The effect of transcutaneous vagus nerve stimulation on pain perception – An experimental study

Brain Stimulation 6 (2013) 202e209

Contents lists available at SciVerse ScienceDirect

Brain Stimulation journal homepage: www.brainstimjrnl.com

Original Articles

The effect of transcutaneous vagus nerve stimulation on pain perception e An experimental study Volker Busch a, *, Florian Zeman b, Andreas Heckel f, Felix Menne a, Jens Ellrich c, d, e, Peter Eichhammer a a

Department of Psychiatry and Psychotherapy, University of Regensburg, Universitätsstraße 84, 93059 Regensburg, Germany Centre for Clinical Studies, Department of Epidemiology, University of Regensburg, Germany Medical Department, Cerbomed GmbH, Erlangen, Germany d Department of Health Science and Technology, Medical Faculty, Aalborg University, Denmark e Institute of Physiology and Pathophysiology, University of Erlangen-Nuremberg, Erlangen, Germany f Department of Neurology, University of Heidelberg, Germany b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 1 February 2012 Received in revised form 19 March 2012 Accepted 17 April 2012 Available online 22 May 2012

Background: Recent preclinical work strongly suggests that vagus nerve stimulation efficiently modulates nociception and pain processing in humans. Most recently, a medical device has offered a transcutaneous electrical stimulation of the auricular branch of the vagus nerve (t-VNS) without any surgery. Objective: Our study investigates whether t-VNS may have the potential to alter pain processing using a controlled design. Methods: Different submodalities of the somatosensory system were assessed with quantitative sensory testing (QST) including a tonic heat pain paradigm in 48 healthy volunteers. Each subject participated in two experimental sessions with active t-VNS (stimulation) or sham t-VNS (no stimulation) on different days in a randomized order (crossed-over). One session consisted of two QST measurements on the ipsiand contralateral hand, each before and during 1 h of a continuous t-VNS on the left ear using rectangular pulses (250 mS, 25 Hz). Results: We found an increase of mechanical and pressure pain threshold and a reduction of mechanical pain sensitivity. Moreover, active t-VNS significantly reduced pain ratings during sustained application of painful heat for 5 min compared to sham condition. No relevant alterations of cardiac or breathing activity or clinical relevant side effects were observed during t-VNS. Conclusions: Our findings of a reduced sensitivity of mechanically evoked pain and an inhibition of temporal summation of noxious tonic heat in healthy volunteers may pave the way for future studies on patients with chronic pain addressing the potential analgesic effects of t-VNS under clinical conditions. Ó 2013 Elsevier Inc. All rights reserved.

Keywords: Pain thresholds Quantitative sensory testing Autonomic nervous system Neuromodulation Analgesia

Introduction Abbreviations: ALL, Dynamic mechanical allodynia; ANS, Autonomic nerve system; BDI, Beck depression inventory; bpm, Beats per minute; cm, Centimeters; cpm, Cycles per minute; CDT, Cold detection threshold; CPT, Cold pain threshold; HPT, Heat pain threshold; HR, Heart rate; MDT, Mechanical detection threshold; MPS, Mechanical pain sensitivity; MPT, Mechanical pain threshold; NRS, Numeric rating scale; PHS, Paradoxical heat sensation; PPT, Pressure pain threshold; QST, Quantitative sensory testing; RD, Respiration depth; RF, Respiration frequency; SCID, Structured clinical interview for psychiatric diseases; SCL, Skin conductance level; SOMS-2a, Screening for somatoform symptoms of the last 2 years; STAI, State and trait anxiety inventory; THP, Tonic heat pain; TSL, Thermal sensory limen; t-VNS, Transcutaneous vagus nerve stimulation; VDT, Vibration detection threshold; WDT, Warm detection threshold; WUR, Wind up ratio; mS, Micro-Siemens. The study was sponsored by the device manufacturer Fa. Cerbomed (Erlangen, Germany). Apart from that, no financial or personal relationship and affiliations relevant to the subject matter of the manuscript have occurred within the last two years or are expected in future. No otherwise grants or funding have been paid. * Corresponding author. Tel.: þ499419410. E-mail address: [email protected] (V. Busch). 1935-861X/$ e see front matter Ó 2013 Elsevier Inc. All rights reserved. doi:10.1016/j.brs.2012.04.006

Recent work suggests that the vagus nerve, traditionally considered a purely parasympathetic efferent nerve, provides an exceeding important route for information into the central nervous system [1]. In the past few years, vagus nerve stimulation (VNS) has been developed as a method to physically alter relevant brain functions, thus offering a clinically useful non drug-based anticonvulsive and antidepressant therapy option [1,2]. The known anatomic projections of the vagus nerve and its association with many different brain functions involved in the perception of pain suggest that VNS might also have applications in the therapy of different pain syndromes. Several experimental animal studies in mammals have demonstrated an inhibitory effect of VNS on the electric response of spinal nociceptive neurons as well as on nociceptive behavior [3e7]. The neurophysiological data

V. Busch et al. / Brain Stimulation 6 (2013) 202e209

from these animal experiments are supported by some observational studies in humans, suggesting a pain-modulating effect of vagus nerve stimulation under conventional VNS. With regard to headache syndromes, several case reports [8e10] and one observational report on 4 therapy-refractory migraine and 2 cluster headache patients exist [11] underlining a reduction of headache frequency or intensity following VNS in patients with seizures, who concurrently suffered from migraine. Invasive VNS is an approved treatment for drug-resistant epilepsy [2]. Besides recognized clinical efficacy there are some disadvantages including the irreversible nature of the electrode implant in the majority of cases, electrode fractures, deep wound infections, transient vocal cord palsy, cardiac arrhythmia under test stimulation, electrode malfunction, and posttraumatic dysfunction of the stimulator [12]. Frequent side effects of chronic, invasive VNS such as hoarseness, cough, dyspnea, and pain are mainly due to bidirectional stimulation of efferent and afferent fibers within the mixed cervical branch of the vagus nerve. Invasiveness and adverse events of VNS have hampered the conduction of clinical trials in other indications than epilepsies. The recently introduced technique of transcutaneous vagus nerve stimulation (t-VNSÒ) [13] combines selective, non-invasive and reversible access to vagus nerve afferents with a low risk profile. t-VNS targets the cutaneous receptive field of the auricular branch of the vagus nerve at the outer ear (inner side of the tragus) [14] and has been shown to activate cerebral vagal patterns in f-MRI studies [15,16]. Several lines of evidence from anatomical and clinical studies reveal the topographic anatomy and the functional impact of the auricular branch of the vagus nerve on the autonomic nervous system [17]. Both invasive and transcutaneous VNS excite thick-myelinated fibers of vagus nerve branches that project to the main therapeutic target the nucleus of the solitary tract in the brainstem. Preclinical data emphasize equivalent anticonvulsive effects of both methods [18]. Based upon the common mode of action of invasive and transcutaneous VNS and first clinical data, the t-VNSÒ device received CE approval for the intended use in drug-resistant epilepsy and depression. Our study aimed at investigating pain perception during a t-VNS approach in a sample of 48 healthy subjects using a randomized, controlled, double-blinded cross-over design. For assessing different submodalities of peripheral and central nociception, we used the quantitative sensory testing procedure developed by the German Research Network on Neuropathic Pain including a tonic heat pain paradigm to obtain a full sensory profile of each single subject [19]. We also investigated whether t-VNS had an effect on the parameters of the autonomic nervous system (skin conductance levels, heart rates and respiration activity). Methods Study The study was approved by the local ethics committee (University of Regensburg, Germany, Proposal Nr. 09/119). Informed consent was obtained from all subjects. Subjects Forty-eight healthy subjects were finally enrolled in the study. All subjects were undergraduate students from the local university. They underwent a neurological examination and were interviewed by a psychiatrist, who additionally administered the SCID-1 Screening instrument [20,21]. Exclusion criteria were the history of a migraine, low back pain or other (chronic) pain syndromes, any cardiac or respiratory diseases, psychiatric disorders, neurological

203

syndromes or the use of psychotropic drugs. In addition, subjects were excluded with BDI Scores  18 indicating a depressive disorder [22], SOMS-2A scores  7 indicating a high physical disability [23] and STAI scores indicating increased state anxiety as compared to a normative sample of students of the same age (41) [24]. All subjects had to be without any acute pain medication for at least one week before starting the investigation. Design and randomization We used a randomized, controlled and double-blinded study design. All subjects received an active t-VNS and an inactive t-VNS (sham) using a cross-over design. After enrollment, all subjects were randomized to one of these two t-VNS branches (activeesham or shameactive). Since all pain measurements were conducted on the ipsi- and contralateral side, as described below, the sequence of measurements on both sides was randomized in the same way (ipsilateralecontralateral or contralateraleipsilateral). The experimenter, who analyzed the data, was blinded for this randomization procedure. Vagus nerve stimulation The transcutaneous vagus nerve stimulator (STV02, Cerbomed, Erlangen, Germany) consisted of a small stimulation unit and a bipolar stimulation electrode placed into the left concha at the inner side of the tragus by direct contact on the skin. The electrode was placed on an acrylic body for a comfortable fit in the pinna [25]. The skin was cleaned with a small disposable alcohol pad [26]. The stimulator could be plugged into a docking station for uploading stimulation programs and settings from a PC or storing events during a stimulation on a hard disc drive. The stimulus was a modified monophasic rectangle impulse with a pulse width of 250 mS. Stimulation frequency was kept at 25 Hz, which is known to activate vagal nerve fibers [27]. The stimulation amplitude (current intensities) could be varied between 0.25 and 10 mA. During the sham stimulation, a current intensity of 0.0 mA served as the control condition. The non-nociceptive t-VNS predominantly addressed A-beta fibers within the vagal auricular branch. Due to habituation effects a re-adjustment of stimulation intensity was necessary, which was performed on every subject within the first 5 min before starting the intervention. The current was elevated step-by-step up to a level where a constant tingling sensation was reached. The stimulation intensity was always kept below a pain threshold, thus avoiding any pricking or burning sensations. After the adjustment phase the stimulation parameters kept stable in the active group and returned to 0.0 mA in the sham group within 30 s. During the study, one of the experimenters reprogrammed the VNS device settings with a laptop computer connected via the docking station. The experimenter who interacted with the subjects and conducted all pain measurements was blinded to all device settings. The subjects were not informed about the respective order or the protocol of the stimulation sequences. Moreover, they were not instructed about the effects to be expected during the different stimulation conditions. It was explained to all subjects that both stimulation interventions would be equivalent types of “nerve stimulation”, though they might perceive them differently. Pain measurements The standardized quantitative sensory testing (QST) battery [19] assembles a comprehensive list of robust and validated short form tests representing all relevant submodalities of the somatosensory

204

V. Busch et al. / Brain Stimulation 6 (2013) 202e209

system. Furthermore, a tonic heat pain paradigm was administered exhibiting a sustained pain component [28]. A trained experimenter conducted the QST measurements according to a standardized protocol and using standardized equipment [19]. Tonic heat pain (THP): A series of contact heat pulses with a fixed frequency of repetition was applied on the dorsum of the forearm. THP measurements were conducted on the ipsilateral (left) and contralateral (right) side of the t-VNS. The sequence of THP measurements on both sides was randomized in the same way as used in the conventional QST measurements (ipsilaterale contralateral or contralateraleipsilateral). The pulse started from a level of 0.5  C below the individual heat pain threshold and increased to 0.5  C above the individual pain threshold resulting in a temperature range of 1  C. The subjects were stimulated with the pulsating heat for 5 min and were instructed to rate their pain every 20 s on the numeric rating scale. Three ratings per minute were averaged resulting in 5 final pain ratings over the course of the THP test (one rating per minute). Thermal thresholds (CDT, WDT, CPT, and HPT) for cold and hot stimuli [29] were measured using a TSA 2001-II thermal sensory testing device (Medoc, Israel). Each temperature threshold was measured by four consecutive sessions. The first session was not considered for analyses [30]. The mean threshold temperature of three sessions was calculated. Stimuli were terminated when the subject pressed a button. Cut-off temperatures were 0 and 50  C. The baseline temperature was 32  C. Furthermore, the subjects were asked if they perceived paradoxical heat sensations (PHS) during the thermal sensory limen (TSL) procedure of alternating warm and cold stimuli. Mechanical detection thresholds (MDT) were measured using modified von Frey filaments (Optihair2-Set, Marstock Nervtest, Germany) exerting forces between 0.25 and 512 mN. Mechanical pain thresholds (MPT) were measured with pinprick stimulators exerting forces between 8 and 512 mN. The Vibration detection thresholds (VDT) were tested with a Rydel-Seiffer 64 Hz tuning fork (8/8 scale) placed over the processus styloideus ulnae. Mechanical pain sensitivity (MPS) was measured using the same pinprick stimulators that were applied five times in a balanced order. The subjects were asked to rate pain for each stimulus on a numerical rating scale (NRS) ranging from 0 indicating “no pain” to 100 “most pain”. Pressure pain thresholds (PPT), defined as the minimum pressure which induces pain or discomfort [31], were tested with a pressure gauge device (FDN200, Fa. Wagner Instruments, USA) with a rod area of 1 cm2 continuously pressed into the thenar with an increasing ramp of 50 kPa/s (w0.5 kg/cm2s). For the assessment of dynamic mechanical allodynia (ALL), a set of three light tactile stimulators was used: a cotton wisp (force 3 mN), a cotton wool tip (force 100 mN) and a standardized brush (force 200e400 mN). The tactile stimuli were applied with a single stroke of approximately 2 cm in length over the skin. In addition, the subjectively perceived intensity of a single pinprick stimulus (256 mN) was compared with that of a series of 10 repetitive pinprick stimuli of the same physical intensity (1/s applied within an area of 1 cm2). The Wind up ratio (WUR) was calculated as the ratio: mean rating of the entire series divided by the mean rating of the five single stimuli. ANS parameters Changes in electrodermal activity and skin conductance levels (SCL) are related to changes in eccrine sweating, which are in turn related to activity of the sympathetic branch of the autonomic nervous system [32]. The SCL the non-dominant hand (on the palmar surface of the middle phalanx of the ring finger) were recorded using a constant-voltage device (Biofeedback Expert 2000, Schuhfried, Vienna, Austria), according to Venables and

colleagues (range: 0e50 mS; digital resolution: 0.024 mS) [32]. Heart rates (HR) were recorded using the same device (digital resolution: 0.004 beats per minute). The respiratory module measuring respiration depths (RD) and frequencies (RF) (resolution 0.2 mm, measurement range 20 cm) was a 2-channel device using a strain gauge belt, fixed around the subject’s upper abdomen. To ensure the same and correct position, the device was mounted 5 cm above the umbilicus, directly on the skin. The device produced a digitalized signal of the analog respiration movements (RD in mm/cycle, RF in cycles per minute). Time markers were synchronized with the recording of the physiological data on a common timeline. Laboratory environment The experimental room was sound-attenuated and provided with a diffuse light during the entire session. Room temperature was kept stable at 20  C [33]. The subjects were encouraged to sit comfortably in a chair and rest their hands in their laps. They looked at a spot on the wall or kept their eyes open throughout the procedure. During the experiment, the subjects were not able to watch the computer screen. Study procedure Each subject was stimulated and measured on two days at intervals of 48 h. At first, somatosensory and pain measurements were done on the ipsilateral (left) and the contralateral (right) side, or vice versa (overall 45 min). After a short break (5 min), the recording of the baseline ANS parameters was affiliated (5 min). Then, the stimulation sequence (active t-VNS or sham t-VNS) started, as described above, with an initial adjustment phase (5 min). Subsequently, a stable t-VNS stimulation period with a constant current intensity was maintained (20 min). ANS parameters were recorded during that period. Finally, the pain measurements providing the same test and side sequence were repeated (45 min) while the t-VNS continued, resulting in a total stimulation time at stable current intensities of approximately 1 h (Fig. S1). The recording of the ANS parameters was stopped before the pain measurements were repeated, since the perception of experimental pain is known to have an impact on sympathetic activity [34]. To limit any novelty effects and to get accustomed to the laboratory situation, all subjects were familiarized with both stimulation interventions (active and sham) in a training session prior to the experiment in the laboratory room with the same equipment which was used later in the study. Statistics Outcome variables, appropriate statistical methods and data handling rules were pre-defined in a statistical analysis plan (SAP). The statistical analyses were performed using PASW for Windows 18.0.2 (SPSS Science, Chicago, Illinois, USA). The sample size was estimated using standard software (G-Power; University of Duesseldorf, Germany). The significance level was set to P ¼ 0.05 (twosided). The primary analyses involving repeated measures ANOVA models were performed without correction of the Type I family wise error rate. This strategy was selected because this was a preliminary pilot study and we wanted to utilize a liberal definition of statistical significance to avoid potential misinterpretation of the data. Since this was a pilot study with multiple primary endpoints, no exact sample size calculation was possible. According to Cohen’s effect size d [35] with an assumption of small (d ¼ 0.3) to medium (d ¼ 0.5) effects, a sample size of 48 subjects was considered to be appropriate for the investigation.

V. Busch et al. / Brain Stimulation 6 (2013) 202e209

To assign a subject to one of the two t-VNS interventions (activeesham or shameactive) and to one of the two side sequences (ipsilateralecontralateral or contralateraleipsilateral), Efron’s Biased Coin randomization with Efron’s P ¼ 0.7 was used [36]. Continuous raw data were summarized as mean and standard deviation. Categorical raw data were expressed as frequency counts and percentages. Tonic heat pain Before VNS effects on tonic heat pain ratings over the course of 5 min were analyzed, the five time points were summarized into one variable “time” by calculating the area under the curve using the trapezoid rule [37]. The AUCs were calculated for sham t-VNS as well as active t-VNS (each before and during the stimulation). Then, two-way analyses of variance for repeated measures using the same within-subject factors and baseline values as covariates were performed as described above. Partial eta squared h2P [38] was used for the measurement of effect size for repeated measure ANOVAs. Thermal and mechanical thresholds A log-transformation was addressed to most of the QST parameters (CDT, WDT, TSL, MDT, MPT, MPS, WUR and PPT) according to the recommendation of the QST standard protocol [19]. Two-way analyses of variance (ANOVA) for repeated measures were then used to investigate the effect of the stimulation on QST parameters. Within-subject factors were “stimulation” (active vs. sham) and “side” (ipsilateral vs. contralateral). The differences of the mean baseline values (pooled ipsi- and contralateral sides) were used as covariates to eliminate confounding effects and raise the power of the results [39]. Because both within-subject factors were dichotomous, no test of sphericity was needed. In the case of an interaction effect between the main factors, a one-way analysis of variance for repeated measures with the factor “stimulation” was performed for each body side. For the illustration of mean QST values, the geometric mean was calculated (de-log transformation) as recommended in Ref. [40,41]. Respiration depths and rates, heart rates and skin conductance levels were tested for normal distribution using Kolmogorove Smirnov Tests. Paired Student t-tests were conducted for the comparison of parameter changes between both interventions (active t-VNS vs. sham t-VNS). Results

205

the sham t-VNS, the same increase of pain perception could be observed on both sides. In contrast, significantly reduced pain ratings in each of the 5 min without the tendency to further increase over time were found on the ipsi- and contralateral side (Fig. 1). Using ANOVA, a significant effect of the main factor “stimulation”, but not for the main factor “side” or the interaction of “stimulation * side” was found indicating an inhibitory effect of t-VNS on both sides (Table 2). Partial eta square (effect size) was h2P ¼ 0:237. Mechanical detection and pain stimuli ANOVA showed a significant effect for the interaction of “stimulation * side” for MPT as well as for MPS, both indicating an ipsilateral decrease of nociceptive perception for mechanical pain stimuli following left-sided t-VNS (Figs. S2 and S3). Partial Eta Square for MPS was h2P ¼ 0:144 and for MPT was h2P ¼ 0:144. Furthermore, the interaction of “stimulation * side” for PPT was significant, indicating an ipsilateral decrease of deep tissue pain following t-VNS (Fig. S4). Partial Eta Square for PPT was h2P ¼ 0:091. MDT showed a significant side effect but no effect of the vagus nerve stimulation. Although a slight wind up following repetitive trains of pinprick stimuli was present in the subjects, the wind up ratios (WUR) were not altered by t-VNS (Table 2). Thermal detection and pain stimuli Thermal detection and pain thresholds (including CDT, WDT, CPT and HPT) as well as the sensory limen for temperature changes (TSL) did not show any significant changes during either active or sham t-VNS (Table 2). ANS parameters The respiration parameters (RD, RF), heart rates (HR) as well as the skin conductance levels (SCL) were normally distributed. The respiration frequencies and heart rates were not significantly altered during active or sham t-VNS. The respiration depths decreased during active and sham t-VNS, though these changes were only slight. Skin conductance levels increased significantly following both interventions to a similar extent. Though RD and SCL changed during the stimulation, these changes were not specific to the type of stimulation, since the comparison of these changes between sham t-VNS and active t-VNS did not reveal any significant differences (Table 3).

Subjects Tolerance and side effects Two subjects were not compliant to the protocol and were excluded from further analyses. They were replaced by two new subjects recruited at the end of the study. The final sample consisted of 48 healthy students (24 female, 24 male) with a mean age of 23.3  2.1 years (range 20e28 years). All of these subjects completed the study. The mean BDI score of the final sample was 2.1  2.1. The mean SOMS-2a score was 1.2  1.6. The mean STAI (state) score was 31.2  6.4. Pain measurements Only one subject revealed VDT less than 8/8 before t-VNS. Likewise, only one subject showed a slightly increased baseline ALL, which remained stable before and after the interventions. Only two subjects stated PHS during the sensory limen procedure. Therefore, these parameters were not used for further analyses (Table 1). Tonic heat pain The baseline pain ratings during the application of tonic heat over 5 min increased on both sides and to a similar extent. During

The t-VNS was well tolerated and no subject discontinued the stimulation. Mean stimulation intensity during the active VNS was 1.6 mA  1.5 mA. Sensations at the site of stimulation (number of patients, active/sham) were a feeling of “slight pain” (2/1), “pressure” (8/6), “prickling” (12/2), “itching” (10/1) or “tickling” (7/2). During the active stimulation, one subject stated a “strange feeling” in his stomach region and one subject stated a transient “irritation of swallowing”. No serious adverse events including acoustic or vestibular reactions were observed. No subject reported on any symptoms after stopping the stimulation. Discussion Summary of findings Based on a randomized, double-blinded and controlled crossover design, our study focused on the effect of t-VNS on human pain processing by measuring a set of relevant submodalities of the somatosensory system using a comprehensive quantifiable testing

206

V. Busch et al. / Brain Stimulation 6 (2013) 202e209

Table 1 Raw data of all QST parameters and areas under the curve for THP, each before and during t-VNS (active or sham) on the ipsilateral (left) and the contralateral (right) side of the stimulation. Parameters

Active stimulation Before

CDT ( C) 

WDT ( C) TSL (D C) PHS (x/3) CPT ( C) HPT ( C) MDT (mN) MPT (mN) MPS (rating) ALL (ratio) WUR (ratio) VDT (x/8) PPT (kPa) THP (rating)

Ipsilateral Contralateral Ipsilateral Contralateral Ipsilateral Contralateral Ipsilateral Contralateral Ipsilateral Contralateral Ipsilateral Contralateral Ipsilateral Contralateral Ipsilateral Contralateral Ipsilateral Contralateral Ipsilateral Contralateral Ipsilateral Contralateral Ipsilateral Contralateral Ipsilateral Contralateral Ipsilateral Contralateral

Sham stimulation During

Before

During

M

Sd

M

Sd

M

Sd

M

Sd

30.27 30.32 33.87 33.93 3.04 2.79 0.10 0.06 14.84 13.83 43.39 43.71 1.41 1.62 197.90 216.13 2.72 2.17 0.04 0.04 3.01 2.93 8.00 8.00 413.25 449.01 19.49 20.06

1.50 1.12 1.24 0.90 2.37 1.05 0.59 0.32 8.91 8.78 3.08 2.91 1.71 2.17 176.28 186.43 4.36 3.09 0.27 0.27 2.57 2.44 0.00 0.00 126.02 171.88 9.87 9.83

29.66 29.72 34.62 34.49 3.45 3.49 0.04 0.04 14.84 14.25 42.45 43.00 1.59 1.47 169.93 139.69 2.09 2.47 0.05 0.06 3.23 3.13 8.00 8.00 431.51 432.93 15.16 15.14

1.36 1.28 1.36 1.20 1.68 1.66 0.29 0.20 8.44 8.77 2.80 2.46 1.85 2.23 156.95 108.05 2.77 3.62 0.33 0.43 2.68 2.84 0.00 0.00 146.37 136.01 8.14 8.35

30.59 30.51 33.93 33.96 2.46 2.71 0.02 0.08 14.31 13.68 43.71 43.23 1.27 1.00 213.74 204.57 2.86 2.15 0.05 0.03 2.99 2.88 8.00 7.99 412.97 430.36 19.90 19.57

0.80 1.07 1.11 0.68 1.26 1.41 0.14 0.35 8.59 8.34 3.10 3.12 1.65 0.77 181.56 150.38 4.55 3.05 0.31 0.21 2.86 2.10 0.00 0.05 116.74 155.53 10.03 10.95

29.77 29.84 34.56 34.50 3.53 3.46 0.06 0.02 15.56 14.61 42.89 42.96 1.92 1.67 150.38 170.16 3.16 3.04 0.05 0.04 3.23 3.00 8.00 7.99 402.28 444.99 18.38 17.09

1.33 1.09 1.18 0.85 1.59 1.47 0.32 0.14 7.51 8.32 2.56 2.72 2.36 3.06 149.98 152.78 4.40 4.42 0.30 0.26 2.98 2.61 0.00 0.05 116.42 162.62 10.99 9.02

CDT, cold detection threshold; WDT, warm detection threshold; TSL, thermal sensory limen; CPT, cold pain threshold; HPT, heat pain threshold; MDT, mechanical detection threshold; MPT, mechanical pain threshold; MPS, mechanical pain sensitivity; ALL, dynamic mechanical allodynia; WUR, wind up ratio; VDT, vibration; PPT, pressure pain threshold; THP, tonic heat pain; M, mean; Sd, standard deviation;  C, Degrees Celsius; mN, 1 N/1000; kPa, kilo Pascal.

protocol (including a tonic heat pain paradigm) before and during active t-VNS or sham t-VNS applications in a sample of 48 healthy volunteers. We found a bilateral inhibition of pain sensitivity during 5 min of tonic heat pain following 1 h of a continuous t-VNS. Moreover, we could demonstrate an ipsilateral increase of mechanical and pressure pain thresholds and a reduction of mechanical pain sensitivity. In contrast, no thermal or mechanical detection thresholds were altered due to t-VNS. Our results indicate a t-VNS-induced inhibition of pain perception in response to different experimental pain modalities not interfering with innocuous somatosensory processing. Since the inhibitory modulation of pain sensitivity was not observable following the sham intervention, we suggest a t-VNS specific effect. We can further exclude the possibility that these findings were due to a learning effect, since the intervention as well as the test sequence was randomized. In contrast to our results, Borckardt and colleagues found pronociceptive effects following conventional VNS (lower pain tolerance) in a small observational study on patients with depression (n ¼ 8). Pain thresholds or stimulus-response functions were not examined. We would prefer to interpret these results cautiously, since question and design of that study is only marginally comparable to our present work. Interestingly, the experimental data did not mirror some of the clinical statements of the subjects in that study. Two patients reported on an improvement of low back pain, one patient stated a significant relief of his tension-type headache. In concert with our findings, Kirchner and colleagues investigated experimental pain in patients with epilepsy (n ¼ 10) and found a reduction of tonic pressure pain and heat pain associated with trains of consecutively applied heat pain stimuli (WUR), whereas

heat pain thresholds remained unchanged following long-term conventional VNS {Kirchner, 2000 #10301}. In summary, the findings of the previously published studies suggest that VNS may not effect pain perception consistently, but rather depends on pain modalities, which are investigated, or mood states of the VNS recipients during the stimulation. Our study is the first investigating the effects of a transcutaneous VNS on pain processing in healthy subjects. Small to moderate effect sizes for tonic heat pain and pain threshold changes following one single hour of t-VNS may suggest a clinical relevance of our results and may serve as rationale to investigate t-VNS in clinical pain syndromes. Pain measurements THP In a former study, tonic heat pain was found to increase pain perception, though the peripheral C and A-delta nociceptor activity decreased [42], suggesting an involvement of central mechanisms of pain encoding following tonic heat by temporal summation [43]. Our data indicate that t-VNS may have influenced pain processings mainly on a central rather than on a peripheral level. Both spinal and supraspinal mechanisms may have contributed to the pain inhibitory effect of t-VNS in our sample. The assumption of a vagal influence on spinal segments is strengthened by animal studies showing that the activation of vagal afferents decreased the activity of second-order nociceptive neurons in the spinothalamic and spinoreticular tract of the spinal cord [44,45] as well as in the trigeminal nuclear complex [46]. Similar to our results, Kirchner and colleagues found an increase of tonic pain thresholds in epilepsy patients during an invasive VNS,

V. Busch et al. / Brain Stimulation 6 (2013) 202e209

207

solitary nucleus, the parabrachial nucleus, the periaqueductal gray, and secondarily to the amygdala, the hypothalamus and other limbic regions. Accordingly, vagal afferents carry any sensory information initially to the solitary nucleus [48] and reach higher brain structures that are associated with mood regulation, emotional processing and even pain perception [49]. Therefore, it is worth considering, whether vagus nerve stimulation, at least to some extent, might have a significant impact on higher brain regions, such as affective brain areas like the amygdalae, which are involved in the perception of tonic pain [50]. A recent f-MRI study revealed a significant decrease in the activity of limbic brain areas (including the amygdalae) along with a significant improvement of well-being induced by t-VNS in 22 healthy subjects [15], thus underlining the significance of a supraspinal vagal influence on pain perception. MPT/MPS Mechanical pain thresholds slightly decreased following active and sham t-VNS. However, the attenuation of the decrease following t-VNS on the left side suggests an inhibitory effect on threshold determination of mechanical pain stimuli. Likewise, the pain sensitivity to mechanical stimuli significantly decreased following active t-VNS, thus indicating a flattening of the stimulusresponse pattern to mechanical pain stimuli.

Figure 1. Heat pain ratings over 5 min of tonic heat application. Error bars represent the mean pain ratings (SEM) on the NRS (0e10) of verum and sham VNS, each on the ipsilateral (left) and contralateral (right) side of stimulation. NRS ¼ numeric rating scale, VNS ¼ vagus nerve stimulation.

though they used a tonic mechanical pain paradigm [47]. The pain ratings following noxious tonic heat in our subjects were inhibited bilaterally, further arguing for the assumption of a central impact on vagus nerve stimulation, since the left vagus nerve was described to enter the medulla bilaterally [48]. A possible supraspinal impact of t-VNS on nociception may be derived from the anatomical course of vagal as well as nociceptive nerve branches. Spinothalamic neurons project to the medulla, the

Table 2 Two-way repeated analyses of variance using “stimulation” (active vs. sham) and “side” (ipsilateral vs. contralateral) as within-subject factors and the difference of the baseline values for each stimulation as covariate. As described in the methods, for most of the QST parameters a log-transformation had to be addressed. Parameters*

Stimulation F1,

(log)

CDT (log) WDT (log) TSL CPT HPT (log) MDT (log) MPT (log) MPS (log) WUR (log) PPT THP

46

0.01 0.05 1.71 1.38 0.38 1.96 0.04 2.41 0.75 0.43 14.30

Side

P

F1,

ns(0.97) ns(0.82) ns(0.20) ns(0.25) ns(0.54) ns(0.17) ns(0.84) ns(0.13) ns(0.39) ns(0.52) <0.001

0.22 0.43 0.42 1.46 1.28 4.64 0.07 2.01 1.49 3.80 0.87

46

Interaction stimulation * side P

F1,

ns(0.84) ns(0.52) ns(0.52) ns(0.23) ns(0.26) 0.037 ns(0.80) ns(0.16) ns(0.23) ns(0.06) ns(0.36)

0.05 0.10 0.03 0.15 1.15 0.75 7.72 6.55 0.02 4.61 1.01

46

P ns(0.83) ns(0.76) ns(0.86) ns(0.70) ns(0.29) ns(0.39) 0.008 0.014 ns(0.97) 0.037 ns(0.32)

F, ANOVA test statistic with degrees of freedom (model, error); CDT, cold detection threshold; WDT, warm detection threshold; TSL, thermal sensory limen; CPT, cold pain threshold; HPT, heat pain threshold; MDT, mechanical detection threshold; MPT, mechanical pain threshold; MPS, mechanical pain sensitivity; WUR, wind up ratio; PPT, pressure pain thresholds; THP, tonic heat pain; ns, non-significant; P, significance. Bold value signifies P < 0.05.

PPT The threshold determination of pressure pain stimuli was thought to reflect the more generalized excitability of second-order neurons in the spinal cord rather than the purely peripheral nociceptor activity from deep tissues [51]. This view is additionally supported by the fact that general deep pain sensitivity was found to be altered in different chronic pain syndromes, regardless of local pain complaints [52e54]. Similarly, in a study on healthy women the temporal summation of afferent nerve signals due to mechanically induced pressure pain of deep muscles was found to be more pronounced compared to a painful stimulation of the skin [55]. Therefore, the ipsilateral increase in pressure pain thresholds following t-VNS on the left side in our study may indicate a sidespecific reduction of central pain processing in our subjects. However, though the interaction of “stimulation * side” was significant, we would prefer to interpret these results cautiously. The disordinal curve progression demonstrates that changes following active t-VNS on the ipsilateral side were similar to the changes following sham t-VNS on the contralateral side. Therefore, these changes may be explained by the natural variability of these values rather than a specific vagal stimulation effect. WUR We did not find lowered pain sensitivity to repetitive pain stimuli following t-VNS. Since we investigated healthy subjects, who demonstrated only a slight wind up phenomenon initially, we think that a potential t-VNS effect on WUR might have emerged in a sample of patients presenting allodynia/hyperalgesia (defined as a sensitization of central pain signaling neurons due to temporal summation effects, i.e. in neuropathic pain syndromes) [56]. Thermal thresholds Though we found an attenuation of mechanical pain perception, no changes of thermal pain perception could be detected following t-VNS in our sample. These results are not contradictory, since both mechanical and thermal pain within the same subjects was often reported to be correlated relatively low [57,58]. A recent factor analysis identified different response patterns to experimental pain modalities (i.e. thermal vs. mechanical) suggesting that the concept of “human pain sensitivity” does not reflect a uniform pain response tendency [59].

208

V. Busch et al. / Brain Stimulation 6 (2013) 202e209

Table 3 Parameters of the autonomic nervous system before and during active and sham t-VNS using paired Student t-tests. Parameter changes due to the intervention were compared between sham and verum t-VNS. Active stimulation

Before

RD (cm) RF (cpm) HR (bpm) SCL (mS)

Sham stimulation

During

T

M

Sd

M

Sd

0.76 16.27 65.94 1.36

0.26 3.52 8.56 1.12

0.68 16.85 64.33 2.27

0.20 2.69 9.30 1.93

3.18 1.77 2.64 6.12

P

0.003 ns (0.08) ns (0.11) <0.001

Before

Comparison of the changes (active vs. sham) During

M

Sd

M

Sd

0.74 16.65 66.44 1.34

0.30 2.91 8.35 0.96

0.68 16.89 65.73 2.52

0.21 2.59 8.20 1.92

T

P

T

P

2.74 0.74 1.44 6.50

0.009 ns (0.46) ns (0.16) <0.001

0.51 0.89 1.44 1.43

ns ns ns ns

(.0.61) (.0.37) (.0.16) (.0.16)

RD, respiration depths in cm; RF, respiration frequencies in cycle per minute; SCL, skin conductance levels in micro-Siemens; HR, heart rates in beats per minute; M, mean; Sd, standard deviation; T, student t-test Statistic; P, significance. Bold value signifies P < 0.05.

Detection thresholds Non-noxious somatosensory perception was not altered by t-VNS, suggesting that the vagus nerve stimulation may have little influence on the somatosensory processing transmitted via A-beta fibers in healthy humans. ANS parameters The skin conductance levels increased during both interventions indicating a marginal sympathetic shift, which may have occurred because the subjects were strained or attentive to the stimulation procedure. These data substantiate the fact that there was no relaxation effect, which might have contributed to the effect of t-VNS on a reduced pain perception. Respiration depths decreased during active and sham t-VNS. However, these changes were only slight and do not point to a specific effect of t-VNS on respiration depths. Tolerance and side effects In addition, no serious adverse events were observed during 1 h of continuous t-VNS. Some mild unpleasant sensation during the stimulation discontinued immediately after the stimulation was stopped. Moreover, some of these sensations cannot unambiguously be referred to the vagus nerve stimulation, as they were also perceived during the sham intervention. For instance, sensations of “pressure” may have been caused to some extent by the mounting of the transcutaneous vagus nerve stimulator in the pinna. Taken together, no relevant alteration of cardiac or breathing activity or clinical relevant side effects following 1 h of t-VNS with mean current intensities of 1.6 mA were observed in our sample. Limitations of the study One could argue about the possibility of accidental trigeminal nerve stimulation in our study due to the heterogeneous sensory innervation of the outer ear. Peuker and colleagues investigated the complete course of nerve supply of the outer ear in 14 cadavers and found four different nerves being involved in the innervation of the outer side of the tragus (N.auricularis mj 45%, N.auriculotemporalis 9% and a mixture of trigeminal and vagus nerves in all other cases) [14]. However, the transcutaneous stimulation of vagus nerve fibers in our study was conducted at the inner side of the tragus being part of the cavum conchae. Anatomical studies revealed that the auricular branch of the vagus nerve solely provided 45% ramification for the cavity of conchae (100% for the cymba conchae). A mixture of the auricular branch of the vagus nerve and great auricular nerve fibers was found in the remaining 55%. No region with triple innervation involving trigeminal nerve fibers was found [14]. Therefore, though we cannot definitely exclude a stimulation of some auricular nerve fibers, a trigeminal nerve fiber involvement during t-VNS at the inner side of the tragus is unlikely.

Though we were able to show a reduction of pain sensitivity during vagus nerve stimulation using a transcutaneous approach in healthy humans, we are aware of the preliminary value of our study. Notwithstanding these promising results, a clinical relevance for the treatment of patients suffering from chronic pain disorders cannot be deduced from our preliminary data. Therefore, further studies should address the pain-modulating effects of t-VNS on both clinical and experimental pain modalities in different chronic pain syndromes (i.e. neuropathic pain or migraine). Conclusion Taken together, our results demonstrate that t-VNS decisively influences pain processings in healthy humans. To the best of our knowledge, this study is the first demonstrating an inhibitory effect of a continuous transcutaneous vagus nerve stimulation on different pain modalities in healthy subjects. Based on a controlled experimental study design, we could contribute to the most recent work identifying an inhibitory effect of t-VNS, specific for mechanical and tonic heat pain stimuli, whereas the non-noxious somatosensory perception was not affected. In addition, detailed analyses of different submodalities of the somatosensory system suggest an impact of tVNS on central pain processing rather than on peripheral nociceptor activity. Since the stimulation procedure of 1 h of continuous t-VNS was well tolerated and no relevant alterations of the autonomic nervous systems occurred during the stimulation procedure, t-VNS may hold the potential to efficiently alleviate pain, especially in those patients suffering from chronic pain syndromes and lacking significant drug mediated clinical effects. Thus, we think that our pilot study may provide the rational basis for further investigations of t-VNS in patients with different chronic pain syndromes. Authors approval All authors have personally reviewed and given final approval of the version submitted, and neither the manuscript, nor its data have been previously published or are currently under consideration of publication. Supplementary data Supplementary data related to this article can be found online at doi:10.1016/j.brs.2012.04.006. References [1] Beekwilder JP, Beems T. Overview of the clinical applications of vagus nerve stimulation. J Clin Neurophysiol 2010;27(2):130e8. Epub 2010/05/28. [2] Hayes W. Vagus nerve stimulation for epilepsy. Health technology assessment e final report. Portland, Oregon: Center for Evidence-based Policy at Oregon Health & Science University; 2009.

V. Busch et al. / Brain Stimulation 6 (2013) 202e209 [3] Randich A, Gebhart GF. Vagal afferent modulation of nociception. Brain Res Brain Res Rev 1992;17(2):77e99. Epub 1992/05/01. [4] Maixner W, Bossut DF, Whitsel EA. Evaluation of vagal afferent modulation of the digastric reflex in cats. Brain Res 1991;560(1e2):55e62. Epub 1991/09/27. [5] Ammons WS, Blair RW, Foreman RD. Vagal afferent inhibition of spinothalamic cell responses to sympathetic afferents and bradykinin in the monkey. Circ Res 1983;53(5):603e12. Epub 1983/11/01. [6] Bohotin C, Scholsem M, Multon S, Martin D, Bohotin V, Schoenen J. Vagus nerve stimulation in awake rats reduces formalin-induced nociceptive behaviour and fos-immunoreactivity in trigeminal nucleus caudalis. Pain 2003;101(1e2):3e12. Epub 2003/01/01. [7] Ren K, Randich A, Gebhart GF. Vagal afferent modulation of a nociceptive reflex in rats: involvement of spinal opioid and monoamine receptors. Brain Res 1988;446(2):285e94. Epub 1988/04/19. [8] Hord ED, Evans MS, Mueed S, Adamolekun B, Naritoku DK. The effect of vagus nerve stimulation on migraines. J Pain 2003;4(9):530e4. Epub 2003/11/26. [9] Sadler RM, Purdy RA, Rahey S. Vagal nerve stimulation aborts migraine in patient with intractable epilepsy. Cephalalgia 2002;22(6):482e4. Epub 2002/07/23. [10] Mauskop A. Vagal nerve stimulation for refractory migraine. Neurology 2001; 56(8):A65e6. [11] Mauskop A. Vagus nerve stimulation relieves chronic refractory migraine and cluster headaches. Cephalalgia 2005;25(2):82e6. Epub 2005/01/22. [12] Spuck S, Tronnier V, Orosz I, Schonweiler R, Sepehrnia A, Nowak G, et al. Operative and technical complications of vagus nerve stimulator implantation. Neurosurgery 2010;67(2 Suppl. Operative):489e94. Epub 2010/11/26. [13] Ventureyra EC. Transcutaneous vagus nerve stimulation for partial onset seizure therapy. A new concept. Childs Nerv Syst 2000;16(2):101e2. Epub 2000/02/09. [14] Peuker ET, Filler TJ. The nerve supply of the human auricle. Clin Anat 2002; 15(1):35e7. Epub 2002/02/09. [15] Kraus T, Hosl K, Kiess O, Schanze A, Kornhuber J, Forster C. BOLD fMRI deactivation of limbic and temporal brain structures and mood enhancing effect by transcutaneous vagus nerve stimulation. J Neural Transm 2007;114(11): 1485e93. Epub 2007/06/15. [16] Chae JH, Nahas Z, Lomarev M, Denslow S, Lorberbaum JP, Bohning DE, et al. A review of functional neuroimaging studies of vagus nerve stimulation (VNS). J Psychiatr Res 2003;37(6):443e55. Epub 2003/10/18. [17] Ellrich J. Transcutaneous vagus nerve stimulation. Eur Neurol Rev 2011;6: 262e4. [18] He W. A new concept of transcutaneous vagus nerve stimulation for epileptic seizure. Soc Neurosci Abstr 2009;4(539). [19] Rolke R, Baron R, Maier C, Tolle TR, Treede RD, Beyer A, et al. Quantitative sensory testing in the German Research Network on Neuropathic Pain (DFNS): standardized protocol and reference values. Pain 2006;123(3):231e43. Epub 2006/05/16. [20] Wittchen HU, Zaudig M, Fydrich T. SKID e Strukturiertes Klinisches Interview für DSM-IV. Göttingen: Hogrefe-Verlag; 1997. [21] Riley J, Benson M, Gremillion H. Sleep disturbance in orofacial pain patients: pain-related or emotional distress? Cranio 2001;19:106e13. [22] Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for measuring depression. Arch Gen Psychiatry 1961;4:561e71. Epub 1961/06/01. [23] Rief W, Hiller W. Toward empirically based criteria for the classification of somatoform disorders. J Psychosom Res 1999;46(6):507e18. Epub 1999/ 08/24. [24] Spielberger CD, Gorsuch RL, Lushene RE. Manual for the State-Trait Anxiety Inventory. Palo Alto; 1970. [25] Dietrich S, Smith J, Scherzinger C, Hofmann-Preiss K, Freitag T, Eisenkolb A, et al. A novel transcutaneous vagus nerve stimulation leads to brainstem and cerebral activations measured by functional MRI. Biomed Tech (Berl) 2008; 53(3):104e11. Epub 2008/07/08. [26] Venables PH, Christie I. Electrodermal activity. In: Martin I, Venables PH, editors. Techniques in psychophysiology. Chichester: Wiley; 1973. [27] Vonck K, Boon P, D’Have M, Vandekerckhove T, O’Connor S, De Reuck J. Longterm results of vagus nerve stimulation in refractory epilepsy. Seizure 1999; 8(6):328e34. Epub 1999/10/08. [28] Lautenbacher S, Roscher S, Strian F. Tonic pain evoked by pulsating heat: temporal summation mechanisms and perceptual qualities. Somatosens Mot Res 1995;12(1):59e70. Epub 1995/01/01. [29] Fruhstorfer H, Lindblom U, Schmidt WC. Method for quantitative estimation of thermal thresholds in patients. J Neurol Neurosurg Psychiatry 1976;39(11): 1071e5. Epub 1976/11/01. [30] Yarnitsky D, Sprecher E, Zaslansky R, Hemli JA. Multiple session experimental pain measurement. Pain 1996;67(2e3):327e33. Epub 1996/10/01. [31] Fischer AA. Pressure algometry over normal muscles. Standard values, validity and reproducibility of pressure threshold. Pain 1987;30(1):115e26. Epub 1987/07/01.

209

[32] Venables PH, Martin I. The relation of palmar sweat gland activity to level of skin potential and conductance. Psychophysiology 1967;3(3):302e11. Epub 1967/01/01. [33] Boucsein W. Elektrodermal activity. Berlin: Springer; 1980. [34] Janig W. The sympathetic nervous system in pain. Eur J Anaesthesiol Suppl 1995;10:53e60. Epub 1995/05/01. [35] Cohen J. Statistical power analysis for the behavioral sciences. In: Hillsdale N, editor. Lawrence Earlbaum Associates; 1988. [36] Efron B. Forcing a sequential experiment to be balanced. Biometrika 1971;58: 403e17. [37] Matthews JN, Altman DG, Campbell MJ, Royston P. Analysis of serial measurements in medical research. BMJ 1990;300:230e5. [38] Keppel G. Design and analysis: a researcher’s handbook. Englewood Cliffs, NJ: Prentice Hall; 1991. [39] Wallenstein S, Fleiss JL. The two-period crossover design with baseline measurements. Commun Stat Theory Methods 1988;17(10). ISSN: 1532-415X: 3333e43. [40] Bland JM, Altman DG. Statistics notes. Logarithms. BMJ 1996;312(7032):700. Epub 1996/03/16. [41] Bland JM, Altman DG. The use of transformation when comparing two means. BMJ 1996;312(7039):1153. Epub 1996/05/04. [42] Adriaensen H, Gybels J, Handwerker HO, Van Hees J. Suppression of C-fibre discharges upon repeated heat stimulation may explain characteristics of concomitant pain sensations. Brain Res 1984;302(2):203e11. Epub 1984/06/08. [43] Andrew D, Greenspan JD. Peripheral coding of tonic mechanical cutaneous pain: comparison of nociceptor activity in rat and human psychophysics. J Neurophysiol 1999;82(5):2641e8. Epub 1999/11/24. [44] Thies R, Foreman RD. Inhibition and excitation of thoracic spinoreticular neurons by electrical stimulation of vagal afferent nerves. Exp Neurol 1983; 82(1):1e16. Epub 1983/10/01. [45] Chandler MJ, Hobbs SF, Bolser DC, Foreman RD. Effects of vagal afferent stimulation on cervical spinothalamic tract neurons in monkeys. Pain 1991; 44(1):81e7. Epub 1991/01/01. [46] Bossut DF, Maixner W. Effects of cardiac vagal afferent electrostimulation on the responses of trigeminal and trigeminothalamic neurons to noxious orofacial stimulation. Pain 1996;65(1):101e9. Epub 1996/04/01. [47] Kirchner A, Birklein F, Stefan H, Handwerker HO. Left vagus nerve stimulation suppresses experimentally induced pain. Neurology 2000;55(8):1167e71. Epub 2000/11/09. [48] Nemeroff CB, Mayberg HS, Krahl SE, McNamara J, Frazer A, Henry TR, et al. VNS therapy in treatment-resistant depression: clinical evidence and putative neurobiological mechanisms. Neuropsychopharmacology 2006;31(7): 1345e55. Epub 2006/04/28. [49] Berthoud HR, Neuhuber WL. Functional and chemical anatomy of the afferent vagal system. Auton Neurosci 2000;85(1e3):1e17. Epub 2001/02/24. [50] Borckardt JJ, Kozel FA, Anderson B, Walker A, George MS. Vagus nerve stimulation affects pain perception in depressed adults. Pain Res Manag 2005; 10(1):9e14. Epub 2005/03/23. [51] Fernandez-de-las-Penas C, Galan-del-Rio F, Fernandez-Carnero J, Pesquera J, Arendt-Nielsen L, Svensson P. Bilateral widespread mechanical pain sensitivity in women with myofascial temporomandibular disorder: evidence of impairment in central nociceptive processing. J Pain 2009;10(11):1170e8. Epub 2009/07/14. [52] Birklein F, Handwerker HO. Complex regional pain syndrome: how to resolve the complexity? Pain 2001;94(1):1e6. Epub 2001/09/29. [53] Desmeules JA, Cedraschi C, Rapiti E, Baumgartner E, Finckh A, Cohen P, et al. Neurophysiologic evidence for a central sensitization in patients with fibromyalgia. Arthritis Rheum 2003;48(5):1420e9. Epub 2003/05/15. [54] Sterling M, Jull G, Vicenzino B, Kenardy J. Sensory hypersensitivity occurs soon after whiplash injury and is associated with poor recovery. Pain 2003;104(3): 509e17. Epub 2003/08/21. [55] Nie H, Arendt-Nielsen L, Andersen H, Graven-Nielsen T. Temporal summation of pain evoked by mechanical stimulation in deep and superficial tissue. J Pain 2005;6(6):348e55. Epub 2005/06/10. [56] Ziegler EA, Magerl W, Meyer RA, Treede RD. Secondary hyperalgesia to punctate mechanical stimuli. Central sensitization to A-fibre nociceptor input. Brain 1999;122(Pt 12):2245e57. Epub 1999/12/03. [57] Arendt-Nielsen L, Lautenbacher S. Pathophysiology of pain perception. In: Lautenbacher S, Fillingim R, editors. Assessment of pain perception. New York: Kluwer Academic; 2004. p. 25e42. [58] Janal MN, Glusman M, Kuhl JP, Clark WC. On the absence of correlation between responses to noxious heat, cold, electrical and ischemic stimulation. Pain 1994;58(3):403e11. Epub 1994/09/01. [59] Hastie BA, Riley 3rd JL, Robinson ME, Glover T, Campbell CM, Staud R, et al. Cluster analysis of multiple experimental pain modalities. Pain 2005;116(3): 227e37. Epub 2005/06/21.