Positional and dimensional relation of tendons around the first metatarsal bone with hallux valgus

It was aimed to reveal whether the positions and dimensions of the extrinsic and intrinsic muscle tendons related to the hallux around the first metatarsal bone are affected by the severity of hallux valgus (HV) and whether tendon positional changes and tendon sizes affect each other. In formalin-fixed 46 feet, three HV angle subgroups (normal, mild, and moderate/severe) were defined. Width, thickness, and cross-sectional area (CSA) of tendons of the extensor hallucis longus (EHL) and brevis (EHB), abductor hallucis (AH), and flexor hallucis longus (FHL) were measured. On the clock model created in coronal plane, positional variations of each tendon were determined. In the moderate/severe HV group, thickness and CSA of the EHB, width and CSA of the AH were smaller, compared to mild HV. Width and CSA of the FHL were smaller in moderate/severe HV than in the normal. Regardless of HV, the width and CSA of the FHL were greater in cases where the FHL was located more lateral, and the width of both FHL and AT were greater in cases where AH located was more plantar. The smaller tendon size of two intrinsic (one plantar and one dorsal) and one extrinsic muscle in the moderate/severe HV group indicates that changes in the tendons are evident in cases of high severity of HV but not in cases of mild HV. Accordingly, the changes do not appear to be due to a factor limited to only one aspect of the foot. It is recommended to consider the possible biomechanical effects of AH, FHL, and EHB tendon dimensional weakness in surgical planning in moderate/severe HV cases.


Introduction
Hallux valgus deformity (HV) is a common and progressive clinical problem characterized by lateral deviation of the hallux and medial deviation of the first metatarsal bone (MTB) [1,2]. In addition to other factors, some specific anatomical and biomechanical factors such as impaired balance between the muscles around the joint were also held responsible for the initiation and progression of this deformity [1][2][3][4].
The positions of the insertions of abductor hallucis (AH) and adductor hallucis tendons, which are located in the plantar-medial and plantar-lateral aspects of the metatarsophalangeal joint (MTPJ), respectively, change as the hallux valgus deformity increases [4]. When the phalanx deviates laterally and pronates, the adductor hallucis muscle that attaches to the plantar-lateral part of the base of the proximal phalanx becomes a deforming force for the joint. Because the AH insertion is plantarly displaced, the resistance function of this muscle against valgus becomes less effective [2,5]. The placement of the extensor hallucis longus (EHL), extensor hallucis brevis (EHB), and flexor hallucis longus (FHL) tendons relative to the axes of MTPJ shifts laterally as a result of the lateral deviation of the hallux. It has been suggested that this new position of the tendons may increase the severity of the deformity [3,5,6].

3
The literature on morphometric definitions and analysis of positional changes in tendons is quite limited.
In pathologies that require the reorganization of muscle balances in the foot, weakening of some muscles and hypertrophy of others are considered as examples of adaptation [7]. USG and electromyography studies in cases with HV showed that some intrinsic foot muscles differ from normal cases in terms of size (cross-sectional area (CSA), thickness, width), quality (fat infiltration, muscle fiber types, etc.), and function [6,[8][9][10][11][12][13]. It is not clear whether this deformity is occuring as a result of aging-related muscle weakness, or whether muscle weakness develops due to the deformity [13]. There are uncertainties about the extent to which changes in muscle belly size and other factors (age, disuse, etc.) are reflected in tendon morphometry [14][15][16][17].
Extrinsic muscles are considered as important determinants of foot function [18]. Isometric function of the FHL plays a role in the redistribution of force from the hind foot to the front, significantly affecting the ground contact strength of the hallux and the pressure under the forefoot [19,20]. The effective pressure to the ground by the toes is important for the management of ankle balance [21]. In all HV groups, lateral displacement of the EHL and EHB moment arms and plantar-lateral deflection of the FHL cause a 'bowstring' effect on the MTPJ, which becomes adductor force for the hallux [16, 17,22]. AH, the main force supporting the joint capsule medially, plays an important role in maintaining first MTPJ stability through isometric contraction [23][24][25]. The degree of displacement of the AH, which shifts toward the plantar and transforms into a flexor rather than an abductor, is associated with the hallux valgus angle (HVA) and the clinical severity of the deformity [3,5,6,22,26]. The direction of the cause-effect relationship between HV and structural and/ or functional weakness of the AH muscle is still unclear. One view is that after changes in the joint, the AH becomes dysfunctional due to its new position, and then the structure and function of the muscle become weaker. The other view is that AH is atrophied or weakened due to disuse for any reason, so it cannot adequately resist the force from the adductor muscles and an abduction-adduction imbalance occurs, as a result of which HV develops [6,9,13].
Plasticity in skeletal muscle in response to chronic loading or prolonged disuse may be reflected in its histological and mechanical properties [14][15][16]. It is controversial whether the electromyographic (EMG) and ultrastructural changes in intrinsic muscles in HV result from chronic ischemia due to increased pressure in the foot [10], or whether they develop secondary to HV deformity [2,24].
Open or percutaneous/minimally invasive surgical procedures can be preferred for the surgical treatment of HV [27][28][29]. New anatomical references may be needed for minimally invasive surgical procedures that can aid in the three-dimensional orientation and localization of structures that cannot be directly seen [29,30]. Additionally, partial or complete transfer of tendons around the hallux can be used in the repair of iatrogenic hallux varus deformity that may develop after corrective surgical treatment of HV [31][32][33][34]. The literature evaluating whether there is a structural change due to HV in the tendons that are planned to be transferred for this repair is quite limited [32].
In this study, the hypotheses that changes in various degrees of HV in the first ray bones of the foot were reflected in changes in the relative position and size of tendons around the MTB and that tendon positional variations were reflected in tendon sizes were evaluated. A modified form of the 'clock method' used by Dalmau-Pastor et al. (2020) to show the localization of neurovascular structures around the MTB [28] was used to standardize the definition of tendon positions.

Materials and method
The study was approved by the Board of Ethics of Mersin University (approval number: 2021-96), and supported by Mersin University Scientific Research Projects Unit (PN:2021-1-TP2-4313). 10% formalin-fixed 46 lower extremities of adult cadavers and amputed lower extremities (aged between 43-84, mean: 67.73 ± 11.54) (19 females and 27 males, and 20 left 26 right) from the inventory of Anatomy Department Laboratory of Mersin University were included. Feet with diffuse pathology around the big toe, except for HV, and feet that had previously undergone surgery were excluded. Digital caliper with 0.01 mm precision (MARCAL 16 ER, Mahr, Gottingen, Germany) and goniometer set (Lafayetta brand) were used for measurements. All measurements repeated by two different observers (FÇ and TK) were analyzed by the interobserver reliability test.
After removal of skin at the dorsal and plantar aspect of the foot, the tendons revealed. The description of the parameters used in the study is as follows:

Parameters about the size of tendons
EHL-W, T, CSA: width, thickness, and cross-sectional area of extensor hallucis longus tendon at CP level AT-W, T, CSA: width, thickness, and cross-sectional area of the accessory tendon of EHL at CP level EHB-W, T, CSA: width, thickness, and cross-sectional area of extensor hallucis brevis tendon at CP level AH-W, T, CSA: width, thickness, and cross-sectional area of abductor hallucis tendon at CP level FHL-W, T, CSA: width, thickness, and cross-sectional area of flexor hallucis longus tendon at CP level The width measurements of each tendon are shown in Fig. 2a-d. CSA was found with the ellipsoid area calculation formula using the width and thickness data of each tendon.

Parameters about the positions of the tendons at the CP level
EHL-EHB: closest distance between the tendons of extensor hallucis longus and extensor hallucis brevis EHL-AT: closest distance between the tendons of extensor hallucis longus and accessory tendon EHL-AH: closest distance between the tendons of extensor hallucis longus and abductor hallucis EHL-FHL: closest distance between the tendons of extensor hallucis longus and flexor hallucis longus EHL-DPAr: closest distance between the tendon of extensor hallucis longus and DPAr.

Hallux valgus angle measurement
The principle for the radiological HVA measurement [4] was adapted to cadavers; at first, the midpoints of the proximal and distal ends of both the first metatarsal bone and the proximal phalanx were marked with pins. Then two straight lines were created by passing a thread from the midpoint pins of each bone. The angle between the two stretched fibers was measured with a goniometer (Fig. 1a). In this study, we modified Dalmau-Pastor et al. (2018) clock model to describe the localization of tendons. To describe the location of the nerves, they cut the foot 1 cm proximal to the MTPJ in the coronal plane (approximately 1.5-2 cm ahead of that described in our study), defining a clock pattern in the circle around the MTB on the section. [27]. In this study, unlike them, the coronal plane was designed using a wire, without cutting the toe. The wire was inserted from dorsal to plantar at the level of DPAr, bent to wrap the toe medially, and its two ends were joined on the medial side to form a ring in the coronal plane. The vertical position of the part where the wire was advanced from the dorsal to the plantar at the DPAr point was maintained during the measurement phase by a second researcher, ensuring that the ring was perpendicular to the bone axis. Midpoint of EHL was marked as 12 on the clock model, and 0° on the 360° circle (Fig. 1a-c). The data in the ring of the right foot were transferred to the clock model as a "mirror image", thus left and right feet could be analyzed together. To adapt the measurements of the closest distance of each structure to the EHL to the clock and circle models, the following simple proportion formulas were used: Formula for clock model: hour distance to point 12 = Measurement to EHL (mm) × 12/Length of the wire (mm) Formula for 360° circle: degree to "0" point in 360° circle = Measurement to EHL(mm) × 360/Length of the wire (mm) The location of structures with respect to the mean values is indicated on the clock model and the 360° circle as shown in Fig. 1a, b, respectively.

Statistical analysis
Parameters with a significant difference between feet with and without HV were grouped. The data were found as normally distributed, and independent t test was used for comparison of two groups. ANOVA was used to evaluate whether there is a difference among the HVA subgroups (normal, mild HV and moderate/severe HV) for the parameters. The homogeneity of variances was demonstrated by Levene's test (p > 0.05). Tukey test was used for post hoc evaluation. Correlations between parameters were analyzed with the Pearson's correlation test. Statistically significance level was accepted as 0.05 for comparisons and 0.01 for correlations.
The agreement between the measurements of two different observers (FÇ and TK) was evaluated with intraclass correlation coefficient for interobserver using 95% estimated confidence intervals, and high agreement was found between the observers.

Findings about tendon sizes and their changes according to the HVA groups
EHL and FHL width were greater in males than in females (p = 0.017, p = 0.002). But there was no difference for the other tendons between the gender, and also between the sides (p > 0.05).
Both foot length and distance from medial malleol to CP have correlation with the CSA of AH (r = 0.682, p = 0.0001 and r = 0.410, p = 0.005) and FHL (r = 0.462, p = 0.001 and r = 0.417, p = 0.004) respectively. Length of the wire around the CP had a correlation with only the CSA of the FHL (r = 0.425, p = 0.003), but not others. There was no meaningful correlation for the CSA of EHL, EHB and AT with those measurements for the foot sizes (p > 0.01).
For EHB, AH, and FHL, some measures of tendon size were decreased in moderate/severe HV ( Table 2). The other tendon size parameters did not change according to the degree of HV (p > 0.05). Accessory tendon of EHL was found in 86.95% of feet. Presence or dimensions of the AT also did not differ according to degree of HVA (p > 0.05).

Findings regarding the position of tendons and whether they change according to the degree of HV
EHL-AH and EHL-FHL distances were greater in male than in female (p = 0.002, p < 0.001). There was no difference for the other tendons between the gender and sides (p > 0.05). There was no significant difference between the genders in terms of any parameter related to the positional placement of the tendons in the clock model and 360 circle model (p > 0.05).
In 41.9% (13 cases) of 31 cases with AH located close to the plantar in the clock model (between 8-9), FHL was close to the plantar-lateral; in 58.1% (18 cases), the FHL was close to the plantar-medial. In 92.9% (13 cases) of 14 cases with AH located close to the dorsal (between 9-10), FHL was close to the plantar-lateral; in only 1 (5.3%) case, FHL was close to the plantar-medial. According to the analysis by Pearson Chi-square test, the difference between the AH groups regarding FHL location was found statistically significant (p = 0.001).
No significant difference was found in terms of HVA degrees for the measured distances of the tendons to each other and their position on the clock model (p > 0.05). While both the direct measurement for the EHL-DPAr distance and the placement of it on the clock model were significantly smaller in moderate/severe HV than the mild HV (p = 0.015, p = 0.022, respectively).
There was no normal case in the group where FHL was located at the 5-6 o'clock position (more plantar-lateral). One of the four cases with FHL at this location had mild HV and three had moderate-to-severe HV. The only case in which AH was located between 7 and 8 was moderate HV.

3
The positions of the tendons on the clock model are categorized as in Table 3. The width and CSA of FHL were found to be greater in cases with more plantar/lateral location of FHL than in cases with medial location. The CSA, width, and thickness EHB were found to be larger in those with converging more distally with the EHL than those with fused more proximally. The widths of both FHL and AT were found to be greater in cases with AH located more plantarly than those with AH located more dorsally (Table 3). Any tendon size did not change significantly according to the positional variation of EHL and AT (p > 0.05).

Discussion
The clock model used to describe the position of the tendons eliminated the effect of the foot size factor, making it possible to categorize the variations in tendon positions and to evaluate these variations in terms of interaction with HVA, severity, and tendon sizes. Our findings regarding the smaller sizes of AH, FHL, and EHB tendons in the moderate/severe HV group were remarkable. The findings that the dimensions of some tendons vary according to the positional variations of the tendons were presented in terms of their contribution to the knowledge about the biomechanical properties of hallux.
While studies investigating the relation between HV development or progression and muscle weakness or atrophy mostly focus on the belly dimensions of intrinsic foot muscles [6,8,[11][12][13], studies on tendon sizes are limited. We interpreted our data on tendon sizes, taking into account the knowledge that muscle size is directly related to muscle strength and that tendon CSA correlates well with the physiological CSA of the associated muscles [35]. Stewart et al. (2013) found that the dorsoplantar thickness, mediolateral width, and CSA of the AH muscle mass decreased in the second-and third-grade HV groups compared to the normal group. As they did not find any significant difference between mild, moderate, and severe HV groups, they suggested that these changes in AH may be a result of muscle disuse due to HV and may occur early in the development of the deformity [24]. However, in our study, the tendon width and CSA of AH were found to be significantly smaller in the moderate/severe HV group than in the mild HV group. Our findings regarding EHB thickness and CSA support the idea that the HV-related adaptive response in the size of these tendons is markedly decreased in advanced HV severity, but not (maybe even increased) in low HV severity. Stewart et al. (2013) pointed out that the size characteristics of other muscles (i.e., flexors and extensors of the hallux) that have lost their normal anatomical relations with the first MTPJ, which may play a role in the development of HV, should also be examined [24]. The FHL tendon, one of the extrinsic muscles, showed a change similar to the intrinsic muscle in the severity of advanced HV, with one difference. The CSA and width of the FHL tendon were smaller in the moderate/severe HV group than in the normal group but were not different from these two groups in the mild HV group.
All these can be interpreted in two ways in terms of cause-effect relationship: changes in the dimensions of AH, EHB, and FHL are HV-related adaptive responses that are not noticeable in small HVAs but becomes evident in advanced HVAs, or all these tendons are already smaller for a reason independent of HV, which is a factor that facilitates the transition of HV from mild to advanced. For the latter suggestion, it may be necessary to find a common reason that can explain the decrease of the tendons of all two intrinsic muscles (one plantar and one dorsal) and one extrinsic muscle. Considering that there is no difference in various HV severities for EHL and AT, it becomes difficult to say that a general/common factor (genetic, metabolic, etc.) that can change the structure of all tendons may be responsible for advanced HV.
Muscle atrophy and decrease in muscle strength may be due to aging or long-term disuse. Advanced age is discussed in the literature as a factor that can both change the quality of AH and increase the risk of HV [8,17]. Hoffmeyer et al. (1988) explain the HV-induced alteration of the intrinsic muscles of the foot by chronic ischemia caused by increased pressure on the foot [10]. But it seems difficult to explain the changes we found in both the plantar and dorsal intrinsic and extrinsic muscle tendons of the foot with this suggestion. Based on all of these, the first suggestion that the changes in tendons develop secondary to the biomechanical effects of bone and joint deformation in advanced HV seems to be stronger.
The decrease in the stiffness of the intrinsic muscles of the foot is associated with a decrease in the resistance capacity against external loading and may result in HV [13]. In chronic overuse injuries, the internal structure of tendons that exceed at least two joints, cross the apex of the convex or concave skeletal curve, and have areas of insufficient vascular nutrition, are exposed to repetitive stress and multiple microtrauma, resulting in tendon degeneration [36]. AH, EHB, and FHL tendons, which are compatible with the anatomical part of this definition, may have suffered from a kind of chronic overuse injury and shrunk with a consequent degeneration as a result of the anatomical and biomechanical changes associated with HV. In the surgical repair of some hallux pathologies, the tendons in the region are used partially or completely [4,25,28,[30][31][32][33][34]. We recommend that the dimensional weakness of the AH, FHL, and EHB tendons be considered when planning surgical treatment or evaluating the efficacy of treatment in moderate-to-severe HV cases. Natsis et al. (2017) stated that the incidence of HV in feet with AT was higher than in feet without AT (almost 2/3); however, they did not find evidence of a relationship between the presence or morphological features of AT and the severity of HV [26]. Similarly, no evidence was found in our study showing the presence or dimensions of AT to be associated with HVA or severity.
Identifying positional variations of tendons around the first MTB may be important for its role in the clinical course of HV [29,37,38]. The tension of the EHL tendon, which should keep the hallux in extension, increases with lateral deviation in the HV. Thus, it is thought that the changing joint mechanics also reduces the active contraction capacity of the flexor muscles against the ground [6]. For the proximal part of the first MTB, we found no evidence that the position of the tendons relative to each other was altered by HVA or severity. It is known that the distal end of the first MTB is deviated medially and the intermetatarsal angle increases in HV [3,5]. The fact that EHL is closer to DPar in moderate/severe HV than in mild HV in our direct measurements and clock model supports that EHL deviates laterally at the proximal level of MTB.
We could not find any study with which we can compare our data on the reflections of tendon positional variations defined by the clock model on tendon dimensions. According to our results, a more plantarly/laterally located FHL around the proximal part of the MTB has greater FHL width and CSA than a more medially located FHL. Small tendon size was associated to the moderate/severe HV group. But we found no evidence for this level to suggest that the FHL shifted laterally in the HV or that positional variation is associated with tendon shrinkage. There was also no evidence that positional variations of FHL affect EHL dimensions. It was found that if AH was anatomically located close to the plantar, FHL and AT were larger, and FHL was more plantar-laterally located. But no evidence was found to directly correlate the positional variation of AH with HV.
If the level of EHB attachment to EHL was more distal, all dimensional parameters of EHB were greater than those more proximal, and these variations were not associated with foot length or HV. It can be assumed that all these determinations regarding the reflections of the positional variations of FHL, AH, EHB serve to maintain the delicate balances in the function of the arch of the foot and the hallux. Detailed morphometric studies including the distal end of the MTB and consisting of large series are needed to prove whether the deterioration in the positional variation and size relations of the tendons in the region lead to HV formation. Our findings on the positional variations of FHL and AH may provide new insights into the role of tendon anatomy in the pathogenesis of HV. Knowing the HV-related changes in the localization of FHL and EHL, which are at risk in some surgical approaches to HV [37][38][39], may contribute to the process.

The limitations of the study
As all groups were exposed to the same fixative solution, the possible effects of fixation on measurements were ignored in comparisons. It is known that the geometry of the forefoot changes while carrying weight [40]. In this study, however, we had to measure HVA in non-weight bearing foot conditions. It is unclear whether the morphometric adaptive response of the belly and tendon of the muscle to the biomechanical factors related to HV parallels at certain stages of the deformity or not and, we did not evaluate the dimension of muscle bellies according to HV groups. Lastly, relationship between the positional features and dimensions of the tendons was evaluated only in the plane passing through the DPAr point (as a reference of pulse area on the dorsum of the foot, and first MTB proximal part). We suggest that when this relationship is evaluated at the distal end of MTB, where the positional changes of tendons are more prominent in HV, new significant morphometric data for HV can be obtained.

Conclusion
The clock model used in the study reduced the effect of the foot size factor and facilitated the evaluation of variations in tendon size and position. One of the remarkable outcomes of the study is that although the tendon sizes of two intrinsic (AH, EHB) and one extrinsic (FHL) muscle are smaller in the moderate/severe HV group, there is no significant difference for EHL and AT. Another one is that the positional variations of FHL and AH are reflected in the dimensions of some tendons. The width and CSA of the more plantarly/laterally located FHL are greater than the more medially located FHL, and the FHL and AT are greater in cases where the AH is located closer to the plantar. Although FHL is more plantarly/laterally located in cases where AH is located closer to the plantar, there is no evidence to link these positional variations with HV for the proximal level of MTB in this study.