Randomised clinical trial: pregabalin attenuates experimental visceral pain through sub-cortical mechanisms in patients with painful chronic pancreatitis

Authors

  • S. S. Olesen,

    1. Mech-Sense, Department of Gastroenterology, Aalborg Hospital, Aarhus University Hospital, Denmark.
    Search for more papers by this author
  • C. Graversen,

    1. Mech-Sense, Department of Gastroenterology, Aalborg Hospital, Aarhus University Hospital, Denmark.
    Search for more papers by this author
  • A. E. Olesen,

    1. Mech-Sense, Department of Gastroenterology, Aalborg Hospital, Aarhus University Hospital, Denmark.
    Search for more papers by this author
  • J. B. Frøkjær,

    1. Mech-Sense, Department of Gastroenterology, Aalborg Hospital, Aarhus University Hospital, Denmark.
    2. Department of Radiology, Aalborg Hospital, Aarhus University Hospital, Denmark.
    Search for more papers by this author
  • O. Wilder-Smith,

    1. Pain and Nociception Neuroscience Research Group, Department of Anesthesiology, Pain and Palliative Medicine, Radboud University Nijmegen Medical Centre, The Netherlands.
    2. Center for Sensory-Motor Interactions (SMI), Department of Health Science and Technology, Aalborg University, Denmark.
    Search for more papers by this author
  • H. van Goor,

    1. Pain and Nociception Neuroscience Research Group, Department of Surgery, Radboud University Nijmegen Medical Center, The Netherlands.
    Search for more papers by this author
  • M. Valeriani,

    1. Center for Sensory-Motor Interactions (SMI), Department of Health Science and Technology, Aalborg University, Denmark.
    2. Division of Neurology, Ospedale Pediatrico Bambino Gesù, IRCCS, Rome, Italy.
    Search for more papers by this author
  • A. M. Drewes

    1. Mech-Sense, Department of Gastroenterology, Aalborg Hospital, Aarhus University Hospital, Denmark.
    2. Center for Sensory-Motor Interactions (SMI), Department of Health Science and Technology, Aalborg University, Denmark.
    Search for more papers by this author

Dr S. S. Olesen, Mech-Sense, Department of Gastroenterology and Hepatology, Aalborg Hospital, Aarhus University Hospital, Mølleparkvej 4, 9000 Aalborg, Denmark.
E-mail: sso@mech-sense.com

Abstract

Aliment Pharmacol Ther 2011; 34: 878–887

Summary

Background  Pregabalin has a broad spectrum of analgesic and antihyperalgesic activity in both basic and clinical studies. However, its mechanisms and sites of action have yet to be determined in humans.

Aims  To assess the antinociceptive effect of pregabalin on experimental gut pain in patients with visceral hyperalgesia due to chronic pancreatitis and to reveal putative changes in corresponding central pain processing as assessed by evoked brain potentials.

Methods  Thirty-one patients were randomly assigned to receive increasing doses of pregabalin or placebo for three consecutive weeks. Perceptual thresholds to electrical stimulation of the sigmoid with recording of corresponding evoked brain potentials were obtained at baseline and study end. The brain source localisations reflecting direct neuronal activity were fitted by a five-dipole model projected to magnetic resonance imaging of the individuals’ brains.

Results  As compared to placebo, pregabalin significantly increased the pain threshold to electrical gut stimulation from baseline (= 0.02). No differences in evoked brain potential characteristics were seen, neither after pregabalin nor placebo treatment (all > 0.05). In agreement with this, brain source locations remained stable during study treatment (all > 0.05).

Conclusion  Pregabalin was superior to placebo for attenuation of experimental visceral pain in chronic pancreatitis patients. We suggest its antinociceptive effects to be mediated primarily through sub-cortical mechanisms. (ClinicalTrials.gov, number NCT 00755573).

Introduction

Pain is the leading symptom in chronic pancreatitis (CP) and driven by a multifactorial aetiology.1 Studies have shown that patients with CP have alterations in central pain processing mimicking those seen in other chronic and neuropathic pain disorders, including sensitisation, cortical reorganisation and alterations in endogenous pain modulation.2–6 In a recent publication patients with CP showed hypersensitivity to experimental stimulation of the lower gut and alterations in central pain processing manifested as modifications of the corresponding evoked brain potentials (EPs).5 These findings probably reflect a generalised sensitised state of the pain system, thus providing a rationale for novel therapies modulating sensory neurotransmission from the gut.

Pregabalin has received considerable attention as an analgesic in clinical studies of chronic and neuropathic pain.7 Accordingly, in a recent publication we demonstrated pregabalin’s efficacy in treatment of pain in patients with CP.8 Although the clinical efficacy of pregabalin is well documented for chronic pain conditions, its analgesic mechanisms of action are not completely understood.7In vitro studies indicate that pregabalin binds selectively to the alpha-2-delta subunit of voltage-dependent calcium channels, thereby blocking the influx of calcium into presynaptic nerve terminals. This reduces release of excitatory neurotransmitters including glutamate, noradrenalin and substance P on spinal second order neurons, hence dampening pain transmission upstream in the central nervous system (CNS).9–14 Taken together, pregabalin appears to have a broad spectrum of analgesic and antihyperalgesic activity in both basic and clinical studies, although its mechanisms and sites of action have yet to be determined in humans.

Neuroimaging methods based on indirect measures of neuronal activity, such as functional magnetic resonance imaging (f-MRI) (changes in hemodynamic responses) or positron emission tomography (PET) (changes in metabolic responses), have been used extensively to study pain processing.15, 16 Although these methods possess excellent spatial resolutions and have greatly contributed to our knowledge of the structural basis of the pain system, their temporal resolutions are relatively poor (of the order of several seconds). Consequently, to address the dynamics of pain processing (i.e. the pain-specific sequential brain activation underlying pain perception) a method with high temporal resolution is needed, such as evoked brain potentials (EPs) which measure neuronal activity directly in real-time. Evoked brain potentials are the electrical manifestation of the brain’s response to an external stimulus.17 Pain EPs have been used widely to study gastrointestinal pain and reflect analgesic effects in clinical and experimental studies.18, 19 The location of brain centres underlying EPs can be determined by multi-channel electroencephalography (EEG) recordings combined with inverse modelling and structural brain imaging based on MRI. This method possesses the opportunity to study pain specific cortical activation dynamically, as it reflects the sequential activation of neuronal pain networks in real time.20

The current study was performed to assess the antinociceptive effect of pregabalin to experimental gut pain in patients with CP and to reveal putative changes in corresponding central pain processing assessed by EPs. We hypothesised that treatment with pregabalin would attenuate visceral pain thresholds and modify central pain processing. The aims of the study were: (i) to assess the antinociceptive effect of pregabalin to electrical stimulation of the sigmoid, (ii) to assess drug-induced modifications of the corresponding EPs, and (iii) to investigate drug effects on EP scalp topographies and underlying brain source localisations.

Methods

Patients

Patients were recruited for an investigator initiated double-blind, placebo-controlled, parallel-group study of increasing doses of pregabalin conducted in the Netherlands (department of Surgery, Radboud University Nijmegen Medical Center) and Denmark (department of Gastroenterology, Aalborg Hospital, Aarhus University Hospital).8 This study presents the neurophysiologic investigations obtained in patients recruited at Aalborg Hospital at baseline (before pregabalin treatment) and at the end of the study. Approval from the local Ethics Committee (N-20080028MCH) was obtained prior to study start and all patients provided written informed consent.

The criteria for inclusion were a diagnosis of CP based on The Mayo Clinic diagnostic criteria,21 abdominal pain typical for chronic pancreatitis (i.e. dull epigastric pain, eventually radiating to the back) and chronic pain (i.e. pain ≥ 3 days per week for at least 3 months). Patients were not eligible if they suffered from other acute or generalised chronic pain syndromes (e.g. irritable bowel syndrome), had a history of major depression, were previously diagnosed with moderate to severe renal impairment, were treated with pregabalin during the previous 4 months, had hypersensitivity or known allergy to pregabalin, had any clinically significant cardiac rhythm abnormality or any evidence of untreated myocardial ischemia or injury.

Patients taking concomitant analgesic medication were included if pain treatment was stable before start of the study. Patients on stable opioid medication and patients on non-opioid analgesics were included. No new analgesic therapies were initiated at any time during the trial.

Eligible patients visiting the outpatient hospital clinic (between October, 2008 and April, 2010) were recruited.

Study design and treatment

The study consisted of a 3-week double-blind treatment (pregabalin or placebo) titrated to analgesic effect and tolerability. Electrical stimulation of the sigmoid with simultaneous recording of EPs was done at baseline and at the end of the treatment period to investigate putative psychophysical and neurophysiological effects of pregabalin treatment.

Enrolled patients were stratified depending on the occurrence of diabetes mellitus (DM). Based on a computer-generated pseudo-random code, patients in each stratum were randomised to receive either pregabalin or matching placebo. Pfizer Clinical Research Operations prepared identical, coded medication bottles containing identical capsules of pregabalin or placebo. After randomisation patients received escalating doses of pregabalin (300–600 mg/day) or matching placebo capsules for 3 weeks. Study medication was titrated at 3-day intervals based on response and tolerability. All patients followed the same oral dosing schedule. Daily dosages were split into two equivalent doses, one administered in the morning and one in the evening. If unacceptable side effects were experienced by the patient, a single downward dose titration was allowed, after which the patient remained on the final dosage during the remainder of the study period. A minimum end dose of 300 mg/day was required; otherwise the subject was withdrawn from the study.

Brain imaging

All subjects were imaged on a 3T MR scanner (Signa HDx, General Electrics, Milwaukee, WI, USA) equipped with an 8-channel head coil. T1-weighted 3D BRAVO-sequence images (FOV 25 × 25 cm, 320 × 320 matrix, 1.0 mm slice thickness, whole head coverage, flip angle 14°, repetition time 9.0 ms, echo time 3.6 ms) were recorded to obtain anatomical images for co-registration with the EEG data. Images were iso-voxeled and translated into the Talaraich space using commercial software (Brainvoyager QX version 2.1.2, Brain Innovation V.V., Maastricht, The Netherlands). Axial T2-weighted FLAIR-sequence images (5 mm slice thickness, whole brain coverage, repetition time 8802 ms, echo time 127 ms, inversion time 2200 ms) were evaluated for atrophy, ischemic and white matter lesions by an expert radiologist.

Visceral stimulations

A 40 cm long probe with stainless steel electrodes mounted at the tip was used for electrical stimulation of the sigmoid. Detailed information on the probe has been described previously.22 An enema containing docosat and sorbitol (Klyx, Ferring, Copenhagen, Denmark) was administered 30 min prior to the experiment. During stimulations patients rested on their left side in the supine position. Stimulation was carried out via an 8.5 cm long anoscope (Cat. No. E-03. 19. 925, Heine Optotechnik, Herrsching, Germany). The stimulation probe was advanced through the anoscope and positioned for stimulation at the rectosigmoid junction.23 To ensure sufficient mucosal contact, the impedance across the electrodes was kept below 3 kΩ throughout stimulations. A computer-controlled constant-current stimulator (IES 230, JNI Biomedical APS, Klarup, Denmark) was used for the electrical stimulation. Each stimulus was applied as a single pulse of 2 ms duration. For each subject we first determined the sensory threshold by increasing the stimulus intensity in steps of 0.5 mA. Random sham stimuli having the same or lower intensity were given to blind the subject for the increase in stimulus intensity. When the sensory threshold was identified, the procedure was repeated to establish the pain threshold and tolerance threshold. Finally, thirty stimuli with an interstimulus interval of 5 s were applied at the pain threshold with simultaneous recordings of EPs.

Evoked brain potentials

The EEG was recorded from 62 electrodes mounted according to the extended international 10–20 system (SynAmp, Neuroscan, El Paso, TX, USA) and a standard EEG cap was used (Quick-Cap International, Neuroscan, El Paso, TX, USA). The impedance of the electrodes was kept below 5 kΩ. Electroencephalography signals were recorded with a sampling frequency of 1000 Hz. The recordings were obtained in a dimmed room, and all unnecessary electrical equipment was turned off to avoid contamination of the signal. Patients were instructed to rest, with their eyes open and gaze fixed on a remote object. Evoked brain potentials were produced by averaging the EEG signals from the repetitive electrical stimulations.

EP analysis

Analysis of the continuous EEG data was done using commercial software (Neuroscan version 4.3.1, Neuroscan, El Paso, TX, USA). The procedure included the following preprocessing steps: (i) Band stop filtering (49–51 Hz), (ii) epoching (−100 ms to 350 ms poststimulus), (iii) manual artefact rejection, (iv) averaging of accepted sweeps, (v) band pass filtering (0.5–70 Hz) and (vi) re-referencing.

Latencies and amplitudes for each EP component were analysed at the Cz and Fz electrodes (referenced to linked ear electrodes) and at the T7 and T8 electrodes (referenced to contra-lateral ear electrodes). These electrodes were favoured since cerebral activation following lower gut stimulation has previously been reported in the centro–frontal areas (Cz and Fz electrodes) and temporal areas (T7 and T8 electrodes).24 The EP analysis was guided by simultaneous topographic mapping based on spline interpolation, which shows the scalp distribution derived from all 62 channels simultaneously.

Dipolar Source Modelling

Dipolar source modelling was performed by brain electrical source analysis (BESA) (BESA Research 5.3, MEGIS Software GmbH, Gräfelfing, Germany). BESA calculates the potential distributions over the scalp from preset voltage dipoles within the brain; followed by evaluation of agreement between the recorded and calculated field distributions. The percentage of data that cannot be explained by the model is expressed as residual variance (RV). A symmetric constraint was applied to bilateral sources based on the symmetry assumption of the two hemispheres.25 An individual head model was built for each subject by co-registration of the individual Talairaich transformed MRI and EEG recordings. This procedure also allowed projection of the spatial position of the dipoles to the individual MRI, whereby dipole localisation was adjusted for individual brain anatomy and pathology (e.g. cortical atrophy) enabling comparison between subjects. In four of placebo-treated patients and three of pregabalin-treated patients an individual MRI of the brain could not be obtained and a standard spheric 4-shell head model was used.

The inverse modelling procedure was done using a sequential strategy.26 First, the grand mean across patients at baseline was analysed. The model derived from this analysis was then applied and modified to the traces from the individual subjects from baseline and post-treatment recordings. The latency interval from 40 ms to 240 ms was considered because it included all the well-shaped potentials.

Statistics

A pre hoc power calculation based on the reported experimental endpoints was not performed because the study was part of a randomised clinical trial evaluating the clinical efficacy of pregabalin in chronic pancreatitis, powered for a clinical primary endpoint; i.e. change in clinical pain score.8

All evaluations of stimulation thresholds, EP components and dipole calculations were performed blindly. The results are expressed as mean ± SD unless otherwise indicated. The demographic data, clinical pain relief and structural MRI data were analysed with a Fisher’s exact test or Student’s t-test as appropriate. Changes in perceptual thresholds and EP characteristics were compared with two-way anova. Dipole localisations were compared by a mixed anova model with coordinate (x vs. y vs. z) as within subject factor and treatment group (pregabalin vs. placebo) as between subject factor. Wald tests were used for post hoc analysis. Normality was checked by QQ-plots and the assumption of variance homogeneity by Levene`s test. P < 0.05 was considered statistically significant. The software package stata version 11.0 (StataCorp LP, College Station, Texas, USA) was used for the statistical analysis.

Results

Of a total of 109 patients assessed for eligibility, 31 patients were randomised and received study medication. Two patients were lost to follow-up due to unacceptable side effects (dizziness and confusion) and three patients could not tolerate the visceral stimulations. Hence, 26 patients were considered for the final analysis (n = 13 in the pregabalin group and n = 13 in the placebo group). Detailed information on study enrolment is provided on-line as supporting information (Figure S1). Baseline demographics and clinical characteristics are shown in Table 1. The groups were comparable at baseline except for gender, with more female patients in the pregabalin group 8 (62%) compared to the placebo group 2 (15%) (= 0.04).

Table 1.   Patient characteristics at baseline
 Pregabalin (n = 13)Placebo (n = 13)
  1. BMI, body mass index; CP, chronic pancreatitis; VAS, Visual analogue scale.

  2. = 0.04 comparing pregabalin and placebo treated patients.

Mean age (years)52 ± 1053 ± 12
Men*511
BMI (kg/m2)22.5 ± 5.622.3 ± 3.2
Aetiology (alcohol/other)6/78/5
Duration of CP (months)89 ± 39117 ± 92
Pancreatic morphology
 Calcifications36
 Dilated pancreatic duct74
 Ductal concrements43
 Cysts42
Diabetes mellitus53
Pancreatic exocrine insufficiency65
Mean pain score prior to treatment (VAS)4.2 ± 2.53.6 ± 2.3
Concomitant analgesic medication (none/weak-analgesics/opioids)2/5/61/3/9
Final pregabalin dose (300 mg/600 mg)5/8N/A

After 3 weeks of study treatment pregabalin-treated patients reported an average pain reduction of −2.5 ± 2.8 points on the VAS scale compared to −1.4 ± 1.7 points in the placebo group (= 0.2).

Brain imaging

Four patients in the placebo group (31%) and six patients in the pregabalin group (46%) had evidence of white matter lesions (= 0.7). Two patients in the placebo group (15%) and two patients in the pregabalin group (15%) had evidence of cortical atrophy (= 1.0). Finally, one patient in each group (8%) had evidence of ischemic lesions (= 1.0).

Experimental pain scores

All subjects reported painful sensations from electrical stimulations of the sigmoid. The impedance between the stimulation electrodes was less than 3 KΩ in all experiments ensuring good electrode contact. Under baseline conditions, there were no differences in stimulation thresholds between patients randomised to receive pregabalin or placebo (= 0.7, = 0.4) –Figure 1a. Following treatment, pregabalin was more effective in attenuating experimental visceral pain than placebo (= 6.3, = 0.02) –Figure 1b. Post hoc analyses are reported in Figure 1.

Figure 1.

 Upper panel (a): Baseline perceptual thresholds to electrical stimulation of the sigmoid in pregabalin-treated patients (grey bars) and placebo-treated patients (white bars). Lower panel (b): Changes in perceptual thresholds after 3 weeks of pregabalin (grey bars) or placebo treatment (white bars). *< 0.05.

No correlations between changes in clinical and experimental pain scores were seen for neither pregabalin-treated patients (= 0.13, = 0.7), nor placebo-treated patients (r = −0.26, = 0.4).

Visceral-evoked potentials

Temporal recordings (C7 and C8 electrodes) were dominated by a biphasic (negative–positive) complex (N1/P1) following electrical gut stimulation. In both the central and frontal leads (Cz and Fz electrodes) dominant peaks were seen at increased latencies as a biphasic (negative–positive) complex (N2/P2). Latencies and amplitudes of the EPs are reported in Table 2. In both the pregabalin group and placebo group latencies and peak-to-peak amplitudes of the EPs did not differ between baseline and study end: T7-electrode latencies (= 0.4), T7-electrode amplitudes (= 0.3), T8-electrode latencies (= 0.2), T8-electrode amplitudes (= 0.2), Cz-electrode latencies (= 0.8), Cz-electrode amplitudes (= 0.9), Fz-electrode latencies (= 1.0) and Fz-electrode amplitudes (= 1.0). Representative recordings from the Cz electrode with corresponding topographic mapping are illustrated in Figure 2.

Table 2.   Latencies and amplitudes of evoked brain potentials*
  Right temporal (T7-electrode)Left temporal (T8-electrode)Central (Cz-electrode)Frontal (Fz-electrode)
N1P1N1P1N2P2N2P2
  1. * The amplitudes were measured from the negative to the positive peak of the biphasic complex (i.e. N1/P1 for the temporal electrodes and N2/P2 for the central and frontal electrodes).

Pregabalin (n = 13)
BaselineLatencies (ms)98 ± 39135 ± 4695 ± 30139 ± 31207 ± 38289 ± 28209 ± 39282 ± 46
Amplitudes(μV)3.9 ± 2.03.6 ± 1.55.6 ± 3.55.6 ± 3.8
Post-treatmentLatencies (ms)101 ± 32142 ± 3898 ± 31143 ± 38207 ± 33303 ± 44214 ± 34306 ± 33
Amplitudes(μV)3.8 ± 2.63.8 ± 1.76.2 ± 5.56.5 ± 5.1
Placebo (n = 13)
BaselineLatencies (ms)103 ± 27146 ± 4192 ± 24126 ± 26196 ± 37288 ± 39197 ± 30276 ± 32
Amplitudes(μV)7.3 ± 3.86.4 ± 3.67.6 ± 4.56.5 ± 4.1
Post-treatmentLatencies (ms)112 ± 41159 ± 39103 ± 30137 ± 38206 ± 35296 ± 36207 ± 37295 ± 35
 Amplitudes (μV)5.7 ± 2.75.0 ± 3.18.0 ± 4.57.3 ± 4.7
Figure 2.

 Evoked brain potentials (EPs) following painful electrical stimulation of the sigmoid in patients with chronic pancreatitis treated with pregabalin (a) or placebo (b). Data from the central electrode (Cz) are shown. Dotted lines represent the baseline EPs and solid lines represent the EPs at study end. Individual components of the evoked brain potentials are labelled N2 and P2 and their corresponding topographic maps are shown.

Dipolar Source Modelling

Dipolar coordinates obtained from source analysis of the EPs projected to individual Talaraich transformed brain images are reported online as supporting material (Table S1). In both pregabalin and placebo-treated patients, dipole activities were located consistently in bilateral insula, cingulate gyrus and bilateral secondary somatosensory area (SII). The dipole localisations are illustrated in Figure 3. The RV values obtained in the individual subjects were satisfactory and only in one patient did the RV exceed 15%; however, exclusion of this patient from the analysis did not alter the results. In both the pregabalin group and the placebo group dipole localisations remained unchanged after study treatment: insula (= 0.02; = 0.9), SII (= 0.3; = 0.6) and cingulate gyrus (= 0.1; = 0.7). Also, no differences were seen for dipole latencies (= 0.03; = 0.9) or dipole strengths (= 1.7; = 0.2).

Figure 3.

 Locations of brain sources evoked by painful stimulation of the sigmoid at baseline (white) and at study end (black). (a) Pregabalin-treated patients. (b) Placebo-treated patients. In both groups dipole locations remained unchanged following study treatment. The average of the individual brain source coordinates (issued from individual brain imaging translated into the Talaraich space) is projected onto a standard MRI.

Discussion

The effect of pregabalin treatment on experimental visceral pain processing was studied in patients suffering from visceral pain due to chronic pancreatitis. An antinociceptive effect of pregabalin on electrical evoked pain from the gut was observed. The evoked brain potentials corresponding to gut stimulations remained unchanged after pregabalin treatment. Congruently with this, brain source localisations underlying the EPs remained stable. These findings may suggest a predominantly sub-cortical antinociceptive effect of pregabalin on experimental visceral pain in human.

Methodological considerations

A gender difference between groups was seen after randomisation with more female patients in the pregabalin group. We have, however, refrained from comparisons in subgroups because of their insufficient size. Previous visceral pain studies showed no gender differences in the experimental thresholds following rectal stimulation.27, 28 Furthermore, all the included female patients were either postmenopausal or had amenorrhoea due to their chronic disease. Gender differences in pain perception have mainly been reported between male patients and fertile female patients and are believed to involve differences in sex hormones such as oestrogen, although this is a topic of ongoing discussion.29, 30 Finally, we saw no differences between groups when comparing baseline perceptual thresholds or brain activity following gut stimulations. Taken together, we do not consider the difference in gender between groups to be of major importance for our findings.

A limited number of patients were included, thus introducing the possibility of a type 2 error. However, due to the comprehensive and unpleasant neurophysiological investigations it was not feasible to include a larger cohort of patients. Most studies based on parallel methods have included a comparable number of patients and under most circumstances changes in neurophysiological parameters were more sensitive than those of psychophysical parameters.18, 19, 31, 32 Consequently, if any electrophysiological changes were evident after pregabalin treatment, we would have expected to find them in this study where a clear psychophysical response was seen.

Translation from experimental to clinical studies has many limitations, which relate to the phasic and nonphysiological nature of experimental stimulations.33 On the other hand, most analgesics alleviate experimental pain stimuli,18, 19 and especially it has been shown that visceral stimulation in healthy volunteers can serve as a bridge to reflect the analgesic potency of opioids in patients with chronic pancreatitis.34

Structural MRI of the brain was used to build individual head models for inverse modelling. The advantage of real head models compared with spherical head models is that differences in underlying brain anatomy and pathology are considered.35 Although not of major importance in studies of healthy volunteers (where differences in brain morphology may be negligible), this is particularly important in CP patients where chronic pain, alcohol abuse and DM may induce structural alterations of the brain.35

Visceral pain processing after pregabalin treatment

The present study was not designed to evaluate the clinical efficacy and tolerability of pregabalin per se, but to explore the antinociceptive mechanisms underlying pregabalin analgesia in visceral pain. For detailed information on clinical efficacy and side-effects the reader is therefore referred to (Ref. 8). In agreement with results from the clinical part of this study published elsewhere, pregabalin-treated patients had greater pain relief compared with the placebo group.8 However, the results did not reach statistical significance due to the smaller number of patients entering the neurophysiological assessment.

Patients had increased pain thresholds to experimental gut stimulation following pregabalin treatment compared with placebo, while sensation thresholds were not modified by pregabalin. Consequently, it is most likely that the effect of pregabalin is due to a pain specific mechanism, rather than being related to a generalised decrease in neuronal excitability. These findings support findings from previous basic and clinical studies of visceral hypersensitivity attenuated by pregabalin treatment.9, 10, 14, 36 Accordingly, an antihyperalgesic effect of pregabalin was recently demonstrated in a human model of acid induced oesophageal hypersensitivity.37 In another study, pregabalin increased distension pain thresholds to normal levels in patients with irritable bowel syndrome and rectal hypersensitivity.36

To our knowledge, no previous studies have addressed the effect of pregabalin on EPs. The lack of changes in EP characteristics is a surprising finding, since we used an adaptive pain threshold paradigm (i.e. individual baseline pain thresholds were adapted to treatment effects on subsequent experimental days). Hence, we would have expected to find an increase in amplitudes corresponding to the increase in stimulus intensity after pregabalin treatment.38 However, the lack of EP modifications could be explained by a predominantly sub-cortical antinociceptive effect of pregabalin, where the increased stimulus intensity would be counterbalanced by a decreased spinal and/or brainstem excitability, resulting in a zero net effect on the cortical response. This hypothesis translates previous reports from animal models where pregabalin was shown to reduce the pool of excitatory neurotransmitters in the spinal cord.13, 14 However, data from animal studies cannot be uncritically translated to man. First of all animal studies are mainly based on motor reflexes or behavioural responses and such data can only partly be interpolated to pain, which is a net result of complex sensory, affective and cognitive processing. Second, there are major differences between the effects of drugs across species (and even strains), and this limits generalisation of findings.39 Also data from human experimental pain studies support a predominantly sub-cortical effect of pregabalin. Previously we have demonstrated an effect of gabapentin (another alpha-2-delta binding agent with a clinical analgesic profile comparable to pregabalin) on temporal summation to electrical skin stimulation, which is thought to be an effect on spinal wide dynamic neurons like the wind-up phenomena seen in animal studies.40 Furthermore, a single dose of gabapentin was shown to reduce thalamic activity in a study based on functional MRI.32

The idea of a predominantly sub-cortical antinociceptive effect of pregabalin will pose the question of how blocking a target channel that is ubiquitous in the central nervous system exerts effects only through sub-cortical mechanisms.41 An explanation for this may be found in an animal study by Suzuki and co-workers.42 They demonstrated that the actions of gabapentin were dependent on descending facilitation from the brainstem through serotonergic pathways culminating on presynaptic spinal receptors.42 The clinical correlates of these findings are seen in many chronic pain disorders, including CP, where altered descending pain modulation (i.e. increased facilitation along with impaired descending inhibition) are believed to play a central role for induction and maintenance of pain.5, 43 Thus, a state-dependent or permissive interaction with brainstem-spinal mechanisms may underlie the analgesic efficacy of pregabalin. Furthermore, this hypothesis explains that alpha-2-delta ligands only modulate abnormal pain function without effect on phasic noxious physiological activity.40, 44

In agreement with EP results, no changes were seen in localisation of brain sources obtained from inverse modelling. As the inverse modelling is derived directly from the EPs this was an expected finding. Brain source activities were located in bilateral insula, cingulate gyrus and bilateral in SII. These localisations are similar to findings from previous studies of lower gut stimulation based on functional MRI, positron emission tomography and inverse modelling of EPs.15, 45

Conclusions

An antinociceptive effect of pregabalin on experimental visceral pain was observed in patients with chronic pancreatitis. We suggest its antinociceptive mechanisms of action to be mediated primarily through sub-cortical mechanisms. This knowledge may be used in future studies and clinical strategies where combinations of analgesics with peripheral and central actions are used in treatment of visceral pain.

Acknowledgements

Declaration of personal interests: None. Declaration of funding interests: The study was funded by a free grant from Pfizer Research and Development, Hertha Christensen’s Foundation and Christenson-Ceson’s Family Foundation.

Ancillary