The effects of a physiotherapist-led exercise intervention on peripheral neuropathy among people living with HIV on antiretroviral therapy in Kigali, Rwanda

Background HIV-associated peripheral neuropathy (PN) is common in people living with HIV. Its management is mostly symptomatic utilising pharmacological approaches. Objectives This study determined the effects of an exercise intervention on PN among Rwandan people living with HIV receiving antiretroviral therapy (ART). Methods A 12-week single-blinded randomised controlled trial using the Brief Peripheral Neuropathy Screen (BPNS) as the assessment tool tested the effects of an exercise intervention on PN, followed by a 12-week non-intervention period. A total of 120 people with HIV- associated PN on ART were randomised to an exercise or no exercise group. Both groups continued receiving routine care. A bivariate analysis using Pearson’s chi-square test for significant differences in PN symptoms and signs, between groups, at baseline, after the 12 weeks intervention and 12 weeks post-intervention using generalised linear regression models to determine predictors of treatment outcomes was undertaken, utilising an intention-to-treat analysis (alpha p ≤ 0.05). Results At 12 weeks, the intervention group compared to the control: neuropathic pain 70% versus 94% (p < 0.005), PN symptoms severity – mild and/or none in 85% versus 60% (p < 0.001) and radiation of PN symptoms reduced, 80% versus 37% (p < 0.001). There were no differences in PN signs at 12 weeks intervention and at 12 weeks post-intervention. Having changed the antiretroviral (ARV) and having developed PN symptoms after the start on ARVs predicted treatment improvement, while demographic factors did not predict any treatment outcome. Conclusion A physiotherapist-led exercise intervention improved PN symptoms, but with non-significant improvement in PN signs. Factors related to early diagnosis and treatment of PN were facilitators for the improvement of PN symptoms. Clinical implications Physiotherapist-led exercises should be integrated into the routine management of people living with HIV on ART with PN symptoms.

more than placebo. However, the authors concluded that rhNGF is not clinically available and smoked cannabis cannot be recommended for routine treatment, thus new strategies for managing HIV-PN are required (Phillips et al. 2014).
Non-pharmacological management that includes exercise programmes and lifestyle changes is recommended (Ahmad & Goucke 2002). However, most of the non-pharmacological management benefits described are for diabetes-related PN, and these include nutritional supplements, warm water footbaths, rest, elevation of limbs, massage and exercises (Taylor et al. 2007). There is limited literature on the effects of these non-pharmacological modalities, particularly exercise for the management of HIV-PN. The available literature on exercise and PN showing beneficial results for diabetesrelated PN (Adeniyi et al. 2010) describes strengthening and balance training incorporated into aerobic exercise programmes (Tofthagen, Visovsky & Berry 2012;White, Pritchard & Turner-Stokes 2004). Thus, it is hypothesised that exercise would also be effective in the prevention or reduction of PN symptoms and signs because of various causes, including HIV-PN (Dobson, McMillan & Li 2014).
The existing evidence shows that exercises are safe and affordable in PLHIV Nixon et al. 2005;O'Brien et al. 2010). The systematic review by Nixon et al. (2005) examined the safety and effectiveness of exercise among PLHIV. They indicated that 'exercise appears to be safe and may lead to significant reductions in depressive symptoms and potentially clinically important improvements in cardiopulmonary fitness' (Nixon et al. 2005:2). Similar findings were identified in a systematic review by O'Brien et al. (2010). Mutimura et al. (2008) found that an exercise intervention improves components of quality of life (QoL) of PLHIV ).
There is a paucity of reported effects of aerobic exercise interventions specifically for PLHIV with PN. Hence, it became important to test the effects of a physiotherapist-led exercise (PTExs) intervention in a randomised controlled trial (RCT). Thus, the purpose of this study was to determine if an exercise intervention has an effect on PN among PLHIV on ART in Rwanda, given that exercise is a simple inexpensive intervention that could potentially be rolled out to PLHIV to improve PN.

Study design, site, participants and selection
This RCT included participants who attended four randomly selected ART clinics situated in health centres and hospitals in Kigali, Rwanda. The exercise intervention was provided to the participants at the Kigali Health Institute's physiotherapy clinic and fitness centre. Included in the study were male and female PLHIV on ART aged 18 and above; thus, the participants in the study were people already diagnosed with HIV at the HIV Voluntary Counselling and Test programmes at health centres and hospitals in Rwanda. The study recruited those with PN and on ART from the ART clinics.
According to the American College of Sports Medicine guidelines for exercise prescription for PLHIV (Sonya 2006), participants were excluded if they had active opportunistic infections, a clinical history contraindicating exercise (Ortiz 2014), could not walk without support, had musculoskeletal and central neurological impairments, a known history of diabetes or substance abuse, vitamin B12 deficiency, tuberculosis, renal failure, hypothyroidism and any other pathologies associated with neuropathy.
The sample considered all 164 PLHIV who had been randomly screened in a prevalence study on PN prior to this trial, by the same authors (Tumusiime et al. 2014). Hence, potential participants were consecutively invited from the 164 PLHIV with PN on ART. Out of the 164 participants, 139 responded positively to the invitation: 19 were excluded based on the above-mentioned exclusion criteria for this study and 120 were then eligible to participate. Sample size computation was done to check if the 120 had sufficient power to detect a difference in PN between the two groups (effect size; p = 0.05). The calculation was done by allowing for a dropout of a maximum of 30% per group based on previous similar studies such as that of Mutimura et al. (2008). With this consideration, the sample size of 120, randomised to 60 per group, had 80% power to detect the effect.
A total of 120 participants were assessed at baseline and were then block randomised into 60 in the experimental and 60 in the control group, using computer-generated random numbers. Concealed allocation to groups was assured by an independent research assistant placing the numbers into opaque sealed envelopes which were then given to eligible participants by a second research assistant. The study was single blinded, the assessor (first author) being blinded to group allocation, until after the last assessments.
The experimental group continued receiving Routine Health Care (RHC), antiretrovirals (ARVs) and other prophylactic medications, namely analgesics, antidepressants, multivitamins, as well as routine medical consultations, CD4 testing and counselling services, from the respective ART clinics plus physiotherapy-led exercises (RHC + PTExs). The control group received only RHC.
The PTExs comprised an exercise intervention that was identified from the literature O'Brien et al. 2010;White et al. 2004) and was recommended as safe and beneficial to PLHIV. The PTExs consisting of aerobic group exercises were given to the experimental group in a well-organised gymnasium. The group included both female and male participants. The intervention included a 15-min warm-up of walking (slow to brisk walk plus full range of upper and lower limb flexibility exercises); 15 min of mobility training with self-stretching (mainly dynamic stretching) in standing, lying and long sitting; 10 min of muscle conditioning with isometric exercises in various starting positions; 10 min of balance exercises and finally 10 min of cooling down (stretching and deep breathing).
Participants' blood pressures and respiratory rates were carefully monitored before and after the exercise sessions based on standard procedures as used in a similar study . Each session was 60 min in duration, offered 3 times a week for 12 weeks. In addition, participants were educated on how to exercise on their own at home. This was so that after the supervised exercise sessions of 12 weeks, participants would continue exercising on their own at home. They were assessed at baseline, after the 12 weeks of the intervention and then again after a further 12 weeks. A qualified physiotherapist, trained by the first author, conducted the exercise programme. Participants were reminded about the programme telephonically at regular intervals, and monthly calls were made to the control group by a designated research assistant after the baseline assessments to remind them of the assessments at 12 and 24 weeks after the start of the study. Assessments were done by the blinded first author. The BPNS which had been adapted in a study previous to this one (Tumusiime et al. 2014) and a checklist for demographic and health status characteristics were used to evaluate the PN.

Measures
The BPNS assessed the primary outcome of interest, namely the PN. The Brief Peripheral Neuropathy Screen is both a self-report and objective PN screening tool, which is a valid and reliable instrument for diagnosis of PN, specifically DSP (Cherry et al. 2005), with a specificity of 88% and sensitivity of 78% (Cherry et al. 2005). The BPNS examines subjective and objective outcomes consistent with PN. It has been used in several clinical trial protocols, in particular by the AIDS Clinical Trial Group (ACTG) (Evans et al. 2011), and evaluated in large-scale studies. It was identified to precisely discover PLHIV who have the highest grade of peripheral nervous system dysfunction and pathology. Cherry et al. (2005) concluded that the presence of both symptoms and signs on the BPNS provides a useful operational criterion for HIV-PN in the era of ART. One of the studies by Mehta et al. (2011) implemented the validated BPNS in patients receiving ART in Mombasa, Kenya, and they recommended that the tool can be easily integrated into routine care by general practitioners in an outpatient HIV clinic with limited resources. The tool was, therefore, realistic to use in a resource-limited country like Rwanda because of the high cost and lack of specialised instruments for nerve conduction, such as electromyography or sural nerve testing and electrophysiological changes. However, the tool was piloted for adaptation to assess PN among PLHIV on ART in Rwanda (Tumusiime et al. 2014a).
As described in the above study (Tumusiime et al. 2014a), BPNS is both a self-report and objective PN screening tool. It is both a valid and reliable instrument for diagnosis of PN specifically DSP, where 'PN is defined as the combination of at least one subjective neuropathy grade greater than 0 and either reduced or no sense of vibration and/or reduced/no ankle deep tendon reflex, bilaterally' (Tumusiime et al. 2014b). At both baseline and post-intervention, the participants were requested to score the 'presence and severity of symptoms, using a scale of 1 (mild) to 10 (severe) for each leg separately'. The symptoms assessed were as follows: 'pain, aching, or burning in feet and/or legs; pins and needles in feet and/or legs; and numbness in feet and/or legs'.
The single highest score of the six scores (three for each leg) was transformed to a: [S]ubjective PN grade as follows: symptoms absent = grade 0, score of 1-3 = grade 1, score of 4-6 = grade 2, and score of 7-10 = grade 3. (Mehta et al. 2011:491) Symptoms had to be bilateral to be graded as ≥ 1. 'Objective findings (signs) included in the BPNS are the loss of the sense of vibration and abnormal ankle deep tendon reflexes'. The vibration sense perception was assessed with a 128-Hz tuning fork, 'maximally struck and applied at the great toe distal interphalangeal joint of each foot'. The vibration sense was defined as 'normal for a vibration felt for > 10 s, mild loss for a vibration felt for 6-10 s, moderate loss for a vibration felt for ≥ 5 s, and severe loss for no feeling of vibration'. The ankle tendon reflexes were defined as 'absent, hypoactive, normal, hyperactive, or clonus'.
For testing of the ankle reflex, the participant was in a sitting position and the assessor (first author) used one hand to place the participant's foot into dorsiflexion at 90°. Holding the reflex hammer in the other hand, the assessor struck the ankle tendon. The tendon reflex was felt in the assessor's hand as planter flexion of the participant's foot. Improvement was determined as the differences in positive change between the two groups, that is the reduced percentages of people with PN, which was considered as an improvement in PN.
A demographic and health status questionnaire was utilised to obtain basic data from the participants.

Data analysis
This included tabulations for frequencies and percentages for categorical variables and measures of central tendency (means) and measures of variability (standard deviations) for continuous variables. A bivariate analysis using Pearson's chi-square test for significant differences between the experimental and control groups, at baseline, after 12 weeks of PTExs and at 12 weeks post-intervention, was done. Because we were interested in overall group changes over time, we adopted generalised estimating equations (GEE) from the generalised linear models with a logit canonical link function to identify the significant predictors of treatment outcome. The predictors tested included the demographic and health status-related characteristics: age, gender, level of education, marital status, occupation, duration since HIV diagnosis, CD4 cell count levels, duration on ART, treatment regime (ARVs), ARV regimen changes and whether PN started before or after the start on ARVs. Adjusted odds ratios and their 95% confidence interval (95% CI) values were http://www.sajp.co.za Open Access reported on magnitude of association and prediction on PN as binomial outcomes. Significance level acceptance was set at p < 0.05.
Data were analysed using STATA (version 11; STATA Corp, College Station, TX, USA). All analyses were done as intention to treat, which maintained the original group composition achieved by randomisation.

Ethical considerations
An ethical clearance certificate (protocol number M080812) was obtained from the Human Research Ethics Committee of the University of the Witwatersrand. As the data were collected in Rwanda, national ethical clearance (no. 032IRB052011) was also obtained from the Institutional Review Board at Kigali Health Institute and the proposal was scientifically approved by the National Commission for Control of HIV and AIDS (approval letter no. 0137/ CNLS/2011/S.E) in Rwanda. Participants gave written signed informed consent prior to their participation in the study and gave permission to use their medical records or files. This study was retrospectively registered with the Pan African Clinical Trial Registry (PACTR201803003147805).

Results
Out of 120 participants who were randomised into the intervention and control groups, 93% participated in all exercise sessions (Figure 1).
There were no significant differences in the demographic and health status characteristics between the experimental and control groups at baseline (Table 1).
Figures 2-4 and 7 demonstrate the increased percentage of participants whose symptoms significantly improved in the exercise intervention group as compared to the control group both after the intervention and again after a further 12 weeks.
There were no significant differences in the signs of PN between the intervention and control groups at the end of the 12 weeks of intervention and again after a further 12 weeks, as can be seen in Figures 5 and 6.
The results of the univariate and multivariate analyses determined the predictors for treatment outcome (

Discussion
The exercise intervention reduced PN symptoms among PLHIV on ART who exercised for 12 weeks, and this reduction was sustained for a further 12 weeks. Therapeutic exercises have known beneficial effects on painful neuromuscular conditions, reduce functional activity limitations and improve QoL through enhancing participation in activities of daily living (ADL) (Cherry et al. 2005), but there is insufficient existing information on the effects of exercise on PN symptoms in PLHIV on ART particularly in resourcelimited settings (Taylor et al. 2007).
Exercise is considered to be a key strategy by rehabilitation professionals to address the disabling health-related consequences of HIV . Several studies with exercise interventions on PLHIV have been conducted, including studies identified in systematic reviews (Maharaj & Chetty 2011;Taylor et al. 2007;Nixon et al. 2005). However, in all of these studies there is a dearth of evidence on the effects of exercise on PN among PLHIV on ART. To our knowledge, there is only one study that has been reported and that is a pilot study, without a control group and it has inconclusive results (Sandoval, Runft & Roddey 2010).   In our study, there were reductions in the neuropathic pain, numbness, paraesthesia and distribution of the symptoms along the lower extremities. To our knowledge, this study is the first one to show that a physiotherapist-led exercise intervention can reduce PN symptoms in PLHIV on ART in Rwanda. The findings are similar to the few existing studies, particularly the study by Sandoval et al. (2010) who found that a community exercise intervention reduced activity limitations among PLHIV with inflammatory peripheral neuropathy; however, their study was not a RCT.
The mechanism of how exercise reduces symptoms of PN is not yet clear. Exercise has been known to have beneficial effects such as an antidepressant effect, reduction of anxiety and inducement of sleep for those with insomnia, among others (Schuch et al. 2014;Suna et al. 2015). The antidepressant effect (Schuch et al. 2014;Suna et al. 2015) might play a role and is one of the treatments of choice for neuropathic pain in PLHIV (Eng & Reime 2014;Lana, Lérida & Mendoza 2008). The antidepressant effect of exercise is likely a result of the increased blood circulation that enhances the discharge of neuro-chemical pain inhibitors (Baron, Binder & Wasner 2010;Brix Finnerup, Hein Sindrup & Staehelin Jensen 2013). These discharged pain inhibitors inhibit pain receptors by enhancing the closing of the pain gate at the spinal dorsal horn of the small diameter nerves (Dobson et al. 2014), thus relieving pain. In addition, exercise leads to a feeling of relaxation, hence improving mood. In cases of pain, the feeling of good mood as a result of exercise inhibits the nociceptive mechanism and releases neuro-chemical pain inhibitors, hence modifying pain perception (Dworkin et al. 2003).

FIGURE 4:
The percentages of participants with neuropathic pain at baseline, 12 and 24 weeks, in the intervention and control groups.    The exercises were prescribed and supervised by a physiotherapist, and the usual therapeutic exercise techniques were applied. These included stretching for joint mobility, isometric strengthening, and aerobic exercises for 'warm-up' and finally 'cool down' at the end of every session. Stretching exercises after warming up are known for their effect on improving flexibility (Golan 2008;Yaksh & Sorkin 2005) and enhancing pain-free range of movements about a joint to promote better performance (Hess & Woollacott 2005;Zakas et al. 2006). Neuromuscular flexibility has been reported as a benefit of stretching before performing strength training exercises (Marek et al., 2005;Rubini, Costa & Gomes 2007;Sigal et al. 2004).
The stretching exercises might have induced muscle relaxation, consequently improving pain-free movements (Rubini et al. 2007). For example, most of the participants anecdotally reported being able to squat on a pit latrine, and walk up and down stairs more easily after the intervention while they had expressed difficulty with these activities before the intervention. Therefore, with the participants able to achieve pain-free movements, they reported reduced PN symptoms.
Furthermore, stretching exercises are reported to improve neural tissue flexibility. The stretching mechanism includes 'sliding and tensioning techniques, which are thought to enhance nerve gliding and restore neural tissue mobility' (Dobson et al. 2014). This may have induced nerve flexibility and mobility (Oskay et al. 2010), resulting in reduced pain during activities. The reduction of pain was of considerable benefit to these participants.
To our knowledge, the only study that has tested the effects of exercise on neuropathic pain in PLHIV is a pilot study by Sandoval et al. (Dobson et al. 2014), with few participants (Sandoval et al. 2010), and no control group, and it did not show changes in pain. In our study, the effect of PTExs on neuropathic pain included reduced sensations such as burning, pins and needles and tingling sensations, numbness, painful cold, pains in the form of 'electric shocks' hyper and hypo-aesthesia to touch and pin prick.
However, there were no significant improvements found in the PN signs, namely sense of vibration and ankle tendon reflex. Sense of vibration and tendon reflexes are mediated by large sensory nerves, unlike PN symptoms which occur as a result of small sensory nerve fibre damage (O'Brien et al. 2008). In addition, the development of PN signs is progressive and takes some time, though the time frame is uncertain. Consequently, it is possible that the treatment and relief or recovery of PN signs may occur after a longer duration exercise intervention than the one conducted here. The improvements shown in the PN symptoms may be because the recovery is quicker in small nerve fibres than in large fibres. Thus, the change in the improvement of the PN signs may need to be evaluated after a longer period of intervention. It seems that the improvement of the PN symptoms was as a result of the improvement in the neuro-tissue flexibility and in mobility, and possibly an improvement in the nerve regeneration or growth.
The improvement attained by the participants during the 12 weeks of exercise was maintained 12 weeks post-exercise, although there was a slight decline in effect. This is likely attributable to the probable continuation of exercises by the participants after the structured exercise programme, though this was not assessed. Further studies could evaluate a home exercise programme among PLHIV with PN. This could encourage PLHIV to continue exercising as a lifetime strategy to prevent and reduce the impact of PN (Devigili et al. 2008).
Some factors significantly predicted treatment outcome. The predictors for improvement of PN with exercises included health-related characteristics, namely the duration

The onset of PN symptoms and signs
Before starting on ARVs ---- of HIV infection, ART regimen changes and PN that developed post-ART initiation. The participants diagnosed with HIV for 7 years and longer were likely to have less improvement as a result of the exercise intervention. Because HIV infection might be one of the causes of PN (Dobson et al. 2014), this implies that living with the infection for a long time results in more severe PN or that prolonged exposure to HIV may render the nerve damage permanent. Early screening of the PN immediately after a person is diagnosed, and treating appropriately is important. Peripheral neuropathy is a chronic and disabling condition and early management that includes physiotherapist-led exercises might give relief and better health outcomes.
The participants who developed PN after initiating ART were likely to significantly improve compared to those who developed PN before initiating ART. The PN that develops after initiating ART is more likely to be related to ART than HIV. A change in the drug regimen (if the regimen contains medication associated with PN such as stavudine or tenofovir) may facilitate improvement of the pain, and with exercise as a supplementary treatment the improvement may be quicker. It is, therefore, important to include exercise in addition to drug regimen changes in the management of PN that occurs as a result of ART.
The participants in whom the ARV regimen was not changed were also likely to improve. This is possible because it is either that the PN was not severe, the PN had just started or it was not suspected to be augmented with a particular toxic ARV type (Dubey et al. 2013).

Conclusions and recommendations
Physiotherapist-led exercise seems to improve PN symptoms as the results of this study indicate that all the PN symptoms among the participants who exercised for 12 weeks improved as compared to the control group which had no improvement. Importantly, this improvement was maintained for a further 12 weeks post-intervention. This is likely a result of the probable continuation of exercise by the participants after the structured exercise programme. Factors related to the early diagnosis and treatment of PN were predictors for improvements of PN.
However, there is a need for further studies to establish the mechanism of how exercise improves PN symptoms. This may lead to a better understanding of how important exercises are in the management of PN among PLHIV.
In addition, studies on the physiological effects of exercise on the toxicity of the ARVs, particularly at the mitochondrial level, may be useful.
Finally, there is a need to investigate and test exercise interventions over prolonged periods of time, to establish if there is an effect on nerve recovery, in particular the larger nerve fibres for the effect on PN signs.