation process; (2) bias due to deviations from intended interventions; (3) bias due to missing outcome data; (4) bias in measurement of the outcome and (5) bias in the selection of the reported result. Based on the assessed bias in these domains, an overall risk-of-bias judgement was made for each outcome as follows: overall ‘low risk’ (all domains were at ‘low risk’); overall ‘some concern’ (no more than two ‘some concern’ domains and no ‘high-risk’ domains); overall ‘high risk’ (three or more ‘some concern’ domains or one or more ‘high-risk’ domains). Effect measures and synthesis methods Meta-analyses and subgroup analyses were performed by using the Cochrane Review Manager Web (RevMan Web, V.4.26.0).32 Meta-analysis was performed for all outcomes, where more than one article measured the same outcome. To explore the sources of heterogeneity and variations in the exercise efficacy, subgroup analyses were performed on outcomes for which an adequate number of articles was identified. Subgroups were based on (1) Exercise intervention modality—dynamic, static and multimodal resistance programmes; (2) Age—adolescent and adult participants and (3) Sex—male and female participants. Dichotomous outcomes were analysed with a random effect Mantel-Haenszel method, presenting risk ratio (RR) alongside 95% CIs. Continuous outcomes were analysed with a random effect inverse-variance method, presenting standardised mean difference (SMD) alongside 95% CIs. SMD effect size was interpreted as small (SMD=0.2), medium (SMD=0.5) and large (SMD=0.8).33 Random effect models were used, as moderate to substantial heterogeneity was expected due to differences in characteristics of the included articles. I2 statistics were used to assess the heterogeneity among articles. I2 statistics were interpreted using rough thresholds reported in the Cochrane Handbook for Systematic Reviews of Interventions: 0%–40% (might not be important); 30%–60% (may represent moderate heterogeneity); 50%–90% (may represent substantial heterogeneity) and 75%–100% (considerable heterogeneity).34 For subgroups in which statistical heterogeneity was detected, post hoc meta-regression was performed in order to relate specific study-level variables to the statistical heterogeneity. Meta-regression was performed using Stata (V.18, 2023).35 Sensitivity analyses were conducted by excluding articles with (1) healthy general population (2) cluster randomised and quasi-experimental designs and (3) overall ‘high risk’ of bias. Reporting bias and certainty assessments For outcomes with 10 or more included articles in the meta-analysis, potential reporting bias was assessed both visually with the examination of funnel plots and statistically using Egger’s test. Certainty of evidence was assessed by using the GRADEprofiler Guideline Development Tool software36 and by following the guidelines in the GRADE Handbook.37 The starting quality of evidence was rated as ‘high’ and could be downgraded one or two levels for each of the following domains: (1) risk of bias; (2) imprecision; (3) inconsistency; (4) indirectness and (5) publication bias. Reporting completeness of the exercise interventions The CERT38 was used to assess reporting completeness of the exercise interventions. CERT consists of 16 items over 7 sections: what (materials), who (provider), how (delivery), where (location), when, how much (dosage), tailoring (what, how) and how well (planned, actual). Each item was scored individually as a ‘0’ (not/inadequately described) or ‘1’ (adequately described). The resulting score ranges from 0 to 19, with a higher score indicating a more detailed description. Equity, diversity and inclusion statement The author group is composed of one male and three female researchers and represents a combination of junior (one author), mid-career (one author) and senior (two authors) researchers from two academic institutions in Sweden. This systematic review included athletes and a healthy general population from any nationality, sex, gender, age and competitive level. Analyses included considerations of age, sex and gender. However, due to the absence of gender-specific data in the articles, analyses based on gender were not possible. Results Study selection The systematic search yielded a total of 15 931 articles. Of these, 40 full-text articles were screened according to the inclusion/exclusion criteria, and 14 were further excluded (reasons for exclusion available in online supplemental file 4). In addition, seven articles were identified through the citation search of included articles, systematic and narrative reviews, of which all seven were excluded as they failed to fulfil inclusion criteria (online supplemental file 4). Finally, 26 articles3964 met all inclusion criteria and were included in the review (figure 1). Study identification and selection PRISMA flow diagram. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses. Study characteristics Detailed study characteristics are available in online supplemental file 5. 20 articles394446 47 49 50 52 included athletes in various sports while 6 articles45 48 51 56 57 59 included healthy general population. Sample sizes across articles ranged from 10 to 2452 (median n=32), and the median age was 20 years (range 10.8–38.6 years). In total, 96% of the participants were males and 4% were females. 10 articles evaluated the efficacy of dynamic resistance programmes.4547 51 53 5761 63 All 10 evaluated the efficacy on isometric neck strength while 2 also evaluated the efficacy on linear and rotational head accelerations.47 63 Six articles evaluated the efficacy of multimodal resistance programmes on isometric neck strength.39 44 48 54 56 64 Five articles evaluated the efficacy of static resistance programmes.40 43 46 49 55 All five evaluated the efficacy on isometric neck strength while one also evaluated the efficacy on linear head acceleration.55 Three articles evaluated the efficacy of neuromuscular warm-up programmes (balance, resistance, plyometric and sport-specific landing and cutting exercises) on SRC incidence.41 42 50 All neuromuscular warm-up programmes included specific neck exercises. One article evaluated the efficacy of a multimodal resistance and mobility programme on isometric neck strength.52 Additionally, one article evaluated a dynamic resistance and skill programme on isometric neck strength, and linear and rotational head accelerations.62 No studies assessed the impact of RT on SRC incidence. Furthermore, no articles assessed the efficacy of exercise interventions on concussion incidence, or linear and rotational head accelerations in the healthy general population, specifically. The median duration of the exercise interventions was 8 weeks (range 4–42 weeks), and the median number of training sessions per week was 3 (range 1–5 sessions). Intervention details are available in online supplemental file 6. Risk of bias 13 articles (50%) had high risk of bias, 12 articles (46%) had some concern and 1 article (4%) had overall low risk of bias. Inadequate randomisation process was the main source of bias in 27% of the articles, inadequate measurement of the outcome in 12% and deviations from the intended intervention in 4% of the articles (table 1). Table 1Risk of bias assessment (RoB 2 tool)Author and publication yearDomain 1: Randomisation processDomain 2: Deviations from intended interventionsDomain 3: Missing outcome dataDomain 4: Measurement of the outcomeDomain 5: Selection of the reported resultsOverall risk-of-biasAttwood et al40, 2022 Attwood et al41, 2018 Barden et al42, 2022 Barrett et al43, 2015 Becker et al44, 2019 Burnett et al45, 2005 Deng et al46, 2022 Eckner et al12, 2018 Fiebert et al48, 2004 Geary et al49, 2014 Hamlin et al39, 2020 Hislop et al50,2017 Leggett et al51, 1991 Maconi et al52, 2016 Mansell et al53, 2005 Müller et al54, 2021 Peek et al55, 2022 Pollock et al56, 1993 Salmon et al57, 2013 Stump et al58, 1993 Taylor et al59, 2006 Tsuyama et al60, 2006 Versteegh et al61, 2020 Wahlquist et al62, 2022 Waring et al63, 2022 Wilson et al64, 2021 RoB 2 criteria: +=indicates a low risk of bias; ?= Indicates that there are some concerns; –=indicates a high risk of bias.There were no differences in risk of bias between outcomes within trials.RoB 2Risk-of-Bias tool Risk of bias assessment (RoB 2 tool) RoB 2 criteria: +=indicates a low risk of bias; ?= Indicates that there are some concerns; –=indicates a high risk of bias.There were no differences in risk of bias between outcomes within trials. RoB 2Risk-of-Bias tool Synthesis of results Three articles did not provide baseline or follow-up data to be included in the meta-analyses.51 52 60 Meta-analyses of between 2 and 20 articles (n=37–4466) were performed separately for the four outcomes. A non-significant effect size was observed for neuromuscular warm-up programmes on SRC incidence (RR 0.69; 95% CI 0.39 to 1.23; I2 69%; low-quality evidence), and for RT on linear head acceleration (SMD −0.43; 95% CI −1.26 to 0.40; I2 63%; very low-quality evidence) or rotational head acceleration (SMD 0.08; 95% CI −0.61 to 0.77; I2 0%; low-quality evidence) (online supplemental file 7, figures 1–3). A large effect size was observed for RT on isometric neck strength (SMD 0.85; 95% CI 0.57 to 1.13; I2 62%; high-quality evidence) (figure 2, table 2). Forest plot of the efficacy of resistance training on isometric neck strength. Table 2GRADE summary of findingsOutcomesNo of participants(articles)Risk of biasInconsistencyIndirectnessImprecisionPublication biasSummary of findingsCertaintyPooled effect95% CISRC incidence4466(3 articles)X–X–XRR 0.690.39 to 1.23⊕⊕◯◯Low*†Linear head acceleration89(3 articles)––X–XSMD −0.43−1.26 to 0.40⊕◯◯◯Very low‡§¶Rotational head acceleration37(2 articles)–XX–XSMD 0.08−0.61 to 0.77⊕⊕◯◯Low**††Isometric neck strength638(20 articles)XXXXXSMD 0.850.57 to 1.13⊕⊕⊕⊕HighGRADE criteria: X=no serious limitations; –=serious limitation.Bolded values represent the observed effect sizes of exercise interventions on the outcome measures.*Downgraded one level for inconsistency due to substantial unexplained heterogeneity (I2=69%).†Downgraded one level for imprecision due to wide CIs which crossed both large benefits, and no benefits (95% CI 0.39 to 1.23).‡Downgraded one level for risk of bias in included articles.§Downgraded one level for inconsistency due to substantial unexplained heterogeneity (I2=63%).¶Downgraded one level for imprecision due to wide confidence intervals which crossed both large benefits and no benefits (95% CI –1.26 to 0.40).**Downgraded one level for risk of bias in included articles.††Downgraded one level for imprecision due to wide confidence intervals which crossed both moderate benefits, and no benefits (95% CI –0.61 to 0.77).GRADEGrading of Recommendations Assessment, Development and EvaluationRRrisk ratioSMDstandardised mean differenceSRCsport-related concussion GRADE summary of findings GRADE criteria: X=no serious limitations; –=serious limitation. Bolded values represent the observed effect sizes of exercise interventions on the outcome measures. Downgraded one level for inconsistency due to substantial unexplained heterogeneity (I2=69%). Downgraded one level for imprecision due to wide CIs which crossed both large benefits, and no benefits (95% CI 0.39 to 1.23). Downgraded one level for risk of bias in included articles. Downgraded one level for inconsistency due to substantial unexplained heterogeneity (I2=63%). Downgraded one level for imprecision due to wide confidence intervals which crossed both large benefits and no benefits (95% CI –1.26 to 0.40). Downgraded one level for risk of bias in included articles. Downgraded one level for imprecision due to wide confidence intervals which crossed both moderate benefits, and no benefits (95% CI –0.61 to 0.77). GRADEGrading of Recommendations Assessment, Development and EvaluationRRrisk ratioSMDstandardised mean differenceSRCsport-related concussion The three articles which were not included in the meta-analyses reported statistically significant increases in isometric neck strength, following the exercise interventions.51 52 60 Two articles reported a significant increase in isometric neck extension strength in six out of eight51 and four out of eight tested angles,60 respectively. One article reported a significant increase in isometric neck flexion, extension and rotation strength.52 Subgroup and meta-regression analyses Subgroup analyses were performed for isometric neck strength, while meta-regression was performed for seven subgroups where statistical heterogeneity was detected. Exercise modality Large effect sizes were observed for dynamic and multimodal resistance programmes on isometric neck strength (SMD 0.86; 95% CI 0.46 to 1.26; I2 34% and SMD 1.13; 95% CI 0.64 to 1.62; I2 66%, respectively). In contrast, a medium effect size was observed for static resistance programmes (SMD 0.74; 95% CI 0.32 to 1.16; I2 44%) (online supplemental file 7, figure 4). Heterogeneity in the dynamic resistance group was reduced from 34% to 12% when the number of sessions per week was used as a covariate, however, without statistically significant regression coefficient (slope 0.26; 95% CI −0.02 to 0.55) (online supplemental file 8, figure 1). Heterogeneity in the multimodal (I2 66%) and static (I2 44%) resistance groups could not be explained by any participant or intervention characteristics factors as covariates. Age A large effect size was observed for RT on isometric neck strength in adult participants (≥19 years) (SMD 1.19; 95% CI 0.53 to 1.84; I2 68%), whereas, a medium effect size was observed in adolescent participants (<19 years) (SMD 0.73; 95% CI 0.31 to 1.16; I2 66%) (online supplemental file 7, figure 5). Heterogeneity in adolescent participants was reduced from 66% to 45% when age was used as a covariate, with statistically significant regression coefficient (slope 0.20; 95% CI 0.05 to 0.36). Heterogeneity in adult participants (I2 68%) could not be explained by any participant or intervention characteristics factors as covariates (online supplemental file 8, figure 2). Sex A large effect size was observed for RT on isometric neck strength in male participants (SMD 0.89; 95% CI 0.59 to 1.19; I2 43%), whereas, a medium effect size was observed in female participants (SMD 0.65; 95% CI 0.06 to 1.24; I2 63%) (online supplemental file 7, figure 6). Heterogeneity in male (I2 43%) and female (I2 63%) groups could not be explained by any participant or intervention characteristics factors as covariates. Sensitivity analyses The pooled effect estimates remained similar for all sensitivity analyses performed (online supplemental file 9). Reporting biases Visual and statistical assessments were performed for isometric neck strength, where no evidence of potential publication bias was found (Egger’s test: p=0.16) (online supplemental file 10). Reporting completeness of the exercise interventions The final CERT scores ranged from 4 to 16 with a median of 11.5 (IQR 9–13). All 26 articles reported on ‘generic or tailored’ (item 14a) while no articles reported on ‘motivation strategies’ (item 6). ‘Rules for exercise progression’ (items 7a) and ‘starting levels’ (item 15), which are important for the accurate replication and implementation of exercise interventions, were reported in 13 (50%) and 10 (38%) articles, respectively (table 3). Table 3Reporting completeness of the exercise interventions (CERT) assessmentWhat (exercise equipment)2. Who (qualifications)How (individual or group)How (supervision)How (adherence measuring and reporting)How (motivation strategies)7 a. How (progression rules)7b. How (programme progression)How (exercises)How (home component)How (non-exercise component)How (adverse events)Where (setting)When, how much (exercise description)14a. Tailoring (generic or tailored)14b. Tailoring (description of tailoring)Tailoring (starting level)16a. How well (adherence assessment)16b. How well (delivery as planned)Total scoreAttwood et al4020221111101110111111001 15 Attwood et al4120180111100101001011001 10 Barden et al4220220111100101001011011 11 Barrett et al4320151101001110110111100 12 Becker et al4420191000000110100111001 8 Burnett et al4520051001001110100111000 9 Deng et al4620221010101110111111011 14 Eckner et al1220181101001111010111001 12 Fiebert et al4820041011001111000111011 12 Geary et al4920141111000011001110000 9 Hamlin et al3920201111001110101111100 13 Hislop et al5020170111100101001011011 11 Leggett et al5119911000000111001111100 9 Maconi et al5220161001101110001111111 13 Mansell et al5320051111101101000111101 13 Müller et al5420211010100011011111011 12 Peek et al5520221011000010101111000 9 Pollock et al5619930001001100000111100 7 Salmon et al5720131101101111101111111 16 Stump et al5819931000000010001010000 4 Taylor et al5920061101101100101111101 13 Tsuyama et al6020060000001100100111100 7 Versteegh et al61 20201111100101011111111 15 Wahlquist et al62 20221011000100001111000 8 Waring et al6320221101000111001111000 10 Wilson et al6420211011100111011111011 14 Total score 21 13 14 20 12 0 13 22 17 13 10 7 18 22 26 24 10 9 15 CERT criteria: 1=adequately described; 0=not/inadequately described.CERTConsensus on Exercise Reporting Template Reporting completeness of the exercise interventions (CERT) assessment CERT criteria: 1=adequately described; 0=not/inadequately described. CERTConsensus on Exercise Reporting Template Discussion This systematic review and meta-analysis included 26 articles that evaluated the efficacy of exercise interventions of different modalities on outcomes related to SRC. Non-significant effect sizes were observed for neuromuscular warm-up programmes on SRC incidence, and for RT on linear and rotational head accelerations, with certainty of evidence ranging from very low to low. A large effect size was observed for RT on isometric neck strength, with 20 articles providing high certainty of evidence. The results of the CERT assessment suggest that exercise interventions were generally not reported in sufficient detail to allow accurate replication in future studies. SRC incidence The result of the meta-analysis indicated that neuromuscular warm-up programmes did not have a significant effect on SRC incidence. There could be several reasons for this. Although similar in study population (male rugby athletes) and intervention content (neuromuscular warm-up exercises), the included articles varied considerably in intervention duration (14 weeks vs 42 weeks), and definition of SRC injury (>24 hours vs >8 days). This likely contributed to the substantial statistical heterogeneity (I2 69%) and differences in effect sizes (mean RR range: 0.36–1.11). Another possible explanation for the observed lack of efficacy could be due to the dose–response relationship. While one article reported a 60% reduction in SRC incidence in the intervention group when compared with the control group,41 another article reported a similar reduction in SRC incidence (59%), but only when analyses included teams that implemented the programme at least three times a week.50 The authors of both articles reasoned that the observed reduction in SRC incidence was due to the inclusion of neck strengthening exercises within the neuromuscular warm-up programmes at all stages of the intervention period.41 50 As the association between greater neck muscle strength and reduced SRC incidence is not definitively established, it is difficult to evaluate whether neck muscle exercises alone could produce such results. It is possible that the inclusion of neck muscle exercises within multimodal injury prevention programmes could yield more beneficial results in reducing SRC than isolated neck muscle exercises alone; however, this remains to be investigated in future studies. Linear and rotational head accelerations The results of the meta-analyses indicated that RT did not have a significant effect on linear and rotational head accelerations. Limited data for linear head acceleration (n=3 articles; 89 participants) and rotational head acceleration (n=2 articles; 37 participants) contributed to the overall uncertainty, making it difficult to draw robust conclusions about the efficacy of exercise interventions on these outcomes. In addition, only one article provided a sample size calculation while the remaining articles were likely underpowered. Therefore, future research endeavours should examine these outcomes in greater detail. Isometric neck strength As all included articles on isometric neck strength implemented exercise interventions in the form of RT, the effect sizes observed in this review further reinforce the well-established benefits of RT on muscle strength.65 The change in muscle strength occurs due to a number of neuromuscular adaptations induced by RT.65 Neurologically, RT leads to the recruitment of an increased number and firing rate of motor units,66 heightened reflex potentiation67 and improved synchronisation,68 whereas muscular adaptations involve an increase in a cross-sectional area of the muscle65 and selective hypertrophy of fast-twitch fibres.69 70 Although the articles included in this review implemented RT with large differences in exercise prescriptions (eg, exercise modality, frequency, duration, intensity), meta-analysis showed an overall large effect size for RT on isometric neck strength. When separated by the exercise modality, all three modalities were efficacious in increasing isometric neck strength, with large effect sizes following dynamic and multimodal resistance programmes, and a medium effect size following static resistance programmes. This indicates that training-induced adaptations, and subsequent neck strength gains, can be achieved with a variety of RT prescriptions. While RT was efficacious in increasing isometric neck strength in both males and females, the articles in this analysis predominantly included male participants, and the comparatively lower effect size observed in females may be attributed to the evident disparity in the representation of participants. Given the well-established benefits of RT on overall muscle strength in females,7173 there is a compelling need to enhance the involvement of this demographic in research aimed at exploring outcomes related to prevention of SRC. This would contribute to greater equality in participant representation and result in a more accurate estimation of the effect sizes in the measured outcomes. Similarly, the comparatively lower effect size for RT in adolescent participants when compared with adults cannot be definitively attributed to differences in physiological adaptations to RT between these two age groups. It is likely that the observed discrepancy is due to significant variations in RT prescriptions across all articles included in the analysis. Therefore, future research should align RT protocols with current recommendations on RT prescriptions7476 in order to maximise strength gains in neck musculature. It is important to note that as the definitive association between greater neck muscle strength and reduced SRC risk remains uncertain, the efficacy of RT on isometric neck strength observed in this review does not inherently imply a reduced risk of SRC. To substantiate such a claim in the future, trials directly investigating the efficacy of neck strength interventions on SRC incidence are warranted. Reporting completeness of the exercise interventions Generally, exercise interventions were not reported in sufficient detail. No articles described ‘motivational strategies’ (item 6). Motivation is an important factor for exercise adherence in athletes,77 78 therefore, the lack of information on motivational strategies could limit the efficacy of otherwise well-designed exercise interventions. Another commonly under-reported aspect was ‘starting level’ (item 15), which was lacking in 62% of articles. Since athletes have varying fitness levels, and abilities, accurately estimating the starting level can prevent situations where the initial exercise stimulus is too challenging for some, and not challenging enough for others. Also, only 50% of articles adequately described ‘rules for exercise progression’ (item 7a), which is essential for stimulating continuous neuromuscular adaptations.79 Thus, the reporting of important aspects of exercise interventions is consistently insufficient, therefore, future research should aim to improve reporting completeness in order to promote study replication and facilitate further implementation. Strengths and limitations of this review To our knowledge, this systematic review is the first to evaluate the efficacy of exercise interventions not only on SRC incidence but also on other outcomes potentially related to SRC. Throughout the review process, the authors diligently adhered to established guidelines to enhance transparency and the overall credibility of the systematic review. This review has some limitations to acknowledge. First, the combination of neck strength directions into a single composite score allowed for a comprehensive evaluation of overall neck strength and facilitated the interpretation of the data. Moreover, as some articles measured isometric neck strength in non-traditional directions (eg, anterolateral neck flexion, posterolateral extension) by calculating composite scores, these values were also included in the analyses. Nevertheless, this approach may have obscured asymmetries in strength between different directions, potentially overlooking findings related to muscle balance and function. Furthermore, the composite score may have reduced the sensitivity to changes in specific areas of neck strength over time or in response to interventions. Second, moderate to substantial heterogeneity was observed in the majority of meta-analyses and subgroup analyses. In some subgroups, the observed heterogeneity could not be explained by any participant or intervention characteristics included as covariates, and may be related to other relevant study-level characteristics not available for extraction. Moreover, as the post hoc meta-regression was conducted with a limited number of articles, generalisability and robustness of the results should be interpreted with caution. Third, the RoB2 tool was originally developed for RCTs, therefore, its applicability in quasi-experimental studies may be limited. Since, in this review, the RoB2 tool was employed to evaluate the methodological quality of all included articles, irrespective of study design, this could result in high-quality articles being downgraded due to criteria that are not tailored to their design. Fourth, this review only included articles published in English, which may have introduced language bias due to the exclusion of any non-English articles relevant to the topic. Implications for future research Several key areas were identified in this review that warrant attention for future research. More adequately powered and rigorous trials are needed, particularly for investigating the efficacy of exercise interventions on SRC incidence, and linear and rotational head accelerations. Given that less than 5% of participants were females, future studies should make greater efforts to ensure equality in participant representation. Since articles included in this review predominantly included athletes from rugby, soccer and American football, future research should broaden its scope beyond these sports to encompass other sports where SRC frequently occurs (eg, ice hockey, handball). Future exercise interventions should align with evidence-based exercise prescription recommendations (eg, exercise modality, duration, frequency and load), but still maintain a sufficient level of simplicity to facilitate adherence and easy implementation in real-world sport settings. Finally, it is essential that researchers improve reporting completeness of exercise interventions in order to advance the understanding of exercise interventions, promote study replication and facilitate further implementation. Conclusion This systematic review and meta-analysis demonstrated non-significant effects for neuromuscular warm-up programmes on SRC incidence, and for RT on linear and rotational head accelerations. Large effect size was observed following RT on isometric neck strength. Dynamic and multimodal resistance programmes seem to be more efficacious in increasing isometric neck strength than static resistance programmes, with higher effect sizes for males than females and adults than adolescents. It is important to note that the observed efficacy of RT on isometric neck strength does not necessarily translate to a reduced SRC risk. Therefore, more adequately powered and rigorous trials are needed to determine the efficacy of exercise interventions on SRC incidence, as well as on linear and rotational head accelerations. Researchers need to include more females, different sports with high risk of SRC and follow exercise prescription recommendations, as well as improve exercise reporting to allow for replication of research and implementation in real-world sport settings. supplementary material