Measuring Plantar Pressures in Diabetic Patients With Lateral Forefoot Amputation

Diabetes mellitus (DM) imposes an escalating healthcare burden in Singapore, where diabetes-related amputation rates remain among the highest globally, and are continuing to increase. An index foot amputation often marks the beginning, rather than the end, of DM foot complications, where between 16-65% of patients undergo further amputation within five years. Notably there is a higher rate of reamputation associated with lateral ray/toes amputation compared to medial or other ray amputations (39% vs 25-27%), and as high as 60% of patients experience repeat ulceration after a 5th ray amputation. This suggests a very high risk of complications specifically associated with lateral ray amputations.

The lateral foot is essential for absorbing ground reaction forces, maintaining balance, and generating propulsion. Consequently, the loss of the lateral rays destabilizes the foot, leading to secondary foot deformities and increased peak plantar pressures (PPP) occurring at remaining metatarsal or midfoot regions. These mechanical shifts are the primary drivers of the "amputation cascade", associated with ulceration and infection and thereafter repeated amputations. Therefore, understanding pressure patterns in diabetes patients is critical for understanding factors contributing to further foot complications, and pave the way for developing strategies against it .

Gap in the literature Despite the biomechanical importance of the lateral ray and a high rate of complication, there is a distinct lack of empirical data regarding the biomechanical features of this specific group. It is also unclear why literature almost only exist for 1st ray and transmetatarsal amputations. Based on preliminary data extracted from diabetic foot management clinic in a Podiatry clinic in a tertiary hospital, there is approximately 25-30% of patients with lateral ray amputations that are undergoing active treatment for diabetic foot complications, forming a substantial proportion of patients and burden on the health system.

This study seeks to compare plantar pressure patterns of patients with such lateral ray amputations to those with intact feet. Complementing the existing data of medial ray and transmetatarsal amputation, we hope that results from this study can enable better identification of individuals at higher risk, guide targeted offloading, and enhance postoperative surveillance in the short term. Long term wise, the findings may inform risk stratification tools, support the incorporation of plantar pressure data into diabetic foot monitoring programmes, and improve efficiency in reducing ulcerations and healthcare burden.

MAIN AIMS

1. To quantify the magnitude of peak pressure in diabetic individuals following lateral forefoot amputation compared to age-, BMI- and neuropathy matched diabetic controls with intact feet.
2. To map the anatomical locations of peak pressure points following the loss of the lateral rays.
3. To develop a pressure pattern classification system by identifying and quantifying the prevalence of specific loading phenotypes

METHODOLOGY This study will use a cross-sectional observational design involving individuals with diabetes who have a lateral forefoot amputation, compared to an age , BMI and neuropathy-matched diabetic control group without amputations. Lateral amputations will include at least 5th toe amputation up to 3rd ray resection. A small prospective look at the reulceration rates for both groups of people will also be noted for any association to plantar pressure values.

The total use of the participant's time will be about 3 hours. The participant may need to visit the hospital up to 2 times. All of the study procedures can be completed in the same day if the participant prefer.

Clinical evaluation for all participants:

• Age, sex
• Height, Weight, Body-mass index (BMI)
• Medical history, diabetes duration, HbA1c levels
• Presence of peripheral arterial disease
• Ulcer history (location, most recent recurrence)
• Neuropathy assessment (using 10g monofilament)
• Ankle, subtalar joint and midfoot joint range of motion measurements
• Identification of pre-ulcerative callus locations.

Published studies on partial foot amputations report small, heterogeneous samples without effect size reporting. However, based on the mean difference and standard deviation published in these studies on ray amputations, a 95% confidence level with anticipated small sample variability, and power analysis (alpha = 0.05, Power = 0.80), the minimum required sample size is approximately 35 participants per group for a total of 70 participants

Protocol:

Demonstration and practice runs will be provided to familiarize participants with the walkway to reduce "targeting" (altering gait to hit the sensor). Participants will perform the second-step protocol, which has been identified as the most efficient method for obtaining reliable plantar pressure data in diabetic populations. For participants in the amputation group, the foot with the lateral amputation will be tested; while the matched controls will test with their dominant foot for standardization. Due to sensor size limitations, one step will be captured per pass on the walkway. Participants will complete a round trip to the starting point after each step. Steps will be collected continuously until 5-6 valid footprints are recorded for the foot of study in each trial. If the subject shows obvious gait deviations, abnormalities or "targeting", the steps will be considered invalid and excluded from analysis. 3 trials will be taken for each participant as per the common research practice, to generate a single representative dataset for each participant.

During the trial, participants will walk at their self-selected natural gait speed. While higher speeds typically correlate with increased peak pressures, a natural pace is prioritized to ensure patient safety and represent "real-world" routine circumstances. Walking speed will be calculated via video analysis of a 1m calibrated zone. These recordings will also be used to check for step validity and to note any gait features needed to explain observations. If there is a significant difference in gait speed between groups, it will be included as a covariate in later analysis.

Data processing A mask is applied over the pressure profile to demarcate between the anatomical locations of interest, namely toes, metatarsals, midfoot and heel. Using the built-in software in the pressure mat, the peak pressure values will be obtained for each region of interest on the mask. The dependent variables at each of the regions of interest will be recorded down for each participant and each trial to be used for statistical analysis. The pressure heatmap and Center of pressure line will be generated automatically by the software.

To ensure the integrity of the matching process, Descriptive Statistics (medians and interquartile ranges) will be used to summarize participant characteristics. To confirm that the control and amputation groups are effectively matched, Mann-Whitney U tests will be performed on age, BMI, and neuropathy scores as the data is likely to be non-normally distributed. Should significant differences arise in baseline variables (assessed from clinical evaluation), for example with gait speed or BMI, these will be formally treated as covariates in the primary analysis to isolate the effect of the amputation.

Given the expected non-normal distribution of plantar pressure data, a Generalized Linear Model (GLM) will be employed for group comparisons. This model allows for the inclusion of amputation status as the primary predictor while statistically controlling for potential confounders such as gait speed and BMI. This ensures that observed pressure elevations are attributable to altered foot morphology rather than gait speed or systemic factors. To maintain statistical stringency due to multiple anatomical regions of interest, a statistical corrections will be applied to the significance threshold (p-value) to reduce type 1 error.

Visual analytics, including plantar pressure heatmaps and center of pressure lines, will be generated as pictures from the software. Based on the location of the highest mean peak pressure, each foot will be classified into a pattern type named with the location of the peak pressure in the mask. If there is more than one location of equally high peak pressure, then it will be classified into a combination type. The proportion of feet in each pattern type in each group can then be calculated with simple statistics. These will provide clinical context to the quantitative data, illustrating shifts in the progression of the pressure wave throughout the stance phase in the presence of lateral ray resections.

To explore the biomechanical compensations that may be associated with high pressure at certain locations, Spearman's Rank Correlation will be used to assess the association between joint range of motion and the location of peak pressure. This non-parametric approach is to determine if restricted joint mobility in the ankle or midfoot is associated with the presence of peak pressure points at the metatarsals or plantar styloid area.