Effects of metatarsal bar and heel lift on foot biomechanics during gait: implications for performance-related biomechanics

Abstract

[Purpose] This study investigated the biomechanical effects of metatarsal bars (MTB) and heel lifts (HL) on lower-limb kinematics and kinetics during gait. [Participants and Methods] Twenty healthy adults underwent three-dimensional gait analysis. Ankle dorsiflexion angle, internal plantarflexion moment, medial longitudinal arch (MLA) height, and gravity movement (GM) were evaluated at 0%, 12%, 31%, and 50% of the walking cycle (WC) and at peak stance. Data were analyzed using two-way repeated-measures ANOVA (condition × WC), followed by Holm-adjusted post-hoc comparisons. [Results] Compared with HL, MTB significantly increased ankle dorsiflexion and decreased MLA height at 31% and 50% WC. Conversely, HL significantly reduced the plantarflexion moment at 12% WC but increased it at 50% WC and peak stance. No significant differences were found in GM, though MTB tended to shift GM posteriorly and HL anteriorly during the latter half of the stance phase. [Conclusion] MTB and HL exhibit distinct phase-specific biomechanical strategies. MTB increases dorsiflexion and reduces MLA height during mid-to-late stance (31–50% WC), suggesting a shock absorption-oriented strategy. In contrast, HL modulates plantarflexion moments to facilitate forward progression. These findings emphasize selecting orthotic components based on specific biomechanical demands at different gait phases.

INTRODUCTION

Foot orthoses have evolved since the late nineteenth century and are widely used to correct foot alignment and manage pain. Among the developments in the twentieth century, the Root paradigm had a strong influence. Root et al. systematized the mechanisms by which orthoses enhance foot function based on foot biomechanics, thereby facilitating its application to various foot disorders and accelerating the transition from ready-made devices to standardized custom-made orthoses¹⁾. Subsequent studies were conducted according to this theory.

Kevin Kirby advanced a biomechanical perspective on triplanar rearfoot motion and proposed the rotational equilibrium theory, which links the subtalar joint axis position to foot deformity tendencies²⁾. Michael O. Seibel emphasized in Foot Function: A Programmed Text that subtalar pronation–supination significantly affects load transfer during gait. Further, he analyzed foot mechanics primarily in the frontal plane³⁾. In 1997, Thomas C. Michaud comprehensively synthesized the custom-made orthotic theory by integrating static assessment (neutral position) and dynamic evaluation during gait⁴⁾. In Japan, Makoto Iritani developed an original theory influenced by Root and popularized completely custom-made orthoses domestically. However, international recognition has been limited by the lack of studies published in English.

Since 2010, advancements in three-dimensional scanning and thermoforming have enabled the application of more precise custom-made orthoses⁵⁾. Recently, artificial intelligence-assisted gait analysis systems have been developed, resulting in improved orthosis fitting and outcome measurement⁶⁾. Nevertheless, the biomechanical perspective on the ankle–foot complex and gait strategies—which are a central part of physiotherapy—remains indispensable.

Although previous studies have examined the overall effects of foot orthoses, the region-specific biomechanical effects of individual orthotic components remain insufficiently understood. In particular, it is necessary to clarify how specific orthotic regions influence kinematic and kinetic outcomes during gait.

In the sagittal plane, MTB and HL are well-known orthotic interventions. The MTB is used to reduce forefoot peak pressure and redistribute load. Static measurements in young women significantly reduced forefoot plantar pressures (maximum/mean/minimum) with increased contact area, indicating redistribution rather than a simple pressure drop⁷⁾. Under dynamic conditions, effects under the second metatarsal head decreased more with an MTB than with a metatarsal pad. Further, oblique orientation was more effective than a perpendicular one. Thus, placement (angle/position) is a key determinant of treatment effect⁸⁾. In patients at high risk of diabetic ulceration, systematic comparisons of custom insoles showed that pressure-reducing effects depend mainly on MTB position/shape. The most effective layout placed the leading edge at approximately 77% of the foot length, just distal to the metatarsal heads⁹⁾. Thus, MTBs reduce forefoot peak pressure under both static and dynamic conditions, with the optimization of their position and angle being critical. In HL orthoses, heel or shoe lifts are commonly used for noninvasive, simple corrections for leg length discrepancy (LLD). LLD can induce pelvic tilt, spinal alignment changes, and gait asymmetry and is considered a risk factor of musculoskeletal pain¹⁰⁾. Systematic reviews have reported that heel/shoe lifts have benefits for pain relief and improved function in adults with common musculoskeletal complaints¹¹⁾. Biomechanically, HLs reduce ankle dorsiflexion, Achilles tendon strain, and tendon force^{12, 13)}. Clinically, randomized trials have shown that HLs can promote superior short-term improvements compared with eccentric exercise in mid-portion Achilles tendinopathy¹⁴⁾. Some case series have revealed immediate and short-term pain relief in insertional tendinopathy with 20-mm lifts¹⁵⁾. Thus, HLs are applied primarily to decrease Achilles tendon stress and to correct alignment in LLD.

Clinically, physiotherapists analyze posture and gait across the sagittal, frontal, and transverse planes. Among these, MTB and HL are commonly used orthotic interventions that primarily influence sagittal-plane mechanics. Furthermore, sagittal-plane kinematic and kinetic variables can generally be evaluated with greater reliability and reproducibility than frontal or transverse plane foot motions during three-dimensional gait analysis. Therefore, the present study focused specifically on sagittal-plane biomechanical variables during gait and examined the effects of MTB and HL on lower-limb kinematics and kinetics.

PARTICIPANTS AND METHODS

Twenty healthy adults underwent sagittal-plane gait analysis using a three-dimensional motion analysis system. The sample size was determined via a priori power analysis using G*Power (version 3.1.9.7) (Heinrich-Heine-Universität, Düsseldorf, Germany). To detect a significant interaction effect in a two-way repeated-measures analysis of variance (ANOVA) with a medium effect size (f=0.25), an alpha level of 0.05, and a power of 0.80, the minimum required sample size was 14 participants. Considering potential dropouts and ensuring robust statistical power, this study recruited 20 participants.

The participants were all university students of the same academic year. Twenty healthy university students (13 men and 7 women) performed steady-state walking under MTB and HL conditions for comparison in the sagittal-plane experiment. All participants were recreationally active, engaging in exercise approximately 1–2 times per week. The Ethics Committee of Gunma Paz University approved all procedures (approval no. PAZ14-22). All participants provided written informed consent. Table 1 shows the participants’ physical characteristics.

Table 1. Participant characteristics.

Variable	Sagittal-plane group (n=20)
Age (years)	21.15 ± 0.36
Sex (Male/Female)	13/7
Height (cm)	166.36 ± 6.74
Weight (kg)	59.73 ± 8.79
BMI (kg/m2)	21.49 ± 1.93
Physical activity level	Moderate (1–2 sessions/week)

Values are presented as mean ± SD unless otherwise indicated. Physical activity level was classified based on self-reported habitual exercise frequency (approximately one to two times per week). BMI: body mass index.

This study examined how MTB and HL interventions by foot orthoses affect sagittal-plane gait kinematics and kinetics. Subtle modifications in foot orthoses, which influence athletic performance, are commonly observed in both sports and clinical settings. Therefore, in the current study, thinner interventions were adopted compared with commonly used MTB and HL. Each orthosis was made of 3-mm-thick PORON^® Soft (Rogers Inoac Corporation, Aichi, Japan), a high-density, high-performance urethane foam. The MTB orthosis was approximately 4 cm long and 10 cm wide and was attached to the metatarsal head. Meanwhile, the HL orthosis was approximately 5 cm long and 5 cm wide and was attached to the calcaneus. Both the MTB and HL interventions were applied bilaterally to both feet in all participants during gait measurement. The orthotic components were directly affixed to the plantar surface of the foot using double-sided adhesive tape and remained securely attached throughout the walking trials. This fixation method is commonly used in clinical foot orthotic therapy during gait assessment and was adopted in the present study to reproduce a clinically practical condition (Fig. 1a). This study measured dynamic changes in medial longitudinal arch of the foot height (MLA height), and measurements were taken barefoot to eliminate the influence of shoes.

Fig. 1.

Experimental setup and outcome variables.

(a) The left image shows a metatarsal bar (MTB), and the right image shows a heel lift (HL). The orthoses were fabricated using 3-mm-thick PORON^® Soft. (b) Gravity movement (GM), defined as the signed anteroposterior distance between the knee joint axis and the vertical projection of the whole-body center of gravity during the stance phase. Values are defined as negative when the center of gravity is located posterior to the knee joint axis and positive when it is located anterior to the axis. (c) Medial longitudinal arch (MLA) height, calculated as the perpendicular distance from the navicular marker to the line connecting the first metatarsal head and calcaneal markers.

To measure gait, the starting position was set 3.5 m before the force platform and the ending position 3.5 m beyond it. The participants performed three free walking trials, and the average values in the trials were analyzed. A three-dimensional motion analysis system (Vicon Motion Systems, Oxford, UK) with nine infrared cameras (T10 Vicon Motion System, Nexus1.8.5 sampling: 100 Hz) captured whole-body kinematics using 35 reflective markers. Three force platforms (AMTI: ADVANCED MECHANICAL TECHNOLOGY INC., Watertown, MA, USA) were installed mid-walkway to acquire kinetic data. Data were time-normalized to one walking cycle (0–100%). The order of the MTB and HL conditions was randomized across participants. A brief interval was provided between conditions for pad replacement and adjustment of the measurement system settings. Prior to data collection, participants performed 5–6 practice walking trials to familiarize themselves with the experimental environment and walking task.

Ankle dorsiflexion angle, internal plantarflexion moment, MLA height, and gravity movement (GM) were evaluated. These variables were selected because MTB and HL primarily influence sagittal-plane foot and ankle mechanics during gait. In addition, sagittal-plane kinematic and kinetic variables can generally be evaluated with greater reliability and reproducibility than frontal or transverse plane foot motions during three-dimensional gait analysis. Therefore, the present study focused specifically on sagittal-plane biomechanical parameters. Each parameter was sampled at 0%, 12%, 31%, and 50% of the walking cycle and at the stance-phase peak. The time points followed Perry’s rocker classification (heel rocker: 0–12%, ankle rocker: 12–31%, and forefoot rocker: 31–61%)¹⁶⁾. Peak values were defined as the maximum during stance for ankle angle, internal plantarflexion moment, and GM activity and as the minimum during stance for MLA height.

MLA height was calculated by adding three markers to the plug-in-gait full model: on the calcaneus and the first metatarsal (positioned 19 mm above the floor) and on the navicular prominence. Using the first metatarsal as a reference, MLA height was defined as the perpendicular distance from the calcaneus-base line to the navicular bone^17,18,19) (Fig. 1b).

GM was defined as the signed anteroposterior distance between the knee joint axis and the vertical projection of the whole-body center of gravity during the stance phase in the sagittal plane. Values were defined as negative when the center of gravity was located posterior to the knee joint axis and positive when it was located anterior to the axis. This parameter was introduced to quantify the relative anterior–posterior position of the body with respect to a functional joint reference, which cannot be determined from center-of-mass trajectory alone^{20, 21)} (Fig. 1c).

Normality was assessed using the Shapiro–Wilk test. Minor deviations from normality were observed in MLA height; however, parametric analyses were conducted with consideration of the robustness of repeated-measures ANOVA to violations of normality.

A two-way repeated-measures ANOVA was conducted to assess the main effects of condition and time and their interaction. If significant effects were detected, post-hoc pairwise comparisons were performed using Holm correction to control type I error due to multiple comparisons. Statistical analyses were performed using R (version 4.3.2; R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

First, a post-hoc power analysis was performed to evaluate the statistical power of the observed interaction. Based on the actual sample size (N=20) and the observed effect size (f=0.37), the achieved statistical power (1−β) was 0.999. This indicates that the study had excellent sensitivity for detecting the biomechanical differences between the conditions.

The effects of orthotic condition (MTB vs. HL) and walking cycle (WC) on lower-limb kinematic and kinetic parameters were analyzed using a two-way repeated-measures ANOVA (Tables 2 and 3).

Table 2. Results of two-way repeated measures analysis of variance (ANOVA) for sagittal plane kinematic and kinetic variables during the stance phase.

Variable	Effect	df	F	p-value	partial η2 (95% CI)
Ankle dorsiflexion angle (°)	Condition	1,19	6.64	0.018*	0.26 [0.01–1.00]
	WC	4,76	61.75	<0.001*	0.76 [0.66–1.00]
	Condition × WC	4,76	1.47	0.22	0.07 [0.00–1.00]
Internal plantarflexion moment (Nm/kg)	Condition	1,19	0.33	0.57	0.02 [0.00–1.00]
	WC	4,76	729.4	<0.001*	0.97 [0.97–1.00]
	Condition × WC	4,76	7.69	<0.001*	0.29 [0.13–1.00]
MLA height (mm)	Condition	1,19	6.64	0.0184	0.26 [0.03–1.00]
	WC	4,76	61.75	<0.001*	0.76 [0.69–1.00]
	Condition × WC	4,76	1.47	0.22	0.07 [0.00–1.00]
GM (mm)	Condition	1,19	0.025	0.877	0.001 [0.00–1.00]
	WC	4,76	104.1	<0.001*	0.85 [0.79–1.00]
	Condition × WC	4,76	2.7	0.037*	0.12 [0.00–1.00]

Results of two-way repeated measures ANOVA examining the effects of condition (metatarsal bar [MTB] vs. heel lift [HL]) and walking cycle (WC) on ankle dorsiflexion angle, internal plantarflexion moment, medial longitudinal arch (MLA) height, and gravity movement (GM). Values are presented as F-values with corresponding p-values and partial eta squared (η²p) with 95% confidence intervals. WC: walking cycle; MLA: medial longitudinal arch; GM: gravity movement (anteroposterior distance relative to the knee joint axis).

Table 3. Comparison of sagittal plane kinematic and kinetic variables between MTB and HL at each walking cycle point.

	Ankle dorsiflexion angle (°)				Internal plantarflexion moment (Nm/kg)
WC (%)	MTB	HL	p-value	d (dz)	MTB	HL	p-value	d (dz)
0	−0.96 ± 3.25	−1.75 ± 2.54	0.319	0.23	0.00 ± 0.01	0.00 ± 0.02	0.936	0.02
12	0.72 ± 3.09	−0.95 ± 3.17	0.081	0.41	0.13 ± 0.13	0.10 ± 0.14	0.038*	0.5
31	9.04 ± 3.84	7.60 ± 3.66	0.006*	0.7	0.57 ± 0.18	0.54 ± 0.20	0.064	0.44
50	13.36 ± 5.29	12.51 ± 4.87	0.034*	0.51	1.33 ± 0.14	1.36 ± 0.15	0.044*	0.48
Peak	14.68 ± 5.11	13.66 ± 4.93	0.066	0.44	1.37 ± 0.14	1.40 ± 0.15	0.028*	0.53
Peak (%)	46.40 ± 1.83	46.75 ± 1.61			47.55 ± 1.40	47.90 ± 1.14
	MLA height (mm)				GM (mm)
WC (%)	MTB	HL	p-value	d (dz)	MTB	HL	p-value	d (dz)
0	23.67 ± 5.38	23.93 ± 5.25	0.319	0.23	130.48 ± 27.60	126.54 ± 21.11	0.076	0.42
12	22.47 ± 4.47	22.91 ± 4.48	0.081	0.41	113.68 ± 31.01	111.40 ± 30.13	0.43	0.18
31	21.60 ± 4.91	22.40 ± 5.11	0.006*	0.7	71.73 ± 18.88	72.68 ± 18.67	0.89	0.03
50	19.54 ± 5.17	20.19 ± 5.02	0.034*	0.51	179.29 ± 13.53	183.77 ± 12.09	0.068	0.43
Peak	18.18 ± 5.45	19.17 ± 5.21	0.066	0.44	179.93 ± 13.21	184.13 ± 11.73	0.092	0.4
Peak (%)	50.50 ± 1.75	49.0 ± 2.83			49.35 ± 1.01	49.45 ± 0.92

Comparison of ankle dorsiflexion angle, internal plantarflexion moment, medial longitudinal arch (MLA) height, and gravity movement (GM) between metatarsal bar (MTB) and heel lift (HL) at each walking cycle (WC) point. Values are presented as mean ± standard deviation. P-values were obtained using paired t-tests (Holm-adjusted), and effect sizes are expressed as Cohen’s dz. WC: walking cycle; MLA: medial longitudinal arch; GM: gravity movement *p<0.05. Effect sizes were interpreted as small (0.2), moderate (0.5), and large (0.8).

For ankle dorsiflexion angle, significant main effects of condition (p=0.018) and WC (p<0.001) were observed, with no significant interaction. Post-hoc analysis revealed that dorsiflexion was greater with MTB at 31% and 50% of the stance phase (p<0.05, dz=0.51–0.70).

For internal plantarflexion moment, a significant main effect of WC (p<0.001) and a significant interaction (p<0.001) were observed. Post-hoc comparisons showed that MTB resulted in a greater moment at 12% of the stance phase (p=0.038, dz=0.50), whereas HL showed greater plantarflexion moments at 50% of the stance phase and at peak stance (p<0.05, dz=0.48–0.53).

For MLA height, the main effects of condition (p=0.018) and WC (p<0.001) were significant, with no significant interaction observed. Post-hoc analysis indicated that MTB resulted in a significantly lower MLA height at 31% and 50% of the stance phase (p<0.05, dz=0.51–0.70).

For GM, only the significant main effect of WC (p<0.001) and the small interaction effect (p=0.037) were observed, with no remarkable differences between conditions in the post-hoc comparisons.

Overall, the MTB increased ankle dorsiflexion and reduced MLA height, whereas HL increased plantarflexion moment during the mid to late stance.

DISCUSSION

This study investigated the effects of commonly used foot orthotic interventions, namely MTB and HL, on lower-limb kinematics and kinetics during gait.

The significant main effect of condition was observed for the ankle dorsiflexion angle, with no significant interaction with WC. This finding indicates that the effects of MTB and HL were not dependent on specific gait phases and were consistently present throughout the stance phase. MTB resulted in greater ankle dorsiflexion across stances, with post-hoc differences observed at 31% WC and 50% WC. These differences did not represent phase-specific effects but instead reflected points at which between-condition differences became more apparent. Therefore, MTB may be an important intervention in terms of altering ankle dorsiflexion range of motion.

In contrast, a significant interaction between condition and WC was observed for the internal plantarflexion moment, indicating phase-dependent effects. In particular, HL reduced the plantarflexion moment at 12% WC during the loading response phase. Greater plantarflexion moments were observed at 50% WC and peak WC during the terminal stance. This pattern suggests that the transient reduction in early stance may function as a preparatory phase for subsequent force generation during the terminal stance. Such temporal changes may indirectly reflect mechanisms involving muscle–tendon unit stretch, including stretch reflex-like behavior and elastic energy storage. Therefore, HL can be an important intervention in terms of regulating the pattern of ankle joint moment generation.

For MLA height, the significant main effect of condition was observed with no significant interaction. This result indicates that MTB consistently reduced the MLA height of the foot compared with HL throughout the stance phase. Hence, MTB can decrease the structural support of the MLA of the foot. Although post-hoc differences were observed at 31% WC and 50% WC, these were not phase-specific effects. Rather, they indicated that arch lowering becomes more apparent from mid-stance to terminal stance.

For the GM, MTB had a tendency toward posterior displacement and HL toward anterior displacement during the latter half of the stance phase. However, a significant main effect of condition was not observed. Although an interaction was present, post-hoc comparisons revealed no substantial differences. Based on these findings, despite localized biomechanical changes at the foot and ankle, GM remains relatively stable.

Overall, these findings suggest a trade-off between mobility and stability. MTB increases ankle dorsiflexion while reducing the MLA height of the foot, indicating enhanced mobility but reduced structural support. In contrast, HL preserves the structure of the MLA of the foot and increases plantarflexion moment during the terminal stance, suggesting that it contributes to mechanical stability.

MTB increased ankle dorsiflexion across most of the stance phases. Considering that MTB is designed to redistribute load toward the forefoot^7,8,9), this increase in dorsiflexion may reflect enhanced shock absorption mechanisms. In addition, the MLA height was reduced, indicating that increased dorsiflexion and arch lowering may act synergistically to absorb impact. Alternatively, arch lowering may reflect compensatory movements such as foot pronation. Although not statistically significant, GM was more likely to shift posteriorly during the latter half of the stance phase, which may reflect a shock absorption strategy. Therefore, MTB may be a shock absorption-oriented biomechanical strategy.

HL shortens the functional length of the Achilles tendon and limits ankle dorsiflexion^{12, 13)}. In this study, HL increased plantarflexion moment during the terminal stance and showed a tendency toward anterior GM displacement during the latter half of the stance phase. According to these findings, HL may contribute to forward progression during the terminal stance. Therefore, HL can be a forward progression-oriented biomechanical strategy.

This study has several limitations. The PORON material used in this study may have been thinner than materials typically used for MTB and HL. Additionally, the participants were healthy adults, which may limit generalizability of the findings to other populations. Although the order of conditions was randomized and practice walking trials were performed, the same participants underwent both conditions; therefore, the potential influence of carryover or learning effects cannot be completely excluded. Subtle adaptations associated with repeated walking trials may have affected the results. In addition, because no baseline barefoot condition was included, it was not possible to determine whether the observed differences resulted from increases in the MTB condition or decreases in the HL condition relative to baseline.

In conclusion, the results of this study demonstrate that even a small 3-mm foot orthotic modification can alter kinematic and kinetic parameters. MTB may reflect a shock absorption-oriented strategy, whereas HL may reflect a forward progression-oriented strategy. Clinically, these findings suggest that careful consideration is required when prescribing and designing foot orthoses. Further research is needed to clarify how region-specific orthotic modifications contribute to these biomechanical changes and how they influence performance.

Conflict of interest

Authors declare no conflicts of interest associated with this manuscript.