3016 Diabetes Care Volume 37, November 2014

Spinal Cord Stimulation and Pain Rachel Slangen,1 Nicolaas C. Schaper,2 Catharina G. Faber,3 Elbert A. Joosten,1 Relief in Painful Diabetic Carmen D. Dirksen,4,5 Robert T. van Dongen,6 Alfons G. Kessels,4 and Peripheral Neuropathy: A Maarten van Kleef1,7 Prospective Two-Center Randomized Controlled Trial

Diabetes Care 2014;37:3016–3024 | DOI: 10.2337/dc14-0684

OBJECTIVE Painful diabetic peripheral neuropathy (PDPN) is a common complication of diabetes mellitus. Unfortunately, pharmacological treatment is often partially effective or accompanied by unacceptable side effects, and new treatments are urgently needed. Small observational studies suggested that spinal cord stimulation (SCS) may have positive effects. 1Department of Anesthesiology and Pain Med- icine, Maastricht University Medical Centre, RESEARCH DESIGN AND METHODS Maastricht, the Netherlands 2 We performed a multicenter randomized clinical trial in 36 PDPN patients with Department of Internal Medicine, Maastricht University Medical Centre, Maastricht, the severe lower limb pain not responding to conventional therapy. Twenty-two Netherlands patients were randomly assigned to SCS in combination with the best medical 3Department of Neurology, Maastricht Univer- sity Medical Centre, Maastricht, the Netherlands treatment (BMT) (SCS group) and 14 to BMT only (BMT group). The SCS system was 4 fi Department of Clinical Epidemiology and Medical implanted only if trial stimulation was successful. Treatment success was de ned Technology Assessment, Maastricht University as ‡50% pain relief during daytime or nighttime or “(very) much improved” for Medical Centre, Maastricht, the Netherlands pain and sleep on the patient global impression of change (PGIC) scale at 6 5CAPHRI School of Public Health and Primary

EMERGING TECHNOLOGIES AND THERAPEUTICS months. Care, Maastricht University, Maastricht, the Netherlands 6 RESULTS Department of Anesthesiology, Pain, and Palli- ative Care, Radboud University Medical Centre, Trial stimulation was successful in 77% of the SCS patients. Treatment success was Nijmegen, the Netherlands observed in 59% of the SCS and in 7% of the BMT patients (P < 0.01). Pain relief 7Department of Anesthesiology, Free University during daytime and during nighttime was reported by 41 and 36% in the SCS group of Amsterdam, Amsterdam, the Netherlands and 0 and 7% in the BMT group, respectively (P < 0.05). Pain and sleep were “(very) Corresponding authors: Rachel Slangen, rachel. ” [email protected], and Maarten van Kleef, much improved in 55 and 36% in the SCS group, whereas no changes were seen in [email protected]. the BMT group, respectively (P < 0.001 and P < 0.05). One SCS patient died because Received 17 March 2014 and accepted 9 July of a subdural hematoma. 2014. CONCLUSIONS Clinical trial reg. no. NCT01162993, clinicaltrials .gov. Treatment success was shown in 59% of patients with PDPN who were treated This article contains Supplementary Data online with SCS over a 6-month period, although this treatment is not without risks. at http://care.diabetesjournals.org/lookup/ suppl/doi:10.2337/dc14-0684/-/DC1. Painful diabetic peripheral neuropathy (PDPN) is a common complication of diabetes A slide set summarizing this article is available online. mellitus (DM), and prevalence of PDPN ranges from 10 to 26% (1–3). In many patients, ’ © 2014 by the American Diabetes Association. the pain is of such intensity that it has a major impact on patients health-related quality Readers may use this article as long as the work of life (HRQoL) and functional ability, including interference with general activity, mood, is properly cited, the use is educational and not mobility, work, social relations, sleep, and enjoyment of life (4). for profit, and the work is not altered. care.diabetesjournals.org Slangen and Associates 3017

Currently, PDPN can only be treated and the treatment algorithm of Jensen cooperation in the study; pregnancy; symptomatically; a variety of medica- et al. (6). Invasive therapy, such as intra- severe cardiac or pulmonary failure tions, alone and in combination, is often thecal drug delivery, was not allowed. A (.NYHA classification II); unstable blood only partially effective or is accompa- steering committee was responsible for glucose control (change in hemoglobin nied by unacceptable side effects (5– the design of the study and for developing A1c [HbA1c] .1.0% in the 3 months pre- 7). Given the limited effectiveness of the protocol (available at ClinicalTrials. ceding the trial); and use of oral anti- pharmacological treatment, alternative gov, NCT01162993). The study was per- coagulation that could not be stopped treatment strategies have been sought formed at the outpatient pain clinics of for a period of 10 days around the im- to alleviate pain, including spinal cord the Maastricht University Medical Centre plantation procedure. stimulation (SCS). In SCS, an electrode (MUMC+) and Radboud University Medical is positioned posterior in the epidural Center (Radboud UMC). The study was ap- Randomization space to the dorsal column at the level proved by the medical ethics committee of Prior to randomization, all patients of the nerve roots that transmit the no- the MUMC+ and the institutional review gave written informed consent. Patients ciceptive information from the painful board of the Radboud UMC. An indepen- were randomized between two groups area. The epidural lead is connected dent data and safety monitoring board re- by a computerized randomization. An to a battery producing an electrical cur- viewed the study procedures and outcomes. independent data manager designed rent, which induces paraesthesia, a the randomization in a 3:2 ratio to the sensation that suppresses the pain Patients SCS group or BMT group, with stratifica- according to the Gate Control theory Patients with DM between 18 and 80 tion according to age, sex, type of DM, (8). Patients can reduce or increase the years of age, suffering from moderate and severity of the PDPN according to intensity of the electric current by to severe PDPN present in the lower the MDNS (16). This randomization ratio means of a device that uses radio fre- limbs according to the Michigan Dia- was chosen to facilitate the inclusion of quency transmission. betic Neuropathy Score (MDNS) (16), participants. Treatment with SCS or BMT Several observational studies sug- were recruited from the outpatient was only performed for 6 months. The gested that SCS may have a positive ef- pain clinics and the diabetic outpatient surgical procedure for implantation of fect on pain in PDPN patients, with 23 clinics of the two participating centers the SCS system was performed by four out of 31 patients (74%) reporting a pain and were screened for eligibility. Inclu- pain specialists (two at MUMC+ and two relieving effect of $50% after 1 year sion criteria were as follows: insufficient at Radboud UMC), all with extensive ex- of treatment (9–13). Three studies pain relief and/or unacceptable side ef- perience of .5 years. reported a sustained treatment effect fects with drug treatment according to (2–7years)ofSCSinPDPNpatients the guidelines and the algorithm de- Lead Placement, Trial Stimulation, (9,10,14). As SCS is a relatively expensive scribed for PDPN by Jensen et al. (6), and Criteria for Permanent SCS therapy (;12,000 Euros per patient) including antidepressants, antiepileptic Implantation and not without risks, such as technical drugs, opioids, or a combination of Implantation of the SCS octapolar lead complications, there is a need for ran- these therapies, pain present for .12 (Octad lead; Medtronic, Minneapolis, domized controlled trials (RCTs) con- months, with a mean pain intensity dur- MN) was performed using local anes- firming its effectiveness in PDPN. ing daytime or nighttime on a numeric thesia and antibiotic prophylaxis. The Therefore, we performed a multicen- rating scale (NRS) of 5 or higher, and, if patient was placed in prone position, ter RCT to determine whether SCS treat- necessary, a psychological assessment and by using fluoroscopy, the epidural ment in combination with best medical was performed. spacewasenteredwithaTuohyneedle treatment (BMT) (SCS group) is more Exclusion criteria were as follows: at the lumbar level. Subsequently, the successful compared with BMT only neuropathic pain most prevalent in the lead was advanced through the needle (BMT group). Treatment success was upper limbs (NRS .3); neuropathy or and connected to an external program- defined as a $50% pain relief in pain chronic pain of other origin than DM; mable stimulator (N’Vision; Medtronic). intensity during daytime or nighttime, recent neuromodulation therapy (,1 The position of the lead over the tho- or an improvement for pain and sleep month before the intake-visit); drug, racic level and settings of the external of $6 in the score of the patient global medication, or alcohol (.5units/day) stimulator were tailored for each pa- impression of change (PGIC) scale. abuse; insufficient cooperation from tient in order to reach optimal paraes- the patient (little motivation, under- thesia coverage. Thereafter, the lead RESEARCH DESIGN AND METHODS standing, or communication); blood was anchored to the paraspinal fascia Study Design clotting disorder; immune deficiency; of the interspinous ligament and an ex- This study was designed as a prospective peripheral vascular disease with no pal- tension lead was threaded through the multicenter RCT to assess the effective- pable foot pulses at both feet (inclusion skin, fixed, and connected to an external ness of SCS in combination with BMT was possible if pulses were absent, but stimulator (External Neurostimulator (SCS group) compared with BMT only Doppler ankle-brachial index was be- Trialing System; Medtronic). After im- (BMT group) in patients with PDPN in tween 0.7 and 1.2 in both feet); active plantation, patients were admitted to the lower limbs on pain, HRQoL, and func- foot ulceration; life expectancy ,1 year; the hospital for 24 h. Patients were dis- tional ability. Prior to inclusion, all pa- pacemaker; local infection or other skin charged from the hospital if no change in tients were treated with BMT according disorders at site of incision; psychiatric position of the lead was seen after X-ray to the international guidelines (5,7,15) problems potentially interfering with verification. 3018 Spinal Cord Stimulation in PDPN Diabetes Care Volume 37, November 2014

After a 2-week trial stimulation, the which assesses two global and eight spe- randomized, even those patients without a spinal cord stimulator (Synergy Versitrel cific qualities of neuropathic pain (19). postbaseline measurement. Dropouts or PrimeAdvanced; Medtronic) was im- HRQoL was measured with the EuroQol were classified as failures for SCS. The effi- planted if the NRS for the intensity of five dimensions (EQ-5D) (20). In this cacy analysis was a logistic regression of pain during daytime or nighttime for instrument, a utility score can be com- the difference between the SCS group the last 4 days of the trial period was puted for each of five health states, and the BMT group in the proportion of at least 50% lower than the baseline basedonpreferencesfromageneral patients with a success of treatment over score, or if there was a score of 6 or population in the U.K., i.e., the U.K. tariff the course of 6 months. For the PGIC score higher (“much improved” or “very much (21). Additionally, patients can rate their (dichotomized as a score of ,6or$6), improved”) on the PGIC scale for pain current health on a visual analog scale there was no pretreatment data and and sleep. After antibiotic prophylaxis, (22). In addition, the Medical Outcomes therefore only differences between the the spinal cord stimulator was placed Study SF-36 (MOS SF-36) (23,24) was two groups were calculated. We per- just cranial to the greater gluteal muscle used to evaluate patients’ perceived formed a multivariate logistic regression in a subcutaneous pocket or in the ante- health state, which is composed of 36 to adjust for the effect of possible un- rior abdominal wall and was connected questions and standardized response balanced distributions of the baseline by a tunneled new sterile extension lead choices, organized into eight multi-item characteristics (age, sex, and MDNS score) to the stimulation lead, under local an- scales that can be averaged to form two on the primary outcome measure. esthesia. The first extension lead was re- composite scores, the physical compo- Linear mixed model analysis was per- moved. The patient remained in the nent score (PCS) and the mental compo- formed to analyze the differences in hospital for 24 h while antibiotic prophy- nent score (MCS). trends of the outcome measures. A ran- laxis was continued. If trial stimulation The impact of PDPN on sleep quality dom intercept regression model exam- was unsuccessful, the stimulation lead and quantity was assessed using the ined the differences in linear rate of was removed. Medical Outcomes Study Sleep Scale. A change between the SCS group and the nine-item Sleep Problems Index was BMT group over the course of 6 months. Outcome Measurements calculated, and quantity of sleep was This model measured each patient’sde- Basic demographic data and PDPN his- scored as the average hours of sleep viation from the population average tory were obtained, including duration per night reported over the past week. change over time and included a random of DM, duration of painful symptoms, Sleep was scored as optimal in cases of intercept for each subject. For this model, type of DM, length, weight, age, sex, 7–8 h of sleep per night (25,26). all available data (baseline and 3 and 6 and glycohemoglobin (HbA1c). Mood was assessed with the Beck months) were used. The model estimated Outcome measures were assessed at Depression Inventory (BDI), a self-report fixed effects for treatment. All statistical baseline and at 3 and 6 months for all inventory for measuring the severity of analyses were performed using SPSS Sta- patients; SCS patients received an extra depression (27). Finally, we documented tistics for Windows, version 20.0. assessment to evaluate the trial stimu- complications and (serious) adverse lation at 2 weeks. events of SCS. RESULTS Treatment success of SCS at 6 months Medication use was registered and Patients was predefined in the protocol as fol- wasscoredintermsofincreased,re- Between 1 February 2010 and 28 Febru- lows: $50%reliefofpainintensityon duced, no change in medication use, ary 2013, we screened 110 patients with an NRS for 4 days (17) during daytime switched to other medication, or com- PDPN at the MUMC+ and at the Radboud or nighttime or a score of $6ona7- pletely stopped with medication. UMC. Study enrollment, randomization, point Likert scale (1 = very much worse and follow-up are shown in Fig. 1. Patient and 7 = very much improved) of the PGIC Statistical Analysis baseline demographics and clinical char- scale for pain and sleep (18,19). A score Data from prospective pilot studies were acteristics are presented in Table 1. of 6 or higher on the PGIC indicates a used to estimate the required sample size In one patient, the implantation of clinically important difference. (10–12,28). The prespecified definition of the stimulation lead was complicated In addition, we measured pain sever- clinically relevant pain relief was a reduc- by a dural puncture, and the procedure ity, pain interference with daily life, pain tion of 50% on the NRS and/or improve- of this patient was immediately stopped characteristics, HRQoL, pain interfer- ment on the PGIC for pain and sleep. We (see below). Therefore, 21 of the 22 pa- ence with sleep, sleep quality and quan- calculated that a sample of 40 patients tients who were assigned to the SCS tity, mood, and registered medication with PDPN was required in order to group underwent a 2-week trial stim- use at all time points. detect a difference in success of 40% be- ulation period. One patient was not The modified Brief Pain Inventory for tween SCS and BMT compared with BMT. willing to fill out questionnaires after Diabetic Peripheral Neuropathy was We assumed a proportion of success in negative trial stimulation. After 6 weeks, used to measure pain severity and pain the SCS group of 50% and in the BMT another patient withdrew from the interference in daily life, which were av- group of 10%, with a power of 80% study due to infection and removal of eraged to form two composite scores, and a two-sided type I error of 0.05, al- the SCS system. the Pain Severity Index (PSI) and the lowing for 10% lost to follow-up. Pain Interference Index (PII) (18). The statistical analysis was carried out Results of Trial Stimulation Neuropathic pain qualities were mea- according to the intent-to-treat (ITT) prin- Trial stimulation was successful in 17 of sured using the Neuropathic Pain Scale, ciple, which included all patients who were the 22 SCS group patients (77%). All 17 care.diabetesjournals.org Slangen and Associates 3019

baseline characteristics, the OR for treat- ment success was 18.8 (95% CI 2.1–170.2). Sex was the only covariate that had a significant influence on the outcome, and adjusting for sex increased the OR to 24.7 (2.4–250.2). The results at 3 months were similar to the results ob- tained at 6 months and are presented in Supplementary Table 1. Results of pain severity, pain inter- ference with daily life, qualities of pain, HRQoL, pain interference with sleep quality and quantity, and mood in both groups are presented in Table 2 and Supplementary Table 1. Both the PSI and PII showed significant differences between the SCS and the BMT group (P , 0.001 and P , 0.008, respectively). The characteristics of neuropathic pain were improved in the SCS group compared with the BMT group, with the exception of deep pain. After 6 months, the mean utility score of the EQ-5D was increased in the SCS group with 0.25, and in the BMT group, the change was 0.00; this difference was not statistically significant (P , 0.776). No significant differences be- tween the SCS and BMT groups were seen on the visual analog scale (VAS) of the EQ-5D (Fig. 3) and the MCS and PCS of the MOS SF-36. Mood, as assessed with the BDI, did not differ between the two groups. Sleep quan- Figure 1—Randomization and follow-up. tity, optimal sleep, and sleep quality were of relative poor quality in both groups, but no significant effect of SCS was observed (Table 2). patients reported pain relief during day- SCS group was reduced by 3.1 points at 6 Seven out of 22 patients (32%) of the time of at least 50% as compared with months as compared with no change in SCS group were able to reduce their pain baseline, and 12 patients (55%) showed pain score in the BMT group (P , 0.001) medication, and for 2 of them, the SCS reduced pain of $50% during nighttime. (Fig. 2). In total, nine patients (41%) re- became the sole treatment for their All 17 patients had a minimal score of 6 ported $50% pain relief during daytime PDPN pain. Twelve patients (55%) did (much improved) for the PGIC for pain in the SCS group, as compared with 0% not change their medication in combina- and 14 patients (64%) for the PGIC for in the BMT group (P , 0.001). Mean tion with SCS treatment. Four patients sleep. pain at night was reduced by 2.4 points (29%) in the BMT group reported an in- in the SCS group compared with 0.9 creased use of medication compared with Outcome Measurements points on the NRS in the BMT group baseline, and one patient changed to an- ITT analysis was performed for 22 pa- (P , 0.003) (Fig. 2). Eight patients (36%) other category of neuropathic pain med- tients assigned to the SCS group and showed $50% pain relief during night- ication. In 9 out of 14 (64%) patients, 14 patients assigned to the BMT group. time in the SCS group, as compared with medication use was unchanged as com- Outcome data were available for 19 pa- one patient (7%) in the BMT group (P , pared with baseline. tients (86.4%) in the SCS group and 14 0.01). Of the 22 SCS patients, 12 (55%) patients in the BMT group (100%) at 6 scored $6 (much improved) on the Complications and (Serious) Adverse months. PGIC for pain (P , 0.001) and 8 (36%) Events Treatment success of SCS was ob- scored $6 on the PGIC scale for sleep Serious adverse events were noted in served in 13 out of 22 patients (59%) (P , 0.011). None of the BMT patients two patients. Implantation of the lead and in 1 out of 14 BMT patients (7%) showed any difference on the PGIC scale for test stimulation was complicated (P , 0.009) (Table 2). The mean pain for pain or sleep. After adjusting for pos- by a dural puncture, causing a postdural score on the NRS during daytime in the sible unbalanced distributions of the puncture headache in one patient 3020 Spinal Cord Stimulation in PDPN Diabetes Care Volume 37, November 2014

– Table 1—Baseline demographics and clinical characteristics of the study population 0.61 [95% CI 0.56 0.66]) (31,32). In the Characteristic SCS group (N = 22) BMT group (N =14) SCS patients, the utility score improved from 0.25 to 0.50, which was mainly due 6 6 Age (years) 57.1 12.4 56.5 8.0 to improvements on three dimensions Male sex (n [%]) 15 (68) 9 (64) of the EQ-5D: daily activities, pain, and Type of DM mood. As pain and mood both have a rel- Type 1 diabetes (n [%]) 3 (14) 1 (7) atively large weight in the regression- Type 2 diabetes (n [%]) 19 (86) 13 (93) based calculation of the utility score, Years of DM (years) 12.7 6 10.1 12.6 6 7.2 they largely accounted for the ob- Years of pain due to DM (years) 6.0 6 5.1 4.9 6 3.6 served major improvement in the SCS † 6 6 HbA1c (mmol/mol) 66.9 22.1 68.6 29.4 group. Nevertheless, probably due to † 6 6 HbA1c (%) 8.3 2.0 8.4 2.7 the large variability in the data, the dif- ‡ 6 6 BMI 29.0 4.3 30.3 5.4 ferences between the SCS and BMT pa- MDNS score§ tients were not significantly different. MDNS score 0 3 3 Remarkably, patients’ valuation of their MDNS score 1 4 3 MDNS score 2 9 5 current health state on the VAS of the MDNS score 3 6 3 EQ-5D showed no differences between the two groups. Patients in the SCS Plus-minus values are means 6 SD. †HbA1c is presented in mmol/mol and %. ‡BMI is the weight in kilograms divided by the square of the height in meters. §MDNS consists of a standard group showed hardly any improve- neurological examination, including vibration sense, light touch, pin-prick, tendon reflexes, and ments, despite reported pain relief and muscle strength, and standard nerve conduction study of two motor nerves (peroneal and improved functioning in daily activities median) and three sensory nerves (sural, median, and ulnar) performed on the nondominant site. Classes are defined as 0 = no neuropathy, 1 = mild neuropathy, 2 = moderate neuropathy, andmoodat6monthsascompared and 3 = severe neuropathy. with baseline. The difference between the general population (i.e., the utility scores) and patients’ valuations (i.e., the VAS) has been explained by the ad- (male, 65 years of age). After conserva- and nighttime pain. No differences aptation phenomenon (33). Where tive treatment was started, the patient were observed in HRQoL after SCS treat- PDPN patients are likely to adapt to was discharged from the hospital. Three ment. Finally, SCS treatment was not changes in their health state, healthy days after the procedure, the headache without any risk as two serious adverse individuals usually do not anticipate ad- suddenly increased and within minutes, events, which were both treatment aptation effects and thus value those the patient became unresponsive. The related, occurred. Significant pain re- health states differently (33,34). An- patient was transferred to the nearest lief during daytime was shown in 41%, other explanation for the discrepancy hospital where a CT scan showed a large as compared with 55–60% in previous between the utility scores and VAS in subdural hematoma over the left hemi- pilot studies, at 6 months (10–12). PDPN patients is that they may have sphere with a diameter of 26 mm, caus- Furthermore, 55% of the SCS group included dimensions in the valuation ing a midline shift of 19.8 mm. Despite showed a clinically important difference of their current health state other immediate surgical evacuation of the onthePGICscoreforpain.Thelarger than those included in the descriptive subdural hematoma, the patient did number of patients with an effect on system of the EQ-5D and subsequently not regain consciousness and died 10 PCIC can be explained by the fact weighted them differently (35). This days after surgery. that an increase of $6scoreonthe also explains the nonsignificant change Another patient contracted an infec- PGIC scale is associated with 30% inMCSandPCSoftheMOSSF-36.As tion of the SCS system 6 weeks after the pain reduction on the NRS scale and discussed below, the number of partic- SCS system implantation, which was is experienced as clinically relevant ipants in this RCT was limited; therefore subsequently removed. Despite treat- (17,29,30). In this study, 11 out of 22 studies with a larger number of partic- ment with antibiotics, the patient re- patients (50%) reported $30% pain re- ipants are necessary to determine the covered slowly but not completely. lief on the NRS during daytime. During effects of SCS on HRQoL. The patient also developed an auto- nighttime, $50% pain relief was noted Success of SCS seems highly dependent nomic neuropathy. After a consulta- in 36%, which was accompanied by a on the use of strict selection criteria and tion with the patient, it was decided nonsignificant1-hincreaseinsleeping the use of a 2-week trial stimulation with that the SCS system would not be time in the SCS patients. No differences predefined criteria to evaluate the effi- reimplanted. were observed in HRQoL after SCS cacy of the trial stimulation. We used treatment. strict selection criteria (36) to include pa- CONCLUSIONS As expected, our patients showed a tients, including an NRS pain intensity $5 This prospective multicenter RCT involv- poor HRQoL, and the participants re- measured on an NRS at the lower extrem- ing patients with moderate to severe ported utility scores of 0.25 (SCS group) ities for at least 6 months, treated with PDPN in the lower limbs showed that and 0.33 (BMT group) at baseline, which best medical therapy and patients with a after 6 months, SCS treatment reduces were substantially lower compared with clear history of PDPN. Causes other than pain more effectively compared with other studies that reported higher utility diabetic peripheral neuropathy were ex- BMT only. SCS reduced both daytime scores in PDPN patients (with a mean cluded, and if necessary, a psychological care.diabetesjournals.org Slangen and Associates 3021

Table 2—Outcomes in the SCS group compared with BMT group at 6 months, ITT analysis Baseline 6 months SCS group BMT group SCS group BMT group Outcomes (N =22) (N =14) (N =19) (N =14)

HbA1c (mmol/mol)† 66.9 6 22.1 68.6 6 29.4 64.2 6 15.1 72.1 6 25.2

HbA1c (%)† 8.3 6 2.0 8.4 6 2.7 10.9 6 8.0 8.8 6 2.3 NRS $50% pain reduction day (n/total n [%]) 9/22 (41) 0/14 (0)*** NRS $50% pain reduction night (n/total n [%]) 8/22 (36) 1/14 (7)** PGIC for pain (n/total n [%]) 12/22 (55) 0/14 (0)*** PGIC for sleep (n/total n [%]) 8/22 (36) 0/14 (0)* Treatment success (n/total n [%]) 13/22 (59) 1/14 (7)** mBPI-DPN† PSI 7.1 6 1.5 6.3 6 1.8 4.0 6 2.8 6.5 6 2.1*** PII 6.0 6 1.9 5.3 6 2.0 3.5 6 2.6 5.5 6 1.5** NPS‡ Deep pain 8.2 6 1.5 7.6 6 1.5 6.3 6 2.9 7.4 6 1.8 Surface pain 6.0 6 3.3 6.5 6 2.4 3.6 6 3.4 6.4 6 2.1** Intensity 8.0 6 1.5 7.6 6 1.5 4.3 6 3.0 7.3 6 2.0*** Unpleasantness 7.9 6 1.8 7.6 6 1.7 5.4 6 2.8 7.5 6 1.6** Coldness 4.2 6 3.6 4.9 6 4.0 2.2 6 2.7 5.6 6 3.1* Hotness 6.9 6 2.7 6.7 6 3.5 2.9 6 3.3 6.3 6 2.5** Dullness 7.6 6 2.1 7.7 6 2.4 4.6 6 3.5 7.6 6 1.7*** Sharpness 7.9 6 1.9 7.3 6 2.4 5.0 6 3.3 7.7 6 1.7** Sensitivity 7.6 6 2.5 6.5 6 2.4 5.0 6 3.4 7.1 6 2.2* Itching 3.9 6 3.1 3.6 6 3.3 1.6 6 2.6 4.2 6 3.2* EQ-5D§ Utility scores 0.25 6 0.31 0.33 6 0.32 0.50 6 0.33 0.33 6 0.29 Current health 53.9 6 18.5 54.6 6 16.7 57.6 6 24.3 56.5 6 14.2 MOS SF-36| MCS 44.7 6 13.5 45.3 6 11.8 49.3 6 11.5 46.7 6 12.0 PCS 27.9 6 7.5 31.7 6 7.9 32.3 6 10.5 30.5 6 7.4 MOS-SS¶ SPI-9 56.1 6 15.6 53.0 6 17.2 40.3 6 19.6 52.9 6 16.5 Quantity of sleep (h) 5.1 6 1.9 6.1 6 2.4 6.6 6 1.8 5.4 6 2.4 Optimal sleep (n [%]) 6 (27) 0 (0) 5/22 (23) 3/14 (21) BDI# 13.2 6 7.3 13.7 6 6.4 13.0 6 9.8 14.4 6 6.3 Plus-minus values are means 6 SD. †The modified Brief Pain Inventory- Diabetic Peripheral Neuropathy (mBPI-DPN) assesses the severity of pain and its impact on daily functioning and forms two composite scores: the PSI and the PII. Zero means “no pain’” or “does not interfere” and 10 “pain as bad as you can imagine” or “completely interferes.”‡The Neuropathic Pain Scale (NPS) includes two items that assess the global dimensions of pain intensity and unpleasantness and eight items that assess the specific qualities of neuropathic pain, in which 0 is “no pain or not” and 10 is “the most sensation imaginable.” §EQ-5D is a health status measure with respect to mobility, self-care, usual activities, pain or discomfort, and anxiety or depression (no problems, some problems, and extreme problems). A utility score of 1 represents “perfect health,” 0 represents “death,” and ,0 “worse than death.” Self-rated current health status was assessed using a vertical, visual analog scale on which 0 indicates worst possible health and 100 indicates perfect health. |MOS SF-36 measures PCS and MCS and is converted to a 0–100 scale, with higher scores indicating higher levels of functioning or well-being. ¶The Medical Outcomes Study Sleep Scale (MOS-SS) measures sleep quality and quantity. Sleep Problems summary 9 (SPI-9) scores range from 0 to 100 and higher scores indicate worse outcomes. #Severity of depression was measured by the BDI, and higher total scores indicate more severe depressive symptoms. *P , 0.05 for the between-group difference in the ITT analysis. **P , 0.01 for the between-group difference in the ITT analysis. ***P , 0.001 for the between-group difference in the ITT analysis.

assessment prior to inclusion was per- both peripheral and central mechanisms, DM (38–40). Whether involvement of formed. A 2-week trial stimulation with have been implicated in the development the spinal cord is a primary or secondary predefined criteria was used to evaluate of pain in diabetic peripheral neuropathy. event in diabetic peripheral neuropathy is the efficacy of SCS. A negative trial stim- A possible explanation for the substantial not yet clear (39). In cases of spinal cord ulation was noted in 23% of the SCS pa- number of negative trial stimulations atrophy, we would expect differences tients, which is in line with previously might be the presence of permanent cen- in various neurological tests between published studies (10–12). These patients tral changes in the spinal cord. Three ob- patients with a positive and those with a reported ,10% pain relief during day- servational studies reported a smaller negative trial stimulation. However, no time and nighttime, as well as no change mean spinal cord area index at the cer- differences in muscle strength, reflexes on the PGIC scale. As summarized by vical and thoracic level in patients with (knee and achilles tendon reflex), vibra- Tesfaye et al. (37), several anatomical or diabetic peripheral neuropathy as com- tion sense, and joint position in the lower pathophysiological changes, including pared with control patients without extremities were found. Thus, our data 3022 Spinal Cord Stimulation in PDPN Diabetes Care Volume 37, November 2014

Figure 2—Mean pain scores at daytime and nighttime. ITT analysis.

indicate that given the number of nega- following implantation. In one patient, an before dural tearing and showed full tive trial stimulations, a trial stimulation infection occurred 6 weeks after SCS sys- neurological recovery after emergent should always be performed when SCS is tem implantation, and the patient re- craniotomy (44). Although extensive ex- considered. sponded to antibiotic management and amination of our patient’srecordwas Although SCS is effective in most pa- removal of the SCS system. In the second performed, no predictors for this devas- tients, the treatment is not without risks. patient, however, a dural puncture during tating serious adverse event could be Review studies report SCS to be a safe the trial phase resulted in postdural punc- found. Except for this lethal complication, intervention in neuropathic pain patients ture headache and subsequent develop- the complication rate in this study is com- (41,42). The most common complications ment of a subdural hematoma with lethal parable to that reported in literature (41– related to SCS are hardware related (lead outcome. To the best of our knowledge, a 43). Nevertheless, given the invasive migration, lead fracturing, connection subdural hematoma is an extremely rare nature of SCS, our data underscore that failure, and discomfort), infection, sub- complication of SCS. After extensive re- SCS reduced pain but should be applied cutaneous hematomas, and cerebro- view of the literature, we found one as a last resort treatment and only should spinal fluid leak (41–43). In our study, case report of a subdural hematoma after be carried out in specialized centers by a severe complications were noted in two an SCS implant, where the patient suf- specialist with excellent experience in patients (9%) during a period of 6 months fered from minor head injury within 24 h performing SCS treatment.

Figure 3—EQ-5D utility score and EQ-5D current health. care.diabetesjournals.org Slangen and Associates 3023

Our study has some limitations that Duality of Interest. This study was supported 11. Pluijms WA, Slangen R, Bakkers M, et al. need to be discussed. Patients with failure by Medtronic, which provided a grant for the Pain relief and quality-of-life improvement after employment of R.S. for 3 years. No other of pharmacological treatment and with spinal cord stimulation in painful diabetic poly- potential conflicts of interest relevant to this neuropathy: a pilot study. Br J Anaesth 2012; high pain levels were screened and in- article were reported. 109:623–629 cluded. Therefore, the results cannot be Medtronic was not involved in the analysis and 12. Tesfaye S, Watt J, Benbow SJ, Pang KA, generalized to patients with less pain. interpretation of the data or in writing the Miles J, MacFarlane IA. Electrical spinal-cord Some of the screened patients were re- manuscript. stimulation for painful diabetic peripheral neu- – ferred to our clinic because of failure of Author Contributions. R.S. contributed to the ropathy. Lancet 1996;348:1698 1701 study design and analysis and interpretation 13. Kumar K, Toth C, Nath RK. Spinal cord stim- previous treatment. Although some of of the data and wrote the manuscript. N.C.S., ulation for chronic pain in peripheral neuropa- the referring centers used a somewhat dif- C.G.F., E.A.J., and C.D.D. contributed to the thy. Surg Neurol 1996;46:363–369 ferent treatment algorithm, all patients study design, analysis and interpretation of the 14. Slangen R, Pluijms WA, Faber CG, Dirksen had been treated with pharmacological data, and writing the manuscript. R.T.v.D. CD, Kessels AG, van Kleef M. Sustained effect of contributed to the study design and reviewed treatment as described in the current spinal cord stimulation on pain and quality of the manuscript. A.G.K. contributed to the life in painful diabetic peripheral neuropathy. guidelines, including antidepressants, anti- study design and analysis and reviewed the Br J Anaesth 2013;111:1030–1031 convulsants, and/or opioids. As pain sever- manuscript. M.v.K. contributed to the study 15. Bril V, England J, Franklin GM, et al.; Amer- ity is not always related to the extent of design and interpretation of the data and ican Academy of Neurology; American Associa- neurologic abnormalities on clinical exam- reviewed the manuscript. M.v.K. is the guar- tion of Neuromuscular and Electrodiagnostic antor of this work and, as such, had full access ination, the MDNS score was evenly dis- Medicine; American Academy of Physical Med- to all the data in the study and takes re- icine and Rehabilitation. Evidence-based guide- tributed in our patients, leading to the sponsibility for the integrity of the data and line: treatment of painful diabetic neuropathy: possibility of referral of more severely af- the accuracy of the data analysis. report of the American Academy of Neurology, fected patients. A major limitation of the the American Association of Neuromuscular study, like all studies using conventional References and Electrodiagnostic Medicine, and the Amer- SCS, was the fact that this study was not 1. Abbott CA, Malik RA, van Ross ERE, Kulkarni ican Academy of Physical Medicine and Reha- bilitation [published correction appears in blinded, since paraesthesia accompanies J, Boulton AJ. Prevalence and characteristics of painful diabetic neuropathy in a large community- Neurology 2011;77:603]. Neurology 2011;76: SCS. Moreover, it is unethical and too in- based diabetic population in the U.K. Diabetes 1758–1765 vasive to implant a SCS system that is sub- Care 2011;34:2220–2224 16. Feldman EL, Stevens MJ, Thomas PK, Brown sequently not used. Finally, PDPN is a 2. Daousi C, MacFarlane IA, Woodward A, MB, Canal N, Greene DA. A practical two-step progressive disease in which symptoms Nurmikko TJ, Bundred PE, Benbow SJ. Chronic quantitative clinical and electrophysiological as- sessment for the diagnosis and staging of diabetic will be present in not only the feet and painful peripheral neuropathy in an urban com- munity: a controlled comparison of people with neuropathy. Diabetes Care 1994;17:1281–1289 legs but will gradually progress proximally and without diabetes. Diabet Med 2004;21: 17. Dworkin RH, Turk DC, Wyrwich KW, et al. toward the hands and arms. We excluded 976–982 Interpreting the clinical importance of treatment patients with pain in the upper extremities 3. Davies M, Brophy S, Williams R, Taylor A. The outcomes in chronic pain clinical trials: IMMPACT – as only the lower extremities were stimu- prevalence, severity, and impact of painful di- recommendations. J Pain 2008;9:105 121 abetic peripheral neuropathy in type 2 diabetes. 18. Zelman DC, Gore M, Dukes E, Tai KS, lated. Hence, it remains to be determined Diabetes Care 2006;29:1518–1522 Brandenburg N. Validation of a modified version if SCS at both cervical and thoracal/lumbar 4. Barrett AM, Lucero MA, Le T, Robinson RL, of the Brief Pain Inventory for painful diabetic pe- spinal levels will be effective. Dworkin RH, Chappell AS. Epidemiology, public ripheral neuropathy. J Vasc Nurs 2005;23:97–104 These limitations notwithstanding and health burden, and treatment of diabetic pe- 19. Galer BS, Jensen MP. Development and pre- liminary validation of a pain measure specificto including the risks, our findings show that ripheral neuropathic pain: a review. Pain Med 2007;8(Suppl. 2):S50–S62 neuropathic pain: the Neuropathic Pain Scale. in PDPN patients, SCS in combination 5. Dworkin RH, O’Connor AB, Audette J, et al. Neurology 1997;48:332–338 with BMT results in clinically relevant Recommendations for the pharmacological 20. Kind P, Dolan P, Gudex C, Williams A. Var- pain relief over a 6-month period. management of neuropathic pain: an overview iations in population health status: results and literature update. Mayo Clin Proc 2010; from a United Kingdom national questionnaire 85(Suppl.):S3–S14 survey. BMJ 1998;316:736–741 Acknowledgments. The authors thank the 6. Jensen TS, Backonja MM, Hernandez´ Jimenez´ 21. Dolan P, Roberts J. Modelling valuations for members of the multidisciplinary steering com- S, Tesfaye S, Valensi P, Ziegler D. New perspec- EQ-5D health states: an alternative model using mittee for designing and guiding the protocol. tives on the management of diabetic peripheral differences in valuations. Med Care 2002;40: The authors specially thank M.N. Janssen, MD, neuropathic pain. Diab Vasc Dis Res 2006;3:108– 442–446 Maastricht University Medical Centre, Maastricht, 119 22. EuroQol Group. EuroQol–a new facility for the Netherlands; R.T. van Dongen, MD, PhD, 7. Snedecor SJ, Sudharshan L, Cappelleri JC, the measurement of health-related quality of Radboud University Medical Center, Nijmegen, Sadosky A, Mehta S, Botteman M. Systematic life. Health Policy 1990;16:199–208 the Netherlands; and M.A. Steegers, MD, PhD, review and meta-analysis of pharmacological 23. Aaronson NK, Muller M, Cohen PD, et al. Radboud University Medical Center, Nijmegen, the therapies for painful diabetic peripheral neu- Translation, validation, and norming of the Netherlands, who implanted most of the SCS ropathy. Pain Pract 2014;14:167–184 Dutch language version of the SF-36 Health Sur- systems, and W. Lardenoye-Somers, Maastricht 8. Melzack R, Wall PD. Pain mechanisms: a new vey in community and chronic disease popula- University Medical Centre, Maastricht, the Nether- theory. Science 1965;150:971–979 tions. J Clin Epidemiol 1998;51:1055–1068 lands; E. Jans-Meers, Maastricht University Med- 9. Daousi C, Benbow SJ, MacFarlane IA. Electri- 24. van der Zee K, Sanderman R, Heyink J. The ical Centre, Maastricht, the Netherlands; I. Arnts, cal spinal cord stimulation in the long-term psychometric qualities of the MOS 36-item Radboud University Medical Center, Nijmegen, the treatment of chronic painful diabetic neuropa- short form survey in a Dutch population. Netherlands; and P. Verstegen, Radboud Univer- thy. Diabet Med 2005;22:393–398 Tijdschr Soc Gezondheidsz 1993;71:183–192 [in sity Medical Center, Nijmegen, the Netherlands for 10. de Vos CC, Rajan V, Steenbergen W, van der Dutch] the guidance of the patients. The authors are Aa HE, Buschman HP. Effect and safety of spinal 25. Viala-Danten M, Martin S, Guillemin I, Hays indebted to the colleagues who referred candi- cord stimulation for treatment of chronic pain RD. Evaluation of the reliability and validity dates for the study to our department and to caused by diabetic neuropathy. J Diabetes Com- of the Medical Outcomes Study Sleep Scale the patients who participated in the study. plications 2009;23:40–45 in patients with painful diabetic peripheral 3024 Spinal Cord Stimulation in PDPN Diabetes Care Volume 37, November 2014

neuropathy during an international clinical trial. 31. Luo N, Johnson JA, Shaw JW, Feeny D, distal symmetrical polyneuropathy. Diabetes Health Qual Life Outcomes 2008;6:113 Coons SJ. Self-reported health status of the gen- Care 2013;36:2456–2465 26. Zelman DC, Brandenburg NA, Gore M. Sleep eral adult U.S. population as assessed by the EQ- 38. Eaton SEM, Harris ND, Rajbhandari SM, impairment in patients with painful diabetic pe- 5D and Health Utilities Index. Med Care 2005; et al. Spinal-cord involvement in diabetic ripheral neuropathy. Clin J Pain 2006;22:681– 43:1078–1086 peripheral neuropathy. Lancet 2001;358:35–36 685 32. Doth AH, Hansson PT, Jensen MP, Taylor RS. 39. Selvarajah D, Wilkinson ID, Davies J, Gandhi 27. Beck AT, Ward CH, Mendelson M, Mock J, The burden of neuropathic pain: a systematic R, Tesfaye S. Central nervous system involve- Erbaugh J. An inventory for measuring depres- review and meta-analysis of health utilities. ment in diabetic neuropathy. Curr Diab Rep sion. Arch Gen Psychiatry 1961;4:561–571 Pain 2010;149:338–344 2011;11:310–322 28. Kumar K, Taylor RS, Jacques L, et al. The 33. Menzel P, Dolan P, Richardson J, Olsen JA. 40. Selvarajah D, Wilkinson ID, Emery CJ, et al. effects of spinal cord stimulation in neuropathic The role of adaptation to disability and disease Early involvement of the spinal cord in diabetic pain are sustained: a 24-month follow-up of the in health state valuation: a preliminary norma- peripheral neuropathy. Diabetes Care 2006;29: prospective randomized controlled multicenter tive analysis. Soc Sci Med 2002;55:2149–2158 2664–2669 trial of the effectiveness of spinal cord stimula- 34. Redelmeier DA, Rozin P, Kahneman D. Un- 41. Compton AK, Shah B, Hayek SM. Spinal cord tion. Neurosurgery 2008;63:762–770; discus- derstanding patients’ decisions. Cognitive and stimulation: a review. Curr Pain Headache Rep sion 770 emotional perspectives. JAMA 1993;270:72–76 2012;16:35–42 29. Farrar JT, Young JP Jr, LaMoreaux L, Werth 35. Ubel PA, Loewenstein G, Jepson C. Whose 42. Bendersky D, Yampolsky C. Is Spinal Cord JL, Poole RM. Clinical importance of changes in quality of life? A commentary exploring discrep- Stimulation Safe? A Review of Its Complications. chronic pain intensity measured on an 11-point ancies between health state evaluations of pa- World Neurosurg. 11 July 2013 [Epub ahead of numerical pain rating scale. Pain 2001;94:149– tients and the general public. Qual Life Res print] 158 2003;12:599–607 43. Jeon YH. Spinal cord stimulation in pain 30. Farrar JT, Pritchett YL, Robinson M, Prakash 36. Atkinson L, Sundaraj SR, Brooker C, et al. management: a review. Korean J Pain 2012;25: A, Chappell A. The clinical importance of Recommendations for patient selection in spi- 143–150 changes in the 0 to 10 numeric rating scale for nal cord stimulation. J Clin Neurosci 2011;18: 44. Chiravuri S, Wasserman R, Chawla A, Haider worst, least, and average pain intensity: analy- 1295–1302 N. Subdural hematoma following spinal cord ses of data from clinical trials of duloxetine in 37. Tesfaye S, Boulton AJ, Dickenson AH. Mech- stimulator implant. Pain Physician 2008;11:97– pain disorders. J Pain 2010;11:109–118 anisms and management of diabetic painful 101