Determinants and associated factors for hallux valgus: a systematic review and meta-analysis

Abstract

Background

Hallux valgus (HV) is a prevalent deformity with inconsistent evidence regarding its etiology. The objective of this review is to systematically synthesize the associated factors of HV.

Methods

Following PRISMA guidelines, five databases (PubMed, Web of Science, Cochrane, Embase, CNKI) were searched up to November 2025. We included observational studies (including cross-sectional, case-control, and prospective cohort designs) assessing factors associated with HV. Meta-analysis was performed and expressed as odds ratios (OR) if two or more articles assessed the same factor.

Results

Twenty-three studies were included. We identified eleven factors independently associated with increased odds of sustaining hallux valgus (HV) (in order of highest to lowest OR): (1) female sex (OR 2.57, 95% CI 1.83–3.52); (2) foot pain (OR 2.36, 95% CI 1.61–3.46); (3) narrow-toed shoes (OR 2.25, 95% CI 1.64–3.07); (4) knee pain (OR 2.04, 95% CI 1.55–2.68); (5) scoliosis (OR 1.84, 95% CI 1.14–2.98); (6) high-heeled shoes (OR 1.65, 95% CI 1.27–2.15); (7) osteoporosis (OR 1.50, 95% CI 1.13–1.99); (8) knee osteoarthritis (OR 1.47, 95% CI 1.05–2.07); (9) history of stroke (OR 1.26, 95% CI 1.04–1.53); (10) flatfoot (OR 1.19, 95% CI 1.07–1.33); and (11) age (OR 1.11, 95% CI 1.03–1.20). No significant associations were found for BMI or heart problems.

Conclusion

HV is a multifactorial condition primarily associated with female sex, aging, and mechanical constriction from narrow footwear, which poses a greater risk than high heels. Systemic conditions like scoliosis and osteoarthritis also contribute significantly. Interventions should be multifactorial, with a strong emphasis on footwear modification.

Trial registration

PROSPERO CRD420251218756.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12891-026-10054-1.

Background

The structural integrity of the human foot represents an evolutionary masterpiece of bipedal adaptation, yet it remains susceptible to progressive deformities that significantly impair mobility and quality of life. Among these, hallux valgus stands as a premier clinical challenge [1], characterized by a tri-planar subluxation of the first metatarsophalangeal joint [2]. Far from a purely cosmetic issue, it is a prevalent condition with significant clinical and public health implications. Epidemiological studies estimate that around 19% of adults have HV, with prevalence increasing markedly with age [3].

Women are disproportionately affected [4] – HV is roughly two to three times more common in females than in males – reflecting both intrinsic predispositions and footwear habits [5]. Cultural factors like footwear use also influence HV occurrence: the deformity is rare in populations that go barefoot yet much more frequent in shoe-wearing societies, underscoring the role of external pressures on the foot [5, 6]. As well as being a major contributor to the costs for forefoot surgery, HV has been linked to functional disability, including foot pain and disease [7, 8], impaired gait patterns [9], poor balance [10], and falls in older adults [11, 12].

While the development of HV is believed to be multifactorial, the exact etiology remains unclear. Previous studies have suggested that intrinsic biological predispositions (such as female sex [3, 13] and aging) and extrinsic mechanical forces (such as footwear habits [14–16] and foot structure [17–19]) are important factors. However, previous epidemiological studies have yielded conflicting results. For instance, the relationship between body mass index (BMI) and HV remains highly inconsistent across different populations and sexes [6]. Furthermore, the extent to which systemic conditions (e.g., osteoporosis) and structural misalignments (e.g., scoliosis) contribute to HV is not yet fully established [20 ,21]. These gaps highlight the critical need for a comprehensive synthesis of the current literature.

Therefore, the primary objective of this systematic review and meta-analysis is to synthesize the best available observational evidence to identify and quantify the determinants and associated factors of HV. By addressing the inconsistencies in previous research, we aim to provide a more definitive, evidence-based consensus to distinguish modifiable from non-modifiable determinants. Clarifying these associations—particularly regarding footwear—is crucial for guiding clinical screening, patient education, and targeted preventive strategies to mitigate the burden of HV.

Methods

This systematic review with meta-analysis was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [22, 23]. The protocol for this systematic review was registered in PROSPERO (CRD420251218756).

Literature search and study selection

PubMed, Web of Science, Cochrane Library, Embase, and CNKI databases were systematically searched for studies on determinants and associated factors for hallux valgus in adults. The search covered all records from the inception of each database to November 2025.

Search strategy

Using PubMed as an example, the search strategy combined terms such as: (“hallux valgus“[Title/Abstract] OR “hallux abductovalgus“[Title/Abstract] OR bunion[Title/Abstract] OR “Hallux Valgus“[Mesh] AND(“risk factor*“[Title/Abstract] OR etiolog*[Title/Abstract] OR pathogenes*[Title/Abstract] OR mechanis*[Title/Abstract] OR determinan*[Title/Abstract] OR associat*[Title/Abstract] OR correlat*[Title/Abstract] OR predispose*[Title/Abstract] OR biomechanic*[Title/Abstract] OR “associated with“[Title/Abstract] OR “Risk Factors“[Mesh])

This search syntax was appropriately adapted for the specific controlled vocabularies and search interfaces of the other databases.

Eligibility criteria

Criteria for studies to be included were: (1) observational studies, including cohort (prospective or retrospective, with any follow-up duration), case–control, and cross-sectional designs; (2) inclusion of adolescents and/or adults (no upper age limit) from the general population, clinical, or occupational settings; (3) assessment of at least one potential associated factor for hallux valgus (e.g., footwear type, age, sex, BMI, family history, physical activity, and foot morphology); and (4) reporting of hallux valgus presence, incidence, or severity assessed clinically and/or radiographically (e.g., hallux valgus angle, intermetatarsal angle).

Exclusion criteria were: (1) non-human studies (animal, cadaveric, or in vitro studies); (2) conference abstracts, reviews, editorials, case reports/series, or other non-original research; (3) studies without clear hallux valgus outcome definition or without extractable data on the exposure–outcome association; and (4) full text not available.

Data extraction

The screening process followed the PRISMA 2020 guidelines [24, 25]. Two reviewers (Yuan Ma and Guanyue Zhang) independently screened titles/abstracts and then full texts. Disagreements were resolved by consensus or, when necessary, consultation with a third reviewer (Shengnian Zhang).

Data were extracted independently using a standardized form, including: first author, publication year, study design, sample size, participant characteristics (specifically the number of male and female participants, age, BMI), definition and assessment of hallux valgus, associated factor(s) evaluated (e.g., demographic, biomechanical, functional, footwear-related), and the corresponding effect estimates (preferably adjusted ORs) with their 95% confidence intervals. When required information was missing or unclear, we contacted study authors to request additional data.

Risk of bias assessment

Risk of bias was assessed independently by two reviewers (Yuan Ma and Guanyue Zhang), with disagreements resolved by consensus or consultation with a third reviewer (Shengnian Zhang). Because this review focused on exposure/risk-factor evidence, the ROBINS-E tool was used to evaluate bias from confounding, participant selection, exposure classification, departures from intended exposures, missing data, outcome measurement, and selective reporting [26]. Each domain and overall risk of bias were rated as low, moderate, serious, or critical.

Data synthesis and statistical analysis

Meta-analysis was conducted when at least two independent studies assessed the same associated factor with sufficiently comparable definitions and provided, or allowed computation of, an odds ratio (OR). Statistical analyses were performed using R software (version 4.5.2; R Foundation for Statistical Computing, Vienna, Austria). Meta-analyses were primarily conducted using the ‘meta’ and ‘metafor’ packages. For dichotomous outcomes, associations were synthesized using odds ratios (ORs) with 95% confidence intervals (CIs).

If an OR was not directly reported, ORs were derived from available regression coefficients (when applicable) or converted using established methods [27]. All ORs were transformed to the natural logarithmic scale (ln OR), and corresponding standard errors were calculated prior to pooling. For each associated factor with ≥ 2 studies, pooled estimates were computed using inverse-variance weighting. Statistical heterogeneity across studies was evaluated using the I² statistic and Cochran’s Q-test.

Potential publication bias was evaluated through visual inspection of funnel plot symmetry and quantified using Egger’s linear regression test [28]. Following the recommendations of the Cochrane Handbook [29], assessment of publication bias was only performed for analyses or subgroups that included at least 10 studies to ensure adequate statistical power [30]. If significant bias was detected (P < 0.05), the trim-and-fill method was planned to assess the impact of potential missing studies. Additionally, a leave-one-out sensitivity analysis was conducted for all outcomes to evaluate the robustness of the pooled results [31].

Certainty of evidence (GRADE Assessment)

The overall certainty of evidence for each associated factor was evaluated using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework. Given that all included studies were observational, the initial certainty of evidence started at “Low.” The evidence was downgraded by one level to “Very Low” if there were serious concerns regarding inconsistency (e.g., significant heterogeneity, I² > 50%) or severe risk of bias. Conversely, the evidence could be upgraded by one level if there was a large magnitude of effect (defined as OR > 2.0 or < 0.5) [32–35].

Result

A total of 3,219 records were identified through database searching, including PubMed (n = 1,861), Web of Science (n = 262), EBSCOhost (n = 956), CNKI (n = 6), and the Cochrane Library (n = 134). After removing 996 duplicate records, 2,223 unique records remained for title and abstract screening. Of these, 389 reports were retrieved for full-text review. Following full-text screening, 286 records were excluded as off-topic, leaving 103 reports assessed for eligibility. Subsequently, 80 reports were excluded due to ineligible subjects (n = 10), ineligible outcome measures (n = 51), unobtainable or incomplete data (n = 18), or duplicate publications (n = 1). Ultimately, 23 studies were included in the systematic review and meta-analysis [5, 6, 36–56] (Fig. 1).

Fig. 1.

PRISMA flow diagram of study selection

Study characteristics

A total of 23 studies published between 2008 and 2025 were included in this meta-analysis. The included studies represented a wide geographic distribution, with the majority conducted in Japan (n = 6), the United States (n = 6), and China (n = 2) Other studies originated from the United Kingdom (n = 3), Australia (n = 3), and Israel (n = 1), with single studies from South Korea, and Indonesia.

In terms of study design, the analysis incorporated cross-sectional studies, case-control studies, and prospective cohort studies. Sample sizes varied significantly across the included literature, ranging from small clinical samples (e.g., n = 28) to large population-based cohorts (e.g., n = 3,868). The study populations were diverse, comprising community-dwelling older adults, university students, clinic patients, and specific occupational or athletic groups such as dancers and sales promotion women.

The diagnostic criteria for hallux valgus (HV) exhibited heterogeneity among the studies. Diagnosis was primarily established using radiographic measurements (with Hallux Valgus Angle thresholds typically set at ≥ 15°or ≥ 20°), the Manchester Scale (grades 3 or 4), validated line-drawing instruments, or clinical diagnosis/self-reporting (Table 1).

Table 1.

Study characteristics

Study(Author, Year)	Country	Study Design	Age (Mean ± SD/Median+Range)		Population		Sample Size (n)		HV Case (n)	Hallux Valgus Diagnostic Criteria
			Case	Control	Case	Control	Case	Control
Huang P, 2016 [36]	China	Case-control	57.50 ± 11.36	53.03 ± 6.26	Patients with HV	Healthy controls	3 M, 47 F	3 M, 27 F	50	Clinical diagnosis
Steinberg, N, 2015 [55]	Israel	Cross-sectional	12(8–16)		Non-professional dancers	Non-dancers	1336 F	226 F	709	Clinical diagnosis
Dufour, A. B, 2014 [40]	American	Prospective cohort	NR		Residents of Framingham		1352 M, 1725 F		1046	Clinical diagnosis
Nakao, H, 2023 [50]	Japan	Cross-sectional	52.0 ± 21.9		Participants of a foot health survey		353 M, 511 F		154	The summed score of the right and left Manchester scale was ≥ 4
Menz, H. B, 2016 [48]	Australia	Case-control	65.6 ± 9.9		Women aged 50–89 years		2627 F		1172	Validated line-drawing instrument; HV defined as ≥ grade 3 on either foot
Golightly, Y. M, 2015 [42]	American	Cross-sectional	68.4 ± 8.9		African American residents	Caucasian residents	125 M, 331 F	358 M, 688 F	938	Footprint HV angle ≥ 15°
Nguyen, U. S, 2010 [6]	American	Case-control	77.9 ± 5.6		Older persons living around Boston		214 M, 386 F		277	Visually HV angle ≥ 15°
Okuda, H, 2014 [52]	Japan	Cross-sectional	18.7 ± 0.6		University students		343 F		102	Footprint HV angle ≥ 15°
Yavuz, Metin, 2009 [56]	American	Case-control	55.2 ± 14.7	53.6 ± 18.7	Clinic patients		1 M, 13 F	5 M, 9 F	14	Self-reported
Soemarko, D. S, 2019 [54]	Indonesia	Cross-sectional	29 (18–49)	28 (19–50)	Women wearing high heels	Women wearing flat shoes	99 F	92 F	191	HV angle ≥ 15°
Galica, A. M, 2013 [41]	American	Case-control	64.5 ± 9.8		Community-dwelling older adults		2149 M, 2681 F		1123	HV angle ≥ 15°
Munteanu, S. E, 2017 [49]	Australia	Cross-sectional	57.9 ± 12.8		Female twins from Australian Twin Registry		260 F		70	Grades 2, 3, or 4 on the Manchester Scale
Matsumoto, T, 2024 [46]	Japan	Prospective cohort	73.9 ± 6.5	73.6 ± 6.1	Older adults with HV	Healthy controls	12 M, 89 F	8 M, F69	101	Radiography HV angle > 20°
Menz, H. B, 2023 [47]	United Kingdom	Prospective cohort	62.3 ± 8.0		Adults aged > 50 years		1243 M, 1027 F		349	Validated line-drawing instrument; HV defined as ≥ grade 3 on either foot
Lee, H. Y, 2022 [44]	American	Case-control	54.2 ± 14.0	43.0 ± 17.5	Patients with HV	Healthy controls	14 M, 16 F	6 M, 24 F	30	HVA > 15° and IMA > 9° via WBCT
Hurn, S. E, 2025 [43]	Australia	Case-control	53.7 ± 19.3	50.0 ± 20.1	Adults with HV	Healthy adults	7 M, 23 F	5 M, 15 F	30	The summed score of the right and left Manchester scale was ≥ 4
Roddy, E, 2008 [53]	United Kingdom	Cross-sectional	63.2 ± 13.4	54.5 ± 14.0	Adults with HV	Healthy adults	341 M, 853 F	1270 M, 1404 F	1194	Self-reported validated line-drawing scale
Cho, N. H, 2009 [39]	Korea	Prospective cohort	59.1 ± 8.5		Rural farming community resident		245 M, 318 F		364	Radiography HV angle ≥ 15°
Nishimura, A, 2014 [51]	Japan	Cross-sectional	76.1 ± 6.9	75.3 ± 6.1	Village resident with HV	Village resident without HV	21 M, 99 F	114 M, 169 F	120	Clinical diagnosis
Tao, C H, 2014 [37]	China	Case-control	37.10 ± 5.35	36.27 ± 5.64	Adults with HV	Healthy adults	103 F	103 F	103	Clinical diagnosis
Abhishek, A, 2010 [38]	United Kingdom	Cross-sectional	62.6 ± 13.7	54.6 ± 14.0	Adults with HV	Healthy adults	239 M, 477 F	857 M, 921 F	686	Validated line-drawing instrument; HV defined as ≥ grade 3 on either foot
Liu, Z, 2024 [45]	Japan	Cross-sectional	15.7 ± 1.5		Adolescent dancesport athletes		73 M, 202 F		151	HV angle ≥ 20°
Matsuoka, M, 2025 [5]	Japan	Prospective cohort	NR		Community-dwelling older adults		835 M, 664 F		352	HV angle ≥ 15°

Synthesis of results

Meta-analyses consisting of between 2 and 17 studies, were performed separately for each HV associated factor. The number of included studies for each meta-analysis is presented below.

Sex

Based on sixteen studies, females had increased odds of sustaining HV compared to males (OR 2.57, 95% CI 1.83–3.52 S, p < 0.001) (Fig. 2).

Fig. 2.

Sex differences in the odds of sustaining HV (female sex is to the right)

Age

Based on six studies, the odds for sustaining HV increased by 0.11 for each yearly increase in age (OR 1.11, 95% CI 1.03–1.20, p = 0.006) (Fig. 3).

Fig. 3.

age differences in the odds of sustaining HV

Body mass index

Based on eleven studies, no significant association was found between obesity (BMI > 30 kg/m²) and the risk of hallux valgus compared to those with a BMI < 25 kg/m² (OR 0.93, 95% CI 0.75–1.14, p = 0.47). Similarly, evidence from four studies indicated that being overweight (BMI 25–30 kg/m²) was not significantly associated with decreased odds of HV relative to the normal-weight group (OR 0.88, 95% CI 0.74–1.04, p = 0.08) (Fig. 4).

Fig. 4.

Differences in the odds of sustaining HV between those with a BMI > 25 compared to < 25

Non-musculoskeletal comorbidities

Meta-analysis of two studies revealed that osteoporosis significantly increased the odds of developing HV (OR 1.50, 95% CI 1.13–1.99, p = 0.005). Conversely, heart problems showed no significant association with HV risk (OR 0.97, 95% CI 0.71–1.32, p = 0.83). Notably, history of stroke was also associated with increased odds of sustaining HV (OR 1.26, 95% CI 1.04–1.53, p = 0.02), although the number of studies was limited (Fig. 5).

Fig. 5.

Differences in the odds of sustaining HV between those with and those without osteoporosis, heart problems and stroke

Limb pain

Based on seven studies, the presence of foot pain was associated with increased odds of hallux valgus (OR 2.36, 95% CI 1.61–3.46, p < 0.001). Furthermore, four studies indicated that knee pain also contributed to a higher likelihood of developing hallux valgus (OR 2.04, 95% CI 1.55–2.68, p = 0.001) (Fig. 6).

Fig. 6.

Differences in the odds of sustaining HV between those with and those without foot pain and knee pain

High heels

Based on ten studies, wearing high-heeled shoes was found to be more likely to cause hallux valgus compared to flat shoes(OR 1.65, 95% CI 1.27–2.15, p < 0.001) (Fig. 7).

Fig. 7.

Differences in the odds of sustaining HV between individuals wearing high-heeled shoes and those wearing flat shoes

Narrow toe-box

Based on ten studies, wearing narrow-toed shoes was found to be associated with an increased risk of hallux valgus compared with wide-toed shoes (OR 2.25, 95% CI 1.64–3.07, p < 0.001) (Fig. 8).

Fig. 8.

Differences in the odds of sustaining HV between those wearing narrow-toed shoes and those wearing wide-toed shoes

Flatfoot

Evidence from four studies indicated that individuals with flat feet had significantly higher odds of developing hallux valgus compared to those with normal arches (OR 1.19, 95% CI 1.07–1.33, p = 0.001) (Fig. 9). No significant heterogeneity was found across the included studies (P_{heterogeneity} = 0.65, I² = 0%), suggesting a consistent association.

Fig. 9.

Odds of sustaining HV associated with flatfoot deformity compared to normal arch height

Scoliosis

Based on three studies, individuals with scoliosis were found to be significantly more likely to develop hallux valgus compared to those without spinal deformities (OR 1.84, 95% CI 1.14–2.98, p = 0.01) (Fig. 10). No significant heterogeneity was observed between these studies (P_{heterogeneity} = 0.17, I² = 43.6%).

Fig. 10.

“Differences in the odds of sustaining HV between individuals with scoliosis and healthy controls

Knee Osteoarthritis

Based on three studies, individuals with knee osteoarthritis were found to be significantly more likely to develop hallux valgus compared to those without knee osteoarthritis (OR 1.47, 95% CI 1.05–2.07, p = 0.026) (Fig. 11).

Fig. 11.

Differences in the odds of sustaining HV between individuals with knee osteoarthritis and those without knee osteoarthritis

Sensitivity analysis and publication bias

To evaluate the stability and robustness of the pooled results, a leave-one-out sensitivity analysis was performed for all primary outcomes (except for those with less than 2 included studies). For the association between each factor and HV, the systematic exclusion of any single study did not materially alter the pooled odds ratio or the direction of the effect size, with the statistical significance remaining consistently below 0.05(except scoliosis). The 95% CIs of the re-calculated ORs remained largely overlapping with the original estimates. These findings indicate that our meta-analysis results are robust and not driven by any single outlier study (Supplementary Fig. 1–11).

Statistical heterogeneity was evaluated using the I² statistic and Cochran’s Q-test. For the Scoliosis group (I² = 43.6%, P = 0.17, moderate heterogeneity was observed. No heterogeneity was detected in the flat feet (I² = 0%, P = 0.65), osteoporosis (I² = 0%, P = 0.72), heart problem (I² = 0%, P = 0.83), and stroke (I² = 0%, P = 0.81) groups. In contrast, Substantial heterogeneity was noted for sex (I² = 96.0%, P < 0.001), age (I² = 85.5%, P < 0.001), foot pain (I² = 90.0%, P < 0.001), knee pain (I² = 85.4%, P < 0.001), high heels (I² = 61.3%, P = 0.006), narrow toe-box (I² = 65.7%, P = 0.002) and osteoarthritis (I² = 78.7%, P = 0.009).

The methodological quality of the included studies was rigorously evaluated using the RoB-E tool. The assessment covered seven distinct domains, and the results are summarized in Supplementary Fig. 12 (Traffic Light Plot) and Supplementary Fig. 13 (Summary Plot).

Overall, 5 studies (21.7%) were judged to be at low risk of bias, indicating high methodological quality across all domains. 15 studies (65.2%) were classified as having a moderate risk of bias, primarily due to concerns in the domains of confounding and exposure classification. 3 studies (13.0%)^37,48,49—were rated as having a serious risk of bias. Detailed domain-level findings are presented in Supplementary Table S1.

Assessment of publication bias via funnel plots and Egger’s test was conducted only for the narrow-toe-box and high heels group (n = 10 studies). The funnel plot displayed a generally symmetrical distribution Supplementary Figs. 14 and 15, and Egger’s regression test yielded a P-value of 0.61and 0.14, suggesting no significant publication bias. For other associated factors with fewer than 10 studies, formal bias testing was omitted to avoid unreliable results due to low statistical power.

GRADE assessment of evidence

The GRADE evidence profile for all eleven factors associated with hallux valgus is summarized in (Table 2). Due to the observational nature of the included studies, the overall certainty of evidence ranged from “Very Low” to “Low”. Factors such as female sex, foot pain, narrow-toed shoes, and knee pain demonstrated a large magnitude of effect (OR > 2.0), which warranted an upgrade; however, concurrent high heterogeneity (I² > 50%) offset this, resulting in a final “Low” certainty rating. Factors with low heterogeneity (osteoporosis, stroke, flatfoot, scoliosis) maintained a “Low” certainty level, whereas factors with high heterogeneity and moderate effect sizes (high-heeled shoes, knee osteoarthritis, age) were downgraded to “Very Low” certainty.

Table 2.

GRADE summary of findings for factors associated with hallux valgus

Associated Factor	No. of Studies	Odds Ratio (95% CI)	Heterogeneity (I2)	Certainty of Evidence (GRADE)	Rationale for Rating (Up/Downgrades)
Female Sex	16	2.57 (1.83–3.52)	96.00%	Low	Started as Low. Downgraded (-1) for severe inconsistency. Upgraded (+ 1) for large effect (OR > 2).
Foot Pain	7	2.36 (1.61–3.46)	90.00%	Low	Started as Low. Downgraded (-1) for severe inconsistency. Upgraded (+ 1) for large effect (OR > 2).
Narrow-toed Shoes	10	2.25 (1.64–3.07)	65.70%	Low	Started as Low. Downgraded (-1) for moderate inconsistency. Upgraded (+ 1) for large effect (OR > 2).
Knee Pain	4	2.04 (1.55–2.68)	61.90%	Low	Started as Low. Downgraded (-1) for moderate inconsistency. Upgraded (+ 1) for large effect (OR > 2).
Scoliosis	3	1.84 (1.14–2.98)	43.60%	Low	Started as Low. No severe inconsistency or risk of bias to warrant downgrade.
High-heeled Shoes	10	1.65 (1.27–2.15)	61.30%	Very Low	Started as Low. Downgraded (-1) for moderate inconsistency (I2 > 50%).
Osteoporosis	2	1.50 (1.13–1.99)	0.00%	Low	Started as Low. No heterogeneity (I2= 0%).
Knee Osteoarthritis	3	1.47 (1.05–2.07)	78.70%	Very Low	Started as Low. Downgraded (-1) for severe inconsistency (I2 > 50%).
History of Stroke	2	1.26 (1.04–1.53)	0.00%	Low	Started as Low. No heterogeneity (I2= 0%).
Flatfoot	4	1.19 (1.07–1.33)	0.00%	Low	Started as Low. No heterogeneity (I2= 0%).
Age	6	1.11 (1.03–1.20)	85.50%	Very Low	Started as Low. Downgraded (-1) for severe inconsistency (I2 > 50%).

GRADE Grading of Recommendations Assessment, Development and Evaluation. All evidence profiles started at “Low” certainty due to the observational design of the included studies. Upgrades were applied for large effect sizes (OR > 2.0). Downgrades were applied for inconsistency/heterogeneity (I² > 50%)

Discussion

Synthesis and interpretation of associated factors

This systematic review and meta-analysis synthesize current evidence on the determinants and associated factors of hallux valgus (HV). Our findings indicate that HV is a multifactorial deformity associated with a combination of intrinsic biological factors—including female sex, advancing age, and systemic comorbidities—and extrinsic mechanical exposures such as footwear type and musculoskeletal comorbidities.

Our meta-analysis of sixteen studies confirms that females exhibit significantly higher odds of developing HV (OR = 2.57), consistent with global prevalence estimates [6, 57]. This marked sex disparity likely reflects a combination of anatomical, hormonal, and footwear-related factors [57, 58]. Similarly, the association between increasing age and HV is well established and supported by our pooled estimates. Progressive degeneration of periarticular soft tissue structures has been proposed as a contributing pathway [58, 59], though the directionality of this relationship in observational data warrants caution.

Our pooled analysis found no significant association between HV and either obesity or overweight status. This null finding is consistent with large cohort data which report an inverse association in women (demonstrating a 30–45% lower HV risk compared to normal-weight counterparts) [6], while the direction appears reversed in men. These divergent sex-specific patterns suggest that BMI’s relationship with HV may be substantially confounded by footwear behavior.

Our analysis identifies osteoporosis as a significant associated factor for HV (OR = 1.50; 95% CI: 1.13–1.99). This association is plausibly related to reduced bone quality at the first metatarsal [20]. A statistically significant association was also observed between prior stroke and HV (OR = 1.26; 95% CI: 1.04–1.53), potentially due to post-stroke neuromuscular impairment and altered gait loading [60, 61]. Conversely, no significant association was identified between heart disease and HV, indicating the absence of a meaningful direct effect [62].

Our meta-analysis confirms significant associations between HV and foot pain, knee pain, and knee osteoarthritis. These associations are consistent with prior literature linking first MTPJ dysfunction to altered load distribution and compensatory proximal joint loading [63–65]. However, given the predominantly cross-sectional nature of the included studies, reverse causation cannot be excluded. These painful symptoms and proximal joint degenerations are likely concurrent associated factors rather than primary causative agents.

Both high-heeled footwear (OR = 1.65) and narrow toe-box shoes (OR = 2.25) were identified as significant associated factors. The higher odds ratio associated with narrow toe-boxes suggests that transverse forefoot compression may be more strongly associated with HV development than heel elevation alone [58 ,66]. Furthermore, pooled evidence demonstrates significant associations between HV and structural misalignments, including pes planus (OR = 1.19) and scoliosis (OR = 1.84). The underlying relationship between spinal alignment and forefoot deformity remains poorly characterized but highlights the importance of whole-body kinetic chain assessment.

Limitations of included studies

Several limitations inherent to the included original studies must be acknowledged. First, there was considerable variability in the diagnostic criteria for hallux valgus. While some studies utilized strict radiographic measurements (e.g., a hallux valgus angle > 15°), others relied on clinical observation or self-reported questionnaires, which could introduce misclassification bias and contribute to the observed heterogeneity. Second, the vast majority of the included studies utilized cross-sectional or case-control designs. Consequently, while robust associations were identified, true causal relationships cannot be definitively established. Third, several studies failed to adequately adjust for critical unmeasured confounders, such as genetic predisposition, familial history, or detailed lifetime footwear habits, which might lead to the overestimation of certain effect sizes.

Limitations of the systematic review

There are also limitations specific to this systematic review and meta-analysis. First, although we conducted a comprehensive search across five major databases (including English and Chinese literature), potential language and publication biases cannot be entirely ruled out, even though Egger’s test did not detect significant small-study effects for most factors. Second, significant statistical heterogeneity was observed in the pooled analyses of certain associated factors (e.g., sex, age, and footwear). Finally, as demonstrated by our GRADE assessment, the overall certainty of the evidence was categorized as “Low” to “Very Low.” This is an expected limitation when synthesizing observational data, underscoring the necessity for readers to interpret these findings as strong associations rather than definitive causal pathways.

Clinical implications

These findings support a multifactorial model of HV. For high-risk groups—particularly older females and those with osteoporosis—early conservative intervention is recommended. Evidence supports the use of orthoses and toe separators for symptom management [67], alongside the avoidance of constrictive footwear [59]. In advanced cases, surgical correction remains indicated, though potential complications necessitate careful patient selection [68]. Ultimately, the increasing evidence for effective conservative management supports a broader shift toward non-operative, holistic approaches [67].

Future directions

Future research must prioritize large-scale, prospective longitudinal cohort studies to clarify whether conditions like flatfoot or scoliosis contribute independently to forefoot deformity, or whether they co-develop. Additionally, sex-stratified analyses are required in future research to untangle the confounding effects of footwear behavior on factors like BMI. Finally, standardizing HV diagnostic criteria across cohorts will be crucial for reducing heterogeneity in future meta-analyses.

Conclusion

Hallux valgus is a multifactorial condition primarily associated with female sex, aging, mechanical constriction from narrow footwear, and systemic structural conditions like scoliosis and flatfoot. While the observational nature of the current evidence limits causal inferences, identifying these determinants highlights the necessity for holistic, multifactorial interventions focusing on early screening and footwear modification. Future prospective longitudinal studies utilizing standardized diagnostic criteria are required to definitively establish causation and refine targeted prevention strategies.

Abbreviations

HV

Hallux valgus

Authors' contributions

G.Z. was responsible for the conception and design of the study and data collection; G.Z. and Y.M. were involved in the processing and statistical analysis of data; G.Z., Y.M., and S.Z. were involved in the drafting of the manuscript; and all authors contributed to the interpretation of the data for the work and revising it critically for important intellectual content. All the authors finally approved the manuscript. All authors have read and agreed to the published version of the manuscript.

Funding

The authors received no financial support for the research, authorship, and publication of this article.

Data availability

All data relevant to the study are included in the article or are available as supplementary files.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.