Introduction

Leprosy is an infectious disease caused by Mycobacterium leprae and its prevalence remains significant in countries such as Brazil and India1,2. Leprosy can lead to a neuropathic condition with impairment of sensation and muscle activation, resulting in a series of sensory-motor dysfunctions and lesions in the upper and lower limbs of patients3,4.

Changes in gait and balance resulting from peripheral polyneuropathy are already widely known. Multibacillary leprosy often develop diffuse and silent neuropathies, which can lead to these changes. Previously, our group reported that leprosy patients with higher tactile sensitivity loss had higher plantar pressure in forefoot regions and that higher pressures in the forefoot regions were correlated with greater sway in the static balance evaluation of the leprosy patients5.

The loss of sensitivity, particularly in the heel area, can alter an individual’s proprioception and consequently affect their balance and centre of pressure shift on the feet. This alteration can also lead to changes in gait, disrupting the distribution of plantar pressure. Patients with leprosy and impaired plantar sensitivity may exhibit higher plantar pressure peaks in both static and dynamic postures compared to healthy individuals, particularly in the heel and forefoot regions6,7,8. Although leprosy causes sensory and motor problems, limited research has examined postural control in patients8,9,10. Due to a lack of studies, there’s still no consensus among experts as to whether leprosy affects balance compared to people who don’t have the disease11,12.

An important physiological measure that is relevant to the assessment of balance control in leprosy patients is known as anticipatory postural adjustment (APA). This complex process involves the activation of mechanisms that occur prior to the initiation of movement. Its purpose is to recalibrate the distribution of forces within the body with the ultimate goal of achieving body stability13,14. The generation of APAs involves intricate and complex co-ordination between sensory and motor information, which can be affected by a disease that affects the transmission of this information along the nervous system15,16.

The assessment of APA requires the use of a video capture system (VCS); however, the significant financial investment required for this technology can be a significant barrier to the study of postural change17,18. This barrier takes on added significance in the context of a disease that is prevalent in impoverished communities and often receives insufficient attention from government agencies, resulting in limited financial resources for effective control19,20, such as leprosy. In recent years, inertial sensors (IS) integrated into wearable devices and smartphones have been validated to collect and analyse APAs in different populations, including adults18,21,22,23, the elderly, and individuals with Parkinson’s disease15,17,24,25,26. This progress opens up the possibility of using these sensors as viable alternatives, offering a more economically efficient approach to studying APAs in the context of leprosy. By developing novel strategies that are not only cost effective, but also maintain the required validity for APA collection, the field of leprosy research could potentially see advances.

The current study represents a pioneering effort in the study of APA within the leprosy population. The aim was to compare APAs with healthy individuals registered using a gold standard method (VCS) and the use of the IS. We also investigated the validity and reproducibility of APAs registered in leprosy patients.

Materials and methods

Subjects

This cross-sectional study was conducted at the Laboratory of Human Motricity (LEMOH), located at the Institute of Health Sciences of the Federal University of Pará, Brazil, and followed the STROBE (Study of Observational Studies in Epidemiology) guidelines. The study involved 60 participants: 30 healthy individuals in the control group (CG) (19 males, 11 females, mean age 44.5 years ± 8.77) and 30 individuals with leprosy in the leprosy group (LG) (17 males, 13 females, mean age 44.6 years ± 12.76). CG participants had no history of infectious or chronic degenerative diseases affecting sensorimotor functions. LG participants had a bacilloscopic diagnosis of leprosy and were being treated at the Tropical Medicine Unit of the Federal University of Pará. LG members had multibacillary leprosy and were on multi-drug therapy or had type 1 or type 2 leprosy reactions. Subjects were excluded if they had grade 2 disability, hyperkeratosis, prolonged plantar ulcers, foot amputation or reabsorption, and arthrosis with painful gait and claw toes. The Human Research Ethics Committee at FUP approved the protocol and informed consent was obtained from all subjects prior to study participation. All procedures were performed in accordance with relevant guidelines and regulations.

We examined the tactile sensitivity of the plantar skin of the foot utilizing Semmes Weinstein monofilaments with forces of 0.2, 2, 4, 10, and 300 g. The experimenter applied a light touch to the skin using the monofilaments at eight distinct areas of the foot: Six in the forefoot, one in the midfoot, and one in the hindfoot. Participants were asked to report the perception of the lightest touch. Each monofilament was tested in three trials, progressing from the lightest to the heaviest. The tactile threshold was defined as the lightest monofilament that the subject could reliably perceive. Leprosy patients were classified as having peripheral neuropathy if their tactile threshold exceeded 0.2 g.

Recording of body acceleration during anticipatory postural adjustments for step initiation

A VCS (Simi motion, Simi, Germany) and a portable device containing IS (MetaMotionC, Mbientlab, USA) were used to record centre of mass (COM) accelerations in the antero-posterior, medio-lateral and superior-inferior axes before, during and after step initiation. Video analysis was considered the gold standard method for these measurements.

For video recording, a reflective marker was attached to the fifth lumbar vertebra (L5) to mark COM displacements during the experiment, and another reflective marker was placed on the left heel to mark the moment of heel lift during step initiation. Three cameras with a sampling frequency of 120 Hz (Simi Motion, Simi, Germany) were used for video recording.

For the inertial assessment, an IS (approximately 25 mm × 4 mm diameter, 0.2 g) was attached to the lower lumbar spine at the level of the L5 vertebra. The IS contained a triaxial accelerometer with a sampling rate of 100 Hz.

For the experimental protocol, participants were instructed to stand in a bipedal stance, barefoot, with their arms at their sides. The first procedure was designed to elicit VCS and IS recordings. In this procedure, participants were asked to perform a vertical jump and the moment of peak acceleration upon landing was considered as time zero for the time series obtained from both devices. All subsequent time points were then referenced to this time zero in order to align the data from the two devices.

To record the APA for step initiation, participants stood on a 2 m platform, feet side by side, heels 6 cm apart. Each participant was informed that the experimenter would give a signal (a red LED light would be turned on), after which the participant would initiate a two-step walk on the platform, always starting with the left foot to which the reflective marker was attached. Participants were also instructed to look at a fixed mark on the wall at eye level and at a distance of three metres while walking. All participants underwent prior training trials to understand the task. Ten valid trials were collected for each participant.

Signal processing

Text files containing time series of COM acceleration from both recording devices and time series of left heel displacement from the video recording were exported. The acceleration time series from the video recording was obtained by a double integration procedure using the reflective marker positioning information. Offline analysis was performed using custom computer routines programmed in the Octave/Matlab programming language.

First, the accelerometer time series of L5 from both devices and the left heel displacement time series from the video recording were synchronised. The moment of the first step in each trial performed by the participant was then identified. In this first step, the moment of heel lift was identified in the left heel displacement time series and this moment was considered to be the start of the step (Tzero). The criterion used to indicate heel lift-off was a displacement greater than the mean plus 2 times the standard deviation of a baseline of heel position before the start of the first step. The baseline was taken between 1.5 and 1 s before heel lift.

In the mediolateral accelerometric time series of L5 from both recording devices, an interval between 1.5 s before step onset and 1 s after step onset was selected to identify the presence of APAs in the recordings. A second-order Butterworth low-pass filter with a cut-off frequency of 30 Hz was applied to filter the mediolateral accelerometric signals in L5. Two experienced experimenters independently analysed each trial and determined the presence of reliable APAs in the recordings.

After identifying the presence of APAs in the recordings, the following parameters were measured (Fig. 1):

  1. 1.

    APA onset This is the time interval between the onset of the APA and Tzero (step initiation). The time of APAonset was determined as the first mediolateral deviation greater than two standard deviations from the baseline mean.

  2. 2.

    APAamp This parameter represents the maximum mediolateral acceleration of the COM before Tzero (step initiation).

  3. 3.

    PEAKtime This is the time taken to reach the peak amplitude of the APA from its onset.

Figure 1
figure 1

Time series of left heel displacement (A) and medio-lateral acceleration at L5 (B) obtained from the image capture system. Dashed line at 0 s represents the moment of heel-off (Tzero), red dashed line indicates the initiation of the initial swing phase (APA), and the interval between this point and Tzero represents APAonset (red arrow). The acceleration interval between the baseline and the maximum acceleration of the Center of Mass (COM) during APA represents APAamp (blue arrow), and the time interval between the initiation of APA and the moment of maximum acceleration of the COM during APA represents PEAKtime (green arrow).

Full size image

Statistical analysis

Statistical analysis was performed using GraphPad PRISM 9 software. The Shapiro–Wilk test was performed to assess the normality of the variables, and the difference between the measures in both groups was assessed based on the data fitting the normal distribution. The demographic characteristics of the two groups were compared using z-score test for two population proportions to assess the proportion of male and female participants in the two study groups. In addition, Student’s t-test was used to compare age, weight, and height. The Mann–Whitney test was used to assess the number of records with reliable APAs in both groups of participants.

Bland–Altman analysis was performed for the three variables in each group. The bias was calculated as the mean of the differences between the measurements obtained from each device, and the 95% confidence interval for the differences was also calculated. If the value of zero difference between the measurements did not fall within the confidence interval for the differences, the bias was considered significant.

Validation of the IS was performed by Pearson’s linear correlation analysis between the measurements obtained with the IS and the VCS. Correlations with a significance level of p < 0.05 were considered valid. Pearson’s correlation coefficients (r) were interpreted using threshold values as follows 0.00–0.19 “very weak”; 0.20–0.39 “weak”; 0.40–0.59 “moderate”; 0.60–0.79 “strong”; 0.80–1.0 “very strong”27.

The test–retest reliability of the records was calculated using the intraclass correlation coefficient (ICC) with a two-way mixed model, single measures and absolute agreement. ICC values were interpreted according to the Shrout and Fleiss classification, where ICC ≥ 0.75 indicates excellent correlation, 0.74–0.4 indicates fair to high correlation, and ICC ≤ 0.39 indicates low correlatio28,29. In addition, the standard error of measurement (SEM) was calculated as the square root of the mean square error term derived from the analysis of variance. The minimal metrically detectable change (MMDC) was considered as the 95% confidence interval of the SEM to estimate the changes between any two measurements that could be clinically significant30.

Student’s t-test was used to compare the mean APA parameters between the control and leprosy groups, and one-way ANOVA was used to compare the same parameters between the control and leprosy subgroups (normal and altered skin-foot tactile sensitivity). The significance level was set at p < 0.05.

Results

Table 1 shows the comparison of clinical-demographic characteristics of both groups. We observed that the groups were matched in male/female proportion, age, height, and weight. Leprosy patients had a significant proportion of participants with altered tactile sensitivity in the foot (13/30 participants) compared to the controls.

Table 1 Demographic and clinical description of the groups.
Full size table

Waveform analysis

The waveform of APA in the mediolateral axis recorded at L5 during the onset of the step in both groups (control group in red; leprosy group in blue) showed similar patterns (VCS in A and C; IS in B and D) (Fig. 2). The waveforms showed a transient component before the onset of the step. The component related to APA was represented by a negative deflection before the onset of the step with the left foot.

Figure 2
figure 2

Waveforms of anticipatory postural adjustments (dashed rectangles) obtained from the video capture system in the control group (A), inertial sensor in the control group (B), video capture system in the Leprosy group (C), and inertial sensor in the Leprosy group (D). The black lines represent the grand mean of each data group, and the red area (in controls) and blue area (in leprosy) represent the confidence interval of the records for each data group, considering a significance level of 95%.

Full size image

In the control group, the median number of steps preceded by reliable APAs was 8, with an interquartile range of 2. In the leprosy group, the median number of steps preceded by reliable APAs was 6.5, with an interquartile range of 2. The difference in the number of records with reliable APAs was statistically significant (p = 0.017).

Concordance, correlation, comparison, and replicability analysis of accelerometer signals for recording anticipatory postural adjustments at L5 during step initiation obtained from video capture and portable sensor

The concordance analysis of the variables APAonset, APAamp, and PEAKtime for the Leprosy and Control groups are presented in Fig. 3. For the Leprosy group, significant biases were observed for APAonset (VCS: 0.475 ms ± 0.09; IS: 0.505 ms ± 0.09; p < 0.001) and PEAKtime (VCS: 0.331 ms ± 0.07; IS: 0.307 ms ± 0.05; p < 0.008), but not for APAamp (VCS: 0.514 m/s2 ± 0.16; IS: 0.543 m/s2 ± 0.16; p = 0.07). For the Control group, significant biases were observed for APAonset (VCS: 0.44 ms ± 0.08; IS: 0.475 ms ± 0.08; p < 0.001), APAamp (VCS: 0.554 m/s2 ± 0.17; IS: 0.624 m/s2 ± 0.18; p = 0.0007), and PEAKtime (VCS: 0.321 ms ± 0.06; IS: 0.302 ms ± 0.05; p < 0.006).

Figure 3
figure 3

Concordance analysis between the same variables measured from the recordings obtained by the video capture (elements in red) and the inertial sensor (elements in blue). Solid lines represent the bias between the measurements, dashed lines represent the limits of agreement, and the shaded areas around the bias indicate the confidence interval of the bias. The bias was considered significant when the confidence interval includes the value zero of bias. (A) Concordance analysis for APA onset from leprosy group. (B) Concordance analysis for APA onset from control group. (C) Concordance analysis for APA amplitude from leprosy group. (D) Concordance analysis for APA amplitude from control group. (E) Concordance analysis for APA time to peak from leprosy group. (F) Concordance analysis for APA time to peak from control group.

Full size image

Figure 4 shows the correlation between the measurements obtained with the VCS and the IS in both groups of participants. The coefficients of linear correlation (and their statistical significance) between the measurements obtained with both devices are shown. For all measurements and for both groups, there were strong and significant correlations between the measurements taken with both devices. Linear correlation coefficients between measurements obtained with both devices were strong for the variable PEAKtime (0.75; p < 0.001) in the leprosy group and all other variables were very strong: PEAKtime (0.83; p < 0.001) in the control group; APAonset (Leprosy group: 0.96; p < 0.001; Control group: 0.92; p < 0.001), APAamp (Leprosy group: 0.86; p < 0.001; Control group: 0.8; p < 0.001). The result of the ANCOVA indicated that the slopes of the correlations between APAonset (p = 0.85), APAamp (p = 0.053), and PEAKtime (p = 0.86) obtained by both instruments did not differ.

Figure 4
figure 4

Correlation between the measurements obtained with the video capture and the inertial sensor in both study groups. (A) APA onset, (B) APA amplitude, (C) APA time to peak.

Full size image

The assessments of the replicability of measurements using both instruments in both study groups are shown in Table 2. It was observed that there was reasonable to high replicability for the variables APAonset and APAamp, and weak replicability for the variable PEAKtime in both groups and instruments.

Table 2 Intraclass correlation coefficient for the variables obtained from the video capture system and inertial sensor in both groups.
Full size table

The intergroup comparison for the measurements obtained with each instrument is presented in Table 3. The results showed a lack of statistical difference between the groups for the three variables related to the studied APAs.

Table 3 Comparison of mean APA parameters between control and leprosy groups.
Full size table

Discussion

The present study is the first to investigate APA in leprosy compared to healthy individuals and to evaluate the use of two different instruments. The results show similarity in APA between groups, although reliable APA is less common in leprosy patients. In addition, the research shows that VCS and IS measurements are reproducible, despite a bias between them, enriching the understanding of postural measurements in this complex context.

Leprosy is a disease that affects peripheral nerves and skin, obviously affecting somatosensory and motor function31, which could potentially alter anticipatory postural adjustments, as they rely on the proper functioning of these systems32. The results of the present study showed that the APAs recorded in leprosy patients were similar to those of healthy individuals. This could be explained by the clinical condition of the patients, who had a disability grade equal to or less than 1, indicating no significant sensory-motor impairment33. This fact can be correlated with the study by34, where tibial nerve block resulting in sensory loss of the sole of the foot and weakness of intrinsic foot muscles did not significantly alter the kinetics or temporal characteristics of gait initiation. Perhaps in patients with a higher degree of disability it would be possible to distinguish the characteristics of the APA recordings between the groups. Even the leprosy patients, although having similar APAs to the control group, were found to have a lower frequency of APAs than the controls. This observation could reflect subclinical losses that affect the difficulty of consistently activating the motor programme. It was reported that one or more components of APA were absent in approximately a quarter of all gait initiation trials in healthy individuals, the results in the leprosy group fell below this reference33. Multibacillary leprosy is often associated with sensory-motor neuropathy that tends to be more diffuse and silent, sometimes even subclinical, observed only through monofilament evaluation and sensory and motor nerve conduction studies. Disability scale, in this case, would not be an ideal parameter. In the International Classification of Functioning, Disability and Health (ICF) “Disability” is used as an umbrella term for impairments, activity limitations and participation restrictions. Grade 2 should be mild absorption of only one finger, a severe crack because of dryness and contractures, so, it is not the best indicator for evaluation of sensory impairments in feet. We compared the APA parameters from controls to the patients considering their neurological examination of the skin-foot tactile sensitivity. We continued observed no significant differences in the APA parameters among the groups.

The current study also showed that both the gold standard method (VCS) and the low-cost alternative method (IS) were reproducible and showed a high correlation between the values of the variables obtained from both instruments. Similar results were observed in a previous study in which we validated the IS in healthy individuals using the same protocol. In that study, caution was advised when using sensors of different masses, as sensors that are too light may capture more noise, thereby hindering the visualisation of APAs35. These results are promising for advocating the use of low-cost inertial sensors to monitor and record APAs in a marginalised population that may not have easy access to gold standard methods such as video recording.

The great advantage of using a video capture system as a gold standard method for measuring APA compared to inertial measurements from wearable sensors is that both methods evaluate the inertial modifications of the center of mass, allowing for a more direct assessment of agreement between the measurements obtained by both devices. A force platform could also be used as a gold standard method, but this would limit the assessment of agreement between some measurements with the inertial sensors, such as the magnitude measurements of APAs. Some limitations of the current study should be considered, such as the sampling of leprosy patients with a disability grade equal to or less than 1 and of the multibacillary type. Future studies involving paucibacillary patients, those with higher levels of disability, and those with permanent sensorimotor sequelae may provide further insight into the effects of leprosy on anticipatory postural adjustments. Nevertheless, we believe that the current study has an adequate sample size to evaluate the early effects of leprosy on patients.

The inertial recording of anticipatory postural adjustments prior to the initiation of a step is a rapid, simple, inexpensive, yet sophisticated method that can be used to monitor sensorimotor impairments in leprosy patients. This technique could assist in the therapeutic management and rehabilitation of these patients. This type of assessment is not yet included in the official protocols for leprosy patients, but could be particularly useful in cases where there is suspicion or evidence of significant sensory-motor loss.