Supplementary - FMRI and treatment efficacy in PTNP to BJA
24th July 2017
Methods
Study sites and governance
The study was approved by the Oxford Research Ethics Committee C (08/H0606/5) and registered with ClinicalTrials.gov (NCT00610155). It was conducted in accordance with the principles of the Declaration of Helsinki. Patients were recruited from St Mary’s Hospital (Portsmouth, UK) and from Solihull Hospital (Birmingham, UK). Neuroimaging was performed at the Oxford Centre for Functional Magnetic Resonance Imaging of Brain (FMRIB) (Oxford, UK). The behavioural data results presented in the manuscript differ numerically from those reported during an early analysis as detailed in clinicaltrials.gov but do not alter the conclusions reported there. In the manuscript, the behavioural results are from 16 subjects. These subjects have 3 complete neuroimaging data sets and their neuroimaging results are presented in the manuscript. In the clinicaltrial.gov results, behavioural data are from 18 subjects including those subjects who did not complete all of the imaging visits as outlined in Supplementary Figure 1. We decided to present the behavioural results from the 16 subjects, as this would be in line with the imaging results presented in the manuscript.
Inclusion and exclusion criteria
Main inclusion criteria were: i) a confirmed diagnosis of neuropathic pain that has persisted for at least 3 months and was confirmed by a qualified pain specialist; ii) presence of DMA with an intensity of at least 4 on the 11-point Numerical Rating Scale (NRS) at visit 2 after completion of down titration of medication; iii) average pain intensity over the previous week of at least 3 on the NRS at screening; iv) age between 18 and 75 years and v) allodynic site within reach from outside the scanner. Main exclusion criteria were the following: i) trigeminal neuralgia, central pain, phantom-limb pain, painful diabetic neuropathy or chronic lower back pain; ii) any other coexisting pain that the patient could not differentiate from the neuropathic pain; iii) diabetes mellitus; iv) no response to pregabalin (≥300 mg/day), gabapentin (≥1800 mg/day, unless they have successfully responded to pregabalin ≥300 mg/day) or tramadol (≥400 mg/day) in the past (patients who have not previously received pregabalin or tramadol were included); v) patients who in the opinion of the investigator have any medical or psychological conditions or social circumstances that would impair their ability to participate reliably in the study, or who may increase the risk to themselves or others by participating; and vi) claustrophobia, metallic implants, metallic foreign bodies, excessive tattooing or any other contraindication for MRI.
Study procedures
At the first visit after obtaining informed consent, patients were evaluated for study criteria including the presence of dynamic mechanical allodynia (DMA). If deemed appropriate, then down-titration of existing medication was started. These patients attended the second visit to confirm their eligibility. At the third visit, patients were familiarised with the scanning procedure to reduce anxiety during scanning, and a baseline resting state scan and the structural scan were acquired after which the subjects were randomised to the treatment sequence. During this visit and visits 4 and 5, patients received a study medication pack for 14 days that consisted of placebo for the first seven days (washout period) followed by the medication appropriate for the treatment period. Subjects were advised to take the medication at approximately the same time in the morning and evening on each day. They were informed that they would be dosed with placebo, tramadol and pregabalin at various times during the study, but they did not know which treatment they were on during each treatment period. During these visits the study medication was dispensed, the previous medication was collected, pain and drug diaries were reviewed and patients completed pain-related questionnaires. During visits 4, 5 and 6, functional MRI (FMRI) scans were obtained. The first, second and seventh (follow-up) visit took place at the Recruiting Site. Visits 3-6 took place at the neuroimaging site in Oxford.
Delivery of brush and visual stimuli
DMA was elicited outside the scanner during visit 3 and inside the scanner during visits 4–6 while obtaining functional scans. This consisted of brushing of the allodynic site (DMAa) and a control site (DMAc) on the contralateral side of the body with a Somedic brush (three strokes in the same direction – proximal to distal during 5–6 minutes, stroke lengths 6–7 cm repeated 15 times, with a jittered duration of inter-stimulus intervals – total scan duration on each side was 10 minutes). After brush stimulation, patients were asked for an average pain rating on an 11-point NRS of all 15 stimuli. During the scanning sessions, this was followed by a control visual stimulation task consisting of a checkerboard flashing at 8 Hz, 15 seconds on and 15 seconds off, for 5 minutes to assess global haemodynamic influences of study drugs on the fundamental neuro-vascular coupling.
Study questionnaires
We collected the following patient-reported scores using validated questionnaires: Douleur Neuropathique 4,1 Hospital Anxiety and Depression Scale,2 Neuropathic Pain Symptom Inventory,3 State Trait Anxiety Inventory Questionnaire,4 Beck Depression Inventory II5 and Pain Catastrophising Scale.6 Of these, Douleur Neuropathique4 and Hospital Anxiety and Depression Scale were performed during visit 1 as screening tools. The Neuropathic Pain Symptom Inventory was given at visits 3–6. All others were collected during visits 4–6.
Randomisation
The allocation sequence was generated by Pfizer Inc and was not revealed to the patients, the recruiting clinicians or the researchers until the study and all analyses were completed (subject- and investigator-blinded and sponsor-open). The coded sequence corresponding to each randomisation number was kept in a sealed envelope and was broken only after the prediction analysis had been completed and the results reported to the sponsor.
The randomisation numbers were assigned at visit 3 if patients fulfilled the criteria and could tolerate MRI. Patients were consecutively allocated to one of 18 randomisation numbers corresponding to pre-numbered coded sequences of treatment allocation (A, B and C in six possible orders repeated three times; Supplementary Table 1). When replacement patients were recruited, they received randomisation numbers from a ‘mirror list’ (eg, number 3 was replaced with number 103).
All drugs and placebo were over-encapsulated, so that all the capsules looked the same and were arranged in identical blister packs.
Imaging data acquisition and data management
FMRI data was acquired using a 3 Tesla Tim Trio scanner (Siemens, Erlangen, Germany).
The blood oxygen level dependent (BOLD) FMRI images (functional and resting scans) were acquired using an echo-planar imaging sequence with the following parameters: GRAPPA (generalized autocalibrating partially parallel acquisitions) acceleration factor 2, field of view 192 × 192 mm, matrix size 64 × 64, repetition time of 3 seconds, echo time of 30 ms and flip angle of 87º. There were 49 x 3 mm-thick slices resulting in volume coverage in Z direction of 147 mm. Field maps were also collected using a symmetric-asymmetric spin echo sequence with the following parameters: repetition time of 532 ms, echo time 1 of 5.19 ms, echo time 2 of 7.65 ms, delta echo time of 2.46 ms flip angle of 60º matrix and field of view as the echo-planar imaging sequence.
For the structural scans, an MPRAGE (Magnetization Prepared Rapid Gradient Echo) sequence was used with the following parameters: single shot, field of view 192 mm, matrix size 192 × 174, volume coverage in superior-inferior direction 192 mm, repetition time of 2040 ms, echo time of 4.7 ms, inversion recovery of 900 ms, flip angle of 8º and 1 × 1 × 1 mm voxel.
All imaging data were stored, validated, analysed and assessed for quality at FMRIB independently of the sponsor. All clinical data were double-entered and quality checked, and all databases were locked before analysis.
Image analysis
These were analysed using tools in the FMRIB Software Library (FSL), version 4.1.7 (http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/).
Pre-processing steps included motion correction,7 field-map correction of echo-planar image distortion,8 removal of non-brain voxels9 and spatial smoothing using a Gaussian kernel of 5 mm full width at half maximum. A high-pass filter cut-off of 100s was used to remove low-frequency artifacts. Pre-processed data were analysed using a general linear model in FEAT (FMRIB Expert Analysis Tool) in the FMRIB Software Library. For task data, the input functions for stimuli were convolved with a single gamma haemodynamic response function. Head-motion parameters were included as general linear model confounds. These analyses generated the parameter estimates for the regressors that described the BOLD activity evoked by the brushing stimuli, DMAa and DMAc and visual stimulation. Images were registered to structural and standard anatomical structural space (MNI152) using linear registration with FLIRT (FMRIB Linear Image Registration Tool)10 and non-linear registration with FNIRT (FMRIB Non-linear Image Registration Tool).11
The chosen regions for resting-state functional connectivity consisted of the rostral anterior cingulate (rACC), bilateral amygdalae, periaqueductal-grey (PAG) and the brainstem. We applied identical pre-processing to the resting-state scans as was applied to the other functional scans. Head motion and cerebrospinal fluid signal was regressed from the resting-state data. All regions used in functional connectivity analyses, except the PAG, were defined anatomically using the Harvard-Oxford Cortical and Subcortical atlases (http://www.fmrib.ox.ac.uk/fsl/data/atlas-descriptions.html) with a threshold of 50%. The posterior border of the rACC corresponds to that of the perigenual ACC as previously described.12 As automated atlases are not available for the PAG, it was defined using the Duvernoy’s brainstem atlas.13
Use of rescue medication
Permitted rescue medications were paracetamol and codeine. Their usage was similar across the 3 treatment groups, and rescue medication was not permitted on the day of the imaging session. The number of participants who took paracetamol during the treatment period of pregabalin was 8, tramadol was 7 and placebo was 7. The number of participants who took paracetamol/codeine compound preparations during the treatment period of pregabalin was 7, tramadol was 9 and placebo was 9.
Supplementary Table 1 Randomisation sequence
Sequence / Period 1 / Period 2 / Period 3I / A / B / C
II / B / C / A
III / C / A / B
IV / A / C / B
V / C / B / A
VI / B / A / C
vii
Supplementary - FMRI and treatment efficacy in PTNP to BJA
24th July 2017
Supplementary Table 2 Characteristics of patients with post-traumatic neuropathic pain at the baseline (visit 3)
Subject / Sex / Age / Duration of pain (years) / Triggering factor / Side / Site of pain / Previous use of / Concomitant medicationpregabalin / gabapentin / tramadol
1 / Male / 49 / 4 / surgery / L / hand / N / N / N / Y
2 / Female / 49 / 2 / trauma / R / knee / N / N / N / N
3 / Female / 28 / 7 / trauma / L / knee / N / N / Y / N
4 / Female / 36 / 7 / trauma / R / ankle / N / N / N / N
5 / Female / 33 / 3 / surgery / L / groin / N / N / N / N
6 / Male / 53 / 9 / trauma / R / hand / Y / N / Y / Y
7 / Female / 56 / 1 / surgery / R / hand / N / N / N / N
8 / Male / 50 / 1 / post-ulcer / L / ankle / N / Y / N / N
9 / Male / 58 / 2 / trauma / L / ankle / Y / N / Y / Y
10 / Female / 52 / 7 / trauma / L / knee / N / N / N / Y
11 / Female / 41 / 1 / surgery / L / chest / N / N / N / Y
12 / Female / 38 / 1.5 / surgery / R / abdomen / N / N / Y / N
13 / Female / 57 / 3 / surgery / R / abdomen / N / N / N / Y
14 / Male / 45 / 4 / trauma / R / foot / N / Y / N / Y
15 / Female / 51 / 6 / unknown / L / hand / N / N / N / Y
16 / Female / 37 / 4 / unknown / R / abdomen / N / Y / N / Y
Duration of pain is time from onset of neuropathic pain; L = left; R = right; concomitant medication with nonsteroidal anti-inflammatory drugs or cyclo-oxygenase-2 inhibitors
vii
Supplementary - FMRI and treatment efficacy in PTNP to BJA
24th July 2017
References:
1 Bouhassira D, Attal N, Alchaar H, et al. Comparison of pain syndromes associated with nervous or somatic lesions and development of a new neuropathic pain diagnostic questionnaire (DN4). Pain 2005; 114: 29-36
2 Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta psychiatrica Scandinavica 1983; 67: 361-70
3 Bouhassira D, Attal N, Fermanian J, et al. Development and validation of the Neuropathic Pain Symptom Inventory. Pain 2004; 108: 248-57
4 Spielberger C, Gorsuch R, Lushene R, Vagg P, Jacobs G. Manual for the State-Trait Anxiety Inventory (form Y). Palo Alto (CA): Consulting Psychologists Press; 1983 1983
5 Beck AT, Steer R. Beck Depression Inventory Manual. New York: NY, Psychological Corporation, 1987
6 Sullivan MJL. The Pain Catastrophizing Scale: development and validation. Psychol Assess 1995; 7: 527-32
7 Jenkinson M, Bannister P, Brady M, Smith S. Improved optimization for the robust and accurate linear registration and motion correction of brain images. Neuroimage 2002; 17: 825–41
8 Jenkinson M. Fast, automated, N-dimensional phase-unwrapping algorithm. Magn Reson Med 2003; 49: 193-7
9 Smith SM. Fast robust automated brain extraction. Human brain mapping 2002; 17: 143-55
10 Jenkinson M, Smith S. A global optimisation method for robust affine registration of brain images. Med Image Anal 2001; 5: 143–56