We designed a single-blind, randomized clinical trial to examine the effects of customized insoles on gait in patients with hemiplegia. Eligible patients were randomized into ‘conventional gait training + customized insoles’ (Group A) or ‘conventional training’ (Group B) at a 1:1 ratio using a computer-generated random table. We put trial instructions and groupings in sealed envelopes and the participants were randomly assigned in order.
Each patient was evaluated by the same assessor who was unaware of the group assignment. The primary outcome measures were conducted three times (baseline[T0], 4 weeks from baseline[T1] and 4 weeks after completion of the intervention[T2]) and the secondary were performed twice(baseline[T0] and 4 weeks from baseline[T1]).
From July 2019 to July 2020, a total of 50 patients were recruited in the outpatient and ward of the rehabilitation department of our hospital. Patient recruitment included the following steps: (1) a patient's attending doctor was required to be acquainted with the inclusion and exclusion criteria, as well as the screening of potential participants and recommending them to the primary researcher; (2) the patients were then judged on their clinical characteristics; (3) the primary researcher explained the trial aims to a potential patient and discussed the rehabilitation objectives with them; and (4) the patients or their relatives were required to sign informed consent forms.
Inclusion criteria for subjects were as follows: (1) meeting the diagnostic criteria for cerebral infarction or cerebral hemorrhage; (2) unilateral limb paralysis from the first stroke; (3) time of onset: 1–12 months after stroke; (4) aged 40–80 years; (5) involved lower limb motor function ≥ Brunnstrom grade III; (6) ability to walk at least 10 meters with or without auxiliary tools; (7) no other diseases or complications that might affect rehabilitation training; (8) stable vital signs.
Exclusion criteria were: (1) a history of any other additional diseases that could influence ambulation; (2) diabetic foot and peripheral neuropathy; (3) MMSE < 17 points; (4) severe communication impairment; and (5) other factors that might prevent participation in the experiment.
All patients underwent conventional gait training. During ambulation, sophisticated therapists performed manipulations to help patients suppress excessive muscle tension, stimulate muscle activity and promote normal movement patterns. In addition, patients received instructions about weight-shifting, involved limb weight-bearing, balance training and various intensive exercises, as functional activities (such as standing up from a chair, turning around, crossing obstacles). Both groups received 40-min training sessions each week for 4 weeks. Only patients in the experimental group were required to wear insoles for a minimum of 1 hour every day and recommended to continue wearing at 4 weeks follow-up after the completion of treatment, but it was not mandatory.
Customization processes of insoles
The process included relevant biomechanical assessments, prescription formulation and the manufacture of insoles (see Figure 1).
Step 1: The assessment mainly consisted of gait observation, plantar pressure test (F-Scan®, Techstrom, Korea), and the use of the Najjarine Assessment System (NAS).Gait observation required the therapist to observe condition of each segment of movement chain from the front, back and side respectively, especially the movement of the involved lower limb and foot .During the plantar pressure test, the patient stood naturally on the electronic plantar pressure plate and remained while static data were collected, then stepped over the plate to collect dynamic data. Gait observation and plantar pressure test can reveal neuromuscular abnormalities of the foot and ankle after central nervous system injury. With NAS, leg length, forefoot to rearfoot position and the calcaneal angle when standing, which reflects patients’ foot structure and the biomechanical status of lower limbs, can be obtained.
Step 2: Based on the results of above assessments, the individualized prescription of insoles could be determined for stroke patients with hemiplegia. A pair of prefabricated insoles, with a 5° lateral wedge, and an arch support made from high-density ethyl vinyl acetate (produced by Jiangsu Suyun Medical Equipment Co. Ltd, Jiangsu, China), were distributed to each patient in the experimental group. Typically, we added a 2° or 4° forefoot pad on the paretic forefoot to promote ankle dorsiflexion, a 4° or 6° inversion ramp pad to promote the paretic foot’s pronation movement and a metatarsal dome pad to alleviate paretic forefoot plantar pressure on the involved side. In addition, a 4° forefoot pad was added on the lateral of the non-paretic forefoot to increase the stability of the uninvolved foot. All the above pads were attached to the plantar surface of the insoles (see Figure 2).
Step 3：The producers chose and cut appropriate-size prefabricated insoles to suit patients’ shoes. All pads were pre-cut and attached to the insole with double-sided tape. After the insoles were made, each patient tried them on and appropriate adjustments and corrections were performed if necessary. Finally, the insoles were molded specifically by heating for 40 seconds with a heat gun and were then shaped by a qualified physiotherapist to maintain the subtalar joint in a neutral standing position. Furthermore, some modifications could be made according to the patient’s condition in the later period.
All of the above steps were performed by qualified therapists.
Tinetti Gait Scale (TGS) measurement was our primary indicator , which was used to evaluate the patient's gait. It consisted of 6 items, including 2 items related to coordinated gait components, 5 items related to compensation strategies and 1 item related to temporal aspects of gait. TGS ranged from 0 to 12 points, representing most deviations to normal.
- Plantar pressure test (%) : Items included weight-bearing of the involved foot (normal: 50%) and the involved forefoot (normal: 27.5%) during static standing, weight-bearing of the involved foot (normal: 50%) and the involved forefoot (normal: 35%) during walking, and gait cycle percentage (early stance phase, normal: 20%; mid-stance phase, normal: 40%; and late stance phase, normal: 20%).
- 6-min walking test (6MWT) : It was carried out indoors along a long, flat, straight enclosed corridor. The length of the walking track was 30 meters and subjects walked as fast as they could for 6 min, and the walking distance was measured.
- Lower Extremity Fugl-Meyer assessment (FMA-LE) : There were 17 items in this assessment, of which 2 items related to reflex activity, 11 items to synergistic movements and 3 items to coordination. The scoring of each item was based on a sequential score of 3 points (0, unable to complete; 1, partially completed; 2, completely completed), except for the 2 reflection items.
- Berg Balance Scale (BBS) : It was a list of 14 items, and each item was composed of a five-point ordinal scale from 0 to 4; 0 represented the lowest level of function and 4 the highest level of function.
- Barthel Index (BI) assessment : It covered 10 domains of functioning (activities): bowel and bladder control, as well as assistance with grooming, toilet use, feeding, transfer, walking, dressing, climbing stairs and bathing. Each activity had 5 dependency levels, ranging from 0 (unable to perform) to 5, 10 or 15 (completely independent).
The sample size calculation was detailed in the Appendix 1(Figure 3，4).
Data analysis was performed using SPSS Statistics (version 20.0, IBM, USA). Patients’ demographic and clinical characteristics, including age, gender, course and classification of stroke, involved side (left/right) and the Brunnstrom stage, are given by the number of cases (%) for categorical data and mean (SD) for continuous variables (Table 1). Tinetti Gait Scale was described as the number of cases (%) and the plantar pressure test, Lower Extremity Fugl-Meyer assessment (FMA-LE) and Berg Balance Scale (BBS), which did not conform to a normal distribution, as well as discontinuous variables such as the Barthel Index (BI) assessment, are presented as the mean (95% confidence interval (CI)). Continuous variables such as the 6-min walking test (6MWT) with a normal distribution are reported as the mean (SD).
Chi-square test was conducted for the comparison in the change of Tinetti Gait Scale (TGS). The Wilcoxon rank sum test was used to analyze differences between the two groups and the Wilcoxon signed-rank test was employed within the groups for discontinuous variables such as Tinetti Gait Scale (TGS) and Barthel Index (BI) assessments. Based on the assumption of normal distribution and homogeneity of variance, we conducted an independent sample t-test to analyze the difference in the change of 6-min walking test (6MWT) between the two groups. Continuous variables, such as the change in plantar pressure test, Lower Extremity Fugl-Meyer assessment (FMA-LE) assessment and Berg Balance Scale (BBS)，did not exhibit a normal distribution and Wilcoxon signed-rank test was used within the groups and Wilcoxon rank sum test between groups. The significant level was set as α = 0.05.