Neuromuscular electrical stimulation for the treatment of diabetic sensorimotor polyneuropathy: A prospective, cohort, proof-of-concept study

Objective: To assess a potential ef ﬁ cacy signal, safety and feasibility of neuromuscular electrical stimulation (NMES) therapy as an adjunct to standard care in patients with diabetic sensorimotor polyneuropathy (DSPN). Methods: In this single-centre, prospective, cohort, proof-of-concept study, 25 patients with DSPN consented to at least one daily 30-minute NMES therapy session (Revitive ® IX) for 10 weeks, with 20 patients completing the study. The primary outcome measure was nerve conductivity assessed using a nerve conduction study of the sural, super ﬁ cial peroneal, common peroneal and tibial nerves at 10 weeks compared to baseline. Secondary outcomes included super ﬁ cial femoral artery (SFA) haemodynamics during NMES therapy compared to rest and quality-of-life at 10 weeks compared to baseline. Results: At 10 weeks, there were signi ﬁ cant increases in sural sensory nerve action potential amplitude and conduction velocity ( p < 0.001), super ﬁ cial peroneal sensory nerve action potential amplitude ( p = 0.001) and conduction velocity ( p = 0.002), common peroneal nerve conduction velocity ( p = 0.004) and tibial nerve compound muscle action potential amplitude ( p = 0.002) compared to baseline. SFA volume ﬂ ow and time-averaged mean velocity signi ﬁ cantly increased ( p ≤ 0.003) during NMES compared to rest. Patient-reported Michigan Neuropathy Screening Instrument scores signi ﬁ cantly decreased ( p = 0.028) at 10 weeks compared to baseline. Three unrelated adverse events occurred, and 15 participants adhered to treatment. Conclusions: NMES therapy as an adjunct to standard care for 10 weeks signi ﬁ cantly increased lower limb nerve conductivity in patients with DSPN and may be bene ﬁ cial in the treatment of DSPN.


Introduction
Diabetes affects half a billion people globally and this is forecast to rise to 780 million by 2045 [47].Diabetic neuropathy is the most common complication, with a lifetime prevalence of more than 50% in people with diabetes [35].It typically manifests as symmetrical, lengthdependent diabetic sensorimotor polyneuropathy (DSPN), resulting in a variety of lower limb symptoms including neuropathic pain, dysesthesia, numbness, weakness and unsteadiness.These symptoms arise from a combination of neuronal and microvascular dysfunction [41].DSPN is an independent risk factor for neuropsychological disorders [36], falls [24], ulceration [54], amputation [1] and mortality [44], and is associated with considerable healthcare costs.In England, ulceration and amputation costs alone in people in diabetes account for an estimated £1 billion of the annual National Health Service (NHS) healthcare budget [23].
The management of DSPN is challenging, and many people suffer with chronic symptoms and develop further foot complications.Painful symptoms are often managed with anticonvulsants and antidepressants, despite their variable analgesic efficacy and associated side effects, and little is offered to patients experiencing other, non-painful symptoms [43].Enhanced glycaemic control has the highest level of evidence for the prevention and treatment of DSPN.However, a Cochrane meta-analysis of 17 randomised controlled trials found preventative effects were only significant in people with Type 1 diabetes and not in people with Type 2 diabetes [7].Enhanced glycaemic control can be achieved through a variety of interventions such as insulin, antidiabetic medications, pancreas transplant, bariatric surgery, self-monitoring of blood glucose (for example, via flash and continuous glucose monitors) and lifestyle modifications, but each has its own set of risks, including an increased risk of severe hypoglycaemic episodes [43].
Participation in supervised exercise programmes has been shown to significantly improve nerve conductivity, glycaemic control and symptoms in patients with DSPN and may have disease-modifying effects [46,4].However, these programmes are difficult to implement in underfunded healthcare systems with limited resources, and this is compounded by poor patient adherence [40].It is critical these programmes align with exercise advice for people with DSPN, which emphasise light to moderate activities such as walking, cycling, swimming and chairbased exercises and the avoidance of high-impact, strenuous or prolonged weight-bearing activities [12,2].Other non-pharmacological interventions include spinal cord stimulation, which is licensed for refractory neuropathic pain despite very low-quality evidence for its use in people with chronic pain [31], as defined by Grading of Recommendations Assessment, Development, and Evaluation (GRADE) guidelines [5].Similarly, transcutaneous electrical nerve stimulation (TENS) and frequency-modulated electromagnetic neural stimulation (FREMS) have been recommended for neuropathic pain, although there is insufficient evidence supporting their efficacy [55].
Neuromuscular electrical stimulation (NMES) is a novel, non-pharmacological, non-invasive adjunctive intervention for people with DSPN.It elicits transcutaneous electrical impulses, but unlike TENS, it elicits impulses of sufficient intensity (i.e. of significant amplitude, duration, shape) to depolarise neurons and cause involuntary muscle contraction.NMES may produce a physiological response comparable to voluntary exercise, improving physical fitness, muscle strength and blood circulation [8,33,15,3,37,51,49]. Technological advancements have enabled NMES to be elicited through small, user-friendly, portable devices that can be used at home while seated in a chair, offering similar muscle movements to recommended chair-based exercises without the need for supervision or travel required in supervised exercise programmes.
The objective of this study was to assess a potential efficacy signal, safety and feasibility of a NMES device as an adjunct to standard of care in patients with DSPN.It was hypothesised that 10 weeks of daily, home-based NMES therapy as an adjunct to standard of care would significantly increase lower limb nerve conductivity in people with DSPN by directly stimulating peripheral nerves and indirectly by increasing blood circulation.While the exact mechanisms are not fully understood, low-level mechanical or electrical stimulation of peripheral nerves may enhance their ability to detect and conduct weak signals [28].Additionally, muscle contraction produced by NMES may increase blood circulation, particularly at a microvascular level, which may contribute to peripheral nerve repair in people with DSPN [10].

Study design
This was a single-centre, prospective, cohort, proof-of-concept study sponsored by Imperial College London.It was approved by East Midlands -Nottingham 1 Research Ethics Committee (14/EM/1302) and the Health Research Authority in England [16].As the device is CE marked and indicated for people with diabetes, this study was classified as a post-market clinical trial, and Competent Authority approval was not required.

Setting
This study took place at Charing Cross Hospital, Imperial College Healthcare NHS Trust in London, United Kingdom (UK).Participants were recruited from diabetic foot, vascular surgery and neurophysiology clinics.

Participants
Adults (≥18 years) with a diagnosis of diabetes according to the 2006 World Health Organisation (WHO) definition [52], presenting with DSPN based on a score of ≥7 on the validated Michigan Neuropathy Screening Instrument (MNSI) questionnaire [13] and exhibiting moderately controlled blood pressure, as judged by the investigator, were eligible to participate.Blood pressure requirements were specified as a precautionary measure, to address any safety concerns related to application of NMES therapy in people with hypertension, though recent evidence suggests NMES can be safely applied without affecting blood pressure [19].
Patients were screened for other non-diabetic causes of neuropathy through blood tests for folate, vitamin B 12 , thyroid function, syphilis and human immunodeficiency virus (HIV), and by taking a history of their alcohol consumption.If other causes of neuropathy were found to be positive, patients were excluded from the study (screening failures), given appropriate counselling and their general practitioner (GP) was notified.A full list of pre-defined inclusion and exclusion criteria are listed in Table 1.

Sample size
The target sample size was 20 participants.As this was a proof-ofconcept study, a power calculation was not performed.For the study objective, the chosen target was realistic to achieve in the study setting and to provide pilot data to assess a potential efficacy signal, feasibility and safety of a NMES device as an adjunct to standard of care in people with DSPN.

Neuromuscular electrical stimulation (NMES) device
The Revitive® IX (Actegy Ltd, Bracknell, UK) was the NMES device under investigation (Fig. 1).It is a CE marked, class IIa medical device that provides NMES via large conductive rubber footpad electrodes.It is intended to increase blood circulation in the legs, ankles and feet and is commercially available to the general public.In particular, it is indicated to improve blood circulation in people with diabetes, where blood circulation problems are common.
The device can be switched on or off, and the intensity and duration of the NMES can be adjusted using the remote control or by manually pressing the control buttons on the device.Intensity ranges from level 0 to 99 and a standard NMES programme is 30 min in duration.Each programme consists of 10 unique waveforms, arranged in a sequence of 15 waveforms that are repeated twice.The frequencies of the waveforms range from 1 to 50 Hz and the pulse durations of the waveforms range from 370 to 940 μs.A wide pulse pattern of 940 μs is intended to allow for venous refilling.The maximum current output of the device is 15 mA at 500Ω resistance.The on-and-off duration of NMES is waveformdependent and ranges from 1.9 to 7.1 s and 1.0 to 1.5 s, respectively (approximate on-and-off ratio ranges from 2:1 to 14:3).The IsoRocker feature can also be enabled during NMES therapy which allows the Revitive® IX to tilt back and forth as lower limb muscles contract and relax.This accentuates natural heel toe raises and the calf muscle pump at the ankle.
At the Week 0 (baseline) study visit, participants were instructed to place the device on the floor, sit with their knees bent at a 90-degree angle and place their bare feet on the device to allow NMES to be applied to the plantar surface of the feet (Fig. 1B).The intensity level was set to 1 and it was systematically increased in increments of 1 until the sensory (the intensity level which stimulation was first felt by the participant) and the motor (the intensity level which calf muscle contraction was first observed by the assessor) thresholds were established.The sensory threshold relied on participant verbal sensory feedback, whereas the motor threshold was verified by the investigator as the first visible muscle twitches of the medial gastrocnemius.
Participants were asked to use the device at the intensity level that was twice their individual motor threshold, or as high as they could tolerate while producing visible but not painful lower limb contractions.They were told to use the device at home for at least one 30-minute session (but no more than six sessions) per day for an intervention period of 10 weeks and to record their device usage in a self-report patient diary.There were no time constraints on when they could use the device, but time of usage was recorded in the diary.As a result, per-protocol treatment adherence required a minimum of 70 sessions during the intervention period.Halfway through the intervention period at 5 weeks participants were invited back to the research clinic to see how they were tolerating the device.

Standard of care
Standard of care was not 'standardised' but defined as the treatment available locally at the clinical trial site for DSPN, such as glycaemic control, lifestyle modifications and pain pharmacotherapies.

Outcomes
The primary outcome measure was a change in lower limb nerve conductivity assessed using a bilateral nerve conduction study of the sural, superficial peroneal, common peroneal and tibial nerves at 10 weeks compared to Week 0 (baseline).All nerve conduction studies were undertaken at Week 0 and Week 10 by a single, independent consultant neurophysiologist using the Dantec® Keypoint® System (Optima Medical, London, UK).Nerve conduction parameters measured included conduction velocity, sensory nerve action potential (SNAP) amplitude, compound muscle action potential (CMAP) amplitude, minimum F wave latency and H reflex. Conduction velocity was calculated using distance and latency.
Nerve conduction studies were conducted in accordance with a local laboratory protocol informed by the British Society for Clinical Neurophysiology guidelines [48].The timing of the test appointments (e.g.morning or afternoon) were determined pragmatically based on participant and neurophysiologist availability.Participants were advised to eat and drink as normal on the day of testing, as these factors are not known to significantly influence nerve conduction study results [21].Before testing, participants assumed a supine position, and the temperature of their feet was measured using a thermometer.In instances where the temperature fell below 30 °C, warm water immersion for a minimum of 5 min was employed to bring the feet to the required temperature, followed by a return to the supine position.Subsequently, NuPrep® gel (Weaver and Company, Colorado, United States of America) was used to clean the areas where the stimulating and recording electrodes were to be placed.Handheld bipolar electrodes were then sequentially placed over the sural, superficial peroneal, common peroneal and tibial nerves with recording electrodes placed over the corresponding muscle.Stimulation parameters were adjusted to elicit a response, which was displayed on a computer monitor as nerve conduction parameters with corresponding waveforms.Multiple stimulations and recordings were performed to ensure accurate and reproducible results.
Secondary outcome measures included a change in haemodynamics during NMES therapy compared to rest.Haemodynamics were not expected to change across the 10-week intervention period but were measured before and during a 30-minute NMES therapy session at both Week 0 (baseline) and Week 10.Haemodynamic testing took place after nerve conduction studies and similarly had no fasting requirements.Superficial femoral artery (SFA) haemodynamics were assessed using duplex ultrasonography.A Philips iU22 ultrasound machine (Philips Electronics UK, Guildford, UK) was programmed with a vascular protocol and a probe was positioned at a 60°angle to measure SFA diameter, time-averaged mean velocity and volume flow in the lower limb.Rest measurements were taken after participants assumed a seated position for at least 15 min and then subsequent measurements during NMES therapy were collected every 5 min of the 30-minute session and mean values were recorded.Additionally, foot and hand haemodynamics were assessed using laser doppler flowmetry.A dual-channel moorVMS-LDF device (Moor Instruments, Axminster, UK) with two probes was used to measure skin temperature and microvascular blood perfusion (flux) on the dorsum of the foot and the hand.Similar rest measurements were taken, and then continuous measurements were taken during the full 30-minute NMES therapy session and mean values were recorded.
Changes in disease-specific quality-of-life were assessed using the MNSI (both the Part A questionnaire and Part B physical examination) and Problem Areas in Diabetes (PAID) questionnaire at 10 weeks compared to Week 0 (baseline).Changes in general quality-of-life were assessed using the 5-level EQ-5D version (EQ-5D-5 L and EQ-5D-VAS) and Short Form-36 (SF-36) at 10 weeks compared to Week 0 (baseline).Device experience, adverse events (AEs) and adherence to treatment were also assessed at 10 weeks.Device experience was recorded on a 10-point numerical rating scale, with '1' denoting 'dissatisfied' and '10' denoting 'very satisfied' (Appendix A).AEs were recorded in study case report forms and adherence to treatment was recorded in self-report patient diaries.

Statistical analyses
All participants who were confirmed as eligible at screening were included in the statistical analyses, except for two patients who specifically requested to be withdrawn and for their data not to be used in the study unless it was related to safety (for more information, see Recruitment and Safety sections).Raw and scored data were entered into an Excel® database (Microsoft, Redmond, Washington, United States of America) and basic statistical analyses were run.R Statistical Software (V4.2.2, R Core Team 2022, Vienna, Austria) was used to perform further statistical analyses, specifically the mice (multivariate imputation by chained equations) package (V3.15.0) was used for multiple imputation.
Visual inspection was used to determine normality of data and confirmatory Shapiro-Wilk tests were performed.Summary statistics such as n, mean, standard deviation (SD), median, interquartile range (IQR) were presented where appropriate.Paired parametric data were analysed using t-test statistics and paired non-parametric data were analysed using Wilcoxon-signed rank tests.A parametric sensitivity analysis was performed on the primary outcome data to assess the robustness of the results.
There were no missing primary outcome data, but there were some missing secondary outcome data (haemodynamic and quality-of-life).Multiple imputation using chained equations was performed to handle this because the data were assumed to be missing at random (MAR).This technique can be reasonably accurate for small sample sizes [6,26,14], but it has the potential for bias [6,27].Specifically, missing values for duplex ultrasonography data at Week 0 and Week 10 were imputed, generating 15 and 10 imputed datasets, respectively.Missing values for laser doppler flowmetry data at Week 0 and Week 10 were imputed, generating 15 datasets each, while missing quality-of-life data at Week 10 generated 10 imputed datasets.Considering the potential for bias, sensitivity analyses were performed on the original datasets with missing data to further assess the robustness of the results.Except for when multiple hypothesis tests were performed and a Bonferroni correction was used, statistical significance was set at p ≤ 0.05.

Recruitment
Between April and October 2015, 26 patients were pre-screened for study eligibility and 25 patients consented to participate in the study.One participant was not consented because they did not have neuropathy and three participants were excluded at screening because they showed signs of neuropathy that was not caused by diabetes (screening failures).No potential participants were excluded based on their blood pressure.Two participants withdrew from the study before completion (Week 10) because of AEs classified as unrelated to the study device − for more information, see Safety section.These two participants requested that only their safety outcome data be reported, with all other outcome data excluded from analyses.As a result, 20 participants completed the study with primary outcome data (Fig. 2).Overall, participants were considered elderly, overweight and had elevated HbA1c and blood pressure.The majority of participants were male, had previously smoked and had a history of peripheral arterial disease (Table 2).

Lower limb nerve conductivity
The normality of nerve conduction parameters was tested using Shapiro-Wilk tests.The results revealed a departure from normality for all parameters (p < 0.05, Appendix B Table B.1). Parameters were presented as median (interquartile range) and Wilcoxon-signed rank tests of paired data were performed.There were significant increases in sural nerve SNAP amplitude and conduction velocity (both p < 0.001), superficial peroneal nerve SNAP amplitude (p = 0.001) and conduction velocity (p = 0.002), common peroneal nerve conduction velocity (p = 0.004) and tibial nerve CMAP amplitude (p = 0.002) after 10 weeks of NMES device use compared to Week 0 (baseline).The greatest increases in nerve conduction parameters between Week 0 and Week 10 were observed in the sural nerve (Table 3).A parametric sensitivity analysis, presenting parameters as mean (standard deviation) and employing t-test statistics, affirmed significant increases in the sural nerve SNAP amplitude and conduction velocity (both p < 0.001), superficial peroneal nerve SNAP amplitude and conduction velocity (both p < 0.001) and tibial nerve CMAP amplitude (p = 0.002) after 10 weeks of NMES device use compared to Week 0. The sensitivity analysis did not detect a significant increase in common peroneal nerve conduction velocity (p = 0.009) but did detect a significant increase in H reflex (p = 0.004) after 10 weeks of NMES device use compared to Week 0 (Table 4).

Safety
Three AEs were reported: back pain, a bleeding varicose vein and a ruptured Baker's cyst.The investigators did not classify these AEs as serious adverse events (SAEs) or related to the study device.Two participants withdrew from the study before completion (Week 10) due to back pain and a bleeding varicose vein (Fig. 1).

Feasibility
Adherence data were available for 17 (85%) of the 20 participants who completed the intervention period, with three participants not returning their patient diaries.Fifteen participants (75%) adhered to the treatment as per-protocol (a minimum of 70 30-minute NMES sessions across the 10-week intervention period).During the intervention period, the mean total number of hours spent using the study device was 116.85 and the mean total number of sessions was 198.65, which was more than double the minimum requirement.For those who completed the device experience numerical rating scale, the overall experience of the study device was positive, with a mean device satisfaction score of 8.31 (scores range from 1 to 10, with higher scores indicating more satisfaction), but 7 participants (35%) had missing data.

Haemodynamics
The normality of haemodynamic parameters was tested using Shapiro-Wilk tests.The results revealed a departure from normality for time-averaged mean velocity and volume flow of the SFA as well as foot temperature and foot flux (p < 0.05, Appendix B Table B.2).These parameters were presented as median (interquartile range) and    6).

Quality-of-life
The normality of quality-of-life parameters was tested using Shapiro-Wilk tests.The results revealed a departure from normality for MSNI (Part B examination), PAID and EQ-5D-VAS questionnaires (p < 0.05, Appendix B Table B.3).These parameters were presented as median (interquartile range) and Wilcoxon-signed rank tests of paired data were performed.For all other parameters, where no evidence suggested a deviation from normal distribution, they were presented as mean (standard deviation) and paired t-tests were performed.As shown in Table 7, there was a significant decrease in MNSI (Part A questionnaire) scores (p = 0.028) after 10 weeks of NMES device use compared to Week 0 (baseline).No significant changes were observed in the remaining questionnaires.The sensitivity analysis with original datasets affirmed a significant decrease in MNSI (Part A questionnaire) scores (p = 0.028) and no significant changes in the remaining questionnaires (Table 8).

Main findings
This single-centre, prospective, cohort, proof-of-concept study demonstrated that daily use of a NMES device as an adjunct to standard of care for 10 weeks significantly increased lower limb nerve conductivity in people with DSPN, as measured by a nerve conduction study.Specifically, NMES improved sensory nerve function, with both the SNAP amplitudes and conduction velocities of the sural and superficial nerves increasing significantly.These findings may suggest greater axonal integrity and myelination, as well as an increase in the number and synchrony of sensory axons [30].The recruitment of sensory neurons induced by NMES is atypical, given its traditional activation of motor units.Yet this distinctive outcome may be explained by the inclusion of a wide pulse (940 μs) waveform within the Revitive® IX program.Previous studies have indicated that wide pulse patterns may depolarise sensory neurons more effectively than motor neurons [9,32,11], highlighting the potential capabilities of NMES programs.
In the motor nerves, the primary analysis demonstrated a significant increase in the conduction velocity of the common peroneal nerve, which may suggest an improvement in axonal integrity, myelination and neuromuscular junction integrity at the common peroneal nerve.This finding, however, was not replicated in the parametric sensitivity analysis, so should be interpreted with caution.Common peroneal nerve conduction velocity has been linked to glycaemic control [7], however HbA1c was only assessed at Week 0 (baseline), thus it is unclear whether this had an effect.Additionally, the CMAP amplitude of the tibial nerve significantly increased in both the primary and sensitivity analyses, which may suggest an increase in the number and synchrony of motor axons, neuromuscular junctions and muscle fibres at the tibial nerve [30].
Three adverse events were reported in total, resulting in two participants withdrawing from the study.Despite this, the investigators classified no AEs as serious or related to the study device (Revitive® IX), and the device was considered low risk for this patient group, provided that contraindications were followed.The Revitive® IX may lead to skin irritation beneath the stimulation electrodes or hypersensitivity owing to NMES, but no participants reported any of these problems.Overall, the treatment protocol was deemed feasible with moderate overall adherence, and many participants exceeded the minimum required number of sessions during the intervention period.
The primary analysis revealed that time-averaged mean velocity and volume flow of the SFA significantly increased during NMES therapy compared to rest at both Week 0 and Week 10.However, the statistical significance observed in the time-averaged mean velocity at Week 10 was not replicated in the sensitivity analysis.Therefore, caution should be exercised in interpreting this particular result.These findings may suggest that NMES transiently enhances arterial blood flow.This may be facilitated by the calf muscle pump, which exerts energy creating a demand for oxygen and nutrients that can be supplied by arterial inflow in the lower limb [3].Furthermore, foot flux significantly increased during NMES therapy compared to rest at both Week 0 and Week 10.This may suggest that NMES transiently promotes blood microcirculation.Microvascular disease has been identified as a distinct driver of DSPN pathogenesis and has been shown to be associated with important clinical outcomes, such as nerve conductivity and pain [42].It has been previously postulated that as few as 10 transcutaneous electrotherapy sessions in people with DSPN may have the ability to enhance circulation at the vasa nervorum and promote nerve function, however further clinical and biochemical studies are needed to test this hypothesis [10].
The MNSI (Part A questionnaire) score, the only quality-of-life measure included which focused on neuropathy, significantly improved over the 10-week intervention period in both the primary and sensitivity analyses to below the threshold for a positive neuropathy screening test [35].While this signifies a clinically meaningful change associated with considerable symptomatic relief, it is crucial to acknowledge that, given the uncontrolled nature of the study, the extent to which these improvements may be attributed to a placebo effect remains uncertain.Moreover, it has been proposed that the threshold for a positive neuropathy screening test should be decreased to a more sensitive score of 4 [17], and notably, the mean Week 10 score in this study met this threshold.

Strengths and limitations of the study
One of the main strengths of this study was the objective primary outcome measure, which involved measuring lower limb nerve conductivity using a nerve conduction study, which was conducted by a single independent consultant neurophysiologist.This is particularly important because nerve conduction studies are known to have high inter-assessor variability [53].Although this individual was not blinded to treatment allocation, blinding should be considered to strengthen future studies.Furthermore, the study achieved its target sample size, with only two participants withdrawing due to unrelated AEs.
A limitation of this study was that it did not control for treatment effects associated with standard of care, including lifestyle modifications and pain pharmacotherapies.As outlined in the study protocol, standard of care was defined as the treatment available locally at the clinical trial site for DSPN.All participants were recruited from a single site, so a consistent level of standard of care can be assumed, albeit potentially varied in treatment modalities.It is recommended that future studies record standard of care at baseline and any changes during the intervention period and actively control for associated effects.Another limitation was that adherence was measured using a patient diary and was therefore vulnerable to self-report bias.Future studies may consider monitoring adherence through a device use tracker or application.
In addition, this study did not include a measure of small fibre function, as nerve conduction studies only assess large fibre function.Given that NMES had a significant effect on large fibre function, it would be of interest to investigate the effect on small fibres, as damage to these fibres have been linked to pain and other sensory abnormalities [34].Assessing small fibre function could provide a more complete understanding of the effect of NMES on DSPN.This could be addressed in future studies by incorporating quantitative sensory testing or a nerve biopsy to assess intraepidermal nerve fibre density.
This study was conducted in a single-centre, it was uncontrolled and no power calculations were performed to determine the sample size, therefore the true treatment effect is uncertain, and the findings may not be generalisable to other settings.However, these limitations are inherent to a small pilot study, and the significant effects measured suggest that a larger, multi-centre, randomised, sham-controlled trial should be considered.The absence of a control arm was largely due to the unavailability of a suitable sham device.However, a current trial is underway to compare the efficacy of a NMES device to a newly developed sham device in people with DSPN [22].

Comparisons with other studies
To the best of the authors' knowledge, this is the largest study to date assessing a potential efficacy signal of a NMES device as an adjunct to standard of care in patients with DSPN, with lower limb nerve conductivity as the primary outcome measure.There has been one previous randomised controlled trial conducted by Williams et al. (as part of a PhD thesis) [50], which included complete primary outcome lower limb  PAID scores range from 0 to 100 points, with higher scores indicating more diabetes distress.EQ-5D-5 L scores range from 0 to 1, with higher scores indicating better health.EQ-5D-VAS scores range from 0 to 100, with higher score indicating better health.SF-36 PCS and MCS scores range from 1 to 100, with higher scores indicating better quality of life.y Median (interquartile range).
A Wilcoxon-signed rank test of paired data.B Paired t-test.
* Statistical significance was set at p ≤ 0.05.
nerve conductivity data for six participants.In contrast to this study, no significant differences were found within the 'NMES device' and 'no device' groups at 10 weeks for any of the nerve conduction parameters measured in the sural, superficial peroneal, common peroneal and tibial nerves, likely owing to the small number of participants in each group.Also, the NMES therapy was set to a minimum threshold to induce muscle twitching in the foot, whereas in this study, the intensity that caused contraction of the medial gastrocnemius was recorded, and participants were encouraged to use the NMES device at twice this level.Having said that, the Williams et al. [50] protocol required at least 4 h of NMES therapy per day, which was far longer than in this study.However, compliance data was not reported, so it is unclear whether participants met this requirement [50].This study adds to the body of evidence that NMES may improve neuropathy symptoms in people with DSPN [18,20,25,39].However, akin to this study, much of the research in this area has been uncontrolled, apart from one randomised controlled trial, which found a significant improvement in neuropathy symptoms after NMES therapy when compared to TENS therapy [39].As the field of NMES therapy advances, future studies should incorporate sham-controls or, at the very least, comparative studies with other neuromodulation devices to enhance the rigour of their findings.There has been more research into TENS therapy in people with DSPN, but these have primarily focused on pain outcomes.A meta-analysis that included 12 studies revealed a statistically significant reduction in pain scores with TENS therapy (mean difference −0.44; 95% CI −0.79 to −0.09), but the methodological quality of the studies was generally low [45].Pilot, randomised, sham-controlled studies have demonstrated significant improvements in postural and protective sensation measures in people with DSPN after 6 weeks of TENS therapy, possibly through increased perfusion [28,29].These outcomes may be crucial to consider in future investigations of NMES in people with DSPN.
The haemodynamic observations in this study align with the literature that NMES transiently increases blood circulation [3,49,38,37].In healthy individuals, it has been shown that NMES therapy (Revitive® IX) significantly increases venous and arterial haemodynamics (venous blood flow p = 0.014; venous time-averaged mean velocity p = 0.065; arterial blood flow p < 0.001; arterial time-averaged mean velocity p < 0.001) compared to rest.Improvement in arterial haemodynamics following NMES therapy has also been demonstrated in people with lower limb pathology.SFA blood flow and time-averaged mean velocity significantly increased during the first use of NMES therapy compared to rest in people with peripheral arterial disease (p < 0.02) [3].

Conclusion
After 10 weeks, a NMES device as an adjunct to standard of care significantly increased lower limb nerve conductivity in people with DSPN, specifically in the sural, superficial peroneal, common peroneal and tibial nerves.NMES may be beneficial in the treatment of DSPN, but further high-quality research in the form of randomised, sham-controlled trials is needed to better understand its effects.

Table 1
Inclusion and exclusion criteria.

Table 2
Baseline characteristics of study participants.
BPM, beats per minute; SD, standard deviation.Wilcoxon-signed rank tests of paired data were performed.For all other parameters, where no evidence suggested a deviation from normal distribution, they were presented as mean (standard deviation) and paired t-tests were performed.As shown in Table5, time-averaged mean velocity and volume flow of the SFA increased significantly during NMES device use compared to rest at both Week 0 and Week 10 (p < 0.001, except for Week 10 volume flow where p = 0.003).There were no significant differences in SFA diameter during NMES device use

Table 5
Haemodynamic parameters at rest and during neuromuscular electrical stimulation at Week 0 and Week 10 (n = 20).

Table 3
Nerve conduction parameters at Week 0 and Week 10 after NMES device use (n = 20).

Table 4
Sensitivity analysis of nerve conduction parameters at Week 0 and Week 10 after NMES device use (n = 20).to rest.There were significant increases in foot flux and hand temperature during NMES device use compared to rest at both Week 0 (p < 0.001) and Week 10 (p < 0.001), but not in foot temperature nor hand flux.The sensitivity analysis with original datasets affirmed a significant increase in time-averaged mean velocity of the SFA during NMES device use compared to rest at Week 0 (p = 0.004), but not at Week 10.Additionally, it affirmed a significant increase in volume flow of the SFA, foot flux and hand temperature during NMES device use compared to rest at both Week 0 and Week 10 (p ≤ 0.003) (Table * Paired t-test; statistical significance was set at p ≤ 0.004 (Bonferroni correction).compared

Table 6
Sensitivity analysis of haemodynamic parameters at rest and during neuromuscular electrical stimulation with missing data at Week 0 and Week 10 (n = 20).

Table 8
Sensitivity analysis of disease-specific and generic patient-reported quality-of life-outcomes at Week 0 and Week 10 after NMES device use (n = 20).
y Median (interquartile range).§ Mean (standard deviation).A Wilcoxon-signed rank test of paired data.B Paired t-test.* Statistical significance was set at p ≤ 0.05.

Table 7
Disease-specific and generic patient-reported quality-of life-outcomes at Week 0 and Week 10 after NMES device use (n = 20).