Lower-limb biomechanical alterations across movement tasks in individuals with chronic ankle instability: An umbrella review of systematic reviews and meta-analyses Authors: Bojie Xuan, Dong Sun, Yang Song, Fengping Li, Zhanyi Zhou, Zilong Wang, Jiazhong Zhu, Kuiyu Chen, Rui Zhao, Dongxu Wang, Diwei Chen, Danica Janicijevic, Monèm Jemni, Yaodong Gu Published: 2026 ABSTRACT Chronic ankle instability (CAI) may reflect altered regulation of mechanical loading and neuromuscular control across the lower limb, but review-level evidence remains fragmented by task and outcome. This umbrella review synthesised systematic reviews and meta-analyses on biomechanical and neuromuscular alterations during movement tasks in individuals with CAI, functional or mechanical ankle instability, recurrent ankle sprain, or ankle-sprain copers. PubMed/MEDLINE, Scopus, Web of Science Core Collection, selected EBSCOhost databases, the Cochrane Library, and ScienceDirect were searched; 18 reviews were included, of which 17 informed the main synthesis and one contributed supplementary dual-task evidence. Evidence was most concentrated for landing and jump-landing tasks, where reduced pre-landing peroneal activation, greater or faster vertical loading, and stiffer sagittal-plane landing mechanics were commonly reported. Walking evidence suggested greater ankle inversion, hip adduction, and knee valgus moment, whereas perturbation landing evidence suggested delayed peroneus longus activation and greater inversion mechanics. Coper comparisons may suggest more preserved task-specific movement profiles after ankle sprain. Overall, the evidence base is sufficient to identify recurring patterns of altered lower limb biomechanics in CAI, but insufficient to support highly definitive conclusions for most task outcome domains.Protocol registrationPROSPERO CRD420261454010 Introduction Lateral ankle sprain is one of the most common musculoskeletal injuries in physically active and athletic populations.1, 2, 3, 4, 5 Although many individuals recover after an initial sprain, a substantial proportion continue to experience persistent symptoms and develop chronic ankle instability (CAI). CAI is characterised by recurrent ankle sprain, repeated episodes of giving way, perceived ankle instability, pain, reduced self-reported function, and limited participation in sport or physical activity.6, 7, 8 By contrast, ankle sprain copers sustain a lateral ankle sprain but do not develop recurrent giving way, persistent perceived instability, or substantial functional limitation, making them a clinically useful comparator for understanding preserved recovery after ankle sprain. CAI has traditionally been viewed as a disorder of the ankle joint itself. Local impairments such as ligamentous laxity, restricted dorsiflexion, altered peroneal muscle function, and reduced ankle strength are commonly described.9,10 However, proprioception and postural control depend on integrated sensorimotor regulation rather than purely local joint status. Impaired ankle sensory feedback, stability, or muscle activation may therefore alter joint motion, loading, muscle activity, and postural control at the foot, ankle, knee, hip, or trunk. In this review, biomechanical and neuromuscular alterations refer to changes in movement strategy, force absorption, joint control, and muscle activation during functional tasks.11,12 From this perspective, CAI may be better understood as a lower limb movement system condition than as an isolated ankle disorder. Movement task context is central when evaluating CAI related biomechanics. Walking and running may reveal locomotor adaptations during relatively low demand activity, whereas gait initiation challenges preparation for the transition from standing to stepping.13 Landing and jump landing tasks require preparation for ground contact, impact absorption, and stability, while perturbation landing adds the need to respond quickly to expected or unexpected inversion stress. Sport manoeuvres such as cutting, side cutting, and stop jump tasks require rapid deceleration, redirection, and frontal plane control.14,15 Different tasks may therefore expose different components of CAI related dysfunction, particularly when movement demands increase. A growing number of systematic reviews and meta-analyses have examined biomechanical or neuromuscular deficits in CAI and related ankle instability conditions.16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33 These reviews have addressed walking and running biomechanics,24,25 gait initiation,27,31 landing and jump-landing mechanics,18,20,22,28,29,33 perturbation landing,19 sport manoeuvres and proximal kinetic-chain adaptations,17,20,30 ankle sprain coper comparisons,21,26,32 functional-task electromyography,23 and dual-task gait or postural control.16 However, the evidence remains fragmented because most reviews focus on a single task, a specific biomechanical outcome domain, or a particular comparator group. It therefore remains difficult to determine which task contexts have the strongest evidence, which biomechanical alterations are most consistent, whether distal ankle impairments are accompanied by proximal kinetic chain changes, and how much confidence should be placed in the available evidence. An umbrella review is warranted because evidence on CAI biomechanics is now spread across several systematic reviews and meta-analyses that differ in task, outcome, comparator, and primary study inclusion. Without review level synthesis, apparent consistency across reviews may reflect repeated use of the same primary studies rather than independent confirmation. A task-based umbrella review can therefore clarify where evidence is strongest, where uncertainty remains, and how methodological quality, risk of bias, evidence certainty, and study overlap affect current conclusions. Therefore, this umbrella review aimed to systematically identify, synthesise, and appraise systematic reviews and meta-analyses examining task-specific mechanical loading, lower limb biomechanics, and neuromuscular regulation during movement tasks in individuals with CAI or related ankle instability conditions. From a mechanobiology in medicine perspective, this review considered how task-specific mechanical inputs and neuromuscular responses may be altered in CAI and how these features could inform biomechanical targets for prevention, rehabilitation, and clinical translation. Using a Population, Intervention/Exposure, Comparator, Outcomes and Study design (PICOS) framework,34,35 participants were individuals with CAI or related ankle instability conditions, including copers where relevant; the exposure was movement or functional tasks; comparators were healthy or non-CAI controls, copers, or unaffected limbs; outcomes were lower limb kinematic, kinetic, ground reaction force, electromyographic, postural stability, or neuromuscular measures; and eligible study designs were systematic reviews, meta-analyses, and reviews of reviews. Methods and materials Design and reporting framework This study was designed as an umbrella review of systematic reviews and meta-analyses. The review aimed to identify, synthesise, and appraise review level evidence on lower limb biomechanical alterations during movement and functional tasks in individuals with CAI or related ankle instability conditions. Eligible evidence syntheses included systematic reviews, meta-analyses, and reviews of reviews that examined kinematic, kinetic, ground reaction force, electromyographic, postural control, or other neuromuscular outcomes during movement tasks. Reviews of reviews were included for completeness of evidence mapping but were not used as primary sources for quantitative synthesis. When overlap with eligible systematic reviews or meta-analyses was identified, their findings were treated as supplementary evidence only and were excluded from corrected covered area calculations to avoid double counting of primary studies. The review was reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 statement, with adaptations where appropriate for an umbrella review design.36,37 Because the unit of inclusion was the published review rather than the individual primary study, data extraction, methodological appraisal, risk of bias assessment, certainty assessment, and primary study overlap assessment were conducted at the review or task outcome level. Findings from included meta-analyses were extracted as reported by the original authors, and primary study data were not statistically pooled again in the present umbrella review. Protocol and registration The protocol for this umbrella review was registered with the International Prospective Register of Systematic Reviews (PROSPERO CRD420261454010). The registered protocol prespecified the review question, eligibility criteria, outcomes of interest, search strategy, methodological quality appraisal, risk of bias assessment, certainty assessment, and planned approach to evidence synthesis. No substantive deviations from the registered protocol were made. Search strategy A systematic literature search was conducted in PubMed/MEDLINE, Scopus, Web of Science Core Collection, selected EBSCOhost databases, and the Cochrane Library from database inception to 1 Ma y 2026. ScienceDirect was searched as a supplementary publisher platform source to maximise the retrieval of potentially relevant review articles. No restrictions on publication year were applied. The search strategy was structured around three core concepts: ankle instability, biomechanical or neuromuscular outcomes, and systematic review or meta-analysis publication type. The ankle instability concept included terms related to CAI, functional ankle instability, mechanical ankle instability, recurrent ankle sprain, recurrent ankle sprains, and ankle instability. The biomechanical outcome concept included terms related to biomechanics, kinematics, kinetics, electromyography, EMG, muscle activation, ground reaction force, plantar pressure, center or centre of pressure, joint moment, joint angle, dynamic postural stability, postural control, and neuromuscular outcomes. These terms were combined with review filters related to systematic review, meta-analyses, umbrella review, and review of reviews. Database specific syntax, including truncation symbols and field tags adapted for each platform, was used where appropriate. Records retrieved from all databases and supplementary sources were exported for deduplication and screening. The full database specific search strategies and numbers of records retrieved are provided in Supplementary Table S1. Eligibility criteria Eligibility criteria were defined according to population, review type, movement task, biomechanical or neuromuscular outcome, and comparator. Reviews were considered eligible if all of the following criteria were met. Eligible reviews included participants with CAI or related ankle instability conditions, including functional ankle instability, mechanical ankle instability, recurrent ankle sprain, recurrent ankle sprains, ankle instability, or ankle sprain copers when comparisons with CAI or ankle instability groups were reported. Reviews involving mixed musculoskeletal populations were eligible only when ankle instability specific data could be identified or interpreted separately. Eligible publications were systematic reviews, meta-analyses, or reviews of reviews that synthesised primary studies relevant to CAI or related ankle instability conditions. Reviews with or without quantitative meta-analyses were eligible, provided that systematic methods had been used for literature searching, study selection, and evidence synthesis. Reviews were eligible when lower limb biomechanics or neuromuscular control was examined during movement or functional tasks. Eligible tasks included walking, running, gait initiation, landing, jump landing, expected or unexpected perturbation landing, cutting, side cutting, stop jump, dynamic balance, sport manoeuvres, and other functional tasks requiring lower limb movement control. Eligible outcomes included lower limb kinematics, kinetics, joint angles, joint moments, ground reaction forces, loading variables, plantar pressure variables, centre of pressure measures, dynamic postural stability, electromyography, muscle activation, muscle onset timing, co-contraction, neuromuscular control, and proximal kinetic-chain variables and biomechanical variables that have been discussed in relation to potential anterior cruciate ligament injury mechanisms. Eligible comparisons included CAI or ankle instability groups versus healthy or non-CAI controls, CAI versus ankle sprain copers, copers versus controls, and affected versus unaffected limb comparisons in individuals with CAI, provided that these comparisons were reported within eligible movement task contexts. Exclusion criteria Publications were excluded if they were narrative reviews, reviews without systematic methods, scoping reviews without systematic evidence synthesis, protocols, editorials, commentaries, conference abstracts with insufficient extractable information, original experimental studies, randomised trials, observational primary studies, case reports, or other publications that were not review articles. Reviews were also excluded when their primary focus fell outside the scope of movement task lower limb biomechanics in CAI or related ankle instability conditions. This applied to reviews focused exclusively on treatment effects, rehabilitation interventions, external ankle supports, bracing, taping, orthoses, pain, clinical outcome scores, isolated strength measures, muscle morphology, muscle architecture, static balance, static postural control, or isolated sensorimotor or neurophysiological outcomes without eligible movement task biomechanical or neuromuscular outcomes. Reviews involving broader ankle injury or musculoskeletal populations were excluded when CAI specific or ankle instability specific findings could not be separated. Reviews were also excluded when biomechanical outcomes could not be distinguished from functional performance outcomes that were not biomechanical. Full text screening decisions and exclusion reasons are provided in Supplementary Table S2. Study selection All records retrieved from database and supplementary searches were imported into Zotero for reference management and duplicate removal. Duplicate records were identified using the duplicate detection function in Zotero and were manually checked where necessary. After deduplication, two reviewers independently screened the titles and abstracts of all remaining records against the predefined eligibility criteria. Records that clearly failed to meet criteria relating to the review question, publication type, population, task, or outcome were excluded at this stage. The full texts of potentially eligible reports were then retrieved and independently assessed by the same two reviewers. Final inclusion decisions were made after full text assessment. Disagreements between the reviewers at either the title and abstract stage or the full text stage were resolved through discussion and consensus. The study selection process is reported in the Results section and summarised in Fig. 1. Full text screening decisions and exclusion reasons are provided in Supplementary Table S2.Fig. 1PRISMA 2020 flow diagram of study selection. PRISMA 2020 flow diagram of study selection. Data extraction Data were extracted using a predefined data extraction form. For each included review, data were independently extracted by two reviewers, and the extracted information was checked for accuracy. Any discrepancies were resolved through discussion and consensus. For each included review, the following information was extracted: first author, year of publication, journal, review type, databases searched, search end date, language restrictions, protocol registration status, number of included primary studies, participant characteristics, diagnostic or eligibility criteria for CAI or related ankle instability conditions, comparison groups, movement tasks, biomechanical or neuromuscular outcome domains, synthesis method, quality or risk of bias tool used in the original review, and main conclusions. For reviews with meta-analyses, outcome level quantitative findings were extracted as reported by the original authors. Extracted data included the effect metric, pooled effect estimate, 95% confidence interval, p value, heterogeneity statistics including I2 where available, number of included studies or effects, participant numbers where reported, and direction of effect.38,39 For reviews without meta-analysis, task specific findings and conclusions were extracted narratively and organised by movement task and biomechanical outcome domain. The present umbrella review did not reanalyse primary study data or pool them statistically again. Quantitative estimates were therefore interpreted as review level findings from the included meta-analyses rather than as newly generated pooled effects. Detailed outcome-level estimates extracted from the included meta-analyses were used for the quantitative summary presented in the Results section. Methodological quality The methodological quality of each included systematic review or meta-analysis was appraised using A Measurement Tool to Assess Systematic Reviews 2 (AMSTAR 2).40 AMSTAR 2 comprises 16 items that evaluate key methodological domains, including the definition of the research question and eligibility criteria, protocol registration, justification of eligible study designs, comprehensiveness of the literature search, duplicate study selection and data extraction, reporting of excluded studies, description of included studies, appraisal of primary study risk of bias, reporting of funding sources, appropriateness of meta-analytic methods, consideration of risk of bias in synthesis and interpretation, investigation of heterogeneity, assessment of publication bias, and reporting of conflicts of interest. Each AMSTAR 2 item was rated as yes (Y), partial yes (PY), no (N), cannot answer (CA), or not applicable (NA). Items related to meta-analytic methods, the influence of risk of bias on meta-analysis, and publication bias assessment were coded as NA for reviews that did not conduct a meta-analysis. Overall methodological confidence for each review was classified according to AMSTAR 2 guidance as high, moderate, low, or critically low. Particular attention was given to AMSTAR 2 critical domains, including protocol registration, search comprehensiveness, reporting of excluded studies, risk of bias assessment, appropriateness of meta-analysis, consideration of risk of bias in interpretation, and publication bias assessment.41 The AMSTAR 2 appraisal was independently completed by two reviewers. Disagreements were resolved through discussion and consensus. Item level AMSTAR 2 ratings and overall methodological confidence judgements are provided in Supplementary Table S3. Risk of bias The risk of bias in the included systematic reviews and meta-analyses was assessed using the Risk Of Bias In Systematic Reviews (ROBIS) tool 42, 43, 44. ROBIS was applied at the review level to evaluate whether limitations in review conduct could have biased the review findings. Four ROBIS domains were assessed. Domain 1 addressed eligibility criteria. Domain 2 addressed identification and selection of studies. Domain 3 addressed data collection and study appraisal. Domain 4 addressed synthesis and findings. Each domain was judged as low concern, some concerns, or high concern. An overall risk of bias judgement was then assigned for each review using the same categories. The ROBIS assessment was independently completed by two reviewers. Disagreements were resolved through discussion and consensus. Domain level ROBIS judgements and overall review level risk of bias assessments are provided in Supplementary Table S4. Certainty of evidence The certainty of evidence was assessed using a modified Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach.45,46 Because review level evidence was synthesised across heterogeneous movement tasks and biomechanical outcomes, certainty was assessed at the task outcome level rather than at the level of each included review. Each task outcome conclusion was treated as a clinically and biomechanically meaningful evidence unit, such as ankle inversion angle during walking, peroneal activation before landing, peroneus longus activation latency during perturbation landing, or hip strength deficits in CAI. For each task outcome domain, certainty was judged across five GRADE dimensions: risk of bias, inconsistency, indirectness, imprecision, and publication bias. Risk of bias reflected the methodological limitations of the included reviews and, where reported, the quality or risk of bias assessments of the underlying primary studies. Inconsistency reflected heterogeneity in effect estimates, direction of findings, movement task protocols, and outcome definitions. Indirectness reflected the applicability of the population, comparator, task, and biomechanical outcome to the review question. Imprecision reflected the number of included studies or effects, sample size, confidence interval width, and proximity of the estimate to the null effect. Publication bias was considered when it had been assessed in the included reviews. Otherwise, it was judged as unclear. Because the primary evidence base consisted predominantly of observational laboratory studies, the initial certainty for most task outcome domains was set at low. Certainty could be downgraded to very low when serious concerns were identified in one or more GRADE dimensions. Certainty was classified as high, moderate, low, or very low, although high or moderate certainty was not assigned unless it was clearly supported by the consistency, precision, directness, and methodological robustness of the evidence. The certainty of evidence for each task outcome domain was independently assessed by two reviewers. Disagreements were resolved through discussion and consensus. Summary certainty ratings are presented in Table 4, and the detailed GRADE rationale for each task outcome domain is provided in Supplementary Table S5. Primary-study overlap and corrected covered area Primary study overlap across included reviews was assessed using a citation matrix. For each included systematic review or meta-analysis, the list of primary studies was extracted, harmonised, and coded against all other included reviews. Primary study identifiers were standardised by author and publication year. When multiple studies shared the same first author and publication year but represented distinct primary studies, suffixes were added to preserve separate study identities. Corrected covered area (CCA) was calculated to quantify the degree of primary study overlap across reviews.47,48 CCA was calculated using the following formula:CCA=N−rr(c−1)×100%where N represents the total number of primary study occurrences across reviews, r represents the number of unique primary studies, and c represents the number of reviews included in the calculation. CCA was calculated for all included reviews, for the main synthesis after excluding the supplementary dual task review, and for task specific review clusters. Overlap was interpreted as slight when CCA was 0 to 5%, moderate when CCA was greater than 5% to 10%, high when CCA was greater than 10% to 15%, and very high when CCA was greater than 15%.49 The complete citation matrix, CCA summary, pairwise overlap analysis, primary study lists, and extraction notes are provided in Supplementary Table S6, which is submitted as a separate Excel supplementary file. Data synthesis Findings were synthesised narratively and organised according to movement task domain and biomechanical outcome domain. Movement task domains included walking and running, gait initiation, landing, jump landing, perturbation landing, sport manoeuvres and proximal kinetic chain adaptations, coper comparison and dynamic discrete tasks, functional task electromyography, and supplementary dual task gait and postural control evidence. Biomechanical outcome domains included kinematics, kinetics or joint moments, ground reaction force or loading variables, electromyography or muscle activity, centre of pressure or postural stability outcomes, and biomechanical variables related to proximal control or discussed in relation to potential anterior cruciate ligament injury mechanisms. Included reviews with meta-analyses contributed quantitative outcome level estimates as reported by the original review authors. The present umbrella review did not pool primary study data again, recalculate pooled effects, or combine effect estimates across reviews. Therefore, reported pooled estimates were interpreted as review level findings rather than newly generated meta-analytic results. When different effect metrics were reported across reviews, including standardised mean differences, weighted mean differences, Cohen's d, and Fisher's Z, these metrics were presented separately and were not directly compared in magnitude. Reviews without meta-analysis contributed narrative evidence and evidence map data. Findings were summarised by task outcome domain, with attention to the direction and consistency of reported biomechanical alterations, the number and type of supporting reviews, review level methodological quality, risk of bias, certainty of evidence, and primary study overlap. Key quantitative outcome level estimates are summarised in Table 3, and detailed outcome-level estimates extracted from the included meta-analyses were used for the quantitative summary presented in the Results section. Results Search results and study selection The search identified 847 records from database searches and supplementary sources. After 470 duplicate records had been removed, 377 records underwent title and abstract screening. Of these, 344 records were excluded because they did not meet the predefined eligibility criteria. The remaining 33 reports were retrieved and assessed in full text. Following full text assessment, 18 reviews were included in the umbrella review.16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33 Of these, 17 contributed to the main synthesis of lower limb biomechanics during movement tasks, and one was retained for a supplementary narrative synthesis of dual task gait and postural control evidence. The study selection process is presented in Fig. 1. Full text screening decisions and reasons for exclusion are provided in Supplementary Table S2. Characteristics of included reviews The included reviews were published from 2017 to 2026 and comprised systematic reviews with and without meta-analysis. Together, they covered movement tasks and biomechanical outcome domains relevant to CAI, functional ankle instability, mechanical ankle instability, recurrent ankle sprain, ankle instability, and ankle sprain copers where relevant. Their characteristics are summarised in Table 1.Table 1Characteristics of included systematic reviews and meta-analyses.Table 1ReferenceReview typeSearch sources/date as reportedIncluded primary studies/samplePopulation/comparisonMovement task(s)Biomechanical outcome domain(s)Moisan et al., 2017Systematic reviewPubMed, Embase, CINAHL, AMED, SPORTDiscus; Jan-Feb 2016; English/French24 studies; 8 muscle-activity, 14 kinematic, and 7 kinetic studiesCAI vs healthy controlsWalking; runningKinematics; kinetics; EMG/muscle activityLuan et al., 2024Systematic review and meta-analysisPubMed, Embase, Cochrane Library, Web of Science, EBSCO, PEDro; to 30 Jun 2023; updated 3 Ma y 2024; no language restriction13 studies; 729 participantsCAI/ankle instability vs non-CAI controls or copersWalkingKinematics; kineticsYousefi et al., 2023Systematic reviewPubMed, ScienceDirect, Scopus, Web of Science, Google Scholar; searched 1 Jun-18 Jul 2022; English6 studiesCAI vs healthy controlsGait initiationCOP/APAs; kinematics; EMG/muscle activityMortezanejad et al., 2024Systematic reviewPubMed, ScienceDirect, Web of Science, Google Scholar, Scopus; 1990-20235 studiesCAI vs healthy controls; affected vs non-affected leg comparisonsGait initiationKinetics/COP; EMG/muscular measuresTheisen & Day 2019Systematic reviewSPORTDiscus, PubMed, CINAHL; search terms included ankle instability and landing kinematics6 studies; 338 participantsCAI vs controlsLandingKinematics; proximal-control interpretationSimpson et al., 2019Systematic reviewPubMed, ScienceDirect, Scopus, SPORTDiscus; to 31 Jan 2018; reference lists searched20 articlesCAI vs non-CAI controlsUnilateral landing; jump-landingKinematics; GRF/loading; EMG/muscle activity; dynamic postural stabilityJeon et al., 2021Systematic review and meta-analysisPubMed, CINAHL, SPORTDiscus, Web of Science; to Dec 2019; English; peer-reviewed11 studies in meta-analysisAI/CAI/FI/MI/recurrent ankle sprain vs healthy/normal controlsLandingEMG/muscle activation; GRF/loadingChan et al., 2022Systematic review and meta-analysisPubMed, MEDLINE, Embase, CINAHL; searched 26 Sep 202121 studiesCAI vs healthy controlsUnilateral jump-landingKinematics; EMG/muscle activity; dynamic postural stabilityHakimi Poor et al., 2022Systematic reviewGoogle Scholar, ScienceDirect, PubMed, Scopus, ProQuest, Cochrane Library; to 2021; English15 articlesCAI/ankle instability participants; controls where reportedUnilateral jump-landingKinematicsHe et al., 2024Systematic review and meta-analysisWeb of Science, Scopus, PubMed, SPORTDiscus, Cochrane Library; May 202414 studies; 564 participantsCAI vs healthy controlsLanding on flat surfaceKinematics; kinetics/joint torques; GRF/loading; proximal-control/potential ACL-mechanism variablesGong et al., 2025Systematic review and meta-analysisPubMed, Embase, Cochrane Library, Web of Science, Scopus; to 30 Nov 202413 studies; 207 CAI and 215 controlsCAI vs healthy controlsExpected/unexpected inversion-perturbation landingKinematics; kinetics; EMG/muscle activityXu et al., 2022Systematic reviewPubMed, Web of Science, SPORTDiscus; inception to Apr 202232 studies: 5 side-cutting, 3 stop-jump, 24 landingCAI vs healthy controlsSide-cutting; stop-jump; landingKinematics; kinetics/joint moments; proximal-control/potential ACL-mechanism biomechanicsDeJong et al., 2020Systematic review and meta-analysisMEDLINE/PubMed and CINAHL; updated to 3 Jun 201955 studiesCAI/FI vs healthy controlsStrength, balance, jumping, gait, and other functional assessmentsProximal neuromuscular/biomechanical outcomes; kinematics; EMG; strengthYu et al., 2022Systematic reviewScienceDirect, PubMed, Web of Science; to 30 Nov 202023 articlesCAI vs copersLocomotor and functional tasksKinematics; kinetics; dynamic postural controlJeon et al., 2025Systematic review and meta-analysisPubMed, CINAHL, SPORTDiscus, Web of Science; inception to Aug 202213 studies; 227 copers, 291 CAI, 159 controlsCopers vs CAI and controlsLanding; walking/runningKinematics; kineticsMohamadi et al., 2026Systematic review and meta-analysisPubMed, Scopus, ScienceDirect, MEDLINE, Web of Science; inception to Mar 202410 studiesCAI vs copers vs healthy subjectsSEBT; jump landing; side cutting; Biodex dynamic balanceNeuromuscular control; kinematics; kinetics; muscle activity; postural controlLabanca et al., 2021Systematic reviewPubMed, Web of Science, Cochrane Library; studies published 2000-202063 articlesCAI vs healthy controlsGait; running; perturbations; landing/hopping; cutting/turning; stance; SEBT; lunges; kicking; Y-balance; squattingEMG/muscle activationChoi et al., 2023Systematic review and meta-analysisPubMed, MEDLINE, CINAHL, SPORTDiscus; inception to Oct 20229 studies in systematic review; 5 in meta-analysisCAI vs healthy controlsDual-task gait and postural controlPostural control; gait; cognitive-motor/dual-task performanceNote: AI, ankle instability; APAs, anticipatory postural adjustments; CAI, chronic ankle instability; COP, centre of pressure; EMG, electromyography; FI, functional instability; GRF, ground reaction force; MI, mechanical instability; SEBT, Star Excursion Balance Test. Study counts and samples are presented as reported by the original reviews. For He et al., 2024, the Results section reported 14 studies and 564 participants, whereas the abstract reported 13 studies and 542 participants; the Results-section value was retained for consistency with the study list and the overlap analysis. For Moisan et al., 2017, outcome-category counts are not mutually exclusive. Table 1Characteristics of included systematic reviews and meta-analyses.Table 1ReferenceReview typeSearch sources/date as reportedIncluded primary studies/samplePopulation/comparisonMovement task(s)Biomechanical outcome domain(s)Moisan et al., 2017Systematic reviewPubMed, Embase, CINAHL, AMED, SPORTDiscus; Jan-Feb 2016; English/French24 studies; 8 muscle-activity, 14 kinematic, and 7 kinetic studiesCAI vs healthy controlsWalking; runningKinematics; kinetics; EMG/muscle activityLuan et al., 2024Systematic review and meta-analysisPubMed, Embase, Cochrane Library, Web of Science, EBSCO, PEDro; to 30 Jun 2023; updated 3 Ma y 2024; no language restriction13 studies; 729 participantsCAI/ankle instability vs non-CAI controls or copersWalkingKinematics; kineticsYousefi et al., 2023Systematic reviewPubMed, ScienceDirect, Scopus, Web of Science, Google Scholar; searched 1 Jun-18 Jul 2022; English6 studiesCAI vs healthy controlsGait initiationCOP/APAs; kinematics; EMG/muscle activityMortezanejad et al., 2024Systematic reviewPubMed, ScienceDirect, Web of Science, Google Scholar, Scopus; 1990-20235 studiesCAI vs healthy controls; affected vs non-affected leg comparisonsGait initiationKinetics/COP; EMG/muscular measuresTheisen & Day 2019Systematic reviewSPORTDiscus, PubMed, CINAHL; search terms included ankle instability and landing kinematics6 studies; 338 participantsCAI vs controlsLandingKinematics; proximal-control interpretationSimpson et al., 2019Systematic reviewPubMed, ScienceDirect, Scopus, SPORTDiscus; to 31 Jan 2018; reference lists searched20 articlesCAI vs non-CAI controlsUnilateral landing; jump-landingKinematics; GRF/loading; EMG/muscle activity; dynamic postural stabilityJeon et al., 2021Systematic review and meta-analysisPubMed, CINAHL, SPORTDiscus, Web of Science; to Dec 2019; English; peer-reviewed11 studies in meta-analysisAI/CAI/FI/MI/recurrent ankle sprain vs healthy/normal controlsLandingEMG/muscle activation; GRF/loadingChan et al., 2022Systematic review and meta-analysisPubMed, MEDLINE, Embase, CINAHL; searched 26 Sep 202121 studiesCAI vs healthy controlsUnilateral jump-landingKinematics; EMG/muscle activity; dynamic postural stabilityHakimi Poor et al., 2022Systematic reviewGoogle Scholar, ScienceDirect, PubMed, Scopus, ProQuest, Cochrane Library; to 2021; English15 articlesCAI/ankle instability participants; controls where reportedUnilateral jump-landingKinematicsHe et al., 2024Systematic review and meta-analysisWeb of Science, Scopus, PubMed, SPORTDiscus, Cochrane Library; May 202414 studies; 564 participantsCAI vs healthy controlsLanding on flat surfaceKinematics; kinetics/joint torques; GRF/loading; proximal-control/potential ACL-mechanism variablesGong et al., 2025Systematic review and meta-analysisPubMed, Embase, Cochrane Library, Web of Science, Scopus; to 30 Nov 202413 studies; 207 CAI and 215 controlsCAI vs healthy controlsExpected/unexpected inversion-perturbation landingKinematics; kinetics; EMG/muscle activityXu et al., 2022Systematic reviewPubMed, Web of Science, SPORTDiscus; inception to Apr 202232 studies: 5 side-cutting, 3 stop-jump, 24 landingCAI vs healthy controlsSide-cutting; stop-jump; landingKinematics; kinetics/joint moments; proximal-control/potential ACL-mechanism biomechanicsDeJong et al., 2020Systematic review and meta-analysisMEDLINE/PubMed and CINAHL; updated to 3 Jun 201955 studiesCAI/FI vs healthy controlsStrength, balance, jumping, gait, and other functional assessmentsProximal neuromuscular/biomechanical outcomes; kinematics; EMG; strengthYu et al., 2022Systematic reviewScienceDirect, PubMed, Web of Science; to 30 Nov 202023 articlesCAI vs copersLocomotor and functional tasksKinematics; kinetics; dynamic postural controlJeon et al., 2025Systematic review and meta-analysisPubMed, CINAHL, SPORTDiscus, Web of Science; inception to Aug 202213 studies; 227 copers, 291 CAI, 159 controlsCopers vs CAI and controlsLanding; walking/runningKinematics; kineticsMohamadi et al., 2026Systematic review and meta-analysisPubMed, Scopus, ScienceDirect, MEDLINE, Web of Science; inception to Mar 202410 studiesCAI vs copers vs healthy subjectsSEBT; jump landing; side cutting; Biodex dynamic balanceNeuromuscular control; kinematics; kinetics; muscle activity; postural controlLabanca et al., 2021Systematic reviewPubMed, Web of Science, Cochrane Library; studies published 2000-202063 articlesCAI vs healthy controlsGait; running; perturbations; landing/hopping; cutting/turning; stance; SEBT; lunges; kicking; Y-balance; squattingEMG/muscle activationChoi et al., 2023Systematic review and meta-analysisPubMed, MEDLINE, CINAHL, SPORTDiscus; inception to Oct 20229 studies in systematic review; 5 in meta-analysisCAI vs healthy controlsDual-task gait and postural controlPostural control; gait; cognitive-motor/dual-task performanceNote: AI, ankle instability; APAs, anticipatory postural adjustments; CAI, chronic ankle instability; COP, centre of pressure; EMG, electromyography; FI, functional instability; GRF, ground reaction force; MI, mechanical instability; SEBT, Star Excursion Balance Test. Study counts and samples are presented as reported by the original reviews. For He et al., 2024, the Results section reported 14 studies and 564 participants, whereas the abstract reported 13 studies and 542 participants; the Results-section value was retained for consistency with the study list and the overlap analysis. For Moisan et al., 2017, outcome-category counts are not mutually exclusive. Characteristics of included systematic reviews and meta-analyses. Note: AI, ankle instability; APAs, anticipatory postural adjustments; CAI, chronic ankle instability; COP, centre of pressure; EMG, electromyography; FI, functional instability; GRF, ground reaction force; MI, mechanical instability; SEBT, Star Excursion Balance Test. Study counts and samples are presented as reported by the original reviews. For He et al., 2024, the Results section reported 14 studies and 564 participants, whereas the abstract reported 13 studies and 542 participants; the Results-section value was retained for consistency with the study list and the overlap analysis. For Moisan et al., 2017, outcome-category counts are not mutually exclusive. Seventeen reviews contributed to the main synthesis of lower limb biomechanics during movement tasks. They addressed walking or running,24,25 gait initiation,27,31 landing and unilateral jump landing,18,20,22,28,29,33 expected or unexpected inversion perturbation landing,19 sport manoeuvres and proximal kinetic-chain adaptations,17,20,30 coper comparisons and dynamic discrete tasks,21,26,32 and functional-task electromyography.23 Choi et al., 2023 was retained for supplementary narrative synthesis because it focused on dual task gait and postural control rather than core joint level lower limb biomechanics.16 The number of primary studies varied substantially, ranging from 5 in the gait initiation review by Mortezanejad et al., 2024 to 63 in the functional task electromyography review by Labanca et al. 2021.23,27 Several reviews focused on single task domains, whereas others addressed broader proximal, coper, or neuromuscular domains. Reviews were mainly distributed across sports medicine, biomechanics, rehabilitation, physiotherapy, and movement science sources. Geographical distribution of underlying primary studies was not consistently reported and was therefore not synthesised separately. Evidence map by task and biomechanical outcome domain The evidence map showed that coverage was uneven across movement tasks and biomechanical outcome domains. Landing related tasks had the broadest evidence coverage across kinematics, kinetics or joint moments, ground reaction force or loading variables, electromyography or muscle activity, centre of pressure or postural stability outcomes, and biomechanical variables related to proximal control or discussed in relation to potential anterior cruciate ligament injury mechanisms (Fig. 2 and Table 2). By contrast, several task domains were supported by fewer reviews, reflecting the more limited evidence available for gait initiation, perturbation landing, electromyography across tasks, and dual task postural control outcomes.Fig. 2Evidence map of included reviews by movement task and biomechanical outcome domain.Table 2Evidence map of included reviews by movement task and biomechanical outcome domain.Table 2Movement task domainKinematicsKinetics/joint momentsGRF/loadingEMG/muscle activityCOP/postural stabilityProximal controlReview countWalking/runningMoisan et al., 2017; Luan et al., 2024; Jeon et al., 2025Moisan et al., 2017; Luan et al., 2024; Jeon et al., 2025-Moisan et al., 2017-DeJong et al., 2020 (gait-related proximal functional-assessment outcomes)4Gait initiationYousefi et al., 2023Mortezanejad et al., 2024-Yousefi et al., 2023; Mortezanejad et al., 2024Yousefi et al., 2023; Mortezanejad et al., 2024-2LandingTheisen and Day 2019; Simpson et al., 2019; He et al., 2024; Jeon et al., 2025He et al., 2024; Jeon et al., 2025Simpson et al., 2019; Jeon et al., 2021; He et al., 2024Simpson et al., 2019; Jeon et al., 2021Simpson et al., 2019Theisen and Day 2019; He et al., 20245Jump-landingChan et al., 2022; Hakimi Poor et al., 2022--Chan et al., 2022Chan et al., 2022-2Perturbation landingGong et al., 2025Gong et al., 2025-Gong et al., 2025--1Sport maneuvers/proximal adaptationsXu et al., 2022; DeJong et al., 2020Xu et al., 2022-DeJong et al., 2020-Xu et al., 2022; DeJong et al., 20202Coper-comparison/dynamic discrete tasksYu et al., 2022; Jeon et al., 2025; Mohamadi et al., 2026Yu et al., 2022; Jeon et al., 2025; Mohamadi et al., 2026-Mohamadi et al., 2026Yu et al., 2022; Mohamadi et al., 2026-3Functional EMG tasks (cross-task)---Labanca et al., 2021--1Dual-task/postural control----Choi et al., 2023-1Note: A review could contribute to more than one movement-task domain and more than one biomechanical outcome domain. The review count represents the number of unique reviews contributing to each row; counts are not mutually exclusive and should not be summed across rows. DeJong et al., 2020 is listed in walking/running only for gait-related proximal functional-assessment outcomes, not as a task specific walking/running review. Labanca et al., 2021 is displayed as a cross-task EMG review to avoid repeating it across all tasks it covered. Choi et al., 2023 was retained for supplementary synthesis of dual-task gait/postural-control evidence. AI, ankle instability; CAI, chronic ankle instability; COP, centre of pressure; EMG, electromyography; GRF, ground reaction force. A dash indicates that no included review was assigned to that task-outcome cell in the evidence map.Table 3Key quantitative outcome-level estimates reported by included meta-analyses.Table 3Review(s)Task/domainComparisonOutcome domainKey quantitative estimate(s)Direction/interpretationLuan et al., 2024WalkingCAI vs non-CAI controlsKinematics; kineticsAnkle inversion angle: WMD 3.71° (95% CI 3.15 to 4.27); hip adduction angle: WMD 1.60° (0.09 to 3.11); knee valgus moment: WMD 0.07 Nm/kg (0.01 to 0.13; I2 = 93%).Greater frontal-plane ankle and proximal-joint alterations during walking in CAI.Jeon et al., 2021LandingAI/CAI vs healthy/normal controlsEMG; GRF/loadingPre-landing peroneal activation: SMD -0.63 (95% CI -0.95 to −0.31; I2 = 31.17%); peak vertical GRF: SMD 0.21 (0.01 to 0.40; I2 = 0%); time to peak vertical GRF: SMD -0.51 (−0.72 to −0.29; I2 = 0%).Lower preparatory ankle-stabiliser activation with greater and faster vertical loading in AI/CAI.Chan et al., 2022Unilateral jump-landingCAI vs healthy controlsKinematics; EMGPost-landing ankle plantarflexion: SMD 0.33 (95% CI 0.02 to 0.65); knee flexion: SMD -0.67 (−0.97 to −0.37); hip flexion: SMD -0.52 (−0.96 to −0.07); pre-landing peroneus longus activation: SMD -0.77 (−1.17 to −0.36).Stiffer sagittal-plane landing pattern and reduced pre-landing peroneus longus activation in CAI.He et al., 2024Landing/proximal-control biomechanicsCAI vs healthy controlsGRF/loading; kinematics; kineticsPeak vertical GRF: SMD 0.30 (95% CI 0.07 to 0.53); hip flexion angle: SMD -0.30 (−0.51 to −0.17); trunk lateral flexion: SMD 0.47 (0.05 to 0.90); hip extension moment: SMD 0.47 (0.09 to 0.84); knee extension moment: SMD 0.39 (0.02 to 0.77).Greater impact loading and altered proximal landing mechanics in CAI; interpretation regarding potential ACL injury mechanisms remains indirect.Gong et al., 2025Expected/unexpected inversion-perturbation landingCAI vs healthy controlsEMG; kinematics; kineticsPre-landing tibialis anterior activity: SMD 0.28 (95% CI 0.03 to 0.54); peroneus longus latency: SMD 1.35 (0.90 to 1.80); sagittal-plane co-contraction index: SMD 0.41 (0.06 to 0.77); ankle inversion angle: SMD 0.56 (0.30 to 0.81); ankle inversion ROM: SMD 0.83 (0.42 to 1.24); knee extension moment: SMD 0.71 (0.32 to 1.11).Impaired neuromuscular preparation and greater post-landing inversion/extension mechanics in CAI.DeJong et al., 2020Functional assessments/proximal adaptationsCAI vs healthy controlsProximal strength; kinematicsHip extension strength: Cohen d 0.62 (95% CI 0.34 to 0.91); hip abduction strength: d 0.52 (0.24 to 0.81); hip external-rotation strength: d 0.93 (0.59 to 1.27). Dynamic-balance knee flexion angle: d −1.44 (−3.95 to 1.07; p = 0.26; I2 = 96.29%).Triplanar hip strength deficits in CAI; dynamic-balance knee-flexion result was not significant.Jeon et al., 2025Landing; walking/runningCopers vs CAI and controlsKinematics; kineticsCompared with CAI, copers showed smaller landing ankle frontal-plane differences (|SMD| 0.31 to 0.40), greater sagittal ankle mechanics during landing and less plantarflexion moment during walking (|SMD| 0.29 to 0.46), and proximal landing differences across knee/hip outcomes (|SMD| 0.25 to 0.73).Copers generally demonstrated biomechanics closer to controls than to CAI.Mohamadi et al., 2026Dynamic discrete tasksCAI vs copers vs healthy subjectsNeuromuscular control; kinematics; kinetics; muscle activity; postural controlAcross 10 studies, pooled SMD-based findings indicated decreased muscle activity, lower postural-control/balance indices, lower kinematic and kinetic parameters, and lower reach distance in CAI; copers and healthy subjects were similar for most parameters.CAI showed greater dynamic-task neuromuscular deficits than copers and healthy subjects.Choi et al., 2023Dual-task postural control/gaitCAI vs controls; single-task vs dual-taskPostural controlControl mediolateral single-task vs dual-task effect: Fisher Z 0.732 (95% CI 0.073 to 1.392; p = 0.029; I2 = 88.25%). CAI mediolateral effect: Fisher Z 0.068 (−0.284 to 0.419; p = 0.706). CAI vs controls under dual-task: Fisher Z −0.053 (−0.246 to 0.140; p = 0.592).Supplementary evidence; no clear CAI-control dual-task difference.Note: This main-text table presents selected key quantitative estimates from included meta-analyses; detailed outcome-level extraction is provided in Supplementary Table S8. Estimates are shown as reported by the original reviews and were not statistically re-pooled in the present umbrella review.Table 4Summary of findings and certainty of evidence by task-outcome domain.Table 4Task/domainOutcomeSupporting review(s)Effect estimate/summaryDirection/interpretationCertainty of evidenceWalkingAnkle inversion angleLuan et al., 2024WMD 3.71° (95% CI 3.15 to 4.27)Greater ankle inversion during walking in CAILowWalkingHip adduction angleLuan et al., 2024WMD 1.60° (95% CI 0.09 to 3.11)Greater hip adduction during walking in CAILowWalkingKnee valgus momentLuan et al., 2024WMD 0.07 Nm/kg (95% CI 0.01 to 0.13); I2 = 93%Greater knee valgus moment; high heterogeneityVery lowGait initiationCOP displacement/APA timingYousefi et al., 2023; Mortezanejad et al., 2024Narrative findings across 5–6 studiesAltered COP displacement, APA timing, and muscle-onset organisation in CAIVery lowLandingPre-landing peroneal activationJeon et al., 2021; Chan et al., 2022SMD −0.63 and SMD −0.77 in separate meta-analysesLower feed-forward peroneal activation in CAI/AILowLandingPeak vertical GRFJeon et al., 2021; He et al., 2024SMD 0.21 and SMD 0.30 in separate meta-analysesHigher impact loading in CAI/AILowLandingTime to peak vertical GRFJeon et al., 2021SMD −0.51 (95% CI −0.72 to −0.29)Shorter time to peak vertical GRF in AILowJump-landingSagittal-plane landing kinematicsChan et al., 2022Ankle plantarflexion SMD 0.33; knee flexion SMD −0.67; hip flexion SMD −0.52Stiffer/less-flexed landing pattern in CAILowPerturbation landingPeroneus longus activation latencyGong et al., 2025SMD 1.35 (95% CI 0.90 to 1.80)Delayed peroneus longus activation in CAILowPerturbation landingAnkle inversion angle/ROMGong et al., 2025Angle SMD 0.56; ROM SMD 0.83Greater post-landing inversion mechanics in CAILowPerturbation landingKnee extension momentGong et al., 2025SMD 0.71 (95% CI 0.32 to 1.11)Higher knee extension moment in CAILowProximal-control landingHip flexion and trunk lateral flexionHe et al., 2024Hip flexion SMD −0.30; trunk lateral flexion SMD 0.47Reduced hip flexion and increased trunk lateral flexion in CAI; relevance to potential ACL injury mechanisms is indirectLowProximal adaptationsHip strength deficitsDeJong et al., 2020Cohen's d 0.52–0.93 for hip strength outcomesHip strength lower in CAI; positive values coded as controls strongerLowCopers vs CAIDynamic-task lower-limb biomechanicsYu et al., 2022; Jeon et al., 2025; Mohamadi et al., 2026Multiple pooled and narrative comparisonsCopers generally showed biomechanics closer to controls than to CAILowDual-task supplementaryDual-task postural control/gaitChoi et al., 2023Fisher Z estimates mostly non-significant for CAI-control comparisonsNo clear CAI-control difference under dual-task conditionsVery lowNote: Certainty was assessed using a modified GRADE approach at the task-outcome level. Most evidence started at low certainty because primary studies were mainly observational laboratory studies and was downgraded for concerns in risk of bias, inconsistency, indirectness, imprecision, or publication bias. Detailed rationale is provided in Supplementary Table S5. Effect estimates are reported as extracted from included meta-analyses and were not re-pooled. SMD, standardised mean difference; WMD, weighted mean difference; CI, confidence interval; CAI, chronic ankle instability; AI, ankle instability; COP, centre of pressure; APA, anticipatory postural adjustment; GRF, ground reaction force; ROM, range of motion; ACL, anterior cruciate ligament. Evidence map of included reviews by movement task and biomechanical outcome domain. Table 2Evidence map of included reviews by movement task and biomechanical outcome domain.Table 2Movement task domainKinematicsKinetics/joint momentsGRF/loadingEMG/muscle activityCOP/postural stabilityProximal controlReview countWalking/runningMoisan et al., 2017; Luan et al., 2024; Jeon et al., 2025Moisan et al., 2017; Luan et al., 2024; Jeon et al., 2025-Moisan et al., 2017-DeJong et al., 2020 (gait-related proximal functional-assessment outcomes)4Gait initiationYousefi et al., 2023Mortezanejad et al., 2024-Yousefi et al., 2023; Mortezanejad et al., 2024Yousefi et al., 2023; Mortezanejad et al., 2024-2LandingTheisen and Day 2019; Simpson et al., 2019; He et al., 2024; Jeon et al., 2025He et al., 2024; Jeon et al., 2025Simpson et al., 2019; Jeon et al., 2021; He et al., 2024Simpson et al., 2019; Jeon et al., 2021Simpson et al., 2019Theisen and Day 2019; He et al., 20245Jump-landingChan et al., 2022; Hakimi Poor et al., 2022--Chan et al., 2022Chan et al., 2022-2Perturbation landingGong et al., 2025Gong et al., 2025-Gong et al., 2025--1Sport maneuvers/proximal adaptationsXu et al., 2022; DeJong et al., 2020Xu et al., 2022-DeJong et al., 2020-Xu et al., 2022; DeJong et al., 20202Coper-comparison/dynamic discrete tasksYu et al., 2022; Jeon et al., 2025; Mohamadi et al., 2026Yu et al., 2022; Jeon et al., 2025; Mohamadi et al., 2026-Mohamadi et al., 2026Yu et al., 2022; Mohamadi et al., 2026-3Functional EMG tasks (cross-task)---Labanca et al., 2021--1Dual-task/postural control----Choi et al., 2023-1Note: A review could contribute to more than one movement-task domain and more than one biomechanical outcome domain. The review count represents the number of unique reviews contributing to each row; counts are not mutually exclusive and should not be summed across rows. DeJong et al., 2020 is listed in walking/running only for gait-related proximal functional-assessment outcomes, not as a task specific walking/running review. Labanca et al., 2021 is displayed as a cross-task EMG review to avoid repeating it across all tasks it covered. Choi et al., 2023 was retained for supplementary synthesis of dual-task gait/postural-control evidence. AI, ankle instability; CAI, chronic ankle instability; COP, centre of pressure; EMG, electromyography; GRF, ground reaction force. A dash indicates that no included review was assigned to that task-outcome cell in the evidence map. Evidence map of included reviews by movement task and biomechanical outcome domain. Note: A review could contribute to more than one movement-task domain and more than one biomechanical outcome domain. The review count represents the number of unique reviews contributing to each row; counts are not mutually exclusive and should not be summed across rows. DeJong et al., 2020 is listed in walking/running only for gait-related proximal functional-assessment outcomes, not as a task specific walking/running review. Labanca et al., 2021 is displayed as a cross-task EMG review to avoid repeating it across all tasks it covered. Choi et al., 2023 was retained for supplementary synthesis of dual-task gait/postural-control evidence. AI, ankle instability; CAI, chronic ankle instability; COP, centre of pressure; EMG, electromyography; GRF, ground reaction force. A dash indicates that no included review was assigned to that task-outcome cell in the evidence map. Table 3Key quantitative outcome-level estimates reported by included meta-analyses.Table 3Review(s)Task/domainComparisonOutcome domainKey quantitative estimate(s)Direction/interpretationLuan et al., 2024WalkingCAI vs non-CAI controlsKinematics; kineticsAnkle inversion angle: WMD 3.71° (95% CI 3.15 to 4.27); hip adduction angle: WMD 1.60° (0.09 to 3.11); knee valgus moment: WMD 0.07 Nm/kg (0.01 to 0.13; I2 = 93%).Greater frontal-plane ankle and proximal-joint alterations during walking in CAI.Jeon et al., 2021LandingAI/CAI vs healthy/normal controlsEMG; GRF/loadingPre-landing peroneal activation: SMD -0.63 (95% CI -0.95 to −0.31; I2 = 31.17%); peak vertical GRF: SMD 0.21 (0.01 to 0.40; I2 = 0%); time to peak vertical GRF: SMD -0.51 (−0.72 to −0.29; I2 = 0%).Lower preparatory ankle-stabiliser activation with greater and faster vertical loading in AI/CAI.Chan et al., 2022Unilateral jump-landingCAI vs healthy controlsKinematics; EMGPost-landing ankle plantarflexion: SMD 0.33 (95% CI 0.02 to 0.65); knee flexion: SMD -0.67 (−0.97 to −0.37); hip flexion: SMD -0.52 (−0.96 to −0.07); pre-landing peroneus longus activation: SMD -0.77 (−1.17 to −0.36).Stiffer sagittal-plane landing pattern and reduced pre-landing peroneus longus activation in CAI.He et al., 2024Landing/proximal-control biomechanicsCAI vs healthy controlsGRF/loading; kinematics; kineticsPeak vertical GRF: SMD 0.30 (95% CI 0.07 to 0.53); hip flexion angle: SMD -0.30 (−0.51 to −0.17); trunk lateral flexion: SMD 0.47 (0.05 to 0.90); hip extension moment: SMD 0.47 (0.09 to 0.84); knee extension moment: SMD 0.39 (0.02 to 0.77).Greater impact loading and altered proximal landing mechanics in CAI; interpretation regarding potential ACL injury mechanisms remains indirect.Gong et al., 2025Expected/unexpected inversion-perturbation landingCAI vs healthy controlsEMG; kinematics; kineticsPre-landing tibialis anterior activity: SMD 0.28 (95% CI 0.03 to 0.54); peroneus longus latency: SMD 1.35 (0.90 to 1.80); sagittal-plane co-contraction index: SMD 0.41 (0.06 to 0.77); ankle inversion angle: SMD 0.56 (0.30 to 0.81); ankle inversion ROM: SMD 0.83 (0.42 to 1.24); knee extension moment: SMD 0.71 (0.32 to 1.11).Impaired neuromuscular preparation and greater post-landing inversion/extension mechanics in CAI.DeJong et al., 2020Functional assessments/proximal adaptationsCAI vs healthy controlsProximal strength; kinematicsHip extension strength: Cohen d 0.62 (95% CI 0.34 to 0.91); hip abduction strength: d 0.52 (0.24 to 0.81); hip external-rotation strength: d 0.93 (0.59 to 1.27). Dynamic-balance knee flexion angle: d −1.44 (−3.95 to 1.07; p = 0.26; I2 = 96.29%).Triplanar hip strength deficits in CAI; dynamic-balance knee-flexion result was not significant.Jeon et al., 2025Landing; walking/runningCopers vs CAI and controlsKinematics; kineticsCompared with CAI, copers showed smaller landing ankle frontal-plane differences (|SMD| 0.31 to 0.40), greater sagittal ankle mechanics during landing and less plantarflexion moment during walking (|SMD| 0.29 to 0.46), and proximal landing differences across knee/hip outcomes (|SMD| 0.25 to 0.73).Copers generally demonstrated biomechanics closer to controls than to CAI.Mohamadi et al., 2026Dynamic discrete tasksCAI vs copers vs healthy subjectsNeuromuscular control; kinematics; kinetics; muscle activity; postural controlAcross 10 studies, pooled SMD-based findings indicated decreased muscle activity, lower postural-control/balance indices, lower kinematic and kinetic parameters, and lower reach distance in CAI; copers and healthy subjects were similar for most parameters.CAI showed greater dynamic-task neuromuscular deficits than copers and healthy subjects.Choi et al., 2023Dual-task postural control/gaitCAI vs controls; single-task vs dual-taskPostural controlControl mediolateral single-task vs dual-task effect: Fisher Z 0.732 (95% CI 0.073 to 1.392; p = 0.029; I2 = 88.25%). CAI mediolateral effect: Fisher Z 0.068 (−0.284 to 0.419; p = 0.706). CAI vs controls under dual-task: Fisher Z −0.053 (−0.246 to 0.140; p = 0.592).Supplementary evidence; no clear CAI-control dual-task difference.Note: This main-text table presents selected key quantitative estimates from included meta-analyses; detailed outcome-level extraction is provided in Supplementary Table S8. Estimates are shown as reported by the original reviews and were not statistically re-pooled in the present umbrella review. Key quantitative outcome-level estimates reported by included meta-analyses. Note: This main-text table presents selected key quantitative estimates from included meta-analyses; detailed outcome-level extraction is provided in Supplementary Table S8. Estimates are shown as reported by the original reviews and were not statistically re-pooled in the present umbrella review. Table 4Summary of findings and certainty of evidence by task-outcome domain.Table 4Task/domainOutcomeSupporting review(s)Effect estimate/summaryDirection/interpretationCertainty of evidenceWalkingAnkle inversion angleLuan et al., 2024WMD 3.71° (95% CI 3.15 to 4.27)Greater ankle inversion during walking in CAILowWalkingHip adduction angleLuan et al., 2024WMD 1.60° (95% CI 0.09 to 3.11)Greater hip adduction during walking in CAILowWalkingKnee valgus momentLuan et al., 2024WMD 0.07 Nm/kg (95% CI 0.01 to 0.13); I2 = 93%Greater knee valgus moment; high heterogeneityVery lowGait initiationCOP displacement/APA timingYousefi et al., 2023; Mortezanejad et al., 2024Narrative findings across 5–6 studiesAltered COP displacement, APA timing, and muscle-onset organisation in CAIVery lowLandingPre-landing peroneal activationJeon et al., 2021; Chan et al., 2022SMD −0.63 and SMD −0.77 in separate meta-analysesLower feed-forward peroneal activation in CAI/AILowLandingPeak vertical GRFJeon et al., 2021; He et al., 2024SMD 0.21 and SMD 0.30 in separate meta-analysesHigher impact loading in CAI/AILowLandingTime to peak vertical GRFJeon et al., 2021SMD −0.51 (95% CI −0.72 to −0.29)Shorter time to peak vertical GRF in AILowJump-landingSagittal-plane landing kinematicsChan et al., 2022Ankle plantarflexion SMD 0.33; knee flexion SMD −0.67; hip flexion SMD −0.52Stiffer/less-flexed landing pattern in CAILowPerturbation landingPeroneus longus activation latencyGong et al., 2025SMD 1.35 (95% CI 0.90 to 1.80)Delayed peroneus longus activation in CAILowPerturbation landingAnkle inversion angle/ROMGong et al., 2025Angle SMD 0.56; ROM SMD 0.83Greater post-landing inversion mechanics in CAILowPerturbation landingKnee extension momentGong et al., 2025SMD 0.71 (95% CI 0.32 to 1.11)Higher knee extension moment in CAILowProximal-control landingHip flexion and trunk lateral flexionHe et al., 2024Hip flexion SMD −0.30; trunk lateral flexion SMD 0.47Reduced hip flexion and increased trunk lateral flexion in CAI; relevance to potential ACL injury mechanisms is indirectLowProximal adaptationsHip strength deficitsDeJong et al., 2020Cohen's d 0.52–0.93 for hip strength outcomesHip strength lower in CAI; positive values coded as controls strongerLowCopers vs CAIDynamic-task lower-limb biomechanicsYu et al., 2022; Jeon et al., 2025; Mohamadi et al., 2026Multiple pooled and narrative comparisonsCopers generally showed biomechanics closer to controls than to CAILowDual-task supplementaryDual-task postural control/gaitChoi et al., 2023Fisher Z estimates mostly non-significant for CAI-control comparisonsNo clear CAI-control difference under dual-task conditionsVery lowNote: Certainty was assessed using a modified GRADE approach at the task-outcome level. Most evidence started at low certainty because primary studies were mainly observational laboratory studies and was downgraded for concerns in risk of bias, inconsistency, indirectness, imprecision, or publication bias. Detailed rationale is provided in Supplementary Table S5. Effect estimates are reported as extracted from included meta-analyses and were not re-pooled. SMD, standardised mean difference; WMD, weighted mean difference; CI, confidence interval; CAI, chronic ankle instability; AI, ankle instability; COP, centre of pressure; APA, anticipatory postural adjustment; GRF, ground reaction force; ROM, range of motion; ACL, anterior cruciate ligament. Summary of findings and certainty of evidence by task-outcome domain. Note: Certainty was assessed using a modified GRADE approach at the task-outcome level. Most evidence started at low certainty because primary studies were mainly observational laboratory studies and was downgraded for concerns in risk of bias, inconsistency, indirectness, imprecision, or publication bias. Detailed rationale is provided in Supplementary Table S5. Effect estimates are reported as extracted from included meta-analyses and were not re-pooled. SMD, standardised mean difference; WMD, weighted mean difference; CI, confidence interval; CAI, chronic ankle instability; AI, ankle instability; COP, centre of pressure; APA, anticipatory postural adjustment; GRF, ground reaction force; ROM, range of motion; ACL, anterior cruciate ligament. Evidence on walking and running was represented across kinematic and kinetic domains, with additional evidence relating to muscle activity and gait-related proximal functional assessment outcomes.24,25 Evidence on gait initiation was concentrated in centre-of-pressure displacement, anticipatory postural adjustments, kinematics, and muscle onset or muscle activity outcomes.27,31 Perturbation landing was represented by one recent meta-analysis that synthesised kinematic, kinetic, and muscle activity outcomes during expected and unexpected inversion perturbation landing.19 Evidence on sport manoeuvres and proximal adaptations was derived primarily from reviews addressing side cutting, stop jump, landing, gait-related proximal outcomes, and broader proximal kinetic-chain adaptations.17,20,30 One review on dual task gait and postural control was retained as supplementary evidence because it addressed cognitive motor and postural control performance rather than core joint level lower limb biomechanics.16 Similarly, Labanca et al., 2021 was presented as a review of electromyography across tasks to avoid repeated counting across the functional tasks covered in that review.23 The plotting data used to generate Fig. 2 are provided in Supplementary Table S7. Primary-study overlap across reviews Primary study overlap was moderate overall, although the degree of overlap varied across task specific review clusters. Across all 18 included reviews, the citation matrix contained 353 primary study occurrences and 185 unique primary studies, yielding an overall CCA of 5.34%, which was interpreted as moderate overlap. After exclusion of the supplementary dual task review, the 17 reviews included in the main synthesis contained 344 primary study occurrences and 177 unique primary studies, yielding a CCA of 5.90%, which also indicated moderate overlap. Task cluster analyses demonstrated greater overlap within closely related evidence domains. The cluster comprising landing, jump landing, perturbation landing, and proximal-control or potential ACL-mechanism-related biomechanics had a CCA of 10.16%, indicating high overlap. The dynamic task cluster comparing copers with individuals with CAI showed very high overlap (CCA 19.70%), and the functional electromyography plus proximal adaptation cluster also showed very high overlap (CCA 22.92%). These findings indicate that the overall evidence base was not dominated by excessive duplication of primary studies, although overlap was more substantial in narrower domains in which reviews addressed similar tasks, populations, or outcome constructs. Therefore, primary study overlap was considered when the task specific findings and certainty of evidence were interpreted. The complete citation matrix, CCA summary, pairwise overlap analysis, primary study lists, and extraction notes are provided in Supplementary Table S6. Task specific biomechanical findings Walking and running Evidence on walking and running was derived primarily from Moisan et al., 2017 and Luan et al. 2024.24,25 For walking, the evidence indicated greater frontal plane alterations at the ankle in individuals with CAI or ankle instability, particularly greater ankle inversion. Proximal frontal plane differences were also reported, including greater hip adduction and knee valgus moment. Evidence specific to running was more limited and was supported mainly by findings from systematic reviews rather than by pooled estimates. Overall, locomotor evidence suggested persistent alterations in frontal plane ankle control and proximal joint control, with stronger quantitative support for walking than for running. Gait initiation Two systematic reviews examined gait initiation in individuals with CAI. Yousefi et al., 2023 synthesised evidence on centre of pressure displacement, anticipatory postural adjustments, ankle and foot kinematics, and muscle activation, whereas Mortezanejad et al., 2024 focused on kinetic and muscular changes during gait initiation.27,31 Collectively, these reviews indicated altered preparatory postural organisation before step execution, including changes in centre of pressure displacement, anticipatory postural adjustment timing, and muscle onset organisation. Because the evidence base was small and the findings were mainly synthesised narratively, evidence on gait initiation was interpreted as suggestive rather than definitive. Landing and jump-landing Landing and jump landing represented the most extensively studied movement task domain. Theisen and Day 2019, Simpson et al., 2019, Jeon et al., 2021, Chan et al., 2022, Hakimi Poor et al., 2022, and He et al., 2024 collectively addressed lower limb kinematics, ground reaction force variables, muscle activation, dynamic postural stability, and proximal-control landing variables.18,20,22,28,29,33 Across these reviews, reduced peroneal activation before landing, greater or more rapid vertical loading, and a stiffer sagittal-plane landing strategy were commonly reported in individuals with CAI or ankle instability. This landing strategy was characterised by greater plantarflexion after landing and reduced knee and hip flexion. Notably, He et al., 2024 further indicated that landing alterations associated with CAI may also involve proximal biomechanical variables that have been discussed in relation to potential anterior cruciate ligament injury mechanisms, including reduced hip flexion, increased trunk lateral flexion, and greater hip or knee extension moments. Perturbation landing Evidence on perturbation landing was provided by Gong et al., 2025, which examined expected and unexpected inversion perturbation landing.19 This review indicated that altered neuromuscular preparation and mechanical responses after landing were evident in individuals with CAI. Reported alterations included greater tibialis anterior activity before landing, delayed peroneus longus activation, increased sagittal plane co-contraction, greater ankle inversion angle and range of motion, and greater knee extension moment. Collectively, these findings suggest that perturbation landing may expose CAI related deficits in rapid lateral ankle protection and frontal plane stabilisation. Sport manoeuvres and proximal kinetic-chain adaptations Evidence on sport manoeuvres and proximal kinetic-chain adaptations was derived primarily from Xu et al., 2022, DeJong et al., 2020, and He et al. 2024.17,20,30 Xu et al., 2022 reviewed side cutting, stop jump, and landing tasks and reported altered proximal lower-limb biomechanics in individuals with CAI. DeJong et al., 2020 synthesised outcomes related to the trunk, hip, thigh, and knee across functional assessments and reported proximal impairments, particularly deficits in hip strength. He et al., 2024 reported proximal landing alterations including reduced hip flexion, increased trunk lateral flexion, and greater hip or knee extension moments.20 These variables were interpreted cautiously as biomechanical features discussed in relation to potential anterior cruciate ligament injury mechanisms rather than as direct evidence of future injury risk. Collectively, these reviews support a kinetic-chain interpretation in which CAI-related alterations may involve the hip, knee, and trunk rather than being confined to the ankle. Copers versus CAI Three reviews contributed evidence comparing ankle sprain copers with individuals with CAI. Yu et al., 2022 reviewed locomotor biomechanics and dynamic postural control, Jeon et al., 2025 synthesised kinematic and kinetic characteristics during landing, walking, and running tasks, and Mohamadi et al., 2026 examined neuromuscular control during dynamic discrete tasks.21,26,32 Across these reviews, copers were generally characterised by biomechanical or neuromuscular profiles that were closer to those of healthy controls than to those of individuals with CAI. This pattern suggests that coper status may be associated with better preserved or compensatory movement strategies after lateral ankle sprain. Supplementary dual-task evidence Choi et al., 2023 was retained as supplementary evidence because it examined dual task gait and postural control outcomes rather than core joint level lower limb biomechanics.16 Its findings did not indicate clear differences between CAI and control groups under dual task conditions, although a significant mediolateral difference between single task and dual task conditions was reported in the control group. Because this evidence domain primarily reflected cognitive motor and postural control performance, it was interpreted separately from the main synthesis of lower limb biomechanics. Key quantitative findings from included meta-analyses Key quantitative estimates are summarised in Table 3, and selected standardised or transformed estimates are displayed in Fig. 3. Full extracted outcome level estimates are provided in Supplementary Table S8.Fig. 3Selected standardised outcome-level effect estimates reported by included meta-analyses. Selected standardised outcome-level effect estimates reported by included meta-analyses. Quantitative evidence from meta-analyses was available for several movement-task domains, with the most consistent pooled findings observed for landing-related tasks. Across landing and unilateral jump-landing studies, the main signals were reduced pre-landing peroneal activation,22,28 greater or faster vertical loading,20,22 greater plantarflexion after landing,28 and reduced hip and knee flexion in individuals with CAI or ankle instability.20,28 Perturbation landing evidence indicated delayed peroneus longus activation, greater tibialis anterior activity, increased co-contraction, greater ankle inversion mechanics, and greater knee extension moment.19 Walking-related pooled findings were reported as weighted mean differences and suggested greater ankle inversion, hip adduction, and knee valgus moment.24 Meta-analyses of proximal outcomes and coper comparisons further indicated hip strength deficits17 and better preserved dynamic-task biomechanics in copers than in individuals with CAI.21,26,32 Fig. 3 presents only selected standardised or transformed estimates that could be displayed on a common metric scale. Weighted mean differences from walking analyses and pooled summaries based on standardised mean difference ranges from coper comparison reviews were not plotted because they were not directly comparable on a single effect size axis. Because effect metrics, task protocols, comparator definitions, and outcome time windows differed substantially across reviews, no new pooled summary estimate was calculated. Overall, quantitative evidence most consistently suggested altered preparatory muscle activation, increased or faster loading, and stiffer sagittal plane landing mechanics in CAI, with additional task specific evidence for altered locomotor mechanics, proximal deficits, and differential coper movement strategies. Methodological quality of included reviews AMSTAR 2 appraisal showed that all included reviews had low or critically low methodological confidence. Of the 18 reviews, 13 were rated as critically low and 5 as low, with none rated as moderate or high. Item level ratings are shown in Fig. 4 and Supplementary Table S3.Fig. 4AMSTAR 2 item level methodological quality appraisal heatmap. AMSTAR 2 item level methodological quality appraisal heatmap. Recurring limitations included absence of prospective protocol registration or a clearly stated a priori protocol, lack of excluded study lists with justifications, limited reporting of funding sources for primary studies, and unclear duplicate data extraction. Among reviews with meta-analysis, publication bias was inconsistently assessed or reported, and several reviews did not clearly explain how primary study risk of bias informed interpretation. Although search comprehensiveness and basic review structure were generally adequate, the AMSTAR 2 profile indicated variable methodological transparency. These findings suggest that the review level evidence base should be interpreted cautiously, particularly where conclusions rely on reviews with multiple critical weaknesses. Risk of bias in reviews ROBIS assessment indicated that review level risk of bias was generally low or of some concern. Of the 18 included reviews, 3 were rated as low overall concern and 15 as having some concerns. No review was rated as high overall concern. Domain level assessments are shown in Fig. 5 and Supplementary Table S4.Fig. 5ROBIS domain level risk of bias assessment heatmap. ROBIS domain level risk of bias assessment heatmap. Reviews with low overall concern generally had clear eligibility criteria, adequate study identification and selection, appropriate data collection and appraisal procedures, and transparent synthesis methods. Reviews with some concerns typically showed limitations in study identification and selection, data collection and appraisal, or synthesis and findings. Across ROBIS domains, eligibility criteria were usually judged as low concern, suggesting that review questions and inclusion criteria were generally well defined. Some concerns were more frequent for identification and selection of studies, data collection and study appraisal, and synthesis and findings, mainly because of incomplete reporting of search or selection procedures, limited detail on duplicate processes or appraisal methods, and variable clarity in synthesis and interpretation. Certainty of evidence Certainty ratings are presented in Table 4 and Fig. 6, with detailed GRADE rationale in Supplementary Table S5. Certainty was assessed across risk of bias, inconsistency, indirectness, imprecision, and publication bias. One level was downgraded for serious concerns and two levels for very serious concerns. Primary study overlap within task clusters was also considered as a potential source of inconsistency and imprecision.Fig. 6GRADE certainty of evidence by task outcome domain. GRADE certainty of evidence by task outcome domain. Across 15 task outcome conclusions assessed using modified GRADE, 12 were rated as low certainty and 3 as very low certainty. No domain provided moderate or high certainty evidence. This indicates that, although review level findings suggested recurring biomechanical alterations in CAI, confidence in the precision and stability of these conclusions remained limited. Very low certainty conclusions related to gait initiation and postural preparation, walking knee valgus moment, and supplementary dual task gait and postural control evidence. Downgrading reflected combined methodological and evidence base limitations, including observational laboratory designs, heterogeneous task protocols and outcome definitions, small samples, imprecision in pooled or narrative findings, and limited or unclear publication bias assessment. The remaining conclusions provided low certainty evidence. Several findings showed consistent directional signals across reviews, but certainty was limited by observational designs, between study heterogeneity, modest sample sizes, review level methodological limitations, and primary study overlap, particularly in landing related, coper comparison, and neuromuscular evidence domains. Overall, the evidence base is sufficient to identify recurring patterns of altered lower limb biomechanics in CAI, but insufficient to support highly definitive conclusions for most task outcome domains. Discussion Principal findings This umbrella review synthesised systematic reviews and meta-analyses on lower limb biomechanical and neuromuscular alterations during movement tasks in individuals with CAI or related ankle instability conditions. Current review-level evidence suggests that CAI may involve task-specific mechanical and neuromuscular alterations across locomotor, gait initiation, landing, perturbation, sport manoeuvre, proximal kinetic-chain, and coper comparison tasks. This pattern supports a movement-system and translational mechanobiology perspective, although confidence remains limited by low to very low evidence certainty. Evidence was most concentrated for landing and jump landing tasks. Across these tasks, reduced pre-landing peroneal activation, greater or faster vertical ground-reaction-force loading, and stiffer sagittal-plane landing mechanics with reduced knee and hip flexion were commonly reported. Walking evidence suggested greater ankle inversion, greater hip adduction, and greater knee valgus moment, gait initiation evidence indicated altered anticipatory postural organisation, and perturbation landing evidence showed delayed peroneus longus activation and greater inversion mechanics. Coper comparisons may suggest more preserved task-specific movement profiles after ankle sprain. These findings should be interpreted in light of the evidence profile. Most included reviews had low or critically low AMSTAR 2 confidence, ROBIS assessments frequently indicated some concerns, and modified GRADE ratings were low or very low for all task-outcome conclusions. Thus, the synthesis identifies recurring biomechanical signals that may inform mechanobiological targets, but inference remains limited by heterogeneous task protocols, modest laboratory samples, review-level methodological limitations, and primary-study overlap within related evidence domains. Task specific kinetic chain interpretation and coper evidence The task-specific pattern of findings suggests that biomechanical alterations in CAI should not be interpreted as uniform across all movements.21,32 This interpretation is consistent with Donovan and Feger 2017, who showed that ankle frontal-plane kinematics may differ across functional tasks rather than following a single task-independent alteration pattern.50 Thus, movement tasks may act as different stress tests for sensorimotor control. Low-demand locomotor tasks mainly revealed frontal-plane control alterations, whereas landing and perturbation tasks exposed deficits in feedforward ankle stabilisation, impact attenuation, and rapid protective responses.51 This distinction is clinically relevant because individuals with CAI may appear functional during simple assessments but show altered mechanics when movement speed, loading, or uncertainty increases.52 Walking evidence primarily suggested altered frontal-plane control. Greater ankle inversion may reflect a more vulnerable foot and ankle position during repeated stance phases.53 Concurrent findings of greater hip adduction and knee valgus moment indicate that locomotor adaptations are not confined to the ankle but may involve redistribution of control demands across the limb.54 Running evidence was less extensive, so running-specific conclusions remain tentative. Gait initiation findings suggest altered anticipatory postural control, because changes in centre of pressure displacement, anticipatory postural adjustment timing, and muscle onset organisation may reflect impaired preparation before stepping.55 Landing and jump landing findings suggest impaired preparation for ground contact and reduced capacity to attenuate load. Reduced peroneal activation before landing may limit lateral ankle readiness at initial contact, while greater peak vertical loading, shorter time to peak vertical ground reaction force, and reduced hip and knee flexion suggest a stiffer strategy with less whole limb shock absorption. Perturbation landing evidence strengthens this interpretation by showing delayed peroneus longus activation, altered tibialis anterior activity, increased co-contraction, and greater inversion mechanics after landing. These findings are clinically meaningful because recurrent ankle sprains often occur during rapid, constrained, or unexpected movements. Together, these findings support a kinetic-chain interpretation of CAI. Local features such as ligamentous laxity, restricted dorsiflexion, peroneal dysfunction, and recurrent giving way remain important, while proprioceptive and postural-control deficits highlight the role of sensorimotor integration. The present synthesis suggests that these impairments may extend to the knee, hip, and trunk. Evidence from walking, landing, sport manoeuvres, and proximal adaptation reviews indicates that distal instability may interact with proximal control demands, especially during single-limb loading, deceleration, redirection, and frontal-plane stability tasks.56 CAI may therefore be better understood as a lower limb movement system condition in which ankle instability and proximal control influence one another. Coper evidence adds a useful contrast to this interpretation. Copers share a history of lateral ankle sprain with individuals with CAI but do not develop recurrent giving way, persistent perceived instability, or functional limitation.57 In this umbrella review, copers generally showed biomechanical and neuromuscular profiles closer to healthy controls than to individuals with CAI.58 Across locomotor, landing, and dynamic discrete tasks, they tended to show smaller frontal plane ankle alterations, less pronounced proximal joint deviations, and fewer neuromuscular control deficits.59 This pattern suggests that successful recovery may involve preserved or restored coordination across the ankle and proximal joints. These findings shift attention from isolated deficits toward biomechanical targets and potential protective mechanisms. More effective sagittal-plane landing mechanics, controlled frontal-plane ankle motion, timely peroneal responses, and coordinated proximal control may represent clinically relevant targets for rehabilitation and prevention. However, current evidence cannot determine whether these profiles are causes of successful recovery, consequences of recovery, or markers of lower symptom burden. Prospective studies after initial lateral ankle sprain are needed to test whether early restoration of coper-like movement patterns reduces CAI risk. Methodological quality, certainty, and overlap The methodological and evidential profile of the included reviews is central to interpreting the task specific findings. Although recurring biomechanical alterations were identified across locomotor, landing, perturbation, sport manoeuvre, and coper comparison tasks, their credibility depends on review quality, risk of bias, evidence certainty, and independence of the underlying primary evidence. AMSTAR 2 indicated low or critically low methodological confidence in all included reviews. Common limitations included absent prospective registration, incomplete lists of excluded studies, limited reporting of primary study funding, unclear duplicate data extraction, and inconsistent publication bias assessment.60,61 ROBIS findings were less severe but still indicated some concerns, mainly in study identification and selection, data collection and appraisal, and synthesis and findings.62 These limitations do not invalidate the evidence, but they reduce confidence in the transparency and reproducibility of review level conclusions. The modified GRADE assessment further supports cautious interpretation, as all task outcome conclusions were rated as low or very low certainty. This reflected the predominance of observational laboratory studies, small samples, heterogeneous task protocols, variable outcome definitions, imprecision, and limited or unclear publication bias assessment.63 Primary study overlap adds another interpretive constraint. Although overall CCA indicated moderate overlap, overlap was higher within narrower clusters, particularly landing related, coper comparison, functional electromyography, and proximal adaptation reviews. Thus, agreement across reviews may partly reflect repeated inclusion of the same primary studies rather than independent replication. Taken together, the findings are best viewed as consistent biomechanical signals rather than high certainty causal conclusions. Current evidence suggests task specific lower limb alterations in CAI, especially during landing and perturbation tasks, but methodological limitations, low certainty ratings, and cluster specific overlap restrict the strength of inference. Clinical and research implications The findings may inform a task-specific kinetic-chain approach to CAI assessment and rehabilitation, but they should be interpreted in light of low to very low evidence certainty and the limited clinical accessibility of many biomechanical measures. Three-dimensional kinematics, joint moments, electromyography, and centre-of-pressure trajectories are useful research outcomes, but they are not readily available in many clinical environments. The immediate clinical implication is therefore not that clinicians should measure all laboratory variables routinely, but that rehabilitation assessment should consider whether simpler, clinically feasible tasks can capture the same movement-control problems. Future research should identify and validate clinically feasible proxies for the task-specific deficits identified here, such as observable frontal-plane ankle control during step-down, hop-landing, jump-landing, or change-of-direction tasks; field-based landing-quality scores; balance or reach tests with clear movement-quality criteria; and wearable or video-based measures where available. Rehabilitation implications should likewise be framed around trainable movement capacities rather than laboratory signals alone. For example, reduced pre-landing peroneal activation should be interpreted as evidence supporting exercises that challenge timely lateral ankle stabilisation, perturbation response, landing control, and sport-specific deceleration, not as a recommendation that clinicians directly train EMG timing without appropriate instrumentation. Programmes may therefore combine ankle strength and range of motion work with progressive perturbation, landing-retraining, hip and trunk control, and change-of-direction tasks. For mechanobiology and clinical translation, future primary studies should improve comparability by using standardised definitions of CAI and copers, including previous lateral ankle sprain history, recurrent giving way, perceived instability, self-reported function, mechanical laxity where assessed, time since injury, limb dominance, and physical activity level. Task protocols also require clearer reporting of walking or running speed, footwear, surface, landing height, approach velocity, jump-landing constraints, perturbation timing, cutting angle, tested limb, familiarisation procedures, and whether the assessment is feasible in laboratory, field, or routine clinical settings. More complete biomechanical reporting is needed to define reproducible mechanical inputs and targets for future synthesis. Studies should provide means, standard deviations, sample sizes, effect sizes, confidence intervals, time windows, joint coordinate conventions, normalisation procedures, and raw or supplementary data where possible. Evidence should also expand beyond landing tasks, especially for running, cutting, change of direction, perturbation, and sport-specific manoeuvres. Greater attention to female athletes, sex-specific analyses, sport subgroups, competitive level, and prospective cohorts after initial lateral ankle sprain is also needed. Future systematic reviews and meta-analyses should follow stronger methodological standards, including prospective protocol registration, duplicate screening and extraction, lists of excluded studies with reasons, reporting of primary study funding, publication bias assessment where feasible, and explicit evaluation of primary study overlap. These improvements would help distinguish robust task-specific biomechanical targets from findings driven by methodological limitations, heterogeneity, or repeated inclusion of the same primary studies. Strengths and limitations Our umbrella review provides a broad task-based synthesis of systematic reviews and meta-analyses on lower limb biomechanics in CAI. By organising evidence across locomotor, gait initiation, landing, perturbation, sport manoeuvre, proximal adaptation, coper comparison, functional electromyography, and supplementary dual task domains, it clarifies where evidence is concentrated, where findings converge, and where gaps remain. A further strength is the integration of evidence mapping with methodological appraisal. AMSTAR 2 assessed methodological quality, ROBIS assessed review level risk of bias, and modified GRADE evaluated certainty at the task outcome level. Primary study overlap was also assessed using citation matrices and CCA. Together, these approaches allow recurring findings to be interpreted in relation to review quality, certainty, and independence of the underlying evidence. Several limitations should also be considered. As with all umbrella reviews, conclusions depend on the scope, quality, search strategies, eligibility criteria, and reporting completeness of the included reviews. Although multiple major databases and a supplementary publisher platform were searched, relevant reviews may have been missed because of indexing differences, terminology variation, publication bias, or incomplete reporting in titles and abstracts. The evidence base was heterogeneous in CAI definitions, participant characteristics, comparator groups, task protocols, biomechanical variables, and outcome reporting. This limited direct comparison across reviews. The present review also did not reanalyse primary study data. Quantitative estimates were extracted as reported by the original meta-analyses and interpreted as review level findings. Therefore, this synthesis could not adjust pooled effects for overlap, harmonise effect metrics, or update estimates using primary study data. Although overall overlap was moderate, overlap was higher within several task clusters, meaning that convergence across reviews may partly reflect repeated primary studies rather than independent replication. AMSTAR 2, ROBIS, and modified GRADE also involve reviewer judgement. The dual task review was retained only as supplementary evidence because it focused on cognitive motor gait and postural control rather than core joint level biomechanics. Its conclusions should therefore be considered contextual and not directly comparable with the main biomechanical task domains. Conclusions Current review-level evidence suggests that CAI may involve task-specific mechanical and neuromuscular alterations rather than uniform lower-limb dysfunction. The most consistent signals were observed during landing and perturbation tasks, with possible changes in preparatory peroneal activation, vertical loading, sagittal-plane shock absorption, and inversion control. Coper evidence may point to preserved coordination as a clinically relevant comparator. Because certainty remains low to very low and primary-study overlap is present, these findings should be treated as translational targets for future assessment research and rehabilitation development rather than definitive causal mechanisms or immediately deployable clinical measurement standards. Better standardised task protocols, biomechanical reporting, and clinically feasible proxy measures are needed to strengthen future review-level evidence. Ethical approval This study does not contain any studies with human or animal subjects performed by any of the authors. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Data availability All data generated or analysed during this study are included in this article and its supplementary files. Declaration of generative AI and AI-assisted technologies in the manuscript preparation process During the preparation of this work, the authors used OpenAI tools to improve language readability. After using these tools, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article. CRediT authorship contribution statement Bojie Xuan: Conceptualization, Formal analysis, Investigation, Methodology, Visualization, Writing – original draft. Dong Sun: Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Supervision, Writing – review & editing. Yang Song: Conceptualization, Investigation, Methodology, Supervision, Writing – review & editing. Fengping Li: Investigation, Methodology, Visualization, Writing – review & editing. Zhanyi Zhou: Investigation, Methodology, Visualization, Writing – review & editing. Zilong Wang: Investigation, Methodology, Writing – review & editing. Jiazhong Zhu: Investigation, Methodology, Writing – review & editing. Kuiyu Chen: Investigation, Methodology, Writing – review & editing. Rui Zhao: Investigation, Methodology, Writing – review & editing. Dongxu Wang: Investigation, Methodology, Writing – review & editing. Diwei Chen: Investigation, Methodology, Writing – review & editing. Danica Janicijevic: Conceptualization, Supervision, Writing – review & editing. Monèm Jemni: Conceptualization, Supervision, Writing – review & editing. Yaodong Gu: Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Supervision, Writing – review & editing.