- Email: [email protected]
Quantitative relationship of first metatarsophalangeal head morphology to hallux rigidus and hallux valgus
G Model FAS 1054 No. of Pages 5
Foot and Ankle Surgery xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Foot and Ankle Surgery journal homepage: www.elsevier.com/locate/fas
Quantitative relationship of first metatarsophalangeal head morphology to hallux rigidus and hallux valgus James D. Michelson, MD* , Jessie W. Janowski, MD, Mark D. Charlson, MD Department of Orthopaedics and Rehabilitation, University of Vermont Medical Center, Burlington, VT, USA
A R T I C L E I N F O
A B S T R A C T
Article history: Received 10 February 2017 Received in revised form 28 April 2017 Accepted 30 April 2017 Available online xxx
Background: Previous qualitative studies have linked first metatarsal head morphology with hallux valgus (HV) and hallux rigidus (HR). This study used a quantitative measurement of 1st MT radius of curvature to assess if HR MT heads were flatter than HV heads. Methods: Weight bearing foot films were used in HV, HR, and normal patients (no forefoot complaints) to measure the metatarsal head radius of curvature (normalized by dividing the radius of curvature by the first metatarsal length to adjust for magnification and foot size). Results: Radiographs from 299 feet were analyzed (105 normal, 57 HR, and 137 HV). The mean normalized radius of curvature was smaller in HV than HR, with normal feet in between (p < .05 for all comparisons). Metatarsal head curvature did not vary with age, weight, or BMI. Conclusion: These quantitative measurements are consistent with qualitative observations, validating the use of subjective metatarsal head morphology assessments. © 2017 European Foot and Ankle Society. Published by Elsevier Ltd. All rights reserved.
Keywords: Hallux valgus Hallux rigidus Metatarsal morphology
1. Introduction Many theories exist for the etiology of hallux valgus and hallux rigidus. Recent laboratory and clinical data suggests that hallux rigidus is caused by reduced excursion of the flexor hallucis longus due to entrapment at the fibro-osseous tunnel. Shoe wear, trauma (or micro-trauma), and underlying genetic predisposition have also been cited as etiologies for hallux valgus [1,2]. In general, patients develop either primary hallux rigidus or hallux valgus, but not both (although end stage hallux valgus can involve the development of arthritis). Coughlin et al characterized 137 feet with hallux rigidus [3]. They noted that 73% of their patients had a chevron or flat shaped metatarsal head. How the shape was determined was not characterized. Mancuso et al. [4] noted that hallux valgus patients had higher incidence of round head shape as compared to normal patients. They, too, did not quantify the shape of the metatarsal head. The current study was undertaken to determine if the aforementioned qualitative subjective observations regarding the shape of the 1st metatarsal in patients with hallux valgus and hallux rigidus could be verified using quantitative
* Corresponding author at: Stafford Hall, Room 426, 95 Carrigan Dr, Burlington, VT, 05401, USA. E-mail address: [email protected] (J.D. Michelson).
measurements based on the 2-dimensional morphology of the 1st metatarsal. The hypothesis was that hallux rigidus subjects would have 1st metatarsal heads with a higher radius of curvature (indicating a flat 1st MT head articular surface) and higher deviation from circularity (consistent with the subjective peaking described [3]), whereas hallux valgus patients will exhibit a smaller radius of curvature (a rounder 1st MT head) with less deviation from pure circularity. Furthermore, patients without any known hallux pathology will show 1st MT radius of curvatures intermediate between these two pathologic conditions. 2. Methods Weight-bearing foot films were measured for normal, hallux valgus and hallux rigidus patients. All patients seen in 2011 by the senior authors (JDM, MDC) with the relevant diagnoses for their clinic visit and accompanying radiographs were chosen. Since all patients were those of the 2 senior authors (JDM, MDC), both of whom are fellowship-trained F&A surgeons, there was consistency in diagnostic classification. All hallux rigidus patients had clinical and radiographic evidence of a dorsal 1st metatarsal head osteophyte with limitation of dorsiflexion at the 1st MTP joint. All hallux valgus patients had increased hallux valgus angulation (>20 ) and increased 1st-2nd inter-metatarsal angulation on both clinical and radiographic examination. The patients in the normal hallux group had presented with conditions that had no associated forefoot pathology. This resulted in 137 patients with hallux valgus,
http://dx.doi.org/10.1016/j.fas.2017.04.017 1268-7731/© 2017 European Foot and Ankle Society. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: J.D. Michelson, et al., Quantitative relationship of first metatarsophalangeal head morphology to hallux rigidus and hallux valgus, Foot Ankle Surg (2017), http://dx.doi.org/10.1016/j.fas.2017.04.017
G Model FAS 1054 No. of Pages 5
2
J.D. Michelson et al. / Foot and Ankle Surgery xxx (2017) xxx–xxx
57 with hallux rigidus. There was no attempt made to assess the severity of clinical symptoms, as it was not thought to affect the underlying 1st metatarsal head morphology. Normal controls were randomly selected from weight-bearing foot radiographs obtained during the same period of time in the Orthopaedic outpatient center with conditions that had no associated forefoot pathology. All radiographs were performed in a single radiography suite by the same team of radiography technicians using a consistent technique. The radiographs were standardized by normalizing the radius of curvature of each subject to the length of the first metatarsal. All radiographs are obtained from the institutional PACS system. Data regarding patient demographics (age, gender, height, weight), primary diagnosis (hallux valgus, hallux rigidus, normal) and co-morbidities were obtained from the electronic medical record system. Not all patients had BMI data available. Measurements of the radiographic images were performed on ImageJ software (v 1.48, available from NIH at http://imagej.nih. gov/ij/index.html). For each foot the first metatarsal head contour was measured with 7 points along the distal articular surface (1 far medial, 1 middle, 1 far lateral, with additional points in between the extremes and the middle). When there were osteophytes present, the far medial/lateral points were located at the end of the articular surface (Fig. 1). The medial and lateral extent of the joint line was determined by the points where there was a discontinuity in the local arc of the articular surface. This is seen in Fig. 1(A & B), where the outer most points on the articular surface (numbered 1 and 7) are located were there is an inflection in the curvature (most easily appreciated in Fig. 1B, where the surface is seen to abruptly deviate away from the inscribed radius of curvature arc at the medial and lateral aspects of the metatarsal head). The length
was measured from the midpoint of the distal and proximal articulations (using the dorsal lip for the proximal joint). All radius of curvature measurements were normalized by length of the 1st metatarsal to account for differences in foot sizes. There is a small variation in the measured metatarsal length on the AP view due to variation in the declination angle of the metatarsal. This amounts to a maximum variation of 3.4% in the measured metatarsal length, assuming an average declination angle of 15 , from the measured length on a lateral radiograph. However, the lateral radiograph will have a different magnification of the 1st metatarsal than the AP view, so using that length to normalize the AP metatarsal head radius of curvature would introduce an unknown amount of error into the calculations. As such, the best method to limit the experimental error was to maintain a consistent magnification between the radius of curvature and the normalizing length by using the AP radiographs for both measurements. Divergence from circularity for the 1st MT heads was determined by measuring the deviation of the 7 points on the MT head articular surface from the resulting arc that was calculated using ImageJ in the determination of the radius of curvature. This was calculated (Excel, Microsoft, Inc., Redmond, WA) as the average of the unbiased standard deviations of all the points from the radius of curvature arc defined by ImageJ. A higher average standard deviation corresponds to a larger divergence from the perfect circle calculated for the 1st MT head radius of curvature. Two independent investigators performed the measurements. There was no significant differences found between the raters (ANOVA), so the aggregate average measurements were used for all subsequent analyses.
Fig. 1. Using ImageJ software (see Section 2), 7 points were placed on the articular surface of the 1st metatarsal head (A). Care was taken to avoid areas of deformity due to exostosis formation on the medial and lateral periphery of the joint. The medial and lateral extent of the joint line was determined by the point where there was a discontinuity in the local arc of the articular surface. (see Section 2) The ImageJ software would then inscribe a best-fit circle that contained the articular surface points (B). This example is from a patient with the clinical diagnosis of hallux rigidus.
Please cite this article in press as: J.D. Michelson, et al., Quantitative relationship of first metatarsophalangeal head morphology to hallux rigidus and hallux valgus, Foot Ankle Surg (2017), http://dx.doi.org/10.1016/j.fas.2017.04.017
G Model FAS 1054 No. of Pages 5
J.D. Michelson et al. / Foot and Ankle Surgery xxx (2017) xxx–xxx
3
Table 1 Demographics of the study population.
Gender^
Female Male Total
Age (years)^^ BMI (kg/m*2)^^
Normal
Hallux rigidus
Hallux valgus
Total
p-value*
77 (33.8%) 28 (39.4%) 105 (35.1%)
31 (13.6%) 26 (36.6%) 57 (19.1%)
120 (52.6%) 17 (23.9%) 137 (45.8%)
228 (100.0%) 71 (100.0%) 299 (100.0%)
<.001
48.5 (45.7, 51.3) 31.7 (29.6, 33.8)
51.9 (48.6, 55.3) 26.4 (25.1, 27.8)
54.3 (51.5, 57.1) 27. (25.6, 28.4)
51.8 (50.1, 53.6) 28.6 (27.6, 29.7)
0.015 <.001
Abbreviation: BMI—body mass index. * Chi-square test used for categorical data, with ANOVA used for continuous data. ^ Values are number of patients (% of gender within foot type group). ^^ Values are mean (95% confidence interval).
The study population consisted of 299 patients (228 female, 71 male) with a mean age of 51.8 years (95% CI: 50.1–53.6 years). As seen in Table 1, women were significantly over-represented in the hallux valgus group compared to the hallux rigidus or normal groups (p < .001). The normal group population had a significantly higher BMI than the other groups (p = .001). The normal group was also significantly older than the hallux valgus group (p = .012) but not the hallux rigidus group. There were no significant differences between the hallux valgus and hallux rigidus groups with respect to age or BMI. There were no significant differences between the groups with respect to prevalence of pes planus, pes cavus,
inflammatory arthritis, tendon disorders (including posterior tibial tendon dysfunction), or non-inflammatory arthritis. More patients had diabetes in the normal group than in either other group. When the radius of curvature of the distal articular surface of the hallux metatarsal was normalized to the 1st metatarsal length (Fig. 2), the radius of curvature was significantly lower in those with hallux valgus compared to either normal radiographs or patients with hallux rigidus (p < .001 for both comparisons). The hallux rigidus patients had higher radius of curvature than normal patients (p < .007) and hallux valgus patients (p < .001). When controlling for the study group, there was no variation in the radius of curvature related to gender or BMI (p > .05 for both). Given the unequal gender distribution between hallux rigidus and hallux valgus, a 2-way ANOVA was performed to determine the relative contributions of foot type and gender to the metatarsal head radius of curvature. The graphic results are shown in Fig. 3, in which the radii of curvature are plotted separately for males and females. Both gender and foot type were significantly related to the radius of curvature (p < .001 for both), but there was no significant interaction between these two parameters (p = .433), indicating that their contributions to the radius of curvature were independent. While females had a lower radius of curvature than males, both overall and within each foot type (Fig. 3), the same pattern of curvature distribution between the foot types was seen in both genders. Hallux valgus feet had lower radius of curvature than either normal (p = .004) or hallux rigidus (p < .001) feet. However, when controlling for gender, there was no significant difference between normal and hallux rigidus radius of curvature (p = .462).
Fig. 2. The radius of curvature for the 1st metatarsal bone for the three foot types. The radius of curvature is normalized by the length of the metatarsal to compensate for radiographic magnification and foot size (see Section 2). The error bars represent 95% confidence intervals. All comparisons between the foot types are significant.
Fig. 3. The radius of curvature for the 1st metatarsal bone for the three foot types, plotted separately for females and males. The radius of curvature is normalized by the length of the metatarsal to compensate for radiographic magnification and foot size (see Section 2). The error bars represent 95% confidence intervals.
2.1. Statistical methods Categorical variable comparisons were done using Chi-square statistics, with demographic continuous variables analyzed using Student’s t-test. Comparisons of the measurements between metatarsal head types were performed using nonparametric Kruskal–Wallis ANOVA, since the parameters were not normally distributed. Pairwise comparisons were performed with the Mann–Whitney U-test. In all cases, p < .05 was considered statistically significant. 2.2. Ethics This study was reviewed and approved by the University of Vermont IRB. 3. Results
Please cite this article in press as: J.D. Michelson, et al., Quantitative relationship of first metatarsophalangeal head morphology to hallux rigidus and hallux valgus, Foot Ankle Surg (2017), http://dx.doi.org/10.1016/j.fas.2017.04.017
G Model FAS 1054 No. of Pages 5
4
J.D. Michelson et al. / Foot and Ankle Surgery xxx (2017) xxx–xxx
The deviation from circularity was assessed by calculating the unbiased standard deviation of the points on the articular surface from a perfect circle (Fig. 4). The lowest deviation from circularity was shown in hallux valgus feet, with normal feet showing an intermediate amount of deviation and hallux rigidus feet demonstrating the greatest amount of deviation from circularity. 4. Discussion The current study quantifies the morphological differences in the 1st metatarsal head in patients having the highly prevalent conditions of hallux valgus (HV) and hallux rigidus (HR). Compared to one another, feet with HR demonstrated metatarsal heads that had higher radius of curvature and flatter than HV feet. This is consistent with earlier studies that have characterized the hallux metatarsal shape with respect to these two conditions. In assessing the relationship of the 1st metatarsal head to HV, several investigators have classified the 1st metatarsal head morphology subjectively into 3 categories: oval (or round), flat (or square), and chevron shaped (or square with a ridge). There has not been any attempt to further define these categories quantitatively. Coughlin and Jones [1] reported on 122 operated on for HV, in which 71% were felt to have an oval shape to the 1st metatarsal head. Mancuso et al. [4] had similar findings when they compared weight-bearing radiographs for 110 HV patients to 102 normal feet. Round 1st metatarsal heads were seen in 91% of their HV patients and 20% of the normal feet. Using a Mose template to define the circularity of the 1st MT head, Okuda et al. [5] that in a population consisting of 40 HV surgical patients and 60 normal feet, round heads were seen in 78% of the HV patient and only 1.7% of the normal controls. In a prospective survey of people over 65 years old, D’Arcangelo et al. [6] found the prevalence of a round 1st metatarsal head increased from 29% in those without HV to 87% in people with severe HV. In contrast, the studies of 1st metatarsal head morphology in hallux rigidus report a high prevalence of flatter or chevron shaped heads. Coughlin and Shurnas reported that 73% of the 137 feet operated upon for hallux rigidus had a chevron or flat shaped metatarsal head [3]. Similarly, Beeson et al. [7], found that in 180 feet with HR, 73% were either flat or chevron shaped. In a
Fig. 4. The deviation of the articular surfaces from pure circularity is measured calculating the unbiased standard deviation of the points on the articular surface from the derived circle that defined the articular radius of curvature (see Section 2). The standard deviations are normalized by the 1st metatarsal length to compensate for radiographic magnification and foot size (see Section 2). The error bars represent 95% confidence intervals. Statistical significant differences: # — p < .02 between Normal and HV subjects, ## — p < .001 between HV and HR subjects.
cadaveric study involve 120 specimens, Stein et al. [8] found that those with evidence of hallux rigidus had a larger radius of curvature, which they quantified on plain radiographs. However, their measurements were performed on non-weight bearing first metatarsals and the measurements were not normalized to account for foot size. There is a single imaging study comparing the appearance of HV and HR feet which was performed in 24 patients, only 15 of which had pure evidence of either condition [9]. They reported three of 11 HV patients had a round head compared to two of four HR patients. The study of Yoshioka et al., hypothesized that as the joint naturally sits in 16–20 of extension and 15 of valgus it is subjected to subluxation forces in excess of one body weight during normal gait [10]. In the current study, hallux valgus patients were found to have significantly smaller normalized radii of curvature from that of hallux valgus patients. From a biomechanical basis this supports the hypothesis that the valgus stresses across the joint in a hallux valgus patient would lead to lateral subluxation of the joint (increased valgus) due to the smaller metatarsal head curvature, whereas in the broader hallux rigidus metatarsal head, this would not lead to subluxation since the articulation has more inherent mediallateral stability [7]. The current study is the first to quantify the 1st metatarsal head morphology as it relates to underlying hallux clinicopathology (hallux valgus and rigidus). Although this demonstrates a measurement tool that is unlikely to find use in the clinical arena, it does serve to prove that the qualitative assessments that have been made by various other investigators previously are anatomically accurate. Since the more subjective categorization has demonstrated good inter-observer reliability [6], the continued use of the oval/flat/chevron classification groups for reporting research findings is justified. A long-term prospective study following patients for several years with correlation of progression to their disease state to hallux metatarsal head curvature will be needed to more firmly establish the biomechanical-pathologic links that the current study implies. 4.1. Limitations of the study The weight bearing foot radiographs did not include a marker ball to control for magnification. This was dealt with by normalizing the radius of curvature to the length of the metatarsal. This serves to cancel any changes that are due to magnification, as both parameters have the same magnification. This is similar to the methodology used by other investigators [11,4]. Differences in hindfoot and forefoot structure could potentially cause transverse joint curvature to be measured at relatively different angles. However, the transverse profile of the hallux metatarsal head (as seen on the AP foot radiograph) does not appreciably change through the relatively small variation in 1st metatarsal inclination (5–20 from the horizontal plane). The low variability of the results also supports the notion that metatarsal inclination does not have a large effect on the measurements. The study is limited to 2dimensional images of the foot. In the absence of 3-dimensional images for all of the subjects, it is not possible to know if systematic differences in the sagittal plane or complex 3-dimensional morphology can also be related to the clinical conditions affecting the hallux. While this is an important basic biomechanical research question regarding the etiologies of hallux rigidus and hallux valgus, the purpose of the current study is restricted to validating the subjective assessment made clinically and reported by others [7,3,8,1,4,6,5]. Our radiographic analysis of the shape of the metatarsal head was performed on a representative group of northeastern
Please cite this article in press as: J.D. Michelson, et al., Quantitative relationship of first metatarsophalangeal head morphology to hallux rigidus and hallux valgus, Foot Ankle Surg (2017), http://dx.doi.org/10.1016/j.fas.2017.04.017
G Model FAS 1054 No. of Pages 5
J.D. Michelson et al. / Foot and Ankle Surgery xxx (2017) xxx–xxx
American patients. Consequently, our findings in this relatively homogenous population cannot necessarily be generalized to other populations. 5. Conclusion The 1st MT head is quantitatively flatter in patients with HR than those with HV, and shows more deviation from pure circularity (e.g. — a chevron morphology) in the HR patients. This is consistent with previous qualitative observations, and validates the use of these qualitative observations to categorize metatarsal head morphology. Conflicts of interest The authors declare that they have no conflict of interest. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
5
References [1] Coughlin MJ, Jones CP. Hallux valgus: demographics, etiology, and radiographic assessment. Foot Ankle Int 2007;28:759–77. [2] Lucas DE, Hunt KJ. Hallux rigidus: relevant anatomy and pathophysiology. Foot Ankle Clin 2015;20:381–9. [3] Coughlin MJ, Shurnas PS. Hallux rigidus: demographics, etiology, and radiographic assessment. Foot Ankle Int 2003;24:731–43. [4] Mancuso JE, Abramow SP, Landsman MJ, Waldman M, Carioscia M. The zeroplus first metatarsal and its relationship to bunion deformity. J Foot Ankle Surg 2003;42:319–26. [5] Okuda R, Kinoshita M, Yasuda T, Jotoku T, Kitano N, Shima H. The shape of the lateral edge of the first metatarsal head as a risk factor for recurrence of hallux valgus. J Bone Joint Surg Am 2007;89:2163–72. [6] D'Arcangelo PR, Landorf KB, Munteanu SE, Zammit GV, Menz HB. Radiographic correlates of hallux valgus severity in older people. J Foot Ankle Res 2010;3:20. [7] Beeson P, Phillips C, Corr S, Ribbans WJ. Cross-sectional study to evaluate radiological parameters in hallux rigidus. Foot (Edinb) 2009;19:7–21. [8] Stein G, Pawel A, Koebke J, Muller LP, Eysel P, Thelen U, et al. Morphology of the first metatarsal head and hallux rigidus: a cadaveric study. Surg Radiol Anat 2012;34:589–92. [9] Schweitzer ME, Maheshwari S, Shabshin N. Hallux valgus and hallux rigidus: MRI findings. Clin Imaging 1999;23:397–402. [10] Yoshioka Y, Siu DW, Cooke TD, Bryant JT, Wyss U. Geometry of the first metatarsophalangeal joint. J Orthop Res 1988;6:878–85. [11] Munuera PV, Dominguez G, Castillo JM. Radiographic study of the size of the first metatarso-digital segment in feet with incipient hallux limitus. J Am Podiatr Med Assoc 2007;97:460–8.
Please cite this article in press as: J.D. Michelson, et al., Quantitative relationship of first metatarsophalangeal head morphology to hallux rigidus and hallux valgus, Foot Ankle Surg (2017), http://dx.doi.org/10.1016/j.fas.2017.04.017
Foot and Ankle Surgery xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Foot and Ankle Surgery journal homepage: www.elsevier.com/locate/fas
Quantitative relationship of first metatarsophalangeal head morphology to hallux rigidus and hallux valgus James D. Michelson, MD* , Jessie W. Janowski, MD, Mark D. Charlson, MD Department of Orthopaedics and Rehabilitation, University of Vermont Medical Center, Burlington, VT, USA
A R T I C L E I N F O
A B S T R A C T
Article history: Received 10 February 2017 Received in revised form 28 April 2017 Accepted 30 April 2017 Available online xxx
Background: Previous qualitative studies have linked first metatarsal head morphology with hallux valgus (HV) and hallux rigidus (HR). This study used a quantitative measurement of 1st MT radius of curvature to assess if HR MT heads were flatter than HV heads. Methods: Weight bearing foot films were used in HV, HR, and normal patients (no forefoot complaints) to measure the metatarsal head radius of curvature (normalized by dividing the radius of curvature by the first metatarsal length to adjust for magnification and foot size). Results: Radiographs from 299 feet were analyzed (105 normal, 57 HR, and 137 HV). The mean normalized radius of curvature was smaller in HV than HR, with normal feet in between (p < .05 for all comparisons). Metatarsal head curvature did not vary with age, weight, or BMI. Conclusion: These quantitative measurements are consistent with qualitative observations, validating the use of subjective metatarsal head morphology assessments. © 2017 European Foot and Ankle Society. Published by Elsevier Ltd. All rights reserved.
Keywords: Hallux valgus Hallux rigidus Metatarsal morphology
1. Introduction Many theories exist for the etiology of hallux valgus and hallux rigidus. Recent laboratory and clinical data suggests that hallux rigidus is caused by reduced excursion of the flexor hallucis longus due to entrapment at the fibro-osseous tunnel. Shoe wear, trauma (or micro-trauma), and underlying genetic predisposition have also been cited as etiologies for hallux valgus [1,2]. In general, patients develop either primary hallux rigidus or hallux valgus, but not both (although end stage hallux valgus can involve the development of arthritis). Coughlin et al characterized 137 feet with hallux rigidus [3]. They noted that 73% of their patients had a chevron or flat shaped metatarsal head. How the shape was determined was not characterized. Mancuso et al. [4] noted that hallux valgus patients had higher incidence of round head shape as compared to normal patients. They, too, did not quantify the shape of the metatarsal head. The current study was undertaken to determine if the aforementioned qualitative subjective observations regarding the shape of the 1st metatarsal in patients with hallux valgus and hallux rigidus could be verified using quantitative
* Corresponding author at: Stafford Hall, Room 426, 95 Carrigan Dr, Burlington, VT, 05401, USA. E-mail address: [email protected] (J.D. Michelson).
measurements based on the 2-dimensional morphology of the 1st metatarsal. The hypothesis was that hallux rigidus subjects would have 1st metatarsal heads with a higher radius of curvature (indicating a flat 1st MT head articular surface) and higher deviation from circularity (consistent with the subjective peaking described [3]), whereas hallux valgus patients will exhibit a smaller radius of curvature (a rounder 1st MT head) with less deviation from pure circularity. Furthermore, patients without any known hallux pathology will show 1st MT radius of curvatures intermediate between these two pathologic conditions. 2. Methods Weight-bearing foot films were measured for normal, hallux valgus and hallux rigidus patients. All patients seen in 2011 by the senior authors (JDM, MDC) with the relevant diagnoses for their clinic visit and accompanying radiographs were chosen. Since all patients were those of the 2 senior authors (JDM, MDC), both of whom are fellowship-trained F&A surgeons, there was consistency in diagnostic classification. All hallux rigidus patients had clinical and radiographic evidence of a dorsal 1st metatarsal head osteophyte with limitation of dorsiflexion at the 1st MTP joint. All hallux valgus patients had increased hallux valgus angulation (>20 ) and increased 1st-2nd inter-metatarsal angulation on both clinical and radiographic examination. The patients in the normal hallux group had presented with conditions that had no associated forefoot pathology. This resulted in 137 patients with hallux valgus,
http://dx.doi.org/10.1016/j.fas.2017.04.017 1268-7731/© 2017 European Foot and Ankle Society. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: J.D. Michelson, et al., Quantitative relationship of first metatarsophalangeal head morphology to hallux rigidus and hallux valgus, Foot Ankle Surg (2017), http://dx.doi.org/10.1016/j.fas.2017.04.017
G Model FAS 1054 No. of Pages 5
2
J.D. Michelson et al. / Foot and Ankle Surgery xxx (2017) xxx–xxx
57 with hallux rigidus. There was no attempt made to assess the severity of clinical symptoms, as it was not thought to affect the underlying 1st metatarsal head morphology. Normal controls were randomly selected from weight-bearing foot radiographs obtained during the same period of time in the Orthopaedic outpatient center with conditions that had no associated forefoot pathology. All radiographs were performed in a single radiography suite by the same team of radiography technicians using a consistent technique. The radiographs were standardized by normalizing the radius of curvature of each subject to the length of the first metatarsal. All radiographs are obtained from the institutional PACS system. Data regarding patient demographics (age, gender, height, weight), primary diagnosis (hallux valgus, hallux rigidus, normal) and co-morbidities were obtained from the electronic medical record system. Not all patients had BMI data available. Measurements of the radiographic images were performed on ImageJ software (v 1.48, available from NIH at http://imagej.nih. gov/ij/index.html). For each foot the first metatarsal head contour was measured with 7 points along the distal articular surface (1 far medial, 1 middle, 1 far lateral, with additional points in between the extremes and the middle). When there were osteophytes present, the far medial/lateral points were located at the end of the articular surface (Fig. 1). The medial and lateral extent of the joint line was determined by the points where there was a discontinuity in the local arc of the articular surface. This is seen in Fig. 1(A & B), where the outer most points on the articular surface (numbered 1 and 7) are located were there is an inflection in the curvature (most easily appreciated in Fig. 1B, where the surface is seen to abruptly deviate away from the inscribed radius of curvature arc at the medial and lateral aspects of the metatarsal head). The length
was measured from the midpoint of the distal and proximal articulations (using the dorsal lip for the proximal joint). All radius of curvature measurements were normalized by length of the 1st metatarsal to account for differences in foot sizes. There is a small variation in the measured metatarsal length on the AP view due to variation in the declination angle of the metatarsal. This amounts to a maximum variation of 3.4% in the measured metatarsal length, assuming an average declination angle of 15 , from the measured length on a lateral radiograph. However, the lateral radiograph will have a different magnification of the 1st metatarsal than the AP view, so using that length to normalize the AP metatarsal head radius of curvature would introduce an unknown amount of error into the calculations. As such, the best method to limit the experimental error was to maintain a consistent magnification between the radius of curvature and the normalizing length by using the AP radiographs for both measurements. Divergence from circularity for the 1st MT heads was determined by measuring the deviation of the 7 points on the MT head articular surface from the resulting arc that was calculated using ImageJ in the determination of the radius of curvature. This was calculated (Excel, Microsoft, Inc., Redmond, WA) as the average of the unbiased standard deviations of all the points from the radius of curvature arc defined by ImageJ. A higher average standard deviation corresponds to a larger divergence from the perfect circle calculated for the 1st MT head radius of curvature. Two independent investigators performed the measurements. There was no significant differences found between the raters (ANOVA), so the aggregate average measurements were used for all subsequent analyses.
Fig. 1. Using ImageJ software (see Section 2), 7 points were placed on the articular surface of the 1st metatarsal head (A). Care was taken to avoid areas of deformity due to exostosis formation on the medial and lateral periphery of the joint. The medial and lateral extent of the joint line was determined by the point where there was a discontinuity in the local arc of the articular surface. (see Section 2) The ImageJ software would then inscribe a best-fit circle that contained the articular surface points (B). This example is from a patient with the clinical diagnosis of hallux rigidus.
Please cite this article in press as: J.D. Michelson, et al., Quantitative relationship of first metatarsophalangeal head morphology to hallux rigidus and hallux valgus, Foot Ankle Surg (2017), http://dx.doi.org/10.1016/j.fas.2017.04.017
G Model FAS 1054 No. of Pages 5
J.D. Michelson et al. / Foot and Ankle Surgery xxx (2017) xxx–xxx
3
Table 1 Demographics of the study population.
Gender^
Female Male Total
Age (years)^^ BMI (kg/m*2)^^
Normal
Hallux rigidus
Hallux valgus
Total
p-value*
77 (33.8%) 28 (39.4%) 105 (35.1%)
31 (13.6%) 26 (36.6%) 57 (19.1%)
120 (52.6%) 17 (23.9%) 137 (45.8%)
228 (100.0%) 71 (100.0%) 299 (100.0%)
<.001
48.5 (45.7, 51.3) 31.7 (29.6, 33.8)
51.9 (48.6, 55.3) 26.4 (25.1, 27.8)
54.3 (51.5, 57.1) 27. (25.6, 28.4)
51.8 (50.1, 53.6) 28.6 (27.6, 29.7)
0.015 <.001
Abbreviation: BMI—body mass index. * Chi-square test used for categorical data, with ANOVA used for continuous data. ^ Values are number of patients (% of gender within foot type group). ^^ Values are mean (95% confidence interval).
The study population consisted of 299 patients (228 female, 71 male) with a mean age of 51.8 years (95% CI: 50.1–53.6 years). As seen in Table 1, women were significantly over-represented in the hallux valgus group compared to the hallux rigidus or normal groups (p < .001). The normal group population had a significantly higher BMI than the other groups (p = .001). The normal group was also significantly older than the hallux valgus group (p = .012) but not the hallux rigidus group. There were no significant differences between the hallux valgus and hallux rigidus groups with respect to age or BMI. There were no significant differences between the groups with respect to prevalence of pes planus, pes cavus,
inflammatory arthritis, tendon disorders (including posterior tibial tendon dysfunction), or non-inflammatory arthritis. More patients had diabetes in the normal group than in either other group. When the radius of curvature of the distal articular surface of the hallux metatarsal was normalized to the 1st metatarsal length (Fig. 2), the radius of curvature was significantly lower in those with hallux valgus compared to either normal radiographs or patients with hallux rigidus (p < .001 for both comparisons). The hallux rigidus patients had higher radius of curvature than normal patients (p < .007) and hallux valgus patients (p < .001). When controlling for the study group, there was no variation in the radius of curvature related to gender or BMI (p > .05 for both). Given the unequal gender distribution between hallux rigidus and hallux valgus, a 2-way ANOVA was performed to determine the relative contributions of foot type and gender to the metatarsal head radius of curvature. The graphic results are shown in Fig. 3, in which the radii of curvature are plotted separately for males and females. Both gender and foot type were significantly related to the radius of curvature (p < .001 for both), but there was no significant interaction between these two parameters (p = .433), indicating that their contributions to the radius of curvature were independent. While females had a lower radius of curvature than males, both overall and within each foot type (Fig. 3), the same pattern of curvature distribution between the foot types was seen in both genders. Hallux valgus feet had lower radius of curvature than either normal (p = .004) or hallux rigidus (p < .001) feet. However, when controlling for gender, there was no significant difference between normal and hallux rigidus radius of curvature (p = .462).
Fig. 2. The radius of curvature for the 1st metatarsal bone for the three foot types. The radius of curvature is normalized by the length of the metatarsal to compensate for radiographic magnification and foot size (see Section 2). The error bars represent 95% confidence intervals. All comparisons between the foot types are significant.
Fig. 3. The radius of curvature for the 1st metatarsal bone for the three foot types, plotted separately for females and males. The radius of curvature is normalized by the length of the metatarsal to compensate for radiographic magnification and foot size (see Section 2). The error bars represent 95% confidence intervals.
2.1. Statistical methods Categorical variable comparisons were done using Chi-square statistics, with demographic continuous variables analyzed using Student’s t-test. Comparisons of the measurements between metatarsal head types were performed using nonparametric Kruskal–Wallis ANOVA, since the parameters were not normally distributed. Pairwise comparisons were performed with the Mann–Whitney U-test. In all cases, p < .05 was considered statistically significant. 2.2. Ethics This study was reviewed and approved by the University of Vermont IRB. 3. Results
Please cite this article in press as: J.D. Michelson, et al., Quantitative relationship of first metatarsophalangeal head morphology to hallux rigidus and hallux valgus, Foot Ankle Surg (2017), http://dx.doi.org/10.1016/j.fas.2017.04.017
G Model FAS 1054 No. of Pages 5
4
J.D. Michelson et al. / Foot and Ankle Surgery xxx (2017) xxx–xxx
The deviation from circularity was assessed by calculating the unbiased standard deviation of the points on the articular surface from a perfect circle (Fig. 4). The lowest deviation from circularity was shown in hallux valgus feet, with normal feet showing an intermediate amount of deviation and hallux rigidus feet demonstrating the greatest amount of deviation from circularity. 4. Discussion The current study quantifies the morphological differences in the 1st metatarsal head in patients having the highly prevalent conditions of hallux valgus (HV) and hallux rigidus (HR). Compared to one another, feet with HR demonstrated metatarsal heads that had higher radius of curvature and flatter than HV feet. This is consistent with earlier studies that have characterized the hallux metatarsal shape with respect to these two conditions. In assessing the relationship of the 1st metatarsal head to HV, several investigators have classified the 1st metatarsal head morphology subjectively into 3 categories: oval (or round), flat (or square), and chevron shaped (or square with a ridge). There has not been any attempt to further define these categories quantitatively. Coughlin and Jones [1] reported on 122 operated on for HV, in which 71% were felt to have an oval shape to the 1st metatarsal head. Mancuso et al. [4] had similar findings when they compared weight-bearing radiographs for 110 HV patients to 102 normal feet. Round 1st metatarsal heads were seen in 91% of their HV patients and 20% of the normal feet. Using a Mose template to define the circularity of the 1st MT head, Okuda et al. [5] that in a population consisting of 40 HV surgical patients and 60 normal feet, round heads were seen in 78% of the HV patient and only 1.7% of the normal controls. In a prospective survey of people over 65 years old, D’Arcangelo et al. [6] found the prevalence of a round 1st metatarsal head increased from 29% in those without HV to 87% in people with severe HV. In contrast, the studies of 1st metatarsal head morphology in hallux rigidus report a high prevalence of flatter or chevron shaped heads. Coughlin and Shurnas reported that 73% of the 137 feet operated upon for hallux rigidus had a chevron or flat shaped metatarsal head [3]. Similarly, Beeson et al. [7], found that in 180 feet with HR, 73% were either flat or chevron shaped. In a
Fig. 4. The deviation of the articular surfaces from pure circularity is measured calculating the unbiased standard deviation of the points on the articular surface from the derived circle that defined the articular radius of curvature (see Section 2). The standard deviations are normalized by the 1st metatarsal length to compensate for radiographic magnification and foot size (see Section 2). The error bars represent 95% confidence intervals. Statistical significant differences: # — p < .02 between Normal and HV subjects, ## — p < .001 between HV and HR subjects.
cadaveric study involve 120 specimens, Stein et al. [8] found that those with evidence of hallux rigidus had a larger radius of curvature, which they quantified on plain radiographs. However, their measurements were performed on non-weight bearing first metatarsals and the measurements were not normalized to account for foot size. There is a single imaging study comparing the appearance of HV and HR feet which was performed in 24 patients, only 15 of which had pure evidence of either condition [9]. They reported three of 11 HV patients had a round head compared to two of four HR patients. The study of Yoshioka et al., hypothesized that as the joint naturally sits in 16–20 of extension and 15 of valgus it is subjected to subluxation forces in excess of one body weight during normal gait [10]. In the current study, hallux valgus patients were found to have significantly smaller normalized radii of curvature from that of hallux valgus patients. From a biomechanical basis this supports the hypothesis that the valgus stresses across the joint in a hallux valgus patient would lead to lateral subluxation of the joint (increased valgus) due to the smaller metatarsal head curvature, whereas in the broader hallux rigidus metatarsal head, this would not lead to subluxation since the articulation has more inherent mediallateral stability [7]. The current study is the first to quantify the 1st metatarsal head morphology as it relates to underlying hallux clinicopathology (hallux valgus and rigidus). Although this demonstrates a measurement tool that is unlikely to find use in the clinical arena, it does serve to prove that the qualitative assessments that have been made by various other investigators previously are anatomically accurate. Since the more subjective categorization has demonstrated good inter-observer reliability [6], the continued use of the oval/flat/chevron classification groups for reporting research findings is justified. A long-term prospective study following patients for several years with correlation of progression to their disease state to hallux metatarsal head curvature will be needed to more firmly establish the biomechanical-pathologic links that the current study implies. 4.1. Limitations of the study The weight bearing foot radiographs did not include a marker ball to control for magnification. This was dealt with by normalizing the radius of curvature to the length of the metatarsal. This serves to cancel any changes that are due to magnification, as both parameters have the same magnification. This is similar to the methodology used by other investigators [11,4]. Differences in hindfoot and forefoot structure could potentially cause transverse joint curvature to be measured at relatively different angles. However, the transverse profile of the hallux metatarsal head (as seen on the AP foot radiograph) does not appreciably change through the relatively small variation in 1st metatarsal inclination (5–20 from the horizontal plane). The low variability of the results also supports the notion that metatarsal inclination does not have a large effect on the measurements. The study is limited to 2dimensional images of the foot. In the absence of 3-dimensional images for all of the subjects, it is not possible to know if systematic differences in the sagittal plane or complex 3-dimensional morphology can also be related to the clinical conditions affecting the hallux. While this is an important basic biomechanical research question regarding the etiologies of hallux rigidus and hallux valgus, the purpose of the current study is restricted to validating the subjective assessment made clinically and reported by others [7,3,8,1,4,6,5]. Our radiographic analysis of the shape of the metatarsal head was performed on a representative group of northeastern
Please cite this article in press as: J.D. Michelson, et al., Quantitative relationship of first metatarsophalangeal head morphology to hallux rigidus and hallux valgus, Foot Ankle Surg (2017), http://dx.doi.org/10.1016/j.fas.2017.04.017
G Model FAS 1054 No. of Pages 5
J.D. Michelson et al. / Foot and Ankle Surgery xxx (2017) xxx–xxx
American patients. Consequently, our findings in this relatively homogenous population cannot necessarily be generalized to other populations. 5. Conclusion The 1st MT head is quantitatively flatter in patients with HR than those with HV, and shows more deviation from pure circularity (e.g. — a chevron morphology) in the HR patients. This is consistent with previous qualitative observations, and validates the use of these qualitative observations to categorize metatarsal head morphology. Conflicts of interest The authors declare that they have no conflict of interest. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
5
References [1] Coughlin MJ, Jones CP. Hallux valgus: demographics, etiology, and radiographic assessment. Foot Ankle Int 2007;28:759–77. [2] Lucas DE, Hunt KJ. Hallux rigidus: relevant anatomy and pathophysiology. Foot Ankle Clin 2015;20:381–9. [3] Coughlin MJ, Shurnas PS. Hallux rigidus: demographics, etiology, and radiographic assessment. Foot Ankle Int 2003;24:731–43. [4] Mancuso JE, Abramow SP, Landsman MJ, Waldman M, Carioscia M. The zeroplus first metatarsal and its relationship to bunion deformity. J Foot Ankle Surg 2003;42:319–26. [5] Okuda R, Kinoshita M, Yasuda T, Jotoku T, Kitano N, Shima H. The shape of the lateral edge of the first metatarsal head as a risk factor for recurrence of hallux valgus. J Bone Joint Surg Am 2007;89:2163–72. [6] D'Arcangelo PR, Landorf KB, Munteanu SE, Zammit GV, Menz HB. Radiographic correlates of hallux valgus severity in older people. J Foot Ankle Res 2010;3:20. [7] Beeson P, Phillips C, Corr S, Ribbans WJ. Cross-sectional study to evaluate radiological parameters in hallux rigidus. Foot (Edinb) 2009;19:7–21. [8] Stein G, Pawel A, Koebke J, Muller LP, Eysel P, Thelen U, et al. Morphology of the first metatarsal head and hallux rigidus: a cadaveric study. Surg Radiol Anat 2012;34:589–92. [9] Schweitzer ME, Maheshwari S, Shabshin N. Hallux valgus and hallux rigidus: MRI findings. Clin Imaging 1999;23:397–402. [10] Yoshioka Y, Siu DW, Cooke TD, Bryant JT, Wyss U. Geometry of the first metatarsophalangeal joint. J Orthop Res 1988;6:878–85. [11] Munuera PV, Dominguez G, Castillo JM. Radiographic study of the size of the first metatarso-digital segment in feet with incipient hallux limitus. J Am Podiatr Med Assoc 2007;97:460–8.
Please cite this article in press as: J.D. Michelson, et al., Quantitative relationship of first metatarsophalangeal head morphology to hallux rigidus and hallux valgus, Foot Ankle Surg (2017), http://dx.doi.org/10.1016/j.fas.2017.04.017