Exploration of the causal association between osteoarthritis and ankle-foot diseases: evidence from the Mendelian randomization and colocalization analysis

Abstract

Objective

To investigate the causal relationships between osteoarthritis (OA) and the development of specific foot and ankle diseases, aiming to clarify their mutual influences and underlying mechanisms, and to provide a more comprehensive understanding of these conditions.

Methods

To investigate these potential causal associations, we conducted a bidirectional Mendelian Randomization (MR) study and colocalization analysis. The study was based on genome-wide association studies (GWAS) summary data, including eight foot and ankle diseases and three types of OA. We applied the inverse variance weighted (IVW) method and IVW with multiplicative random effects to analyze the genetic data, aiming to determine whether there is a causal relationship between OA and foot and ankle diseases.

Results

The Mendelian Randomization analysis showed that acquired hallux valgus and hallux rigidus increase the risk of OA, particularly knee OA. Conversely, knee OA elevates the risk of acquired hallux valgus, hallux rigidus, and foot fractures. Hip OA is also linked to a higher risk of acquired flat foot. These findings suggest a bidirectional relationship between OA and foot and ankle diseases.

Conclusions

This study provides evidence of bidirectional causal relationships between OA and certain foot and ankle diseases. These findings enhance our understanding of the pathophysiological connections between these conditions and suggest potential areas for targeted interventions and preventive measures. Future research should focus on confirming these genetic associations and exploring their clinical implications in diverse patient populations.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40520-025-03219-w.

Introduction

The feet and ankles play vital roles in supporting the force exerted by the human body, which is almost four to six times the normal weight [1]. The complex anatomical structures of the foot and ankle can transfer the forces to the ground and remain stable in daily activities and movements [2–4]. However, damage to these structures could break the anatomical stability and cause foot and ankle injuries, such as injuries in traffic accidents, or chronic damage caused by complications like autoimmune diseases, congenital deformity, and diabetes. Foot and ankle injuries are commonly seen in our daily lives. Based on the national surveillance database of the United States, one research found an incidence of ankle fracture of 4.22 per 10,000 person-years [5]. Another reported that ankle injuries affected the major part of patients with foot and ankle injuries, and ankle injury had the largest population aged 51 to 60, because older age was more likely to have increased complications and fatality, which contributed to the ankle injury [6].

Osteoarthritis (OA) was one of the complications found to promote the occurrence of foot and ankle injuries. A meta-analysis revealed that patients with knee OA had significantly more pronated foot postures than those without [7]. Individuals with patellofemoral osteoarthritis were also reported to have less ankle dorsiflexion and greater foot mobility magnitude than those without patellofemoral osteoarthritis [8]. Besides, some knee OA patients were reported to have an enduring presence of foot pain even after improvement of knee pain and function by total knee arthroplasty (TKA). And some foot deformities and disorders, including flatfoot, were reported to be associated with knee OA [9–12]. On the other hand, foot and ankle symptoms were also reported to promote osteoarthritis development. One research reported that foot and ankle symptoms in one or both feet were associated with significantly increased odds of developing both knee osteoarthritis symptoms and symptomatic radiographic knee OA [13]. Another research found that individuals with ankle pain were more likely to have symptomatic radiographic knee OA, although foot pain didn’t show a significant influence on OA occurrence [14]. Although all these findings could be partially explained as that the changes in biomechanics of abnormal foot/ankle and knee could influence the function of each other, the present researches are still limited in providing direct evidence to prove the resulting relationships between osteoarthritis on foot/ankle injuries. Given that the global population is aging, the number of osteoarthritis patients is also climbing [15], which is increasing the need to figure out the relationship between osteoarthritis and foot/ankle diseases.

Mendelian randomization (MR) is a novel approach distinct from conventional clinical and epidemiological investigations. It applies measurable genetic variances to explore causal associations between exposures and outcomes, adhering to Mendel’s principles of genetic inheritance. Employing single-nucleotide polymorphisms (SNPs) as variables, MR differs from traditional observational studies, offering potential advantages in reducing the influence of reverse causation and confounding factors [16]. Previous research has applied MR analysis to discover the causal relationships, identify influencing factors, or find potential treatment targets, which benefits the management of various diseases, including osteoarthritis and foot/ankle diseases [17]. However, there is still a lack of research employing MR analysis to explore the relationships between foot/ankle diseases and osteoarthritis. In this research, we aimed to conduct a bidirectional MR analysis to investigate the causal relationships between osteoarthritis and foot/ankle diseases. Osteoarthritis, including all osteoarthritis, hip osteoarthritis, and knee osteoarthritis, and common foot/ankle diseases, such as foot fracture and flat feet, etc., were included to investigate the direct relationships from genetic aspects, which would contribute to the diagnosis and management of foot/ankle diseases and osteoarthritis.

Data and methods

Study design and data source

In this study, we used a bidirectional Mendelian randomization (MR) analysis to investigate the causal association between ankle or foot diseases and osteoarthritis risk. After that, the colocalization analysis was used to determine whether osteoarthritis and ankle or foot diseases shared similar genetic variation. The schematic diagram is shown in Fig. 1. For data selection, the GWAS summary datasets related to osteoarthritis and ankle or foot diseases were used. A total of 8 ankle or foot diseases, along with 3 osteoarthritis, were included. The ankle or foot diseases included ankle enthesopathies, foot fractures, acquired flat foot, charcot foot, acquired clubfoot, congenital deformities, acquired hallux valgus, and hallux rigidus. And the osteoarthritis included osteoarthritis conditions in any joints, the hip joint, and the knee joint. The specific information of these GWAS summary datasets was summarized in Table 1.

Fig. 1.

The schematic diagram of investigating the causal association between ankle or foot disease and osteoarthritis using the Mendelian randomization and colocalization analysis methods

Table 1.

The information of the GWAS summary data related to ankle, foot diseases and osteoarthritis

Diseases	Source	Case (n)	Control (n)	Population
Ankle or foot disease
Ankle enthesopathies	FinnGen database	7930	294,770	European
Foot fracture	FinnGen database	8397	383,369	European
Acquired flat foot	FinnGen database	2996	262,844	European
Charcot foot	FinnGen database	503	297,867	European
Acquired clubfoot	FinnGen database	146	262,844	European
Feet congenital deformities	FinnGen database	756	409,304	European
Acquired hallux valgus	FinnGen database	15,886	262,844	European
Hallux rigidus	FinnGen database	4852	262,844	European
Osteoarthritis
All OA	34,822,786	177,517	649,173	European
Hip OA	34,822,786	36,445	316,943	European
Knee OA	34,822,786	62,497	333,557	European

SNP selection

SNPs strongly associated with OA or different ankle-foot diseases were selected as the instrumental variants (IVs) to analyze the relationships between exposure and outcome variants. IVs were selected according to three criteria: Firstly, the instrumental variants (IV) used as gene proxy for the exposure of various health-related items should be strongly correlated with these exposure items (P1 < 5e-8). Therefore, the P value set for screening out proxied SNPs of all OA, hip OA, and knee OA was set as P1 < 5e-8. However, due to the lack of eligible SNPs, the P value set for screening out proxied SNPs of ankle enthesopathies, foot or ankle fracture, charcot foot, and acquired flat foot is lowered to P1 < 5e-7. And the P value set for screening out proxied SNPs of congenital deformities and acquired clubfoot was lowered to P1 < 5e-6. Previous research reported that using more IVs could cause a larger bias but smaller variance, while applying too few IVs could cause a smaller bias but larger variance. To cope with this, the F values of SNPs were calculated to serve as the selection criteria that IVs with an F value more than 10 were considered as strongly associated with exposure and used for further investigation [18, 19]. Secondly, the IV screened out to proxy exposure should not be significantly associated with outcomes and confounding factors (P2 < 5e-5). Thirdly, the eligible IVs cause an effect on the outcomes through the pathway affecting the exposure. The genetic instruments proxying the ankle or foot diseases were listed in supplementary Tables 1–8, and the genetic instruments proxying osteoarthritis were summarized in supplementary Tables 9–11.

Bi-directional MR analysis and colocalization analysis

The inverse variance weighted (IVW) method, IVW with multiplicative random effects, the Weighted median, and Weighted mode methods were applied to investigate whether there was a causal association between osteoarthritis and ankle or foot diseases. After eliminating unreasonable SNPs by the equivalent unbalance method and screening out strongly correlated SNPS, the SNPs are counted to select the appropriate MR Analysis method. The colocalization analysis in Mendelian randomization is a statistical method used to determine whether two or more phenotypes share a common genetic basis. By examining the genetic overlap, researchers can infer whether the same genetic variants influence both the exposure and the outcome, providing insights into potential causal relationships. The colocalization analysis typically involves assessing the posterior probability (PP) for hypothesis 4, which indicates that there is a shared causal variant influencing both the exposure and the outcome. In this study, a high PP value, set at greater than 80%, suggests strong evidence of colocalization, implying a common genetic basis for the two traits being studied.

Sensitivity analysis

After the MR analysis, the sensitivity analysis was further carried out to detect the heterogeneity during the MR analysis. The Egger-Intercept and its P values for horizontal pleiotropy were determined by the MR-Egger method. And the Cochran Q statistics and the relative P value were determined using the IVW method. In addition, the heterogeneity I² was also calculated using the formula: I2=(Q-df)/Q * 100%. If the calculated I² < 0, we set it to 0. All the analysis was performed using the R software with the relevant R package.

Ethics approval and consent to participate

The GWAS summary data utilized in this research were obtained from a public database, originally sourced from published studies. These studies had previously received ethical approval from their respective institutional ethics committees, including the completion of informed consent procedures. Consequently, our study did not require further ethical clearance.

Results

The MR analysis of ankle or foot disease on osteoarthritis risk

As shown in Fig. 2, the MR analysis using the IVW method indicated that acquired hallux valgus would significantly increase risk of osteoarthritis of all joints (OR = 1.061, 95%CI: 1.027–1.097), hip osteoarthritis (OR = 1.107, 95%CI: 1.044–1.174) and knee osteoarthritis (OR = 1.078, 95%CI: 1.014–1.145). Meanwhile, we found that hallux rigidus would significantly increase the risk of osteoarthritis of all joints (OR = 1.071, 95%CI: 1.021–1.122) and knee osteoarthritis (OR = 1.086, 95%CI: 1.009–1.169). In addition, we found that congenital deformities significantly increased the risk of knee osteoarthritis (OR = 1.078, 95%CI: 1.014–1.145). While we detected no significantly causal effect of the other ankle or foot diseases on osteoarthritis, including ankle enthesopathies, foot or ankle fracture, acquired flat foot, charcot foot, and acquired clubfoot.

Fig. 2.

The causal effect of ankle of foot diseases on the risk of osteoarthritis using the IVW Mendelian randomization

The MR analysis of osteoarthritis on the risk of ankle or foot diseases

As shown in Fig. 3, the IVW MR analysis indicated that osteoarthritis of all joints would significantly increase the risk of foot or ankle fracture (OR = 1.368, 95%CI: 1.104–1.695), acquired flat foot (OR = 2.227, 95%CI: 1.403–3.535), hallux valgus (OR = 1.803, 95%CI: 1.439–2.258) and hallux rigidus (OR = 2.61, 95%CI: 1.806–3.772). And we found that hip OA increased the risk of acquired flat foot (OR = 1.25, 95%CI: 1.062–1.472), hallux valgus (OR = 1.803, 95%CI: 1.439–2.258), and hallux rigidus (OR = 2.61, 95%CI: 1.806–3.772). In addition, we found that knee OA would significantly increase the risk of foot or ankle fracture (OR = 1.159, 95%CI: 1.011–1.328), acquired hallux valgus (OR = 1.334, 95%CI: 1.121–1.587), and hallux rigidus (OR = 1.864, 95%CI: 1.524–2.279). Osteoarthritis seemed to have no causal effect on the other ankle or foot diseases.

Fig. 3.

The causal effect of osteoarthritis on the risk of ankle or foot diseases using the IVW Mendelian randomization

Sensitivity and colocalization analysis

The sensitivity analysis results between MR analysis of ankle or foot diseases and osteoarthritis were summarized in the supplementary Table 12. The significant heterogeneity was detected during the MR analysis of ankle enthesopathies and OA of any joints, acquired flat foot and all OA, as well as hallux valgus and OA. Meanwhile, we found that heterogeneity existed in the MR analysis process about the causal effect of all OA on congenital deformities, acquired flat foot, hallux rigidus, and hallux valgus. And the significant heterogeneity was detected in the MR analysis process about the causal effect of hip OA on acquired flat foot, hallux rigidus, hallux valgus, and knee OA on ankle enthesopathies, acquired flat foot, and hallux valgus (supplementary Table 13). The colocalization analysis between osteoarthritis and ankle or foot diseases was summarized in Supplementary Table 14. The colocalization analysis indicated that no shared causal genetic variation was detected between osteoarthritis and ankle or foot diseases.

Discussion

Some studies found the relationship between foot and ankle diseases and osteoarthritis, which alerts surgeons to the possibility of concurrent conditions when diagnosing and treating one of them. However, it is still not clear whether causal associations between these two types of diseases exist. In this study, we adapted a bio-directional MR analysis to investigate the causal relationship between foot and ankle diseases and osteoarthritis. We found that Acquired Hallux valgus causally promoted the occurrence of all kinds of OA. Hallux rigidus causally promoted all OA and Knee OA, and congenital deformities had a significant causal effect on Knee OA. In addition, OA also had a positive causal effect on foot and ankle diseases, especially for Acquired Hallux rigidus and Hallux rigidus. These results proved that foot and ankle diseases and osteoarthritis could causally influence the occurrence of each other from a genetic view, indicating that doctors should pay more attention to potential cooccurrence, and the prevention and treatment of one would be beneficial for the other.

Some researchers have reported that foot deformities were associated with the development of osteoarthritis. For example, many patients undergoing TKA have both knee and ankle pathology, and the changes of mechanical axis angles of the knee are accompanied by increased shifting of the hindfoot into valgus, which means that fixed hindfoot valgus deformity should be examined in patients with knee OA [20]. Besides, ankle dorsiflexion and calcaneal angle were found to be significantly different between knee OA or hip OA patients and patients without OA, while patients showed different foot posture characteristics between knee OA and hip OA [21]. However, the causal associations between foot deformities and OA were still not discussed. We firstly found that congenital foot deformity could causally promote the occurrence of knee OA, with no causal effects on hip and all OA, which was consistent with the results of those clinical observational studies and further explained the causal relationship between foot deformities and knee OA from genetic view.

Hallux valgus is a progressive deformity that affects 12%-65% of individuals over the age of 65. Altering foot morphology and function leads to changes in gait dynamics and plantar pressure, particularly in the forefoot region, causing pain and decreased walking functionality in patients [22, 23]. Previous research has reported the associations between Hallux valgus and osteoarthritis. A community-based Cross-Sectional Study investigated the potential risk factors for Hallux valgus and found that participants with knee or hip OA had significantly higher odds of Hallux valgus. Older individuals also had higher odds of Hallux valgus, while aging was also important for OA development [24]. Bilateral pes planus and hallux valgus increased the pain of women with Knee OA, Kellgren-Lawrence grade 2 and 3, and also increased their disability levels [12]. In addition, researchers applied a foot-toe orthosis to treat Hallux Valgus and discovered a positive effect on the moment and joint motion of the knee during walking in knee OA patients, indicating that Hallux Valgus therapy could benefit knee OA treatment [25]. Due to the limitations of clinical observational studies, previous studies were not able to discuss the causal effects between Hallux Valgus and OA, and little research focused on different types of OA. With the help of MR analysis, we could now reveal this relationship. Our results proved the bi-directional causality between Acquired Hallux Valgus and all kinds of OA.

Hallux rigidus is a kind of degenerative arthritis that occurs in the first metatarsophalangeal joint, which can cause pain, stiffness, and joint enlargement [26]. Different studies have reported that knee OA was associated with the occurrence of Hallux rigidus. A Japanese cross-sectional study found that radiographic knee OA (grade 2 or higher) was significantly associated with Hallux rigidus, indicating knee OA as an independent risk factor for Hallux rigidus [27]. Another cross-sectional study also revealed that the severity of knee OA was significantly associated with the presence or absence of Hallux rigidus [28]. However, little was known about the relationship between Hip/All OA and Hallux rigidus. Our research found that there was a bidirectional causality between Hallux rigidus and All/Hip/Knee OA, except that Hallux rigidus didn’t significantly influence the occurrence of Hip OA, which further proved the effects of knee OA on Hallux rigidus and provided more new information.

Furthermore, we discovered that all OA causally promoted the occurrence of Acquired flat feet and Foot fracture, Hip OA causally promoted Acquired flat feet, and Knee OA causally promoted Foot fracture. Acquired flat feet were also defined as the progressive collapsing foot deformity (PCFD), which referred to a deformity showcasing a lack of propulsive gait and a partial or complete flattening of the medial arch of the foot during weight bearing [29–31]. Acquired flat feet mainly affected middle-aged or elderly women with high body mass index and caused foot pain, movement dysfunction, and impairment of life quality [32, 33]. Previous research has reported that flat feet were associated with the development of OA. Bilateral, not unilateral, flat feet were found to be associated with worse knee pain in Knee OA patients, adjusted for age, sex, body mass index, and tibiofemoral joint K/L grade [34]. Another study based on computer tomography revealed that more severe flat feet might occur when knee OA progresses the meanwhile, knee OA patients with severe flat feet had an increased knee valgus alignment, indicating that flat feet and knee OA could influence the progress of each other [35]. However, there were still no prospective studies to investigate the causal relationship between flat feet and OA. Our results firstly showed the bi-directional causal relationships from a genetic view, which could benefit diagnosing and treating patients with Acquired flat feet or OA when they had the other disease.

Only limited research has focused on the effect of OA on Foot fractures. One study investigated the predictors of ankle and foot fractures in women aged more than 65 years and found that self-reported OA was associated with ankle fracture [36]. Another research reported that knee OA was relevant to abnormalities in plantar pressure and center of pressure in patients’ feet, which was related to pain and functional disability. Foot fractures were not discussed [37]. However, numerous studies have proved that OA significantly increases the risk of bone fractures. OA patients with knee symptoms were more likely to experience one or more falls and fractures [38]. Our study firstly reported the significant causal effects of all OA and knee OA on Foot fractures. Both these OA diseases promoted the occurrence of Foot fractures, suggesting that clinicians should pay attention to whether foot fractures simultaneously happen when they receive OA patients with a history of trauma. Our research has several advantages. This is the first study to investigate the causal relationships between Foot and ankle diseases and OA, finding that Foot/ankle diseases and OA could causally influence each other. Furthermore, MR analysis reduced the cost of elucidating the relationship between diseases, thereby identifying potential topics for clinical investigation. For instance, based on our findings, it is suggested that researchers conduct prospective cohort studies to investigate the incidence of ankle-foot diseases in OA participants or explore the effects of foot diseases on the incidence, symptoms, and prognosis of OA. All these findings could help guide the clinical management of both conditions. In addition to knee exercise, OA patients should also focus on foot function exercise. Orthopedic surgeons should pay greater attention to the potential comorbidity of ankle or foot injuries in individuals with OA.

Limitations

Although we employed the MR-Egger method to detect the horizontal pleiotropy, the possibility of residual bias due to unknown or undetected horizontal pleiotropy could not be entirely ruled out. In addition, this research was conducted based on the European population. It should not be ignored that the effects of exposures on outcomes could change in different populations. Besides, although we have found out the bidirectional relationships between Foot/ankle diseases and OA, more convincing clinical investigations with larger populations are still needed to validate these genetic findings.

Conclusion

This study revealed the potential causal relationships between osteoarthritis (OA) and specific foot and ankle diseases, such as acquired hallux valgus, hallux rigidus, and foot fractures, indicating the pathophysiological connections between OA and foot diseases. Although further validation is required to confirm this bidirectional causality due to the observational nature of the data and the inherent limitations of colocalization methods, clinicians could consider the potential comorbidity of these conditions when developing treatment plans. These potential associations suggested extra attention to or preventive measures for patients with OA or foot diseases. Future researches are suggested to focus on validating these genetic associations through larger-scale clinical studies and in different patient populations, thereby contributing to the management and potentially improved outcomes for individuals affected by both OA and foot and ankle diseases.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

T Lei and JH Wang were responsible for Conceptualization. JH wang, H Liu and ZW Zhu were involved in Investigation. X Tong, JY Nan, and RZ Chen were responsible for Writing-original draft. T Lei and JH Wang were responsible for Writing-review & editing. Yuxin Chang contributed the most during the revision process. All authors have read and agreed to the published version of the manuscript. Thanks for the contribution of Long Hua in the research investigation.

Funding

Natural Science Foundation of China (Grant No. 82360437) supported this research. China Postdoctoral Science Foundation (Grant No. 2025M772075) supported this research.

Data availability

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

Declarations

Conflict of interest

The authors declare that no any commercial or financial relationships could be a potential conflict of interest.

Consent for publication

All authors have read and agreed to the published version of the manuscript.

Patient involvement statement

Not applicable.

Clinical trial number

not applicable.