Research Paper
Pain hypersensitivity and spinal nociceptive hypersensitivity in chronic pain: prevalence and associated factors Michele Curatoloa,*, Monika Muller ¨ b,c, Aroosiah Ashrafb, Alban Y. Nezirid, Konrad Streitbergerb, Ole K. Andersene, e Lars Arendt-Nielsen
Abstract Hypersensitivity of pain pathways is considered a relevant determinant of symptoms in chronic pain patients, but data on its prevalence are very limited. To our knowledge, no data on the prevalence of spinal nociceptive hypersensitivity are available. We studied the prevalence of pain hypersensitivity and spinal nociceptive hypersensitivity in 961 consecutive patients with various chronic pain conditions. Pain threshold and nociceptive withdrawal reflex threshold to electrical stimulation were used to assess pain hypersensitivity and spinal nociceptive hypersensitivity, respectively. Using 10th percentile cutoff of previously determined reference values, the prevalence of pain hypersensitivity and spinal nociceptive hypersensitivity (95% confidence interval) was 71.2 (68.3-74.0) and 80.0 (77.0-82.6), respectively. As a secondary aim, we analyzed demographic, psychosocial, and clinical characteristics as factors potentially associated with pain hypersensitivity and spinal nociceptive hypersensitivity using logistic regression models. Both hypersensitivity parameters were unaffected by most factors analyzed. Depression, catastrophizing, pain-related sleep interference, and average pain intensity were significantly associated with hypersensitivity. However, none of them was significant for both unadjusted and adjusted analyses. Furthermore, the odds ratios were very low, indicating modest quantitative impact. To our knowledge, this is the largest prevalence study on central hypersensitivity and the first one on the prevalence of spinal nociceptive hypersensitivity in chronic pain patients. The results revealed an impressively high prevalence, supporting a high clinical relevance of this phenomenon. Electrical pain thresholds and nociceptive withdrawal reflex explore aspects of pain processing that are mostly independent of sociodemographic, psychological, and clinical pain-related characteristics. Keywords: Pain sensitivity, Central sensitization, Quantitative sensory tests, Nociceptive withdrawal reflex, Prevalence
1. Introduction There is overwhelming evidence for hypersensitivity of nociceptive pathways in pain states, leading to pain amplification.22 Pain hypersensitivity can be measured in humans using quantitative sensory tests (QSTs). Spinal nociceptive hypersensitivity is typically assessed as electromyography response to an electrical stimulus applied to the lower extremity (nociceptive withdrawal reflex).18 Groups of patients with different pain conditions have higher pain sensitivity than groups of pain-free subjects.3 Pain patients display lower thresholds of nociceptive withdrawal reflex, compared with healthy subjects.9 Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article. a
Department of Anesthesiology and Pain Medicine, University of Washington, Seattle, WA, USA, b Department of Anesthesiology and Pain Therapy, University Hospital of Bern, Inselspital, Bern, Switzerland, c Center for Sensory–Motor Interaction, Institute of Social and Preventive Medicine, University of Bern, Bern, Switzerland, d Department of Obstetrics and Gynecology, Cantonal Hospital of Winterthur, Winterthur, Switzerland, e Center for Sensory–Motor Interaction, Department of Health Science and Technology, University of Aalborg, Aalborg, Denmark
Because of the subjective nature of the pain experience, the measurement of pain sensitivity has to rely on self-report; pain thresholds are therefore commonly used to assess pain hypersensitivity.4 In contrast, the measurement of nociceptive withdrawal reflex does not rely on self-report; it is therefore primarily an objective parameter of nociceptive spinal sensitivity.18 A clinically relevant question is how frequently hypersensitivity occurs in chronic pain patients. In 2 case–control studies on chronic low back and neck pain, pain and nociceptive reflex thresholds to electrical stimulation had the best discriminatory value for hypersensitivity among 26 different tests.11,13 These methods are therefore promising to study the epidemiology of hypersensitivity states. Importantly, we are not aware of data on the prevalence of nociceptive spinal hypersensitivity, a phenomenon that has been studied extensively in basic and clinical research and has an established role in the determination of pain states. The primary aim of this study was to determine the prevalence of pain hypersensitivity and nociceptive spinal hypersensitivity in chronic pain patients. The secondary aim was to analyze whether demographic, psychosocial, and pain-related factors are associated with pain hypersensitivity and spinal nociceptive hypersensitivity.
Institutional URL: http://depts.washington.edu/anesth/. *Corresponding author. Address: Department of Anesthesiology and Pain Medicine, University of Washington, 1959 NE Pacific St, Seattle, WA 98195, USA. Tel.: 11 206 543 2568; fax: 11 206 543 2958. E-mail address: [email protected] (M. Curatolo).
2.1. Participants
PAIN 156 (2015) 2373–2382 © 2015 International Association for the Study of Pain http://dx.doi.org/10.1097/j.pain.0000000000000289
November 2015
·
Volume 156
2. Methods
·
Number 11
We studied consecutive patients who were referred to the Pain Clinic of the University Department of Anesthesiology and Pain www.painjournalonline.com
2373
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
2374
·
PAIN®
M. Curatolo et al. 156 (2015) 2373–2382
Therapy of the University of Bern, Switzerland. A multimodal patient assessment that included sociodemographic, psychological, and clinical parameters, as well as QST, had been initiated in 2008. This study was conducted on patients presenting to the pain clinic from July 1, 2011, to June 30, 2013. All patients were informed in written form on the background of the QSTs and on the potential use of their data for scientific reasons. All participants gave consent before the tests. Approval was obtained by the institutional review board, ie, the Ethics Committee of the Canton of Bern, Switzerland (Study Number: 2684). Exclusion criteria were pain duration less than 3 months, neurological disorders of the lower extremity to be tested (palsy, paresthesia, polyneuropathy), any conditions potentially affecting the neurological function of the lower extremity (eg, diabetes mellitus or alcohol abuse), pain or vascular disorders of the lower extremity to be tested (including radicular pain), language problems, pregnancy, tumor pain (defined as pain in the region of infiltration by a primary tumor or metastasis), psychiatric diseases other than unipolar depressive disorder, and rheumatic inflammatory diseases. 2.2. Sociodemographic, psychological, and clinical characteristics Sociodemographic, psychological, and clinical characteristics were recorded by the time of first consultation (Table 1). Part of them were collected for descriptive purposes. A selection of these variables was used for the statistical analyses pertaining to the study aims (see “Statistical analyses” section). Sociodemographic variables were gender, age, body mass index (BMI), and work-related parameters. Psychological characteristics were depression and catastrophizing, assessed with the Beck Depression Inventory Fast Screen15 and the Catastrophizing Scale of the Coping Strategies Questionnaire,7 respectively. Clinical characteristics were history of trauma or surgery related to pain; pain duration; pain intensity immediately before testing; maximum, minimum, and average pain in the preceding 24 hours, rated on a 0 to 10 numerical rating scale (NRS), with 0 indicating no pain and 10 indicating the worst pain imaginable; pain-related life interference as measured by the Multidimensional Pain Inventory8; sleep disturbance as assessed on a 0 to 10 NRS (0 5 no sleep disturbance, 10 5 worst sleep disturbance imaginable); current pain medication, classified as daily intake of any pain medication, opioids, antidepressants, and anticonvulsants (yes/no); and type of pain syndrome. The pain syndrome was categorized as musculoskeletal, neuropathic, orofacial, visceral, noncervicogenic headache, complex regional pain syndrome, phantom limb pain, and atypical pain syndromes (Table 2). Musculoskeletal pain was defined as pain in or around bones, joints, or muscles, characterized by musculoskeletal features such as trigger points, tenderness, and loaddependent or mechanically triggered pain. Musculoskeletal pain was categorized as regional and widespread. Regional musculoskeletal pain included cervical pain, thoracic spine pain, low back pain, shoulder pain, hand pain, hip pain, knee pain, foot pain, and other regional musculoskeletal pain. Widespread musculoskeletal pain was defined as pain felt in the 4 body quadrants during the last week. Neuropathic pain was diagnosed as such when at least one of the following criteria was fulfilled: (1) clinical signs of nerve dysfunction (ie, palsy, sensory loss, paresthesia, allodynia, or hyperalgesia) associated with pain at the territory supplied by the nerve; (2) history of nerve lesion associated with pain in the territory supplied by the nerve; (3) magnetic resonance image finding of
Table 1
Sociodemographic, psychological, and clinical characteristics of patients included in the study and completeness of data. Baseline characteristics Sociodemographic characteristics Gender* Male Female Age, y* BMI, kg/m2* Working status Regular work as usual Reduced work due to pain No work Retired or studying Disability pension No Yes Ongoing litigation No Yes Psychological characteristics Depression (BDI-FS: 0-21, cutoff: 4)* No (score # 4) Yes (score . 4) Catastrophizing (CQS: 0-6)* Clinical characteristics History of trauma related to pain* No Yes History of surgery related to pain* No Yes Pain duration, y* ,1 1-2 .2 Maximum pain in the last 24 h (NRS: 0-10) Minimum pain in the last 24 h (NRS: 0-10) Average pain in the last 24 h (NRS: 0-10)* Pain intensity before testing (NRS: 0-10) Pain-related life interference (MPI: 0-6) Pain-related sleep interference (NRS: 0-10) Daily intake of pain medication* No Yes Daily intake of opioids No Yes Daily intake of antidepressants and/or anticonvulsants No Yes
Mean (SD) or no. of patients (%)
n (total 5 961) 961
422 (43.9) 539 (56.1) 50.0 (15.1) 26.0 (5.2)
961 921 934
284 (30.4) 170 (18.2) 300 (32.1) 180 (19.3) 927 835 (90.1) 92 (9.9) 925 694 (75.0) 231 (25.0) 898 409 (45.5) 489 (54.4) 3.2 (1.4)
913 961
714 (74.3) 247 (25.7) 954 543 (56.9) 411 (43.1) 2.4 (1.0-6.4)* 226 (24.6) 179 (19.5) 514 (55.9) 7.3 (2.0) 4.1 (2.5) 6.2 (2.1) 5.6 (2.5) 4.0 (1.3) 5.3 (3.0)
919
878 872 878 957 918 935 959
236 (24.6) 723 (75.4) 958 680 (71.0) 278 (29.0) 958 677 (70.7) 278 (29.3)
The type of pain syndrome is presented in a separate table. * Variables included in the regression analysis for factors potentially associated with central hypersensitivity. The other variables were collected for descriptive purposes. BDI-FS, Beck Depression Inventory Fast Screen, whereby 0 5 no depression and 21 5 maximum depression; BMI, body mass index; CQS, Catastrophizing Scale of the Coping Strategies Questionnaire, whereby 0 5 no catastrophizing and 6 5 maximum catastrophizing; MPI, Multidimensional Pain Inventory; NRS, numerical rating scale, whereby 0 5 no pain and 10 5 maximum pain.
nerve compression together with pain at its innervation. Neuropathic pain was categorized as postherpetic, posttraumatic nerve lesion, post-thoracotomy, radicular, trigeminal, and other neuropathic pain syndromes. Orofacial pain was defined as pain localized at the facial region or in the mouth, without clear neuropathic features as defined above (trigeminal neuralgia was
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
November 2015
·
Volume 156
·
Number 11
www.painjournalonline.com
Table 2
Classification of type of pain. Type of pain
No. of patients (%)
Musculoskeletal pain Regional musculoskeletal pain Neck pain Thoracic spine pain Low back pain Shoulder pain Hand pain Hip pain Knee pain Foot pain Other regional musculoskeletal pain Widespread pain Neuropathic pain Postherpetic Posttraumatic nerve lesion Post-thoracotomy Radicular Trigeminal neuralgia Other neuropathic pain states Orofacial pain Visceral pain Chronic pelvic pain Inguinal pain Other visceral pain syndromes Other pain syndromes Noncervicogenic headache CRPS Phantom limb pain Atypical pain
637 (66.3) 528 106 27 219 25 12 41 46 21 31 109 102 (10.6) 14 23 22 25 5 13 61 (6.3) 71 (7.4) 16 26 29 90 (9.4) 21 35 3 31
A pain syndrome was defined as atypical when it could not be classified as any of the pain syndromes as defined in the “Methods” section. CRPS, complex regional pain syndrome.
categorized as neuropathic pain). Visceral pain was diagnosed as such when it was located in the deep abdominal or pelvic region, with concomitant history of abdominal/pelvic disease or surgery. Noncervicogenic headache included migraine, cluster headache, and tension-type headache. Complex regional pain syndrome was diagnosed according to the published definition criteria.5 A pain syndrome was defined as atypical when it could not be classified as one of the above-mentioned ones. In case of multiple pain syndromes, all of them were registered. 2.3. Assessments of pain hypersensitivity and spinal nociceptive hypersensitivity Pain threshold to electrical stimulation and the threshold of the nociceptive withdrawal reflex were chosen because in 2 previous
2375
case–control studies, these assessments displayed the best discriminatory value for hypersensitivity among 26 different tests.11,13 At the beginning of the testing session, all the tests were performed on the patients for training purposes. Only when the patients felt familiar with the tests, we formally started the testing procedure. Electrical stimulation was performed through bipolar skin surface Ag/AgCl electrodes placed just distal to the lateral malleolus (innervation area of the sural nerve). The body side was contralateral to the side of most pain. In case of bilateral pain of equal intensity, the side was selected according to a computer-generated random sequence. Electromyography reflex responses to electrical stimulation were recorded from the middle of the biceps femoris and rectus femoris muscles (Ag/AgCl electrodes).10 Stimulation and electromyographic recordings were made by a computercontrolled constant current stimulator (NCS System, Evidence 3102 evo; Neurosoft, Ivanovo, Russia). A 25-millisecond train of five 1-millisecond square-wave impulse (perceived as a single stimulus) was delivered. The current intensity was increased from 1 mA in steps of 1 mA until (1) a biceps femoris reflex with an amplitude higher than 20 mV for at least 10 milliseconds in the 50- to 150millisecond poststimulation interval was detected (nociceptive reflex threshold); and (2) a pain sensation was evoked (pain threshold). These thresholds were measures of spinal nociceptive hypersensitivity and of pain hypersensitivity, respectively. If increasing the current intensity led to intolerable pain in the absence of reflex, only the pain threshold was recorded. 2.4. Statistical analyses The prevalence of pain hypersensitivity and spinal nociceptive hypersensitivity was calculated as the percentage of patients who displayed values equal to or below the 5th, 10th, and 25th percentiles of reference values, which were determined in previous studies on 300 pain-free subjects10,14 (Table 3). The prevalence of hypersensitivity and its corresponding 95% confidence intervals (CIs) were calculated using a binominal distribution. We modeled factors potentially associated with hypersensitivity using logistic regression models based on maximum likelihood estimation. For this purpose, we considered the 10th percentile as cutoff for the presence of hypersensitivity as primary outcome. The factors analyzed were gender, age, BMI, depression, catastrophizing, history of trauma or surgery related to the pain condition, type of pain, pain duration, average pain intensity of the last 24 hours, and daily intake of pain medication. We based the choice of these variables on previous research on factors potentially associated with hypersensitivity and clinical reasoning. Associations between hypersensitivity and
Table 3
Normative values of pain detection and reflex threshold after electrical stimulation, based on data generated in 300 pain-free individuals. Test Electrical pain detection threshold
Electrical reflex threshold
Gender
Age, y
5th percentile
10th percentile
25th percentile
Female Female Male Male Female Female Male Male
20-49 50-80 20-49 50-80 20-49 50-80 20-49 50-80
6.0 8.0 7.0 9.0 9.0 11.0 10.7 12.3
7.0 9.0 7.7 9.0 10.0 13.7 11.7 13.3
7.7 10.0 8.3 10.0 12.0 15.7 13.3 14.3
All values are in milliamperes.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
2376
·
PAIN®
M. Curatolo et al. 156 (2015) 2373–2382
these baseline patient characteristics were expressed as odds ratios (ORs), including 95% CIs. We first modeled univariate associations, then partially adjusted associations including nonpain-related factors such as gender, age, BMI, depression, and catastrophizing scale, and finally computed a fully adjusted model including all baseline characteristics potentially associated with the presence of pain hypersensitivity and spinal nociceptive hypersensitivity. During testing, we noticed that in roughly one-third of the patients, the reflex threshold could not be determined, as pain became intolerable before a reflex was obtained. We estimated the prevalence of hypersensitivity and potentially associated factors based on 3 different assumptions to account for these missing values and make the results of pain and reflex thresholds comparable. In a first step, we conducted a complete case analysis, ie, we restricted the analysis to the patients in whom a reflex threshold could be obtained (n 5 696). In a second step, a multiple imputation analysis was performed, ie, missing values for reflex threshold and other incompletely recorded explanatory variables were imputed; prevalence of hypersensitivity and associated factors were estimated based on the imputed data set.6 We used a multivariate sequential imputation technique based on chained equations generating 15 multiple imputed data sets. Where possible, we imputed continuous values. For analysis, we dichotomized variables as reported in Table 1. We finally performed a third analysis in which patients who did not display a reflex were considered as being not hypersensitive.
3. Results A total of 1508 patients were considered for the testing procedure. Of these, 547 patients were not included (reasons in parentheses): 50 (pain duration less than 3 months), 220 (neurological disorders), 16 (vascular disorders), 45 (tumor pain), 36 (psychiatric diseases other than unipolar depressive disorder), 25 (rheumatic inflammatory disease), and 155 (other reasons, including language problems, logistic reasons, previous bilateral leg amputation, and pregnancy). Nine hundred sixty-one patients were therefore eligible for the analyses. It was possible to determine pain detection threshold after electrical stimulation in all 961 patients. A nociceptive reflex could be evoked in 696 patients (72.4%); in the remaining patients, pain became intolerable at stimulus intensities that did not evoke a reflex. The mean pain and reflex thresholds in these 696 patients with complete records were 6.4 mA (SD 5 3.3) and 8.1 mA (SD 5 4.8), respectively. Tables 1 and 2 report baseline patient characteristics.
3.1. Prevalence of pain hypersensitivity and spinal nociceptive hypersensitivity The prevalence of pain hypersensitivity and spinal nociceptive hypersensitivity at the different cutoffs for reference values is reported in Table 4 and illustrated in Figures 1 and 2. Probability density curves of the normative data and patient data for pain hypersensitivity and spinal nociceptive hypersensitivity are illustrated in Figures 3 and 4, respectively. In the multiple imputation analysis, the prevalence of pain hypersensitivity and spinal nociceptive hypersensitivity (95% CI) with 10th percentile cutoff was 71.2 (68.3-74.0) and 80.0 (77.082.6), respectively. These findings were very similar with the complete case analysis. When considering patients without reflex as being not hypersensitive, the prevalence of spinal nociceptive hypersensitivity at the 10% cutoff dropped to 57.9 (54.8-61.1). 3.2. Factors associated with pain hypersensitivity and spinal nociceptive hypersensitivity The results of the regression analyses for factors potentially associated with pain hypersensitivity and spinal nociceptive hypersensitivity are displayed in Tables 5–7. Table 5 presents the results of the analysis on all cases (n 5 961) and is based on the imputed data set. In the unadjusted analysis, depression, catastrophizing, and pain-related sleep interference were significantly associated with pain hypersensitivity, but not with spinal nociceptive hypersensitivity. Depression, catastrophizing, and pain-related sleep interference were not statistically significant anymore after adjustment for other variables. The average pain intensity was significantly associated with both pain hypersensitivity and spinal nociceptive hypersensitivity in the unadjusted analyses; the significance was not observed anymore in the adjusted analysis for spinal nociceptive hypersensitivity, but was still present for pain hypersensitivity. The association was positive: the higher the pain intensity, the higher the risk of being hypersensitive. All other variables did not display statistically significant correlations. These results were mostly confirmed by the analysis limited on the complete data set (n 5 696, Table 6). The only differences were that pain-related sleep interference was still significantly related to pain hypersensitivity after adjustment and the average pain intensity was not significant anymore in the adjusted analyses on pain hypersensitivity. When patients with missing reflexes were considered as being not hypersensitive, the average pain intensity was no longer significant for nociceptive spinal hypersensitivity (Table 7). Having
Table 4
Prevalence of hypersensitivity as assessed by electrical stimulation and using cutoffs at different normative percentiles. Prevalence in percent (95% CI) Multiple imputation analysis (n 5 961) Pain hypersensitivity Spinal nociceptive hypersensitivity Complete case analysis (n 5 696) Pain hypersensitivity Spinal nociceptive hypersensitivity Patients with missing reflex considered as being not hypersensitive (n 5 961) Pain hypersensitivity Spinal nociceptive hypersensitivity
5th percentile cutoff
10th percentile cutoff
25th percentile cutoff
62.0 (58.8-65.0) 72.3 (69.0-75.5)
71.2 (68.3-74.0) 80.0 (77.0-82.6)
80.7 (78.3-83.2) 87.5 (85.0-90.0)
62.5 (59.0-66.1) 72.8 (69.5-76.2)
72.2 (68.9-75.6) 80.0 (77.1-83.0)
82.0 (79.1-84.9) 87.6 (85.2-90.1)
NA 52.8 (49.6-55.9)
NA 57.9 (54.8-61.1)
NA 63.5 (60.4-66.5)
Pain hypersensitivity was assessed by pain threshold. Spinal nociceptive hypersensitivity was assessed by the nociceptive withdrawal reflex threshold. The table presents the results of the multiple imputation analysis, complete case analysis, and a sensitivity analysis that considered patients without reflex as being not hypersensitive. CI, confidence interval; NA, not applicable.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
November 2015
·
Volume 156
·
Number 11
Figure 1. Prevalence of pain hypersensitivity as assessed by electrical stimulation, using cutoffs at 5th, 10th, and 25th percentiles. The graph reports the results of the multiple imputation analysis (Table 4, n 5 961).
pain for more than 2 years reduced the risk of hypersensitivity, but only in the unadjusted analysis. Daily intake of pain medication reduced the risk of hypersensitivity in both unadjusted and adjusted analyses.
4. Discussion To our knowledge, this is the largest study on the prevalence of pain hypersensitivity in chronic pain patients and the first one on the prevalence of spinal nociceptive hypersensitivity. It revealed an impressively high prevalence for both parameters, supporting the clinical relevance of phenomena that have been extensively investigated in basic and clinical research. Both pain and spinal nociceptive hypersensitivity were not associated with most sociodemographic, psychological, and clinical characteristics.
Figure 2. Prevalence of spinal nociceptive hypersensitivity as assessed by the nociceptive withdrawal reflex, using cutoffs at 5th, 10th, and 25th percentiles. The graph reports the results of the multiple imputation analysis (black bars, n 5 961) and of the analysis that considered patients without reflex as being not hypersensitive (gray bars, n 5 961) (Table 4).
www.painjournalonline.com
2377
Figure 3. Probability density curve of the experimental data collected in chronic pain patients (blue curve) and of the normative data collected in painfree subjects (red curve) for pain hypersensitivity. The vertical lines represent the 5th, 10th, and 25th percentiles of the normative data considered as cutoff values to define the prevalence of pain hypersensitivity.
4.1. Prevalence of pain hypersensitivity In the multiple imputation analysis, the prevalence of pain hypersensitivity (95% CI) using the 5th, 10th, and 25th percentile cutoffs was 62.0% (58.8-65.0), 71.2% (68.3-70.0), and 80.7% (78.3-83.2), respectively (Table 4 and Fig. 1). Very similar results were obtained by the complete case analysis. In a previous study using pressure pain tolerance threshold at the second toe, the prevalence of pain hypersensitivity was substantially lower: at the 5th, 10th, and 25th percentile cutoffs, the prevalence was 17.5%, 24.5%, and 35.5%, respectively.20 It is difficult to draw firm conclusions by comparing the previous with this study, as they have been conducted on 2 different patient populations. However, both studies have been performed in the same setting and using the same data set for calculation of reference values.10,19 It is therefore unlikely that the much higher prevalence observed in this study is entirely due to a much higher pain sensitivity of the patient population analyzed. One possible explanation is the different methodology used to assess hypersensitivity, with electrical pain
Figure 4. Probability density curve of the experimental data collected in chronic pain patients (blue curve) and of the normative data collected in painfree subjects (red curve) for spinal nociceptive hypersensitivity. The vertical lines represent the 5th, 10th, and 25th percentiles of the normative data considered as cutoff values to define the prevalence of spinal nociceptive hypersensitivity.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
2378
Pain hypersensitivity (electrical pain detection threshold)
Gender Male Female Age (per 10 y increase) BMI (cutoff: 25 kg/m2) Normal weight Overweight Depression (BDI-FS, cutoff: 4) No (score # 4) Yes (score . 4) Catastrophizing (CQS: 0-6) History of trauma related to pain No Yes History of surgery related to pain No Yes Type of pain Musculoskeletal Neuropathic Orofacial Visceral Other pain syndromes Pain duration, y ,1 1-2 .2 Average pain (NRS: 0-10) Pain-related sleep inference (NRS: 0-10) Daily intake of any pain medication No Yes
P
Spinal nociceptive hypersensitivity (nociceptive reflex threshold)
Partially adjusted OR (95% CI)
P
Fully adjusted OR (95% CI)
P
Unadjusted OR (95% CI)
P
Partially adjusted OR (95% CI)
P
Fully adjusted OR (95% CI)
P
·
Unadjusted OR (95% CI) 1 1.24 (0.93-1.64) 1.04 (0.95-1.14)
0.14 0.43
1 1.25 (0.94-1.67) 1.04 (0.95-1.15)
1 0.12 1.17 (0.87-1.57) 0.39 1.04 (0.94-1.15)
1 0.30 1.06 (0.75-1.50) 0.41 1.10 (0.97-1.22)
1 0.75 1.06 (0.74-1.50) 0.13 1.09 (0.97-1.22)
1 0.75 0.95 (0.66-1.37) 0.14 1.08 (0.96-1.22)
0.78 0.19
1 0.96 (0.73-1.27)
0.79
1 0.94 (0.70-1.25)
1 0.66 0.90 (0.67-1.21)
1 0.50 1.08 (0.76-1.52)
1 0.67 1.05 (0.74-1.49)
1 0.79 1.05 (0.73-1.52)
0.78
1 1.54 (1.16-2.05) 1.19 (1.07-1.32)
1 0.003 1.30 (0.91-1.84) 0.001 1.14 (1.00-1.29)
1 0.15 1.32 (0.93-1.88) 0.05 1.04 (0.90-1.19)
1 0.12 1.06 (0.73-1.52) 0.62 1.04 (0.92-1.18)
1 0.77 0.99 (0.74-1.49) 0.53 1.04 (0.90-1.21)
1 0.96 1.02 (0.66-1.58) 0.57 0.98 (0.83-1.16)
0.92 0.85
1 0.97 (0.71-1.34)
0.87
NA
NA
1 0.98 (0.69-1.38)
1 0.89 0.76 (0.51-1.12)
0.17 NA
NA
1 0.76 (0.50-1.14)
0.19
1 1.02 (0.77-1.36)
0.88
NA
NA
1 1.03 (0.76-1.40)
1 0.86 0.76 (0.55-1.06)
0.11 NA
NA
1 0.79 (0.55-1.14)
0.22
1 0.93 (0.59-1.47) 1.19 (0.65-2.18) 0.71 (0.43-1.20) 0.82 (0.51-1.31)
0.76 0.57 0.20 0.40
NA
1 1.06 (0.66-1.72) 1.46 (0.77-2.77) 0.83 (0.47-1.46) 0.88 (0.53-1.46)
0.81 0.25 0.51 0.62
1 1.1 (0.58-1.78) 1.12 (0.51-2.46) 0.78 (0.41-1.50) 1.00 (0.56-1.77)
0.97 0.77 0.46 0.99 NA
NA
1 0.99 (0.55-1.77) 1.17 (0.51-2.66) 0.83 (0.41-1.68) 1.13 (0.60-2.10)
0.96 0.71 0.60 0.70
NA NA NA
1 1.07 (0.68-1.67) 1.07 (0.75-1.53) 1.09 (1.00-1.18) 1.05 (0.99-1.11)
0.78 0.72 0.05 0.11
1 0.79 (0.45-1.38) 0.66 (0.41-1.07) 1.09 (1.01-1.18) 1.04 (0.97-1.10)
0.41 0.09 NA 0.04 NA 0.25
NA NA
1 0.84 (0.48-1.50) 0.64 (0.39-1.06) 1.08 (0.97-1.19) 1.08 (0.97-1.19)
0.57 0.08 0.16 0.68
NA
1 0.94 (0.67-1.33)
1 0.47 1.12 (0.75-1.66)
0.57 NA
NA
1 1.01 (0.66-1.56)
0.95
1 0.96 (0.62-1.48) 1.06 (0.75-1.49) 1.15 (1.08-1.23) 1.09 (1.04-1.15)
1 1.20 (0.88-1.65)
NA
0.85 0.76 NA ,0.001 NA ,0.001 NA
0.25
NA
M. Curatolo et al. 156 (2015) 2373–2382
Table 5
Factors associated with pain hypersensitivity and spinal nociceptive hypersensitivity.
The 10th percentile of normative values was used as a cutoff to categorize patients as hypersensitive or not. The table presents unadjusted and adjusted ORs with 95% CIs and P-values with multiple imputation analysis for missing measurements of nociceptive reflex threshold. Partial adjustment was for nonpain-related variables: gender, age, BMI, depression, and catastrophizing; full adjustment was for all variables. The effect of continuous variables is displayed per standardized increase of 1 SD unit. BDI-FS, Beck Depression Inventory Fast Screen, ranging from 0 5 no depression to 21 5 maximum depression; BMI, body mass index; CI, confidence interval; CQS, Catastrophizing Scale of the Coping Strategies Questionnaire, whereby 0 5 no catastrophizing and 6 5 maximum catastrophizing; NA, not applicable; NRS, numerical rating scale, whereby 0 5 no pain and 10 5 worst pain imaginable; OR, odds ratio.
PAIN®
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
November 2015
Table 6
·
Factors associated with pain hypersensitivity and spinal nociceptive hypersensitivity. Unadjusted OR (95% CI)
Spinal nociceptive hypersensitivity (nociceptive reflex threshold)
Partially adjusted OR (95% CI)
P
Fully adjusted OR (95% CI)
P
Unadjusted OR (95% CI)
P
Partially adjusted OR (95% CI)
P
Fully adjusted OR (95% CI)
P
·
0.60 0.32
1 1.21 (0.85-1.72) 1.06 (0.94-1.20)
1 0.30 1.11 (0.76-1.62) 0.33 1.06 (0.94-1.21)
1 0.59 1.01 (0.69-1.47) 0.35 1.09 (0.96-1.24)
1 0.96 1.01 (0.68-1.50) 0.17 1.11 (0.97-1.26)
1 0.96 0.85 (0.56-1.30) 0.14 1.11 (0.96-1.29)
0.45 0.16
1 1.09 (0.78-1.53)
0.61
1 0.99 (069-1.42)
1 0.97 0.97 (0.67-1.42)
1 0.89 1.15 (0.79-1.68)
1 0.46 1.09 (0.73-1.61)
1 0.69 1.09 (0.72-1.66)
0.69
1 1.60 (1.36-2.27) 1.21 (1.06-1.37)
1 0.007 1.34 (0.89-2.03) 0.003 1.15 (0.98-1.33)
1 0.16 1.35 (0.87-2.08) 0.08 1.06 (0.87-1.26)
1 0.18 0.99 (0.67-1.45) 0.54 1.04 (0.90-1.19)
1 0.96 0.90 (0.57-1.44) 0.59 1.05 (0.89-1.24)
1 0.67 0.98 (0.60-1.60) 0.55 0.98 (0.81-1.19)
0.93 0.87
1 1.09 (0.74-1.60)
0.67
NA
NA
1 0.98 (0.62-1.54)
1 0.93 0.74 (0.49-1.11)
0.15 NA
NA
1 0.73 (0.45-1.18)
0.20
1 1.01 (0.72-1.42)
0.94
NA
NA
1 1.07 (0.72-1.59)
1 0.75 0.70 (0.48-1.02)
0.06 NA
NA
1 0.69 (0.45-1.07)
0.10
1 0.79 (0.47-1.34) 1.28 (0.61-2.66) 0.73 (0.40-1.33) 0.89 (0.51-1.53)
0.38 0.51 0.31 0.67
NA
1 0.87 (0.49-1.56) 1.60 (0.71-3.64) 0.72 (0.35-1.47) 1.05 (0.55-2.00)
0.65 0.26 0.37 0.88
1 0.83 (0.45-1.50) 1.11 (0.50-2.47) 0.76 (0.39-1.48) 1.00 (0.53-1.87)
0.55 0.80 0.42 0.99 NA
NA
1 0.89 (0.46-1.72) 0.37 (0.55-3.38) 0.90 (0.41-1.99) 1.19 (0.57-2.46)
0.73 0.50 0.80 0.64
NA NA NA
1 1.14 (0.65-1.98) 1.05 (0.68-1.65) 1.07 (0.97-1.19) 1.08 (1.00-1.16)
0.65 0.81 0.65 0.04
1 0.80 (0.44-1.46) 0.63 (0.38-1.02) 1.08 (0.99-1.18) 1.4 (0.98-1.11)
0.46 0.06 NA 0.07 NA 0.23 NA
NA NA NA
1 0.92 (0.48-1.77) 0.61 (0.37-1.03) 1.10 (0.98-1.24) 1.02 (0.94-1.11)
0.80 0.07 0.11 0.66
NA
1 0.75 (0.49-1.14)
1 0.18 1.02 (0.67-1.54)
0.93 NA
NA
1 0.94 (0.59-1.50)
0.80
1 1.04 (0.63-1.72) 1.08 (0.72-1.62) 1.15 (1.06-1.24) 1.12 (1.06-1.18)
1 0.99 (0.68-1.43)
NA
0.88 0.71 NA ,0.001 NA ,0.001 NA
0.95
NA
The 10th percentile of normative values was used as a cutoff to categorize patients as hypersensitive or not. The table presents unadjusted and adjusted ORs with 95% CIs and P-values from complete case analysis (n 5 612-696). Partial adjustment was for nonpain-related variables: gender, age, BMI, depression, and catastrophizing; full adjustment was for all variables. The effect of continuous variables is displayed per standardized increase of 1 SD unit. BDI-FS, Beck Depression Inventory Fast Screen, ranging from 0 5 no depression to 21 5 maximum depression; BMI, body mass index; CI, confidence interval; CQS, Catastrophizing Scale of the Coping Strategies Questionnaire, whereby 0 5 no catastrophizing and 6 5 maximum catastrophizing; NA, not applicable; NRS, numerical rating scale, whereby 0 5 no pain and 10 5 worst pain imaginable; OR, odds ratio.
www.painjournalonline.com
1 1.09 (0.78-1.53) 1.06 (0.95-1.18)
Number 11
Gender Male Female Age (per 10 y increase) BMI (cutoff: 25 kg/m2) Normal weight Overweight Depression (BDI-FS, cutoff: 4) No (score # 4) Yes (score . 4) Catastrophizing (CQS: 0-6) History of trauma related to pain No Yes History of surgery related to pain No Yes Type of pain Musculoskeletal Neuropathic Orofacial Visceral Other pain syndromes Pain duration, y ,1 1-2 .2 Average pain (NRS: 0-10) Pain-related sleep inference (NRS: 0-10) Daily intake of any medication No Yes
P
Volume 156
Pain hypersensitivity (electrical pain detection threshold)
2379
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
2380
·
PAIN®
M. Curatolo et al. 156 (2015) 2373–2382
Table 7
Factors associated with spinal nociceptive hypersensitivity, considering patients with missing reflex as being not hypersensitive (n 5 845-961). Spinal nociceptive hypersensitivity (nociceptive reflex threshold) Gender Male Female Age (per 10 y increase) BMI (cutoff: 25 kg/m2) Normal weight Overweight Depression (BDI-FS, cutoff: 4) No (score # 4) Yes (score . 4) Catastrophizing (CQS: 0-6) History of trauma related to pain No Yes History of surgery related to pain No Yes Type of pain Musculoskeletal Neuropathic Orofacial Visceral Other pain syndromes Pain duration, y ,1 1-2 .2 Average pain (NRS: 0-10) Pain-related sleep inference (NRS: 0-10) Daily intake of any medication No Yes
Unadjusted OR (95% CI)
P
Partially adjusted OR (95% CI)
P
Fully adjusted OR (95% CI)
P
1 1.10 (0.85-1.43) 1.00 (0.92-1.09)
0.46 0.97
1 1.08 (0.82-1.42) 1.00 (0.99-1.01)
0.58 0.91
1 1.08 (0.81-1.44) 1.00 (0.99-1.01)
0.61 0.99
1 0.86 (0.66-1.12)
0.27
1 0.88 (9.67-1.16)
0.37
1 0.92 (0.69-1.23)
0.60
1 0.99 (0.76.1.29) 1.05 (0.95-1.15)
0.91 0.36
1 0.92 (0.67-1.28) 1.06 (0.95-1.20)
0.64 0.30
1 0.98 (0.70-1.37) 1.08 (0.94-1.23)
0.88 0.26
1 0.85 (0.63-1.14)
0.27
NA
NA
1 0.77 (0.55-1.08)
0.14
1 0.92 (0.71-1.19)
0.52
NA
NA
1 1.01 (0.75-1.36)
0.95
1 1.04 (0.68-1.59) 1.17 (0.68-2.00) 1.04 (0.63-1.71) 1.38 (0.87-2.18)
0.85 0.56 0.88 0.17
NA
1 1.03 (0.65-1.64) 1.01 (0.57-1.81) 1.01 (0.60-2.00) 1.59 (0.93-2.71)
0.90 0.97 0.76 0.09
1 0.92 (0.62-1.38) 0.71 (0.51-0.98) 1.00 (0.94-1.06) 0.99 (0.94-1.03)
0.69 0.04 0.89 0.52
NA NA NA
NA NA NA
1 0.89 (0.58-1.37) 0.74 (0.52-1.04) 1.03 (0.94-1.11) 0.98 (0.92-1.03)
0.59 0.09 0.55 0.43
1 0.67 (0.49-0.91)
0.009
NA
NA
1 0.65 (0.46-0.92)
0.01
NA
The 10th percentile of normative values was used as a cutoff to categorize patients as hypersensitive or not. The table presents unadjusted and adjusted ORs with 95% CIs and P-values. Partial adjustment was for nonpain-related variables: gender, age, BMI, depression, and catastrophizing; full adjustment was for all variables. The effect of continuous variables is displayed per standardized increase of 1 SD unit. BDI-FS, Beck Depression Inventory Fast Screen, ranging from 0 5 no depression to 21 5 maximum depression; BMI, body mass index; CI, confidence interval; CQS, Catastrophizing Scale of the Coping Strategies Questionnaire, whereby 0 5 no catastrophizing and 6 5 maximum catastrophizing; NA, not applicable; NRS, numerical rating scale, whereby 0 5 no pain and 10 5 worst pain imaginable; OR, odds ratio.
threshold possibly being more sensitive than pressure pain tolerance threshold to detect pain hypersensitivity. In this study, the stimulation was performed at a body site that was far distant from the site of pain. Patients with any pain or pathology at the tested site were excluded. We therefore measured widespread pain hypersensitivity, likely reflecting a generalized hyperexcitability of central pain pathways. 4.2. Prevalence of spinal nociceptive hypersensitivity In the multiple imputation analysis, the prevalence of spinal nociceptive hypersensitivity (95% CI) using the 5th, 10th, and 25th percentile cutoffs was 72.3% (69.0-75.5), 80.0% (77.0-82.6), and 87.5% (85.0-90.0), respectively (Table 4 and Fig. 2). Very similar results were obtained by the complete case analysis. A nociceptive reflex could not be evoked in 37.6% of all patients, as pain became intolerable at stimulus intensities that did not evoke a reflex. It is unclear why the reflex cannot be evoked in a subset of patients, but this phenomenon has been repeatedly observed in previous studies.1,2,21 In a further analysis that considered these patients as being not hypersensitive, the prevalence obviously dropped, but was still significantly high, ranging 52.8% to 63.5% depending on the cutoff (Table 4 and Fig. 2).
As for the determination of pain hypersensitivity, the assessment of the nociceptive reflex was performed at a body site where no injury potentially causing pain was present. We therefore likely assessed a state of generalized spinal nociceptive hyperexcitability. 4.3. Factors associated with pain hypersensitivity and spinal nociceptive hypersensitivity Both pain hypersensitivity and spinal nociceptive hypersensitivity were not associated with most sociodemographic, psychological, and clinical characteristics (Tables 5–7). The few variables that were statistically significant displayed very low ORs, raising doubts on the clinical relevance of the impact. Interestingly, this finding is consistent with the results of a previous study on 300 pain-free subjects using the same paradigms of electrical stimulation.10 This suggests that electrical pain threshold and nociceptive withdrawal reflex explore aspect of sensitization processes that are independent of most sociodemographic, psychological, and clinical characteristics. The variable that was most consistently associated with pain hypersensitivity and spinal nociceptive hypersensitivity was baseline spontaneous pain intensity: the higher the patients
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
November 2015
·
Volume 156
·
Number 11
scored on the NRS, the higher the risk of hypersensitivity with both pain and nociceptive reflex threshold. However, it has to be noticed that the quantitative impact of this variable was modest, as the ORs were only slightly above 1. Pain intensity is the result of multiple and mostly uncontrollable factors, including nociceptive input from damaged tissues, neuroplastic changes, alterations in central pain modulation, and cognitive and affective factors. In a human study, it is not possible to find out which of these factors are responsible for the influence of pain intensity on parameters of pain hypersensitivity and spinal nociceptive hypersensitivity. Independent of the pathophysiologic explanation, it is likely that pain intensity and hypersensitivity processes feed each other in a positive feedback loop: high pain intensity causes more hypersensitivity, which in turn contributes to higher pain levels. Accordingly, acting at this loop by reducing hypersensitivity seems to be a relevant therapeutic aim. The magnitude of pain-related interference with sleep was associated with increased risk of pain hypersensitivity. However, also for this variable, the OR was only slightly above 1, indicating a modest quantitative impact. Nevertheless, the finding suggests that improving sleep quality may be important for the management of pain hypersensitivity, potentially leading to reduction in pain. In a previous study using pressure pain tolerance thresholds, different factors were found to be positively related with pain hypersensitivity.20 They included female gender, duration of pain, and having an atypical pain syndrome. Catastrophizing and baseline pain intensity were marginally not significant, and depression was significant only with the univariate analysis. Remarkably, a study on 300 pain-free subjects14 found female gender to be significantly associated with hypersensitivity to pressure stimulation and catastrophizing was marginally not significant, basically mirroring the results of the aforementioned study conducted on patients.20 Previous research on both painfree subjects and chronic pain patients found catastrophizing to be related with pain thresholds, but not with the nociceptive withdrawal reflex.16,17,21 This suggests that catastrophizing affects pain through supraspinal mechanisms, rather than through descending modulatory processes. Overall, it seems that pain hypersensitivity as assessed by electrical stimulation is less associated with sociodemographic, psychological, and clinical characteristics, compared with pain hypersensitivity as assessed by pressure stimulation. We know that different stimulation modalities explore different dimensions of pain perception.12 However, the limited knowledge of the specific mechanisms explored by different test modalities does not allow clear explanations for these differences. The limited association of demographic, psychological, and clinical characteristics with electrical tests suggests that this modality may explore only a limited part of the complex pain experience. This however is not necessarily a limitation. Both clinicians and researchers may find it useful to include assessments that are minimally affected by confounding and frequently uncontrollable parameters, leading to large unexplained variability. 4.4. Strengths and limitations The main strengths are the large sample size and the combination of measures of both subjective pain sensitivity and spinal nociceptive processes. Considering the overwhelming data from basic and clinical research on hypersensitivity, the study fills an important translational gap. The results pertain to the methodology used. Because we performed a large investigation for a long time period, we could not apply an extensive experimental protocol involving multiple tests. As stated above, different
www.painjournalonline.com
2381
methods to assess hypersensitivity explore different neural mechanisms and pain dimensions, potentially leading to different results in terms of prevalence and its determining variables. Further research will hopefully clarify which modality is most closely related to clinically significant outcomes, such as pain, disability, or course of the disease. Also, the results pertain to a patient population of a tertiary pain clinic. As for any epidemiologic study, their external validity needs to be determined by further studies conducted in different settings and on different patient populations. 4.5. Summary conclusions This study found an impressively high prevalence of pain hypersensitivity and spinal nociceptive hypersensitivity in a population of chronic pain patients attending a tertiary pain center. Depending on the cutoff of reference values and the type of analysis, the prevalence of pain hypersensitivity and spinal nociceptive hypersensitivity ranged from 62.5% to 82.0% and from 52.8% to 87.6%, respectively. Both hypersensitivity parameters were not associated with most baseline variables, suggesting that electrical pain thresholds and nociceptive withdrawal reflex explore aspects of pain processing that are mostly independent of sociodemographic, psychological, and clinical characteristics. Because of the likely influence of hypersensitivity on clinically relevant outcomes, clinical quantification of hypersensitivity may improve our understanding of the pathophysiology of pain in individual patients. Furthermore, the development of effective therapeutic strategies that attenuate the hypersensitivity state is an important target of pain research.
Conflict of interest statement The authors have no conflicts of interest to declare. The study was funded by the Swiss National Science Foundation (32003B_138361), the Danish Research Council for Technology and Production, and the Scientific Funds of the University Department of Anesthesiology and Pain Therapy of the University of Bern.
Acknowledgements M. Muller ¨ and A. Ashraf contributed equally. Article history: Received 6 May 2015 Received in revised form 25 June 2015 Accepted 2 July 2015 Available online 7 July 2015
References [1] Banic B, Petersen-Felix S, Andersen OK, Radanov BP, Villiger PM, Arendt-Nielsen L, Curatolo M. Evidence for spinal cord hypersensitivity in chronic pain after whiplash injury and in fibromyalgia. PAIN 2004;107: 7–15. [2] Biurrun Manresa JA, Fritsche R, Vuilleumier PH, Oehler C, Morch CD, Arendt-Nielsen L, Andersen OK, Curatolo M. Is the conditioned pain modulation paradigm reliable? A test-retest assessment using the nociceptive withdrawal reflex. PLoS One 2014;9:e100241. [3] Curatolo M. Diagnosis of altered central pain processing. Spine 2011;36: S200–204. [4] Curatolo M, Arendt-Nielsen L. Central hypersensitivity in chronic musculoskeletal pain. Phys Med Rehabil Clin N Am 2015;26:175–84. [5] Harden RN, Bruehl S, Stanton-Hicks M, Wilson PR. Proposed new diagnostic criteria for complex regional pain syndrome. Pain Med 2007;8: 326–31.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
2382
·
M. Curatolo et al. 156 (2015) 2373–2382
[6] Harel O, Zhou XH. Multiple imputation: review of theory, implementation and software. Stat Med 2007;26:3057–77. [7] Keefe FJ, Brown GK, Wallston KA, Caldwell DS. Coping with rheumatoid arthritis pain: catastrophizing as a maladaptive strategy. PAIN 1989;37:51–6. [8] Kerns RD, Turk DC, Rudy TE. The West Haven-Yale Multidimensional Pain Inventory (WHYMPI). PAIN 1985;23:345–56. [9] Lim EC, Sterling M, Stone A, Vicenzino B. Central hyperexcitability as measured with nociceptive flexor reflex threshold in chronic musculoskeletal pain: a systematic review. PAIN 2011;152:1811–20. [10] Neziri AY, Andersen OK, Petersen-Felix S, Radanov B, Dickenson AH, Scaramozzino P, Arendt-Nielsen L, Curatolo M. The nociceptive withdrawal reflex: normative values of thresholds and reflex receptive fields. Eur J Pain 2010;14:134–41. [11] Neziri AY, Curatolo M, Limacher A, Nuesch E, Radanov B, Andersen OK, Arendt-Nielsen L, Juni P. Ranking of parameters of pain hypersensitivity according to their discriminative ability in chronic low back pain. PAIN 2012;153:2083–91. [12] Neziri AY, Curatolo M, Nuesch E, Scaramozzino P, Andersen OK, ArendtNielsen L, Juni P. Factor analysis of responses to thermal, electrical, and mechanical painful stimuli supports the importance of multi-modal pain assessment. PAIN 2011;152:1146–55. [13] Neziri AY, Limacher A, Juni P, Radanov BP, Andersen OK, ArendtNielsen L, Curatolo M. Ranking of tests for pain hypersensitivity according to their discriminative ability in chronic neck pain. Reg Anesth Pain Med 2013;38:308–20.
PAIN®
[14] Neziri AY, Scaramozzino P, Andersen OK, Dickenson AH, Arendt Nielsen L, Curatolo M. Reference values of mechanical and thermal pain tests in a pain-free population. Eur J Pain 2011;15:376–83. [15] Poole H, Bramwell R, Murphy P. The utility of the Beck Depression Inventory Fast Screen (BDI-FS) in a pain clinic population. Eur J Pain 2009;13:865–9. [16] Rhudy JL, France CR, Bartley EJ, Williams AE, McCabe KM, Russell JL. Does pain catastrophizing moderate the relationship between spinal nociceptive processes and pain sensitivity? J Pain 2009;10:860–9. [17] Rhudy JL, Maynard LJ, Russell JL. Does in vivo catastrophizing engage descending modulation of spinal nociception? J Pain 2007;8:325–33. [18] Sandrini G, Serrao M, Rossi P, Romaniello A, Cruccu G, Willer JC. The lower limb flexion reflex in humans. Prog Neurobiol 2005;77:353–95. [19] Scaramozzino P, Neziri AY, Andersen OK, Arendt Nielsen L, Curatolo M. Percentile normative values of parameters of electrical pain and reflex thresholds. Scand J Pain 2013;4:120–4. [20] Schliessbach J, Siegenthaler A, Streitberger K, Eichenberger U, Nuesch E, Juni P, Arendt-Nielsen L, Curatolo M. The prevalence of widespread central hypersensitivity in chronic pain patients. Eur J Pain 2013;17: 1502–10. [21] Sterling M, Hodkinson E, Pettiford C, Souvlis T, Curatolo M. Psychologic factors are related to some sensory pain thresholds but not nociceptive flexion reflex threshold in chronic whiplash. Clin J Pain 2008;24:124–30. [22] Woolf CJ. Central sensitization: implications for the diagnosis and treatment of pain. PAIN 2011;152:S2–15.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
Pain hypersensitivity and spinal nociceptive hypersensitivity in chronic pain: prevalence and associated factors Michele Curatoloa,*, Monika Muller ¨ b,c, Aroosiah Ashrafb, Alban Y. Nezirid, Konrad Streitbergerb, Ole K. Andersene, e Lars Arendt-Nielsen
Abstract Hypersensitivity of pain pathways is considered a relevant determinant of symptoms in chronic pain patients, but data on its prevalence are very limited. To our knowledge, no data on the prevalence of spinal nociceptive hypersensitivity are available. We studied the prevalence of pain hypersensitivity and spinal nociceptive hypersensitivity in 961 consecutive patients with various chronic pain conditions. Pain threshold and nociceptive withdrawal reflex threshold to electrical stimulation were used to assess pain hypersensitivity and spinal nociceptive hypersensitivity, respectively. Using 10th percentile cutoff of previously determined reference values, the prevalence of pain hypersensitivity and spinal nociceptive hypersensitivity (95% confidence interval) was 71.2 (68.3-74.0) and 80.0 (77.0-82.6), respectively. As a secondary aim, we analyzed demographic, psychosocial, and clinical characteristics as factors potentially associated with pain hypersensitivity and spinal nociceptive hypersensitivity using logistic regression models. Both hypersensitivity parameters were unaffected by most factors analyzed. Depression, catastrophizing, pain-related sleep interference, and average pain intensity were significantly associated with hypersensitivity. However, none of them was significant for both unadjusted and adjusted analyses. Furthermore, the odds ratios were very low, indicating modest quantitative impact. To our knowledge, this is the largest prevalence study on central hypersensitivity and the first one on the prevalence of spinal nociceptive hypersensitivity in chronic pain patients. The results revealed an impressively high prevalence, supporting a high clinical relevance of this phenomenon. Electrical pain thresholds and nociceptive withdrawal reflex explore aspects of pain processing that are mostly independent of sociodemographic, psychological, and clinical pain-related characteristics. Keywords: Pain sensitivity, Central sensitization, Quantitative sensory tests, Nociceptive withdrawal reflex, Prevalence
1. Introduction There is overwhelming evidence for hypersensitivity of nociceptive pathways in pain states, leading to pain amplification.22 Pain hypersensitivity can be measured in humans using quantitative sensory tests (QSTs). Spinal nociceptive hypersensitivity is typically assessed as electromyography response to an electrical stimulus applied to the lower extremity (nociceptive withdrawal reflex).18 Groups of patients with different pain conditions have higher pain sensitivity than groups of pain-free subjects.3 Pain patients display lower thresholds of nociceptive withdrawal reflex, compared with healthy subjects.9 Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article. a
Department of Anesthesiology and Pain Medicine, University of Washington, Seattle, WA, USA, b Department of Anesthesiology and Pain Therapy, University Hospital of Bern, Inselspital, Bern, Switzerland, c Center for Sensory–Motor Interaction, Institute of Social and Preventive Medicine, University of Bern, Bern, Switzerland, d Department of Obstetrics and Gynecology, Cantonal Hospital of Winterthur, Winterthur, Switzerland, e Center for Sensory–Motor Interaction, Department of Health Science and Technology, University of Aalborg, Aalborg, Denmark
Because of the subjective nature of the pain experience, the measurement of pain sensitivity has to rely on self-report; pain thresholds are therefore commonly used to assess pain hypersensitivity.4 In contrast, the measurement of nociceptive withdrawal reflex does not rely on self-report; it is therefore primarily an objective parameter of nociceptive spinal sensitivity.18 A clinically relevant question is how frequently hypersensitivity occurs in chronic pain patients. In 2 case–control studies on chronic low back and neck pain, pain and nociceptive reflex thresholds to electrical stimulation had the best discriminatory value for hypersensitivity among 26 different tests.11,13 These methods are therefore promising to study the epidemiology of hypersensitivity states. Importantly, we are not aware of data on the prevalence of nociceptive spinal hypersensitivity, a phenomenon that has been studied extensively in basic and clinical research and has an established role in the determination of pain states. The primary aim of this study was to determine the prevalence of pain hypersensitivity and nociceptive spinal hypersensitivity in chronic pain patients. The secondary aim was to analyze whether demographic, psychosocial, and pain-related factors are associated with pain hypersensitivity and spinal nociceptive hypersensitivity.
Institutional URL: http://depts.washington.edu/anesth/. *Corresponding author. Address: Department of Anesthesiology and Pain Medicine, University of Washington, 1959 NE Pacific St, Seattle, WA 98195, USA. Tel.: 11 206 543 2568; fax: 11 206 543 2958. E-mail address: [email protected] (M. Curatolo).
2.1. Participants
PAIN 156 (2015) 2373–2382 © 2015 International Association for the Study of Pain http://dx.doi.org/10.1097/j.pain.0000000000000289
November 2015
·
Volume 156
2. Methods
·
Number 11
We studied consecutive patients who were referred to the Pain Clinic of the University Department of Anesthesiology and Pain www.painjournalonline.com
2373
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
2374
·
PAIN®
M. Curatolo et al. 156 (2015) 2373–2382
Therapy of the University of Bern, Switzerland. A multimodal patient assessment that included sociodemographic, psychological, and clinical parameters, as well as QST, had been initiated in 2008. This study was conducted on patients presenting to the pain clinic from July 1, 2011, to June 30, 2013. All patients were informed in written form on the background of the QSTs and on the potential use of their data for scientific reasons. All participants gave consent before the tests. Approval was obtained by the institutional review board, ie, the Ethics Committee of the Canton of Bern, Switzerland (Study Number: 2684). Exclusion criteria were pain duration less than 3 months, neurological disorders of the lower extremity to be tested (palsy, paresthesia, polyneuropathy), any conditions potentially affecting the neurological function of the lower extremity (eg, diabetes mellitus or alcohol abuse), pain or vascular disorders of the lower extremity to be tested (including radicular pain), language problems, pregnancy, tumor pain (defined as pain in the region of infiltration by a primary tumor or metastasis), psychiatric diseases other than unipolar depressive disorder, and rheumatic inflammatory diseases. 2.2. Sociodemographic, psychological, and clinical characteristics Sociodemographic, psychological, and clinical characteristics were recorded by the time of first consultation (Table 1). Part of them were collected for descriptive purposes. A selection of these variables was used for the statistical analyses pertaining to the study aims (see “Statistical analyses” section). Sociodemographic variables were gender, age, body mass index (BMI), and work-related parameters. Psychological characteristics were depression and catastrophizing, assessed with the Beck Depression Inventory Fast Screen15 and the Catastrophizing Scale of the Coping Strategies Questionnaire,7 respectively. Clinical characteristics were history of trauma or surgery related to pain; pain duration; pain intensity immediately before testing; maximum, minimum, and average pain in the preceding 24 hours, rated on a 0 to 10 numerical rating scale (NRS), with 0 indicating no pain and 10 indicating the worst pain imaginable; pain-related life interference as measured by the Multidimensional Pain Inventory8; sleep disturbance as assessed on a 0 to 10 NRS (0 5 no sleep disturbance, 10 5 worst sleep disturbance imaginable); current pain medication, classified as daily intake of any pain medication, opioids, antidepressants, and anticonvulsants (yes/no); and type of pain syndrome. The pain syndrome was categorized as musculoskeletal, neuropathic, orofacial, visceral, noncervicogenic headache, complex regional pain syndrome, phantom limb pain, and atypical pain syndromes (Table 2). Musculoskeletal pain was defined as pain in or around bones, joints, or muscles, characterized by musculoskeletal features such as trigger points, tenderness, and loaddependent or mechanically triggered pain. Musculoskeletal pain was categorized as regional and widespread. Regional musculoskeletal pain included cervical pain, thoracic spine pain, low back pain, shoulder pain, hand pain, hip pain, knee pain, foot pain, and other regional musculoskeletal pain. Widespread musculoskeletal pain was defined as pain felt in the 4 body quadrants during the last week. Neuropathic pain was diagnosed as such when at least one of the following criteria was fulfilled: (1) clinical signs of nerve dysfunction (ie, palsy, sensory loss, paresthesia, allodynia, or hyperalgesia) associated with pain at the territory supplied by the nerve; (2) history of nerve lesion associated with pain in the territory supplied by the nerve; (3) magnetic resonance image finding of
Table 1
Sociodemographic, psychological, and clinical characteristics of patients included in the study and completeness of data. Baseline characteristics Sociodemographic characteristics Gender* Male Female Age, y* BMI, kg/m2* Working status Regular work as usual Reduced work due to pain No work Retired or studying Disability pension No Yes Ongoing litigation No Yes Psychological characteristics Depression (BDI-FS: 0-21, cutoff: 4)* No (score # 4) Yes (score . 4) Catastrophizing (CQS: 0-6)* Clinical characteristics History of trauma related to pain* No Yes History of surgery related to pain* No Yes Pain duration, y* ,1 1-2 .2 Maximum pain in the last 24 h (NRS: 0-10) Minimum pain in the last 24 h (NRS: 0-10) Average pain in the last 24 h (NRS: 0-10)* Pain intensity before testing (NRS: 0-10) Pain-related life interference (MPI: 0-6) Pain-related sleep interference (NRS: 0-10) Daily intake of pain medication* No Yes Daily intake of opioids No Yes Daily intake of antidepressants and/or anticonvulsants No Yes
Mean (SD) or no. of patients (%)
n (total 5 961) 961
422 (43.9) 539 (56.1) 50.0 (15.1) 26.0 (5.2)
961 921 934
284 (30.4) 170 (18.2) 300 (32.1) 180 (19.3) 927 835 (90.1) 92 (9.9) 925 694 (75.0) 231 (25.0) 898 409 (45.5) 489 (54.4) 3.2 (1.4)
913 961
714 (74.3) 247 (25.7) 954 543 (56.9) 411 (43.1) 2.4 (1.0-6.4)* 226 (24.6) 179 (19.5) 514 (55.9) 7.3 (2.0) 4.1 (2.5) 6.2 (2.1) 5.6 (2.5) 4.0 (1.3) 5.3 (3.0)
919
878 872 878 957 918 935 959
236 (24.6) 723 (75.4) 958 680 (71.0) 278 (29.0) 958 677 (70.7) 278 (29.3)
The type of pain syndrome is presented in a separate table. * Variables included in the regression analysis for factors potentially associated with central hypersensitivity. The other variables were collected for descriptive purposes. BDI-FS, Beck Depression Inventory Fast Screen, whereby 0 5 no depression and 21 5 maximum depression; BMI, body mass index; CQS, Catastrophizing Scale of the Coping Strategies Questionnaire, whereby 0 5 no catastrophizing and 6 5 maximum catastrophizing; MPI, Multidimensional Pain Inventory; NRS, numerical rating scale, whereby 0 5 no pain and 10 5 maximum pain.
nerve compression together with pain at its innervation. Neuropathic pain was categorized as postherpetic, posttraumatic nerve lesion, post-thoracotomy, radicular, trigeminal, and other neuropathic pain syndromes. Orofacial pain was defined as pain localized at the facial region or in the mouth, without clear neuropathic features as defined above (trigeminal neuralgia was
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
November 2015
·
Volume 156
·
Number 11
www.painjournalonline.com
Table 2
Classification of type of pain. Type of pain
No. of patients (%)
Musculoskeletal pain Regional musculoskeletal pain Neck pain Thoracic spine pain Low back pain Shoulder pain Hand pain Hip pain Knee pain Foot pain Other regional musculoskeletal pain Widespread pain Neuropathic pain Postherpetic Posttraumatic nerve lesion Post-thoracotomy Radicular Trigeminal neuralgia Other neuropathic pain states Orofacial pain Visceral pain Chronic pelvic pain Inguinal pain Other visceral pain syndromes Other pain syndromes Noncervicogenic headache CRPS Phantom limb pain Atypical pain
637 (66.3) 528 106 27 219 25 12 41 46 21 31 109 102 (10.6) 14 23 22 25 5 13 61 (6.3) 71 (7.4) 16 26 29 90 (9.4) 21 35 3 31
A pain syndrome was defined as atypical when it could not be classified as any of the pain syndromes as defined in the “Methods” section. CRPS, complex regional pain syndrome.
categorized as neuropathic pain). Visceral pain was diagnosed as such when it was located in the deep abdominal or pelvic region, with concomitant history of abdominal/pelvic disease or surgery. Noncervicogenic headache included migraine, cluster headache, and tension-type headache. Complex regional pain syndrome was diagnosed according to the published definition criteria.5 A pain syndrome was defined as atypical when it could not be classified as one of the above-mentioned ones. In case of multiple pain syndromes, all of them were registered. 2.3. Assessments of pain hypersensitivity and spinal nociceptive hypersensitivity Pain threshold to electrical stimulation and the threshold of the nociceptive withdrawal reflex were chosen because in 2 previous
2375
case–control studies, these assessments displayed the best discriminatory value for hypersensitivity among 26 different tests.11,13 At the beginning of the testing session, all the tests were performed on the patients for training purposes. Only when the patients felt familiar with the tests, we formally started the testing procedure. Electrical stimulation was performed through bipolar skin surface Ag/AgCl electrodes placed just distal to the lateral malleolus (innervation area of the sural nerve). The body side was contralateral to the side of most pain. In case of bilateral pain of equal intensity, the side was selected according to a computer-generated random sequence. Electromyography reflex responses to electrical stimulation were recorded from the middle of the biceps femoris and rectus femoris muscles (Ag/AgCl electrodes).10 Stimulation and electromyographic recordings were made by a computercontrolled constant current stimulator (NCS System, Evidence 3102 evo; Neurosoft, Ivanovo, Russia). A 25-millisecond train of five 1-millisecond square-wave impulse (perceived as a single stimulus) was delivered. The current intensity was increased from 1 mA in steps of 1 mA until (1) a biceps femoris reflex with an amplitude higher than 20 mV for at least 10 milliseconds in the 50- to 150millisecond poststimulation interval was detected (nociceptive reflex threshold); and (2) a pain sensation was evoked (pain threshold). These thresholds were measures of spinal nociceptive hypersensitivity and of pain hypersensitivity, respectively. If increasing the current intensity led to intolerable pain in the absence of reflex, only the pain threshold was recorded. 2.4. Statistical analyses The prevalence of pain hypersensitivity and spinal nociceptive hypersensitivity was calculated as the percentage of patients who displayed values equal to or below the 5th, 10th, and 25th percentiles of reference values, which were determined in previous studies on 300 pain-free subjects10,14 (Table 3). The prevalence of hypersensitivity and its corresponding 95% confidence intervals (CIs) were calculated using a binominal distribution. We modeled factors potentially associated with hypersensitivity using logistic regression models based on maximum likelihood estimation. For this purpose, we considered the 10th percentile as cutoff for the presence of hypersensitivity as primary outcome. The factors analyzed were gender, age, BMI, depression, catastrophizing, history of trauma or surgery related to the pain condition, type of pain, pain duration, average pain intensity of the last 24 hours, and daily intake of pain medication. We based the choice of these variables on previous research on factors potentially associated with hypersensitivity and clinical reasoning. Associations between hypersensitivity and
Table 3
Normative values of pain detection and reflex threshold after electrical stimulation, based on data generated in 300 pain-free individuals. Test Electrical pain detection threshold
Electrical reflex threshold
Gender
Age, y
5th percentile
10th percentile
25th percentile
Female Female Male Male Female Female Male Male
20-49 50-80 20-49 50-80 20-49 50-80 20-49 50-80
6.0 8.0 7.0 9.0 9.0 11.0 10.7 12.3
7.0 9.0 7.7 9.0 10.0 13.7 11.7 13.3
7.7 10.0 8.3 10.0 12.0 15.7 13.3 14.3
All values are in milliamperes.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
2376
·
PAIN®
M. Curatolo et al. 156 (2015) 2373–2382
these baseline patient characteristics were expressed as odds ratios (ORs), including 95% CIs. We first modeled univariate associations, then partially adjusted associations including nonpain-related factors such as gender, age, BMI, depression, and catastrophizing scale, and finally computed a fully adjusted model including all baseline characteristics potentially associated with the presence of pain hypersensitivity and spinal nociceptive hypersensitivity. During testing, we noticed that in roughly one-third of the patients, the reflex threshold could not be determined, as pain became intolerable before a reflex was obtained. We estimated the prevalence of hypersensitivity and potentially associated factors based on 3 different assumptions to account for these missing values and make the results of pain and reflex thresholds comparable. In a first step, we conducted a complete case analysis, ie, we restricted the analysis to the patients in whom a reflex threshold could be obtained (n 5 696). In a second step, a multiple imputation analysis was performed, ie, missing values for reflex threshold and other incompletely recorded explanatory variables were imputed; prevalence of hypersensitivity and associated factors were estimated based on the imputed data set.6 We used a multivariate sequential imputation technique based on chained equations generating 15 multiple imputed data sets. Where possible, we imputed continuous values. For analysis, we dichotomized variables as reported in Table 1. We finally performed a third analysis in which patients who did not display a reflex were considered as being not hypersensitive.
3. Results A total of 1508 patients were considered for the testing procedure. Of these, 547 patients were not included (reasons in parentheses): 50 (pain duration less than 3 months), 220 (neurological disorders), 16 (vascular disorders), 45 (tumor pain), 36 (psychiatric diseases other than unipolar depressive disorder), 25 (rheumatic inflammatory disease), and 155 (other reasons, including language problems, logistic reasons, previous bilateral leg amputation, and pregnancy). Nine hundred sixty-one patients were therefore eligible for the analyses. It was possible to determine pain detection threshold after electrical stimulation in all 961 patients. A nociceptive reflex could be evoked in 696 patients (72.4%); in the remaining patients, pain became intolerable at stimulus intensities that did not evoke a reflex. The mean pain and reflex thresholds in these 696 patients with complete records were 6.4 mA (SD 5 3.3) and 8.1 mA (SD 5 4.8), respectively. Tables 1 and 2 report baseline patient characteristics.
3.1. Prevalence of pain hypersensitivity and spinal nociceptive hypersensitivity The prevalence of pain hypersensitivity and spinal nociceptive hypersensitivity at the different cutoffs for reference values is reported in Table 4 and illustrated in Figures 1 and 2. Probability density curves of the normative data and patient data for pain hypersensitivity and spinal nociceptive hypersensitivity are illustrated in Figures 3 and 4, respectively. In the multiple imputation analysis, the prevalence of pain hypersensitivity and spinal nociceptive hypersensitivity (95% CI) with 10th percentile cutoff was 71.2 (68.3-74.0) and 80.0 (77.082.6), respectively. These findings were very similar with the complete case analysis. When considering patients without reflex as being not hypersensitive, the prevalence of spinal nociceptive hypersensitivity at the 10% cutoff dropped to 57.9 (54.8-61.1). 3.2. Factors associated with pain hypersensitivity and spinal nociceptive hypersensitivity The results of the regression analyses for factors potentially associated with pain hypersensitivity and spinal nociceptive hypersensitivity are displayed in Tables 5–7. Table 5 presents the results of the analysis on all cases (n 5 961) and is based on the imputed data set. In the unadjusted analysis, depression, catastrophizing, and pain-related sleep interference were significantly associated with pain hypersensitivity, but not with spinal nociceptive hypersensitivity. Depression, catastrophizing, and pain-related sleep interference were not statistically significant anymore after adjustment for other variables. The average pain intensity was significantly associated with both pain hypersensitivity and spinal nociceptive hypersensitivity in the unadjusted analyses; the significance was not observed anymore in the adjusted analysis for spinal nociceptive hypersensitivity, but was still present for pain hypersensitivity. The association was positive: the higher the pain intensity, the higher the risk of being hypersensitive. All other variables did not display statistically significant correlations. These results were mostly confirmed by the analysis limited on the complete data set (n 5 696, Table 6). The only differences were that pain-related sleep interference was still significantly related to pain hypersensitivity after adjustment and the average pain intensity was not significant anymore in the adjusted analyses on pain hypersensitivity. When patients with missing reflexes were considered as being not hypersensitive, the average pain intensity was no longer significant for nociceptive spinal hypersensitivity (Table 7). Having
Table 4
Prevalence of hypersensitivity as assessed by electrical stimulation and using cutoffs at different normative percentiles. Prevalence in percent (95% CI) Multiple imputation analysis (n 5 961) Pain hypersensitivity Spinal nociceptive hypersensitivity Complete case analysis (n 5 696) Pain hypersensitivity Spinal nociceptive hypersensitivity Patients with missing reflex considered as being not hypersensitive (n 5 961) Pain hypersensitivity Spinal nociceptive hypersensitivity
5th percentile cutoff
10th percentile cutoff
25th percentile cutoff
62.0 (58.8-65.0) 72.3 (69.0-75.5)
71.2 (68.3-74.0) 80.0 (77.0-82.6)
80.7 (78.3-83.2) 87.5 (85.0-90.0)
62.5 (59.0-66.1) 72.8 (69.5-76.2)
72.2 (68.9-75.6) 80.0 (77.1-83.0)
82.0 (79.1-84.9) 87.6 (85.2-90.1)
NA 52.8 (49.6-55.9)
NA 57.9 (54.8-61.1)
NA 63.5 (60.4-66.5)
Pain hypersensitivity was assessed by pain threshold. Spinal nociceptive hypersensitivity was assessed by the nociceptive withdrawal reflex threshold. The table presents the results of the multiple imputation analysis, complete case analysis, and a sensitivity analysis that considered patients without reflex as being not hypersensitive. CI, confidence interval; NA, not applicable.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
November 2015
·
Volume 156
·
Number 11
Figure 1. Prevalence of pain hypersensitivity as assessed by electrical stimulation, using cutoffs at 5th, 10th, and 25th percentiles. The graph reports the results of the multiple imputation analysis (Table 4, n 5 961).
pain for more than 2 years reduced the risk of hypersensitivity, but only in the unadjusted analysis. Daily intake of pain medication reduced the risk of hypersensitivity in both unadjusted and adjusted analyses.
4. Discussion To our knowledge, this is the largest study on the prevalence of pain hypersensitivity in chronic pain patients and the first one on the prevalence of spinal nociceptive hypersensitivity. It revealed an impressively high prevalence for both parameters, supporting the clinical relevance of phenomena that have been extensively investigated in basic and clinical research. Both pain and spinal nociceptive hypersensitivity were not associated with most sociodemographic, psychological, and clinical characteristics.
Figure 2. Prevalence of spinal nociceptive hypersensitivity as assessed by the nociceptive withdrawal reflex, using cutoffs at 5th, 10th, and 25th percentiles. The graph reports the results of the multiple imputation analysis (black bars, n 5 961) and of the analysis that considered patients without reflex as being not hypersensitive (gray bars, n 5 961) (Table 4).
www.painjournalonline.com
2377
Figure 3. Probability density curve of the experimental data collected in chronic pain patients (blue curve) and of the normative data collected in painfree subjects (red curve) for pain hypersensitivity. The vertical lines represent the 5th, 10th, and 25th percentiles of the normative data considered as cutoff values to define the prevalence of pain hypersensitivity.
4.1. Prevalence of pain hypersensitivity In the multiple imputation analysis, the prevalence of pain hypersensitivity (95% CI) using the 5th, 10th, and 25th percentile cutoffs was 62.0% (58.8-65.0), 71.2% (68.3-70.0), and 80.7% (78.3-83.2), respectively (Table 4 and Fig. 1). Very similar results were obtained by the complete case analysis. In a previous study using pressure pain tolerance threshold at the second toe, the prevalence of pain hypersensitivity was substantially lower: at the 5th, 10th, and 25th percentile cutoffs, the prevalence was 17.5%, 24.5%, and 35.5%, respectively.20 It is difficult to draw firm conclusions by comparing the previous with this study, as they have been conducted on 2 different patient populations. However, both studies have been performed in the same setting and using the same data set for calculation of reference values.10,19 It is therefore unlikely that the much higher prevalence observed in this study is entirely due to a much higher pain sensitivity of the patient population analyzed. One possible explanation is the different methodology used to assess hypersensitivity, with electrical pain
Figure 4. Probability density curve of the experimental data collected in chronic pain patients (blue curve) and of the normative data collected in painfree subjects (red curve) for spinal nociceptive hypersensitivity. The vertical lines represent the 5th, 10th, and 25th percentiles of the normative data considered as cutoff values to define the prevalence of spinal nociceptive hypersensitivity.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
2378
Pain hypersensitivity (electrical pain detection threshold)
Gender Male Female Age (per 10 y increase) BMI (cutoff: 25 kg/m2) Normal weight Overweight Depression (BDI-FS, cutoff: 4) No (score # 4) Yes (score . 4) Catastrophizing (CQS: 0-6) History of trauma related to pain No Yes History of surgery related to pain No Yes Type of pain Musculoskeletal Neuropathic Orofacial Visceral Other pain syndromes Pain duration, y ,1 1-2 .2 Average pain (NRS: 0-10) Pain-related sleep inference (NRS: 0-10) Daily intake of any pain medication No Yes
P
Spinal nociceptive hypersensitivity (nociceptive reflex threshold)
Partially adjusted OR (95% CI)
P
Fully adjusted OR (95% CI)
P
Unadjusted OR (95% CI)
P
Partially adjusted OR (95% CI)
P
Fully adjusted OR (95% CI)
P
·
Unadjusted OR (95% CI) 1 1.24 (0.93-1.64) 1.04 (0.95-1.14)
0.14 0.43
1 1.25 (0.94-1.67) 1.04 (0.95-1.15)
1 0.12 1.17 (0.87-1.57) 0.39 1.04 (0.94-1.15)
1 0.30 1.06 (0.75-1.50) 0.41 1.10 (0.97-1.22)
1 0.75 1.06 (0.74-1.50) 0.13 1.09 (0.97-1.22)
1 0.75 0.95 (0.66-1.37) 0.14 1.08 (0.96-1.22)
0.78 0.19
1 0.96 (0.73-1.27)
0.79
1 0.94 (0.70-1.25)
1 0.66 0.90 (0.67-1.21)
1 0.50 1.08 (0.76-1.52)
1 0.67 1.05 (0.74-1.49)
1 0.79 1.05 (0.73-1.52)
0.78
1 1.54 (1.16-2.05) 1.19 (1.07-1.32)
1 0.003 1.30 (0.91-1.84) 0.001 1.14 (1.00-1.29)
1 0.15 1.32 (0.93-1.88) 0.05 1.04 (0.90-1.19)
1 0.12 1.06 (0.73-1.52) 0.62 1.04 (0.92-1.18)
1 0.77 0.99 (0.74-1.49) 0.53 1.04 (0.90-1.21)
1 0.96 1.02 (0.66-1.58) 0.57 0.98 (0.83-1.16)
0.92 0.85
1 0.97 (0.71-1.34)
0.87
NA
NA
1 0.98 (0.69-1.38)
1 0.89 0.76 (0.51-1.12)
0.17 NA
NA
1 0.76 (0.50-1.14)
0.19
1 1.02 (0.77-1.36)
0.88
NA
NA
1 1.03 (0.76-1.40)
1 0.86 0.76 (0.55-1.06)
0.11 NA
NA
1 0.79 (0.55-1.14)
0.22
1 0.93 (0.59-1.47) 1.19 (0.65-2.18) 0.71 (0.43-1.20) 0.82 (0.51-1.31)
0.76 0.57 0.20 0.40
NA
1 1.06 (0.66-1.72) 1.46 (0.77-2.77) 0.83 (0.47-1.46) 0.88 (0.53-1.46)
0.81 0.25 0.51 0.62
1 1.1 (0.58-1.78) 1.12 (0.51-2.46) 0.78 (0.41-1.50) 1.00 (0.56-1.77)
0.97 0.77 0.46 0.99 NA
NA
1 0.99 (0.55-1.77) 1.17 (0.51-2.66) 0.83 (0.41-1.68) 1.13 (0.60-2.10)
0.96 0.71 0.60 0.70
NA NA NA
1 1.07 (0.68-1.67) 1.07 (0.75-1.53) 1.09 (1.00-1.18) 1.05 (0.99-1.11)
0.78 0.72 0.05 0.11
1 0.79 (0.45-1.38) 0.66 (0.41-1.07) 1.09 (1.01-1.18) 1.04 (0.97-1.10)
0.41 0.09 NA 0.04 NA 0.25
NA NA
1 0.84 (0.48-1.50) 0.64 (0.39-1.06) 1.08 (0.97-1.19) 1.08 (0.97-1.19)
0.57 0.08 0.16 0.68
NA
1 0.94 (0.67-1.33)
1 0.47 1.12 (0.75-1.66)
0.57 NA
NA
1 1.01 (0.66-1.56)
0.95
1 0.96 (0.62-1.48) 1.06 (0.75-1.49) 1.15 (1.08-1.23) 1.09 (1.04-1.15)
1 1.20 (0.88-1.65)
NA
0.85 0.76 NA ,0.001 NA ,0.001 NA
0.25
NA
M. Curatolo et al. 156 (2015) 2373–2382
Table 5
Factors associated with pain hypersensitivity and spinal nociceptive hypersensitivity.
The 10th percentile of normative values was used as a cutoff to categorize patients as hypersensitive or not. The table presents unadjusted and adjusted ORs with 95% CIs and P-values with multiple imputation analysis for missing measurements of nociceptive reflex threshold. Partial adjustment was for nonpain-related variables: gender, age, BMI, depression, and catastrophizing; full adjustment was for all variables. The effect of continuous variables is displayed per standardized increase of 1 SD unit. BDI-FS, Beck Depression Inventory Fast Screen, ranging from 0 5 no depression to 21 5 maximum depression; BMI, body mass index; CI, confidence interval; CQS, Catastrophizing Scale of the Coping Strategies Questionnaire, whereby 0 5 no catastrophizing and 6 5 maximum catastrophizing; NA, not applicable; NRS, numerical rating scale, whereby 0 5 no pain and 10 5 worst pain imaginable; OR, odds ratio.
PAIN®
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
November 2015
Table 6
·
Factors associated with pain hypersensitivity and spinal nociceptive hypersensitivity. Unadjusted OR (95% CI)
Spinal nociceptive hypersensitivity (nociceptive reflex threshold)
Partially adjusted OR (95% CI)
P
Fully adjusted OR (95% CI)
P
Unadjusted OR (95% CI)
P
Partially adjusted OR (95% CI)
P
Fully adjusted OR (95% CI)
P
·
0.60 0.32
1 1.21 (0.85-1.72) 1.06 (0.94-1.20)
1 0.30 1.11 (0.76-1.62) 0.33 1.06 (0.94-1.21)
1 0.59 1.01 (0.69-1.47) 0.35 1.09 (0.96-1.24)
1 0.96 1.01 (0.68-1.50) 0.17 1.11 (0.97-1.26)
1 0.96 0.85 (0.56-1.30) 0.14 1.11 (0.96-1.29)
0.45 0.16
1 1.09 (0.78-1.53)
0.61
1 0.99 (069-1.42)
1 0.97 0.97 (0.67-1.42)
1 0.89 1.15 (0.79-1.68)
1 0.46 1.09 (0.73-1.61)
1 0.69 1.09 (0.72-1.66)
0.69
1 1.60 (1.36-2.27) 1.21 (1.06-1.37)
1 0.007 1.34 (0.89-2.03) 0.003 1.15 (0.98-1.33)
1 0.16 1.35 (0.87-2.08) 0.08 1.06 (0.87-1.26)
1 0.18 0.99 (0.67-1.45) 0.54 1.04 (0.90-1.19)
1 0.96 0.90 (0.57-1.44) 0.59 1.05 (0.89-1.24)
1 0.67 0.98 (0.60-1.60) 0.55 0.98 (0.81-1.19)
0.93 0.87
1 1.09 (0.74-1.60)
0.67
NA
NA
1 0.98 (0.62-1.54)
1 0.93 0.74 (0.49-1.11)
0.15 NA
NA
1 0.73 (0.45-1.18)
0.20
1 1.01 (0.72-1.42)
0.94
NA
NA
1 1.07 (0.72-1.59)
1 0.75 0.70 (0.48-1.02)
0.06 NA
NA
1 0.69 (0.45-1.07)
0.10
1 0.79 (0.47-1.34) 1.28 (0.61-2.66) 0.73 (0.40-1.33) 0.89 (0.51-1.53)
0.38 0.51 0.31 0.67
NA
1 0.87 (0.49-1.56) 1.60 (0.71-3.64) 0.72 (0.35-1.47) 1.05 (0.55-2.00)
0.65 0.26 0.37 0.88
1 0.83 (0.45-1.50) 1.11 (0.50-2.47) 0.76 (0.39-1.48) 1.00 (0.53-1.87)
0.55 0.80 0.42 0.99 NA
NA
1 0.89 (0.46-1.72) 0.37 (0.55-3.38) 0.90 (0.41-1.99) 1.19 (0.57-2.46)
0.73 0.50 0.80 0.64
NA NA NA
1 1.14 (0.65-1.98) 1.05 (0.68-1.65) 1.07 (0.97-1.19) 1.08 (1.00-1.16)
0.65 0.81 0.65 0.04
1 0.80 (0.44-1.46) 0.63 (0.38-1.02) 1.08 (0.99-1.18) 1.4 (0.98-1.11)
0.46 0.06 NA 0.07 NA 0.23 NA
NA NA NA
1 0.92 (0.48-1.77) 0.61 (0.37-1.03) 1.10 (0.98-1.24) 1.02 (0.94-1.11)
0.80 0.07 0.11 0.66
NA
1 0.75 (0.49-1.14)
1 0.18 1.02 (0.67-1.54)
0.93 NA
NA
1 0.94 (0.59-1.50)
0.80
1 1.04 (0.63-1.72) 1.08 (0.72-1.62) 1.15 (1.06-1.24) 1.12 (1.06-1.18)
1 0.99 (0.68-1.43)
NA
0.88 0.71 NA ,0.001 NA ,0.001 NA
0.95
NA
The 10th percentile of normative values was used as a cutoff to categorize patients as hypersensitive or not. The table presents unadjusted and adjusted ORs with 95% CIs and P-values from complete case analysis (n 5 612-696). Partial adjustment was for nonpain-related variables: gender, age, BMI, depression, and catastrophizing; full adjustment was for all variables. The effect of continuous variables is displayed per standardized increase of 1 SD unit. BDI-FS, Beck Depression Inventory Fast Screen, ranging from 0 5 no depression to 21 5 maximum depression; BMI, body mass index; CI, confidence interval; CQS, Catastrophizing Scale of the Coping Strategies Questionnaire, whereby 0 5 no catastrophizing and 6 5 maximum catastrophizing; NA, not applicable; NRS, numerical rating scale, whereby 0 5 no pain and 10 5 worst pain imaginable; OR, odds ratio.
www.painjournalonline.com
1 1.09 (0.78-1.53) 1.06 (0.95-1.18)
Number 11
Gender Male Female Age (per 10 y increase) BMI (cutoff: 25 kg/m2) Normal weight Overweight Depression (BDI-FS, cutoff: 4) No (score # 4) Yes (score . 4) Catastrophizing (CQS: 0-6) History of trauma related to pain No Yes History of surgery related to pain No Yes Type of pain Musculoskeletal Neuropathic Orofacial Visceral Other pain syndromes Pain duration, y ,1 1-2 .2 Average pain (NRS: 0-10) Pain-related sleep inference (NRS: 0-10) Daily intake of any medication No Yes
P
Volume 156
Pain hypersensitivity (electrical pain detection threshold)
2379
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
2380
·
PAIN®
M. Curatolo et al. 156 (2015) 2373–2382
Table 7
Factors associated with spinal nociceptive hypersensitivity, considering patients with missing reflex as being not hypersensitive (n 5 845-961). Spinal nociceptive hypersensitivity (nociceptive reflex threshold) Gender Male Female Age (per 10 y increase) BMI (cutoff: 25 kg/m2) Normal weight Overweight Depression (BDI-FS, cutoff: 4) No (score # 4) Yes (score . 4) Catastrophizing (CQS: 0-6) History of trauma related to pain No Yes History of surgery related to pain No Yes Type of pain Musculoskeletal Neuropathic Orofacial Visceral Other pain syndromes Pain duration, y ,1 1-2 .2 Average pain (NRS: 0-10) Pain-related sleep inference (NRS: 0-10) Daily intake of any medication No Yes
Unadjusted OR (95% CI)
P
Partially adjusted OR (95% CI)
P
Fully adjusted OR (95% CI)
P
1 1.10 (0.85-1.43) 1.00 (0.92-1.09)
0.46 0.97
1 1.08 (0.82-1.42) 1.00 (0.99-1.01)
0.58 0.91
1 1.08 (0.81-1.44) 1.00 (0.99-1.01)
0.61 0.99
1 0.86 (0.66-1.12)
0.27
1 0.88 (9.67-1.16)
0.37
1 0.92 (0.69-1.23)
0.60
1 0.99 (0.76.1.29) 1.05 (0.95-1.15)
0.91 0.36
1 0.92 (0.67-1.28) 1.06 (0.95-1.20)
0.64 0.30
1 0.98 (0.70-1.37) 1.08 (0.94-1.23)
0.88 0.26
1 0.85 (0.63-1.14)
0.27
NA
NA
1 0.77 (0.55-1.08)
0.14
1 0.92 (0.71-1.19)
0.52
NA
NA
1 1.01 (0.75-1.36)
0.95
1 1.04 (0.68-1.59) 1.17 (0.68-2.00) 1.04 (0.63-1.71) 1.38 (0.87-2.18)
0.85 0.56 0.88 0.17
NA
1 1.03 (0.65-1.64) 1.01 (0.57-1.81) 1.01 (0.60-2.00) 1.59 (0.93-2.71)
0.90 0.97 0.76 0.09
1 0.92 (0.62-1.38) 0.71 (0.51-0.98) 1.00 (0.94-1.06) 0.99 (0.94-1.03)
0.69 0.04 0.89 0.52
NA NA NA
NA NA NA
1 0.89 (0.58-1.37) 0.74 (0.52-1.04) 1.03 (0.94-1.11) 0.98 (0.92-1.03)
0.59 0.09 0.55 0.43
1 0.67 (0.49-0.91)
0.009
NA
NA
1 0.65 (0.46-0.92)
0.01
NA
The 10th percentile of normative values was used as a cutoff to categorize patients as hypersensitive or not. The table presents unadjusted and adjusted ORs with 95% CIs and P-values. Partial adjustment was for nonpain-related variables: gender, age, BMI, depression, and catastrophizing; full adjustment was for all variables. The effect of continuous variables is displayed per standardized increase of 1 SD unit. BDI-FS, Beck Depression Inventory Fast Screen, ranging from 0 5 no depression to 21 5 maximum depression; BMI, body mass index; CI, confidence interval; CQS, Catastrophizing Scale of the Coping Strategies Questionnaire, whereby 0 5 no catastrophizing and 6 5 maximum catastrophizing; NA, not applicable; NRS, numerical rating scale, whereby 0 5 no pain and 10 5 worst pain imaginable; OR, odds ratio.
threshold possibly being more sensitive than pressure pain tolerance threshold to detect pain hypersensitivity. In this study, the stimulation was performed at a body site that was far distant from the site of pain. Patients with any pain or pathology at the tested site were excluded. We therefore measured widespread pain hypersensitivity, likely reflecting a generalized hyperexcitability of central pain pathways. 4.2. Prevalence of spinal nociceptive hypersensitivity In the multiple imputation analysis, the prevalence of spinal nociceptive hypersensitivity (95% CI) using the 5th, 10th, and 25th percentile cutoffs was 72.3% (69.0-75.5), 80.0% (77.0-82.6), and 87.5% (85.0-90.0), respectively (Table 4 and Fig. 2). Very similar results were obtained by the complete case analysis. A nociceptive reflex could not be evoked in 37.6% of all patients, as pain became intolerable at stimulus intensities that did not evoke a reflex. It is unclear why the reflex cannot be evoked in a subset of patients, but this phenomenon has been repeatedly observed in previous studies.1,2,21 In a further analysis that considered these patients as being not hypersensitive, the prevalence obviously dropped, but was still significantly high, ranging 52.8% to 63.5% depending on the cutoff (Table 4 and Fig. 2).
As for the determination of pain hypersensitivity, the assessment of the nociceptive reflex was performed at a body site where no injury potentially causing pain was present. We therefore likely assessed a state of generalized spinal nociceptive hyperexcitability. 4.3. Factors associated with pain hypersensitivity and spinal nociceptive hypersensitivity Both pain hypersensitivity and spinal nociceptive hypersensitivity were not associated with most sociodemographic, psychological, and clinical characteristics (Tables 5–7). The few variables that were statistically significant displayed very low ORs, raising doubts on the clinical relevance of the impact. Interestingly, this finding is consistent with the results of a previous study on 300 pain-free subjects using the same paradigms of electrical stimulation.10 This suggests that electrical pain threshold and nociceptive withdrawal reflex explore aspect of sensitization processes that are independent of most sociodemographic, psychological, and clinical characteristics. The variable that was most consistently associated with pain hypersensitivity and spinal nociceptive hypersensitivity was baseline spontaneous pain intensity: the higher the patients
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
November 2015
·
Volume 156
·
Number 11
scored on the NRS, the higher the risk of hypersensitivity with both pain and nociceptive reflex threshold. However, it has to be noticed that the quantitative impact of this variable was modest, as the ORs were only slightly above 1. Pain intensity is the result of multiple and mostly uncontrollable factors, including nociceptive input from damaged tissues, neuroplastic changes, alterations in central pain modulation, and cognitive and affective factors. In a human study, it is not possible to find out which of these factors are responsible for the influence of pain intensity on parameters of pain hypersensitivity and spinal nociceptive hypersensitivity. Independent of the pathophysiologic explanation, it is likely that pain intensity and hypersensitivity processes feed each other in a positive feedback loop: high pain intensity causes more hypersensitivity, which in turn contributes to higher pain levels. Accordingly, acting at this loop by reducing hypersensitivity seems to be a relevant therapeutic aim. The magnitude of pain-related interference with sleep was associated with increased risk of pain hypersensitivity. However, also for this variable, the OR was only slightly above 1, indicating a modest quantitative impact. Nevertheless, the finding suggests that improving sleep quality may be important for the management of pain hypersensitivity, potentially leading to reduction in pain. In a previous study using pressure pain tolerance thresholds, different factors were found to be positively related with pain hypersensitivity.20 They included female gender, duration of pain, and having an atypical pain syndrome. Catastrophizing and baseline pain intensity were marginally not significant, and depression was significant only with the univariate analysis. Remarkably, a study on 300 pain-free subjects14 found female gender to be significantly associated with hypersensitivity to pressure stimulation and catastrophizing was marginally not significant, basically mirroring the results of the aforementioned study conducted on patients.20 Previous research on both painfree subjects and chronic pain patients found catastrophizing to be related with pain thresholds, but not with the nociceptive withdrawal reflex.16,17,21 This suggests that catastrophizing affects pain through supraspinal mechanisms, rather than through descending modulatory processes. Overall, it seems that pain hypersensitivity as assessed by electrical stimulation is less associated with sociodemographic, psychological, and clinical characteristics, compared with pain hypersensitivity as assessed by pressure stimulation. We know that different stimulation modalities explore different dimensions of pain perception.12 However, the limited knowledge of the specific mechanisms explored by different test modalities does not allow clear explanations for these differences. The limited association of demographic, psychological, and clinical characteristics with electrical tests suggests that this modality may explore only a limited part of the complex pain experience. This however is not necessarily a limitation. Both clinicians and researchers may find it useful to include assessments that are minimally affected by confounding and frequently uncontrollable parameters, leading to large unexplained variability. 4.4. Strengths and limitations The main strengths are the large sample size and the combination of measures of both subjective pain sensitivity and spinal nociceptive processes. Considering the overwhelming data from basic and clinical research on hypersensitivity, the study fills an important translational gap. The results pertain to the methodology used. Because we performed a large investigation for a long time period, we could not apply an extensive experimental protocol involving multiple tests. As stated above, different
www.painjournalonline.com
2381
methods to assess hypersensitivity explore different neural mechanisms and pain dimensions, potentially leading to different results in terms of prevalence and its determining variables. Further research will hopefully clarify which modality is most closely related to clinically significant outcomes, such as pain, disability, or course of the disease. Also, the results pertain to a patient population of a tertiary pain clinic. As for any epidemiologic study, their external validity needs to be determined by further studies conducted in different settings and on different patient populations. 4.5. Summary conclusions This study found an impressively high prevalence of pain hypersensitivity and spinal nociceptive hypersensitivity in a population of chronic pain patients attending a tertiary pain center. Depending on the cutoff of reference values and the type of analysis, the prevalence of pain hypersensitivity and spinal nociceptive hypersensitivity ranged from 62.5% to 82.0% and from 52.8% to 87.6%, respectively. Both hypersensitivity parameters were not associated with most baseline variables, suggesting that electrical pain thresholds and nociceptive withdrawal reflex explore aspects of pain processing that are mostly independent of sociodemographic, psychological, and clinical characteristics. Because of the likely influence of hypersensitivity on clinically relevant outcomes, clinical quantification of hypersensitivity may improve our understanding of the pathophysiology of pain in individual patients. Furthermore, the development of effective therapeutic strategies that attenuate the hypersensitivity state is an important target of pain research.
Conflict of interest statement The authors have no conflicts of interest to declare. The study was funded by the Swiss National Science Foundation (32003B_138361), the Danish Research Council for Technology and Production, and the Scientific Funds of the University Department of Anesthesiology and Pain Therapy of the University of Bern.
Acknowledgements M. Muller ¨ and A. Ashraf contributed equally. Article history: Received 6 May 2015 Received in revised form 25 June 2015 Accepted 2 July 2015 Available online 7 July 2015
References [1] Banic B, Petersen-Felix S, Andersen OK, Radanov BP, Villiger PM, Arendt-Nielsen L, Curatolo M. Evidence for spinal cord hypersensitivity in chronic pain after whiplash injury and in fibromyalgia. PAIN 2004;107: 7–15. [2] Biurrun Manresa JA, Fritsche R, Vuilleumier PH, Oehler C, Morch CD, Arendt-Nielsen L, Andersen OK, Curatolo M. Is the conditioned pain modulation paradigm reliable? A test-retest assessment using the nociceptive withdrawal reflex. PLoS One 2014;9:e100241. [3] Curatolo M. Diagnosis of altered central pain processing. Spine 2011;36: S200–204. [4] Curatolo M, Arendt-Nielsen L. Central hypersensitivity in chronic musculoskeletal pain. Phys Med Rehabil Clin N Am 2015;26:175–84. [5] Harden RN, Bruehl S, Stanton-Hicks M, Wilson PR. Proposed new diagnostic criteria for complex regional pain syndrome. Pain Med 2007;8: 326–31.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
2382
·
M. Curatolo et al. 156 (2015) 2373–2382
[6] Harel O, Zhou XH. Multiple imputation: review of theory, implementation and software. Stat Med 2007;26:3057–77. [7] Keefe FJ, Brown GK, Wallston KA, Caldwell DS. Coping with rheumatoid arthritis pain: catastrophizing as a maladaptive strategy. PAIN 1989;37:51–6. [8] Kerns RD, Turk DC, Rudy TE. The West Haven-Yale Multidimensional Pain Inventory (WHYMPI). PAIN 1985;23:345–56. [9] Lim EC, Sterling M, Stone A, Vicenzino B. Central hyperexcitability as measured with nociceptive flexor reflex threshold in chronic musculoskeletal pain: a systematic review. PAIN 2011;152:1811–20. [10] Neziri AY, Andersen OK, Petersen-Felix S, Radanov B, Dickenson AH, Scaramozzino P, Arendt-Nielsen L, Curatolo M. The nociceptive withdrawal reflex: normative values of thresholds and reflex receptive fields. Eur J Pain 2010;14:134–41. [11] Neziri AY, Curatolo M, Limacher A, Nuesch E, Radanov B, Andersen OK, Arendt-Nielsen L, Juni P. Ranking of parameters of pain hypersensitivity according to their discriminative ability in chronic low back pain. PAIN 2012;153:2083–91. [12] Neziri AY, Curatolo M, Nuesch E, Scaramozzino P, Andersen OK, ArendtNielsen L, Juni P. Factor analysis of responses to thermal, electrical, and mechanical painful stimuli supports the importance of multi-modal pain assessment. PAIN 2011;152:1146–55. [13] Neziri AY, Limacher A, Juni P, Radanov BP, Andersen OK, ArendtNielsen L, Curatolo M. Ranking of tests for pain hypersensitivity according to their discriminative ability in chronic neck pain. Reg Anesth Pain Med 2013;38:308–20.
PAIN®
[14] Neziri AY, Scaramozzino P, Andersen OK, Dickenson AH, Arendt Nielsen L, Curatolo M. Reference values of mechanical and thermal pain tests in a pain-free population. Eur J Pain 2011;15:376–83. [15] Poole H, Bramwell R, Murphy P. The utility of the Beck Depression Inventory Fast Screen (BDI-FS) in a pain clinic population. Eur J Pain 2009;13:865–9. [16] Rhudy JL, France CR, Bartley EJ, Williams AE, McCabe KM, Russell JL. Does pain catastrophizing moderate the relationship between spinal nociceptive processes and pain sensitivity? J Pain 2009;10:860–9. [17] Rhudy JL, Maynard LJ, Russell JL. Does in vivo catastrophizing engage descending modulation of spinal nociception? J Pain 2007;8:325–33. [18] Sandrini G, Serrao M, Rossi P, Romaniello A, Cruccu G, Willer JC. The lower limb flexion reflex in humans. Prog Neurobiol 2005;77:353–95. [19] Scaramozzino P, Neziri AY, Andersen OK, Arendt Nielsen L, Curatolo M. Percentile normative values of parameters of electrical pain and reflex thresholds. Scand J Pain 2013;4:120–4. [20] Schliessbach J, Siegenthaler A, Streitberger K, Eichenberger U, Nuesch E, Juni P, Arendt-Nielsen L, Curatolo M. The prevalence of widespread central hypersensitivity in chronic pain patients. Eur J Pain 2013;17: 1502–10. [21] Sterling M, Hodkinson E, Pettiford C, Souvlis T, Curatolo M. Psychologic factors are related to some sensory pain thresholds but not nociceptive flexion reflex threshold in chronic whiplash. Clin J Pain 2008;24:124–30. [22] Woolf CJ. Central sensitization: implications for the diagnosis and treatment of pain. PAIN 2011;152:S2–15.
Copyright Ó 2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
