Central post-stroke pain (CPP) is extremely difficult to relieve and responds very poorly to analgesics targeting neuropathic pain, probably because the mechanisms underlying this pain remain poorly understood.

Stroke pain is traditionally considered to be of central origin and related to changes in the spinal cord and/or brain nociceptive systems. However, a recent study in a small cohort of patients has suggested that the peripheral nervous system (PNS) may have a role in the initiation and persistence of APD.

The main objective of this prospective randomised controlled bicentric study (Raymond Poincaré and Ambroise Paré) in double blind and parallel groups against placebo (3 arms) will be to evaluate the efficacy of two peripheral nerve blocks performed 14 days apart on spontaneous neuropathic pain after stroke. The active treatments used for the blocks will be either lidocaine 20 mg/ml or levobupivacaine 1.25 mg/ml or placebo (saline)

The primary endpoint will be the change in neuropathic pain intensity (assessed on an 11-point pain intensity scale), expressed as a difference in pain intensity between the value obtained before each block and that obtained 45 minutes after, corresponding to the maximum expected effect. Secondary endpoints will include exertional pain, pain quality, % relief, clinical global impression, pain assessment on a patient diary for a fortnight after each block and adverse events.

Patients will be randomised to receive one of 3 study treatments (lidocaine 2%, levobupivacaine 1.25 mg/ml or placebo). The treatment protocol will involve 2 perineural blocks performed 14 days apart. Assessment will continue for up to 2 weeks after each block, i.e. up to one month after the start of treatment. An evaluation of pain will be carried out before the block and after each block, at 45 minutes and at 5 hours, and then daily by the patient on a self-evaluation booklet for the 14 days following each block.

Randomisation will be centralised on a server from a list drawn up in advance by computer rogramme, balanced by blocks of variable size. Allocation between the 3 arms will be done according to a balanced 1:1:1 distribution. Treatments will be numbered from 1 to n, and allocated to patients in the chronological order of their inclusion in the trial.

Patients will be randomised on the day of treatment using a centralised computerised randomisation procedure to receive one of the 3 study treatments (lidocaine 20 mg/ml levobupivacaine 1.25 mg/ml or saline). No matching by age or duration of pain is planned, as randomisation usually results in groups matched at baseline on these criteria. The treatment will be administered over two visits performed 14 days apart by a qualified anaesthetist using the peri-nervous route according to current ecommendations (see above). Only one randomisation will be performed at baseline, so that a patient on active treatment cannot receive placebo at a later date and vice versa (see figure 1).

The investigators plan to randomise 10 patients per group and a total of 30 patients to achieve 90% power with a two sided α risk=0.05,. Given the estimated premature discontinuation rate, the investigators consider it necessary to include 12 patients per group for a total of 36 patients.

This study opens the way to new therapeutic avenues for these patients who often fail all treatments

Inclusion Criteria:

1. Patients aged 18 years and over with no maximum age (blocks are generally very well tolerated in the very elderly)
2. Pain in the upper or lower limb distal enough to be completely covered by a peripheral nerve block
3. Chronic pain for at least 6 months
4. Ischaemic or haemorrhagic stroke for at least 6 months documented clinically and by appropriate imaging (MRI)
5. Post-stroke central neuropathic pain defined as pain occurring in the aftermath of stroke meeting the criteria for probable or defined neuropathic pain according to the NeuPSIG algorithm and with a DN4 screening questionnaire score of at least 4 out of 10.
6. Spontaneous pain intensity greater than or equal to 4 out of 10 on an 11-point numerical scale (EN) at inclusion and randomisation (i.e. just before each block)
7. Patients affiliated to a social security scheme or beneficiaries of such a scheme
8. Stable oral analgesic pharmacological treatment for at least 2 weeks prior to inclusion
9. Acceptance and signing of the informed consent

Exclusion Criteria:

1. Inability or unwillingness to sign an informed consent
2. Person subject to a legal protection measure (safeguard of justice, curatorship, guardianship)
3. Patients with ongoing psychiatric pathology (major depression, psychosis) or cognitive disorders that prevent a good understanding of the protocol and questionnaires
4. Pain that is too widespread in one hemicycle or limb and cannot be adequately covered by blocks
5. Ongoing drug or substance abuse
6. Language (aphasia) or comprehension disorders, illiteracy
7. Moderate to severe renal or hepatic impairment
8. Contraindication to local anaesthetics for use in perineural blocks (infection or acute inflammation in the injection area, known allergy).
9. Pregnancy or breastfeeding
10. Known hypersensitivity to lidocaine, levobupivacaine, amide-linked local anaesthetics or to any of the excipients contained in the specialities used in the study.
11. Patients with recurrent porphyria or severe hypotension contraindicating treatment with lidocaine and/or levobupivacaine
12. Current treatment with antiarrhythmic drugs causing torsades de pointes (amiodarone, disopyramide, quinidinics, sotalol...) or with antiarrhythmic drugs with local anaesthetic activity (mexiletine or class III antiarrhythmic drugs) and cannot be discontinued.
13. Too little pain at the time of the blocks (< 4 out of 10)
14. Need to modify analgesic pharmacological treatment at the beginning or during the study

It is a randomised, double-blind study