Forefoot–rearfoot coupling patterns and tibial internal rotation during stance phase of barefoot versus shod running

1. Introduction 

Excessive tibial internal rotation coupling with rearfoot eversion during the first half stance phase of running was associated with patella-femoral pain syndrome, Achilles tendon pain and shin splint (Clement et al., 1981, Smart et al., 1980, Tiberio, 1987, Viitasalo and Kvist, 1983). The amount of internal tibial rotation is proposed to be related to coupling motion patterns between the forefoot and rearfoot (Lundberg, 1989, Naster et al., 2002). A twisted plate model of the foot suggests that the forefoot produces counter motions with respect to the rearfoot segments during barefoot running (Hunt et al., 2001, Sarraffian, 1993). From heel-strike through footflat, the rearfoot is everted and the forefoot becomes flexible to absorb shock and adapt itself to irregularities in the ground floor surface (Nordin and Frankle, 2001). A cross-correlation between the rearfoot and forefoot motion indicated that rearfoot eversion/inversion was highly correlated to forefoot plantar/dorsiflexion (r<−0.85) and abduction/adduction (r>0.94) with no phase shift during the stance phase of barefoot running (Pohl et al., 2006). Johanson et al. (1994) reported that a large forefoot inversion with respect to the rearfoot results in an abnormal gait pattern when resulting in compensatory subtalar joint pronation. Furthermore, using a mitered hinge model, rearfoot eversion in the frontal plane was found to be coupled with tibial internal rotation during gait (Pohl et al., 2006, Nigg et al., 1993). A high correlation value (r=0.99) was reported between rearfoot eversion and tibial internal rotation during the first 50% stance phase of gait (Pohl et al., 2006, Nigg et al., 1993). Therefore, based on the twisted plate and mitered hinge models, the forefoot and rearfoot coupling motion patterns could contribute to the amount of tibial rotation.

In previous studies the rearfoot and tibia coupling motion was modelled as a single rigid segment because of technical difficulties associated with evaluating the forefoot motion in a shoe condition. Furthermore, in vivo studies on the forefoot motions, subjects were tested in barefoot condition to enable tracking of markers on the forefoot (Pohl et al., 2006, Hunt et al., 2001). Therefore, footwear effects on the three-dimensional forefoot motion coupling with the rearfoot frontal plane motion and their contributions to the tibial rotation remained unknown.

The use of forefoot posting in orthotic interventions to compensate excessive foot pronation is still misunderstood. Clinically, it is believed that abnormal foot pronation is associated with forefoot excessive motions with respect to the rearfoot (Johanson et al., 1994, Tillman et al., 2003). However, Johanson et al. (1994) indicated that posting in the rearfoot was more effective in controlling foot pronation than posting in the forefoot, even in the presence of a forefoot deformity. A better understanding of the forefoot and rearfoot coupling relationships and their contributions to the tibial rotation in asymptomatic feet will provide information of the importance of forefoot posting in the orthotic interventions in controlling excessive tibial rotation.

A number of techniques have been used to examine coupling motion relationships between rearfoot and tibia during dynamic motions. Cross-correlations are based on the assumption that linear relationships exist between two adjacent segments. However, this technique is not useful in determining the degree of linkage between the segments that have a non-linear relationship (Sideway et al., 1995). Rearfoot eversion and tibial internal rotation (EV/TIR) excursion ratio is used to provide a measure of the relative motion between the rearfoot and tibia from heel-strike to the respective peaks around midstance (DeLeo et al., 2004). In the recent studies, the EV/TIR ratio varied between 0.65 in the normal shod (Stacoff et al., 2000) and 2.40 in the barefoot conditions (Pohl et al., 2006). These values suggest that the rearfoot is everted by 1° for every 1.54° and 0.41° tibial internal rotation in shod and barefoot conditions, respectively. In the present study, EV/TIR excursion ratio will be used to determine if the tibia has a relatively greater motion with respect to the rearfoot (Nawoczenski et al., 1995, Nigg et al., 1993, Williams et al., 2001). For example, runners with lower EV/TIR ratios display relatively more tibial internal rotation with respect to the rearfoot eversion rotation and increasing the risk for knee related injuries (McClay and Manal, 1997, Williams et al., 2001). A continuous relative phase angle technique (CRP) was also proposed to describe the coupling motion relationships of two adjacent segments throughout the stance phase (Hamill et al., 1999). This technique indicates the amount of in-phase or out-of-phase relationships between two adjacent segments. Hamill et al. (1999) reported that the relationship between the rearfoot and tibia was more out-of-phase in the strike phase than the rest of stance in a group of healthy runners. However, there is no information regarding to coupling motion patterns of the forefoot and rearfoot during shod running in the literature. Thus, relative phase angle technique will be used to provide quantitative information on the forefoot–rearfoot coupling motion patterns throughout the stance phase of barefoot running versus running with sandals.

With respect to the following three assumptions, sandals were used as footwears in the present study. Firstly, the sandals’ adjustable straps and the bottom midsole designs enable greater changes in the forefoot and rearfoot coupling motion patterns than running shoe. Secondly, the sandal allows tracking of the rearfoot and forefoot surface markers during running trials. Finally, sandals are often used to evaluate the effects of foot orthoses on the rearfoot and tibia coupling motions (Branthwaite et al., 2004, Nawoczenski et al., 1995). However, the confounding effects of the sandal on the outcome measures of these coupling motions were unknown in the literature.

In current study, we hypothesized that tibial internal rotation is increased when the forefoot–rearfoot coupling patterns are modified to a more in-phase relationship with shoe wears during the stance phase of running. The purposes were: (i) to compare the excursion of tibial internal rotation and rearfoot eversion from heel-strike to peak value during the stance phase of running in barefoot versus shod conditions, (ii) to determine differences in mean relative phase angle of the forefoot eversion/inversion and rearfoot eversion/inversion (FFev/in–RFev/in), forefoot dorsi/plantarflexion and rearfoot eversion/inversion (FFd/p–RFev/in), forefoot adduction/abduction and rearfoot eversion/inversion (FFad/ab–RFev/in) during the stance phase of barefoot versus shod running.

2. Methods 

Sixteen able-bodied healthy men having an average age of 28.2 (SD 5.2 years), weight of 82.3 (SD 10.4kg) and height of 179.0 (SD 5.4cm) volunteered. The experimentation procedures were explained to all participants and those who volunteered signed an informed consent form approved by the Hospital Ethics Committee.

2.1. Experimental set-up 

Six cameras (Motion Analysis Corporation, Santa Rosa, USA) were arranged along two arcs on the left and right sides of a force plate (960Hz, AMTI, Model OR6-5) placed in the middle of a 10m runway. The capture volume (0.5×0.5×0.75m3) covered the lower limb motion. Video data were collected using the EVa 4.2 software (Motion Analysis Corporation, Santa Rosa, USA) at 60Hz. Accuracy of the spatial reconstruction was assessed by means of an artificial foot (prosthesis) where markers corresponded to the forefoot and rearfoot. The average angular standard deviation was found about 1.5° in fast motions.

Forefoot, rearfoot and tibia were modelled as three rigid segments. The motion of the forefoot with respect to the rearfoot was defined using Kidder’s et al. (1996) model and the rearfoot motion with respect to the tibia complied with the Joint Coordinate System recommendation (Wu et al., 2002). These joint representations accounted for the functional anatomy of the foot and allowed the greater kinematic analysis than previous simpler models.

Fourteen reflective skin markers (16mm diameter) were attached to the right foot and shank. Of these, eleven markers were fixed on predefined anatomical landmarkers to define the forefoot, rearfoot and tibia coordinate systems as shown in (Fig. 1(a) and (b)) and describe as:

•Forefoot: medial side of the fifth metatarsal head (M5MH), lateral side of the first metatarsal base (L1MB) and head (L1MH), anterior part of the second toe (TOE).

•Rearfoot: posterior calcaneus (POSTC), medial calcaneus (MEDC), lateral calcaneus (LATC).

•Tibia: tibial tubercle (TIBT), head of the fibula (HFIB), medial malleolus (MEDM), lateral malleous (LATM).

View Large Image Download to PowerPoint

Fig. 1. Anterior (a) and posterior (b) views of marker placements in barefoot condition. Lateral view in shod condition (c).

Three technical markers were also placed on the anterior middle aspect of the tibia (ANTT), fifth metatarsal base (M5MB) and middle part of between the second and third metatarsals (M23M). To avoid from marker dropout, skin movement artefact and hidden markers during running trials, the markers at the tibial tubercle, medial malleolus, medial side of fifth metatarsal head, lateral side of the first metatarsal head were removed and calculated as virtual markers after recording a barefoot neutral standing position. The technical markers were used to define the coordination of the virtual markers during running trials.

Three-dimensional joint rotations were calculated using Euler angles. The sequence of rotations was first plantar/dorsiflexion about a fixed media-lateral axis of the proximal segment, abduction/adduction about the floating axis, then inversion/eversion about the anterior–posterior axis of the distal segment. The tibial internal/external rotation corresponded to the rearfoot abduction/adduction motion (Grood and Suntay, 1983). All the kinematic parameters were calculated using a set of programs written in Matlab 7.0.4 from the three-dimensional coordinates previously filtered at 8Hz with a low-pass zero phase shift fourth-order Butterworth filter.

3. Results 

Fig. 2 indicates that the rearfoot eversion and tibial internal rotation occurred from heel-strike to about midstance then the rearfoot inverts and the tibia rotates externally from midstance to toe-off. A similar pattern of rearfoot and tibial rotation coupling motions was observed from heel-strike to toe-off in the barefoot and shod running conditions. The steeper slope of the mean curve indicates a higher rearfoot frontal plane motion than the tibial rotation during stance phase. The EV/TR excursion ratio in the barefoot and shod running was 1.80 and 2.24, respectively (P>0.05). This finding shows that the rearfoot is everted by 1° for every 0.55° and 0.44° tibial internal rotation in the barefoot and shod conditions during the stance phase of running, respectively.

View Large Image Download to PowerPoint

Fig. 2. Angle–angle plot for rearfoot eversion/inversion versus tibial internal/external rotation from heel-strike (HS) to toe-off (TO) in the shod and barefoot conditions.

Table 1 illustrates the mean and standard deviation of the tibial internal rotation and rearfoot eversion in barefoot and shod conditions. Rearfoot eversion excursion increased by 2% while tibial internal rotation excursion decreased by 1.6% in the shod condition when compared to barefoot condition. However, these changes were minimal and statistically insignificant (P>0.05).

Table 1. Mean values (SD) of the excursion of rearfoot eversion and tibial internal rotation (°) in barefoot and shod conditions

	Variables	Barefoot	Shod	P-values
	Rearfoot eversion excursion	−8.8 (2.3)	−9.0 (4.1)	0.79
	Tibial internal rotation excursion	4.1 (2.0)	4.0 (2.0)	0.89
	Eversion/tibial internal rotation ratio	−1.8 (2.0)	−2.2 (1.1)	0.44

P-values compared barefoot and shod conditions.

Table 2, Table 3 present the mean absolute relative phase angles of FFev/in–RFev/in and FFd/p–RFev/in for each interval of the stance phase in the shod and barefoot conditions. No statistical differences were observed in the relative phase angles of FFev/in–RFev/in and FFd/p–RFev/in between shod and barefoot conditions for any intervals of the stance phase (P>0.05).

Table 2. Mean (SD) of the forefoot eversion/inversion and rearfoot eversion/inversion absolute relative angle (°) across the five intervals of stance phase in barefoot versus shod running

	Intervals	Barefoot	Shod	P-values
	Heel-strike	41.3 (30.1)	31.2 (27.4)	0.28
	Foot-flat	25.1 (22.7)	34.4 (23.3)	0.24
	Heel-rise	20.9 (15.3)	25.0 (19.1)	0.48
	Push-off	31.4 (21.3)	30.3 (28.4)	0.88
	Toe-off	27.9 (29.9)	23.8 (32.0)	0.70

P-values compared barefoot and shod conditions.

Table 3. Mean (SD) of forefoot dorsi/plantarflexion and rearfoot eversion/inversion absolute relative angle (°) across the five intervals of stance phase in barefoot versus shod running

	Intervals	Barefoot	Shod	P-values
	Heel-strike	43.8 (32.9)	41.3 (19.5)	0.84
	Foot-flat	21.7 (11.7)	27.5 (18.9)	0.42
	Heel-rise	18.1 (12.8)	22.1 (15.6)	0.55
	Push-off	29.4 (21.9)	34.3 (24.2)	0.55
	Toe-off	43.6 (18.4)	38.6 (18.7)	0.60

P-values compared barefoot and shod conditions.

Statistical analysis for the absolute relative phase angle of FFad/ab–RFev/in showed an interaction effect of the intervals of the stance phase and conditions (P<0.01). Effect size estimation indicated that the intervals of the stance phase contribute to 47% of the total variance (more important factor). In the barefoot condition, the relative phase angle was by 50° and 53° higher and in out-of-phase in the heel-strike compared to the foot-flat (P<0.01) and heel-rise (P<0.01), respectively. Furthermore, a statistically higher (by −22°) in-phase relationship was observed in the heel-rise phase than toe-off (P<0.05).Whereas in the shod condition, significant differences among the intervals were observed between foot-flat (20.7°) and toe-off (41.0°), (P<0.05) as well as between heel-rise (16.0°) and push-off periods (42.1°), (P<0.05), (Table 4). These findings indicate that FFad/ab–RFev/in coupling motion have a more out-of-phase relationship during heel-strike compared to latter intervals during barefoot running. Contrary, a higher out-of-phase relationship (by 26°) of the FFad/ab–RFev/in coupling motion was observed in push-off and toe-off compared to heel-rise during the shod running. Furthermore, a statistical difference between the shod and barefoot conditions was observed in heel-strike (P=0.01). This difference was higher by 37° in the out-of-phase relationship in the barefoot compared to the shod condition. This finding shows that the out-of-phase relationship of FFad/ab–RFev/in at heel-strike in the barefoot condition is modified to a more in-phase relationship with the sandal.

Table 4. Mean (SD) of forefoot adduction/abduction and rearfoot eversion/inversion absolute relative angle (°) across the five intervals in barefoot versus shod running

	Intervals	Barefoot	Shod	P-values
	Heel-strike	71.5 (45.4)	34.5 (28.2)	0.01⁎
	Foot-flat	21.2 (18.2)Heel strike	20.7 (15.7)	0.93
	Heel-rise	17.9 (14.5)Heel strike	16.0 (13.5)	0.67
	Push-off	33.2 (21.1)	42.1(23.7)Heel rise	0.32
	Toe-off	39.9 (27.1)Heel rise	41.0 (25.1)Foot flat	0.91

Significant differences between intervals in each condition are shown by superscript (P<0.05).

⁎ P-values compared barefoot and shod conditions.

4. Discussion 

The first purpose of this study was to compare the excursion of rearfoot eversion and tibial internal rotation from heel-strike to the peak value during the first half stance phase in barefoot versus shod running. The findings showed an insignificant change in the rearfoot eversion excursion and tibial internal rotation by using the sandals when compared to the barefoot running. This is consistent with the results obtained by comparing normal shod and barefoot running via the direct measurement of the markers mounted on bones (Stacoff et al., 2000). However, Stacoff et al. (1991) showed differences in the rearfoot and tibia coupling patterns using skin and shoe mounted markers between barefoot and normal shod running. The functional frontal plane subtalar joint motion and tibial rotation during barefoot running was reported to vary from 8° to 15° and from 3° to 6°, respectively (McClay and Manal, 1997, Pohl et al., 2006, Stacoff et al., 2000). In the present study, the average range of the frontal plane rearfoot motion was 11.4° (SD 4.3) in barefoot and 10.9° (SD 4.9) in the shod condition. Furthermore, the average range of tibial rotation was 5.2° (SD 2.4) and 5.3° (SD 2.9) in the barefoot and shod conditions in healthy runners, respectively. This finding shows that the rearfoot eversion and tibial internal rotation excursions varied in the range reported in the previous studies during the selected running speed. This variation in the mean values may due to different experimental protocols and foot joint models utilized. It is suggested that the effect of normal shod on the rearfoot and tibia coupling motion could be observed when different type of feet are selected and tested during higher running speeds or when cutting movements are performed (Stacoff et al., 2000).

Second purpose of this study was to determine differences in the mean relative phase angle of the forefoot and rearfoot during the five intervals of stance phase between barefoot and shod running. No statistical difference was noted in the mean relative phase relationship of FFev/in–RFev/in during the five intervals of stance phase in barefoot versus shod running. Cornwall and McPoil (2002) showed that forefoot inversion was coupled with rearfoot eversion during the heel-strike phase period of walking; in contrast, Pohl et al. (2006) found that the forefoot was everted with respect to the rearfoot in the heel-strike period of barefoot running. These findings suggest that the forefoot frontal plane motion with respect to rearfoot could vary during different gait patterns (running versus walking). Shod running had no significant effect on the mean relative angle of FFd/p–RFev/in during the five intervals of stance phase. Hunt et al., 2001, Lundberg et al., 1989a, Lundberg et al., 1989b showed that the forefoot sagittal and frontal plane motion patterns were linked to the collapse of the medial longitudinal arch. They believed that the talonavicular joint could contribute to the forefoot sagittal and frontal plane motions and arch behaviours. Generally, the sandals could not change FFev/in–RFev/in and FFd/p–RFev/in coupling patterns as well as tibial internal rotation as compared to barefoot running. However, regarding to the twisted plate and mitered hinge models, significant changes of FFev/in–RFev/in could likely affect on the amount of tibial rotations during running. This needs to further investigations with different footwear structures and foot orthoses.

The mean relative angle of FF ad/ab–RFev/in was different among the five intervals of the stance phase and between the barefoot and shod conditions. A statistically significant coupling relationship was previously reported between forefoot transverse and rearfoot frontal plane motions (Naster et al., 2002, Pohl et al., 2006). Significant changes in the coupling relationships between the forefoot transverse and rearfoot frontal plane motions could not indicate the amount of tibial rotation during dynamic motions. This finding is in contrast to the concept that cutaneous receptors of the forefoot may motivate the contraction of inverting muscles leading to control of the rearfoot and tibial rotations. Rattanaprasert et al. (1999) found that the tibialis posterior muscle support the arch of the foot and the frontal plane rearfoot motion was not affected by the loss of tibialis posterior muscle. They suggested that the motion of the forefoot relative to the rearfoot was mostly about the behaviour of the longitudinal arch. This is in agreement with Buchanan and Davis (2005) who observed a significant relationship between forefoot angle and navicular drop (r=0.55, P<.001) in healthy subjects. Lee et al. (1999) reported that the medial foot length was positively correlated with relative forefoot abduction while Aramantzios et al. (2005) found that the motion at the forefoot relative to the rearfoot is influenced by the mats with different hardness during landing. They suggested that the acting forces can not possibly be compensated by means of muscular actions in the forefoot motion. In general, it is speculated that the transverse plane motion of the forefoot with respect to the rearfoot could due to the flexibility of arch in absorbing shock and adapting to the ground floor surface. Therefore, the effect of sandals on the forefoot transverse plane motion could have a greater contribution to the flexibility of arch than the amount of tibia rotation.

The results of present study suggest that the frontal plane forefoot–rearfoot coupling pattern exhibit a similar trend in the out-of-phase pattern at heel-strike to the in-phase pattern at the midstance in asymptomatic feet. This finding could be compared with the frontal plane forefoot–rearfoot coupling pattern in symptomatic feet during running. Furthermore, variations in the forefoot transverse plane motion and the rearfoot frontal plane relationship could be related to the flexibility of arch in absorbing shock and adapting itself to the ground floor surface. Therefore, small changes in the tibial excessive motion could be expected when forefoot postings in foot orthotics change the forefoot transverse plane motion during running. In general, this finding questions the rational for the prophylactic use of forefoot posting in foot orthoses. Finally, sandal had no significant effects on rearfoot–tibia coupling motions. This result eliminates the possible confounding effects of sandals on the outcome measures of rearfoot–tibia coupling motions when they are tested with foot orthoses.

5. Conclusion 

Significant variations in the forefoot adduction/abduction and rearfoot eversion/inversion coupling patterns could have little effect on the amount of tibial internal rotation excursion. Yet it remains to be determined whether changes in the frontal plane forefoot–rearfoot coupling patterns influence the tibia kinematics for different shoe wears or foot orthotic interventions.