The foot is often described in current foot models by substantially fewer segments than the actually existing anatomic structures [7
]. However, for the determination of instabilities and their specific anatomical localization occurring during the gait cycle, a more detailed analysis seems to be necessary in view of the fact that clinicians claim that instability in hallux valgus is mostly confined within the first tarso-metatarsal joint which indicates surgery in the form of an arthrodesis of this joint [4
The first ray of the foot has an important function with respect to load transfer and stability [2
]. The relevance of instability of the first ray in hallux vagus deformity has been a matter of discussion for decades. The particular pathogenesis of hallux valgus formation and deformity progression has been attributed to first ray instability with variable percentage [4
To date, the diagnosis of a hypomobile or hypermobile first ray has mostly been performed by clinical assessment with a high degree of variance [5
] or by static measurement methods at a limited dorsiflexion force of 55 N [9
]. Consequently, as a result of these approaches the normal mobility of the first ray had been defined between 4 and 8 mm. Standard weight-bearing radiographs only add indirect and mostly inconstant static signs of medial cuneiform-first metatarsal instability (plantar joint opening, localized osteoarthritis or widened first intermetatarsal angle) [5
Up to now, dynamic measurements of forefoot kinematics were mostly performed with opto-electronical methods. The values of the forefoot dorsal flexion in relation to the hindfoot in normals were found to be within the range of 3.0-6.2° [7
Dynamic standard fluoroscopic analysis has been employed in a variety of in-vivo biomechanical evaluations, e.g. normal and anterior cruciate-deficient knee joint kinematic studies or kinematic studies following various types of knee joint arthroplasty [23
]. In most application modes additional data acquisition such as CT scanning of the knee was necessary as a prerequisite for the generation of a mathematical 3-D model utilizing an iterative model-fitting approach. Data from conventional dynamic fluoroscopy itself can not readily be taken for direct measurements due to a substantial amount of magnification and distortion depending on the distance of the corresponding image point from the focus. The novel digital fluoroscopic acquisition tool used in this study allowed a distortion-free and detailed analysis of sagittal motion at the anatomic joints of the medial ray. Such detailed in-vivo analyses have not been possible previously. Of course, this novel method has several limitations and potential errors. The errors arising from the described evaluation method can to be divided into three categories:
(i) errors due to the spatial and temporal resolution of the imaging system,
(ii) errors due to two-dimensional imaging and evaluation,
(iii) errors due to manual positioning and digitalisation of the image.
These errors were estimated in more detail. The 25 frames/s imaging rates only allow the analysis of relatively slow walking speeds. In our chosen set-up, free walking was replaced by a single step analysis, as walking at a continuous speed would mean that the fluoroscope would also have to move. In principal, additional analysis of in-vivo kinematics in the horizontal plane during foot contact phase would be desirable [2
] but could hardly be accomplished by the current technique due to an inevitable mechanical interference of the walking person with the fluoroscopic device turned to a vertical or oblique position.
The main source of errors was expected from (iii). The outline of bone was manually drawn into an image where all contours were well visible. However, some images, in particular during the last phase of roll-over process, were blurry so that errors arose from the transfer of the contours onto the subsequent images. Further, the determination of the medial foot column mobility depended on the correct determination of the talar contour. An angular deviation of 1° of malpositioning of the talar contour would change the first ray mobility result by about 2.5 mm. The errors from the other sources (i) and (ii) were considered to have only a minor influence on the results. Moreover, the errors (ii) and (iii) were minimized as much as possible by copying the bone contours from images where these contours could be well defined to the more blurry images.
A comparison of our results was made with data from the static analyses, with data obtained by the classic marker technique and with data from kinematic analyses by camera. Our data indicate a substantially higher first ray mobility than described in the literature for static measurements [9
]. Taking into account the substantially higher loading during single-leg full weight-bearing as in our experiments, the limited loading during static measurements and the difference between static and dynamic values [7
] this does not appear to be surprising. Between the groups of the healthy volunteers and the patients there were, however, only statistically non-significant differences with 13.6 mm and 13.1 mm, respectively.
More recent studies use an advanced combined 2D-3D model-image registration technique for foot kinematics in healthy subjects [26
], patients with hallux valgus [30
], hallux rigidus [22
], flat-arched feet [31
] and for subjects with ankle arthroplasty [32
]. These studies report 15° of plantarflexion to 20° dorsiflexion with the healthy subjects with and without weight-bearing activities. The calculated angular values of forefoot dorsiflexion for both our two groups of subjects, nevertheless, were comparable to literature data of opto-electronical measurements ranging between 0.7 and 9.3° [7
]. Furthermore, angular measurements are independent of linear measures, such as the individual foot length, and seem to be generally preferable compared to mere distance measures. With 5.3° and 5.6° only statistically non-significant differences between the groups of the healthy volunteers and the patients were found in our study. Present 3D multisegment foot models have been shown to have a very high reliability index for the sagittal plane kinematics. Moreover, they also yield data for the motion within the coronal and horizontal planes. However the adequate marker placement, soft tissue artifacts stereophotogrammetric-based marker position tracking and the basic assumptions of the corresponding foot model do have an influence on the calculations of the corresponding joint rotations [22
The relative rotational movements in the sagittal plane in our study did not show distinguishable differences between both groups. Still, it is noticeable that, in contrast to clinical assumptions [5
], an increased mobility at the first metatarsal-medial cuneiform articulation was not seen in either of our two groups. Compared with the navicular-medial cuneiform articulation and the talo-navicular joint even the smallest rotations were found at the first metatarsal-medial cuneiform articulation. This agrees well with data from in-vitro experiments [15
] and reports from a limited number of invasive in vivo assessment of mid- and forefoot motion during walking [18
] or slow running [28
The groups of the healthy volunteers and the patients differed significantly in the time-point of occurrence of the maximum values of the first ray flexion and the relative rotations of the bones to each other. The motion diagrams within the group of patients reached their maxima with heel rise. In contrast, the motion diagrams of the healthy volunteer group reached their maximum values significantly later which might at least point to increased medial ray flexibility in the patient group despite a comparable total range of motion in both groups.
The translational relative motions between the foot bones are considered to be in the order of magnitude of the measurement precision. With very limited values between 1 and 2 mm a characteristic curve form could not be recognized.
The values of the standard deviation for the first ray flexion and the relative rotational motion of the bones were relatively high within both of our groups of test subjects. This indicates high inter-individual variations within the groups. Maximum rotational motions could be found mainly in the navicular-medial cuneiform articulation and in the talo-navicular articulation which has also been reported by invasive measurements of rear-, mid- and forefoot motion [18