The Effect of Ankle–Foot Orthoses and Their Characteristics on Balance, and Postural Control in Children With Cerebral Palsy: A Systematic Review

ABSTRACT

Background and Aims

Cerebral palsy (CP) frequently compromises postural control and balance in children. Ankle–foot orthoses (AFOs) are routinely prescribed to improve stability, yet the influence of specific design parameters—stiffness, alignment, and individualized tuning—on balance outcomes remains inadequately characterized.

Methods

Adhering to PRISMA 2020 guidelines and registered in PROSPERO (CRD420251170932), this systematic review searched six databases (PubMed, Scopus, Web of Science, Embase, CINAHL, and Cochrane Library) up to October 10, 2025. Two reviewers independently screened interventional and observational studies assessing AFO effects on balance or postural control in children with CP. Methodological quality was appraised using Joanna Briggs Institute tools, evidence certainty graded via GRADE, and findings synthesized qualitatively across AFO types, assessment methods, and outcome domains.

Results

A total of 15 studies (n = 390) were included, predominantly involving ambulatory children with spastic diplegia (GMFCS I–III). Overall evidence certainty was very low for static and dynamic balance. Ground‐reaction AFOs consistently enhanced static stability versus solid AFOs. Hinged AFOs yielded inconsistent dynamic balance effects. Individualized tuning of the AFO‐footwear interface emerged as pivotal for optimizing postural control. Conversely, isolated stiffness modifications and prescribed wearing schedules showed variable or detrimental impacts.

Conclusion

AFOs may improve balance in children with CP, but evidence is constrained by methodological heterogeneity and very low certainty. Optimal outcomes appear contingent upon individualized configuration—integrating biomechanical tuning, stiffness modulation, footwear compatibility, and context‐sensitive wearing protocols. High‐quality, parameter‐specific trials are imperative to advance precision orthotic prescription and maximize functional stability.

Summary

• What's known: Ankle‐foot orthoses (AFOs) are routinely prescribed to improve stability in children with cerebral palsy, yet prior systematic reviews report conflicting evidence regarding their effects on balance despite moderate evidence for gait improvements.

• What's new: This review demonstrates that orthotic effectiveness depends not on AFO type alone but on individualized configuration—particularly shank‐to‐vertical angle tuning (> 10°), stiffness modulation, and footwear integration.

• Clinical implications: Precision orthotic prescription should prioritize individualized AFO–footwear tuning over generic device selection, recognizing that excessive stiffness may impair mediolateral stability and that active, context‐specific use may outweigh prolonged passive wear duration.

1. Introduction

Cerebral palsy (CP) is recognized as the most common cause of motor disability in childhood, representing a significant public health issue globally [1, 2]. The prevalence of CP is estimated to range from 1.5 to over 4 per 1000 live births or 10,000 children, with variations across different geographical regions and socioeconomic statuses [1]. By definition, CP is a group of permanent disorders affecting the development of movement and posture, which are attributed to nonprogressive disturbances that occurred in the developing fetal or infant brain [3]. The primary motor impairments, which include alterations in muscle tone, coordination, and postural control, are often accompanied by a range of comorbidities such as disturbances of sensation, cognition, communication, and perception, as well as epilepsy [2].

Among the spectrum of motor dysfunctions in CP, deficits in postural control and balance are hallmark features that profoundly limit functional independence [4]. These deficits arise from a complex interplay of neuromuscular impairments, including spasticity, muscle weakness, and impaired motor coordination, which disrupt the central nervous system's ability to regulate the body's center of mass (CoM) over its base of support [5]. Biomechanically, the ankle joint is pivotal for maintaining stability, serving as the primary axis for the “ankle strategy”—a key postural response involving fine, distal adjustments to control body sway [6]. In children with CP, however, factors such as spasticity in the plantar flexor muscles and weakness of the dorsiflexors severely compromise this mechanism, leading to delayed or poorly scaled postural responses and an over‐reliance on less efficient hip or stepping strategies [7]. Functionally, these impairments have cascading consequences, hindering the performance of essential daily activities such as stable standing, independent walking, and participation in recreational play [8]. Consequently, the increased risk of falls and reduced mobility can lead to activity limitations and participation restrictions, negatively impacting the overall quality of life for both the child and their family [9].

To address these neuromuscular and biomechanical deficits, ankle–foot orthoses (AFOs) are among the most frequently prescribed interventions in the clinical management of children with CP [10]. The primary goals of AFO prescription are to correct joint alignment, provide mediolateral stability to the foot and ankle complex, manage hypertonia such as spasticity, and ultimately, improve the efficiency and safety of gait [11]. However, AFOs are not a monolithic intervention; they encompass a wide spectrum of designs, including rigid, hinged (articulated), and dynamic or flexible models, each possessing distinct mechanical properties [12]. The therapeutic outcome of an AFO is critically dependent on its specific design characteristics—such as its stiffness, trim lines, and ankle angle—and the precise tuning of the orthosis‐footwear combination [13]. These features directly influence the user's postural control by altering sensory feedback and modifying biomechanical constraints, highlighting that the effectiveness of the intervention is intrinsically linked to its design and customization [13].

Despite the established benefits of AFOs for improving gait parameters in children with CP, their specific effects on balance and postural control remain equivocal [14]. Previous systematic reviews have highlighted this ambiguity; while there is moderate‐to‐strong evidence that AFOs enhance gait kinematics and gross motor function, these reviews concurrently report insufficient or conflicting evidence to form definitive conclusions about their impact on balance [11, 14]. The existing evidence suggests that outcomes may be design‐dependent. For instance, rigid AFOs may enhance static stability at the cost of limiting the ankle movements required for dynamic postural adjustments, whereas hinged AFO (HAFO) might better facilitate dynamic balance by allowing controlled ankle motion [11].

Critically, a significant gap in the literature is the lack of a systematic synthesis focusing on how specific AFO design characteristics—such as stiffness, alignment, and the tuning of the orthosis‐footwear combination—modulate balance and postural control outcomes [12]. The current body of evidence is further fragmented by a preponderance of primary studies with small sample sizes, methodological variability, and heterogeneous patient populations (e.g., different CP types and functional levels), which complicates the synthesis of results and limits their generalizability [15]. Therefore, a comprehensive and systematic review is needed to specifically investigate the influence of AFOs and their distinct characteristics on balance and postural control.

2. Methods

2.1. Protocol and Registration

This systematic review was designed, conducted, and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) 2020 statement (the completed PRISMA 2020 checklist is provided in Supporting Information: File 1) [16]. The protocol for this review was prospectively registered with the International Prospective Register of Systematic Reviews (PROSPERO) (Registration number: CRD420251170932). All objectives, eligibility criteria, and synthesis methods were prespecified in the registered protocol and followed throughout the study.

2.2. Eligibility Criteria

This systematic review included all interventional and observational studies published up to October 10, 2025, that investigated the effects of any type of AFO on measures of balance or postural control in children diagnosed with cerebral palsy. The search was conducted without imposing any language restrictions to ensure a comprehensive retrieval of all relevant literature.

Studies were excluded if they were exploratory or pilot studies, case reports, case series, conference proceedings, or secondary literature such as narrative reviews, systematic reviews, or meta‐analyses. This exclusion aligns with methodological guidance for intervention‐focused reviews, as uncontrolled designs cannot isolate causal effects of AFOs on balance outcomes [17]. Furthermore, studies were excluded if the study population was not exclusively children with cerebral palsy, if balance or postural control were not assessed as outcome measures, or if the research evaluated the combined effects of AFOs with other concurrent interventions (e.g., surgery and pharmacological treatments) where the specific contribution of the AFO could not be isolated. Finally, articles for which the full text could not be obtained were also excluded from the final synthesis.

2.3. Information Sources and Search Strategy

A systematic and comprehensive literature search was conducted across six electronic databases—PubMed, Scopus, Web of Science, Embase, CINAHL, and the Cochrane Library—to identify all relevant studies published up to October 10, 2025. The search strategy was meticulously designed using a combination of text keywords and controlled vocabulary terms, such as Medical Subject Headings (MeSH) and Emtree, structured around three core concepts: CP, AFO, and balance/postural control. These concepts were combined using Boolean operators (AND, OR) to ensure a thorough retrieval of literature without any language restrictions. Additionally, the reference lists of included articles and relevant reviews were manually screened to capture any studies missed by the initial database search. To ensure comprehensive retrieval and avoid excluding studies due to inconsistent age classifications, broad age‐related terms (e.g., “children”, “adolescents”, “pediatric”) were deliberately incorporated. This maximized search sensitivity, capturing all relevant literature irrespective of original population labels. The full, detailed search strategy for each database is provided in Table 1.

Table 1.

Search strategy and number of studies retrieved from each database.

Number	Databases	Search strategy	Number of articles retrieved
1	Pubmed	(“ankle‐foot orthosis”[Title/Abstract] OR “ankle foot orthoses”[Title/Abstract] OR “AFO”[Title/Abstract] OR “orthotic device”[Title/Abstract] OR “orthotic treatment”[Title/Abstract] OR “orthotic intervention”[Title/Abstract] OR “foot orthosis”[Title/Abstract] OR “foot orthoses”[Title/Abstract] OR “lower limb orthosis”[Title/Abstract]) AND (“balance”[MeSH Terms] OR “postural balance”[MeSH Terms] OR “balance”[Title/Abstract] OR “postural control”[Title/Abstract] OR “static balance”[Title/Abstract] OR “dynamic balance”[Title/Abstract] OR “balance performance”[Title/Abstract] OR “balance ability”[Title/Abstract] OR “balance test”[Title/Abstract] OR “equilibrium”[Title/Abstract] OR “center of pressure”[Title/Abstract] OR “CoP”[Title/Abstract] OR “sway”[Title/Abstract] OR “stability”[Title/Abstract] OR “limits of stability”[Title/Abstract] OR “force plate”[Title/Abstract] OR “balance platform”[Title/Abstract] OR “posturography”[Title/Abstract] OR “Berg Balance Scale”[Title/Abstract] OR “Pediatric Balance Scale”[Title/Abstract] OR “timed up and go”[Title/Abstract] OR “TUG”[Title/Abstract] OR “functional reach test”[Title/Abstract] OR “FRT”[Title/Abstract]) AND (“cerebral palsy”[MeSH Terms] OR “cerebral palsy”[Title/Abstract] OR “spastic cerebral palsy”[Title/Abstract] OR “hemiplegic cerebral palsy”[Title/Abstract] OR “diplegic cerebral palsy”[Title/Abstract] OR “children with cerebral palsy”[Title/Abstract] OR “pediatric cerebral palsy”[Title/Abstract] OR “CP”[Title/Abstract] OR “GMFCS”[Title/Abstract] OR “child”[MeSH Terms] OR “adolescent”[MeSH Terms] OR “pediatric”[Title/Abstract] OR “pediatrics”[MeSH Terms])	112
2	Scopus	TITLE‐ABS‐KEY(“ankle‐foot orthosis” OR “ankle foot orthoses” OR “AFO” OR “orthotic device” OR “orthotic treatment” OR “orthotic intervention” OR “foot orthosis” OR “foot orthoses” OR “lower limb orthosis”) AND TITLE‐ABS‐KEY(“balance” OR “postural balance” OR “postural control” OR “static balance” OR “dynamic balance” OR “balance performance” OR “balance ability” OR “balance test” OR “equilibrium” OR “center of pressure” OR “CoP” OR “sway” OR “stability” OR “limits of stability” OR “force plate” OR “balance platform” OR “posturography” OR “Berg Balance Scale” OR “Pediatric Balance Scale” OR “timed up and go” OR “TUG” OR “functional reach test” OR “FRT”) AND TITLE‐ABS‐KEY(“cerebral palsy” OR “spastic cerebral palsy” OR “hemiplegic cerebral palsy” OR “diplegic cerebral palsy” OR “children with cerebral palsy” OR “pediatric cerebral palsy” OR “CP” OR “GMFCS” OR “child” OR “adolescent” OR “pediatric”)	380
3	Embase	(‘ankle‐foot orthosis’/exp OR ‘ankle‐foot orthosis’:ti,ab, kw,de OR ‘ankle foot orthoses’:ti,ab,kw,de OR ‘AFO’:ti,ab,kw,de OR ‘orthotic device’:ti,ab,kw,de OR ‘orthotic treatment’:ti,ab,kw,de OR ‘orthotic intervention’:ti,ab,kw,de OR ‘foot orthosis’:ti,ab,kw,de OR ‘foot orthoses’:ti,ab,kw,de OR ‘lower limb orthosis’:ti,ab,kw,de) AND (‘balance’/exp OR ‘postural balance’/exp OR ‘balance’:ti,ab,kw,de OR ‘postural control’:ti,ab,kw,de OR ‘static balance’:ti,ab,kw,de OR ‘dynamic balance’:ti,ab,kw,de OR ‘balance performance’:ti,ab,kw,de OR ‘balance ability’:ti,ab,kw,de OR ‘balance test’:ti,ab,kw,de OR ‘equilibrium’:ti,ab,kw,de OR ‘center of pressure’:ti,ab,kw,de OR ‘CoP’:ti,ab,kw,de OR ‘sway’:ti,ab,kw,de OR ‘stability’:ti,ab,kw,de OR ‘limits of stability’:ti,ab,kw,de OR ‘force plate’:ti,ab,kw,de OR ‘balance platform’:ti,ab,kw,de OR ‘posturography’:ti,ab,kw,de OR ‘berg balance scale’:ti,ab,kw,de OR ‘pediatric balance scale’:ti,ab,kw,de OR ‘timed up and go’:ti,ab,kw,de OR ‘TUG’:ti,ab,kw,de OR ‘functional reach test’:ti,ab,kw,de OR ‘FRT’:ti,ab,kw,de) AND (‘cerebral palsy’/exp OR ‘cerebral palsy’:ti, ab, kw, de OR ‘spastic cerebral palsy’:ti,ab,kw,de OR ‘hemiplegic cerebral palsy’:ti,ab,kw,de OR ‘diplegic cerebral palsy’:ti,ab,kw,de OR ‘children with cerebral palsy’:ti,ab,kw,de OR ‘pediatric cerebral palsy’:ti,ab,kw,de OR ‘CP’:ti,ab,kw,de OR ‘GMFCS’:ti,ab,kw,de OR ‘child’/exp OR ‘adolescent’/exp OR ‘child’:ti,ab,kw,de OR ‘adolescent’:ti,ab,kw,de OR ‘pediatric’:ti,ab,kw,de)	283
4	Web of Science	TS = (“ankle‐foot orthosis” OR “ankle foot orthoses” OR “AFO” OR “orthotic device” OR “orthotic treatment” OR “orthotic intervention” OR “foot orthosis” OR “foot orthoses” OR “lower limb orthosis”) AND TS = (“balance” OR “postural balance” OR “postural control” OR “static balance” OR “dynamic balance” OR “balance performance” OR “balance ability” OR “balance test” OR “equilibrium” OR “center of pressure” OR “CoP” OR “sway” OR “stability” OR “limits of stability” OR “force plate” OR “balance platform” OR “posturography” OR “Berg Balance Scale” OR “Pediatric Balance Scale” OR “timed up and go” OR “TUG” OR “functional reach test” OR “FRT”) AND TS = (“cerebral palsy” OR “spastic cerebral palsy” OR “hemiplegic cerebral palsy” OR “diplegic cerebral palsy” OR “children with cerebral palsy” OR “pediatric cerebral palsy” OR “CP” OR “GMFCS” OR “child” OR “adolescent” OR “pediatric”)	113
5	CINAHL	((TI “ankle‐foot orthosis” OR TI “ankle foot orthoses” OR TI “AFO” OR TI “orthotic device” OR TI “orthotic treatment” OR TI “orthotic intervention” OR TI “foot orthosis” OR TI “foot orthoses” OR TI “lower limb orthosis”) OR (AB “ankle‐foot orthosis” OR AB “ankle foot orthoses” OR AB “AFO” OR AB “orthotic device” OR AB “orthotic treatment” OR AB “orthotic intervention” OR AB “foot orthosis” OR AB “foot orthoses” OR AB “lower limb orthosis”)) AND ((TI “balance” OR TI “postural balance” OR TI “postural control” OR TI “static balance” OR TI “dynamic balance” OR TI “balance performance” OR TI “balance ability” OR TI “balance test” OR TI “equilibrium” OR TI “center of pressure” OR TI “CoP” OR TI “sway” OR TI “stability” OR TI “limits of stability” OR TI “force plate” OR TI “balance platform” OR TI “posturography” OR TI “Berg Balance Scale” OR TI “Pediatric Balance Scale” OR TI “timed up and go” OR TI “TUG” OR TI “functional reach test” OR TI “FRT”) OR (AB “balance” OR AB “postural balance” OR AB “postural control” OR AB “static balance” OR AB “dynamic balance” OR AB “balance performance” OR AB “balance ability” OR AB “balance test” OR AB “equilibrium” OR AB “center of pressure” OR AB “CoP” OR AB “sway” OR AB “stability” OR AB “limits of stability” OR AB “force plate” OR AB “balance platform” OR AB “posturography” OR AB “Berg Balance Scale” OR AB “Pediatric Balance Scale” OR AB “timed up and go” OR AB “TUG” OR AB “functional reach test” OR AB “FRT”)) AND ((TI “cerebral palsy” OR TI “spastic cerebral palsy” OR TI “hemiplegic cerebral palsy” OR TI “diplegic cerebral palsy” OR TI “children with cerebral palsy” OR TI “pediatric cerebral palsy” OR TI “CP” OR TI “GMFCS” OR TI “child” OR TI “adolescent” OR TI “pediatric”) OR (AB “cerebral palsy” OR AB “spastic cerebral palsy” OR AB “hemiplegic cerebral palsy” OR AB “diplegic cerebral palsy” OR AB “children with cerebral palsy” OR AB “pediatric cerebral palsy” OR AB “CP” OR AB “GMFCS” OR AB “child” OR AB “adolescent” OR AB “pediatric”) OR (MH “Cerebral Palsy”))	84
6	Cochrane Library	(“ankle‐foot orthosis”:ti,ab,kw OR “ankle foot orthoses”:ti,ab,kw OR AFO:ti,ab,kw OR “orthotic device”:ti,ab,kw OR “orthotic treatment”:ti,ab,kw OR “orthotic intervention”:ti,ab,kw OR “foot orthosis”:ti,ab,kw OR “foot orthoses”:ti,ab,kw OR “lower limb orthosis”:ti,ab,kw) AND (“balance”:ti,ab,kw OR “postural balance”:ti,ab,kw OR “postural control”:ti,ab,kw OR “static balance”:ti,ab,kw OR “dynamic balance”:ti,ab,kw OR “balance performance”:ti,ab,kw OR “balance ability”:ti,ab,kw OR “balance test”:ti,ab,kw OR “equilibrium”:ti,ab,kw OR “center of pressure”:ti,ab,kw OR “CoP”:ti,ab,kw OR “sway”:ti,ab,kw OR “stability”:ti,ab,kw OR “limits of stability”:ti,ab,kw OR “force plate”:ti,ab,kw OR “balance platform”:ti,ab,kw OR “posturography”:ti,ab,kw OR “Berg Balance Scale”:ti,ab,kw OR “Pediatric Balance Scale”:ti,ab,kw OR “timed up and go”:ti,ab,kw OR TUG:ti,ab,kw OR “functional reach test”:ti,ab,kw OR FRT:ti,ab,kw) AND (“cerebral palsy”:ti,ab,kw OR “spastic cerebral palsy”:ti,ab,kw OR “hemiplegic cerebral palsy”:ti,ab,kw OR “diplegic cerebral palsy”:ti,ab,kw OR “children with cerebral palsy”:ti,ab,kw OR “pediatric cerebral palsy”:ti,ab,kw OR CP:ti,ab,kw OR GMFCS:ti,ab,kw OR child:ti,ab,kw OR adolescent:ti,ab,kw OR pediatric:ti,ab,kw)	55

2.4. Study Selection

Following the literature search, all retrieved citations were imported into EndNote 2025, and duplicate records were removed. The study selection process was independently conducted by two reviewers (ZG and MA) in two distinct phases. In the first phase, the reviewers screened the titles and abstracts of all unique records against the predefined eligibility criteria. In the second phase, the full texts of the potentially relevant articles were retrieved and assessed for final inclusion. The inter‐rater reliability for both phases was calculated using Cohen's kappa coefficient to quantify the level of agreement. Any disagreements between the reviewers at either stage were resolved through discussion and consensus.

2.5. Data Extraction

Two reviewers (ZG and MA) independently performed the data extraction from all included studies using a standardized and pre‐piloted data form. This form was designed to systematically capture comprehensive information, including: study characteristics (author, year, and design); participant demographics and clinical details (sample size, age, CP topography, GMFCS level); detailed intervention specifics (AFO type, material, tuning parameters, and wearing schedule); outcome data (type of balance measured, assessment tools, and key findings); and methodological information (adherence, adverse effects, and limitations). To ensure accuracy and consistency, the extracted data were cross‐checked between the reviewers, and any discrepancies were resolved through discussion and consensus.

2.6. Risk of Bias Assessment

The risk of bias and methodological quality of all included studies were independently assessed by two reviewers (ZG and MA) using the Joanna Briggs Institute (JBI) critical appraisal tools. The choice of tool was tailored to the specific design of each study; the JBI Checklist for Randomized Controlled Trials, the JBI Checklist for Quasi‐Experimental Studies, and the JBI Checklist for Analytical Cross‐Sectional Studies were used to evaluate biases relevant to each methodology [18]. Inter‐rater reliability for this process was determined using Cohen's kappa coefficient, and any discrepancies between the reviewers were resolved through discussion to reach a consensus.

For the scoring process, each criterion on the checklists was assigned a 1 if the study met the standard or a 0 if it was not met or unclear. A total quality score was then calculated for each article by determining the percentage of criteria that were met. Based on this final score, studies were categorized into three levels of methodological quality: high quality (score ≥ 75%), moderate quality (score 50%–74%), or low quality (score < 50%) [19].

2.7. Grading the Quality of Evidence

The overall certainty of the evidence for each primary outcome was independently graded by two reviewers (ZG and MA) using the Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) framework. Inter‐rater reliability for the final quality ratings was calculated using Cohen's kappa coefficient, and any remaining disagreements were resolved through discussion to reach a consensus [20]. According to this approach, evidence from randomized controlled trials (RCTs) starts as High quality and evidence from observational studies starts as Low quality. This initial rating was then adjusted by assessing five domains that can downgrade the quality (risk of bias, inconsistency, indirectness, imprecision, and publication bias). The final quality of evidence was rated as High, Moderate, Low, or Very Low [20].

2.8. Synthesis of Results

Given the substantial methodological and clinical heterogeneity across the included studies, a quantitative meta‐analysis was not feasible. This decision is consistent with Cochrane guidelines, which advise against statistical pooling when substantial variability precludes a valid quantitative synthesis [21]. Heterogeneity was evident in several key domains, including AFO design characteristics, participant functional levels (GMFCS and topographic distribution), and the outcome measurement tools employed for balance and postural control. Consequently, a structured qualitative synthesis was conducted to summarize and interpret the evidence.

To facilitate a structured analysis and comparison, the extracted data were systematically organized and grouped based on several key factors. The findings were primarily categorized by the type of AFOs (e.g., rigid, hinged, and dynamic), the type of balance assessed (static vs. dynamic), and the balance measurement method employed (e.g., force plate posturography and clinical balance scales). This approach ensures that the clinical significance and precision of orthotic interventions are evaluated beyond simple statistical thresholds, facilitating a robust examination of how specific biomechanical features influence postural control despite study heterogeneity.

2.9. Statistical Analysis

Following the SAMPL guidelines, inter‐rater reliability for study selection and quality assessment was quantified using Cohen's kappa coefficient [22]. All calculations were performed using IBM SPSS Statistics (Version 28.0; IBM Corp., Armonk, NY, USA). For the interpretation of primary study outcomes, an a priori significance level of 0.05 with two‐sided tests was adopted as the standard benchmark. Where available, the synthesis prioritized the extraction of exact p‐values, effect sizes, and 95% confidence intervals (CIs) to evaluate the magnitude and precision of orthotic interventions beyond simple statistical thresholds [23].

3. Results

3.1. Study Selection

The search across six electronic databases initially yielded 1027 records. After removing 309 duplicates and 85 review articles, 633 titles and abstracts were screened. Following the exclusion of 587 irrelevant records, the full texts of the remaining 46 studies, all of which were in English, were assessed for eligibility. Inter‐reviewer agreement was substantial (Cohen's kappa = 0.70), and 15 studies met the inclusion criteria (Figure 1).

Figure 1.

PRISMA flow diagram illustrating the selection process of studies included in the systematic review.

3.2. Study Characteristics

A total of 15 studies with heterogeneous designs were included in this review, comprising three RCTs, six cross‐sectional studies, and a variety of quasi‐experimental and repeated‐measures designs, including randomized cross‐over models. Key study characteristics, including methodological design, participant demographics, and detailed AFO specifications, are summarized in Table 2. The balance outcome measures, assessment protocols, and synthesized comparative findings for each included study are presented in Table 3.

Table 2.

Study characteristics, participant demographics, and ankle–foot orthosis specifications of included studies.

NO	Author & year (Ref.)	Study design	Follow‐up	Aim	Participants, n (age, y ± SD)	CP topography (GMFCS level)	AFO Type	Tuning Parameters	Material	Wearing schedule & h/day	Footwear compatibility	Other treatments	Adherence & adverse effects	Retention rate
1	Bjornson et al. 2024 [24]	RCT crossover	3 mos	Compare individualized vs. nonindividualized AFO‐FC on balance/mobility	21 (6.9 ± 1.9)	Diplegia (II–III)	Rigid AFO‐FC (IAFD vs. NAFD)	IAFD: custom ankle/SVA > 10°, shoe mods; NAFD: 90° ankle	Rigid (Cascade)	≥ 8 h/d instructed; ≥ 8 h/d reported	Standardized skate shoes (modified vs. unmodified)	None concurrent	Parental report ≥ 8 h/d; mild/moderate skin issues; 2 withdrawals	90.5%
2	Wang et al. 2022 [25]	Crossover	6 wks	Compare HAFO vs. FRAFO on static stability & STS	9 (11.1 ± 3.4)	Diplegia (II–III)	HAFO vs. FRAFO	FRAFO: anterior shell for knee ext.; HAFO: hinged DF	FRAFO: 3 mm PP; HAFO: ‐	≥ 4 h/d instructed; actual ‐	—	Prior HAFO use (6 wks); no concurrent Tx	Instructed wear reported; no major AEs	100%
3	Sanad 2021 [26]	RCT	3 mos	Compare solid AFO vs. GRAFO on balance	30 (7.1 ± 0.8)	Diplegia (‐)	Solid AFO vs. GRAFO	—	3 mm PP (post. solid, ant. GRAFO)	Awake except 1 h/d	—	Regular PT (1 h/session, 3×/wk)	Instruction given; no quantitative adherence; no AEs	100%
4	Unes et al. 2021 [27]	Cross‐sectional	—	Compare outcomes by daily AFO wearing time	80 (7.4 ± 3.9)	Uni. & Diplegia (II–III)	Custom PP AFOs (neutral ankle/subtalar)	“Tuned by prescriber”; no protocol	2.5 mm PP, 3 Velcro straps	6–12 h/d (mean 6.1) versus 12–24 h/d (mean 14.1)	—	Excl. recent surgery/Botox	Parental report; no systematic AE monitoring	100%
5	Limpaninlachat et al. 2021 [28]	Cross‐sectional	—	Compare SAFO vs. HAFO on function/activity	26 (9.3 ± 0.5)	Diplegia (II–III)	SAFO vs. HAFO	—	—	> 5 days/wk	—	Excl. surgery (12 mos) & Botox (6 mos)	—	100%
6	Goihl et al. 2021 [29]	Clinical controlled	—	Assess AFO effects on trunk acceleration & walking energy cost	19 (9.6 ± 2.7)	Unilateral (III)	Mixed (carbon n = 4, thermoplastic n = 12, rigid n = 2, functional joint n = 1)	—	Thermoplastic & carbon composite	Tested during walking	Tested “shoes only” vs. “shoes+AFO”	None specified	No fitting/functional problems; no AEs	100%†
7	Bahar‐Özdemir et al. 2021 [30]	Cross‐sectional	—	Compare hinged vs. solid AFOs on balance/gait in hemiplegic CP	42 (9.5 ± 2.2)	Unilateral (III)	Solid (n = 11) versus Hinged (n = 8)	Solid: 0° DF; Hinged: free DF, PF restricted at 0°	Semi‐rigid PP	—	—	Excl. surgery/Botox (6 mos)	—	—
8	Meyns et al. 2020 [31]	Repeated‐measures	4–6 wks	Examine vAFO stiffness effects on gait stability & energy cost	15 (10.0 ± 2.0)	‐ (I–III)	Carbon vAFO with adjustable hinge	3 configs: flexible, stiff, rigid (Nm/deg & ROM quantified)	Full‐carbon prepreg	Tested during gait	Sneakers (flat, flexible soles)	None specified	1 refused rigid; 1 pressure ulcer (flexible) → excluded	86.6%
9	Sanad et al. 2018 [32]	RCT	—	Compare immediate effect of solid AFO vs. GRAFO on balance	30 (6.4 ± 1.7)	Diplegia (‐)	Solid AFO vs. GRAFO	—	3 mm PP (custom‐molded)	During testing	Barefoot during testing	Excl. surgery (6 mos)	—	100%
10	Olama, Khaled et al. 2013 [33]	RCT	6 mos	Evaluate “three‐side support AFO” on standing balance	30 (4.8 ± 0.8)	Diplegia (‐)	Modified 3‐side support AFO + metal supports	—	PP + light metal supports	30 min/session (3×/wk) + ≥ 1 h/d home	—	Identical therapeutic exercise (both groups)	—	100%
11	Degelean et al. 2012 [34]	Cross‐sectional	—	Analyze AFO effect on trunk control & lower‐limb coordination	20 (7.6 ± 1.7)	Diplegia (I–II)	PLS (GMFCS I) vs. Solid (GMFCS II)	Ankle 0° DF, prevents PF	4.8 mm PP	During testing	Inside own shoes	NDT 2–3×/wk; excl. Botox (1 yr)	—	100%
12	Bahramizadeh et al. 2012 [35]	Quasi‐experimental	4 wks	Determine FRAFO effect on static postural control	8 (8.1 ± 2.4)	Diplegia (I–II)	FRAFO with adjustable ankle hinge	Ankle angle ↓≥ 5°/wk to neutral for final test	3 mm PP (custom‐molded)	During testing	“Braced footwear”	—	—	100%
13	Rha et al. 2010 [36]	Cross‐sectional	—	Assess static balance & postural control mechanisms with hinged AFOs	21 (6.1 ± 1.1)	Diplegia (I–III)	Hinged AFO (PF stop, free DF)	Blocked PF, allowed free DF	3 mm PP	During testing	—	—	—	100%
14	Wesdock et al. 2003 [37]	Repeated‐measures	8 wks	Evaluate solid AFO & wedged shoes on standing balance duration	11 (7.0 ± 2.6)	Uni. & Diplegia (II–IV)	Bilateral solid AFOs (+ wedged shoes condition)	Neutral 90° ankle; wedges individualized (avg 2.4 cm)	—	≥ 6 h/d	Athletic shoes (wedges on 2nd pair)	Ongoing school/private PT (uncontrolled)	—	100%
15	Kott et al. 2003 [38]	RCT crossover	—	Examine orthoses effect on upright functional skills	28 (10.6 ± 4.5)	Uni. & Diplegia (I–II)	Hinged, solid, GRAFO, SMOs	—	—	—	—	—	—	60.7%

Abbreviations: AE, adverse event; AFO, ankle–foot orthosis; AFO‐FC, AFO‐footwear combination; CP, cerebral palsy; DF, dorsiflexion; FRAFO, floor‐reaction AFO; GMFCS, Gross Motor Function Classification System; GRAFO, ground‐reaction AFO; HAFO, hinged AFO; IAFD, individualized alignment/footwear design; mos, months; NAFD, nonindividualized alignment/footwear design; NDT, neurodevelopmental treatment; NR, not reported; PF, plantarflexion; PLS, posterior leaf spring AFO; PP, polypropylene; PT, physical therapy; RCT, randomized controlled trial; SAFO, solid AFO; SMO, supramalleolar orthosis; STS, sit‐to‐stand; SVA, shank‐to‐vertical angle; Tx, treatment; Uni., unilateral; vAFO, ventral shell AFO; wks, weeks; y, years.

Symbols: – = range; ‐ = not reported/not applicable; ≥ = greater than or equal to; ↓ = decreased. Values: n (mean ± SD) unless specified.

^† Retention for analytical sample.

Table 3.

Outcome measures, assessment tools & key findings on balance and postural control.

No.	Author & year (Ref.)	Type of balance	Balance measurement tool	Key findings	Control group intervention	Effect direction & magnitude	Statistical Significance	Clinical Relevance Note
1	Bjornson et al. 2024 [24]	Static & dynamic	Pediatric Balance Scale (PBS)	AFO‐FC/IAFD outperformed AFO‐FC/NAFD in total, static, and dynamic subscales	AFO‐FC/NAFD (90° alignment, unmodified shoes)	↑ IAFD > NAFD (large magnitude; exceeded MCID)	p = 0.03 (total); p = 0.042 (static); p = 0.009 (dynamic)	Individualized tuning/footwear yields clinically meaningful balance & mobility gains
2	Wang et al. 2022 [25]	Static & dynamic	3D motion analysis + force plates (CoP, STS kinetics)	No static CoP differences; FRAFO enhanced vertical GRF during STS	HAFO (crossover baseline)	↔ static stability; ↑ FRAFO > HAFO (GRF)	p > 0.05 (static); p = 0.018 (GRF)	FRAFO improves sit‐to‐stand force generation but not quiet‐standing stability
3	Sanad 2021 [26]	Static	Biodex Balance System (A‐P & M‐L indices)	Both improved post‐PT; GRAFO yielded superior post‐treatment indices	Solid AFO + identical PT	↑↑ GRAFO > SAFO (31.5% vs 15.4% A‐P; 43.9% vs 16.3% M‐L)	p < 0.05 (within & between groups)	GRAFO provides greater static stability gains over 3 months
4	Unes et al. 2021 [27]	Dynamic	Pediatric Berg Balance Scale (PBBS)	No balance difference by wear time; longer wear correlated with lower QoL	Within‐group (6–12 h/d vs. 12–24 h/d)	↔ PBBS; ↓ longer wear	p = 0.196 (PBBS); p < 0.001 (correlations)	Extended daily wear ( > 12 h) does not enhance balance and may reduce satisfaction
5	Limpaninlachat et al. 2021 [28]	Dynamic	Functional Reach Test (FRT)	HAFO enabled greater forward reach vs SAFO in independent walkers	SAFO	↑ HAFO > SAFO (subgroup only)	p < 0.05 (independent walkers)	HAFO benefits dynamic weight‐shifting in higher‐functioning diplegia
6	Goihl et al. 2021 [29]	Dynamic	Trunk accelerometry (Sample Entropy, SE)	AFOs normalized trunk acceleration complexity toward TD patterns	TD normative reference	↑ complexity (normalization)	Significant (p < 0.05 per study)	AFOs improve underlying postural control strategy without altering energy cost
7	Bahar‐Özdemir et al. 2021 [30]	Dynamic	PBS, FTSST, NeuroCom (LOS, WA, STS)	AFOs improved functional balance vs barefoot; no difference between hinged vs solid	Healthy controls (normative)	↑ AFO > Barefoot; ↔ Hinged vs Solid	p < 0.05 (AFO vs. barefoot); ns (between types)	Either hinged or solid AFO suffices for functional balance in hemiplegia
8	Meyns et al. 2020 [31]	Dynamic	3D gait analysis (Margins of Stability, MoS)	vAFOs reduced ML stability; higher stiffness further decreased stability & increased trunk compensation	Shoes‐only condition	↓ ML stability with AFO/stiffness	Significant (p < 0.05 implied)	Excessive AFO stiffness may compromise dynamic stability in flexed‐knee gait
9	Sanad et al. 2018 [32]	Static & Dynamic	Biodex Balance System (Overall Stability Index)	GRAFO produced greater immediate balance improvement than solid AFO	Solid AFO	↑ GRAFO > SAFO (28.97% vs 15.80% improvement)	p < 0.05	GRAFO yields superior immediate postural stability gains
10	Olama, Khaled et al. 2013 [33]	Dynamic	Biodex Stability System (Overall, A‐P, M‐L indices)	AFO+exercise outperformed exercise alone across all indices at 6 months	Exercise‐only program	↑↑ AFO+Ex > Ex (moderate‐large magnitude)	p < 0.05	Modified 3‐side support AFO enhances standing balance when combined with PT
11	Degelean et al. 2012 [34]	Dynamic	3D motion capture (trunk angular velocity)	AFOs increased trunk frontal angular velocity but improved intersegmental coordination	TD normative reference	↓ ML trunk control; ↑ limb coordination	Significant	AFOs may trade off trunk stability for improved lower‐limb coordination during gait
12	Bahramizadeh et al. 2012 [35]	Static	Force platform (CoP excursion & velocity)	FRAFO improved knee alignment but increased ML sway velocity; no CoP displacement gain	Barefoot/TD reference	↔ CoP displacement; ↓ ML stability ( ↑ sway velocity)	Significant (ML velocity)	Neutral FRAFO tuning may compromise mediolateral static stability
13	Rha et al. 2010 [36]	Static	Dual force platforms (CoP)	Hinged AFOs did not reduce CoP magnitude but increased ankle strategy contribution for ML balance	TD reference	↔ sway magnitude; ↑ ankle strategy utilization	Significant (mechanism); ns (magnitude)	Hinged AFOs normalize postural control mechanisms without reducing sway quantity
14	Wesdock et al. 2003 [37]	Static	Stopwatch (standing duration)	WAFO improved duration only in subgroup with baseline > 15 s; no overall group effect	AFO alone/No orthosis	↑ (subgroup only); ↔ (overall)	Significant in subgroup	Wedged shoes benefit static balance only in children with moderate baseline stability
15	Kott et al. 2003 [38]	Dynamic	PBS, Standardized Walking Obstacle Course (SWOC)	No significant differences in PBS scores or obstacle stability with vs. without orthoses	Unbraced condition (crossover)	↔ functional balance & stability	p > 0.05	Orthoses may not confer measurable functional balance advantages in heterogeneous CP

Abbreviations: AFO, ankle–foot orthosis; AFO‐FC, AFO‐footwear combination; A‐P, anteroposterior; CI, confidence interval; CoP, center of pressure; CP, cerebral palsy; Ex, exercise; FRT, Functional Reach Test; FTSST, Five Times Sit‐to‐Stand Test; GRF, ground reaction force; HAFO, hinged AFO; IAFD, individualized alignment/footwear design; LOS, Limits of Stability; MCID, minimal clinically important difference; ML, mediolateral; NAFD, nonindividualized alignment/footwear design; NR, not reported; ns, not significant; PBS, Pediatric Balance Scale; PBBS, Pediatric Berg Balance Scale; PT, physical therapy; SAFO, solid AFO; STS, sit‐to‐stand; SWOC, Standardized Walking Obstacle Course; TD, typically developing; vAFO, ventral shell AFO; WAFO, wedged AFO.

Symbols:↑= improvement/increase; ↓= decrease/impairment; ↔ = no statistically significant difference; ↑↑/↓↓ = large magnitude effect. Statistical significance threshold set at p < 0.05 unless otherwise specified. Effect directions and magnitudes synthesized directly from primary study outcomes.

3.3. Risk of Bias Assessment

The risk of bias assessment revealed generally acceptable methodological quality (Table 4), with substantial inter‐rater agreement (Cohen's κ = 0.70). Five RCTs demonstrated moderate quality (Scores 5–9) [24, 26, 32, 33, 38]. Common weaknesses included inadequate randomization, lack of allocation concealment and blinding, and insufficient reporting of baseline comparability—biases particularly difficult to avoid in orthotic trials due to the visible nature of the devices.

Table 4.

Methodological quality and risk of bias assessment of included studies using the JBI critical appraisal checklists.

No	Author's name and year	JBI checklist items for randomized controlled trials													Overall score	Comments
		1	2	3	4	5	6	7	8	9	10	11	12	13
1	Bjornson et al. 2024 [24]	Yes	No	Yes	No	No	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes	9 (69.2%) Moderate quality	Despite its methodological strengths, such as utilizing true randomization, ensuring baseline similarity between groups, and performing an intent‐to‐treat analysis, the study is limited by a high risk of performance and detection bias due to the lack of blinding, alongside an undescribed method of allocation concealment.
2	Sanad 2021 [26]	Yes	No	Yes	No	No	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes	9 (69.2%) Moderate quality	A high risk of performance bias exists due to the unavoidable lack of blinding for participants and providers. Furthermore, the absence of reported allocation concealment and outcome assessor blinding introduces a considerable risk of selection and detection bias.
3	Sanad et al. 2018 [32]	No	No	Yes	No	No	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes	8 (61.5%) Moderate quality	This study has a high risk of bias due to its quasi‐randomization method (odd/even numbers) and a complete lack of blinding for participants, personnel, and assessors.
4	Olama, Khaled et al. 2013 [33]	No	No	Yes	No	No	No	No	Yes	Yes	Yes	Yes	Yes	Yes	7 (53.8%) Moderate quality	The methods for randomization and allocation concealment were not reported, and there was no blinding. Critically, the intervention group received the AFO plus 30 min of additional, specialized training, which serves as a significant confounder.
5	Kott et al. 2003 [38]	Yes	No	Yes	No	No	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes	9 (69.2%) Moderate quality	The primary methodological limitation, inherent to this intervention, is the unavoidable lack of blinding for participants, administrators, and outcome assessors. Additionally, the method of allocation concealment was not described.

No	Author's name and year	JBI checklist items for Quasi‐experimental studies									Overall score	Comments
		1	2	3	4	5	6	7	8	9
1	Wang et al. 2022 [25]	Yes	Yes	Yes	Yes	No	Yes	Yes	Yes	Yes	8 (88.8%) High quality	The primary limitation is the cross‐sectional design, which only assesses the immediate effects of the orthosis without capturing potential long‐term functional adaptations.
2	Goihl et al. 2021 [29]	Yes	Yes	Yes	Yes	No	Yes	Yes	Yes	Yes	8 (88.8%) High quality	The single criterion not met (long‐term follow‐up) is not relevant to this immediate‐effects study design.
3	Meyns et al. 2020 [31]	Yes	Yes	Yes	Yes	No	Yes	Yes	Yes	Yes	8 (88.8%) High quality	There was no longitudinal follow‐up (e.g., 6 months, 12 months) to assess the effects over time.
4	Bahramizadeh et al. 2012 [35]	Yes	Yes	Yes	Yes	No	Yes	Yes	Yes	Yes	8 (88.8%) High quality	Knee angle was one pre/post measure. Postural control (CoP) was only measured post‐intervention (at 4 weeks), not at baseline.
5	Wesdock et al. 2003 [37]	Yes	Yes	No	No	Yes	Yes	Yes	Yes	Yes	7 (77.8%) High quality	The primary limitations are the lack of a separate control group and a significant confounding bias, as co‐interventions (physical therapy) were neither controlled nor standardized during the study period.

NO	Author's name and year	JBI checklist items for cross‐sectional studies								Overall score	Comments
		1	2	3	4	5	6	7	8
1	Unes et al. 2021 [27]	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	8 (100%) High quality
2	Limpaninlachat et al. 2021 [28]	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	8 (100%) High quality
3	Bahar‐Özdemir et al. 2021 [30]	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	8 (100%) High quality
4	Degelean et al. 2012 [34]	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	8 (100%) High quality
5	Rha et al. 2010 [36]	Yes	Yes	Yes	Yes	No	No	Yes	Yes	6 (75%) High quality	The study fails to identify or statistically control for key confounding variables, such as the mix of GMFCS levels or the slight age difference between groups, relying only on exclusion criteria.

Abbreviation: JBI, Joanna Briggs Institute.

Five quasi‐experimental studies achieved high‐quality scores (7 or 8) [25, 29, 31, 35, 37]. The primary limitation involved inadequate control groups or insufficient management of concurrent interventions (e.g., physical therapy), leading to potential confounding bias. However, these studies were otherwise robust regarding intervention description and follow‐up consistency.

The five cross‐sectional studies also showed high quality (scores 6–8) [27, 28, 30, 34, 36]. Methodological concerns centered on the management of confounding factors (e.g., lack of statistical adjustment for age, GMFCS level, or orthotic experience) and the reliability of exposure measurement (e.g., lack of detail on orthotic use duration). Despite this, they exhibited strong rigor in sampling and analysis, offering valuable complementary evidence to this review.

3.4. Grading the Quality of Evidence

The quality of evidence for each primary outcome was graded using the GRADE framework, with substantial inter‐rater reliability between the two reviewers (Cohen's κ = 0.75). Overall, the certainty of evidence for both static and dynamic balance was rated as very low, primarily due to serious concerns regarding risk of bias, inconsistency, and imprecision.

Regarding static balance, studies exhibited methodological weaknesses—including unclear randomization, lack of blinding, and inadequate allocation concealment—alongside substantial inconsistency in AFO design and measurement tools. Although findings often indicated potential benefits in sway reduction, limited sample sizes led to downgrading due to imprecision [24, 25, 26, 32, 35, 36, 37].

Similarly, the evidence for dynamic balance was rated as very low. Beyond the aforementioned methodological limitations, the heterogeneity of assessment methods—ranging from clinical scales (e.g., Pediatric Balance Scale (PBS), Functional Reach Test (FRT)) to laboratory‐based posturography—resulted in inconsistent outcomes. While some studies reported statistically significant improvements (p < 0.05) with specific AFO designs (e.g., hinged or ventral shell), the lack of uniform effects across the literature and absence of long‐term follow‐up limit the strength of definitive conclusions [24, 25, 27, 28, 29, 30, 31, 32, 33, 34, 38]. The results of this assessment are summarized in Table 5.

Table 5.

GRADE evidence profile.

Outcome	No. of studies assessing outcome	Study design	Risk of bias	Inconsistency	Indirectness	Imprecision	Publication bias	Final certainty
Static balance	7	3 RCTs, 3 Quasi‐experimental, 1 Cross‐sectional	Serious	Serious	Not serious	Serious	Undetected	Very Low
Dynamic balance	11	4 RCTs, 3 Quasi‐experimental, 4 Cross‐sectional	Serious	Serious	Not serious	Serious	Undetected	Very Low

Abbreviation: RCT, randomized control trial.

3.5. Main Outcomes

Based on the 15 included studies (n = 390), participant demographics were categorized according to the reported mean age, with both middle childhood (6–8 years) [24, 26, 27, 32, 34, 36, 37] and late childhood (9–12 years) [25, 28, 29, 30, 31, 35, 38] each represented in seven studies, while only one focused on early childhood (3–5 years) [33]. Spasticity was the predominant topographic type; of the 14 studies reporting this, nine focused on diplegia [24, 25, 26, 28, 32, 33, 34, 35, 36], two on unilateral hemiplegia [29, 30], and three included mixed populations [27, 37, 38]. Functional classification via GMFCS (reported in 12 studies) showed a predominantly ambulatory cohort, with Level II identified in 10 studies [24, 25, 27, 28, 31, 34, 35, 36, 37, 38], Level III in nine [24, 25, 27, 28, 29, 30, 31, 36, 37], Level I in five [31, 34, 35, 36, 38], and Level IV in one study [37].

3.6. Balance Assessment Methods

A total of 15 included studies were bifurcated into two main assessment categories: functional, observation‐based clinical scales and quantitative, instrumented laboratory measures. Ten studies utilized laboratory instrumentation to analyze the biomechanical and neuromuscular components of postural control [25, 26, 29, 30, 31, 32, 34, 35, 36, 38]. The predominant method was force‐platform posturography (n = 6), used to quantify Center of Pressure (CoP) variables—specifically displacement, velocity, and sway area [26, 32, 35, 36, 37, 38]. Other technologies included commercial systems (Biodex [26, 32, 33]; NeuroCom [30]) to derive composite stability indices, 3D motion capture for trunk kinematics and Margins of Stability (MoS) [25, 31, 34], and trunk accelerometry to evaluate the complexity (Sample Entropy) of postural adjustments [29].

In parallel, six studies utilized validated clinical ordinal scales to quantify functional balance performance [24, 27, 28, 30, 33, 38], including the PBS [24, 28, 30, 38], Pediatric Berg Balance Scale (PBBS) [27], FRT [25], Five Times Sit to Stand Test (FTSST) [30], Standardized Walking Obstacle Course (SWOC) [38], and timed static standing duration [37]. One study combined both methodologies, highlighting a key methodological dichotomy: clinical scales measure functional capacity (i.e., task performance), whereas laboratory tools assess the underlying quality or strategy of postural control (e.g., kinematic patterns and stability margins) [30].

3.7. Types of Balance Assessed

Balance outcomes were categorized into static (quiet standing) and dynamic (stability during movement) domains. Static balance was evaluated in seven studies [24, 25, 26, 32, 35, 36, 37], while dynamic balance was assessed in eleven [24, 25, 27, 28, 29, 30, 31, 32, 33, 34, 38], with three studies evaluating both concurrently [24, 25, 32].

3.8. Types of AFOs Used

The included studies investigated a wide range of AFO designs, differing in stiffness, mechanical articulation, and intended biomechanical function. Based on design characteristics, the AFOs were categorized into five main types: solid (SAFO), hinged/articulated (HAFO), ground‐reaction (GRAFO/FRAFO), dynamic or carbon composite AFOs, and customized or tuned AFO–footwear combinations. The visual representation of these orthosis types is provided in Figure 2.

Figure 2.

Classification and key characteristics of ankle–foot orthoses (AFOs) identified in the systematic review. All photographs were provided by the authors.

Solid AFOs (SAFOs) were the most commonly investigated orthosis type, reported in eight studies [24, 26, 27, 31, 32, 33, 37, 38]. Hinged or articulated AFOs (HAFOs) were evaluated in six studies [26, 28, 29, 31, 33, 38]. Ground‐Reaction AFOs (GRAFOs/FRAFOs) were examined in four studies [24, 32, 35, 37]. Dynamic or Carbon Composite AFOs were used in two studies [25, 34]. Customized and Tuned AFO–Footwear Combinations were analyzed in two studies [25, 38].

3.9. Tuning, Stiffness, and Alignment Parameters Reported

Specific tuning parameters were detailed in seven studies, categorized as follows:

Fixed Ankle Angle: Five studies set the AFO at a fixed neutral angle (90°, 0° dorsiflexion) [24, 26, 31, 32, 37]. One of these achieved this position via a 4‐week gradual adjustment [32].

Individualized Alignment and SVA: One study customized the ankle angle based on gastrocnemius length and stiffness, targeting an SVA > 10°, compared to a standard AFO set at 90° [38].

HAFOs with Motion Control: Two studies designed HAFOs with a mechanical plantarflexion stop (0°/neutral) while allowing free dorsiflexion [26, 29].

Variable Stiffness Settings: One trial used a full‐carbon ventral shell AFO, specifying quantitative stiffness values (Nm/deg) for three distinct configurations: flexible, stiff, and rigid [34].

Footwear‐Mediated Alignment (Wedges and Sole Modifications): Two studies incorporated footwear modifications as an integral tuning process [30, 38]. Methods included adding custom wedges (avg. 2.4 cm) to solid AFOs to improve knee extension [30] or using individualized sole modifications to restore rockers [38].

3.10. AFO Wearing Schedule and Duration

Only six of the 15 included studies reported specific AFO wearing schedules [24, 25, 27, 28, 33, 37], revealing highly variable protocols. Reported minimum daily usage ranged from 1 h (plus 30 min therapy) [33] to 4 [25], 6 [37], or 8 h per day [24]. Alternatively, one study categorized cohorts by habitual use (6–12 vs. 12–24 h/day) [27], while another reported only frequency ( ≥ 5 days/week) without specifying daily hours [28]. The remaining nine studies lacked data on prescribed wearing schedules.

3.11. Footwear Compatibility

Although the AFO‐Footwear Combination (AFO‐FC) is critical, only six studies reported specific details regarding footwear compatibility [24, 29, 31, 34, 35, 37]. The level of detail varied considerably; two studies treated footwear as a standardized or experimental variable: one used sneakers with flat, flexible soles [31], while another provided high‐top skate shoes with individualized internal and external modifications as part of a tuned AFO‐FC intervention [24]. Additionally, one study used footwear for alignment tuning, comparing standard athletic shoes against custom wedged shoes [37]. The remaining studies reported minimal detail, simply noting that AFOs were worn “inside children's own shoes” [34], as “braced footwear” [35], or with “shoes and AFO” [29].

3.12. Analytical Comparison of AFO Type on Balance

3.12.1. Static Balance

Synthesized evidence indicates that GRAFOs consistently outperform SAFOs in enhancing static stability. While both orthotic types improve stability indices relative to barefoot conditions, GRAFOs yield significantly greater gains (28.97% vs. 15.80%; p < 0.05) [26, 32]. In contrast, findings for floor‐reaction AFOs (FRAFOs) remain contradictory; certain reports document a degradation in mediolateral (M‐L) stability (p < 0.05) [35], whereas others observe no significant differences in static CoP parameters compared to hinged AFOs (HAFOs) (p > 0.05) [25]. For HAFOs, evidence suggests limited efficacy in reducing sway magnitude; however, they appear to modulate postural control mechanisms by significantly increasing the contribution of the ankle strategy to M‐L balance(p < 0.05) [36].

3.12.2. Dynamic Balance

Comparisons between HAFOs and SAFOs produce conflicting outcomes contingent upon the assessment modality. One study reported superior dynamic weight‐shifting (FRT) with HAFOs [28], whereas others found no significant differences across broader functional batteries (PBS, FTSST) (p > 0.05) [30]. Furthermore, the overall functional benefit of orthoses on dynamic balance remains unestablished; randomized trials comparing multiple AFO types against unbraced conditions reported no statistically significant differences in total PBS scores or obstacle‐course stability (p > 0.05) [38]. Biomechanical analyses of postural control quality also yield divergent findings: rigid designs (PLS/solid AFOs) have been linked to increased trunk frontal angular velocity [34], whereas accelerometry‐based studies suggest that various AFO configurations can “normalize” trunk acceleration complexity toward typically developing patterns [29].

3.13. Influence of Intervention Parameters (Tuning, Footwear, and Duration) on Balance Outcomes

Orthotic configuration and usage protocols emerge as critical moderators of balance outcomes, often outweighing the influence of AFO type alone.

Tuning and Stiffness: Individualized tuning protocols that optimize the shank‐to‐vertical angle (SVA > 10°) and customize ankle alignment based on physiological properties demonstrate significantly greater improvements in both static (p = 0.042) and dynamic (p = 0.009) balance subscales compared to standardized (90°) configurations [24]. Conversely, increasing ventral‐shell stiffness correlates negatively with M‐L dynamic stability and promotes compensatory trunk lateroflexion [31].

Footwear Integration: Footwear‐mediated alignment is a vital component of the AFO–footwear combination (AFO‐FC). Interventions incorporating custom shoe wedges or rocker modifications show context‐dependent efficacy; statistically significant improvements are consistently observed in children with higher baseline functional levels (p < 0.05) [37].

Wearing Schedule: Evidence regarding a dose–response relationship remains inconclusive; comparative analyses reveal no significant differences in dynamic balance scores between high‐duration and moderate‐duration cohorts (p = 0.196) [27]. Conversely, structured low‐duration protocols ( ~ 1 h/day) integrated with therapy demonstrate measurable gains [33]. Table 6 provides a comprehensive synthesis of the current evidence regarding AFO effects and key intervention parameters.

Table 6.

Summary of evidence on AFO effects and intervention parameters.

Comparison/parameter	Population (CP/GMFCS)	Assessment method	Key finding (direction) a	Certainty (GRADE)	References
Static balance
GRAFO vs. SAFO	Diplegia/I–III	Biodex Stability Index	↑↑ Greater improvement with GRAFO (A‐P & M‐L)	⊕◯◯◯ Very Low	[26, 32]
FRAFO vs. Barefoot	Diplegia/I–II	Force‐plate CoP	↔/↓ No improvement; ↑ ML sway velocity	⊕◯◯◯ Very Low	[35]
HAFO vs. Barefoot	Diplegia/I–III	CoP + Strategy analysis	↔ No sway reduction; ↑ Ankle strategy contribution	⊕◯◯◯ Very Low	[36]
Dynamic balance
HAFO vs. SAFO	Diplegia & Unilateral/II–III	FRT vs. PBS/NeuroCom	↑ in FRT; ↔ in multidimensional scales	⊕◯◯◯ Very Low	[28, 30]
Increased stiffness	Mixed/I–III	Margin of stability (3D)	↓ Progressive decrease in ML stability	⊕◯◯◯ Very Low	[31]
Intervention parameters
Individualized tuning (SVA > 10°)	Bilateral/II–III	PBS total & subscales	↑↑ Significant gain vs. non‐individualized (90°)	⊕◯◯◯ Very Low.	[24]
Footwear wedging	Mixed/II–IV	Static standing duration	↑ Benefit only in higher‐functioning subgroup	⊕◯◯◯ Very Low	[37]

Abbreviations: AFO, ankle–foot orthoses; CP, cerebral palsy; FRT, Functional Reach Test; GMFCS, Gross Motor Function Classification System; GRAFO, ground‐reaction AFO; HAFO, hinged AFO; ML, mediolateral; PBS, Pediatric Balance Scale; SAFO, solid AFO; SVA, shank‐to‐vertical angle.

^a↑ indicates improvement; ↓ indicates deterioration; ↔ indicates no significant change. Double arrows (↑↑) indicate greater magnitude of improvement compared to the reference condition.

4. Discussion

This systematic review, addressing the equivocal evidence on the effects of AFOs and their design characteristics on balance and postural control in children with CP, synthesized data from 15 heterogeneous studies encompassing 390 predominantly ambulatory participants (GMFCS levels I–III, primarily spastic diplegia, aged 6–12 years). Key findings revealed design‐dependent benefits: GRAFOs consistently outperformed SAFOs in enhancing static stability (e.g., superior sway reduction), whereas HAFOs and FRAFOs yielded mixed or neutral effects; dynamic balance improvements were evident with specific configurations (e.g., HAFOs in functional reach tasks) but conflicted across assessments. Critically, individualized tuning (e.g., optimized SVA > 10° and footwear integration) emerged as a pivotal moderator, yielding superior static and dynamic gains compared to non‐tuned AFOs, while excessive stiffness degraded mediolateral stability and no dose‐response relationship was identified for wearing duration. These results underscore the paramount role of AFO customization over generic designs in optimizing postural outcomes.

4.1. Comparison With Previous Studies

The central finding of this review—that the overall certainty of evidence for the effect of AFOs on both static and dynamic balance is very low—aligns with the conclusions of previous systematic reviews. For instance, a meta‐analysis by Lintanf et al. also reported “conflicting and insufficient” evidence regarding balance outcomes, despite finding moderate evidence for improvements in gait and gross motor function [14]. However, the current review diverges from and expands upon previous work in its analytical approach. Whereas earlier reviews primarily concluded that the evidence was ambiguous, the explicit aim of this study was to systematically investigate why this ambiguity exists by analyzing AFO design characteristics as primary variables.

The primary strength of this review, therefore, is its novel synthesis of intervention parameters. We have demonstrated that the heterogeneity in outcomes is not random but is critically linked to the lack of reporting on essential parameters—a gap previously identified by Eddison et al. [12]. Our review empirically confirms this gap, finding that fewer than half of the included studies provided specific details on tuning, wearing schedules, or footwear. By synthesizing the limited data available on these parameters, this review is the first to systematically demonstrate that individualized tuning of the AFO‐Footwear Combination is a pivotal moderator that yields superior balance outcomes compared to non‐individualized designs.

4.2. Biomechanical Mechanisms of AFOs on Postural Control

The mechanisms by which different AFOs influence balance are directly related to their mechanical properties and how they constrain or facilitate ankle‐joint kinematics [13]. These designs intervene at the ankle but create cascading biomechanical effects up the kinetic chain, altering the postural control strategies employed by the central nervous system (CNS) [39].

SAFOs exert their effect primarily by providing a rigid external structure that enhances mediolateral (M‐L) stability and locks the ankle joint [11]. This sagittal plane rigidity functionally eliminates the “ankle strategy” for postural control [40]. Consequently, the CNS is forced to compensate by prioritizing a “hip strategy” to manage CoM displacements, a mechanism that can be effective for static stability but is less efficient for dynamic adjustments [36]. This aligns with our findings: while one RCT found SAFOs provided an immediate 15.80% improvement in the Overall Stability Index [32], they were found to be significantly inferior to GRAFOs for static balance [26, 32], and one study linked SAFOs to a negative dynamic effect on mediolateral trunk control during walking [34].

HAFOs utilize a mechanical joint, most commonly configured with a plantarflexion stop [30]. The primary mechanism of this design is to block pathological equinus (plantarflexion) while permitting controlled dorsiflexion [36]. By preserving sagittal plane motion in dorsiflexion, HAFOs allow for the continued use of the ankle strategy. This mechanism was uniquely identified in our review, where one cross‐sectional study found that HAFOs (with a plantarflexion stop) did not reduce the quantity of sway but did significantly alter the control mechanism by increasing the ankle strategy's contribution to M‐L balance [36]. This mechanism facilitates a more physiological response to anterior‐posterior perturbations and is critical for dynamic balance tasks such as sit‐to‐stand, which require forward tibial progression [25]. Our findings partially support this, showing HAFOs enabled a significantly greater reach distance on the FRT compared to SAFOs [25], though we found no difference between HAFOs and FRAFOs during a Sit‐to‐Stand task [28].

GRAFOs, or FRAFOs, operate by manipulating the ground reaction force (GRF) vector relative to the knee joint [39]. The rigid anterior shell applies a posteriorly directed force on the proximal tibia, which creates a substantial knee extension moment [35]. This mechanism is specifically intended to correct crouch gait by stabilizing the knee. By providing this distal stability at the knee, the GRAFO provides a stable base for the hip and trunk, thereby reducing the compensatory postural adjustments required to maintain an upright CoM [26]. This mechanism is strongly supported by our review's findings for static balance, where GRAFOs were found to be consistently superior to SAFOs in two separate RCTs [26, 32]. However, this effect is not universal for all anterior‐shell designs, as our review also found contradictory evidence for FRAFOs, which (when tuned to neutral) failed to improve static stability and significantly worsened M‐L sway velocity in one study [35].

Dynamic and Carbon Composite AFOs rely on material flexibility and energy return rather than rigid immobilization [31]. Their mechanism involves providing controlled resistance and assistance throughout the gait cycle, facilitating the natural ankle rockers [13]. This controlled, spring‐like motion is hypothesized to “normalize” postural control patterns, as evidenced by findings of increased complexity (Sample Entropy) in trunk acceleration signals during gait [29]. This finding of “normalized” postural control highlights the potential of flexible designs [29]. However, this mechanism appears highly sensitive to stiffness, as another study in our review demonstrated that increasing the stiffness of a ventral shell AFO (from flexible to rigid) led to a progressive decrease in mediolateral dynamic stability (Margin of Stability), which was correlated with compensatory trunk lateroflexion [31]. By modulating the ankle torque‐angle relationship, these AFOs allow for fine‐tuned control of tibial progression, which is essential for both static and dynamic stability [13].

4.3. Influence and Mechanistic Role of Tuning, Stiffness, Alignment, Wearing Schedule, and Footwear Integration

The biomechanical and clinical effectiveness of AFOs in improving postural control is strongly influenced by their configuration parameters—particularly tuning, stiffness, alignment, wearing schedule, and footwear integration. These parameters determine how the orthosis interacts with neuromuscular systems to stabilize the CoM, which is especially critical in children with CP, where spasticity and muscle weakness impair intrinsic balance mechanisms [12, 13].

4.3.1. Tuning, Stiffness, and Alignment

Findings from this review highlight that individualized tuning, especially when optimizing the SVA and ankle alignment relative to gastrocnemius properties, enhances both static and dynamic balance [24]. Proper tuning ensures efficient tibial progression and physiological GRF alignment, consistent with previous findings [41, 42].

In contrast, neutral or nonindividualized alignment (90° ankle angle) may constrain normal motion and increase mediolateral sway, as shown in untuned floor‐reaction AFOs [35]. Similarly, excessive stiffness in ventral shell AFOs has been linked to reduced mediolateral stability [13, 31], whereas moderate flexibility—as seen in carbon composite AFOs—supports controlled energy return and smoother CoM transitions [29]. Collectively, an optimal balance between stiffness and alignment is crucial to promote energy‐efficient, dynamically stable postural control and prevent compensatory trunk strategies [13].

4.3.2. Wearing Schedule and Duration

The reviewed studies revealed no clear dose‐response relationship between wearing duration and balance improvements. Longer daily use did not yield superior outcomes; in fact, prolonged wear (12–24 h/day) was associated with lower dynamic balance scores [43]. This may reflect reduced active engagement and sensory feedback during continuous passive support. Conversely, short, structured wearing protocols combined with functional activities—such as 1–1.5 h/day use during therapy—proved effective in enhancing balance [33]. These findings, consistent with previous studies, indicate that the context and purpose of use are more influential than total daily duration [12, 44].

4.3.3. Footwear Integration

Footwear plays a pivotal role in the AFO–Footwear Combination (AFO‐FC) by modulating alignment, heel height, and rocker function. Studies demonstrated that individualized shoe modifications such as wedging or rocker restoration optimize SVA and improve balance [24, 26]. In contrast, using non‐standardized or personal shoes introduces variability and may negate orthotic benefits [29, 34]. As highlighted by Owen and Bjornson, the shoe is not an accessory but an integral component influencing stiffness distribution, GRF trajectory, and postural stability [24, 41].

4.4. Clinical Implications

While the overall certainty of evidence for the effect of AFOs on balance was rated as very low, the findings of this systematic review nonetheless have significant implications for clinical practice, primarily by challenging the conventional, categorical approach to AFO prescription. Even with this low certainty, the evidence strongly suggests that clinical decision‐making should shift from a simple choice between AFO types (e.g., solid vs. hinged) to a more nuanced focus on individualized tuning of the AFO‐FC. The superiority of an individually tuned AFO‐FC (with optimized SVA) over a standard 90° orthosis in improving both static and dynamic balance provides some of the highest quality evidence found in this review for this paradigm shift [24]. Clinicians should therefore treat footwear not as an accessory but as an integral component of the intervention, essential for fine‐tuning alignment through features like wedges or sole modifications [24, 37]. Furthermore, the assumption that maximal stiffness equates to maximal stability is contradicted by our findings; evidence shows that excessive stiffness can degrade dynamic mediolateral stability and promote compensatory trunk movements [31].

This review also calls into question common protocols for AFO dosage and assessment. The absence of a clear dose‐response relationship for wearing time suggests that prescribing all‐day wear may not be the optimal strategy for improving balance [27]. The fact that a low‐duration protocol (e.g., 1–1.5 h) used during active therapy sessions yielded significant balance improvements indicates that the context and quality of AFO use (i.e., active functional training) may be more potent than passive duration [33]. Finally, based on these findings, clinicians must be aware of the methodological dichotomy in assessment. An improvement on a clinical scale (e.g., PBS) may not reflect an improvement in the underlying quality of postural control. For instance, HAFOs were found to alter the postural strategy (increasing ankle strategy use) without reducing the quantity of sway. This implies that a comprehensive assessment, beyond functional scales, is required to determine if an AFO is inducing a beneficial or a compensatory biomechanical change [36].

4.5. Limitations and Future Research Suggestions

The conclusions of this systematic review must be interpreted within the context of several key limitations. The primary limitation is the inability to perform a quantitative meta‐analysis due to significant methodological and clinical heterogeneity across the included studies. This heterogeneity was profound, manifesting in the wide variability of AFO designs (e.g., solid, hinged, GRAFO, dynamic), participant characteristics (e.g., GMFCS levels I‐IV, differing topographies), and the diverse array of outcome measures used, which spanned from clinical scales to laboratory‐based posturography. Consequently, the review's main finding is that the overall certainty of evidence for both static and dynamic balance is very low, constraining the strength of any definitive clinical recommendations. Furthermore, in accordance with the registered PROSPERO protocol and established methodological guidance for intervention‐focused reviews, case reports and case series were excluded; while this decision preserves the methodological rigor required for causal inference, it may limit insights into rare adverse events or exceptional clinical responses that such designs can provide.

This review's findings are further limited by the methodological weaknesses inherent in the primary studies themselves, as reflected in the GRADE assessment's “serious” ratings for risk of bias and imprecision. The risk of bias was substantial, particularly in the RCTs, which were universally hampered by a lack of blinding for participants and assessors—a challenge inherent to orthotic trials. Additionally, the prevalence of cross‐sectional designs (6 of the 15 studies) fundamentally limited the ability to establish causal relationships between AFO use and balance outcomes, allowing only for associations to be observed. Furthermore, many studies were characterized by small sample sizes, which were explicitly noted as limitations by the original authors and which directly contribute to the imprecision of the evidence. Furthermore, a notable limitation was the inconsistent reporting of exact p‐values, effect sizes, and confidence intervals in the primary literature, which hinders a robust interpretation of clinical significance. Similarly, the absence of long‐term follow‐up, noted in the GRADE assessment and explicitly mentioned as a limitation in several included studies, further contributes to imprecision and restricts conclusions to immediate or short‐term effects. The most significant limitation, however, was the critical gap in intervention reporting. Fewer than half of the included studies provided specific details on AFO tuning, footwear, or prescribed wearing schedules, and adherence was typically measured by subjective parental report rather than objective monitoring.

Based on these limitations, future research must pivot from asking if AFOs work to how they work. There is a critical need for high‐quality, longitudinal, adequately powered, multicenter RCTs that include long‐term follow‐up periods and adhere to standardized reporting guidelines for intervention parameters. Future studies must meticulously describe and, ideally, quantify AFO‐FC tuning (e.g., SVA, ankle angle), stiffness (e.g., in Nm/deg), and footwear characteristics. These trials should move beyond simple “AFO versus barefoot” comparisons and instead adopt designs that directly compare intervention parameters (e.g., individualized tuned AFO‐FC vs. nontuned AFO‐FC; high‐stiffness vs. flexible‐stiffness designs). To address the issue of adherence, the inclusion of objective wear‐time sensors is essential. Finally, outcome assessment should employ a comprehensive, combined approach—using both functional clinical scales (like the PBS) and laboratory‐based measures (like MoS or kinematics)—to capture changes in both functional capacity and the underlying quality of postural control, while also including patient‐reported outcomes and real‐world mobility metrics.

5. Conclusion

This systematic review demonstrates that while AFOs can influence balance in children with cerebral palsy, the overall certainty of evidence remains very low due to substantial heterogeneity, methodological limitations, and inconsistent reporting of critical intervention parameters. The findings suggest that orthotic effectiveness is not determined by AFO type alone, but rather by individualized configuration—including shank‐to‐vertical angle tuning, stiffness modulation, footwear integration, and context‐specific wearing protocols. Future high‐quality, parameter‐driven trials are essential to move beyond generic device comparisons and establish precision orthotic prescription strategies that optimize postural control and functional outcomes.

Author Contributions

Zeinab Gasavi Nezhad: conceptualization, investigation, writing – original draft, validation, methodology, visualization, software, formal analysis, supervision, data curation, resources, project administration. Amir Reza Vafayi: investigation, methodology, validation, visualization, software, formal analysis, data curation, writing – review and editing. Mokhtar Arazpour: writing – review and editing, conceptualization, investigation, methodology, data curation, resources, software, formal analysis.

Funding

The authors have nothing to report.

Copyright and Permissions

All figures and tables included in this manuscript are original creations by the authors, are adapted from publicly available sources under open licenses (e.g., Creative Commons), or have been properly cited. All direct quotations and paraphrased ideas from the scholarly works of others have been appropriately credited with in‐text citations and corresponding entries in the reference list, ensuring full academic integrity.

Ethics Statement

This study is a review of previously published literature and did not involve any direct interaction with human or animal participants.

Consent

Ethical approval and informed consent were not required for this research.

Conflicts of Interest

The authors declare no conflicts of interest.

Transparency Statement

Zeinab Gasavi Nezhad affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned and registered (PROSPERO) have been explained.

Supporting information

Supporting File

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.