難治性疼痛に対する脊髄刺激療法 無作為化比較試験：PAcStim RCT (PAcStim)

1. Background and Rationale 1.1 Clinical burden and limitation of current care Neuropathic pain (NP) is a severe form of chronic pain caused by lesions or diseases affecting the somatosensory nervous system. It is characterized by burning, shooting pain and allodynia.) The global population prevalence of NP is about 6-10%. The Canadian prevalence of neuropathic pain is 1.9% to 3.4%, meaning that approximately 302,000 to 537,000 Ontarians suffer from NP. This high prevalence is concerning because NP is associated with substantial decreases in quality of life, high rates of comorbidity with sleep problems, depression, and anxiety, and high economic costs to the individual and the society. Current pharmacological treatments for NP, including anticonvulsants, antidepressants and opioids, have a low success rate (25%) and most patients experience adverse effects3 leaving many patients without effective treatment. Non-pharmacological and interventional treatments provide additional approaches to managing NP, including spinal cord stimulation, peripheral nerve stimulation, dorsal root ganglion stimulation, intrathecal drug delivery systems, sympathetic nerve blocks, radiofrequency ablation, epidural steroid injections, and chemical neurolysis, but many patients remain refractory to available NP therapies.

1.2 Mechanism of action of spinal cord stimulation (SCS) Spinal cord stimulation (SCS) modulates pain signaling via electrical stimulation of dorsal column fibres, interfering with transmission of nociceptive signals to the brain. Neuromodulation approaches such as SCS are often effective for treating chronic refractory NP syndromes but long-term benefits are limited in some patients. There is evidence to support benefits of SCS in diabetic NP and its superiority over repeat surgery for patients with history of failed previous back surgery syndrome (FBSS) and the cost-benefits of SCS also support its use.

1.3 Paresthesia-based vs paresthesia-free spinal cord stimulation Modern SCS uses epidural leads positioned over the dorsal columns, programmed to either paresthesia-based (PB-SCS) typically below 80 Hz, or paresthesia-free (PF-SCS) modes (500-10,000 Hz high-frequency.

Paresthesia-based SCS (PB-SCS), with stimulating frequencies lower than 80 Hz, has been extensively used to treat NP in the limbs with or without axial pain with mean reduction in pain intensity of 60%.8 However, PB-SCS suffers from limitations including up to 40% of patients reporting attenuation of benefit at one year after implant and 10% to 15% experiencing painful or uncomfortable paresthesias.9 The recent increase in availability and use of paresthesia-free SCS (PF-SCS) modes that use high stimulating frequencies (500-10,000 Hz) are potentially more efficacious tools to treat NP. PF-SCS has a number of advantages over PB-SCS. PF-SCS stimulates analgesic pathways within the nervous system without stimulating fibers responsible for paresthesias (i.e., without tingling), making it more tolerable for patients. PF-SCS is likely more effective than PB-SCS, which simply masks pain signals with paresthesia. In the pivotal SENZA-RCT, HF10 demonstrated superior responder rates versus traditional low-frequency SCS at 3 months (back pain: 84.5% vs 43.8%; leg pain: 83.1% vs 55.5%) and maintained superiority through 12-24 months, with no paresthesias reported in HF10 recipients.

Quality-of-life analyses from the SENZA program show greater improvements (Oxygen Desaturation Index (ODI), sleep quality, global functioning, clinician impression of change) with HF10 than with conventional SCS at 12 months. Finally, because PF-SCS is imperceptible, it is less susceptible to habituation and may have more durable effects.

OHTA assessment shows that both PF-SCS and PB-SCS are effective, with some indication that PF-SCS may be more effective. The OHTA report concludes that peripheral nerve stimulation for chronic neuropathic pain likely reduces pain by 2.5-4 points on a 0-10 scale, achieves ≥50% pain relief in 50-70% of patients, improves quality of life (EQ-5D, SF-36), yields over 80% patient satisfaction, is generally safe with low serious adverse event rates, and may be cost-effective in Ontario (ICER $30,000-$50,000/QALY), though evidence remains limited and high-quality blinded trials are needed. (However these may all over-estimate the effects because blinding was not possible.

Our research group has recently synthesized evidence supporting PF-SCS in complex regional pain syndrome, a type of NP, and found very promising effects, with all of the 13 studies identifying a reduction in pain intensity (ranging from 30% to 100%). However, the 13 studies included only 62 patients in total, and only one study was a randomized controlled trial. Similarly, our own unpublished, observational data is very promising. Among 61 patients, the mean decrease in pain after 6 months of PF-SCS was 2.5 points on a 0 to 10 numeric rating scale for pain. However, none of these studies were fully blinded, leading to high risk of bias and the potential that the promising effects are due to placebo effects rather than true treatment effects..Patients with chronic pain frequently express preferences for non-drug options and paresthesia-free stimulation due to comfort, sleep/position independence, and device usability. HTA processes and patient-experience inputs in Ontario echoed positive patient-reported benefits and preferences for neuromodulation (including PNS), further supporting randomized evaluation of PF-SCS under blinding.

An important feature of PF-SCS is the potential to allow for a fully blinded RCT, which was not possible in RCTs testing PB-SCS. This is important from a research perspective to reduce bias in evaluation of outcomes,9 as pain research is prone to overestimation of treatment effects without proper blinding Indeed, only one blinded RCT testing the effect of SCS on NP has been published. In this 50-patient cross-over RCT, patients received two 3-month periods of PF-SCS and two 3-month periods of placebo in random order but the trial nurse who set the stimulator settings and collected data was not blinded. The study found no difference in the reduction of pain associated disability with PF-SCS but pain was not evaluated as the primary outcome. The authors interpreted this finding of equal decrease in disability index across placebo and PF-SCS periods as evidence for the placebo effect, suggesting that previous open-label studies showing an effect of PF-SCS on pain may not have reflected true treatment effects. However, our group and others have raised concerns about the methodology of this study and the study did not adhere to recommendations for SCS study design in important ways. Of primary concern are that there was no washout between treatment periods, which could lead to contamination between treatments and bias results towards the null and the effect of the treatments was calculated against a single baseline measurement taken at the start of the study rather than a baseline taken at the start of each treatment period. Thus, any beneficial effect of previous treatment periods could be reflected in subsequent change scores, potentially making subsequent placebo treatment periods appear beneficial, again biasing the study result towards the null. To directly address the shortcoming of the previous trial, in this trial we will conduct a Placebo versus Active Spinal Cord Stimulation (PAcStim) RCT for refractory NP, a blinded, multi-centre, cross-over RCT in 90 patients who will receive SCS implantation. Patients will be randomly allocated to receive 6 weeks of active stimulation (PF-SCS) followed by 2-week washout and 6 weeks of placebo (i.e., low amplitude) stimulation, or vice-versa.