Variation in the location of the shoe sole flexion point influences plantar loading patterns during gait

Female staff and students between the age of 20 and 40 were recruited at the Fontys University for Applied Sciences via e-mail. All participants had to have a shoe size (European) between 38 and 41. Participants were excluded if they (i) had a Foot Posture Index outside the normal range (<0 - >6) [13, 14], (ii) had rheumatoid arthritis, (iii) had diabetic neuropathy or (iv) were wearing custom made orthotic devices. The Medical Ethics Committee of the Vrije Universiteit Medical Centre approved the study (2009/267) and informed consent was obtained from all participants.

Three shoes (Bata Industrials^© type EVA: laced work shoes, nubuck leather upper, PU-sole) with different positions of the sole flexion point were worn by all participants. To create differences in the sole flexion point, the outer soles of the shoes were incised over the full width of the shoe up to the inner sole. Although we acknowledge that other aspects of shoe structure could influence the sole flexion point (such as sole hardness and the flexibility of the upper), the incision approach enabled us to standardise the shoes as much as possible while retaining the overall integrity of the shoe. The tread pattern is scaled proportionally to shoe size. The location of MTPJ 1 was palpated for four people, two in the smallest size and two in the largest size. The location of MTPJ 1 in respect to the outer sole tread pattern was assessed. The chosen orientation of the incision was a straight line similar to the orientation of the axis that combines motion of the MTP 1 and 2 joint. One shoe was incised directly underneath the MTPJ 1 and 2 axis (hereafter referred to as sub-MTPJ); one (depending on the shoe size) 2–3 cm proximal to the MTPJs (hereafter referred to as prox-MTPJ) and one at the level of the tarso-metatarsal joints (hereafter referred to as midfoot) (Figure 1). All incisions were first made in the smallest sized shoe and parallel to the orientation of the MTP 1 and 2 joint axis. The location of the incisions for size 38 in respect to the tread pattern was copied for the larger sizes.

Prior to the study, the shoe size of the right foot was established by means of a shoe size calliper (heel to longest toe), and this size was used for both feet regardless of the size of the left foot. The corresponding size insoles for the plantar pressure assessment were placed in the shoe. All shoes were tied by the same researcher in order to diminish the influence of the lace tightness on the outcome as much as possible. Participants were asked to walk in the shoes for two minutes prior to data collection. During data collection participants walked over an eight metre walkway at their own comfortable speed. The order of presentation of the three shoe conditions was randomised.

In-shoe plantar pressures were measured using the Pedar-X system (Novel gmbh, Münich, Germany), which consists of 99 capacitive sensors arranged in a grid and embedded within a thin flexible insole approximately 2 mm thick. A previous study has demonstrated acceptable reliability of this system with the exception of the area under the toes [15]. Data of twelve steps were obtained after excluding initiation, termination and turning steps [16]. Previous findings of Pataky et al.[17] highlighted the importance of choosing the edges of masks that are congruent with plantar anatomy. Based on an average peak pressure template of 104 ‘normal and healthy’ feet [18], anatomically correct masks were therefore created to evaluate the pressure-related outcome measures to reflect the more proximal location of MTPJs 4 and 5 and the 4^th and 5^th toes (Figure 2).

Five pressure-related variables were selected a priori for analysis: peak pressure (PP), contact time (CT), pressure–time integral (PTI) and the ratio PP and PTI between masks MTPJ1 and MTPJ2-3 (PP and PTI ratio). PP and CT were derived directly from the software. PTI was calculated according the method described by Melai et al.[19]:

The PP and PTI ratio were determined by dividing the PP or PTI of the MTPJ2-3 mask by the PP or PTI of the MTPJ1 mask, to provide an indication of lateral shift in forefoot loading. Walking speed and BMI were measured to test for confounding.

All data were tested for normalcy; the data were defined to be normal distributed when skewness > 1.0 (IBM^© SPSS^© statistics version 20.0.0). Skewed data were log transformed. For each mask all variables were compared across the three different shoe conditions using multilevel model linear regression (MLwiNversion 2.26 [20]). The different outcome measures were set as level 1, the participants were set as level 2 [21]. Random intercepts and random slopes were added to the basic model and changes of the -2log likelihood ratio were used to evaluate the best model. The confidence interval was set at 95%.

Walking with the prox-MTPJ versus the sub-MTPJ sole flexion point produced changes in PP for two masks (Figure 3). The PP increased in the MTPJ 4–5 mask (avg 154.2 kPa from 145.2 kPa p = 0.028) and decreased in the medial midfoot mask (avg. 91.5 kPa from 99.9 kPa p = 0.008) (Table 1). No other changes were found between these two sole flexion points for any of the other variables. Several differences were found when walking in the shoe with a midfoot flexion point. Significant decreases were found in PP medial midfoot (94 kPa p = 0.023), medial hindfoot (227 kPa p = 0.023) and lateral hindfoot (231 kPap = 0.01) when walking with a midfoot sole flexion point versus the sub-MTPJ sole flexion point (resp. 99.9, 237, 243.3 kPa). No significant changes were found comparing the midfoot to the prox-MTPJ flexion point.

Contact time was found to decrease in the midfoot (medial: 591 ms p = 0.013 and lateral 607.2 ms p < 0.000) wearing the midfoot sole flexion point shoe versus both other flexion points (sub-MTPJ resp. 608.4 and 628.7 ms; prox-MTPJ resp. 606.7 and 624.1 ms). The lateral hindfoot was found to have a lower CT (415.4 ms p = 0.025) in the midfoot shoe versus the sub-MTPJ shoe (444.8 ms).

Reduction of the PTI in the medial midfoot (0.957 Ns/cm² p = 0.003), lateral midfoot (1.252 Ns/cm² p = 0.046), medial hindfoot (3.628 Ns/cm² p = 0.003) and lateral hindfoot (3.633 Ns/cm² P = 0.006) were found in the midfoot flexion point versus the sub-MTPJ sole flexion point (resp. 1.09, 1.514, 3.895 and 3.852 Ns/cm²) and the prox-MTPJ (resp. 1.09, 1.628, 3.809 and 3.914 Ns/cm²).

Walking in the midfootsole flexion point shoe showed a trend towards a lateral shift in forefoot loading (PTI MTP2-3/MTP1 = 1.09) compared to the sub-MTPJ (PTI MTP2-3/MTP1 = 1.03) but this trend was not significant (p = 0.056).