artigo de dor journal of pain

8/3/2019 Artigo de Dor Journal of Pain

1/14

Pain Sensitivity Risk Factors for Chronic TMD: Descriptive

Data and Empirically Identified Domains from the OPPERA

Case Control Study

Joel D. Greenspan,* Gary D. Slade,y,z,x Eric Bair,x,{,# Ronald Dubner,* Roger B. Fillingim,

Richard Ohrbach, Charlie Knott,** Flora Mulkey,{ Rebecca Rothwell,x

and William Maixnerx,#

*Department of Neural and Pain Sciences, and Brotman Facial Pain Center, University of Maryland Dental School,Baltimore, Maryland.yDepartment of Dental Ecology, zDepartment of Epidemiology, xCenter for Neurosensory Disorders, {Department of

Biostatistics, and#Department of Endodontics, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina.Department of Community Dentistry & Behavioral Science, University of Florida, Gainesville, Florida.

Department of Oral Diagnostic Services, University at Buffalo, Buffalo, New York.**Battelle Memorial Institute, Durham, North Carolina.

Abstract: Many studies report that people with temporomandibular disorders (TMD) are more sensitiveto experimental pain stimuli than TMD-free controls. Such differences in sensitivity are observed in re-

mote body sites as well as in the orofacial region, suggesting a generalized upregulation of nociceptive

processing in TMD cases. This large case-control study of 185 adults with TMD and 1,633 TMD-free con-

trols measured sensitivity to painful pressure, mechanical cutaneous, and heat stimuli, using multiple

testing protocols. Based on an unprecedented 36 experimental pain measures, 28 showed statistically sig-

nificantly greater pain sensitivity in TMD cases than controls. The largest effects were seen for pressure

pain thresholds at multiple body sites and cutaneous mechanical pain threshold. The other mechanical

cutaneous pain measures and many of the heat pain measures showed significant differences, but

with lesser effect sizes. Principal component analysis (PCA) of the pain measures derived from 1,633 con-trols identified 5 components labeled: 1) heat pain ratings; 2) heat pain aftersensations and tolerance; 3)

mechanical cutaneous pain sensitivity; 4) pressure pain thresholds; and 5) heat pain temporal summation.

These results demonstrate that compared to TMD-free controls, chronic TMD cases are more sensitive to

many experimental noxious stimuli at extracranial body sites, and provide for the first time the ability to

directly compare the case-control effect sizes of a wide range of pain sensitivity measures.

Perspective: This article describes experimental pain sensitivity differences between a large sampleof people with chronic TMD and non-TMD controls, using multiple stimulus modalities and measures.

Variability in the magnitude and consistency of case-control differences highlight the need to consider

multiple testing measures to adequately assess pain processing alterations in chronic pain conditions.

2011 by the American Pain Society

Key words: QST, chronic facial pain, heat pain, pressure pain, mechanical pain.

Anumber of studies have reported chronic TMD

cases to be more sensitive than pain-free controlsto experimental pain stimuli, using a variety of

protocols and pain measures.23-25,33,38 However, suchdifferences are not observed in all studies.3,5,38 Amongstudies showing significant case-control differences,

Supported by NIH grant U01DE017018. This material was also supportedby the North Florida/South Georgia Veterans Health System, Gainesville,FL. The OPPERA program also acknowledges resources specifically pro-videdfor thisproject by the respectivehost universities: Universityat Buf-falo, University of Florida, University of Maryland-Baltimore, andUniversity of North Carolina-Chapel Hill.

Roger Fillingim and Gary Slade are consultants and equity stock holders,and William Maixner is a cofounder and equity stock holder in Algynom-ics, Inc.,a company providing research services in personalized pain med-ication and diagnostics. Richard Ohrbach, Joel Greenspan, Charles Knott,Ronald Dubner, EricBair,Flora Mulkey and Rebecca Rothwelldeclare thatthey have no conflicts of interest. Portions of these data were presented

atthe 2010Annual ScientificMeetingof theAmericanPain Society in Bal-timore, MD.Supplementary data accompanying this article are available online at

jpain.org and sciencedirect.com.Address reprint requests toJoel D. Greenspan, Department ofNeuralandPain Sciences, 8 South, Univ. of Maryland Dental School, Baltimore, MD21201. E-mail: [email protected]

1526-5900/$36.00

2011 by the American Pain Society

doi:10.1016/j.jpain.2011.08.006

T61

The Journal of Pain, Vol 12, No 11 (November), Suppl. 3, 2011: pp T61-T74

Available online atwww.sciencedirect.com
http://jpain.org/http://sciencedirect.com/mailto:[email protected]://dx.doi.org/10.1016/j.jpain.2011.08.006http://www.sciencedirect.com/http://www.sciencedirect.com/http://dx.doi.org/10.1016/j.jpain.2011.08.006mailto:[email protected]://sciencedirect.com/http://jpain.org/


2/14

chronic TMD cases often are more pain sensitive inremote body sites as well as in the orofacial region, sug-gesting a generalized upregulation of nociceptive inputprocessing in this patient population.

It is not clear which circumstances result in finding sig-nificant differences between TMD cases and controls inexperimental pain sensitivity. In previous studies, various

modes of assessing experimental pain sensitivity re-vealed differences between TMD cases and controls, in-cluding mechanical and heat stimuli, as well ascutaneous and deep tissue stimuli. Case-control differ-ences also have been observed using assessments ofthreshold, tolerance, suprathreshold ratings, and tempo-ral summation (TS) of pain.34 However, there is currentlyno indication that a particular type of stimulus or assess-ment method is more or less likely to identify case-control differences in experimental pain sensitivity.

Yet, important differences exist among various types ofstimuli and among different assessment protocols. Differ-ent modalities of experimental pain stimulation (ie, me-

chanical, thermal) recruit different groups of nociceptorsandactivateonly partially overlappingCNS pathways. Fur-thermore, different pain assessment measures (ie, thresh-olds, suprathreshold ratings, TS) engage different aspectsof pain perception and thus, at some level, different CNSprocesses.1 Hence, there is only modest correlation be-tween the same types of assessments of different stimulusmodalities, or between different types of assessments ofthe same stimulus modality.2,17,18,21 This suggests thatconditions affecting experimental pain sensitivity mayhave differential effects upon components of the painprocessing system, showing varying results depending

upon the experimental protocol and the particular

aspect of pain evaluated.The Orofacial Pain: Prospective Evaluation and Risk As-

sessment (OPPERA) study is a prospective cohort study de-signed to identify causal determinants of TMD pain. Thisstudy is based on a conceptual model in which enhancedpain sensitivity represents a critically important intermedi-ate phenotypic risk factor for development of TMD.8 Thefindings presented below describe the baseline assess-ment of pain sensitivity in a large group of TMD-free con-trols as well as a cohort of TMD cases. A wide range of

quantitative sensory testing (QST) was conducted in orderto assess multiple aspects of experimental pain sensitivity.While previous research has found that some aspects of

pain sensitivity are associated with TMD, no studies haveassessed this characteristic with the broad array of mea-sures used here. Furthermore, in addition to case-controlcomparisons of individual QST measures, we have con-ducted principal component analysis to characterize la-tent constructs of pain sensitivity.

Methods

Study Setting and ParticipantsAs described elsewhere,35 the OPPERA baseline case-

control study used advertisements, emails, flyers, andword-of-mouth to recruit people who had chronic TMD

(cases) and people who did not (controls). Theywere recruited between May 2006 and November 2008

from communities in and around academic health cen-ters at 4 US study sites: Baltimore, MD; Buffalo, NY;Chapel Hill, NC; and Gainesville, FL. At each study site,the target was to recruit 800 controls and variable num-bers of cases based on local operational requirements,for a total of 3,200 controls and 200 cases. The actualnumber enrolled was 3,263 controls and 185 cases.

The classification of TMD was based on the Research Di-agnostic Criteria for Temporomandibular Disorder.9 Insummary, cases met all 3 of the following criteria: duringthe telephone interview, 1) pain reported with sufficientfrequencyin the cheeks, jaw muscles,temples or jaw jointsduring the preceding 6 months (at least15 days in the pre-ceding month and at least 5 days per month in each of the5 months preceding that); during the examination, 2) painreported in the examiner-defined orofacial region for atleast 5 days out of the prior 30 days; and 3) pain reportedin at least3 masticatory muscles or at least 1 temporoman-dibular joint in responseto palpation of the orofacial mus-cles or maneuver of the jaw. Examiners defined the

orofacial region by touching the following anatomicalareas bilaterally: temporalis, preauricular, masseter,poste-rior mandibular, and submandibular. Controls met all 6 ofthe following criteria: during the telephone interview, 1)pain reported infrequently in the cheeks, jaw muscles,temples or jaw joints (no orofacial pain in the precedingmonth and no more than 4 days per month in any ofthe 5 months preceding that); 2) no more than 4 head-aches per month within the preceding 3 months; 3) neverdiagnosed with TMD; 4) no use of night guard occlusalsplint; and during the examination, 5) pain reported inthe examiner-defined orofacial region for no more than

4 days in the prior 30 days; 6) classified as having neither

myalgia nor arthralgia. However, controls could be posi-tive or negative with respect to pain in response to palpa-tion or jaw maneuver. Additional studywide criteria for allstudy participants were: aged 18 to 44 years; fluent in En-glish; negative responses to each of 10 questions regard-ing significant medical conditions; no history of facialinjury or surgery; not receiving orthodontic treatment;not pregnant or nursing.

This analysis uses data from all 185 recruited TMD casesand one-half of the 3,263 recruited controls (1,633 peo-ple). The controls for this analysis were selected at ran-dom so that data from people in the reserved samplecould be used for validation studies that will be reported

elsewhere. The accompanying paper gives a more de-tailed account of study recruitment, case-classificationmethods, and inclusion and exclusion criteria.35

TheOPPERA study wasreviewed andapprovedby insti-tutional review boards at each of the 4 study sites and atthe data coordinating center, Battelle Memorial Institute.Allstudyparticipantsverballyagreed to a screening inter-view done by telephone, and they provided informed,

signed consent for all other study procedures.

Psychophysical ProtocolsQuantitative sensory testing was conducted in 3 sen-

sory domains, in the following order: pressure pain, me-chanical cutaneous (pricking) pain, and heat pain.

T62 The Journal of Pain OPPERA Study: QST Differences in TMD


3/14

Pressure Pain

Pressure pain thresholds (PPT) were assessed usinga commercially available pressure algometer (Somedic;Horby, Sweden). Five body sites were tested, bilaterally,in the following order: 1) the center of the temporalismuscle; 2) the center of the masseter muscle; 3) overlyingthe temporomandibular joint (TMJ); 4) the center of the

trapezius muscle; and 5) overlying the lateral epicondyle.The protocol involved manual application of the algo-meter, with which the examiner would increase pressureat a steady rate (30 kPa/s), until the participant indicatedfirst pain sensation by pressing a button. The first trial ateach test site was considered a practice trial, and was ex-cluded from data analysis. In subsequent trials, the pres-sure level at the time of the button press was recorded asa threshold estimate. If no response was given at thepoint the stimulus reached 600 kPa, a value of 600 wasused as the threshold value. This procedure was repeatedat the same test site until either: 1) 2 values were re-corded within 20 kPa of one another; or 2) 5 trials were

administered. In either case, the mean of the 2 closestvalues was recorded as the threshold estimate.

The examiners from each of the 4 study sites partici-pated in a reliability exercise at an early stage of this pro-

ject. PPTs were assessed on a group of 16 subjects (notpart of the main study), each of whom was tested by 2 ex-aminers at each of 5 body sites. The overall ICC was .91,ranging from .87 to .94.

Mechanical Cutaneous (Pinprick) Pain

Pricking pain sensitivity was assessed using a set ofweighted probes, manufactured locally, matching those

used by the German Neuropathic Pain Network.32

Thisset of probes had a flat contact area of .2-mm diameter,and exerted forces between 8 and 512 mN. Stimuli wereapplied to the dorsum of the digits 2 to 4. Measures in-

cluded pain threshold, ratings of pain intensity in re-sponse to the 2 largest stimulus intensities, and TS ofpain. Pain threshold was derived using an adaptive stair-case method,6 calculated as the geometric mean of 5 se-ries of ascending and descending stimulus intensities. Ifsubjects gave 2 no responses in a row using the 512-mN probe, the staircase was halted and a value of 512was used as the threshold value.

After threshold determination, suprathreshold cuta-

neous mechanical pain sensitivity was assessed usinga protocol similar to that of the German Neuropathic

Pain Network.32 Participants judged the pain intensityevoked by suprathreshold stimuli, verbally reportinga number between 0 and 100, without a visual reference.Participants were instructed that 0 represented nopain, while 100 represented the most intense painimaginable. Participants reported pain intensity aftera single stimulus (applied for approximately .5 sec), andthen again after a series of 10 stimuli were applied at1-second intervals. For the series of 10 stimuli, partici-pants were asked to report an overall pain intensity forthe series of stimuli. At 15 and 30 seconds after the

series-of-10 stimuli was administered, participants wereasked to rate the pain intensity of any residual sensation

at the stimulated finger. Participants were also asked ifany residual nonpainful sensations were present at the30-second time point. This testing series (a single stimu-lus followed by a series-of-10) was conducted 4 timeswith the 256-mN probe, and then with the 512-mNprobe. TS of pricking pain was calculated as the differ-ence between the rating of the series-of-10 stimuli and

the rating of the single stimulus. For these suprathres-hold rating protocols, participants were first given prac-

tice runs on a site distant from subsequent testing, inorder to verify the participants understanding of theprotocol. In the course of this testing, if the participantreported 100, the protocol was halted. The participantwas then offered the option of continuing with thenext series, or omitting the rest of this set of tests. Theparticipant was also instructed that s/he could stop test-ing at any time by telling the examiner to stop.

Heat Pain

Heat pain sensitivity was assessed using a commerciallyavailable thermal stimulator (Pathway; Medoc; RamatYishai, Israel). Stimuli were applied on the ventral fore-arm. Heat pain threshold was determined using a proto-col similar to that for PPT. The ATS thermode (2.56 cm2)was manually placed in contact with the skin at a temper-ature of 32C. After a few seconds, the temperature in-creased at a rate of .5C/second until the participantpushed a button indicating s/he just then felt a pain sen-sation. The temperature of the thermode at the time of

the button press was recorded as a threshold estimate.This was repeated 4 times, moving the thermode toa new site on the forearm each time. Following this,

pain tolerance was estimated using the same protocol.The sole difference was that the participant was in-structed to press the button when s/he could no longertolerate the pain. This was repeated 4 times, movingthe thermode for each trial. For both threshold and tol-erance testing, a ceiling temperature was set at 52 C,which was entered as the threshold or tolerance estimateif the participant failed to press the button on a giventrial. For both the threshold andtolerance protocols,par-ticipants were first given practice runs on a site distantfrom subsequent testing, in order to verify the partici-pants understanding of the protocol.

Following heat pain tolerance testing, participants

judged the pain intensity evoked by suprathresholdheat stimuli, verbally reporting a number between0 and 100. As with pricking pain ratings, participantswere instructed that 0 represented no pain, while100 represented the most intense pain imaginable.Participants were told that they would receive 10 ther-mal stimuli in a row, and would be verbally cued to re-port their peak pain intensity after each stimulus.Practice trials were administered to provide the partici-pant a sense of the timing of stimulus delivery, and toverify understanding of the protocol. The CHEPS ther-mode (5.73 cm2) was manually placed on the skin ata temperature of 38C, and then a series of 10 tempera-

ture pulses were given at 2.4- to 2.5-second interstimulusintervals. For the first series of stimuli, the peak

Greenspan et al The Journal of Pain T63


4/14

temperature was 46C, with a ramp rate of 20C/second,and a hold time of 750 msec at the peak temperature.The participant was cued to report when the tempera-ture just started to decline after reaching the peak. At in-tervals of 15 and 30 seconds after the 10th thermal pulse,the participant was asked to rate the pain intensity ofany lingering sensation using the same 0 to 100 scale. Fol-

lowing this, the thermode was moved to another loca-tion on the forearm, and the same protocol wasconducted with a peak temperature of 48C. This was fol-lowed by another test series with a 50C peak tempera-ture at a third location. In the course of the testing, ifthe participant reported 100, the thermode was removedand that series of thermal stimuli was halted. The partic-ipant was then offered the option of continuing with thenext series, or omitting the rest of this set of tests. Theparticipant was also instructed that s/he could stop test-ing at any time by telling the examiner to stop.

An inadvertent protocol variation at 1 of the studysitesusing a different baseline temperature (38C) for

heat pain threshold and tolerance protocolsled to sep-arate evaluation of these data. Comparison of thresholddata from the 1 site versus the others revealed a clear dif-ference in the distributions, indicating that the valuesfrom the 1 site shouldbe excluded from analysis. Compar-ison of tolerance data showed very little differenceamong the sites, indicating that these data could be com-bined without a biasimposed by the procedural variation.

Statistical Methods

Missing Data Imputation

Imputation for missing values used an expectation-maximization method (as described in Slade et al35)which finds maximum likelihood estimates using a para-metric model for incomplete data. For most summarymeasures, the criteria for imputation on QST itemswere based on having at least 50% of the data for anytype of measure for a given participant. Specifically,

data were imputed for individuals with at least 8 validvalues among the 16 measurements from the 256 mNand 512 mN probes on single stimulus and series of 10stimulus measurements, whereas data are not imputedfor individuals with fewer than 8 valid values. Similarly,individuals with at least 8 of the 16 aftersensation mea-

surements at 15 and 30 seconds for 4 trials each were in-cluded in imputation for poststimulus ratings. For PPTdata, imputation was conducted for individuals with atleast 10 of the 20 measures on the 5 body sites: tempora-lis, masseter, TMJ, trapezius, and lateral epicondyle, each

measured twice bilaterally. For thermal measures, datawere imputed for individuals with at least 4 of the com-bined 8 threshold and tolerance ratings from tempera-tures 1 to 4. Similarly, imputation was done forindividuals with at least 3 of the 6 aftersensation mea-surements at both 15 and 30 seconds from temperatures46, 48, and 50C. An exception to this 50% criterion wasmade for thermal ratings, whereby data were imputed

for individuals with at least 1 of the 30 measurementsfrom the thermal trains at 46, 48, and 50C. The rationale

for this criterion stems from the data collection protocolwhich required a halt in the procedure when participantsgave a rating of 100, with the option to continue at sub-sequent temperatures. Thus, an individual with only 1rating at 46C is informative, and their exclusion wouldcertainly bias results. The statistics regarding data impu-tation for each variable are presented in Supplementary

e-Tables 1 and 2.

Group Comparisons

Descriptive statistics for each summary score were gen-erated. Statistical significance of differences in meanscores was evaluated using an ANOVA derived froma least squares general linear model in which study sitewas a covariate. Study site was used as a covariate be-cause operational requirements during recruitment cre-ated different proportions of cases among sites. Therelationship between each summary score and occur-rence of TMD was expressed as the standardized odds ra-

tio (SOR), calculated from an unconditional, binarylogistic regression model. To achieve this, the summaryscore was transformed to a unit-normal deviate, andstudy site was a covariate. The transformation meantthat odds ratios could be interpreted as the relativechange in odds of TMD for each standard deviation ofchange in the summary score. In order to be able to di-rectly compare SORs derived from threshold/tolerancemeasures with all the others, the inverse of threshold/tol-erance SORs are presented, so that for all measures,a higher value represents greater pain sensitivity forTMD patients versus controls. A second logistic regres-sion model generated a fully-adjusted estimate of the re-

lationship, using additional covariates of age (in years),gender, and race/ethnicity (dichotomized as white ornon-white). A third logistic regression model using theimputed dataset calculated standardized odds ratiosfor each summary score, with adjustment for study site,age, gender, and race/ethnicity.

All P values were computed without adjustment formultiple tests, and we therefore refrain from nominat-ing P = .05 as a threshold for statistical significance. Inthis papers case-control analysis, 40 measures were in-vestigated and, therefore, Bonferroni correction for theprobability of type I error would yield a critical P value

of .05/40 = .00125. Using the same rationale, rejection

of the null hypothesis concerning odds ratios would oc-cur only if the 99.875% confidence interval excludedthe null value of 1. In general, though, we avoid drawingconclusions about statistical significance of associations,even with correction for multiple tests, because these pa-pers report only univariate- or demographically-adjustedresults. Furthermore, the Bonferroni adjustment is prob-ably overly conservative in this setting, where severalmeasures are moderately correlated. Instead, we will re-serve judgments about statistical significance to subse-

quent papers that will use multivariable modeling toconsider multiple characteristics simultaneously, as pro-posed in the OPPERA heuristic model.

Of the 185 TMD cases, only 19 had arthralgia with nomyalgia, and only 9 had myalgia with no arthralgia



5/14

(the remaining 157 cases were classified with both myal-gia and arthralgia). The small numbers with myalgiaalone or with arthralgia alone precluded any useful anal-ysis of those TMD subtypes.

Principal Component Analysis

Additionally, we applied principal component analysis(PCA) to this data set to reduce the dimensionality and toidentify putative latent variables. The approach beganwith 4 steps, widely used in exploratory principal compo-nent analysis:27 1) variable selection; 2) evaluation of thecorrelation matrix; 3) extraction of principal compo-nents; and 4) varimax rotation and interpretation of fac-tor loadings. This model is being used for purelyexploratory purposes. In particular, this model will notbe used to calculate summary scores for the various com-ponents.

Our primary interest focused on PCA loadings in con-

trols for 3 reasons: 1) there are many more controls

than cases, thereby improving statistical power to iden-tify factors and estimate loadings; 2) our underlying con-ceptual model of TMD22 proposes that relationshipsbetween putative risk factors might alter following on-set of chronic pain, and we wanted to evaluate that pos-sibility; and 3) the long-term goal of OPPERA is toidentify risk factors for TMD among controls,so it is desir-able to identify a reduced set of variables for that group.Thus, we fit separate PCA models for the TMD cases andcontrols.

We included the variables identified in Tables 1 and 2in the model, with 2 exceptions related to cutaneousmechanical pain, and 1 exception related to heat pain.

The Overall ratings of 10 stimuli1 for each of the 2probe intensitieswere excluded since they werederivative of the TS measures, which were consideredmore important. Given that heat pain thresholds from1 of the sites was excluded from analysis, we elected toremove that variable from the PCA. This allowed us toinclude all the other data from all 4 sites. We elected tokeep all of the remaining 33 variables in the modeleven though some of the variables were highlycorrelated with one another. We wanted to determinewhich measures could be combined into latentvariables, and this required that all the variables beretained in the model. Although this has the potential

to increase the variance of our estimates of the PCAloadings, this was not a major concern given our verylarge sample size. As shown below, we estimated thevariance of our PCA loadings using bootstrapping, andthese variances were uniformly low despite thecorrelations among the input variables. Also, since weare not creating summary scores, there is no conceptualproblem with having multiple items that measuresimilar constructs loading on a single component.

After imputing missing values as described earlier (seepreceding section), there were 291 participants who stillhad at least 1 missing value among the variables that weincluded in the PCA model. Moreover, some study sites

had more participants with missing values than others.Given the danger of bias associated with dropping 291

participants from the model, we performed a secondround of imputation. First, we imputed any missing ob-servations of mechanical cutaneous pain variables basedon the observed mechanical cutaneous pain variables foreach participant. We also imputed missing heat pain vari-ables based on the observed heat pain variables. We usedthe EM algorithm to perform this imputation under the

assumption that the data are multivariate normal, as de-scribed previously. Finally, we imputed any remaining

observations of any of the variables in the model basedon all the observed QST variables for that participantacross all domains, namely: pressure pain, cutaneous me-chanical pain, and heat pain. After performing this sec-ond round of imputation, there were no remainingparticipants with missing data. PCA was performed onthis fully imputed data set.

A scree plot was generated to estimate the number ofcomponents to include in the model (Fig 1). The varianceexplained by each principal component decreased mostconspicuously after the fifth component, suggesting

that at least 5 components are needed. This estimatewas verified with a parallel analysis.39 Parallel analysis es-timates the number of components to include in a PCAmodel by generating random data sets with the samenumbers of observations and predictor variables as theoriginal data. Eigenvalues were computed for eachrandom data set and averaged over all the data sets.When the average eigenvalue from these randomlygenerated data sets is larger than the corresponding ei-genvalue of the original data, then the principal compo-nent associated with that eigenvalue is likely to berandom noise. The parallel analysis also showed strong

evidence that components 1 through 5 were above the

chance line (Fig 1). Components 6 and 7 were near thechance line, so it was unclear if they should be includedin the model.

We therefore fit PCA models using 5, 6, and 7 compo-nents. The bootstrap confidence intervals for compo-nents 6 and 7 were very wide for many of the loadings,suggesting that the estimated loadings were unstable.(Data not shown) However, the confidence intervals forthe loadings of the first 5 components were very narrow,indicating a stable (and accurate) model. Additionally,

the Cronbachs alpha values for these 5 componentswere highranging from .82 to .94 further support-ing a good reliability of this model. Thus, we report

a model based on 5 components.We fit the PCA models using the R statistical comput-

ing platform. The models for the TMD cases and controlswere fitted separately. All variables were normalized tohave mean 0 and standard deviation 1 prior to fittingthe models. After calculating the PCA eigenvectors, weapplied a varimax rotation to increase the interpretabil-ity of the resulting PCA loadings. (Nonorthogonal rota-tions were also considered, but the resulting loadingswere nearly identical to the loadings produced by thevarimax rotation.) Subsequent descriptions of the PCAloadings refer to the rotated loadings.

We estimated the variance in the PCA loadings of each

model by drawing 1,000 bootstrap samples for each dataset and fitting a PCA model for each replicate. The 95%



6/14

confidence bounds for our PCA loadings were estimatedto be the 2.5 and 97.5% percentiles of the correspondingloading over the 1,000 bootstrap replicates.10

A second PCA was conducted to determine the place ofheat pain thresholds in the component weights. In thiscase, data were derived from only the 3 study sites em-ploying the same testing protocol.

Much of the data from this study is elaborated inSupplementary e-Tables 1 and 2 enumerate missing andimputed data for each QST measure. Supplementarye-Tables 3 to 5 report means and distributions for eachQST measure. Supplementary e-Tables 6 to 9 elaboratethe results of the PCA analyses for both TMD cases andcontrols. Supplementary e-Tables 10 to 15 report thegender, race/ethnicity, and age effects among the con-trol groups QST measures.

Results

Case-Control ComparisonsAll P values were computed without adjustment formultiple tests. A decision was made not to use Bonfer-roni (or other) methods to correct for multiple compar-isons. We avoid drawing firm conclusions aboutstatistical significance of associations, even with correc-tion for multiple tests, as this paper reports only uni-variate- or demographically-adjusted results. Instead,exacting judgments about statistical significance willawait subsequent papers that will use multivariablemodeling to consider multiple characteristics simulta-neously.

For nearly all measures, cases were significantly more

sensitive, on average, than controls (Tables 1 and 2;Supplementary e-Tables 35). At all bodily sites, pressurepain thresholds (PPTs) were lower in cases than controls.Exceptions were found for heat pain TS measures, whichfailed to show statistically significant case-control differ-ences for all 6 measures. Additionally, the ratings of indi-vidual heat pain stimuli at 48 and 50C failed to showstatisticallysignificant case-controldifferences, althoughthe same trend was apparent as the significant differ-ence found for ratings of 46C. Consideration of the indi-vidual trial ratings reveal that for all trials, cases rated theheat pain as more intense than controls, while the in-crease in ratings with repetition was only slightly greater

for cases than controls (Fig 2).Standardized odds ratios (SORs [inverted for thresh-

old/tolerance measures]) were above 1.0, signifyingthat greater odds of TMD were associated with greaterratings of pain or lower threshold/tolerance. In all in-

stances, with the exception of the heat pain TS mea-sures, the ratings of 48C stimuli, and heat painthreshold (with imputed values), the 95% CI of theSORs excluded the value of 1.0, reflecting a statisticallysignificant difference between the groups. Among thethreshold and tolerance measures, SORs for the PPTswere the highest (ranging from 2.384.05), particularlyfor the cranial sites. Furthermore, the 95% CIs for

SORs of trapezius and cranial site PPTs are distinctly out-side the range of SORs for any of the rating measures,

indicating a statistically significant difference betweenthe PPTs and any of the rating measures. The SOR formechanical cutaneous pain threshold (2.01) was similarto that for PPT of the lateral epicondyle (2.38), withlargely overlapping 95% CIs. In contrast, the 95% CIsfor both of these measures did not overlap with thatof heat pain threshold, indicating a statistically signifi-

cant difference in SORs for mechanical pain versusheat pain thresholds. Among all the suprathreshold rat-

ing measures (ratings of single stimuli, aftersensations,and TS; ranging from 1.071.44), 95% CIs overlappedconsiderably, suggesting no statistically significant dif-ference among any of these measures in their abilityto distinguish TMD cases from controls.

The SORs for all QST measures changed little when ad-justed for multiple factors, or comparing results with andwithout missing data imputation (Tables 12).

PCA

The loadings (Table 3) and corresponding 95% confi-dence intervals (Supplementary e-Table 6) show thePCA results derived from the control participants. Com-ponent 1 is primarily based on the heat pain ratings ofsuprathreshold stimuli, with a weaker loading fromheat pain tolerance. Component 2 is primarily based onthe heat pain aftersensation ratings, with weaker load-ings from the mechanical pain aftersensation ratings de-rived from the larger intensity stimulus. Component 3 isprimarily based on the mechanical cutaneous pain mea-sures:ratings, TS, aftersensations, and threshold. Compo-nent 4 is principally based on pressure pain thresholds.Component 5 is primarily based on heat pain TS. Of

note, the mechanical TS derived fromthe higher intensitystimulus did not load above .40 on any of the factors.

A second PCA was conducted on data from controlsubjects, this time including heat pain threshold data,which restricted the PCA to data derived from the 3 sitesusing the same heat pain threshold parameters. A similarset of components were produced, with the highly

weighted variables still defining the 5 components as de-scribed above. Heat pain threshold had the strongestweighting in Component 4 (.49), and weightings below.40 for other components. Thus, it is more strongly asso-ciated with pressure pain thresholds than with any of theother heat pain measures.

PCA results for the TMD cases (Supplementary e-Tables7 and 8) show a similar grouping, with small differences.For one thing, the ordering of the components (based onpercent of variance accounted for) is different, althoughthis reflects very small differences in the actual variance

values. With respect to individual loadings, the mechan-ical pain aftersensation ratings load strongly with bothcomponents 1 and 5 for the TMD cases (correspondingto components 2 and 3 for the controls). In this manner,it split its association between heat pain aftersensationand other mechanical pain measures. In contrast, thesemeasures for the control group only significantly associ-ated with other mechanical pain measures. Additionally,

heat pain intensity ratings load strongly on component 2(component 1 for controls), but also load negatively



7/14


8/14


9/14

on component 3 (component 5 for controls), which oth-erwise has strong positive loadings from heat pain TSmeasures. This negative loading was indicated, but of in-sufficient magnitude to be deemed significant in thecontrol group. Given these apparent differences, a seriesof permutation tests were performed to test the null hy-pothesis that the loadings were the same for bothmodels (Supplementary e-Table 9 [component number-

ingmatching that for controls]). Of the differences notedabove, the loadings of heat pain single stimulus ratings

upon TMD component 3 (control component 5) were sta-tistically significantly different between TMD cases andcontrols at the P = .05 level. However, any adjustmentfor multiple testing would render these differences non-significant. Otherwise, a scattering of other loading dif-ferences were statistically significant at P = .05, but oflow loading overall. Thecorresponding confidence inter-vals for the model based on TMD cases are slightly wider(Supplementary e-Table 6 versus Supplementary e-Table8), but this is not surprising, given that the sample sizewas much smaller. Neither model shows evidence of

high variance or other forms of instability.

Demographic DifferencesAmong controls, women showed greater pain sensitiv-

ity than men for almost all pain measures(Supplementary e-Tables 10 and 11). The exceptionswere the heat pain TS measures, which were not signifi-cantly different between the sexes.

Several pain measures revealed less pain sensitivity fornon-Hispanic whites compared to other racial/ethnicgroups combined (Supplementary e-Tables 12 and 13).

Exceptions to this included the cranial pressure painthresholds, mechanical cutaneous and heat pain thresh-

olds, and some of the heat pain TS measures, which didnot show significant group differences.

A number of pain measures differed significantlyamong age groups (Supplementary e-Tables 14 and 15).In most instances, differences were in the direction ofgreater pain sensitivity for the younger participants.This was observed for some of the pressure pain thresh-olds (marginally), all of the mechanical cutaneous pain

measures, and some of the heat pain measures, moststrongly the aftersensation ratings.

DiscussionThe key findings from this study are as follows:1. Chronic TMD cases were more pain sensitive than

controls to a wide range of mechanical and thermalstimuli, applied to symptomatic and asymptomaticbody sites. This difference was most pronouncedfor pressure pain thresholds, and less pronouncedfor most of the cutaneous mechanical and heat

pain sensitivity measures. One aspect of heat painsensitivitytemporal summationdid not distin-guish TMD cases from controls.

2. Mechanical and heat pain sensitivity measures canbe regarded as reflecting 5 components of experi-

mental pain sensitivity: 1) heat pain ratings; 2)heat pain aftersensations andtolerance; 3) mechan-ical cutaneous pain sensitivity; 4) pressure painthresholds; and 5) heat pain temporal summation.These components were similar for TMD cases andnon-TMD controls.

3. Demographic differences were found for multipleexperimental pain measures, most strongly and

widely for sex differences, somewhat less so for ra-cial/ethnicity differences, and for age differences.

Figure 1. Components resultingfrom principal component analysis(PCA) (verticalbars) and parallel analysis(trianges) conducted ondata from control participants.



10/14

TMD Case-Control Differences inExperimental Pain Sensitivity

A number of studies have reported TMD case-controldifferences in experimental pain sensitivity, typically us-ing 1 or 2 pain measures,23-25,33,38 although notuniversally.3,38 In all instances of significant differences,chronic TMD cases were found to be more painsensitive. This current study supports these previousreports, with a much larger sample size, and alsoprovides the opportunity to compare pain sensitivitydifferences across a large array of tests. With respect toPPTs, the statistical tests and the (reverse coded) SORswould be expected to be highly significant at cranialsites, given that these sites are symptomatic regions forTMD. Yet, the (reverse coded) SORs for the PPTs onextracranial sites are also highly significant, indicatingthat remote muscle hyperalgesia is prominent in TMDcases. The cutaneous mechanical pain threshold (reversecoded) SOR, while also highly significant, was smallerand had 95% CIs that did not overlap with cranial site

PPTs, thus making it less discriminating of case-controldifferences than the PPTs. Furthermore, the heat painthreshold (reverse coded) SOR, while also significant,was smaller still and had 95% CIs that did not overlapwith any of the other thresholds (reverse coded) SORs,making it least discriminating of case-control differences

among all the threshold measures. It should be notedthat themechanical cutaneouspain andheat pain testingonly involved asymptomatic body sites (the upper ex-tremity). As such, they represent an index of remote orwidespread hyperalgesia for the TMD cases.

With respect to suprathreshold ratings of single stim-uli, the SORs for mechanical cutaneous stimuli werehigher than those for heat stimuli, showing the sametrend revealed by thresholds. In fact, while all the me-chanical cutaneous measures showed significant case-control differences, not all the heat pain measures did.Specifically, the fully adjusted heat pain TS SORs showedthe same trend in case-control differences, but did not

reach statistical significance. However, the fact that95% CIs for most of these SORs overlap mitigate against

Figure 2. Ratings of individual heat stimuli (0100 scale) presented at 3 different temperatures, 10 times each. Solid lines = tempo-romandibular disorder (TMD) cases; dashed lines = controls.



11/14

the general hypothesis that suprathreshold cutaneousmechanical pain measures discriminate TMD cases

from controls better than suprathreshold heat painmeasures do.

SORs for ratings of aftersensation pain intensityshowed similar magnitudes of association with TMD asthe single trial ratings. This observation suggests that ex-aggerated sensory perseveration is as prominent as cuta-neous hyperalgesia in TMD cases.

The heat pain TS measureswhether calculated as

a difference in ratings between the first and later stimuli,or as the rate of change in pain ratings over the first few

stimuli in a seriesdid not show statistically significantcase-control differences. This contrasts with earlier re-

ports of increased heat pain TS in women with TMD.25

It is worth noting that the curves depicting TS in Fig 2of this papershowing the higher ratings of pain forTMD cases across all trialslooks very similar to Fig 4Bof Maixner et al.25 In this respect, these results are consis-tent. Also, TS can be calculated in multiple ways, and iscomplicated by ceiling effects of a bounded rating scale,as used here. The current report employed 2 commonly

used measures for heat pain TS, and used a conservativeprocess to adjust for ceiling effects. A subsequent report

Table 3. Component Loadings for Principal Components Analysis Model in Controls*

COMPONENT 1 COMPONENT 2 COMPONENT 3 COMPONENT 4 COMPONENT 5

Pressure pain threshold

Temporalis .11 .05 .10 .89 .02

Masseter .11 .04 .07 .92 .02

TM joint .11 .04 .06 .90 .03

Trapezius .21 .12 .10 .81 .00

Lateral epicondyle .19 .05 .03 .81 .01

Mechanical cutaneous pain

Threshold .17 .06 .41 .37 .07

Single stimulus ratings

256mN probe .26 .08 .71 .11 .04

512mN probe .36 .10 .72 .13 .06

Aftersensation ratings

15 s, 256mN probe .01 .27 .85 .01 .03

30 s, 256mN probe .03 .28 .79 .03 .02

15 s, 512mN probe .11 .40 .80 .05 .01

30 s, 512mN probe .06 .41 .76 .05 .04

Temporal summation

256mN probe .28 .09 .58 .14 .00

512mN probe .37 .14 .31 .16 .10

Heat pain tolerance .46 .25 .12 .37 .08Single stimulus ratings

46C .77 .17 .20 .16 .36

48C .85 .17 .18 .15 .32

50C .88 .14 .12 .13 .21

Ratings of 10 stimuli: area under curve

46C .86 .18 .19 .18 .11

48C .91 .19 .16 .17 .19

50C .83 .18 .14 .17 .27

Thermal aftersensations

15 s, 46C .21 .80 .24 .05 .01

30 s, 46C .23 .85 .25 .07 .03

15 s, 48C .16 .84 .26 .08 .05

30 s, 48C .14 .83 .17 .04 .01

15 s, 50C .17 .88 .19 .09 .0130 s, 50C .09 .86 .24 .07 .04

Temporal summation: highest minus first rating

46C .10 .03 .02 .03 .83

48C .06 .02 .05 .04 .87

50C .33 .02 .01 .02 .72

Temporal summation: slope of line for first three ratings

46C .25 .00 .04 .00 .70

48C .10 .00 .03 .00 .80

50C .18 .03 .01 .02 .74

Cumulative variance .17 .32 .46 .59 .72

Cronbachs alpha .96 .95 .9 .93 .87

NOTE. Bold entries indicate statistical significance.*Data are from 1,633 controls who do not have TMD.



12/14

will examine these data in a more elaborate way, in orderto better account for individual variation in TS profilesand ceiling effects.

It is worth noting that our results are consistent witha recent report, in which TMD cases were not signifi-cantly different from controls in measures of heat painTS, but showed significantly higher heat pain aftersensa-

tion ratings.30

Mechanical pain TS was significantly higher for TMDcases than controls, yet heat pain TS measures did notdiffer significantly between the 2 groups. This may bedue to mechanistic differences between mechanicaland thermal TS. However, protocol differences mayalso be responsible for differences in our TS results.For heat pain, individual stimuli were rated, and TSmeasures were calculated by virtue of rating changesfrom 1 stimulus to the next. For mechanical pain, a rat-ing was given for a single stimulus, and then anotherrating was given for a series of 10 stimuli applied at1 Hz, and the difference between these 2 ratings

were taken as the TS. The extent to which these meth-odological differences could contribute to different re-sults cannot be separated from potential mechanisticdifferences.

In comparing SORs from site-adjusted versus fully-adjusted analyses, the latter typically had slightly highervalues. Thus, the case-control differences were not seri-ously confounded by demographic differences betweenTMD cases and controls. There were few or no changes infully adjusted SORs when comparing analyses conductedwith unimputed versus imputed data. Thus, the processof data imputation performed for this study, while po-

tentially changing the results by way of reducing the

bias of data dropout, had no substantive effect uponthe outcome.

PCA-Derived Components ofExperimental Pain Sensitivity

PCA analysis on control participants yieldeda 5-component solution that was parsimonious and ro-bust. The components were characterized as heat painratings (component 1), heat pain aftersensations andtolerance (component 2), mechanical cutaneous painsensitivity (component 3), pressure pain thresholds(component 4), and heat pain TS (component 5). Of

note, all components are modality specific.Modality specific components have been described

in a previous report of factor analyses of experimentalpain measures.16 Another study evaluating associa-tions among multiple experimental pain measures

found higher associations within than across modali-ties.2 In contrast to these studies, an analysis of painthreshold data across multiple stimulus modalitiesfound only a small fraction of the variance attribut-able to stimulus modality.26 The results reported heresupport the previous studies that show experimentalpain measurement factors or components to be mo-dality specific.

It is reasonable to expect that the experimental painmeasures may separate according to stimulus modality.

The nervous system components that relay nociceptiveinformation show some degree of modality specificity.For instance, along with polymodal nociceptors, whichrespond to both thermal and mechanical stimuli, thereare also nociceptors that only respond to mechanicalstimuli and others that only respond to heat.13,20,36

Furthermore, while many spinal cord and thalamic

neurons respond to both noxious thermal andmechanical stimuli, there are those that only respond tonoxious mechanical stimuli.7,19 Thus, at the individualneuron level, nociceptive information initiated by heatversus mechanical stimuli is partially segregated in theCNS. This separation of neural processing would beexpected to play a role in perceptual differences, andresult in increased intermodality versus intramodalityvariance.

Heat pain TS measures did not load with heat pain rat-ings on a common component, in agreement with a pre-vious study.16 This finding suggests that TS measuresreflect mechanisms that are distinct from those responsi-

ble for simple suprathreshold heat pain perception.Indeed, some differences, such as the key role of theN-methyl D-aspartate (NMDA) receptor in heat pain TS,have been documented.40

The current study is the first to report data on bothmechanical and thermal TS obtained from the sameparticipants. Strikingly, the 2 measures did not load to-gether in the PCA analyses, suggesting significant differ-ences in the mechanisms underlying TS between the 2modalities. However, as noted above, methodologicaldifferences may have played a role in differentiatingthese measures.

Demographic DifferencesIn accordance with many previous reports, women

showed significantly greater pain sensitivity than menin almost all of the QST measures used in this study. A re-cent review11 tabulated results from published studies ofthis topic, and found that 68 to 75% of studies reportedsignificantly lower thresholds or significantly greater TSfor women when tested with modalities used in the pres-ent study. The 1 exception to finding sex differences inthe present study was with respect to TS of heat pain.Of 4 published studies evaluating sex differences inheat pain TS, 3 reported significantly greater TS in

women,12,14,31

and 1 reported no difference.37

Thus,a large majority of laboratory pain studies investigatingsex differences found greater sensitivity among womenthan men, although the magnitude of the differencevaries across studies and across pain measures.

Most pain measures revealed significantly less painsensitivity for non-Hispanic whites than non-white racesand Hispanics, combined. This was true for all the me-chanical cutaneous pain measures except threshold,and for the heat pain measures, except for thresholdand some of the TS measures. Interestingly, the pressurepain thresholds showed significant differences on thetrapezius and lateral epicondyle sites, but not any of

the cranial sites. These results are somewhat consistentwith previous reports of greater pain sensitivity in



13/14

African American participants versus non-Hispanic whiteparticipants, particularly for suprathreshold pain assess-ments.4,29

Fewer pain measures showed significant age effects. Inmost cases, differences were in the direction of greaterpain sensitivity for the younger participants. In compari-son to sex and racial differences in pain measures, age

differences showed smaller mean value differences. Itshould be noted that the age range of participants forthis study (1844) is not large compared to studies specif-ically evaluating the effects of age upon pain. Most stud-ies have reported no significant age effects uponexperimental pain within this age range, but onlywhen participants in this age range were comparedwith more elderly participants.15

Study LimitationsOne should note that the large sample size for this

study provides for significant group differences even

with small mean differences. As one example, case-control differences in thepain intensityratings of cutane-ous mechanical probes were statistically significant, eventhough the mean rating difference was less than 10 onthe 0 to 100 scale (Table 1). This raises a cautionary noteof distinguishing between statistical significant differ-ences and clinically meaningful differences in pain sensi-tivity. At the same time, such results exemplify that QSTmeasures can reveal group differences in pain sensitivitythat are of a magnitude likely to be missed in a clinicalexam.

As is the case for many other studies of demographicdifferences in experimental pain, several of the observed

differences in the present study were either unadjusted,or were adjusted only for study site. Importantly, therewas no adjustment for etiological characteristics (eg, psy-chological states or other demographic variables) thatmight confound these demographic associations. Hence,while these observed differences are useful to illustrate

demographic risk indicators and potential sources ofconfounding in our case-control comparisons, they donot necessarily signify causal effects of demographiccharacteristics.

We formally assessed interexaminer reliability of PPT,as this QST measure was the most likely to be variablyinfluenced by the examiners handling of the stimulus

probe. We found a very good level of reliability (overallICC of .91), giving us confidence in consistent assess-

ments across examiners and sites. A recent study re-ported interexaminer ICCs for several different sensorytesting protocols, including some closely approximatingthose used here.28 They also report a high ICC for PPT(.89), and nearly as high value for heat pain threshold(.87). Somewhat lower values were found for mechanicalcuntaneous threshold (.56) and mechanical cutaneoustemporal summation (.52).

ConclusionsThis large-scale study demonstrated significantly

greater pain sensitivity for people with TMD versus thosewithout TMD in a wide range of mechanical and thermalpain tests. Most of the tests involved stimulation ofasymptomatic body sites on the upper extremity, provid-ing support for the theory of a generalized upregulationof pain processing in chronic pain conditions, even thosewith somatically restricted symptomology.

AcknowledgmentsThe authors would like to thank the OPPERA research

staff for their invaluable contributions to this work. Inaddition, we express our gratitude to the research partic-

ipants who have devoted time and effort in support ofthis research.

Supplementary DataThe supplementary data accompanying this article areavailable online at jpain.org and sciencedirect.com.

References

1. Arendt-Nielsen L, Lautenbacher S: Assessment of painperception, in Lautenbacher S, Fillingim RB (eds): Patho-physiology of Pain Perception. New York, NY, Kluwer Aca-

demic, 2004, pp 25-42

2. Bhalang K, Sigurdsson A, Slade GD, Maixner W: Associa-tions among four modalities of experimental pain inwomen. J Pain 6:604-611, 2005

3. Bragdon EE,LightKC, Costello N, Sigurdsson A, Bunting S,Bhalang K, Maixner W: Group differences in pain modula-tion: Pain-free women compared to pain-free men and towomen with TMD. Pain 96:227-237, 2002

4. Campbell CM, Edwards RR, Fillingim RB: Ethnic differ-ences in responses to multiple experimental pain stimuli.Pain 113:20-26, 2005

5. CarlsonCR, Reid KI, Curran SL, Studts J, Okeson JP, Falace D,

Nitz A, Bertrand PM: Psychological and physiological param-eters of masticatory muscle pain. Pain 76:297-307, 1998

6. Cornsweet TN: The staircase-method in psychophysics.Am J Psychol 75:485-491, 1962

7. Davidson S, Zhang X, Khasabov SG, Simone DA,Giesler GJ Jr: Termination zones of functionally charac-terized spinothalamic tract neurons within the primate

posterior thalamus. J Neurophysiol 100:2026-2037, 2008

8. Diatchenko L, Nackley AG, Slade GD, Fillingim RB,Maixner W: Idiopathic pain disorders - Pathways of vulnera-bility. Pain 123:226-230, 2006

9. Dworkin SF, LeResche L: Research diagnostic criteria fortemporomandibular disorders: Review, criteria, examina-tions and specifications, critique. J Craniomandib Disord 6:301-355, 1992

10. Efron B, Tibshirani R: An Introduction to the Bootstrap.Boca Raton, FL, Chapman & Hall/CRC, 1994

11. Fillingim RB, King CD, Ribeiro-Dasilva MC, Rahim-Williams B, Riley JL III: Sex, gender, and pain: a review of re-

cent clinical and experimental findings. J Pain 10:447-485,2009

http://jpain.org/http://sciencedirect.com/http://sciencedirect.com/http://jpain.org/


14/14

12. Fillingim RB, Maixner W, Kincaid S, Silva S: Sex dif-ferences in temporal summation but not sensory- dis-criminative processing of thermal pain. Pain 75:121-127, 1998

13. Garell PC, McGillis SLB, Greenspan JD: Mechanical re-sponse properties of nociceptors innervating feline hairyskin. J Neurophysiol 75:1177-1189, 1996

14. George SZ, Wittmer VT, Fillingim RB, Robinson ME: Sexand pain-related psychological variables are associatedwith thermal pain sensitivity for patients with chronic lowback pain. J Pain 8:2-10, 2007

15. Gibson SJ, Farrell M: A review of age differences in theneurophysiology of nociception and the perceptual experi-ence of pain. Clin J Pain 20:227-239, 2004

16. Hastie BA, Riley JL III, Robinson ME, Glover T,Campbell CM, Staud R, Fillingim RB: Cluster analysis of mul-tiple experimental pain modalities. Pain 116:227-237, 2005

17. Janal MN,Glusman M, Kuhl JP, Clark WC:On theabsenceof correlation between responses to noxious heat, cold,electrical and ischemic stimulation. Pain 58:403-411, 1994

18. Lautenbacher S, Rollman GB, McCain GA: Multi-methodassessment of experimental and clinical pain in patients withfibromyalgia. Pain 59:45-53, 1994

19. Lenz FA, Gracely RH, Rowland LH, Dougherty PM:A pop-ulation of cells in the human thalamic principal sensory nu-cleus respond to painful mechanical stimuli. Neurosci Lett180:46-50, 1994

20. Lynn B, Faulstroh K, Pierau F-K: The classification andproperties of nociceptive afferent units from the skin ofthe anaesthetized pig. Eur J Neurosci 7:431-437, 1995

21. Lynn B, Perl ER: A comparison of four tests for assessingthe pain sensitivity of different subjects and test areas. Pain3:353-365, 1977

22. Maixner W, Diatchenko L, Dubner R, Fillingim RB,Greenspan JD, Knott C, Ohrbach R, Weir B, Slade GD: Orofa-cial pain: prospective evaluation and risk assessment - TheOPPERA study. J Pain 12(Suppl 3):T4-T11, 2011

23. Maixner W, FillingimR, BookerD, Sigurdsson A: Sensitiv-ity of patients with painful temporomandibular disorders toexperimentally evoked pain. Pain 63:341-351, 1995

24. Maixner W, Fillingim R, Kincaid S, Sigurdsson A,Harris MB: Relationship between pain sensitivity and restingarterial blood pressure in patients with painful temporo-mandibular disorders. Psychosom Med 59:503-511, 1997

25. Maixner W, Fillingim R, Sigurdsson A, Kincaid S, Silva S:

Sensitivity of patients with painful temporomandibular dis-orders to experimentally evoked pain: Evidence for alteredtemporal summation of pain. Pain 76:71-81, 1998

26. Neddermeyer TJ, Fluhr K, Lotsch J: Principle componentsanalysis of pain thresholds to thermal, electrical, and me-chanical stimuli suggests a predominant common sourceof variance. Pain 138:286-291, 2008

27. Pett MA, Lackey NR, Sullivan JJ: Making Sense ofFactor Analysis: The Use of Factor Analysis for InstrumentDevelopment in Health Care Research. Thousand Oaks, CA,Sage Pub., 2003

28. Pigg M, Baad-Hansen L, Svensson P, Drangsholt M, List T:Reliability of intraoral quantitative sensory testing (QST).Pain 148:220-226, 2010

29. Rahim-Williams FB, Riley JL III, Herrera D, Campbell CM,Hastie BA, Fillingim RB: Ethnic identity predicts experimen-tal pain sensitivity in African Americans and Hispanics. Pain129:177-184, 2007

30. Raphael KG, Janal MN, Ananthan S, Cook DB, Staud R:Temporal summation of heat pain in temporomandibulardisorder patients. J Orofac Pain 23:54-64, 2009

31. Robinson M, Wise E, Gagnon C, Fillingim R, Price D: In-fluences of gender role and anxiety on sex differences intemporal summation of pain. J Pain 5:77-82, 2004

32. Rolke R, Baron R, Maier C, Tolle TR, Treede RD, Beyer A,Binder A, Birbaumer N, Birklein F, Botefur IC: Quantitativesensory testing in the German Research Network on Neuro-

pathic Pain (DFNS): Standardized protocol and referencevalues. Pain 123:231-243, 2006

33. Sarlani E, Grace EG, Reynolds MA, Greenspan JD: Evi-dence for up-regulated central nociceptive processing in pa-tients with masticatory myofascial pain. J Orofac Pain 18:41-55, 2004

34. Sarlani E, Greenspan JD: Evidence for generalized hyper-algesia in temporomandibular disorders patients. Pain 102:221-226, 2003

35. Slade GD, Bair E, By K, Mulkey F, Baraian C, Gonzalez Y,Gordon S, Ribeiro-Dasilva MC, Pei Feng Lim, Greenspan JD,Dubner R, Fillingim RB, Maixner W, Diatchenko L, DampierD, Knott C, Ohrbach R: Study methods, recruitment, socio-demongraphic findings and demographic generalizability

in the OPPERA study. J Pain 12(Suppl 3):T12-T26, 2011

36. Slugg RM, Meyer RA, Campbell JN: Response of cutane-ous A- and C-fiber nociceptors in the monkey to controlled-force stimuli. J Neurophysiol 83:2179-2191, 2000

37. Staud R, Robinson ME, Vierck CJ Jr, Price DD: Diffusenoxious inhibitory controls (DNIC) attenuate temporalsummation of second pain in normal males but not innormal females or fibromyalgia patients. Pain 101:167-174,2003

38. Svensson P, List T, Hector G: Analysis of stimulus-evokedpain in patients with myofascial temporomandibular paindisorders. Pain 92:399-409, 2001

39. Thompson B, Daniel LG: Factor analytic evidence for theconstruct validity of scores: A historical overview and someguidelines. Educ Psychol Meas 56:197-208, 1996

40. Vierck CJ, Cannon RL, Fry G, Maixner W, Whitsel BL:Characteristics of temporal summation of second pain sensa-tions elicited by brief contact of glabrous skin bya preheatedthermode. J Neurophysiol 78:992-1002, 1997


artigo de dor journal of pain

Documents