A study of the immediate effects of glycerine-filled insoles, contoured prefabricated orthoses and flat insoles on single-leg balance, gait patterns and perceived comfort in healthy adults

Thirty healthy adults (17 men, 13 women) with a mean (SD): age, 24.3 (2.5) years; height, 178.3 (10.0) cm; and body mass, 72.4 (12.4) kg, were recruited from the university staff and student base, in response to advertisements. All individuals who were willing to wear footwear interventions for the duration of the test procedures were included. Exclusion criteria were current use of shoe insoles or foot orthoses; history of lower limb or foot injuries, including pain, within the previous 12 months; any circulatory or sensory conditions affecting the feet or lower limbs; inability to walk 10 m unassisted; any other musculoskeletal, sensory or neurological impairments which are known to alter balance or gait performance; unable to understand or communicate in English and thus provide informed consent and follow instructions. Ethical approval was granted by the Medical Research Ethics Committee at The University of Queensland (#2014000140). Written informed consent was obtained from all participants. Standing foot positioning was quantified using the 6-item Foot Posture Index (FPI) [26] to characterise the study sample. Mean (SD) FPI score for the test leg was 4.6 (3.3). Normal foot posture (FPI score: 0 to +5) was observed in 66.7 % of participants (N = 20). 26.7 % of the sample (N = 8) presented with pronation (FPI score: +6 to +9), with the remaining 6.6 % (N = 2) reported to be highly pronated (FPI score: 10+).

All test procedures were conducted in a University laboratory. Using a within-subject repeated measures study design, balance and gait tasks were assessed for all participants during three different randomised shoe inserts. The order in which the SEBT and level-ground walking tests were performed was also randomised, with single-leg standing completed last due to laboratory constraints. Randomisation schedules were generated using an online list randomiser (www.​random.​org/​lists/​). To allow for familiarisation with the test procedures, all participants completed 1–2 practice trials of each balance and gait task prior to data collection.

StandEO tests were performed on either the right (N = 14) or left leg (N = 16) (randomly presented). As our participants were healthy, and did not present with unilateral lower limb impairments, we assumed that their response to shoe insoles and foot orthoses would be comparable between-legs. Participants were instructed to stand in the centre of a force platform on their randomly allocated test leg, with their opposite leg flexed to ~90° at the knee, with the legs not touching and arms folded across their chest [27], over 30 s. Participants were asked to look straight ahead at a black, circular, 15 cm diameter, visual target, positioned at eye level, located 3 m from the centre of the force platform. When a participant touched down with their raised foot, this was considered to be a balance failure and the test was discarded and repeated. For each insert, three StandEO repetitions were performed, and the average of each was calculated. A 10 s rest period was given between repetitions to prevent fatigue.

The SEBT is a reliable assessment of dynamic stability [28, 29] which has been used in previous studies to explore the effects of foot orthoses on balance performance [30]. Participants stood with either their right (N = 14) or left (N = 16) heel (randomly presented) in the centre of an 8-point star marked on the floor with tape (with each point set at a 45° angle), with their arms folded across their chest as described by Kinzey et al. [28]. All participants were given the standardised instruction to “Reach as far as you can along the line without moving your standing foot, keeping your heel down. When touching the line with your reaching foot, try not to step or place all your weight down: lightly touch the ground then return to the starting position”. Three repetitions were performed along each of eight directional lines: anterior, anteromedial, medial, posteromedial, posterior, posterolateral, lateral and anterolateral. All participants began the test by reaching in the anterior direction, and were thereafter randomised to completing the remaining reaches in either a clockwise (N = 16) or anticlockwise (N = 14) sequence. A marker was placed on the floor at the point where the most distal aspect of the reaching foot touched the ground. The distance from the centre of the star to the marker was determined using a tape measure, and averaged across the three repetitions for each direction. All participants were given a 10 s rest period between repetitions. Tests were discarded and repeated if a participant lost their balance, touched outside the directional lines, or bore weight through their reach foot.

Temporospatial gait parameters were measured during level ground walking, at a comfortable self-selected speed in each of the three insert conditions, using an instrumented walkway system (GAITRite®, CIR Systems, Inc., Havertown, PA 19083, USA). The GAITRite® system is an electronic walkway, approximately 8.2 m long, with the active area being 0.61 m wide and 7.32 m long. The GAITRite® system is a reliable and valid tool for measuring gait performance in healthy adults [31]. To allow for acceleration and deceleration at the beginning and end of each gait trial, a taped line was placed 2 metres before and 2 metres after the GAITRite® walkway. Participants were positioned at the start line and given the following standardised instructions: “Starting at the blue line, walk along the length of the walkway at your usual, comfortable speed and continue until you pass the second blue line at the end of the walkway”. For each temporospatial measure, the average of three ~12 m gait trials was calculated for each insert.

The glycerine-filled shoe insoles (BestSole Inc., Boynton Beach, FL 33424, USA) were ~2 mm thick (at their most consistent thickness), composed of ethylene-vinyl acetate plastic with polyester knit fabric overlay. Channels, extending from the forefoot to heel regions, were filled with 100 % fluid United States Pharmacopeia ingestible grade glycerine. The flat (control) insoles were also constructed from ethylene-vinyl acetate plastic with polyester knit fabric overlay, but did not contain fluid-filled channels. These control insoles were used to isolate the effects of the glycerine-fluid on balance and walking, as there were no other differences in the composition or characteristics of the flat insoles relative to the glycerine-filled insoles. The contoured prefabricated foot orthoses investigated in this study (Orthaheel, Triplanar motion control, Vionic Group LLC, San Rafael, CA 94903, USA) were selected from a range of popular, commercially available inserts, designed for providing motion control. All the inserts were available in nine shoe sizes ranging from a women’s size 4.5 to men’s size 16 (United States sizing). The insoles and orthoses were worn within participants’ own closed shoes and were fitted by the investigators after removing the original footwear inserts.

Force data during the StandEO task was collected using a Kistler force platform (Model 9296AA, Kistler, Alton, UK). Data were sampled at 1,000 Hz (Power1401 Data Acquisition System, Cambridge Electronic Design, UK) and low-pass filtered (20 Hz, 4th order Butterworth filter) off-line using Matlab (The Mathworks, Nathick, USA). Balance measures during StandEO included CoP total path velocity (mm s⁻¹), and the range and standard deviation of movement in mediolateral (ML Range, MLSD) and anterior-posterior (AP Range, APSD) directions (mm), and were analysed over the first 29 s of the task. CoP path velocity (mm s⁻¹) represents the speed at which the centre of mass is moving, with a higher value indicating more rapid and potentially unstable movement [32]. CoP range (mm) is the maximum range of movement during the task in each ML and AP axes, with higher values indicating greater sway [32]. CoP SD (mm) corresponds to the variability about the average position of the CoP coordinate, with higher values suggesting greater exploratory (or less controlled) behaviour and movement patterns which are not constrained [33]. For the SEBT, reach distance in each of eight directions was measured in cm. Temporospatial gait parameters, measured using the GAITRite® system, were overall walking velocity (cm s⁻¹), cadence (steps min⁻¹), step length (cm), base of support (cm), double-limb support time (% gait cycle), toe in/out angle (°). After completing all balance and gait tasks for one insert, participants were asked to rate their perceived level of comfort on a 100 mm visual analogue scale (with 0 mm being extremely uncomfortable and 100 mm being extremely comfortable) [34]. Upon completion of all test procedures, participants were then asked to rank the relative comfort of each of the three inserts, from 1^st to 3^rd with 1^st being most comfortable and 3^rd being least comfortable. The three inserts were listed vertically and displayed in a randomised order for each participant. Both the ranking and rating scales have been previously validated and are reported to be reliable measures of footwear comfort [34].

Statistical analyses were performed using SPSS version 22 (SPSS Inc, Chicago, IL 60606, USA). Data were examined for normality and homogeneity of variance. Three independent repeated measures multivariate analyses (MANOVA) were performed to determine any between-condition differences in CoP measures during StandEO, SEBT performance, and temporospatial gait parameters. The within-subjects factor was insert condition (glycerine-filled insole, prefabricated orthosis, flat insole). The dependent variables were CoP movement (CoP velocity, ML range, MLSD, AP range, APSD); SEBT reach distance (in anterior, anteromedial, medial, posteromedial, posterior, posterolateral, lateral, anterolateral directions); and temporospatial gait parameters (velocity, cadence, step length, base of support, double-limb support, toe in/out). Where a significant multivariate effect was observed, follow-up univariate tests (ANOVA) were performed. Repeated measures ANOVA were used to explore any between-condition differences in VAS rating scores for insert comfort. Where the assumption of sphericity was violated, a Greenhouse–Geisser correction was applied. Pearson’s correlation coefficients were used to explore whether VAS comfort rating scores were associated with CoP measures, SEBT reach distance, and temporospatial gait parameters for each of the insert conditions. A chi square test was used to determine any differences in the frequency with which each insert was ranked from most comfortable to least comfortable. The level of significance was set to 0.05, which was adjusted to 0.017 for each of the three MANOVAs.

Four participants were unable to successfully perform the StandEO without touching down with their raised leg, and were therefore excluded from the analyses. Repeated measures MANOVA showed no significant main effect of insert condition on CoP measures during StandEO (Wilks’ λ = 0.660, F (10,92) = 2.125, P = 0.030) (Table 1).

Repeated measures MANOVA showed no significant between-condition differences for SEBT performance in any of the eight reach directions (Wilks’ λ = 0.644, F (16,102) = 1.572, P = 0.090) (Table 2).

No significant multivariate effects of insole condition were observed for any temporospatial gait parameters during level ground walking (Wilks’ λ = 0.604, F (20,98) = 1.405, P = 0.138) (Table 3).

Frequency analyses indicated that the flat insoles were most commonly ranked as the ‘most comfortable’ condition (N = 15), with the glycerine-filled insole most commonly reported to be ‘least comfortable’ (N = 13) (Fig. 1). Chi square test indicated that differences in comfort ranking scores between the three insert conditions were statistically significant (χ² [4, N = 30] = 10.600, P = 0.031). Mean (SD) VAS comfort rating scores were 53.4 (23.2) mm (flat, control insole), 48.2 (23.6) mm (glycerine-filled insole) and 51.8 (24.7) mm (prefabricated orthosis), respectively. Repeated measures ANOVA reported no significant between-condition differences in VAS comfort ratings (F (2,58) = 0.453, P = 0.638). A significant, but weak correlation was observed between VAS rating scores for the flat insole with double-limb support time (left leg) during level-ground walking (r = −0.398, P = 0.029). Similarly, weak to moderate associations were reported between the prefabricated orthosis and posterior reach distance during the SEBT (r = −0.406, P = 0.026) and toe in/out angle during walking (r = −0.368, P = 0.045).

There were several limitations to this study. Firstly, a young asymptomatic population was explored, therefore our findings cannot be generalised to symptomatic clinical populations who may wear insoles or orthoses for therapeutic benefits. We felt it was important to explore the immediate effects of a novel glycerine-filled insole on balance and walking in healthy adults first in order to establish any potentially detrimental effects before applying them to patients with known balance impairments for whom insoles might be a helpful component of their management. Secondly, the viscous glycerine content of the novel insert is less likely to perturb the measures of single-leg standing, controlled voluntary movement during the SEBT, or level-ground walking at a comfortable pace that we studied, when compared to more dynamic activities such as reactive stepping, running, or jumping. Our findings cannot be extrapolated to these more dynamic activities before further research is conducted. Thirdly, due to obvious differences in the composition, structure, and feel of the inserts it was not possible to blind participants from the interventions under investigation. We countered potentially deleterious effects of non-blinding by keeping our participants unaware of the inserts of greatest interest in this study. Finally, participants wore their own footwear during test procedures, in order to reflect usual conditions to which they were already accustomed. However, it is possible that there may have been an interaction between the inserts and shoes.