Pain 140 (2008) 420428 www.elsevier.

com/locate/pain

Habituation and sensitization to heat and cold pain in women with bromyalgia and healthy controls
Bruce W. Smith a,*, Erin M. Tooley a, Erica Q. Montague a, Amanda E. Robinson a, Cynthia J. Cosper a, Paul G. Mullins b
a

Department of Psychology, University of New Mexico, MSC03 2220, Albuquerque, NM 87131, USA b School of Psychology, Bangor University, Addeliad Brigantia, Bangor, LL57 2AS, UK Received 6 June 2008; received in revised form 6 September 2008; accepted 16 September 2008

Abstract The purpose of this study was to examine dierences in habituation to heat and cold pain in women with bromyalgia (FM; n = 33) and in women who were healthy controls (HC; n = 44). Quantitative sensory testing (QST) was used to assess pain thresholds during ve consecutive trials of ascending heat and descending cold stimulation. Anxiety, depression, fatigue, and pain during the previous week were assessed using self-report measures. The overall hypotheses were that there would be dierences between groups in pain thresholds and in the rate of habituation to heat and cold pain stimuli. Multilevel modeling was used to test the hypotheses. There were large overall dierences in pain thresholds, with the FM group showing greater sensitivity to heat and cold pain stimuli compared with the HC group. While habituation occurred in both of the groups for heat pain, the HC group had stronger habituation across trials than the FM group. Conversely, while the HC group habituated to cold pain stimuli, the FM group showed sensitization and had decreased cold pain thresholds across trials (they felt cold pain at higher temperatures). In addition, anxiety, depression, fatigue, and pain were related to decreased heat and cold pain thresholds in the overall sample. However, when group was controlled, none of these variables were related to thresholds or rates of habituation or sensitization. The dierences between women with FM and healthy women in habituation and sensitization may have important implications for the etiology, diagnosis, and treatment of FM and other chronic pain conditions. 2008 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.
Keywords: Fibromyalgia; Habituation; Sensitization; Heat pain; Cold pain

1. Introduction Fibromyalgia (FM) is a chronic pain condition characterized by widespread muscle pain and also by associated fatigue, sleep and cognitive diculties, and depressed mood. The prevalence of FM is estimated to be 2% to 4% and is diagnosed 67 times more often in women than in men [15,38]. The diagnosis of FM is based on decreased pressure pain thresholds at a miniCorresponding author. Tel.: +1 505 277 0643; fax: +1 505 277 1394. E-mail address: bwsmith@unm.edu (B.W. Smith).
*

mum of 11 of 18 specic tender points as dened by the American College of Rheumatology criteria [39]. While the etiology of FM remains largely unknown, there is increasing evidence that FM may involve increased sensitivity to pain mediated by the central nervous system [5,12,15,17]. People with FM experience increased pain sensitivity throughout the body and not just tender points where everyone is more sensitive to pain [24,31,39]. In addition, increased pain sensitivity in FM is not limited to mechanical stimuli, but also includes electrical, heat, and cold stimuli [2,5,11]. Researchers have attempted to increase our understanding of FM by using a variety of experimental

0304-3959/$34.00 2008 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.pain.2008.09.018

B.W. Smith et al. / Pain 140 (2008) 420428

421

models that may reect enhanced central sensitization to pain in FM. In addition to increased sensitivity to many types of pain throughout the body, there is also evidence that FM is characterized by increased temporal summation of painful stimuli [3234]. Temporal summation is when the sensation of pain increases with the repetition (0.31.0 Hz) of brief painful stimuli [13,35]. It is related to windup, in which the dorsal horn neurons respond to repetitive C-ber stimulation [7,27,35]. While temporal summation may involve increased sensitization to painful stimuli in FM, there may also be dierences between people with FM and healthy individuals in the rates of adaptation or habituation to painful stimuli. While sensitization reects an increase, habituation involves a decrease in response to a repeated stimulus [26]. The processes underlying sensitization and habituation may dier in important ways [19], and the inability to habituate to pain may play a role in the development of chronic pain in FM and other painful conditions [9,36]. The purpose of this study was to examine dierences between women with FM and healthy women in habituation to heat and cold stimulation. To accomplish this, we examined changes in the temperatures at which women reported a stimulus as painful during consecutive trials. We examined pain thresholds because they may be more aected in FM than pain tolerance [10] and be related to the tendency to experience pain at low levels of stimulation in chronic pain conditions. In addition, we used longer intervals between painful stimulation than in temporal summation studies to increase the likelihood of evoking habituation rather than sensitization processes [14,19]. Our main hypotheses were that women with FM would show decreased habituation to heat and cold pain stimuli relative to healthy women. As has been found in previous studies [2,5,11], we also expected the FM group to show greater sensitivity to both heat and cold pain stimuli. Finally, we also assessed anxiety, depression, fatigue, and pain to determine whether they were related to pain thresholds or rates of habituation. 2. Methods 2.1. Participants The sample consisted of 77 women who had bromyalgia (FM; n = 33) or were healthy controls (HC; n = 44). The study was conducted in compliance with the Institutional Review Board of the University of New Mexico and informed consent was obtained by trained research assistants. The participants were recruited through newspaper ads and through rheumatology clinics. Only women between the ages of 30 and 60 were recruited for the study. All participants were paid $20 for completing the study.

The women in the FM group had their FM diagnosis conrmed by a physician, had an average pain rating for the past month of at least 20 on a 0100 scale (0 = no pain and 100 = pain as bad at it can be), had pain in all four quadrants (e.g. both arms and legs), and had no other conditions that caused greater pain. The women in the HC group had average pain ratings of less than 20 on a 0100 scale, had no pain-related diagnoses (e.g., FM, osteoarthritis, rheumatoid arthritis), and reported that they were not currently depressed or taking antidepressant medications. The bromyalgia patients were not excluded if depressed or taking antidepressant medications. There were no dierences between the FM and HC groups in age, percent married, or percent Caucasian or other ethnicity. The participants were between the ages of 34 and 59 (M = 48.00, SD = 6.81). The largest percentage were married (49%), while fewer were divorced (33%), never married (16%), or widowed (2%). The majority were Caucasian (66%), while fewer were Hispanic (23%), AsianAmerican (1%), or mixed and others (10%). There were dierences between the FM and HC groups in employment status, education, and income. The FM group also had a lower mean income range than HC group ($40,000-$49,999 vs. $60,000-$69,999, t(75) = 2.394, p < .05). The FM group also had a lower percent of college graduates (29% vs. 69%, v2(1) = 10.89, p < .001) and percent employed than the HC group (48% vs. 88%, v2(1) = 13.28, p < .001). 2.2. Procedure Participants lled out questionnaires that included the measures listed below and were tested for heat and cold pain thresholds using the Contact Heat-Evoked Potential Stimulator (CHEPS; Medoc, Ramat-Yishai, Israel). The CHEPS generates heat or cold sensations using a 3 3-cm contact thermode. This thermode comprises two layers: (1) an external layer consisting of a heating foil that has two thermocouples (electronic thermal sensors) and (2) a lower layer which is a Peltier element with one thermistor (electronic thermal sensors) and one water thermistor. The thermode was placed on the thenar (eshy area of the palm at the base of the thumb) of the right hand. The thenar was selected because it is an easily standardized location with good sensitivity and reproducibility [3]. For the heat pain task, the temperature of the thermode began at 32 C and increased by 0.5 C/s until the participant indicated that the heat sensation became painful. For the cold pain task, the temperature of the thermode began at 32 C and decreased by 0.5 C/s until the participant said that the cold sensation became painful. When the participant indicated that the temperature became painful, the researcher pressed a button

422

B.W. Smith et al. / Pain 140 (2008) 420428

returning the temperature to 32 C in less than a second (40 C/s). There were ve consecutive trials for each task and there was a 30 s interval between each trial. The threshold was assessed once during each trial. Since the rate of temperature change was 0.5 C/s and the starting temperature was 32 C, the length of each trial in seconds was the dierence between 32 and the threshold temperature multiplied by two. The order of presentation of the tasks (e.g., heat rst, cold second or cold rst, heat second) was counterbalanced across participants and there were no order eects in any of the analyses. These analyses involved an order variable for whether participants were given heat or cold rst (heat rst = 1, cold rst = 2). To determine whether there were order eects, we entered this order variable as a predictor of heat and cold pain thresholds in the multilevel analyses described below. There were no main eects for order and no interactions for order with any of the other variables in predicting either heat or cold pain thresholds. 2.3. Measures 2.3.1. Anxiety and depression symptoms Anxiety and depression symptoms were assessed using 18 items from the Mental Health Inventory [37]. This inventory was used because it does not contain items that are likely to be confounded with having chronic pain and it has been well-validated [40]. There were nine anxiety items such as how much of the time have you been a nervous person? and nine depression items such as how much of the time have you felt downhearted and blue? Participants were asked to respond to each of these items with regard to the previous week. The responses were scored on a six-point scale for all items except for the last item on the depression subscale which was scored on a ve-point scale. Cronbachs alpha was .939 for anxiety and .949 for depression. 2.3.2. Fatigue The Fatigue Severity Scale [20] included nine statements (e.g., I am easily fatigued) to assess how much fatigue the participants had experienced and how much it aected their ability to function. The responses were scored on a seven-point scale where 1 = strongly disagree to 7 = strongly disagree. Cronbachs alpha for the nine items was .966. 2.3.3. Pain Pain was assessed with three items using visual analogue scales. The items included: (1) indicate the level of pain you are currently experiencing, (2) indicate the average level of pain you have experienced in the previous week, and (3) indicate the worst pain you have experienced over the previous week. For each of

these items, 0 = no pain and 100 = pain as bad as it could be. We used these items because visual analogue scales have frequently been used with FM patients and these three items have formed a reliable scale [30]. Cronbachs alpha for the three items was .968. 2.4. Data analytic methods Multilevel modeling was used to analyze the repeated measures data and test our predictions. This method is useful for the analysis of data that have a nested hierarchical structure. The heat and cold pain threshold data took a hierarchical form, with ve repeated observations nested within each of the 77 participants. The SPSS 16.0 Mixed program was used for the multilevel analyses, and the specications of the models were based on the guidelines provided by Singer [29] and Peugh and Enders [25]. The heat and cold pain threshold measures were the criterion variables to be predicted in the analyses. There were two basic types of prediction equations in the multilevel analyses: a Level 1 equation, which examined the inuence of within-person variations in pain thresholds, and a Level 2 equation, which tested the eects of group (1 = HC, 2 = FM) and the other individual dierence variables (e.g., anxiety, depression, fatigue, and pain over the past week). In essence, the Level 2 equations address questions regarding between-person dierences and take the following form: Do people who have different scores on a between-persons predictor (e.g., group) have dierent levels on the criterion (e.g., heat and cold pain thresholds)? Level 1 questions address the issue of when rather than who: for example, When a person has the highest threshold, is it on an earlier or later trial? The Level 1 question examined within-person variation with regard to the trial in which the pain threshold was assessed. To prepare for these analyses, the trial number was centered on each participant by subtracting the average trial number for each participant from each trial number. The Level 1 equation was specied as follows: Level 1 : pain thresholdij b0j b1j trial numberij rij where pain thresholdij is the pain threshold for participant j at trial i, b0j is the mean pain threshold for participant j, b1j is the slope of the trial number predicting the pain threshold for participant j, trial numberij is the trial number of participant j at trial i, and rij is the random component of the pain threshold for participant j at time i. Individual dierences in the average level of the pain thresholds were probed through analyses at Level 2. The Level 2 question addresses who rather than when: for example, Who has the greatest increase in heat pain thresholds? The individual dierence variables were used as predictors of variance in the Level 1 pain

B.W. Smith et al. / Pain 140 (2008) 420428

423

thresholds. An example of the initial Level 2 equation for this model is as follows: Level 2 : b0j c00 c01 group u0j where b0j is the mean pain threshold for participant j, c00 is the grand mean pain threshold across all participants, c01 is the slope predicting the mean pain threshold for participant j from whether the participant is in the HC group or the FM group, and u0j is the random component of the mean pain threshold for participant j. Additional Level 2 predictors are not shown but were also examined by adding them to the equation (e.g., c02 = pain, c03 = fatigue, c04 = anxiety, c05 = depression). Cross-level interactions were also probed by examining the eect of individual dierence variables on the relationship between trial number and pain threshold. The Level 2 equation for modeling these interactions was as follows: Level 2 : bij c10 c11 group uij where b1j is the slope of the trial number predicting the pain threshold for participant j, c10 is the grand mean for the slope of trial number predicting the pain threshold across all participants, c11 indicates how the relationship between pain threshold and trial number for participant j changes depending on whether the participant is in the HC or the FM group, and u1j is the random component of the slope of the trial number predicting the pain threshold for participant j. Thus, c11 provides a test of our prediction that the rate of change across trials would vary by group. Adding the other Level 2 variables provided for a test of the interactions between these other variables and the trial number in predicting the pain thresholds. The other specications for the multilevel model were selected following Singer (1998) to identify the best tting model of the variables under study. The intercept and variables that initially showed signicant random eects were kept in the model and modeled using an unstructured covariance matrix. Finally, eect sizes were computed for all the predictors in our hypotheses. According to the guidelines outlined in Singer [29], eect sizes in multilevel modeling can be derived with the variancecovariance parameter estimates. By using these estimates for the intercept, we determined the proportion of explainable between-subjects variance accounted for when the individual dierence variables (e.g., group, anxiety, depression, etc.) were added to the model and the within-subjects variance accounted for when the within-subject variables (e.g., trial number and crosslevel interactions) were added to the model. For the purposes of this study, this parameter estimate of the proportion of explainable variance accounted for when

Table 1 Descriptive statistics and group comparisons for the health variables FM group Mean Pain Fatigue Anxiety Depression 62.68 5.76 3.20 2.77 SD 20.19 1.41 1.21 1.21 HC group Mean 7.56 2.15 1.82 1.40 SD 11.40 0.86 0.61 0.44 13.882 12.849 5.988 6.236 .000 .000 .000 .000 3.36 3.09 1.44 1.50 t p d

Note. df = 75.

a variable is added to the model will be referred to as variance explained (VE). 3. Results 3.1. Descriptive statistics Table 1 displays the means and standard deviations for the FM and HC groups for the health variables (e.g., pain, fatigue, anxiety, depression). There were large and signicant dierences (p < .05) between the groups on all the variables (Cohens d P 1.44). The FM group was higher on pain, fatigue, anxiety, and depression than the HC group. The largest dierences were in pain and fatigue (d = 3.36 and 3.09, respectively) with smaller dierences in anxiety and depression (d = 1.44 and 1.50, respectively). 3.2. Zero-order correlations Table 2 displays the correlations between the mean heat and cold pain thresholds and the health variables. The correlations are presented for the FM and HC groups combined. The heat and cold pain thresholds were strongly correlated with each other (r = .817). The heat pain threshold was moderately negatively related to pain, fatigue, anxiety, and depression. Conversely, the cold pain threshold was moderately positively related to pain, fatigue, anxiety, and depression. Pain, fatigue, anxiety, and depression were all strongly positively related.

Table 2 Correlations between the main study variables for the combined FM and HC groups 1 1. 2. 3. 4. 5. 6.
a * **

2 .323** .368** .340** .354**

Heat threshold Cold thresholda Pain Fatigue Anxiety Depression

.817** .258* .368** .329** .391**

.871** .639** .636**

.674** .694**

.831**

Mean of ve trials. p < .05. p < .01.

424

B.W. Smith et al. / Pain 140 (2008) 420428

Table 3 Descriptive statistics and group comparisons for the heat and cold pain threshold trials FM group Mean Heat pain threshold Trial 1 40.78 Trial 2 41.48 Trial 3 42.22 Trial 4 42.37 Trial 5 42.68 All Trials 41.90 Cold pain threshold Trial 1 13.72 Trial 2 14.52 Trial 3 14.51 Trial 4 14.46 Trial 5 15.09 All Trials 14.46 SD 3.38 3.81 3.81 4.17 4.27 3.78 8.74 9.02 8.93 9.26 9.58 8.94 HC group Mean 43.59 44.82 45.45 45.77 46.17 45.09 7.64 7.30 6.57 6.92 6.01 6.89 SD 2.74 2.79 2.76 2.87 2.71 2.72 6.28 6.25 6.05 7.43 6.28 6.16 3.951 4.326 4.207 3.950 4.031 4.037 3.331 3.874 4.332 3.887 4.652 4.106 .000 .000 .000 .000 .000 .000 .002 .000 .000 .000 .000 .000 0.91 1.06 0.97 0.95 0.98 0.97 0.80 0.93 1.04 0.90 1.12 0.99 t p d

3.3. Decreased habituation to heat pain stimuli in FM The mean heat pain threshold for all participants was 43.67 C (SD = 3.54). Table 3 displays the means and standard deviations for the FM and HC groups for heat threshold pain ratings for each trial and averaged across all trials. As expected, there were large and signicant dierences between the groups for each trial and for the mean of all trials. The Cohens d eect size of the difference was between .91 and 1.06 for the individual trials and .97 for the mean of all trials. Thus, our prediction regarding lower heat pain thresholds (registering pain at lower temperatures) for the FM group relative to the HC group was supported. Next, we used multilevel analyses to examine the individual eects for group and the health variables. First, we entered group and each of the health variables into the equations separately to determine whether they pre-

Table 4 Multilevel analyses with univariate predictors of heat and cold pain thresholds Estimate Heat pain threshold Group 3.1917 Pain .0294 Fatigue .6077 Anxiety 1.0230 Depression 1.2683 Cold pain threshold Group 7.5719 Pain 0.0850 Fatigue 1.4403 Anxiety 4.9399 Depression 4.1814 SE .7543 .0131 .1826 .3488 .3542 1.7563 0.0298 .4317 1.4667 1.4540 df 71.926 71.989 72.009 71.922 71.936 72.000 70.000 71.000 72.000 72.000 t 4.231 2.245 3.328 2.933 3.580 4.311 2.856 3.336 3.368 2.876 p .000 .028 .001 .005 .001 .000 .006 .001 .001 .005 VE 19.39 5.55 13.74 9.81 14.51 19.68 9.28 13.15 10.49 11.44

dicted heat pain thresholds. Table 4 displays the results for the multilevel analyses with group and the health variables as univariate predictors of heat pain thresholds. Group was negatively related to heat pain threshold indicating that the FM group had lower heat pain thresholds (they felt heat pain at lower temperatures; VE = 19.39%). In addition, pain, fatigue, anxiety, and depression were all negatively related to heat pain thresholds accounting for less variance than group (VEs = 5.5514.51%). In addition, we entered each of the demographic variables separately and none of them predicted heat pain thresholds. Next, we specied a multilevel equation predicting heat pain threshold from group, the trial number, the group trial number interaction, and separately added each of the health variables. When group, trial, and group trial interaction were in the equation, none of the health variables still predicted heat pain threshold or interacted with trial to predict heat pain threshold. Group, trial number, the group trial interaction all predicted the cold pain thresholds when controlling for each of the health variables. The nal model is displayed in Table 5. The group trial interaction was signicant (VE = 3.92%). This indicates that the HC group had greater increases in heat pain thresholds (higher temperatures were required to evoke pain) across the trials as compared with the FM group. The mean increase across the ve trials was 1.90 C (from 40.78 to 42.68 C) for the FM group and 2.58 C (from 43.59 to 46.17 C) for the HC group. Thus, the hypothesis that the FM group would show smaller increases in heat pain (less habituation) thresholds relative to the HC group was supported. Finally, we repeated the analysis controlling for trial length. On average, the HC group had longer trials than the FM group (26.2 s vs. 19.8 s) because they had higher heat pain thresholds and it took more time to reach higher thresholds. However, controlling for trial length revealed very little dierence in the group x trial interaction (b = .2638, t = .3.936, p = .000) indicating that
Table 5 Multilevel analyses with multivariate predictors of heat and cold pain thresholds Estimate Heat pain thresholds Group 2.6601 Trial .5305 Group Trial .2666 Cold pain thresholds Group 7.5719 Trial .0940 Group Trial .6311 SE .8554 .0336 .0753 1.7563 .0778 .1570 df 68.891 281.236 281.047 72.000 294.000 294.000 t 3.110 15.782 3.542 4.311 1.209 4.019 p .000 .000 .000 .000 .228 .000 VE 11.19 45.48 3.92 19.68 1.08 4.89

Note. VE, variance explained and refers to proportion of explainable variance accounted for when a variable is added to the model.

Note. VE, variance explained and refers to proportion of explainable variance accounted for when a variable is added to the model.

B.W. Smith et al. / Pain 140 (2008) 420428

425

trial length did not account for group dierences in habituation. 3.4. Increased sensitization rather than habituation to cold pain stimuli in FM The mean cold pain threshold for all participants was 10.25 C (SD = 8.27). Table 3 displays the means and standard deviations for the FM and HC groups for cold pain threshold ratings for each trial and averaged across all trials. As with heat pain thresholds, there were large and signicant dierences between the groups for each trial and for the mean of all trials. The eect size of the dierences was between .80 and 1.12 for the trials and .99 overall. Thus, our prediction regarding higher cold pain thresholds (registering cold pain at higher temperatures) for the FM group relative to the HC group was supported. Next, we used multilevel analyses to examine the individual eects for group and the health variables. First, we entered group and each of the health measures into the equation separately to determine whether they predicted cold pain thresholds. Table 4 displays the results for the multilevel analyses with group and the health measures as univariate predictors of cold pain thresholds. Group was positively related to cold pain threshold indicating that the FM group had lower pain thresholds (they felt cold pain at higher temperatures; VE = 19.68%). In addition, pain, fatigue, anxiety, and depression were all positively related to cold pain thresholds accounting for less variance than group (VEs = 9.2813.15%). In addition, we entered each of the demographic variables separately and none of them predicted cold pain thresholds. Next, we specied a multilevel equation predicting cold pain threshold from group, trial number, the group trial number interaction, and separately added each of the other health variables. When group, trial, and group trial interaction were in the equation, none of the other health variables still predicted cold pain thresholds or interacted with trial to predict cold pain threshold. Group, trial number, the group trial interaction all predicted the cold pain thresholds when controlling for each of the health variables. The nal model is displayed in Table 5. The group trial interaction was signicant (VE = 4.89%). This interaction indicates that there was a signicant difference in the change across the trials between the FM and the HC groups. There was a mean decrease across the ve trials of 1.63 C (from 7.64 to 6.01 C) for the HC group and a mean increase of 1.37 C (from 13.72 to 15.09 C) for the FM group. Thus, the hypothesis regarding the changes in cold pain thresholds was only partially supported in that there was a group dierence in the predicted direction. However, the HC showed

habituation, while the FM group showed sensitization across the trials. Finally, we repeated the analysis controlling for trial length. On average, the HC group had longer trials than the FM group (50.2 s vs. 35.1 s) because they had lower cold pain thresholds and it took more time to reach lower thresholds. However, controlling for trial length revealed very little dierence in the group trial interaction (b = .6563, t = 4.691, p = .000) indicating that trial length did not account for group dierences in habituation and sensitization. 4. Discussion The purpose of this study was to examine habituation to heat and cold pain in women with FM and in healthy controls. As predicted, we found that the FM group had decreased rates of habituation to heat pain stimulation relative to the HC group. However, while we found the expected dierences between the FM and HC groups in the rates of change across cold pain trials, the HC group demonstrated habituation while the FM group became more sensitized. In addition, we examined group dierences in mean heat and cold pain thresholds. As expected, there also were large overall dierences in pain thresholds, with the FM group showing greater sensitivity to heat and cold pain stimuli compared with the HC group. Finally, while pain, fatigue, anxiety, and depression were univariate predictors of heat and cold pain thresholds, none of them were related to thresholds or rates of habituation or sensitization when group was controlled. The examination of changes in response to potentially painful stimulation across time may be important for understanding the development and maintenance of chronic pain. Whether there is an increased or decreased response may depend on the length of the time interval between stimulation [19]. When the interval is brief (e.g., seconds), such as with temporal summation, there may be an increased response or sensitization [7,13]. When the interval is longer (e.g., minutes, hours, or days), there may be a decreased response or habituation [4,36]. Since most studies have examined the eects of repetitions with brief intervals in temporal summation in FM, we wanted to determine whether longer intervals would evoke habituation. We also wanted to see whether there would be dierences in habituation between people with FM and healthy individuals that parallel dierences in sensitization found in temporal summation studies [3234]. We found that habituation did occur in response to both heat and cold stimulation in healthy women. In contrast, there was less pronounced habituation to heat and sensitization to cold stimulation in women with FM. These results suggest that FM may be characterized by decits in habituation processes as well as enhanced

426

B.W. Smith et al. / Pain 140 (2008) 420428

sensitization. This is important because the mechanisms underlying habituation may be dierent from sensitization. Characterizing these dierences may increase our understanding of why people with FM may not adapt as well following painful injuries and ares. The sensitization that occurs in temporal summation has been linked to wind-up in which the dorsal horn neurons respond to repetitive C-ber stimulation [7,27,35]. This type of sensitization may also be related to age and gender and psychological variables such as anxiety and catastrophizing [6,8,13,21,28]. In contrast, habituation may not simply be the ip side of sensitization, and may involve dierent mechanisms and reect dierent explanations for chronic pain in FM. Bingel et al. [4] found neural activation associated with habituation in regions of the brain that have been associated with placebo analgesia and the attentional modulation of pain. Habituation may be supported by higher brain regions whose function may be aected by behavioral as well as pharmacological interventions [1]. The eectiveness of acceptance-based and values-based approaches to chronic pain [2223] may rely on habituation processes that support accepting and moving towards valued goals despite current pain. Moreover, the ability to habituate to injuries and daily aches and pains may be a protective factor against the development of chronic pain. Why did we nd that FM was associated with sensitization to cold stimulation rather than habituation? Although there may be no denite answers at this point, initial clues may be provided by studies that have found dierences in the response of FM patients to heat and cold pain. As in the current study, Desmueles et al. [5] found similar larger mean dierences in cold as compared with heat pain thresholds in people with FM relative to healthy controls. Hurtig et al. [18] examined both heat and cold pain thresholds in FM and found that cold pain thresholds were especially linked to pain intensity, tender points, and sleep quality. They suggested that the dierences may be due to the role of cutaneous veins in mediating cold pain. Berglund et al., [2] found that paradoxical heat sensations were reported for cold stimulation in FM patients. They suggested that greater mean dierences for cold pain thresholds may be due to hypervigilance eects and that the paradoxical sensations may be due to dysfunction in the limbic cortex. Finally, although pain, fatigue, anxiety, and depression over the previous week were related to mean heat and pain thresholds in univariate models, none of these variables were related to thresholds or rates of habituation or sensitization when group was controlled. This suggests that the dierences between the FM and HC groups cannot be simply be explained by or reduced to dierences in recently experienced pain, fatigue, and distress. These results are in contrast to previous ndings showing that anxiety was related to temporal summa-

tion [13]. It may be that a more specic psychological measure such as pain acceptance or catastrophizing would have been independently related to temporal change in our experimental model [6,13]. While we believe that the primary value in this study was to identify dierences in both habituation and sensitization in FM patients, it points to several important next steps for understanding the role of temporal pain processes in FM. First, studies could systematically vary the time intervals to determine when habituation and sensitization occur and see how these times courses vary for people with FM and other chronic pain disorders. Second, experimental measures of habituation could be linked to the likelihood of developing chronic pain conditions and adapting to them. Ideally, researchers could assess habituation and do long-term prospective studies with patients and controls. For example, we plan to follow our participants to determine whether habituation predicts who develops chronic pain in the HC group and who adapts better over time to their condition in the FM group. Third, additional eorts could be made to dierentiate the mechanisms that underlie sensitization and habituation to pain and understand their role in the development of or protection against chronic pain. Functional imaging could be used to better understand the role of the brain and dierentiate it from other central and peripheral mechanisms [4]. Opioid antagonists could be used to determine the extent to which habituation may involve the endogenous opioid system [16]. Fourth, a better understanding of sensitization and habituation might be applied in eorts to improve diagnosis and treatment. Simple tests of habituation could provide an additional diagnostic indicator and more eective treatments may be facilitated by directly targeting or capitalizing on habituation processes. For example, psychosocial interventions that encourage functioning despite pain could be tested to determine whether they enhance habituation [2223]. This study has several limitations. First, this study was conned to women between the ages of 30 and 60 because of the diculty in recruiting enough men with FM. Just as dierences in temporal summation have been found between men and women [8,28], gender differences may also be seen with longer time intervals between stimulation. Second, the pain stimuli were limited to heat and cold and to one time interval. As suggested above, it would be useful to vary the time interval and also examine additional types of pain. In addition, we only examined pain threshold and not the response to higher levels of painful stimulation. It is possible that examining pain over a fuller range of stimulation could reveal dierent patterns with regard to habituation and sensitization. Third, the placement of the thermode was limited to one location on the thenar of the hand. By using the

B.W. Smith et al. / Pain 140 (2008) 420428

427

same area, we may have combined the eects of peripheral and central habituation processes. Moving the thermode to dierent places on the hand or other body surface could have reduced the eects of peripheral habituation mechanisms. However, thermal sensitivities have been shown to vary among body parts [3]. Thus, varying the position of the thermode could confound habituation dierences with dierences due to variations in thermal sensitivities. Future studies of habituation could increase the likelihood of distinguishing between central and peripheral mechanisms by using a larger surface such as the volar side of the forearm and comparing the eects of xed and variable locations [14]. 5. Conclusions In conclusion, this study identied dierences between women with FM and healthy women in habituation and sensitization to heat and cold pain stimulation. Women with FM showed decreased habituation to heat pain stimuli as compared with healthy women. Most striking, women with FM became sensitized while healthy women became habituated to cold pain stimuli. In addition, our results conrmed previous ndings showing large dierences in thermal pain thresholds, especially with cold pain. Also, self-report pain, fatigue, and distress were not related to thresholds or rates of habituation or sensitization when controlling for group. Overall, this study suggests that there may be important dierences between women with FM and healthy women in temporal pain processes and that habituation may be worthy of further study.

Acknowledgements This research was supported by a Cross Campus Collaboration grant (#765110, Bruce W. Smith, Principal Investigator, Paul G.M. Mullins, Co-Principal Investigator) and a Research Allocation Committee grant (#06-17, Bruce W. Smith, Principal Investigator) from the University of New Mexico, Albuquerque, New Mexico. None of the authors of this manuscript have any nancial or other relationships that might lead to a conict of interest. We thank Wilmer Sibbitt, M.D., for his help in recruiting bromyalgia patients.

References
[1] Becerra LR, Breiter HC, Stojanovic M, Fishman S, Edwards A, Comite AR. Human brain activation under controlled thermal stimulation and habituation to noxious heat: an fMRI study. Magn Reson Med 1999;41:104457. [2] Berglund B, Harju E, Kosek E, Lindblom U. Quantitative and qualitative analysis of cold dysesthesia and hyperalgesia in bromyalgia. Pain 2002;96:17787.

[3] Berglund B, Harju E. Master scaling of perceived intensity of touch, cold, and warmth. Eur J Pain 2003;7:32334. [4] Bingel U, Schoell E, Herken W, Buchel C, May A. Habituation to painful stimulation involves the antinociceptive system. Pain 2007;131:2130. [5] Desmeules JA, Cedraschi C, Rapiti E, Baumgartner E, Finckh A, Cohen P. Neurophysiologic evidence for a central sensitization in patients with bomyalgia. Arthritis Rheum 2003;48:14209. [6] Edwards RR, Smith MT, Stonerock G, Haythornthwaite JA. Pain-related catastrophizing in healthy women is associated with greater temporal summation of and reduced habituation to thermal pain. Clin J Pain 2006;22:7307. [7] Eide PK. Wind-up and the NMDA receptor complex from a clinical perspective. Eur J Pain 2000;4:517. [8] Fillingim RB, Maixner W, Kincaid S, Silva S. Sex dierences in temporal summation but not sensory-discriminative processing of thermal pain. Pain 1998;17:3344. [9] Flor H, Diers M, Birbaumer N. Peripheral and electrocortical responses to painful and non-painful stimulation in chronic pain patients, tension headache patients and healthy controls. Neurosci Lett 2004;361:14750. [10] Geissner ME, Casey KL, Brucksch CB, Ribbens CM, Appleton BB, Croord LJ. Perception of noxious and innocuous heat stimulation among health women and women with bromyalgia: association with mood, somatic focus, and catastrophizing. Pain 2003;102:24350. [11] Gracely RH, Grant MAB, Giesecke T. Evoked pain measures in bromyalgia. Best Pract Res Clin Rheumatol 2003;17:593609. [12] Gracely RH, Petzke F, Wolf JM, Clauw DJ. Functional magnetic resonance imaging evidence of augmented pain processing in bromyalgia. Arthritis Rheum 2002;46:133343. [13] Granot M, Granovsky Y, Sprecher E, Nir R, Yarnitsky D. Contact heat-evoked temporal summation: tonic versus repetitivephasic stimulation. Pain 2006;122:295305. [14] Grerath W, Baumgartner U, Treede R. Peripheral and central components of habituation of heat pain perception and evoked potentials in humans. Pain 2007;132:30111. [15] Jacobsen S, Bredkjaer SR. The prevalence of bromyalgia and widespread chronic musculoskeletal pain in the general population (letter). Scand J Rheumatol 1992;21:2613. [16] Janicki P, Libich J, Gumulka W. Lack of habituation of pain evoked potentials after naloxone. Pol J Pharmacol Pharm 1979;31:2015. [17] Henriksson KG. Is bromyalgia a central pain state? J Musculoskelet Pain 2002;10:4557. [18] Hurtig IM, Raak RI, Kendal SA, Gerdle B, Wahren LK. Quantitative sensory testing in bromyalgia patients and in healthy subjects: identication of subgroups. Clin J Pain 2001;17:31622. [19] Kleinbohl D, Trojan J, Konrad C, Holzl R. Sensitization and habituation of AMH and C-ber related percepts of repetitive radiant heat stimulation. Clin Neurophysiol 2006;117:11830. [20] Krupp LB, LaRocca NG, Muir-Nash J, Steinberg AD. The fatigue severity scale. Application to patients with multiple sclerosis and systemic lupus erythematosus. Arch Neurol 1989;46:11213. [21] Lautenbacher S, Kunz M, Strate P, Nielsen J, Arendt-Nielsen L. Age eects of pain thresholds, temporal summation and spatial summation of heat and pressure pain. Pain 2005;115:4108. [22] McCracken LM, Eccleston C. A prospective study of acceptance of pain and patient functioning with chronic pain. Pain 2005;118:1649. [23] McCracken LM, Vowles KE. A prospective analysis of acceptance of pain and values-based action in patients with chronic pain. Health Psychol 2008;27:21520.

428

B.W. Smith et al. / Pain 140 (2008) 420428 [33] Staud R, Cannon RC, Mauderli AP, Robinson ME, Price DD, Vierck CJ. Temporal summation of pain from mechanical stimulation of muscle tissue in normal controls and subjects with bromyalgia syndrome. Pain 2003;102:8795. [34] Staud R, Robinson ME, Price DD. Temporal summation of second pain and its maintenance are useful for characterizing widespread central sensitization of bromyalgia patients. J Pain 2007;8:893901. [35] Torebjork HE, Hallin RG. Identication of aerent C units in intact skin nerves. Brain Res 1974;67:387403. [36] Valeriani M, de Tommaso M, Restuccia D, Le Pera D, Guido M, Iannetti GD. Reduced habituation to experimental pain in migraine patients: a CO2 laser evoked potential study. Pain 2003;105:5764. [37] Veit CT, Ware JE. The structure of psychological distress and well-being in general populations. J Consult Clin Psychol 1983;51:73042. [38] Wolfe F, Ross K, Anderson J, Russell IJ, Hebert L. The prevalence and characteristics of bromyalgia in the general population. Arthritis Rheum 1995;38:1928. [39] Wolfe FW, Smythe HA, Yunas MB, Bennett RM, Bombardier C, Goldenberg DL. The American college of rheumatology 1990 criteria for the classication of bromyalgia. Arthritis Rheum 1990;33:16072. [40] Zautra AJ, Guarnaccia C, Reich JW. Factor structure of mental health measures for older adults. J Consult Clin Psychol 1988;56:5149.

[24] Petzke F, Khine A, Williams D, Groner K, Clauw DJ, Gracely RH. Dolorimetry performed at 3 paired tender points highly predicts overall tenderness. J Rheumatol 2003;30:56774. [25] Peugh JL, Enders CK. Using the SPSS Mixed Procedure to t cross-sectional and longitudinal multilevel models. Educ Psychol Meas 2005;65:71741. [26] Prescott SA. Interactions between depression and facilitation within neural networks: updating the dual-process theory of plasticity. Learn Mem 1998;5:44666. [27] Price DD. Psychological mechanisms of pain and analgesia: progress in pain research and management. Seattle: IASP Press; 1999. pp. 1248. [28] Sarlani E, Greenspan JD. Gender dierences in temporal summation of mechanically evoked pain. Pain 2002;97:1639. [29] Singer JD. Using SAS PROC MIXED to t multilevel models, hierarchical models, and individual growth models. J Educ Beh Stat 1998;24:32355. [30] Smith BW, Zautra AJ. The eects of anxiety and depression on weekly pain in women with arthritis. Pain 2008;138:35461. [31] Sorensen J, Graven-Nielsen T, Henriksson KG, Bengtsson M, Arendt-Nielson L. Hyperexcitability in bromyalgia. J Rheumatol 1998;25:1525. [32] Staud R, Vierck CJ, Cannon RL, Mauderli AP, Price DD. Abnormal sensitization and temporal summation of second pain (wind-up) in patients with bromyalgia syndrome. Pain 2001;91:16575.