- Email: [email protected]
Subtalar neutral position as an offset for a kinematic model of the foot during walking
Available online at www.sciencedirect.com
Gait & Posture 28 (2008) 29–37 www.elsevier.com/locate/gaitpost
Subtalar neutral position as an offset for a kinematic model of the foot during walking Jeff R. Houck *, Josh M. Tome, Deborah A. Nawoczenski Department of Physical Therapy and Center for Foot and Ankle Research, Ithaca College-Rochester Campus, 1100 S. Goodman Street, Rochester, NY 14620, USA Received 19 December 2006; received in revised form 13 September 2007; accepted 18 September 2007
Abstract The lack of a common reference position when defining foot postures may underestimate the ability to differentiate foot function in subjects with pathology. The effect of using the subtalar neutral (STN) position as an offset for both rearfoot and forefoot through comparison of the kinematic walking patterns of subjects classified as normal (n = 7) and abnormally pronated (n = 14) foot postures was completed. An OptotrakTM Motion Analysis System (Northern Digital, Inc.) integrated with Motion Monitor SoftwareTM (Innovative Sports, Inc.) was used to track three-dimensional movement of the leg, rearfoot and first metatarsal segments. Intrarater reliability of positioning the foot into STN using clinical guidelines was determined for a single rater for 21 subjects. Walking data were subsequently compared before and after an offset was applied to the rearfoot and first metatarsal segments. Repeated measures of foot positioning found the STN position to be highly repeatable (intraclass correlation coefficients > 0.9), with peak errors ranging from 1.98 to 4.38. Utilizing STN as the offset resulted in a significant increase in rearfoot eversion ( p = 0.019) during early stance, and greater first metatarsal dorsiflexion ( p < 0.007) across stance in the pronated foot groups that was not observed prior to applying the offset. When applied to subjects with differing foot postures, the selection of a common reference position that is both clinically appropriate and reliable may distinguish kinematic patterns during walking that are consistent with theories of abnormal pronation. # 2007 Elsevier B.V. All rights reserved. Keywords: Foot; Kinematics; Subtalar neutral; Gait
1. Introduction Advances in technology have enabled improved in vivo assessment of foot function across multiple foot segments. Various kinematic models representing multi-segment foot function have been proposed for studies of walking [1–7], however, a number of technical aspects of implementing these foot models remain controversial. Key issues to consider when implementing a kinematic foot model are the potential for skin artifact errors [8–10], the designation of anatomical coordinate systems [3–5], and the definition of a ‘‘neutral’’, or reference position of the foot joints [4,6]. Skin artifact errors arise from movement of the markers relative to the underlying bone or bones. Careful selection of marker * Corresponding author. E-mail address: [email protected] (J.R. Houck). 0966-6362/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2007.09.008
placement on the foot may avoid large skin artifact errors for targeted bones. For the calcaneus and first metatarsal bones, recent studies found skin artifact errors to be minimal, falling within a range of <38 [8–10]. Variability across studies may also be related to the definition of local anatomical coordinate systems. Use of radiographs [3], a special jig [6] or digitized points [4] have been used to improve the identification of local anatomical coordinate systems for subsequent recording of relative motion during walking trials. An additional procedural component to the description of relative motion between anatomic segments is the establishment of a neutral position, or reference position. Various methods have been used to designate the neutral position and include relaxed standing postures, weight-bearing subtalar joint position, alignment to an external frame, alignment of the rearfoot to the tibia, or reference to foot function during the stance period of gait
30
J.R. Houck et al. / Gait & Posture 28 (2008) 29–37
[2,4,6,11]. Using this neutral position, an offset is created that is subsequently applied to the motion data. To date, the optimal approach to define a neutral position of the foot and the application of this offset to subjects with abnormal foot pronation postures have not been established. To reduce noise and make valid comparisons across subjects of different foot postures, a defined neutral position that is both reliable and clinically meaningful is preferred. Clinicians who assess foot posture frequently use an operational definition of neutral termed the ‘subtalar neutral’ (STN) position [12–14]. The operational definition of STN requires clinicians to palpate the talonavicular joint and determine the mid-position, or position of maximum congruency [14]. Even with the degree of nebulousness associated with this definition, well-trained examiners have been able to repeatedly place the rearfoot of healthy subjects in STN position within a few degrees (<38) [15,16]. Investigators have also used the relaxed standing posture as a designated neutral position. Outcomes from two-dimensional (2D) analysis studies have shown mean rearfoot eversion motion to be centered around rearfoot postures observed during relaxed standing posture, and not the STN position [13,17,18]. Interpretation of the results needs to be made with caution, however, because of the limitations associated with 2D analysis, particularly as they apply to out of plane rearfoot rotations at the beginning and end phases of stance. Additionally, because of the large variability in standing postures assumed by subjects, the use of a relaxed standing trial as the neutral position for kinematic models may misrepresent the excursion and endpoints of joint motion that are frequently linked to pathology in pronated foot types. The first purpose of this study was to determine the reliability of rearfoot and forefoot position using a clinical method for establishing a STN position in standing. To improve clinical application, subjects classified as normal and (abnormally) pronated foot postures were included. Five kinematic measurements, frequently implicated in abnormal foot function were evaluated for the first purpose of the study: (1) rearfoot inversion/eversion; (2) dorsiflexion/plantarflexion of the rearfoot relative to the tibia; (3) first metatarsal dorsiflexion/plantarflexion relative to the rearfoot; (4) hallux dorsiflexion/plantarflexion relative to the first metatarsal; and (5) first metatarsal elevation relative to the laboratory coordinate system. The second purpose of this study was to compare two key indicators of foot pronation, rearfoot inversion/eversion and first metatarsal dorsiflexion/plantarflexion [19–21] during walking between subjects classified as normal and pronated foot type using two distinct definitions of the neutral position: relaxed standing and STN. A common reference foot posture is needed to appropriately represent foot function during gait and enable comparisons across subjects of different foot postures. As a clinically relevant position, STN may be more effective as a neutral position than a relaxed standing posture when comparing foot kinematics in subjects with varying foot postures.
Table 1 Subject characteristics
Age (years) Height (cm) Mass (kg) Shoe size a
Normal (n = 7)
Pronator (n = 14)
p-Valuea
22 1 166.2 6.0 66.3 10.9 8.0 0.9
22.2 1.3 167.9 9.0 67.1 15.7 9.4 1.7
0.68 0.62 0.89 0.05
p values indicate result of a Student’s t-test.
2. Methods 2.1. Subjects Twenty-one subjects (seven controls and 14 pronators) consented to participate in this study. A power analysis derived from pilot data of four subjects revealed a sample of 14 subjects, or seven per group was needed to detect differences between pronators and controls with 80% power. To improve clinical relevance, a larger group of subjects classified as pronators were included. With the exception of gender, there were no significant differences between the groups for age, height and mass (Table 1). Of the 21 participants, only three were males. Subjects were considered to be abnormal pronators if they met the inclusion criteria assessed during the clinical exam: a forefoot varus deformity exceeding 108, measured goniometrically in the prone, non-weight-bearing position; rearfoot eversion beyond vertical during weight bearing; and a minimum navicular drop difference of 10 mm or greater from a position of STN to relaxed standing. One or more of these findings have been used as clinical indicators of abnormal pronation in previous investigations [22–24]. The clinical examination was performed by an experienced clinician with greater than 20 years experience (DAN). The study protocol was approved by the Review Board for Human Subjects Research at Ithaca College. Data collection took place in the Movement Analysis Laboratory, Center for Foot and Ankle Research at Ithaca College, Rochester Campus facility. 2.2. Kinematic model The included foot segments were the rearfoot (calcaneus), first metatarsal and hallux. The tibia was also tracked to establish rearfoot position with respect to the tibia. Each segment was tracked by placing three infrared emitting diodes (IREDs) on a thermoplastic molded platform on the skin overlying the calcaneus, first metatarsal, and hallux (Fig. 1). Previous in vitro [9] and in vivo studies [25] reported good repeatability and validity of tracking first metatarsal and hallux segments using skin sensors. Each segment was tracked at 60 Hz using an OptrotrakTM Motion Analysis System (Northern Digital, Inc., CAN). Kinematic data were smoothed using a fourth order, zero phase lag, Butterworth filter with a cut off frequency of 6 Hz. A right-handed Cartesian frame was used to establish coordinate systems for each segment. Anatomical landmarks were digitized on each segment, and defined as follows: Tibia:
Origin: located on the lateral malleolus. y-Axis: joins the lateral malleolus to the fibular head and its positive direction is superior.
J.R. Houck et al. / Gait & Posture 28 (2008) 29–37
31
head of the proximal phalanx with its positive direction superior. The orientation of this axis reflects the anatomical position of the hallux. z-Axis: orthogonal to the x–y plane with its positive direction to the right.
Fig. 1. Infrared emitting diodes (IREDs) on thermoplastic molded platforms placed on the skin overlying the calcaneus, first metatarsal, hallux, and tibia.
x-Axis: orthogonal to the plane defined by the y-axis and medial malleolus with its positive direction anterior. z-axis: orthogonal to the x–y plane with its positive direction to the right. Rearfoot (all points digitized on the floor):
Origin: midpoint of the posterior heel. x-Axis: joins the midpoint of the posterior heel to the tip of the second phalanx and its positive direction points anterior. y-Axis: orthogonal to the plane defined by the x-axis and a point lying on the floor with its positive direction superior. This resulted in a vertical axis aligned with the laboratory coordinate system. z-Axis: orthogonal to the x–y plane with its positive direction to the right. First metatarsal:
Origin: base of the first metatarsal. x-Axis: joins the base of the first metatarsal to the head of the first metatarsal and its positive direction points anterior. y-Axis: orthogonal to the plane defined by the x-axis and another point of equal height off the floor as the first metatarsal head with its positive direction superior. The orientation of this axis reflects the anatomical position of the first metatarsal (metatarsal declination). z-Axis: orthogonal to the x–y plane with its positive direction to the right. Hallux
Origin: base of the proximal phalanx. x-Axis: joins the base of the proximal phalanx to the head of the proximal phalanx and its positive direction points anterior. y-Axis: orthogonal to the plane defined by the x-axis and another point of equal height off the floor as the
Angles were determined using a ZXY sequence as proposed by Cole et al. [26] such that dorsiflexion/plantarflexion represented rotations about the z-axis and inversion/eversion were represented as rotations about the anterior-directed x-axis. The measurements defined for this analysis included the angular orientation of the rearfoot with respect to (w.r.t.) the tibia (dorsiflexion/plantarflexion and eversion/inversion), the first metatarsal w.r.t. the rearfoot (dorsiflexion/plantarflexion), the hallux w.r.t. first metatarsal (dorsiflexion/plantarflexion) and vertical height of the first metatarsal head in the laboratory reference frame. The first metatarsal and hallux angles, as well as the height of the first metatarsal head were included in this analysis because of their presumed relationship to changes in rearfoot alignment. In some cases of abnormal pronation, the medial forefoot and/or metatarsal head may elevate when a subject’s foot is placed in a STN position during the non-weightbearing exam. This is referred to as forefoot varus [27,28]. These same changes may also be observed during standing when the subtalar joint is placed in neutral. Therefore, inclusion of these variables was deemed relevant to the assessment of the STN position. 2.3. Procedures The relaxed standing trial was acquired as subjects placed their feet in a natural, self-selected posture, attempting equal weight bearing on both feet. For the STN trial, a single experienced examiner (DAN) used standard clinical procedures to palpate the STN position. Briefly, the examiner palpated the talonavicular articulation while the subject rolled the rearfoot into inversion and eversion, in effect causing arch elevation and lowering. STN was defined as the mid-position of the talar head relative to the navicular bone. Once the STN position was established by the examiner, subjects were asked to hold that position while a three second trial was collected. The subject then walked in place and the process was repeated a second time. No feedback was given to the examiner or the subject. After completing the standing trials, each subject walked down a walkway at a self-selected natural walking speed. A minimum of five walking trials were collected. A force plate (Kistler Instrument Corp., Amherst, NY, model 9865B) embedded in the floor was used to determine initial contact and toe off phases of stance. 2.4. Definitions of neutral position Two distinct designations of foot postures, established during relaxed standing and STN position were created for the rearfoot and forefoot segments. For the rearfoot, the inversion/eversion position of the calcaneus with respect to the tibia when the foot was placed in the STN position was considered the rearfoot neutral position. A frontal plane offset (uc), calculated as the difference between the STN position and relaxed standing (Fig. 2A and B), was applied to the inversion/eversion patterns of the rearfoot with respect to the tibia. Kinematic patterns without, and with this applied offset are termed uncorrected, and corrected data, respectively and are shown in Fig. 4.
32
J.R. Houck et al. / Gait & Posture 28 (2008) 29–37
Fig. 3. Left foot image of the sagittal plane offset (uM) applied as a rotation around the medial/lateral axis of the first metatarsal segment to account for first metatarsal head elevation [B1 = base of first metatarsal during standing, B2 = base of first metatarsal during subtalar neutral, H1 = head of first metatarsal during standing, H2 = head of first metatarsal during subtalar neutral, H3 = repositioned placement of first metatarsal head to its standing position and uM = angle required to rotate first metatarsal head to its standing position]. Fig. 2. Left foot posterior image of the frontal plane offset (uc), necessary to result in neutral inversion/eversion when the subject was positioned in subtalar neutral, was applied to the inversion/eversion patterns of the rearfoot with respect to the tibia.
degree of correction this would result in a 30–408 shift in plantarflexion/dorsiflexion of the first metatarsal w.r.t. the calcaneus. 2.5. Analysis
For the forefoot, a sagittal plane adjustment to the first metatarsal segment was made to account for potential changes in first metatarsal position that occurred when the foot was placed in the STN (Fig. 3A and B). For subjects with abnormal pronation, it is not uncommon to find elevation of the first metatarsal/medial forefoot, indicating a non-rigid forefoot varus deformity, when the foot is placed in STN [27,28]. Yet, healthy subjects are able to assume the STN position while maintaining contact of the first metatarsal head with the ground. To establish a comparable first metatarsal neutral position across subjects (i.e. the first metatarsal head is on the floor while the mid-foot is manipulated into the STN position), an offset (uM) was applied that plantarflexed the first metatarsal so the head would rest on the floor (Fig. 3A and B and Appendix A). Once the offset was applied to the first metatarsal, the corrected data was calculated by noting the plantarflexion angle of the first metatarsal w.r.t. the calcaneus in the STN position and subtracting this value from the uncorrected data. Depending on the
An intraclass correlation coefficient (ICC) was used to evaluate reliability of the five inter-segmental kinematic measures as subjects were placed in STN. ICC, standard error of the measurement (S.E.M.), and absolute error data are provided for repeated measurements on these variables by the same tester. The ICC, model 3, identified by Portney and Watknins as appropriate for intrarater reliability was used in this study [29]. To estimate the expected error in degrees, two approaches were used including calculating the percent agreement and the standard error of the measurement (S.E.M.). Absolute errors were the absolute values of the difference between trial one and trial two for each subject. In previous studies errors in rearfoot inversion/eversion of 28 were clinically desirable [15,16]. Therefore the percentage of trials resulting in differences of 18 or less and 28 or less were reported. The standard error of the measurement assumes a normal distribution of errors and therefore is a metric that is useful for understanding what errors are expected across subjects, where errors are expected to stay within 1 S.E.M.
J.R. Houck et al. / Gait & Posture 28 (2008) 29–37
68% of the time [29]. The calculation of the S.E.M. is included in the following equation: pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi standard error of the measurement ¼ S:D: 1 ICC value
33
time. Actual errors were below 28 or 2 mm 81% of the time for all variables.
(1)
3.2. Walking where S.D. is the pooled standard deviation of trial one and trial two and ICC value is the correlation coefficient from the ICC statistic. For the second purpose of the study, kinematic data from each walking trial were interpolated to 1% increments of stance and then averaged across five trials to gain a representative pattern for each variable for each subject. First, the data were processed without applying an offset (uncorrected). Second, the definitions of neutral for the rearfoot and first metatarsal using the STN position were applied to the data (corrected). Defined points across stance (initial contact and peaks) were selected to compare patterns statistically before, and after applying the STN definition of neutral. To assure the same point of stance were used in the analysis between groups, peaks were first determined from the patterns of the pronator group, and then the same point of stance of the normal group was used. For rearfoot eversion/inversion the stance points used were initial contact, peak eversion (28% of stance) and peak inversion (96% of stance). For first metatarsal dorsiflexion/plantarflexion the stand points used were initial contact, peak dorsiflexion (73% of stance) and peak plantarflexion (100% of stance). A two-way, mixed linear model which included a fixed factor of groups (normal versus pronator) and repeated factor of stance point were used to compare patterns. Due to the imbalance in sample sizes maximum likelihood estimation was used. In the presence of group stance point interaction, pairwise comparisons were conducted. For each twoway ANOVA and pairwise comparison an alpha level of 0.05 was utilized to indicate significance. All statistical procedures were completed using SAS version 9.1 software.
3. Results 3.1. Reliability The five measures of foot position during the first and second STN position were highly repeatable (Table 2). The ICC values were greater than 0.9 for all variables. The peak errors ranged from 1.98 to 4.38. The two S.E.M. statistic predicted repeatability errors would be less than 1.4–2.48 96% of the time. Peak errors associated with first metatarsal elevation were as high as 5.3 mm. The two S.E.M. statistic predicted that errors would fall within 2.4 mm 96% of the
The stance points used in the analysis for rearfoot eversion/inversion and first metatarsal dorsiflexion/plantarflexion distinguished the pronator group for both the corrected and uncorrected data (Table 3 and Fig. 4). A significant interaction (group stance point) occurred for both uncorrected ( p = 0.034) and corrected ( p = 0.034) rearfoot eversion/inversion. The pairwise comparisons of the uncorrected rearfoot eversion/inversion data suggested the pronated group was significantly more inverted then the normal group at initial contact (2.1 4.08 vs 2.8 4.08, p = 0.016) and 96% of stance (6.4 4.98 vs 2.0 3.98, p = 0.05). In contrast the corrected rearfoot eversion/ inversion patterns suggested significantly greater rearfoot eversion at 28% of stance ( p = 0.019) for the pronated group (7.5 3.38) compared to those in the normal group (4.0 2.08). The stance points used in the analysis of first metatarsal dorsiflexion/plantarflexion suggested a significant main effect of greater dorsiflexion in the corrected data but not the uncorrected data (Table 3 and Fig. 5). The uncorrected first metatarsal dorsiflexion/plantarflexion data was not significantly different across stance between the pronator and normal groups (normal group = 30.4 5.68 versus pronator group = 28.7 5.28; p = 0.51). In contrast, the corrected first metatarsal dorsiflexion/plantarflexion data across stance (group main effect) indicated significantly greater dorsiflexion ( p = 0.007) of the pronator group (4.6 4.58) compared to the healthy group (1.6 4.58). There was a trend toward a significant interaction ( p = 0.07) in both the corrected and uncorrected first metatarsal dorsiflexion/plantarflexion data.
4. Discussion The findings of this study show the STN position to be highly repeatable within an examiner, and the position may be reliably applied to a clinical population with foot postures
Table 2 Summary of reliability statistics for subtalar neutral position Variable
ICC (95% CI)
S.E.M.
Absolute error (range)
<18 (%)
<28 (%)
Rearfoot Calcaneus eversion Calcaneus dorsiflexion
0.96 (0.89–0.98) 0.94 (0.87–0.98)
1.08 0.78
0.9 0.98 (0.1–4.38) 0.9 0.58 (0.1–1.98)
48 67
90 100
Forefoot First metatarsal elevation First metatarsal dorsiflexion Hallux dorsiflexion
0.96 (0.92–0.99) 0.95 (0.88–0.98) 0.95 (0.89–0.98)
1.2 mm 1.08 1.28
0.7 0.98 (0.01–5.2 mm) 1.1 0.88 (0.01–2.98) 1.3 1.18 (0.2–3.88)
57 a 67 52
a
The percentage below 1 mm or 2 mm, not degrees, is reported in each respective column.
81a 90 81
34
J.R. Houck et al. / Gait & Posture 28 (2008) 29–37
Table 3 Mean (1 S.D.) for rearfoot and forefoot data Normal
Pronator
D
ANOVA
Calcaneal eversion Uncorrected Initial contact Peak () eversion Peak (+) inversion
2.8 4.0 7.3 3.1 2.0 3.9
2.1 4.0* 6.0 4.4 6.4 4.9*
4.9 1.3 4.4
Group = 0.057 Group stance point = 0.034
Corrected Initial contact Peak () eversion Peak (+) inversion
0.4 2.6 4.0 2.0 5.2 1.2
0.5 2.8 7.5 3.3* 4.8 1.8
0.1 3.5 0.4
Group = 0.138 Group stance point = 0.034
2.0 4.5 1.5
Group = 0.512 Group stance point = 0.069
6.6 9.1 4.1
Group = 0.007 Group stance point = 0.069
First metatarsal dorsiflexion Uncorrected Initial contact Peak (+) dorsiflexion Peak () plantarflexion
30.0 5.5 21.3 3.2 39.8 4.2
Corrected Initial contact Peak (+) dorsiflexion Peak () plantarflexion
1.3 5.3 7.4 4.8 11.1 3.6
*
28.0 5.8 16.8 6.1 41.3 8.6 5.3 4.6* 16.5 4.9* 7.0 7.3
p values of less than 0.05 indicated by pairwise comparisons.
associated with abnormal pronation. Unique to this study was the inclusion of the medial forefoot (i.e., first metatarsal segment) in the determination of the neutral position. Prior investigations have not considered the changes in the medial forefoot that may occur with STN positioning, nor the implications of a forefoot offset when assessing movement patterns. The high reliability of both forefoot and rearfoot variables supported the inclusion of both segments when defining the neutral position. Only when applying these neutral offsets (derived from the STN position) did rearfoot eversion and first metatarsal dorsiflexion show strong statistical differences in foot function consistent with clinical theories of pronation. These data support the current practice of using the STN neutral position as a clinical metric for application to kinematic foot models. The data from this study suggest the STN position is a reliable method for estimating a neutral position of the foot. Previous studies examining intrarater reliability suggested clinicians were able to place the rearfoot within 38, 90% of the time [15,16]. Our data is more optimistic suggesting experienced clinicians are able to place the rearfoot within 28, 90% of the time. Unlike previous studies using threedimensional techniques, this study included subjects classified as abnormal pronators based on clinical exam. Difficulties positioning subjects with abnormal foot postures were expected to increase errors relative to other studies. Surprisingly, the errors noted in this study were slightly lower for standing STN, but similar to values obtained in the prone and seated STN positions of healthy subjects for a single examiner [15]. The S.E.M. values suggest that using the STN position to define a reference position is suspect to re-positioning errors of 2–38 for the rearfoot.
Not addressed in previous studies was the reliability of positioning the forefoot. It is not uncommon to find forefoot deviations, frequently reported as forefoot varus, in persons with abnormal pronation during the non-weight-bearing clinical examination. These same deviations may also be observed during the weight-bearing exam when the foot is positioned in STN and the first metatarsal is observed to elevate. To date, there have been no studies that have assessed forefoot corrections in a clinically pronated foot group. Similar to S.E.M. values reported for the rearfoot, repositioning errors of 2–38 were also found for the first metatarsal kinematics. These reliability errors (2–38) are smaller than the expected changes in kinematics associated with many foot and ankle pathologies. Therefore, using the STN position as a neutral position is a useful method for accounting for differences in foot posture. Applying the STN position as a definition of neutral position to the rearfoot and first metatarsal kinematic patterns significantly influenced the interpretation of foot function. The uncorrected data for rearfoot eversion/ inversion of the normal subjects suggested higher rearfoot inversion of the pronator group during initial contact and late stance (Fig. 4), and no differences during early stance—data inconsistent with clinical observations. Only when the data were corrected using the STN position as neutral did the pronator group show greater rearfoot eversion during early stance. Similarly, differences in peak first metatarsal dorsiflexion between the normal and pronator group were amplified when applying the definition of neutral that incorporated the STN position (Fig. 5). Increased first metatarsal dorsiflexion reflects a lowering of the anterior component of the medial longitudinal arch. Again, the findings of a lower arch
J.R. Houck et al. / Gait & Posture 28 (2008) 29–37
Fig. 4. Comparison of uncorrected (relaxed standing) and corrected (STN) kinematics for rearfoot eversion/inversion relative to the tibia. Standard error of the mean error bars are shown.
are consistent with subjects who present clinically with abnormal pronation [30]. The findings from this study show that the STN position can be determined reliably within a session, and when incorporated into a definition of neutral, rearfoot and forefoot kinematics distinguish subjects classified as pronators and normal. Whether the STN position ideally aligns the foot joints requires further investigation of the position of individual joints of the foot (e.g. talonavicular joint). While non-pathologic subjects were able to achieve and maintain STN in this study, the degree to which this method could be implemented is limited to non-fixed deformities of the foot. Other techniques previously mentioned must be implemented for a common reference position between subjects of a fixed-foot population. Further, it is unknown whether the restoration of the arch complex, including the navicular height and first metatarsal declination angle would be a better technique for defining a neutral foot. In this study, the first metatarsal declination angle was accomplished mathematically by performing a 2D correction to the first metatarsal segment in the sagittal plane. Inclusion of the midfoot in this neutral definition may
35
Fig. 5. Comparison of uncorrected (relaxed standing) and corrected (STN) kinematics for first metatarsal dorsiflexion/plantarflexion relative to the rearfoot. Standard error of the mean error bars are shown.
help to further delineate foot kinematics between control subjects and subjects with foot pathologies. In conclusion, the STN position can be reliably applied to subjects classified as having both normal feet and those having abnormal pronation. Applying a defined neutral position that incorporated STN, with both forefoot and rearfoot offsets, resulted in distinctly different kinematic patterns between groups for rearfoot eversion/inversion and first metatarsal dorsiflexion. The technique of establishing a common reference system that is clinically relevant may help to identify kinematic patterns that may be amenable to intervention such as orthotic therapies. STN may be a more effective common reference position than relaxed standing, enabling comparisons across subjects of different foot postures. Further research using methods that are able to precisely define the position of individual foot bones are needed to validate the STN position, and between-session reliability.
Appendix A. Definition of first metatarsal neutral position During relaxed standing and STN, the positions of the first metatarsal base and head were determined from
36
J.R. Houck et al. / Gait & Posture 28 (2008) 29–37
digitized points and transformed into the local coordinate system for each subject, where (xB1, yB1, zB1) = position vector during standing. (xH1, yH1, zH1) = position vector during standing. (xB2, yB2, zB2) = position vector during STN. (xH2, yH2, zH2) = position vector during STN.
of first metatarsal base
of first metatarsal base of first metatarsal head
L2M1 ¼ ðxB2 xH3 Þ2 þ ðyB2 yH3 Þ2 þ ðzB2 zH3 Þ2
(2)
where LM1 = length of first metatarsal during standing. zH1 ¼ zH3
Using the relaxed standing position of the first metatarsal head and its base, u2 about the z-axis was derived by the following equation: xH2 xB2 u2 ¼ arctan (3) yB2 yH2 Similarly, using the new position of the first metatarsal head (xH3, yH3, zH3) and the base of the first metatarsal from the STN position, u3 about the z-axis was derived using the following equation: xH3 xB2 u3 ¼ arctan (4) yB2 yH3 The difference between u2 and u3 (5), reflects the plantarflexion required to rotate the first metatarsal about the z-axis so the metatarsal head rests on the floor with the foot in a STN position. u2 u 3 ¼ uM
There is no conflict of interests of any authors with the presented work in this manuscript.
of first metatarsal head
Because placement into the STN position may result in elevation of the first metatarsal head, the first metatarsal was corrected. The position vectors obtained during a standing trial and STN trial were used to plantarflex the first metatarsal segment so the first metatarsal head would rest on the floor. This was accomplished by rotation of the first metatarsal about its base until the head reached the same vertical and medial/lateral position as during standing. The equations listed below were used to accomplish this.A new metatarsal head position (xH3, yH3, zH3) was calculated from the following equation using the quadratic formula:
yH1 ¼ yH3 ;
Conflict of interest
(5)
where u2 = angle of non-adjusted first metatarsal flexion during STN position. u3 = angle of corrected first metatarsal flexion during STN position. uM = change in first metatarsal flexion due to corrected STN position.
References [1] Carson MC, et al. Kinematic analysis of a multi-segment foot model for research and clinical applications: a repeatability analysis. J Biomech 2001;34(10):1299–307. [2] Hunt AE, et al. Inter-segment foot motion and ground reaction forces over the stance phase of walking. Clin Biomech 2001;16(7):592–600. [3] Kidder SM, et al. A system for the analysis of foot and ankle kinematics during gait. IEEE Trans Rehabil Eng 1996;4(1):25–32. [4] Leardini A, et al. An anatomically based protocol for the description of foot segment kinematics during gait. Clin Biomech 1999;14(8): 528–36. [5] Leardini A, et al. Rear-foot, mid-foot and fore-foot motion during the stance phase of gait. Gait Posture 2007;25(3):453–62. [6] Liu W, et al. Three-dimensional, six-degrees-of-freedom kinematics of the human hindfoot during the stance phase of level walking. Hum Mov Sci 1997;16(2–3):283–98. [7] MacWilliams BA, Cowley M, Nicholson DE. Foot kinematics and kinetics during adolescent gait. Gait Posture 2003;17(3):214–24. [8] Tranberg R, Karlsson D. The relative skin movement of the foot: a 2D roentgen photogrammetry study. Clin Biomech 1998;13(1):71–6. [9] Umberger BR, Nawoczenski DA, Baumhauer JF. Reliability and validity of first metatarsophalangeal joint orientation measured with an electromagnetic tracking device. Clin Biomech (Bristol Avon) 1999;14(1):74–6. [10] Reinschmidt C, et al. Tibiofemoral and tibiocalcaneal motion during walking: external vs. skeletal markers. Gait Posture 1997;6(2):98– 109. [11] Levinger P, Gilleard W. Relationship between static posture and rearfoot motion during walking in patellofemoral pain syndrome: effect of a reference posture for gait analysis. J Am Podiatr Med Assoc 2006;96(4):323–9. [12] Donatelli R, et al. Relationship between static and dynamic foot postures in professional baseball players [see comment]. J Orthop Sports Phys Ther 1999;29(6):316–25. discussion 326–30. [13] McPoil TG, Cornwall MW. Relationship between three static angles of the rearfoot and the pattern of rearfoot motion during walking. J Orthop Sports Phys Ther 1996;23(6):370–5. [14] Sell KE, et al. Two measurement techniques for assessing subtalar joint position: a reliability study. J Orthop Sports Phys Ther 1994;19(3):162–7. [15] Pierrynowski MR, Smith SB. Effect of patient position on the consistency of placing the rearfoot at subtalar neutral. J Am Podiatr Med Assoc 1997;87(9):399–406. [16] Pierrynowski MR, Smith SB, Mlynarczyk JH. Proficiency of foot care specialists to place the rearfoot at subtalar neutral. J Am Podiatr Med Assoc 1996;86(5):217–23. [17] McPoil T, Cornwall MW. Relationship between neutral subtalar joint position and pattern of rearfoot motion during walking. Foot Ankle Int 1994;15(3):141–5. [18] McPoil TG, Cornwall MW. The relationship between static lower extremity measurements and rearfoot motion during walking. J Orthop Sports Phys Ther 1996;24(5):309–14. [19] Boozer MH, Finch A, Waite LR. Investigation of the relationship between arch height and maximum pronation angle during running. Biomed Sci Instrum 2002;38:203–7. [20] Eustace S, et al. Hallux valgus, first metatarsal pronation and collapse of the medial longitudinal arch—a radiological correlation. Skeletal Radiol 1994;23(3):191–4.
J.R. Houck et al. / Gait & Posture 28 (2008) 29–37 [21] Wearing SC, et al. Sagittal plane motion of the human arch during gait: a videofluoroscopic analysis. Foot Ankle Int 1998;19(11):738–42. [22] Mueller MJ, Host JV, Norton BJ. Navicular drop as a composite measure of excessive pronation. J Am Podiatr Med Assoc 1993;83(4):198–202. [23] Lattanza L, Gray G, Kantner R. Closed versus open kinematic chain measurements of subtalar joint eversion: implications for clinical practice. J Orthop Sports Phys Ther 1988;9:310–4. [24] Nawoczenski DA, Ludewig PM. The effect of forefoot and arch posting orthotic designs on first metatarsophalangeal joint kinematics during gait. J Orthop Sports Phys Ther 2004;34(6):317–27. [25] Longworth R, Chockalingam N, Redmond AC. An enhanced protocol to reduce error in electromagnetic tracking of first metatarsophalangeal joint motions. Gait Posture 2006;23(3):391–4.
37
[26] Cole GK, et al. Application of the joint coordinate system to threedimensional joint attitude and movement representation: a standardization proposal. J Biomech Eng 1993;115(4A):344–9. [27] Astrom M, Arvidson T. Alignment and joint motion in the normal foot. J Orthop Sports Phys Ther 1995;22(5):216–22. [28] Chen MJ, et al. Measurement of forefoot varus angle by laser technology in people with flexible flatfoot. Am J Phys Med Rehabil 2003;82(11):842–6. [29] Portney LG, Watknins MP. Foundations of clinical research. Applications to practice, 2nd ed., Upper Saddle River, NJ: Prentice Hall Health; 2000. [30] Tome J, et al. Comparison of foot kinematics between subjects with posterior tibial tendon dysfunction and healthy controls. J Orthop Sports Phys Ther 2006;36(9):635–44.
Gait & Posture 28 (2008) 29–37 www.elsevier.com/locate/gaitpost
Subtalar neutral position as an offset for a kinematic model of the foot during walking Jeff R. Houck *, Josh M. Tome, Deborah A. Nawoczenski Department of Physical Therapy and Center for Foot and Ankle Research, Ithaca College-Rochester Campus, 1100 S. Goodman Street, Rochester, NY 14620, USA Received 19 December 2006; received in revised form 13 September 2007; accepted 18 September 2007
Abstract The lack of a common reference position when defining foot postures may underestimate the ability to differentiate foot function in subjects with pathology. The effect of using the subtalar neutral (STN) position as an offset for both rearfoot and forefoot through comparison of the kinematic walking patterns of subjects classified as normal (n = 7) and abnormally pronated (n = 14) foot postures was completed. An OptotrakTM Motion Analysis System (Northern Digital, Inc.) integrated with Motion Monitor SoftwareTM (Innovative Sports, Inc.) was used to track three-dimensional movement of the leg, rearfoot and first metatarsal segments. Intrarater reliability of positioning the foot into STN using clinical guidelines was determined for a single rater for 21 subjects. Walking data were subsequently compared before and after an offset was applied to the rearfoot and first metatarsal segments. Repeated measures of foot positioning found the STN position to be highly repeatable (intraclass correlation coefficients > 0.9), with peak errors ranging from 1.98 to 4.38. Utilizing STN as the offset resulted in a significant increase in rearfoot eversion ( p = 0.019) during early stance, and greater first metatarsal dorsiflexion ( p < 0.007) across stance in the pronated foot groups that was not observed prior to applying the offset. When applied to subjects with differing foot postures, the selection of a common reference position that is both clinically appropriate and reliable may distinguish kinematic patterns during walking that are consistent with theories of abnormal pronation. # 2007 Elsevier B.V. All rights reserved. Keywords: Foot; Kinematics; Subtalar neutral; Gait
1. Introduction Advances in technology have enabled improved in vivo assessment of foot function across multiple foot segments. Various kinematic models representing multi-segment foot function have been proposed for studies of walking [1–7], however, a number of technical aspects of implementing these foot models remain controversial. Key issues to consider when implementing a kinematic foot model are the potential for skin artifact errors [8–10], the designation of anatomical coordinate systems [3–5], and the definition of a ‘‘neutral’’, or reference position of the foot joints [4,6]. Skin artifact errors arise from movement of the markers relative to the underlying bone or bones. Careful selection of marker * Corresponding author. E-mail address: [email protected] (J.R. Houck). 0966-6362/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2007.09.008
placement on the foot may avoid large skin artifact errors for targeted bones. For the calcaneus and first metatarsal bones, recent studies found skin artifact errors to be minimal, falling within a range of <38 [8–10]. Variability across studies may also be related to the definition of local anatomical coordinate systems. Use of radiographs [3], a special jig [6] or digitized points [4] have been used to improve the identification of local anatomical coordinate systems for subsequent recording of relative motion during walking trials. An additional procedural component to the description of relative motion between anatomic segments is the establishment of a neutral position, or reference position. Various methods have been used to designate the neutral position and include relaxed standing postures, weight-bearing subtalar joint position, alignment to an external frame, alignment of the rearfoot to the tibia, or reference to foot function during the stance period of gait
30
J.R. Houck et al. / Gait & Posture 28 (2008) 29–37
[2,4,6,11]. Using this neutral position, an offset is created that is subsequently applied to the motion data. To date, the optimal approach to define a neutral position of the foot and the application of this offset to subjects with abnormal foot pronation postures have not been established. To reduce noise and make valid comparisons across subjects of different foot postures, a defined neutral position that is both reliable and clinically meaningful is preferred. Clinicians who assess foot posture frequently use an operational definition of neutral termed the ‘subtalar neutral’ (STN) position [12–14]. The operational definition of STN requires clinicians to palpate the talonavicular joint and determine the mid-position, or position of maximum congruency [14]. Even with the degree of nebulousness associated with this definition, well-trained examiners have been able to repeatedly place the rearfoot of healthy subjects in STN position within a few degrees (<38) [15,16]. Investigators have also used the relaxed standing posture as a designated neutral position. Outcomes from two-dimensional (2D) analysis studies have shown mean rearfoot eversion motion to be centered around rearfoot postures observed during relaxed standing posture, and not the STN position [13,17,18]. Interpretation of the results needs to be made with caution, however, because of the limitations associated with 2D analysis, particularly as they apply to out of plane rearfoot rotations at the beginning and end phases of stance. Additionally, because of the large variability in standing postures assumed by subjects, the use of a relaxed standing trial as the neutral position for kinematic models may misrepresent the excursion and endpoints of joint motion that are frequently linked to pathology in pronated foot types. The first purpose of this study was to determine the reliability of rearfoot and forefoot position using a clinical method for establishing a STN position in standing. To improve clinical application, subjects classified as normal and (abnormally) pronated foot postures were included. Five kinematic measurements, frequently implicated in abnormal foot function were evaluated for the first purpose of the study: (1) rearfoot inversion/eversion; (2) dorsiflexion/plantarflexion of the rearfoot relative to the tibia; (3) first metatarsal dorsiflexion/plantarflexion relative to the rearfoot; (4) hallux dorsiflexion/plantarflexion relative to the first metatarsal; and (5) first metatarsal elevation relative to the laboratory coordinate system. The second purpose of this study was to compare two key indicators of foot pronation, rearfoot inversion/eversion and first metatarsal dorsiflexion/plantarflexion [19–21] during walking between subjects classified as normal and pronated foot type using two distinct definitions of the neutral position: relaxed standing and STN. A common reference foot posture is needed to appropriately represent foot function during gait and enable comparisons across subjects of different foot postures. As a clinically relevant position, STN may be more effective as a neutral position than a relaxed standing posture when comparing foot kinematics in subjects with varying foot postures.
Table 1 Subject characteristics
Age (years) Height (cm) Mass (kg) Shoe size a
Normal (n = 7)
Pronator (n = 14)
p-Valuea
22 1 166.2 6.0 66.3 10.9 8.0 0.9
22.2 1.3 167.9 9.0 67.1 15.7 9.4 1.7
0.68 0.62 0.89 0.05
p values indicate result of a Student’s t-test.
2. Methods 2.1. Subjects Twenty-one subjects (seven controls and 14 pronators) consented to participate in this study. A power analysis derived from pilot data of four subjects revealed a sample of 14 subjects, or seven per group was needed to detect differences between pronators and controls with 80% power. To improve clinical relevance, a larger group of subjects classified as pronators were included. With the exception of gender, there were no significant differences between the groups for age, height and mass (Table 1). Of the 21 participants, only three were males. Subjects were considered to be abnormal pronators if they met the inclusion criteria assessed during the clinical exam: a forefoot varus deformity exceeding 108, measured goniometrically in the prone, non-weight-bearing position; rearfoot eversion beyond vertical during weight bearing; and a minimum navicular drop difference of 10 mm or greater from a position of STN to relaxed standing. One or more of these findings have been used as clinical indicators of abnormal pronation in previous investigations [22–24]. The clinical examination was performed by an experienced clinician with greater than 20 years experience (DAN). The study protocol was approved by the Review Board for Human Subjects Research at Ithaca College. Data collection took place in the Movement Analysis Laboratory, Center for Foot and Ankle Research at Ithaca College, Rochester Campus facility. 2.2. Kinematic model The included foot segments were the rearfoot (calcaneus), first metatarsal and hallux. The tibia was also tracked to establish rearfoot position with respect to the tibia. Each segment was tracked by placing three infrared emitting diodes (IREDs) on a thermoplastic molded platform on the skin overlying the calcaneus, first metatarsal, and hallux (Fig. 1). Previous in vitro [9] and in vivo studies [25] reported good repeatability and validity of tracking first metatarsal and hallux segments using skin sensors. Each segment was tracked at 60 Hz using an OptrotrakTM Motion Analysis System (Northern Digital, Inc., CAN). Kinematic data were smoothed using a fourth order, zero phase lag, Butterworth filter with a cut off frequency of 6 Hz. A right-handed Cartesian frame was used to establish coordinate systems for each segment. Anatomical landmarks were digitized on each segment, and defined as follows: Tibia:
Origin: located on the lateral malleolus. y-Axis: joins the lateral malleolus to the fibular head and its positive direction is superior.
J.R. Houck et al. / Gait & Posture 28 (2008) 29–37
31
head of the proximal phalanx with its positive direction superior. The orientation of this axis reflects the anatomical position of the hallux. z-Axis: orthogonal to the x–y plane with its positive direction to the right.
Fig. 1. Infrared emitting diodes (IREDs) on thermoplastic molded platforms placed on the skin overlying the calcaneus, first metatarsal, hallux, and tibia.
x-Axis: orthogonal to the plane defined by the y-axis and medial malleolus with its positive direction anterior. z-axis: orthogonal to the x–y plane with its positive direction to the right. Rearfoot (all points digitized on the floor):
Origin: midpoint of the posterior heel. x-Axis: joins the midpoint of the posterior heel to the tip of the second phalanx and its positive direction points anterior. y-Axis: orthogonal to the plane defined by the x-axis and a point lying on the floor with its positive direction superior. This resulted in a vertical axis aligned with the laboratory coordinate system. z-Axis: orthogonal to the x–y plane with its positive direction to the right. First metatarsal:
Origin: base of the first metatarsal. x-Axis: joins the base of the first metatarsal to the head of the first metatarsal and its positive direction points anterior. y-Axis: orthogonal to the plane defined by the x-axis and another point of equal height off the floor as the first metatarsal head with its positive direction superior. The orientation of this axis reflects the anatomical position of the first metatarsal (metatarsal declination). z-Axis: orthogonal to the x–y plane with its positive direction to the right. Hallux
Origin: base of the proximal phalanx. x-Axis: joins the base of the proximal phalanx to the head of the proximal phalanx and its positive direction points anterior. y-Axis: orthogonal to the plane defined by the x-axis and another point of equal height off the floor as the
Angles were determined using a ZXY sequence as proposed by Cole et al. [26] such that dorsiflexion/plantarflexion represented rotations about the z-axis and inversion/eversion were represented as rotations about the anterior-directed x-axis. The measurements defined for this analysis included the angular orientation of the rearfoot with respect to (w.r.t.) the tibia (dorsiflexion/plantarflexion and eversion/inversion), the first metatarsal w.r.t. the rearfoot (dorsiflexion/plantarflexion), the hallux w.r.t. first metatarsal (dorsiflexion/plantarflexion) and vertical height of the first metatarsal head in the laboratory reference frame. The first metatarsal and hallux angles, as well as the height of the first metatarsal head were included in this analysis because of their presumed relationship to changes in rearfoot alignment. In some cases of abnormal pronation, the medial forefoot and/or metatarsal head may elevate when a subject’s foot is placed in a STN position during the non-weightbearing exam. This is referred to as forefoot varus [27,28]. These same changes may also be observed during standing when the subtalar joint is placed in neutral. Therefore, inclusion of these variables was deemed relevant to the assessment of the STN position. 2.3. Procedures The relaxed standing trial was acquired as subjects placed their feet in a natural, self-selected posture, attempting equal weight bearing on both feet. For the STN trial, a single experienced examiner (DAN) used standard clinical procedures to palpate the STN position. Briefly, the examiner palpated the talonavicular articulation while the subject rolled the rearfoot into inversion and eversion, in effect causing arch elevation and lowering. STN was defined as the mid-position of the talar head relative to the navicular bone. Once the STN position was established by the examiner, subjects were asked to hold that position while a three second trial was collected. The subject then walked in place and the process was repeated a second time. No feedback was given to the examiner or the subject. After completing the standing trials, each subject walked down a walkway at a self-selected natural walking speed. A minimum of five walking trials were collected. A force plate (Kistler Instrument Corp., Amherst, NY, model 9865B) embedded in the floor was used to determine initial contact and toe off phases of stance. 2.4. Definitions of neutral position Two distinct designations of foot postures, established during relaxed standing and STN position were created for the rearfoot and forefoot segments. For the rearfoot, the inversion/eversion position of the calcaneus with respect to the tibia when the foot was placed in the STN position was considered the rearfoot neutral position. A frontal plane offset (uc), calculated as the difference between the STN position and relaxed standing (Fig. 2A and B), was applied to the inversion/eversion patterns of the rearfoot with respect to the tibia. Kinematic patterns without, and with this applied offset are termed uncorrected, and corrected data, respectively and are shown in Fig. 4.
32
J.R. Houck et al. / Gait & Posture 28 (2008) 29–37
Fig. 3. Left foot image of the sagittal plane offset (uM) applied as a rotation around the medial/lateral axis of the first metatarsal segment to account for first metatarsal head elevation [B1 = base of first metatarsal during standing, B2 = base of first metatarsal during subtalar neutral, H1 = head of first metatarsal during standing, H2 = head of first metatarsal during subtalar neutral, H3 = repositioned placement of first metatarsal head to its standing position and uM = angle required to rotate first metatarsal head to its standing position]. Fig. 2. Left foot posterior image of the frontal plane offset (uc), necessary to result in neutral inversion/eversion when the subject was positioned in subtalar neutral, was applied to the inversion/eversion patterns of the rearfoot with respect to the tibia.
degree of correction this would result in a 30–408 shift in plantarflexion/dorsiflexion of the first metatarsal w.r.t. the calcaneus. 2.5. Analysis
For the forefoot, a sagittal plane adjustment to the first metatarsal segment was made to account for potential changes in first metatarsal position that occurred when the foot was placed in the STN (Fig. 3A and B). For subjects with abnormal pronation, it is not uncommon to find elevation of the first metatarsal/medial forefoot, indicating a non-rigid forefoot varus deformity, when the foot is placed in STN [27,28]. Yet, healthy subjects are able to assume the STN position while maintaining contact of the first metatarsal head with the ground. To establish a comparable first metatarsal neutral position across subjects (i.e. the first metatarsal head is on the floor while the mid-foot is manipulated into the STN position), an offset (uM) was applied that plantarflexed the first metatarsal so the head would rest on the floor (Fig. 3A and B and Appendix A). Once the offset was applied to the first metatarsal, the corrected data was calculated by noting the plantarflexion angle of the first metatarsal w.r.t. the calcaneus in the STN position and subtracting this value from the uncorrected data. Depending on the
An intraclass correlation coefficient (ICC) was used to evaluate reliability of the five inter-segmental kinematic measures as subjects were placed in STN. ICC, standard error of the measurement (S.E.M.), and absolute error data are provided for repeated measurements on these variables by the same tester. The ICC, model 3, identified by Portney and Watknins as appropriate for intrarater reliability was used in this study [29]. To estimate the expected error in degrees, two approaches were used including calculating the percent agreement and the standard error of the measurement (S.E.M.). Absolute errors were the absolute values of the difference between trial one and trial two for each subject. In previous studies errors in rearfoot inversion/eversion of 28 were clinically desirable [15,16]. Therefore the percentage of trials resulting in differences of 18 or less and 28 or less were reported. The standard error of the measurement assumes a normal distribution of errors and therefore is a metric that is useful for understanding what errors are expected across subjects, where errors are expected to stay within 1 S.E.M.
J.R. Houck et al. / Gait & Posture 28 (2008) 29–37
68% of the time [29]. The calculation of the S.E.M. is included in the following equation: pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi standard error of the measurement ¼ S:D: 1 ICC value
33
time. Actual errors were below 28 or 2 mm 81% of the time for all variables.
(1)
3.2. Walking where S.D. is the pooled standard deviation of trial one and trial two and ICC value is the correlation coefficient from the ICC statistic. For the second purpose of the study, kinematic data from each walking trial were interpolated to 1% increments of stance and then averaged across five trials to gain a representative pattern for each variable for each subject. First, the data were processed without applying an offset (uncorrected). Second, the definitions of neutral for the rearfoot and first metatarsal using the STN position were applied to the data (corrected). Defined points across stance (initial contact and peaks) were selected to compare patterns statistically before, and after applying the STN definition of neutral. To assure the same point of stance were used in the analysis between groups, peaks were first determined from the patterns of the pronator group, and then the same point of stance of the normal group was used. For rearfoot eversion/inversion the stance points used were initial contact, peak eversion (28% of stance) and peak inversion (96% of stance). For first metatarsal dorsiflexion/plantarflexion the stand points used were initial contact, peak dorsiflexion (73% of stance) and peak plantarflexion (100% of stance). A two-way, mixed linear model which included a fixed factor of groups (normal versus pronator) and repeated factor of stance point were used to compare patterns. Due to the imbalance in sample sizes maximum likelihood estimation was used. In the presence of group stance point interaction, pairwise comparisons were conducted. For each twoway ANOVA and pairwise comparison an alpha level of 0.05 was utilized to indicate significance. All statistical procedures were completed using SAS version 9.1 software.
3. Results 3.1. Reliability The five measures of foot position during the first and second STN position were highly repeatable (Table 2). The ICC values were greater than 0.9 for all variables. The peak errors ranged from 1.98 to 4.38. The two S.E.M. statistic predicted repeatability errors would be less than 1.4–2.48 96% of the time. Peak errors associated with first metatarsal elevation were as high as 5.3 mm. The two S.E.M. statistic predicted that errors would fall within 2.4 mm 96% of the
The stance points used in the analysis for rearfoot eversion/inversion and first metatarsal dorsiflexion/plantarflexion distinguished the pronator group for both the corrected and uncorrected data (Table 3 and Fig. 4). A significant interaction (group stance point) occurred for both uncorrected ( p = 0.034) and corrected ( p = 0.034) rearfoot eversion/inversion. The pairwise comparisons of the uncorrected rearfoot eversion/inversion data suggested the pronated group was significantly more inverted then the normal group at initial contact (2.1 4.08 vs 2.8 4.08, p = 0.016) and 96% of stance (6.4 4.98 vs 2.0 3.98, p = 0.05). In contrast the corrected rearfoot eversion/ inversion patterns suggested significantly greater rearfoot eversion at 28% of stance ( p = 0.019) for the pronated group (7.5 3.38) compared to those in the normal group (4.0 2.08). The stance points used in the analysis of first metatarsal dorsiflexion/plantarflexion suggested a significant main effect of greater dorsiflexion in the corrected data but not the uncorrected data (Table 3 and Fig. 5). The uncorrected first metatarsal dorsiflexion/plantarflexion data was not significantly different across stance between the pronator and normal groups (normal group = 30.4 5.68 versus pronator group = 28.7 5.28; p = 0.51). In contrast, the corrected first metatarsal dorsiflexion/plantarflexion data across stance (group main effect) indicated significantly greater dorsiflexion ( p = 0.007) of the pronator group (4.6 4.58) compared to the healthy group (1.6 4.58). There was a trend toward a significant interaction ( p = 0.07) in both the corrected and uncorrected first metatarsal dorsiflexion/plantarflexion data.
4. Discussion The findings of this study show the STN position to be highly repeatable within an examiner, and the position may be reliably applied to a clinical population with foot postures
Table 2 Summary of reliability statistics for subtalar neutral position Variable
ICC (95% CI)
S.E.M.
Absolute error (range)
<18 (%)
<28 (%)
Rearfoot Calcaneus eversion Calcaneus dorsiflexion
0.96 (0.89–0.98) 0.94 (0.87–0.98)
1.08 0.78
0.9 0.98 (0.1–4.38) 0.9 0.58 (0.1–1.98)
48 67
90 100
Forefoot First metatarsal elevation First metatarsal dorsiflexion Hallux dorsiflexion
0.96 (0.92–0.99) 0.95 (0.88–0.98) 0.95 (0.89–0.98)
1.2 mm 1.08 1.28
0.7 0.98 (0.01–5.2 mm) 1.1 0.88 (0.01–2.98) 1.3 1.18 (0.2–3.88)
57 a 67 52
a
The percentage below 1 mm or 2 mm, not degrees, is reported in each respective column.
81a 90 81
34
J.R. Houck et al. / Gait & Posture 28 (2008) 29–37
Table 3 Mean (1 S.D.) for rearfoot and forefoot data Normal
Pronator
D
ANOVA
Calcaneal eversion Uncorrected Initial contact Peak () eversion Peak (+) inversion
2.8 4.0 7.3 3.1 2.0 3.9
2.1 4.0* 6.0 4.4 6.4 4.9*
4.9 1.3 4.4
Group = 0.057 Group stance point = 0.034
Corrected Initial contact Peak () eversion Peak (+) inversion
0.4 2.6 4.0 2.0 5.2 1.2
0.5 2.8 7.5 3.3* 4.8 1.8
0.1 3.5 0.4
Group = 0.138 Group stance point = 0.034
2.0 4.5 1.5
Group = 0.512 Group stance point = 0.069
6.6 9.1 4.1
Group = 0.007 Group stance point = 0.069
First metatarsal dorsiflexion Uncorrected Initial contact Peak (+) dorsiflexion Peak () plantarflexion
30.0 5.5 21.3 3.2 39.8 4.2
Corrected Initial contact Peak (+) dorsiflexion Peak () plantarflexion
1.3 5.3 7.4 4.8 11.1 3.6
*
28.0 5.8 16.8 6.1 41.3 8.6 5.3 4.6* 16.5 4.9* 7.0 7.3
p values of less than 0.05 indicated by pairwise comparisons.
associated with abnormal pronation. Unique to this study was the inclusion of the medial forefoot (i.e., first metatarsal segment) in the determination of the neutral position. Prior investigations have not considered the changes in the medial forefoot that may occur with STN positioning, nor the implications of a forefoot offset when assessing movement patterns. The high reliability of both forefoot and rearfoot variables supported the inclusion of both segments when defining the neutral position. Only when applying these neutral offsets (derived from the STN position) did rearfoot eversion and first metatarsal dorsiflexion show strong statistical differences in foot function consistent with clinical theories of pronation. These data support the current practice of using the STN neutral position as a clinical metric for application to kinematic foot models. The data from this study suggest the STN position is a reliable method for estimating a neutral position of the foot. Previous studies examining intrarater reliability suggested clinicians were able to place the rearfoot within 38, 90% of the time [15,16]. Our data is more optimistic suggesting experienced clinicians are able to place the rearfoot within 28, 90% of the time. Unlike previous studies using threedimensional techniques, this study included subjects classified as abnormal pronators based on clinical exam. Difficulties positioning subjects with abnormal foot postures were expected to increase errors relative to other studies. Surprisingly, the errors noted in this study were slightly lower for standing STN, but similar to values obtained in the prone and seated STN positions of healthy subjects for a single examiner [15]. The S.E.M. values suggest that using the STN position to define a reference position is suspect to re-positioning errors of 2–38 for the rearfoot.
Not addressed in previous studies was the reliability of positioning the forefoot. It is not uncommon to find forefoot deviations, frequently reported as forefoot varus, in persons with abnormal pronation during the non-weight-bearing clinical examination. These same deviations may also be observed during the weight-bearing exam when the foot is positioned in STN and the first metatarsal is observed to elevate. To date, there have been no studies that have assessed forefoot corrections in a clinically pronated foot group. Similar to S.E.M. values reported for the rearfoot, repositioning errors of 2–38 were also found for the first metatarsal kinematics. These reliability errors (2–38) are smaller than the expected changes in kinematics associated with many foot and ankle pathologies. Therefore, using the STN position as a neutral position is a useful method for accounting for differences in foot posture. Applying the STN position as a definition of neutral position to the rearfoot and first metatarsal kinematic patterns significantly influenced the interpretation of foot function. The uncorrected data for rearfoot eversion/ inversion of the normal subjects suggested higher rearfoot inversion of the pronator group during initial contact and late stance (Fig. 4), and no differences during early stance—data inconsistent with clinical observations. Only when the data were corrected using the STN position as neutral did the pronator group show greater rearfoot eversion during early stance. Similarly, differences in peak first metatarsal dorsiflexion between the normal and pronator group were amplified when applying the definition of neutral that incorporated the STN position (Fig. 5). Increased first metatarsal dorsiflexion reflects a lowering of the anterior component of the medial longitudinal arch. Again, the findings of a lower arch
J.R. Houck et al. / Gait & Posture 28 (2008) 29–37
Fig. 4. Comparison of uncorrected (relaxed standing) and corrected (STN) kinematics for rearfoot eversion/inversion relative to the tibia. Standard error of the mean error bars are shown.
are consistent with subjects who present clinically with abnormal pronation [30]. The findings from this study show that the STN position can be determined reliably within a session, and when incorporated into a definition of neutral, rearfoot and forefoot kinematics distinguish subjects classified as pronators and normal. Whether the STN position ideally aligns the foot joints requires further investigation of the position of individual joints of the foot (e.g. talonavicular joint). While non-pathologic subjects were able to achieve and maintain STN in this study, the degree to which this method could be implemented is limited to non-fixed deformities of the foot. Other techniques previously mentioned must be implemented for a common reference position between subjects of a fixed-foot population. Further, it is unknown whether the restoration of the arch complex, including the navicular height and first metatarsal declination angle would be a better technique for defining a neutral foot. In this study, the first metatarsal declination angle was accomplished mathematically by performing a 2D correction to the first metatarsal segment in the sagittal plane. Inclusion of the midfoot in this neutral definition may
35
Fig. 5. Comparison of uncorrected (relaxed standing) and corrected (STN) kinematics for first metatarsal dorsiflexion/plantarflexion relative to the rearfoot. Standard error of the mean error bars are shown.
help to further delineate foot kinematics between control subjects and subjects with foot pathologies. In conclusion, the STN position can be reliably applied to subjects classified as having both normal feet and those having abnormal pronation. Applying a defined neutral position that incorporated STN, with both forefoot and rearfoot offsets, resulted in distinctly different kinematic patterns between groups for rearfoot eversion/inversion and first metatarsal dorsiflexion. The technique of establishing a common reference system that is clinically relevant may help to identify kinematic patterns that may be amenable to intervention such as orthotic therapies. STN may be a more effective common reference position than relaxed standing, enabling comparisons across subjects of different foot postures. Further research using methods that are able to precisely define the position of individual foot bones are needed to validate the STN position, and between-session reliability.
Appendix A. Definition of first metatarsal neutral position During relaxed standing and STN, the positions of the first metatarsal base and head were determined from
36
J.R. Houck et al. / Gait & Posture 28 (2008) 29–37
digitized points and transformed into the local coordinate system for each subject, where (xB1, yB1, zB1) = position vector during standing. (xH1, yH1, zH1) = position vector during standing. (xB2, yB2, zB2) = position vector during STN. (xH2, yH2, zH2) = position vector during STN.
of first metatarsal base
of first metatarsal base of first metatarsal head
L2M1 ¼ ðxB2 xH3 Þ2 þ ðyB2 yH3 Þ2 þ ðzB2 zH3 Þ2
(2)
where LM1 = length of first metatarsal during standing. zH1 ¼ zH3
Using the relaxed standing position of the first metatarsal head and its base, u2 about the z-axis was derived by the following equation: xH2 xB2 u2 ¼ arctan (3) yB2 yH2 Similarly, using the new position of the first metatarsal head (xH3, yH3, zH3) and the base of the first metatarsal from the STN position, u3 about the z-axis was derived using the following equation: xH3 xB2 u3 ¼ arctan (4) yB2 yH3 The difference between u2 and u3 (5), reflects the plantarflexion required to rotate the first metatarsal about the z-axis so the metatarsal head rests on the floor with the foot in a STN position. u2 u 3 ¼ uM
There is no conflict of interests of any authors with the presented work in this manuscript.
of first metatarsal head
Because placement into the STN position may result in elevation of the first metatarsal head, the first metatarsal was corrected. The position vectors obtained during a standing trial and STN trial were used to plantarflex the first metatarsal segment so the first metatarsal head would rest on the floor. This was accomplished by rotation of the first metatarsal about its base until the head reached the same vertical and medial/lateral position as during standing. The equations listed below were used to accomplish this.A new metatarsal head position (xH3, yH3, zH3) was calculated from the following equation using the quadratic formula:
yH1 ¼ yH3 ;
Conflict of interest
(5)
where u2 = angle of non-adjusted first metatarsal flexion during STN position. u3 = angle of corrected first metatarsal flexion during STN position. uM = change in first metatarsal flexion due to corrected STN position.
References [1] Carson MC, et al. Kinematic analysis of a multi-segment foot model for research and clinical applications: a repeatability analysis. J Biomech 2001;34(10):1299–307. [2] Hunt AE, et al. Inter-segment foot motion and ground reaction forces over the stance phase of walking. Clin Biomech 2001;16(7):592–600. [3] Kidder SM, et al. A system for the analysis of foot and ankle kinematics during gait. IEEE Trans Rehabil Eng 1996;4(1):25–32. [4] Leardini A, et al. An anatomically based protocol for the description of foot segment kinematics during gait. Clin Biomech 1999;14(8): 528–36. [5] Leardini A, et al. Rear-foot, mid-foot and fore-foot motion during the stance phase of gait. Gait Posture 2007;25(3):453–62. [6] Liu W, et al. Three-dimensional, six-degrees-of-freedom kinematics of the human hindfoot during the stance phase of level walking. Hum Mov Sci 1997;16(2–3):283–98. [7] MacWilliams BA, Cowley M, Nicholson DE. Foot kinematics and kinetics during adolescent gait. Gait Posture 2003;17(3):214–24. [8] Tranberg R, Karlsson D. The relative skin movement of the foot: a 2D roentgen photogrammetry study. Clin Biomech 1998;13(1):71–6. [9] Umberger BR, Nawoczenski DA, Baumhauer JF. Reliability and validity of first metatarsophalangeal joint orientation measured with an electromagnetic tracking device. Clin Biomech (Bristol Avon) 1999;14(1):74–6. [10] Reinschmidt C, et al. Tibiofemoral and tibiocalcaneal motion during walking: external vs. skeletal markers. Gait Posture 1997;6(2):98– 109. [11] Levinger P, Gilleard W. Relationship between static posture and rearfoot motion during walking in patellofemoral pain syndrome: effect of a reference posture for gait analysis. J Am Podiatr Med Assoc 2006;96(4):323–9. [12] Donatelli R, et al. Relationship between static and dynamic foot postures in professional baseball players [see comment]. J Orthop Sports Phys Ther 1999;29(6):316–25. discussion 326–30. [13] McPoil TG, Cornwall MW. Relationship between three static angles of the rearfoot and the pattern of rearfoot motion during walking. J Orthop Sports Phys Ther 1996;23(6):370–5. [14] Sell KE, et al. Two measurement techniques for assessing subtalar joint position: a reliability study. J Orthop Sports Phys Ther 1994;19(3):162–7. [15] Pierrynowski MR, Smith SB. Effect of patient position on the consistency of placing the rearfoot at subtalar neutral. J Am Podiatr Med Assoc 1997;87(9):399–406. [16] Pierrynowski MR, Smith SB, Mlynarczyk JH. Proficiency of foot care specialists to place the rearfoot at subtalar neutral. J Am Podiatr Med Assoc 1996;86(5):217–23. [17] McPoil T, Cornwall MW. Relationship between neutral subtalar joint position and pattern of rearfoot motion during walking. Foot Ankle Int 1994;15(3):141–5. [18] McPoil TG, Cornwall MW. The relationship between static lower extremity measurements and rearfoot motion during walking. J Orthop Sports Phys Ther 1996;24(5):309–14. [19] Boozer MH, Finch A, Waite LR. Investigation of the relationship between arch height and maximum pronation angle during running. Biomed Sci Instrum 2002;38:203–7. [20] Eustace S, et al. Hallux valgus, first metatarsal pronation and collapse of the medial longitudinal arch—a radiological correlation. Skeletal Radiol 1994;23(3):191–4.
J.R. Houck et al. / Gait & Posture 28 (2008) 29–37 [21] Wearing SC, et al. Sagittal plane motion of the human arch during gait: a videofluoroscopic analysis. Foot Ankle Int 1998;19(11):738–42. [22] Mueller MJ, Host JV, Norton BJ. Navicular drop as a composite measure of excessive pronation. J Am Podiatr Med Assoc 1993;83(4):198–202. [23] Lattanza L, Gray G, Kantner R. Closed versus open kinematic chain measurements of subtalar joint eversion: implications for clinical practice. J Orthop Sports Phys Ther 1988;9:310–4. [24] Nawoczenski DA, Ludewig PM. The effect of forefoot and arch posting orthotic designs on first metatarsophalangeal joint kinematics during gait. J Orthop Sports Phys Ther 2004;34(6):317–27. [25] Longworth R, Chockalingam N, Redmond AC. An enhanced protocol to reduce error in electromagnetic tracking of first metatarsophalangeal joint motions. Gait Posture 2006;23(3):391–4.
37
[26] Cole GK, et al. Application of the joint coordinate system to threedimensional joint attitude and movement representation: a standardization proposal. J Biomech Eng 1993;115(4A):344–9. [27] Astrom M, Arvidson T. Alignment and joint motion in the normal foot. J Orthop Sports Phys Ther 1995;22(5):216–22. [28] Chen MJ, et al. Measurement of forefoot varus angle by laser technology in people with flexible flatfoot. Am J Phys Med Rehabil 2003;82(11):842–6. [29] Portney LG, Watknins MP. Foundations of clinical research. Applications to practice, 2nd ed., Upper Saddle River, NJ: Prentice Hall Health; 2000. [30] Tome J, et al. Comparison of foot kinematics between subjects with posterior tibial tendon dysfunction and healthy controls. J Orthop Sports Phys Ther 2006;36(9):635–44.