The objective of this study was to evaluate the radiographic assessment of the hallux valgus angle of the margo medialis pedis (medial border of the foot) as an alternative to the metatarsophalangeal hallux valgus angle. Basically the hallux valgus angle of the margo medialis pedis can be derived from any method which provides a clear picture of the foot outline without any distortion. Besides non radiographic methods like foot prints, foot scans or foot outline drawings, which are potential methods for screening purposes or preventive interventions, the hallux valgus angle of the margo medialis pedis can also be measured on x-ray images given that the soft tissue of the foot is sufficiently represented. In this study the hallux valgus angle from the margo medialis pedis and the metatarsophalangeal hallux angle were taken from the same x-ray pictures. This gives the opportunity to measure the metatarsophalangeal angle and the hallux valgus angle of the margo medialis pedis, under identical conditions (weight-bearing, foot position, temperature).
The three hallux valgus measuring methods used in the present study have a very good performance with reliability coefficients above 0.95 (p = 0.001) and low coefficients of variation (Table 1). For all methods intra- and interrater concordance correlation coefficients are higher than 0.96 (p = 0.001). Estimating the results of the metatarsophalangeal methods (SB and CC) from the results of the margo medialis pedis method (MMP) is associated with low random bias, high precision (>0.89) and high accuracy (>0.88) (Table 2). Results are nearly identical for SB and CC method. However, there was a systematic bias of 4.8 degrees, constant over the whole range of measurement for both methods. The hallux valgus angle of the margo medialis pedis is consistently smaller than the metatarsophalangeal angle.
We assume that this systematic bias is due to the anatomical varus position of the metatarsale 1 in respect to the os cuneiforme and os naviculare, which has direct impact on the metatarsophalangeal angle but does not affect the hallux valgus angle of the margo medialis pedis. Further studies will be necessary to prove this assumption.
Contrary to a study by Resch[9], in which an interobserver measurement difference of 6.4 degrees was observed in the measurement of the metatarsophalangeal angle, results of the current study show that independent observers arrived at the same results. The interobserver reliability coefficients were higher than 0.96 for all three measurement methods. There was no essential difference between the methods in this respect, though the reliability was a little higher in the margo medialis pedis method. We assume that the high reliability in our study is due to the great emphasis placed on exact measurement definition, the standardised analysis guidelines and the comprehensive pre-experience of the observers with these measurement methods.
A number of different methods for measuring the metatarsophalangeal angle have been published to date[5–10]. The relevant literature gives standard values for the metatarsophalangeal angle, and for the first and second intermetatarsal angles[12, 22–24].
The influence of the method on the angle values measured was first discussed by Barnicot and Hardy[13], and confirmed by Schneider and Knahr[12]. For this reason we used two methods for the evaluation of the metatarsophalangeal angle: The shaft bisection method[10], a standard method, and a slight variation of the centre base - centre head method[6] with a detailed advice for the construction of the reference point. The reference point construction disregards the cartilaginous joint surface area on the metatarsal head and therefore can also be identified on postoperative x-ray images after removal of the medial exostosis.
Both methods show high reliability and low variability. The variability of the CC method is lower than that of the SB method and the reliability higher probably due to the more strict definition of the measurement procedure for CC method. Data of this study indicate that the two methods (SB and CC) are unbiased estimates of each other (Table 3, Figure 9). Further studies would be necessary to prove how far the expected advantage of the CC method for pre/post operative hallux valgus angle comparisons holds true.
A limitation of the study is that only patients with a symptomatic hallux valgus deformity were included. It should be mentioned that a direct transfer of our results to non-radiographic material necessary for serial screening investigations is not yet possible.
The main aim of this study was to get basic information about the possibility to use the hallux valgus angle from the margo medialis pedis as an alternative to the metatarsophalangeal hallux angle. We prefered to use the same x-ray pictures for both measurement methods as only this guaranties identical external influences on the hallux position. Though systematic differences between margo medialis measurements from radiographic and non-radiographic pictures are unlikely further investigations necessary. In a preliminary additional study we found very similar hallux valgus values from margo medialis pedis measurements based on foot outline drawings and x-ray pictures. 14 feet from staff members were investigated by two raters two times each. Hallux valgus measures varied from 4° to 37,4° for foot outline drawings and from 0° to 37° for x-ray pictures. T-test showed no significant differences between HV from foot outline drawings and x-ray pictures. (t = 1.7, df = 13, p < 0.5) Correlation between methods as well as intra - and interrater reliability were significant at the p = 0.001 level. (methods r = 0.880; intrarater r = 0.991; interrater r = 0.997) It may be assumed that further results will show that the margo medialis pedis hallux valgus angle derived from non-radiological material measurements is suitable for epidemiological screening of the hallux valgus angle.