{"paper_id":"87049d5b-d24a-4476-afa1-a32da3ba8f8d","body_text":"The nociceptive and pain-modulatory pathways enable the brain to regulate pain \nsensations through both inhibitory and excitatory processes [ 1 ]. One of the \noutputs of the central nervous system (CNS) to modulate pain is the endogenous \npain modulation (EPM). Abnormal EPM, either increased pain facilitation and/or \nimpaired pain inhibition, might be associated with the genesis of several chronic \npain mechanisms [ 2 ,  3 ]. EPM can be assessed by using psychophysical methods, such \nas temporal summation of pain (TSP) and conditioned pain modulation (CPM) [ 4 ]. \nTSP evaluates facilitatory modulation through the changes in pain perception \ncaused by a series of repeated noxious stimuli [ 4 ,  5 ]. CPM is a paradigm used to \nassess the pain inhibition pathway based on the “pain inhibits pain” theory \n[ 6 ].\nTemporomandibular disorders (TMD) are a collective term for a heterogeneous \ncondition of musculoskeletal disorders involving pain and/or functional \nlimitations in the masticatory muscles, temporomandibular joints (TMJ), and \nassociated structures in the orofacial region [ 7 ,  8 ]. Multiple studies have \nassessed pain facilitation and modulation in TMD patients in both trigeminal and \nnon-trigeminal areas. Some studies reported higher sensitivity to noxious stimuli \nin TMD patients compared with healthy controls [ 9 ,  10 ]. In addition, certain \nstudies reported lower pain thresholds [ 11 ,  12 ] and increased TSP in TMD patients \nwhen tested outside the trigeminal nerve region [ 13 ]. In contrast, another study \nfound no significant difference in pain thresholds outside the painful area [ 14 ]. \nRegarding CPM responses, TMD cases produced impaired CPM at extra-segmental sites \ncompared with controls [ 15 ,  16 ]. In other studies, however, no significant change \nin the CPM effect was found at pain-free sites between those with and without TMD \n[ 17 ,  18 ].\nMigraine is a head pain frequently characterised by unilateral throbbing pain \nand a variety of neurological and autonomic symptoms, including hypersensitivity \nto light, sound, smell, nausea, cognitive, emotional, and motor disturbances \n[ 19 ]. Studies regarding the investigation of CPM efficiency in migraine patients \nhave reported mixed results. A preliminary study showed disrupted descending pain \nmodulation in both episodic and chronic migraine [ 20 ], while other CPM studies \nindicated that migraine patients typically show only mild or absent inhibitory \nresponses, comparable to those in controls [ 21 ].\nFibromyalgia is characterised by widespread musculoskeletal pain, tenderness, \nand heightened pain sensitivity, often accompanied by other symptoms, such as \nfatigue, sleep disturbances, and cognitive issues [ 22 ]. CPM responses in \nindividuals with fibromyalgia have been somewhat inconsistent. Some studies, on \none hand, have suggested that individuals with fibromyalgia exhibit impaired CPM \nresponses [ 23 ]. On the other hand, other research has reported normal [ 24 ] or \neven enhanced CPM responses in fibromyalgia patients [ 25 ].\nTMD is associated with several comorbidities, particularly migraine and \nfibromyalgia. The three pain conditions have been associated with impaired \nEPM and they occasionally coexist in some patients. To date, no studies have \nmeasured the EPM system in TMD patients with comorbid migraine and fibromyalgia. \nConsidering previous findings, we hypothesised that TMD patients with \ncomorbidities may exhibit greater EPM impairment in the non-trigeminal innervated \narea compared with TMD patients without comorbidities. Therefore, this study \naimed to investigate EPM in TMD patients with comorbid migraine and fibromyalgia \n(FM) in a non-trigeminally innervated area.\n\nA previous meta-analysis, which encompassed 9 studies exploring pain modulation \nin chronic orofacial pain, these studies comprised 14 to 58 samples within their \nrespective groups, with an average sample size of 20 to 30 participants. The \nfindings from these studies demonstrated compromised pain modulation in patients \ncompared with control groups [ 13 ]. A power analysis was conducted using G*Power \nversion 3.1.9.4 (Heinrich-Heine-Universität Düsseldorf, Düsseldorf, \nNRW, Germany) to ascertain the minimum sample size essential for testing the \nstudy hypothesis. The results indicated that to achieve 80% power for detecting \na medium effect, at a significance level of α = 0.05, a total sample \nsize of N = 269 was necessary, equating to 54 individuals per group. Hence, \nconsidering insights from previous studies and the power analysis calculation, it \nwas recommended to conduct the study with a sample size ranging from 15 to 54 \nparticipants per group. Ultimately, the study succeeded in enrolling a total of \n135 participants, which were 30 participants in four groups and 15 participants \nin one group.\nParticipants were recruited between May 2021 and December 2023 from Orofacial \nPain Clinic, King’s College Dental Institute and Headache Clinic at St Thomas’ \nHospital. Ethical approval for the study was obtained from the Health Research \nAuthority (Approval No: 21/WS/0050). Informed consent was mandatory prior to \nparticipation.\nParticipants were recruited into five groups based on history and clinical \nexamination:\n● TMD: chronic myogenous painful TMD patients. TMD diagnosis was based \non the 1st edition of International Classification of Orofacial Pain, 1st edition \n(ICOP) for primary myofascial orofacial pain [ 26 ] and TMD pain presented for more \nthan 3 months.\n● Migraine (MG): chronic migraine patients. Chronic migraine diagnosis \nfollowed the 3rd edition of the International Classification of Headache \nDisorders (ICHD-3) [ 19 ].\n● (TMD + MG): chronic myogenous painful TMD patients with chronic \nmigraine.\n● (TMD + MG + FM): chronic myogenous painful TMD pain patients with \ncomorbid migraine and fibromyalgia. Fibromyalgia diagnosis aligned with the \nAmerican College of Rheumatology (ACR) criteria [ 22 ].\n● Healthy participants.\nThe study was inclusive of individuals aged between 18 and 50 years complying \nwith specific inclusion and exclusion criteria. This selection aimed to mitigate \nthe influence of confounding factors related to the experimental protocol [ 3 ] and \nthe participant characteristics [ 27 ,  28 ,  29 ] in the context of the CPM paradigm.\n● Presence of systemic comorbidities and psychological disorders, for \nexample, fibromyalgia, systemic myofascial pain and chronic fatigue syndrome, \nchronic headaches, migraines, heart arrhythmias, endometriosis, interstitial \ncystitis, irritable bowel syndrome, lower back pain, autoimmune diseases, and \nsleep disorders or diabetes mellitus.\n● Diagnosis of medication overuse for headache, and use nonsteroidal \nanti-inflammatory drugs (NSAIDs) or paracetamol within 12 hours before the \nexperiment.\n● Smoking more than five cigarettes daily.\n● Consuming over 6 cups of caffeinated drinks daily.\n● Indications of substance abuse.\n● Alcohol consumption within 24 hours before the experiment.\n● Irregular menstrual cycles in the case of female participants.\nThe study was divided into three parts, as follows:\nPart 1: Mechanical detection threshold (MDT) and pressure pain threshold (PPT).\nPart 2: Mechanical Temporal Summation (MTS) and decay of after-sensations.\nPart 3: Conditioned Pain Modulation Protocol (CPM).\nTo measure mechanical detection threshold (MDT), von Frey Filaments \n(Semmes-Weinstein) were administered on both the painful site and the hands. The \nprocedure began with the application of a filament with an initial force of 0.008 \ngrammes (0.08 mN). With eyes closed, participants were asked to indicate when \nthey first sensed the touch. The MDT with von Frey Filaments is a standard tool \nfor assessing tactile and mechanical pain thresholds. High test-retest \nreliability has been reported (intraclass correlation coefficients (ICC) \n0.76–0.98), though some studies note only moderate reproducibility (Kappa \n<0.6), influenced by examiner training, site, and protocol [ 30 ,  31 ].\nSubsequently, the pressure pain threshold (PPT) was assessed by a pressure \nalgometer (Pressure algometer, Wagner Pain Test™, Wagner \nInstruments, Greenwich, CT, USA) at the same locations. The PPT by algometry \nshowed high intra- and inter-rater reliability, with excellent concurrent \nvalidity between digital and analogue devices (ICC 0.82–0.99) [ 32 ]. Pressure was \nmanually increased at a rate of 30 kPa per second until participants reported \nfeeling the first sensation of pain. Each stimulation was repeated three times \nand then the average force was calculated.\nThe MTS was performed to evaluate pain facilitation (Fig.  1 ). The test was \nconducted according to the Deutscher Forschungsverbund Neuropathischer Schmerz \n(DFNS) standardised protocol [ 33 ] by using weighted pinprick stimulators \n(Pinprick Stimulator, MRC Systems GmbH, Heidelberg, BW, Germany). Employing the \nrecommended force of 256 mN for the hand [ 34 ], a pinprick stimulus was applied \nover the volar part of the forearm of the dominant hand. The stimulator was \npositioned vertically, perpendicular to the testing surface, and a single \npinprick stimulus was administered. This stimulus was repeated three times, each \nwith a 10-second interval. After each pinprick test, patients were asked to rate \nthe painful sensation on a numerical pain scale (NPS) of 0 to 100, where 0 \nrepresents no pain and 100 signifies the maximum imaginable pain. After a single \npinprick test, this was followed by a train of 10 successive stimuli, 3 cycles of \nthe same force and repeated at a 1/s rate (1 Hertz). The repetition of stimuli \nwas concentrated within a confined area of 1 cm 2 . After every 10 stimuli, \npatients were asked to rate the degree of painful sensation on a numerical pain \nscale. Subsequently, the wind-up ratio (WUR) was computed by dividing the mean \nrating of the three series of 10 stimuli by the mean rating of the three \nindividual stimuli. The MTS has demonstrated good to excellent \ntest-retest reliability in both healthy and clinical populations. For example, \nICC values range from 0.80–0.91 in healthy subjects (hand/back sites) using 2–3 \nrepeated measurements [ 35 ], 0.73–0.89 in low back pain patients [ 36 ], and \n0.63–0.86 in shoulder pain cohorts [ 37 ].\nOverview of mechanical temporal summation and conditioned pain \nmodulation protocol.\nEach participant underwent the CPM paradigm (Fig.  1 ); the pinprick stimulator \nused in the MTS test served as a test stimulus in the CPM test, while cold water \nserved as the conditioning stimulus (CS). This specific cold water conditioning \nstimulus aligned with the recommendations from the CPM guidelines [ 38 ]. The \nselection of a 10 °C intensity followed a precedent set by a previous \nprotocol and has been validated as effectively inducing a CPM effect through \ncorroborating studies [ 39 ,  40 ]. Cold water was prepared within an insulated ice \nbucket, and its temperature was measured with a digital thermometer. Participants \nwere requested to immerse their nondominant hand in cold water at wrist level. \nBlood pressure and pulse rate were measured at the 20-second mark. At the \n30-second point during the immersion, the MTS procedure was applied using the \nsame protocol as previously described, but targeting a different area on the same \nhand. CPM magnitude was calculated as the difference of pain intensity report \nduring the hand immersion in cold water and pain intensity report without hand \nimmersion from the MTS test (CPM = MTS with conditioning pain score − MTS without \nconditioning pain score). A negative CPM value indicated a decrease in pain \nresponse to the TS when a simultaneous CS is applied. The % CPM effect was \ncalculated by determining the percentage reduction in pain perception caused by \nthe conditioning stimulus compared to the baseline stimulus. The formula used \nwas: % CPM effect = ((baseline pain − conditioning stimulus pain)/baseline pain) × 100. A higher % CPM effect \nindicated a stronger pain modulation response, meaning the body is better at \ninhibiting pain perception in response to the conditioning stimulus. For the \nvalidity and reliability of CPM, a meta-analysis reported good intra-session \nreliability (ICC 0.64–0.77 in healthy subjects; ICC 0.77 in patients), but only \nfair inter-session reliability (ICC 0.44–0.59), depending on test and \nconditioning stimuli [ 41 ].\nTo determine an efficient CPM, two approaches were used to determine the % CPM \nthreshold. Firstly, a review conducted by Pud  et al . [ 42 ] revealed that \nthe expected reduction in pain sensitivity corresponding to the Diffuse Noxious \nInhibitory Control (DNIC) effect, averaged approximately 29%. This finding \nimplied that when participants experienced one pain-inducing stimulus, they \nsubsequently experienced another pain-inducing stimulus to a lesser degree. As a \nresult, efficient conditioned pain modulation was characterised as the ability of \nindividuals to suppress at least 29% of pain and this cutoff was applied in a \nlater study [ 43 ]. Secondly, the method proposed by Locke  et al.  [ 44 ] was adopted and further elaborated in the section on pain modulation profile (PMP) \nclassification.\nThe classification of participants into different categories of PMP was achieved \nby applying the cut-off thresholds of WUR and CPM [ 45 ]. PMP refers to different \nways in which an individual’s body responds to and regulates pain signals [ 46 ]. \nIn the context of WUR, the cut-off values were determined by adjusting the method \ndescribed by Vaegter and Graven-Nielsen [ 45 ]. This adjustment process involved \nthe inclusion of normative temporal summation data obtained from a cohort of 110 \nhealthy women. Individual WUR values that exceeded the upper limit (95% \nconfidence interval) of the normative WUR data for the hand (WUR >2.57) were \nidentified as elevated. For the establishment of CPM cut-off criteria, a \nmethodology described by Locke  et al . [ 44 ] was implemented. This \napproach integrated temporal summation (TS) data collected across three \nexperimental cycles, alongside the CPM data. A significant CPM effect was \nrecognised when the percentage increase in TS from the baseline exceeded the \ninherent measurement error. The calculation of the measurement error involved a \nseries of steps. Firstly, the intraclass correlation coefficient (ICC) model 3.3 \nfor mean TS (both single and serial stimulation) within each group (5 groups) \nfrom the three experimental cycles. Then, computed the standard error of \nmeasurement (SEM) for each visit, using the formula: SEM = SD(TS) × \nsquare root of (1 − ICC), and the sum of the SEM with the mean TS for each \nstimulation. Next, we converted this value into a percentage change relative to \nthe mean TS. Finally, we averaged these three TS + SEM relative change percentage \nvalues. Any CPM value above the inherent measurement error indicated a normal or \ngreater effect. By applying those WUR and CPM effect cut-off values, we therefore \nclassified participants into four different PMP categorizations: PMP I: Double \npro-nociception (increased TSP/impaired CPM), PMP II: Inhibitory pro-nociception \n(normal TSP/impaired CPM), PMP III: Facilitatory pro-nociception (increased \nTSP/normal CPM) and PMP IV: Antinociception (normal TSP/normal CPM).\nAll manual data obtained was entered into a secure electronic database. \nCategorical variables were analysed using the chi-square test. Continuous \noutcomes were assessed for the normality, and reported as mean ± standard \ndeviation (SD) with 95% confidence interval. Mean differences among 5 groups \nwere compared using analysis of variance (ANOVA) if data was normally distributed \nand Kruskal-Wallis H test for non-normally distributed data. The significance \nlevel was set at  p   < 0.05. Additionally, the  post hoc  test \nwas conducted when the initial analysis demonstrated statistical significance.\n\nOf the 152 participants, 13 were excluded for specific reasons: 4 had taken \nNSAIDs within 5 hours before the study, 4 had consumed alcohol within the \nprevious 12 hours, 3 had smoked cigarettes before the experiment, and 2 were \nunable to tolerate the water temperature. Therefore, the final number of \nparticipants included in the study was 139. The clinical characteristics of \nparticipants are shown in  Table 1 . Among them, 99 were female and 40 were male, \nwith a mean age of 42.13 years, SD 10.81. Participants had a mean BMI of 23.15, \nSD 3.58. Most participants (96%) were right-handed. The CPM protocol was \nperformed predominantly between 13:00 and 17:00. The average temperature of the \ncold water used in the test was 9.96 °C with an SD of 0.42 °C. \nNo significant differences were observed across the various characteristics of \nthe participants. When examining cases, the TMD + MG + FM group gave the highest \nmean clinical pain score of 60 out of 100.\nGeneral characteristics of study participants.\nTMD: TMD patients; TMD + MG: TMD patients with chronic migraine; MG: Chronic \nmigraine patients; TMD + MG + FM: TMD patients with comorbid migraine and \nfibromyalgia. \nSD: Standard Deviation; TMD: Temporomandibular disorders; MG: \nmigraine; FM: fibromyalgia; N/A: Not Applicable.\nMDT and PPT values and differences were documented in Fig.  2 a,b. Within the \nparticipant pool, individuals with migraine (MG) had the lowest mean threshold of \n0.45 grammes on the face, while the TMD + MG + FM group had the lowest mean \nthreshold of 0.68 grammes on the hand. In contrast, the remaining groups \nexhibited relatively similar mean MDT values for both facial and hand areas. \nSignificant differences of MDT were detected when accessing the painful region, \ndifferentiating between the MG group from the healthy controls, the TMD, as well \nas the TMD + MG groups. In terms of the PPT test, the MG group consistently \nreported the lowest threshold on the face (1.35 kPa), whereas the control group \nshowcased the highest PPT (1.59 kPa). Nevertheless, distinctions were also \napparent in testing on the face between when comparing control group with the MG, \nTMD + MG, and TMD + MG + FM groups. Interestingly, although the TMD + MG + FM \ngroup registered the lowest PPT on the hand, no significant differences were \nobserved across all five groups.\nThe TMD + MG + FM group had the most elevated pain scores of 21 and 43 out of \n100 in the single and serial tests, respectively. In contrast, the healthy \ncontrol group reported the lowest pain scores in both assessments, 15 and 32 out \nof 100, respectively ( Table 2 ). No significant differences were \nfound between the participants when the TS test was examined with a single \nstimulus. However, with serial stimulation, there were notable differences \nbetween the groups. The mean WUR from the TS tests was presented in  Table 2  and \nFig.  2 c, but the result showed no significant differences across the five study \ngroups.\nPain rating scores and outcomes during temporal summation and \nconditioned pain modulation protocol.\n*: Statistical difference ( p   < 0.05); TMD: TMD patients; MG: Chronic \nmigraine patients; TMD + MG: TMD patients with chronic migraine; TMD + MG + FM: \nTMD patients with comorbid migraine and fibromyalgia.\nNRS: Numeric Rating Scale; SD: Standard Deviation; CPM: conditioned pain modulation; TMD: Temporomandibular \ndisorders; MG: migraine; FM: fibromyalgia.\nThe pain scores obtained from the CPM experiment were detailed in both  Table 2  \nand Fig.  2 d. The control group reported the highest pain scores, and these scores \ndiffered significantly from the other four groups for both single and serial \nstimulation, 8 and 21 out of 100, respectively. In the context of within-group \nanalysis ( Table 2 ), the presence of pain inhibition (where the pain score during \nCPM deviated from the baseline pain score in MTS) was evident in all groups for \nboth types of stimulation, except for TMD + MG + FM in both stimulations.\nThe CPM magnitude values for both the single and serial tests were presented in \n Table 2  and Fig.  2 e. No noticeable distinction in CPM magnitude emerged for the \nsingle stimuli. However, in the context of serial stimulation, the control group \nshowed the lowest CPM magnitude (−11), which was significantly different from the \nMG, TMD + MG and TMD + MG + FM groups. Furthermore, apart from the healthy \nparticipants, the TMD group demonstrated the second lowest CPM magnitude (−8) \nduring serial stimulation, followed by the MG (−5), TMD + MG (−4) and TMD + MG + \nFM (−4) groups.\nThe pain score acquired from the MTS and the CPM were used to calculate the \npercentage change in CPM effect. The CPM effect was reversed from CPM magnitude \nwhich meant that higher CPM effect (more positive percentage) indicated higher \npain inhibition ability, while lower CPM percentage indicated lower pain \ninhibition ability. The mean percentage change within each group was reported in \n Table 2  and the significant differences among groups were shown in Fig.  2 f. In \nthe single stimulation, CPM effect in healthy individuals was higher and \ndifferent from other groups of patients. For the serial stimulation, all the pain \npatients, except for TMD patients, exhibited lower CPM effect than controls and \nthe difference was also observed between TMD and TMD + MG groups. The frequency \nof participants who had efficient CPM (the ability of individuals to inhibit at \nleast 29% of pain) were presented in  Table 2  and Fig.  3 . It illustrated that the \nlargest proportion of participants with effective CPM responses were found among \nthe healthy samples, accounting for 70% and 60% for the single and serial \ntests, respectively. In contrast, the lowest proportion of effective CPM \nresponders was observed in the TMD + MG + FM group, with rates of 36% and 21% \nfor the single and serial tests, respectively.\nComparison of (a) mechanical detection threshold (MDT), (b) \npressure pain threshold (PPT) among participants, (c) pain score during \nmechanical temporal summation (TS), (d) pain score during conditioned pain \nmodulation (CPM), (e) conditioned pain modulation magnitude, (f) percentage of \nCPM effect change. Controls (black), TMD Patients (orange), Migraine Patients \n(green), TMD Patients with Comorbid Migraine (blue), TMD Patients with Comorbid \nMigraine and Fibromyalgia (grey). SD: standard deviation; TMD: temporomandibular \ndisorder; MG: migraine; FM: fibromyalgia.\nPercentage of participants with an effective conditioned pain \nmodulation (CPM) response, a minimum of 29% pain inhibition. TMD: \ntemporomandibular disorder; MG: migraine; FM: fibromyalgia.\nThe correlation analysis examined the relationship between the number of \ncomorbidities and the CPM effect. In this study, we have analysed by selecting \nthree groups: the TMD group had no comorbid pain, the TMD + MG group experienced \none comorbid pain, and the TMD + MG + FM group faced two comorbid pain \nconditions. Using Pearson Correlation, we found a significantly negative \ncorrelation between the number of comorbidities and the CPM effect during single \nstimulation. The correlation coefficient ( r ) was −0.312 with  p  \n< 0.001. A similar negative correlation was observed during serial stimulation, \nwhere an  r  was −0.344 with  p   < 0.001, suggesting a meaningful \nyet mildly negative trend.\nThe cutoff value of 2.57 for WUR and the corresponding value for CPM were \nemployed to classify participants into four pain modulation profiles. In the \ncontext of CPM, the threshold values were derived from TS relative change from \nthe baseline for each stimulation. For single stimulation, CPM cut-off was \n16.20% for the control group, 8.48% for the TMD group, 12.35% for the MG \ngroup, 9.50% for the TMD + MG group, and 10.73% for the TMD + MG + FM group. \nThe serial stimulation had CPM cut-off points at 10.52% for the control group, \n12.87% for the TMD group, 12.75% for the MG group, 14.43% for the TMD + MG \ngroup, and 12.77% for the TMD + MG + FM group. Consequently, the distribution of \nPMP classifications within each group was visually presented in Fig.  4 . When \nanalysing the distribution of PMP classification among the five groups, there was \na significant difference among them during serial stimulation (χ 2  = \n28.85,  p  = 0.004), whereas such differences were not present in single \nstimulation (χ 2  = 8.96,  p  = 0.760). Nevertheless, employing \nthe Bonferroni correction in the context of pairwise comparisons within each \ncategory did not provide the capability to identify which specific categories \ndiffer from each other.\nDistribution of pain modulation profile (PMP) in single and \nserial stimulations across groups: controls (a), TMD Patients (b), Migraine \nPatients (c), TMD Patients with Comorbid Migraine (d), TMD Patients with Comorbid \nMigraine and Fibromyalgia (e). TMD: temporomandibular disorder; MG: migraine; FM: \nfibromyalgia; CPM: conditioned pain modulation; TSP: temporal summation of pain.\n\nThis study investigated EPM in individuals with TMD, migraine, and fibromyalgia \nin comparison to a control group of healthy individuals, as well as the \ncorrelation between the CPM responses and the number of comorbidities. A total of \n139 participants were predominantly female, right-handed, with an average age of \n42.13 years and a mean BMI of 23.15. Although there were no significant \ndifferences in the characteristics of the participants, there is evidence that \ngender and age have an impact on EPM. Pain facilitation was more prominent in \nfemales and those of advanced age [ 47 ]. EPM is less efficient among females [ 48 ] \nand older individuals [ 49 ]. Efforts were made to adjust for age, but due to the \npredominantly female composition of the patient population, adjusting for gender \nwas challenging. The study prioritized a larger, more diverse group of \nparticipants while maintaining control over other factors, such as the timing of \nthe test and water temperature. Water temperature was closely monitored using a \ndigital thermometer. The simple design of the CPM setup suggests potential for \npractical use in clinical settings with limited equipment, but the reliability of \nthe results should be considered when interpreting findings.\nThe MDT and PPT tests on the hand showed no significant differences between the \ngroups, but the face told a different story. Migraine patients were most \nsensitive to mechanical stimuli in the masseter muscle compared with other \ngroups, except for TMD + MG + FM. Furthermore, the MG group consistently reported \nthe lowest PPT on the face, with differences evident when comparing the control \ngroup. These findings contradict a prior study comparing headache with other pain \nconditions (TMD, FM, lower back pain, and irritable bowel syndrome) and reported \nthe greatest degree of increased pain sensitivity in TMD and FM [ 50 ]. However, \nthe headache population under the aforementioned study extended beyond \nmigraineurs. Despite our discrepant findings, the rationale underlying the \ngreatest increased sensitivity to pain in migraine compared with other pain \nconditions remains elusive [ 51 ], and we regrettably lacked a definitive \nexplanation for this phenomenon. However, as the number of pain comorbidities \nincreased, as in the TMD + MG and TMD + MG + FM groups, there was a discernible \npattern of lower MDT and PPT. This observation suggests that the combination of \nthese conditions may have a synergistic effect on pain sensitivity, a phenomenon \nconsistent with a previous study [ 50 ]. This may underscore the possible interplay \nbetween multiple pain conditions and their collective effects on pain \nsensitivity.\nThe results from the MTS revealed interesting patterns in different stimulation \nscenarios. The test involved a single stimulus; there were no significant \ndifferences observed among the participants. However, in serial stimulation, \nnotable differences emerged among groups. MTS was found to be higher in \nindividuals with pain conditions when compared with healthy participants, except \nfor migraine patients. These MTS results were consistent with some previous \nstudies involving TSP using pinprick stimuli in non-trigeminal areas among TMD \nindividuals compared with control subjects. Four studies reported an increase in \nMTS over the hand [ 13 ,  16 ], trapezius [ 16 ], and leg [ 52 ], while another study \nfound no increase in MTS over the trapezius and hand areas [ 53 ]. The presence of \nmigraine did not seem to significantly influence MTS responses in TMD patients. \nHowever, when fibromyalgia was present, as seen in the TMD + MG + FM group, it \nled to TS differences when compared with the TMD, TMD + MG and MG groups. This \nphenomenon could be attributed to the manifestation of widespread bodily pain and \ncentral sensitization associated with the presence of fibromyalgia. Additionally, \nthe mean WUR did not show significant differences across the five study groups. \nThis finding is consistent with the recent study that showed no WUR difference \namong TMD patients, TMD patients with migraine, and patients with headaches \nsecondary to TMD [ 54 ]. However, another study yielded contrasting results [ 13 ]. \nThis observation lends support to the idea that temporal summation might be an \ninconsistent phenomenon or that the assessment methods used may not effectively \ncapture the phenomenon of central sensitization [ 55 ].\nIn the evaluation of the CPM experiment, the difference in the change from the \nbaseline in the CS pain score serves as an indicator of pain inhibition within \neach group. Subsequently, the alteration in the CPM magnitude was converted into \na percentage CPM effect to enable a comparison of the effectiveness of EPM among \ngroups. The higher the CPM effect, the more efficient the EPM. In general, \nindividuals with chronic pain conditions, particularly those in the MG, TMD + MG \nand TMD + MG + FM groups, displayed lower CPM responses to dynamic stimulation \nthan healthy participants. The TMD group exhibited less CPM effect only with the \nsingle stimulation. This suggests that the ability to modulate pain in response \nto single and repeated stimuli is less effective in MG and TMD patients with \ncomorbid pain compared with healthy participants. On the other hand, TMD-only \npatients showed a lower CPM effect compared with pain-free subjects with only \nsingle stimulation. When considering the threshold for efficient CPM, defined as \nthe ability to inhibit at least 29% of pain [ 42 ], healthy participants had the \nhighest proportion of efficient CPM responses. In contrast, the TMD + MG + FM \ngroups had the lowest proportion of effective CPM responses. Altogether, our \nresults showed that EPM was impaired in TMD patients with comorbid pain \nconditions and the combination of pain conditions shows a trend to reduce the \nefficacy of CPM.\nTo date, few reports have found an effect of CPM in relation to the combination \nof pain conditions and multiple studies have provided results on the individual \npain condition in relation to EPM and CPM response. In the TMD population, the \nevidence regarding impaired EPM is somewhat mixed. Three studies reported \nimpaired EPM, specifically in the trigeminal region but not in non-trigeminal \nregions [ 13 ,  18 ,  56 ]. Conversely, another study presented contrasting results, \nfinding impaired EPM in both extra-segmental and intra-segmental areas in TMD \npatients [ 57 ]. Additionally, in one study, no EPM effect was observed in any of \nthese sites [ 58 ]. Furthermore, when examining only the outside trigeminal area, \nthree studies were conducted. Among these, an impairment of EPM was found in TMD \npatients when compared with a pain-free population in two studies [ 15 ,  59 ], while \nin one study, no such difference was found [ 10 ]. Most of these studies have \nprovided evidence supporting intact EPM in TMD patients in non-trigeminal areas. \nIn other words, CPM between TMD patients and healthy individuals was not \ndifferent. Our results in the TMD-only group align with these findings. However, \nwhen concurrent pain conditions are present in the TMD population, it may impact \ntheir ability to engage in EPM, resulting in reduced CPM compared with healthy \nindividuals, as demonstrated by our study results. The lack of significant CPM \ndifferences among TMD patients is consistent with a previous study that examined \nCPM in TMD patients with migraine and headache attributed to TMD but not in \ncontrols [ 54 ]. Furthermore, the analysis using Pearson’s correlation revealed a \nnegative correlation between the number of comorbidities and the efficiency of \nCPM, for both single and serial stimulations. This implies that as the number of \ncomorbid pain conditions increases, the ability of individuals to effectively \ninhibit pain decreases. Therefore, the ability to modulate pain might be an \nimportant factor to consider when managing TMD patients with co-existing pain \ndisorders.\nThe distribution of PMP showed a significant difference among the five groups \nfor serial stimulation but not for single stimulation. Interestingly, in both \nserial and single stimulation, the percentage of participants with either \nincreased TSP or impaired EPM (PMP I, II, III) were higher than that in TMD-only \npatients and control subjects. However, applying the Bonferroni correction for \npairwise comparisons within each category made it challenging to pinpoint \nspecific group differences. A limited number of studies have reported the use of \nthe PMP classification in chronic pain patients [ 13 ,  45 ] and found that a small \npercentage of patients had no impaired EPM and no difference in PMP \nclassification between TMD patients and pain-free subjects [ 13 ]. This scarcity of \nusage may stem from its perceived lack of refinement in effectively \ndistinguishing pain patients based on their EPM abilities. However, this study \nsuggests that the PMP categorization approach may be valuable, especially in \ndynamic stimulation paradigms that showed significant pain response differences \nbetween chronic pain patients and healthy controls. This underscores the \npromising applicability of the PMP classification and suggests further \nopportunities for the development of predictive models of pain onset and \npersonalised pain treatment strategies based on the PMP classification [ 60 ].\nIn terms of the different effects of stimulus modality, previous literature did \nnot extensively explore differences in MTS and CPM outcomes between single \n(static) stimuli and serial (dynamic) stimuli. Our findings suggested that the \ncumulative effect of repeated painful stimuli, as seen in the serial test, might \nhave a more pronounced impact on pain modulation. The result was similar to a \nstudy comparing different stimulus modalities for MTS and CPM, including heat and \npressure, as well as single and serial stimuli. This prior study suggested that \nCPM was significantly more influenced by serial stimuli in MTS than by single \nstimuli in MTS [ 61 ]. However, it is crucial to acknowledge that CPM responses as \nwell are highly dependent on the experimental design and can be influenced by \nconsiderations of its reliability [ 41 ]. The use of distinct stimuli or even \nvariations in the intensity of the same stimulus can potentially impact the \nresults of CPM assessments [ 62 ].\nThe discovery of reduced efficacy of CPM in patients suffering from various pain \nsyndromes requires an explanation. This situation theoretically raises a dilemma \nresembling a “chicken and egg” scenario [ 60 ]. It suggests that either the \npatients originally had normal CPM efficiency, but their pain inhibition capacity \nhad been depleted due to the persistent chronic pain, rendering them incapable of \neffectively reducing pain within the CPM protocol setting, or the patients \ninitially had a less efficient CPM. CPM plays an important role in both pain \nresearch and clinical practice, enabling the assessment of a patient’s pain \nmodulation system, predicting treatment outcomes, and guiding tailored pain \nmanagement [ 60 ]. However, challenges, such as individual variability and lack of \nstandardised protocols, need to be addressed to maximise its clinical utility in \nthe treatment of chronic pain.\nThis study has several limitations to be acknowledged. Firstly, the CPM protocol \nused lacks standardisation for specific conditions, so it was adapted for TMD \nbased on tailored evidence. Further research on CPM validation and variability is \nneeded. Secondly, precise control of water temperature was critical. \nUnfortunately, due to limited access to equipment, we had to use an alternative \ninsulated cooling container with a digital thermometer. These adaptations were \ntested and validated prior to the study. Thirdly, the population of TMD patients \nwith comorbid fibromyalgia and migraine was small due to low prevalence and time \nconstraints. Moreover, medications beyond NSAIDs, paracetamol, or psychiatric \ntreatments may not have been fully excluded. While these agents could potentially \ninfluence pain modulation, their impact is likely less immediate compared with \nshort half-life analgesics, and was therefore not specifically controlled for in \nthe study design. Lastly, we did not include a TMD + FM group for comparison, \npreferring the TMD + MG + FM group to assess the impact of comorbidities within \nthe study’s timeframe. Hence, increasing the sample size for the TMD + MG + FM \ngroup and incorporating the TMD + FM group might have potentially modified the \nresults, rendering them more significant.\n\nThe coexistence of comorbid pain conditions, such as migraine and fibromyalgia, \nin TMD patients is correlated with a reduced capacity for endogenous pain \nmodulation. This suggests that TMD patients struggling with concurrent pain \ndisorders may require a more thorough assessment of their pain modulation \nsystems. Targeted treatment strategies that systematically address these \ninterrelated pain conditions might be crucial in improving pain management \noverall. This highlights the importance of a comprehensive treatment approach for \nindividuals with multiple pain-related comorbidities.","source_license":"CC-BY-4.0","license_restricted":false}