Epidemiology and clinical significance of os subtibiale in European and Asian populations: a meta‑analysis with treatment, diagnosis, and management considerations

Introduction

The presence of numerous accessory sesamoids and accessory ossicles in the foot and ankle region is well known to physicians and health care professionals involved in treating and managing lower extremity pain. Among these, os subtibiale (OS) is a rare anatomical variation located near the medial malleolus (Figure 1). Eugen Bircher¹ first described it in 1918, reporting a case of a 29-year-old patient with a left foot injury, in whom bilateral imaging showed triangular findings 2–3 mm below the medial malleolus. Based on the radiographs, Bircher categorically ruled out a traumatic origin. Another source cites an even earlier publication by Wilhelm Pfitzner from 1896 as the first recorded mention of OS.^2,3

In most cases, accessory ossicles are asymptomatic and are discovered incidentally on imaging.⁴ However, after trauma or injury, their radiographic appearance and associated symptoms may mimic other pathologies, leading to potential misinterpretation.^4,5 Although OS is typically asymptomatic, it can occasionally be associated with localized pain, which poses a diagnostic and therapeutic challenge.⁶ The differential diagnosis should include a medial malleolar fracture, post-traumatic ossification of the deltoid ligament complex, and an unfused medial malleolar ossification center.⁶ Accurate differentiation among these conditions is crucial to ensure appropriate management and optimal clinical outcomes.⁶

Moreover, since the initial description, conflicting OS prevalence rates have been frequently reported.^4,6,7 Current knowledge regarding the epidemiology of OS remains scarce and not well established, as the available data presented in multiple studies are heterogeneous. Additionally, small sample sizes and limited methodological rigor reduce the reliability of most estimates.

Aim

This meta-analysis aimed to clarify and systematically summarize all available data on OS prevalence, investigate its sex-related and geographic distribution patterns, and analyze its clinical significance. Our objective was to provide clinicians with a definitive reference for accurate diagnosis and appropriate management of OS by combining scarce and heterogeneous reports into a single report.

Methods

Study design and search strategy

The study protocol was pre-registered in the PROSPERO database (CRD420251164899). Its structure and data reporting strictly adhered to the PRISMA guidelines.⁸ An extensive search of major electronic databases (Pubmed, MEDLINE, Embase, and ScienceDirect) was performed through August 2025 for original articles reporting the prevalence of OS. The search terms used were: “os subtibiale” OR “accessory ossicle” OR “os subtibialis.” There were no restrictions regarding the language or publication date. A manual search of the articles’ reference lists was carried out once full-text articles were acquired to verify if any potenatially relevant publications were initially omitted. The papers written in languages other than English were translated by medical professionals fluent in those languages. A narrative review approach was used to describe and discuss the data regarding the clinical implications (treatment, diagnosis, and management considerations) related to OS.⁹

Study selection and data extraction

After evaluating each full-text publication, abstract, and title, 2 independent investigators (AO and MP) conducted data extraction and study selection procedures. In the case of any disagreements, a consensus was reach through comprehensive discussion, and a senior author (DT) made the final decision. Any articles that offered thorough information on the prevalence of OS were included in the study. Preset spreadsheets were used to collect the necessary data. The obtained data included the OS’s sex distribution, imaging modality (magnetic resonance imaging [MRI], computed tomography [CT], X-ray, or dissection), sample size, and geographic distribution. Studies with insufficient, partial, or difficult-to-extract data, letters to the editor, review articles, conference abstracts, case reports, case series, and studies reporting any findings other than OS (eg, fractures, calcifications, or nonunified apophyses) were excluded. The study selection process was conducted using Mendeley Reference Manager, version 2.93.0 (Elsevier, Amsterdam, the Netherlands).

Statistical analysis

For all statistical analysis and forest plot construction, MetaXL 5.3 by EpiGear International Pty Ltd (Queensland, Australia) was utilized. A random effects model was used to compute the pooled prevalence estimates (PPEs) of OS. Double arcsine transformation was used to stabilize the variances when pooling proportions, as some studies reported proportions close to 0. Each 95% CI was analyzed to identify significant differences within subgroups; a difference was deemed significant if the ranges did not overlap. The χ² test and the I² statistic were used to evaluate the heterogeneity among the included studies.¹⁰ The heterogeneity was interpreted as “may not be significant” at values between 0% and 40%, “may indicate moderate heterogeneity” at 30% to 60%, “may indicate substantial heterogeneity” at 50% to 90%, and “may represent considerable heterogeneity” at 75% to 100% based on the I² statistic.¹⁰ In the case of the χ² test, substantial heterogeneity was defined as a P value below 0.1. To further investigate potential sources of heterogeneity, a sensitivity analysis was carried out using the leave-one-out approach, and subgroup comparisons were made whenever feasible.

The prevalence estimates of OS were evaluated using a Doi plot with the Luis Furuya-Kanamori (LFK) index to identify publication bias, and the findings were interpreted as follows: absolute values between 0 and 1 indicated no asymmetry (no significant small study effect); values greater than 2 meant substantial asymmetry (strongly implying presence of small study effect); and values between 1 and 2 were understood as minor asymmetry (possibly suggesting a small study effect).¹¹

Quality assessment

The Anatomical Quality Assessment (AQUA) tool was used to evaluate the studies’ reliability and quality.¹² The analysis assessed each of the following 5 domains: 1) objective(s) and subject; 2) study design; 3) technique characterization; 4) descriptive anatomy; and 5) reporting of results. All “yes” answers suggested a low risk of bias, whereas a “no” response to any signaling question under any of the categories was deemed to indicate a high risk of bias. The “unclear” option was chosen when the results of the study were too inconsistent to be examined closely.¹² To maintain a high standard of evidence synthesis, the design and execution of this study were aligned with A Measurement Tool to Assess Systematic Reviews 2 criteria.¹³

Results

Initially, the database search yielded 2418 records. After removal of duplicates, 2245 articles were identified. The preliminary search of titles and abstracts resulted in the elimination of 2114 papers. The reference search identified a total of 46 studies which were subsequently evaluated in full-text alongside 131 articles from the database search. Finally, 14 studies (16 082 feet from Europe and Asia) met all eligibility criteria and were included into the quantitative analysis (Figure 2).^4,6,7,14-24 Table 1 summarizes the main characteristics of the 14 studies included in this meta-analysis.

Table 1. Characteristics of the included studies

Author, year	Country	Diagnostic modality	Number of feet examined	Number of feet with os subtibiale
Bizarro et al,13 1921	England	X-ray	100	0
Candan et al,4 2022	Turkey	X-ray	1651	4
Cepecchi et al,15 1964	Italy	X-ray	1361	8
Hachiya et al,16 1961	Japan	X-ray	680	1
Holle et al,17 1948	Germany	X-ray	1000	12
Kalbouneh et al,7 2021	Jordan	X-ray	1000	1
Leimbach et al,18 1937	Germany	X-ray	500	1
Matsui et al,19 1964	Japan	X-ray	213	2
Nikaido et al,20 1959	Japan	X-ray	1315	7
Ochs et al,6 2021	Germany	X-ray	1551	30
Shands et al,21 1931	United States	X-ray	1101	0
Suzuki et al,22 1957	Japan	X-ray	701	1
Tsuruta et al,23 1968	Japan	X-ray	1449	11
Tsuruta et al,24 1981	Japan	X-ray	3460	31

Quality assessment

According to the adopted quality evaluation criteria (AQUA tool), most of the articles had a low risk of bias in most of the categories. Some studies (specifically those published in the early 1900s) were classified as unclear in terms of risk of bias, as they used different patterns for data reporting methods. Only a small number of studies were deemed to carry a high risk of bias (Supplementary material, Table S1).

General prevalence

In total, the PPE of OS in the general population based on the 14 included studies was 0.51% (95% CI, 0.27–0.82; I² = 81.46; P <⁠0.001; Table 2; Figure 3). Out of the 16 082 analyzed feet, OS was found in 109 of the specimens. The LFK index equaled –0.98, which indicated no potential of publication bias (Supplementary material, Figure S1).

Table 2. Prevalence of os subtibiale in different populations

Population	Number of feet	Prevalence of os subtibiale (95% CI)	I2 value	P value
General	16 082	0.51 (0.27–0.82)	81.46	<⁠0.001
Women	3125	0.3 (0.08–1.05)	38.78	0.2
Men	2986	0.57 (0.18–1.82)	72.09	0.03
Europe	4512	0.97 (0.48–1.9)	71.27	0.008
Asia	10 469	0.49 (0.3–0.81)	56.05	0.03

Sex-specific prevalence

There were 3 studies including data on the sex, and they comprised 6111 feet (3125 female and 2986 male). The analysis of sex-related differences in OS prevalence showed that, although the occurrence of the ossicle was nearly twice as frequent in men than in women, based on the overlapping CIs, the difference did not reach significance. In women, the PPE was 0.3% (95% CI, 0.08–1.05; I² = 38.78; P = 0.2), whereas in men, it was –0.57% (95% CI, 0.18–1.82; I² = 72.09; P = 0.03; Table 2)

Geographical distribution

The geographical analysis was restricted, since the data originated only from 2 continents: Europe (5 studies; 4512 feet) and Asia (8 studies; 10 469 feet). The PPE of OS in Europe was 0.97% (95% CI, 0.48–1.9; I² = 71.27; P = 0.008) and was approximately 2-fold higher than in Asia, where it was 0.49% (95% CI, 0.3–0.81; I² = 56.05; P = 0.03; Table 2).

Discussion

Our study was conducted to synthesize the available data on the occurrence of OS, with particular focus on potential variability across demographic parameters. To date, no other systematic review or meta-analysis has provided a comprehensive overview of this anatomical variant.

In contrast, the literature contains detailed publications on different accessory ossicles, assessing their prevalence and clinical implications. Their occurrence varies significantly—the most frequent variants include the accessory navicular bone (12.6% of the cases),⁵ os trigonum (9%),²⁵ and os peroneum (6.6%).²⁶ Conversely, rare variants, such as os vesalianum pedis,²⁷ os supranaviculare,²⁸ os intermetatarseum,²⁹ and os calcaneus secundarius, are estimated to occur in approximately 1% of the population.

Epidemiology

The embryological origin of OS remains a subject of debate. While most accessory ossicles are believed to arise from secondary ossification centers, some authors suggest that OS may originate separately from the secondary ossification center of the medial malleolus, making its developmental pathway uncertain.³⁰

The ambiguity regarding its origin is reflected in significant differences in the prevalence rates reported across various populations. A radiographic study by Candan et al⁴ analyzed 1651 foot images from a Turkish population and reported OS prevalence of 0.24%. This Figure is approximately half the pooled rate estimated in our meta-analysis. However, since our study did not stratify data by ethnicity, this discrepancy may reflect genuine anthropometric differences between specific populations and the global average.

Prevalence appears markedly higher in certain athletic populations. Kinoshita et al³¹ compared professional soccer players with a control group and found that OS was present in 18.1% of the examined athletes. Even in their control group, the prevalence was 2.5%—a value 5-fold greater than the estimate in our study. This may suggest that mechanical stress or specific selection factors in athletic environments influence the observed frequency of this variant.

Clinical implications

Although OS is predominantly an incidental finding, it may sometimes be associated with painful syndromes, highlighting its significance in clinical practice.^4,6,7 Patients typically present with pain, tenderness, or swelling in the medial aspect of the ankle, often following trauma or prolonged overuse. While some cases involve bilateral OS identified after injury, reports documenting such bilateral symptomatic variants remain rare in the literature.^3,32

A primary clinical challenge lies in differentiating OS from acute medial malleolar fractures that can occur in both children and adults. For instance, a case series of 3 pediatric patients with ankle pain and swelling showed that the ossicle was easily misdiagnosed as a malleolar fracture; correct identification was achieved only after thorough investigation.³³ Similarly, Bandyopadhya³⁴ emphasized the importance of distinguishing an anteriorly located OS from a fracture. In that report, an 18-year-old patient remained symptomatic despite conservative management; the bone was properly identified and excised only after intraoperative assessment. Historical data support this recurring issue. Coral³⁵ reported a case in which a 38-year-old man presented with post-traumatic pain and swelling at the medial malleolus. Surgical exploration showed that the suspected fracture fragment observed on radiography was actually OS, thereby proving the surgical intervention to be unwarranted in this case.

Beyond osseous pathology, OS can confound the assessment of soft tissue and ligamentous injuries. Aydin et al³⁶ have highlighted that fact because the ossicle is connected to the deltoid ligament, and ankle trauma affecting the medial compartment may cause OS to be pulled away or detached from the medial malleolus. In these cases, MRI is essential for diagnosing deltoid injury and depicting the ossicle’s characteristics. However, MRI findings can be misleading. Turan et al³⁷ reported a case of a woman with posterior tibial tendon impingement and pain in the posterior aspect of the medial malleolus, which impaired her ability to walk or stand. Such cases may create diagnostic uncertainty, as MRI may suggest other conditions. This particular patient was initially thought to have tenosynovitis due to diffuse edema and fluid within the posterior tibial tendon.

In terms of management, conservative modalities of treatment may not always fully alleviate the symptoms, necessitating more invasive procedures. A 20-year-old semiprofessional soccer player described in a case report by Bellapianta et al³ was prescribed anti-inflammatory medication, rest, stretching, and icing, but showed no clinical improvement over a 6-month period. Eventually, surgical intervention (ankle arthroscopy with debridement of the hypertrophic synovium) was performed, enabling him to return to regular soccer training.

Ultimately, clinical experience plays a crucial role in avoiding these pitfalls. Kim et al³⁸ described 3 patients with chronic pain symptoms (lasting 1–2 y), initially misattributed to medial malleolar avulsion fractures. The first 2 cases were misdiagnosed despite radiographic evaluation, while the third one was promptly identified due to the authors’ prior experience. Radiological differentiation between OS and other structures can be aided by specific characteristics: the ossicle is located posteriorly to the medial malleolus and can be recognized by its significant size and distinct, rounded appearance, whereas most fractures have irregular margins.³⁹ Differentiation is further supported by an absence of a corresponding donor defect on the adjacent tibia.

Limitations and methodological considerations

Significant heterogeneity across the included studies restricts generalizability of our findings. The subgroup analysis demonstrated decreased variability in some strata, but residual disparities persisted, most likely due to certain unmeasured factors. As a matter of importance, significant heterogeneity should be expected in a meta-analysis based on anatomical research due to the inherent heterogeneity of anatomical investigations.⁴⁰ This could also be explained by a variety of factors, such as variations in patient preferences and the expertise of the investigators involved in the diagnosis and assessment of OS.

We performed sensitivity analyses that excluded older studies in order to evaluate the robustness of our findings. These analyses supported the stability of our results by not significantly changing the pooled values or the direction of the observed relationships. Global application of our findings is further limited by a lack of data from some continents. The available data were disproportionately distributed geographically, with the majority of studies coming from Asia and Europe and none from the African and American populations. Thus, our findings might not be globally applicable because anthropometric differences are known to exist among various ethnic groups.

The absence of information that would enable us to determine the pooled prevalence of symptomatic patients with OS was another constraint in our calculations. Without the possibility to determine the proportion of patients with OS who will experience problems associated with this anatomical variation at any point in their lives, we could not draw firm conclusions about the frequency of clinical manifestations and correlations of this ossicle. Therefore, it is necessary to explore this association in future research.

A strength of our meta-analysis lies in the methodological homogeneity of the included studies, as all utilized plain radiography as the primary diagnostic tool. This uniformity allows for a direct comparison of data across different cohorts. However, it must be acknowledged that plain radiography has lower sensitivity than advanced imaging modalities, such as CT or MRI. Consequently, small ossicles may be missed due to overlapping bone structures. Furthermore, the analysis showed that the small study effect and publication bias had comparatively little impact (LFK index, –0.98), which increased the validity of our findings. However, it is crucial to understand that both the sample size and the heterogeneity of the included studies have influenced the distribution of the LFK index in the absence of asymmetry. Therefore, this result should be interpreted with caution even though our values suggest minimal publication bias. To advance the field of evidence-based anatomy, future research must adopt widely recognized standardized checklists when describing anatomical variations.⁴¹ Given that many of our study’s findings rely on diverse anatomical and case reports, implementing such a standardized framework would considerably improve the consistency and quality of reporting in the field.

Finally, the accuracy of our meta-analysis depends on the diagnostic precision of the primary studies. The distinction between a true accessory ossicle and an avulsion fracture of the medial malleolus is often subtle, particularly in patients with a history of trauma. It is possible that some cases classified as OS in the primary literature were, in fact, post-traumatic fragments, which constitutes an inherent limitation of synthesizing retrospective radiographic data.

Conclusions

Awareness of OS as an anatomical variant is essential to avoid diagnostic errors. Correct diagnosis eliminates the need for surgical exploration, sparing the patient potential perioperative complications and functional consequences of prolonged inactivity.