The goal of this clinical trial is to learn if deep brain stimulation (DBS) works to treat severe pain following a stroke in adults. It will also learn about the safety of deep brain stimulation. The main questions it aims to answer are:

• Does DBS lower the pain score in these participants.
• What medical problems do participants have when having DBS? Researchers will compare different settings, to see if DBS works to treat severe post stroke pain.

Participants will:

• Undergo baseline screening procedures and have an MRI scan.
• Have neurosurgery to put the DBS system in
• Have follow up for 10 months
• Visit the clinic at least 5 times in the study for check-ups and tests
• Fill in questionnaires about pain and mood and have check ups remotely

The EPIONE trial is a double blind randomised controlled crossover trial. Comparing DBS stimulation with "Pseudo-ON " stimulation in participants who have central post stroke pain refractory to best medical/non-medical therapy. All participants will have the device on (with different stimulation parameters depending on the phase) during the trial. Blinding refers to knowledge of stimulation parameters. Following the randomised crossover periods, stimulation is further honed (refined) for an additional period of 6 months, called the "optimisation" period. This may include making use of circadian and motion sensing features of the Picostim-DyNeuMo-2 device, for example to increase stimulation during sleep or movement as appropriate to the individual's symptoms.

Existing evidence for the efficacy of DBS for chronic post stroke pain:

Modern DBS efficacy trials often include a single- or double-blind design with a sham stimulation or crossover phase. Strict case ascertainment criteria are also typically used. Unlike these modern trials, most pain DBS trials are uncontrolled case series or case reports, with non-standardised recruitment criteria and considerable heterogeneity between cases, hence the need for trials of DBS for pain that meets current scientific standards.

A number of open-labelled studies have demonstrated moderate efficacy of DBS for chronic pain including CPSP. A meta-analysis published in 2005 described the results of 424 cases pooled from 6 studies. Sites of stimulation included the periaqueductal/periventricular grey (PAG/PVG), the sensory thalamus and the internal capsule (IC). The authors noted that techniques used to assess pain severity were too heterogeneous to compare between studies. The meta-analysis showed that trial stimulation was successful in 50% of patients with central pain (i.e. CPSP), and in the patients with successful trial stimulation who went on to implantation, 58% reported ongoing pain relief. Thus, according to this meta-analysis, the overall percentage of patients with CPSP who benefitted from DBS was 31%, but the proportion receiving benefit was much higher in those with a positive response to test stimulation (58%). This meta-analysis highlights the importance of being able to pre-select patients most likely to have successful trial stimulation, which the investigators hope to address in the present trial.

A subsequent trial reported the results of a series of 56 patients who received DBS for chronic pain, of whom 11 had central post-stroke pain. Only 2/11 had successful test stimulation, but the two who had DBS implanted reported ongoing pain relief. Another open label single centre trial reported early improvements in 69.6% of patients with chronic post-stroke pain who underwent DBS, and of patients who retained their DBS stimulators at 1 year, there was a significant improvement in pain VAS score (p < 0.001). A case series of 4 patients with intractable pharmacologically resistant hemi-body thalamic pain lasting for at least 2 years. Three of these patients had post-stroke pain. Post DBS assessments were at 3,6 and 12 months. Three patients achieved long-lasting pain relief of more than 40% at 3, 6 and 12 months.

A recent assessor-blinded, randomized controlled crossover trial of DBS for post-stroke pain, targeting the ventral striatum/anterior limb of the internal capsule (VS/ALIC) is a unique and pioneering example of a controlled trial of DBS for chronic pain. In this trial of 9 patients, it was demonstrated that DBS was associated with significant improvements in scores on the Montgomery-Asberg Depression Rating Scale, the Beck Depression Inventory and the McGill Affective pain rating index, although the primary outcome of & 50% pain relief, was not achieved. No similar trial design has been employed to test the efficacy of more traditional targets for pain DBS- namely the PAG and the sensory thalamus. The aim of this trial is to build on the methodology demonstrated in previous trials, applying it to PAG and sensory thalamus stimulation.

The explanation for variability in outcomes of different trials of DBS for CPSP is not clear. However, important methodological differences exist between studies over crucial issues of patient selection, pre-operative opioid reduction and assessment of outcome. Blinding and sham stimulation are generally absent from reports of DBS for chronic pain (although blinded clinical evaluations of efficacy have been carried out. It is of great importance to conduct a prospective randomized controlled cross over trial for chronic pain in a selected chronic pain subgroup (i.e. chronic post-stroke pain) using traditional DBS targets (i.e. PAG and sensory thalamus). It is also crucial that careful analysis of pre-operative structural imaging and pain assessments are conducted to help develop algorithms that can predict which patients are likely to have positive responses to test stimulation as this ability would greatly improve operative outcome data and enable surgeons to avoid surgery in patients unlikely to benefit.

Rationale for the EPIONE trial The timing of this trial is apt given a recent evaluation of non-surgical therapies for chronic pain by NICE which indicated many current treatments are poorly effective (NICE NG193). This trial may offer evidence for the role of surgical therapy in chronic pain which has been approved by NICE historically (NICE IPG382) but not found to be cost effective, due to lack of evidence. Deep brain stimulation surgery offers an opportunity to provide randomised, blinded, controlled trial designs since the surgery is not the treatment itself, rather the neurostimulation provided by the implanted system is the treatment. Participants and assessors of the clinical effect can be blinded to stimulation parameters, and participants can be randomised to different settings at different time points. The investigators have sought to use these features of neurostimulation surgery to generate high quality evidence to support (or refute) the use of DBS neurosurgery for chronic pain.

Patients in the trial will undergo MRI brain imaging carried out for surgical planning. The investigators will use structural and DTI scans of brain structure to determine whether there are any pre-existing brain changes that predict whether a patient will have a successful or unsuccessful response to stimulation. Developing algorithms which incorporate imaging data to predict response to test stimulation will help to pre-select patients most likely to benefit from DBS in the future if the trial is successful. It will also help to understand more about the mechanism of action of DBS for pain.

There is clinical equipoise in this trial. Although the investigators have seen in large prospective cohort studies that stimulation can treat pain well in a significant proportion of patients, the investigators do not know if its efficacy can be demonstrated in a randomised controlled trial, which provides the level of evidence required to justify its use in the National Health Service (NHS). The investigators are also aware stimulation can cause unwanted side effects in some patients especially at higher power settings, which might make any treatment effect intolerable - some participants may prefer not to have stimulation. The purpose of comparing two settings during the randomisation/crossover phase, is to assess if appropriate stimulation is effective on average at treating CPSP without intolerable side effects before progressing to stimulation optimisation. Given the economic burden of the disease, economic outcome data will also be collated throughout the trial.

Inclusion criteria

A patient will be eligible for inclusion in this trial if all of the following criteria apply:

• Willing and able to give informed consent for participation in the trial.
• Willing and able to follow pre and post-operative procedures in Oxford.
• Aged 21 years or above.
• Diagnosed as having central post stroke pain of 2 years' minimum duration refractory to best medical/non-medical treatment
• Mean usual VAS (or NRS) pain score > 6/10 despite input from a multidisciplinary pain team.

Exclusion criteria

A patient will not be eligible for the trial if any of the following apply:

• Contraindication for elective general anaesthesia, for example but not limited to severe cardiovascular disease, hyponatraemia, hyperkalaemia, etc.
• Previous implantation of a DBS device with device still in situ.
• Contraindication to MRI
• Contraindication to neurosurgery, e.g. Bleeding disorders, not able to stop anticoagulation safely for perioperative phase (approx. 10 days, 5 days pre-operatively, 5 days postoperatively) Major psychiatric or cognitive disorder that may affect mental capacity that is untreated or may otherwise affect the participant's ability to engage in the trial
• Active skin-based infection or colonisation with a multi-drug resistant organism e.g. methicillin-resistant Staphylococcus aureus (MRSA)
• Requires regular MRI investigations post-operatively
• Likely to require diathermy, ultrasound or transcranial magnetic stimulation post DBS device insertion
• Not tolerant of awake surgery
• Unable to cooperate with device recharging
• Pregnancy or planned pregnancy
• In the investigator's opinion unable to comply with the protocol

Prior to entering the study, participants will have consented to their non-identifiable research data being shared with or by third parties. Data will be de-identified, containing no personally identifiable data (PID). These data are radiological information, de-identified questionnaire or clinical data and electrical traces from the implantable pulse generator (Picostim). There is a plan to make individual participant data and related data dictionaries available on request. All clinical data will be presented at the end of the trial as data listings. These will be checked to confirm the lists accurately represents the data collected during the course of the trial. The clinical investigation data will then be locked and a final data listing produced. The clinical trial report will be based on the final data listings. The locked clinical investigation data may then be used for analysis and publication.

When summary data are published or otherwise made available, requests will be considered, as per the plan (detailed in the CIP)

Blinding maintained using the hand held controller (shows ON at all times) and programs which ensure equivalent battery drain at all stimulation settings.

Pseudo-ON = a low power programme that provides 'sham stimulation' (high frequency, low amplitude stimulation generally ineffective at providing pain relief) and drains battery over time, requiring the participant to recharge the IPG periodically: around 250Hz, less than 0.5mA (with potential to be as low as 0mA), 450uS.