Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2025 Mar 27;31(8):2596–2607. doi: 10.1111/odi.15331

Statistical Validation of Unsupervised Clustering for Adolescent TMD: A Cross‐Sectional Study

Hye Kyoung Kim 1,✉
PMCID: PMC12423482  PMID: 40145361

ABSTRACT

Objective

This study employs unsupervised clustering to identify Temporomandibular disorders (TMD) phenotypes in adolescents, aiming to identify distinct clusters based on biopsychosocial features. It compares these clusters with conventional TMD classifications to assess if this method offers enhanced insights into TMD diagnosis.

Methods

Data from 662 adolescent patients with TMD were analyzed using unsupervised clustering and classified into four groups based on DC/TMD Axis I: 1 (disc displacement), 2 (joint pain), 3 (muscle pain), and 4 (combined joint and muscle pain). Patient‐reported outcomes were measured with instruments including the Brief Pain Inventory, the Pain Catastrophizing Scale, the Symptom Checklist‐90‐Revised, and the Pittsburgh Sleep Quality Index. Statistical analyses validated the clusters against conventional classifications.

Results

Three distinct clusters were identified: High Impact (n = 70), Mild Symptoms (n = 423), and High Catastrophizing (n = 169), each displaying unique patterns in pain severity, pain catastrophizing, psychological distress, and sleep disturbances. Multinomial logistic regression of conventional TMD classifications revealed that only pain severity significantly differentiated the subcategories among these biopsychosocial factors.

Conclusions

The findings underscore the variability in TMD presentations among adolescents and suggest that integrating phenotyping into the conventional diagnostic approach could significantly improve diagnostic accuracy and treatment outcomes, facilitating better management of high‐risk adolescent patients.

Keywords: adolescent, biopsychosocial, diagnosis, phenotype, psychological, temporomandibular joint disorders

1. Introduction

Temporomandibular disorders (TMD) are a major cause of orofacial pain and functional limitations of the masticatory system, significantly affecting the quality of life in adolescents and adults (Pigozzi et al. 2021).

The prevalence of TMD in adolescents is similar to or lower than that in adults (Valesan et al. 2021; Rentsch et al. 2023). Among adults, TMD prevalence is approximately 31% (Valesan et al. 2021), whereas in adolescents, it ranges from 7.3% to 30%, depending on study methods and diagnostic criteria (Christidis et al. 2019; Minervini et al. 2023).

Although the clinical features of TMD in adolescents are generally comparable to those in adults, variations in manifestations exist. Both groups commonly experience symptoms such as headache, temporomandibular joint (TMJ) sounds, jaw pain, and myofascial pain (da Silva et al. 2016; Bertoli et al. 2018). Adolescents are frequently diagnosed with disc displacement with reduction (Christidis et al. 2019; Bertoli et al. 2018), whereas adults also present with conditions like arthralgia, degenerative joint disorders, and ear symptoms, reflecting more pain‐related TMD presentations (Ujin Yap et al. 2021; de Paiva Bertoli et al. 2018). Psychosocial factors, including stress, depression, and anxiety, significantly impact TMD symptoms and severity in adolescents, mirroring findings in adults. These factors are key risk elements for TMD (Ujin Yap et al. 2021; Ekberg et al. 2023; Schiffman et al. 2014).

These psychological factors complicate the diagnosis and management of TMD but are crucial elements to assess the complexity and heterogeneity of the disorder. Current diagnostic and treatment protocols for TMD, originally designed for adults, are often applied to adolescents without considering their unique developmental needs. Adolescence, defined by the World Health Organization as ages 10–19 (World Health Organization 2024), represents a critical phase characterized by unique emotional, physical, and social developments, including heightened emotional reactivity (Jaworska and MacQueen 2015). Evidence indicates that TMD pain in adolescence substantially increases the risk of chronic TMD pain in young adulthood (Nilsson and List 2020). Early recognition and intervention in high‐risk adolescents, addressing comorbid psychosocial factors, may prevent chronic pain progression (Nilsson and List 2020; Gatchel et al. 2006). Thus, there is a pressing need for tailored assessments and interventions specific to adolescent TMD. Accurate diagnosis and identification of high‐risk subgroups within this population are essential.

The Diagnostic Criteria for Temporomandibular Disorders (DC/TMD) is widely used for adult patients with TMD, addressing physical (Axis I) and psychosocial (Axis II) dimensions (Schiffman et al. 2014). Conventional diagnoses based on DC/TMD Axis I include intra‐articular TMD (e.g., disc displacement and degenerative joint diseases), arthralgia, myalgia, myofascial pain with referral, and headache attributed to TMD. Despite its utility, diagnosing TMD remains challenging due to its multifactorial nature and the variability of symptoms across patients. Conventional diagnostic classifications often fail to capture the full spectrum of patients' pain experiences, potentially resulting in suboptimal management strategies.

In clinical practice, patients with identical diagnoses, such as disc displacement, arthralgia, or myalgia, frequently present with varying psychosocial symptoms, resulting in diverse pain experiences and behaviors. These range from straightforward cases to those resistant to treatment. As a result, a diagnosis limited to Axis I may not adequately capture the complexity of a patient's condition. This highlights the potential for an Axis II‐based TMD phenotype to more accurately reflect clinical symptoms and guide treatment planning. In efforts to refine diagnostic criteria for adolescent TMD, recently proposed adaptations of the DC/TMD protocol emphasize a multidimensional assessment approach (Ekberg et al. 2023). International experts in TMD and pain psychology recommend integrating validated pain scales and assessments for psychological factors, including depression, anxiety, catastrophizing, and sleep disorders (Ekberg et al. 2023). These comprehensive evaluations are essential to understanding the interplay between physical symptoms and psychological factors that define adolescent TMD. By incorporating scales such as the brief pain inventory (BPI) for pain measurement (Cleeland and Ryan 1994), the pain catastrophizing scale (PCS) for assessing catastrophizing levels (Sullivan et al. 1995), the symptom checklist‐90 revised (SCL‐90R) for psychological distress (Derogatis 1977), and the Pittsburgh sleep quality index (PSQI) for subjective sleep quality (Buysse et al. 1989), this study aligns with the latest expert recommendations. These tools ensure a biopsychosocial understanding of TMD tailored to the adolescent population.

A PubMed search reveals a stark disparity in research volume on TMD between adolescents and adults, with significantly more studies on adult TMD. While some research explores phenotype clustering in adult TMD based on multifactorial biopsychosocial factors (Bair et al. 2016; Pimenta e Silva Machado et al. 2012), equivalent studies for adolescents remain scarce. This gap underscores the urgent need for focused research on how these factors influence TMD phenotypes in adolescents, which is crucial for developing targeted diagnostic and therapeutic strategies.

In this context, the study employs an unsupervised clustering algorithm using multidimensional pain‐related outcome measures in adolescent patients to identify clusters rooted in Axis II. The objective is to determine whether these clusters effectively differentiate patients' clinical symptom profiles. By comparing these clusters with conventional diagnostic classifications, the study aims to validate their potential in enhancing diagnostic precision and treatment efficacy. The hypothesis posits the existence of distinct phenotypic clusters among adolescent TMD patients identifiable through this approach.

2. Materials and Methods

2.1. Participants and Study Setting

A total of 733 consecutive adolescent patients visited the Orofacial Pain and Oral Medicine Clinic at Dankook University Dental Hospital, Korea, between March 2019 and November 2020 with primary complaints of TMD. The study, approved by the Dankook University Institutional Review Board (DKUDH IRB No. 2024‐06‐001) and conducted in accordance with the Declaration of Helsinki (version 2002), waived Informed consent due to its retrospective nature. Participants included adolescents aged 10–19 who completed questionnaires during their initial visit. Exclusion criteria comprised individuals undergoing invasive dental treatment or orthodontic treatment, those with recent jaw or facial trauma, a history of TMD surgery, or severe comorbid conditions such as leukemia. Patients on medications for psychiatric disorders, neurologic impairments, or systemic joint diseases were also excluded.

Adolescent patients with TMD were retrospectively analyzed based on medical records to identify subgroups reflecting variations in pain‐related outcomes. Diagnoses followed Axis I criteria of the DC/TMD (Schiffman et al. 2014). Patients were categorized into four groups based on clinical diagnoses: Group 1 (disc displacement), Group 2 (Joint disorders), Group 3 (muscle disorders), and Group 4 (combined joint and muscle disorders). Group 1 served as the control group, encompassing cases of disc displacement without arthralgia or degenerative joint disorders. Minor pain symptoms associated with disc displacement were also included in this group.

2.2. Assessment of Patient‐Reported Pain‐Related Outcome Measures

Upon initial clinical evaluation, patients completed a series of validated questionnaires to assess pain‐related outcomes. These included:

  • Brief Pain Inventory (BPI): This instrument evaluates both the severity of pain and its impact on daily functions (Cleeland and Ryan 1994). It features two primary scales: the pain severity scale, which measures the sensory dimension of pain at its worst, least, average, and current levels, and the pain interference scale, which assesses the reactive dimension of how pain affects general activity, mood, work, relationships, sleep, and enjoyment of life. Originally designed to measure walking disability, the scale has been adapted in this study to assess interference with chewing ability. This concise instrument utilizes an 11‐item numerical rating scale based on a 10‐point system.

  • Pain Catastrophizing Scale (PCS): This scale assesses the level of catastrophizing thoughts related to pain across three subdimensions: rumination, magnification, and helplessness (Sullivan et al. 1995). Scores from these subdimensions are combined to calculate a total score, providing a comprehensive measure of an individual's tendency to catastrophize in response to pain. It is a valuable tool for understanding how pain‐related thoughts, perceived as threatening, overwhelming, or unmanageable, influence the overall pain experience (Haythornthwaite et al. 2024).

  • Symptom Checklist‐90‐Revised (SCL‐90R): his diagnostic tool evaluates a wide range of psychological problems and symptoms across nine subdimensions: somatization, obsessive‐compulsive behavior, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety, paranoid ideation, and psychoticism (Derogatis 1977). Among its three global indices, the global severity index (GSI) provides a summary score of the severity of psychological symptoms, serving as a comprehensive indicator of an individual's mental health. The utility of the SCL‐90R has been extensively validated in populations with TMD (Manfredini et al. 2009; Preti et al. 2019; Kim and Kim 2023; Durá et al. 2006).

  • Pittsburgh Sleep Quality Index (PSQI): This instrument evaluates sleep quality and disturbances over a one‐month period (Buysse et al. 1989). It assesses seven components: subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleep medication, and daytime dysfunction. Each component is scored individually, contributing to a cumulative total score that reflects overall sleep quality. The PSQI is one of the most commonly used and validated instruments for assessing subjective sleep quality in patients with TMD (Roithmann et al. 2021).

2.3. Statistical Analysis

Statistical analyses were conducted using SPSS (IBM Statistics for Windows, Version 26.0, IBM Corp., Armonk, NY, USA) and R (Version 4.4.2, R Foundation for Statistical Computing). Continuous variables were tested for normality using the Kolmogorov–Smirnov test, which confirmed that these variables did not follow a normal distribution. Descriptive statistics, including mean, standard deviation, median, and interquartile range, were calculated for all continuous variables. Categorical variables were described using frequencies (n) and percentages (%). To explore factors influencing pain interference in adolescents with TMD, linear regression analysis was performed. Significant factors identified through regression analysis were then used to conduct K‐means clustering, an unsupervised learning technique aimed at categorizing patient samples into homogeneous subgroups based on patient‐reported pain‐related outcome measures. This method is particularly effective in uncovering inherent structures within the dataset by predefining the number k of clusters and iteratively adjusting the cluster centers while assigning data points to the nearest cluster. The initial cluster centers were randomly selected from the dataset, and the algorithm iteratively recalculated the cluster means and reassigned data points to the nearest centers based on Euclidean distance. This iterative process continued until the within‐cluster sum of squares (SSW) was minimized. The efficacy of the clustering was evaluated by examining the percentage of total variance explained by between‐cluster variability, providing a measure of the clustering's effectiveness in differentiating subgroups.

The statistical significance of differences between the identified clusters was assessed using ANOVA for each variable, followed by post hoc Tukey's HSD tests to explore specific inter‐cluster differences.

To examine the relationship between conventional TMD classification and patient‐reported pain‐related outcomes, a multinomial logistic regression analysis was conducted. Additionally, Chi‐square and Fisher's exact tests were performed to assess differences in subject distribution across TMD classifications and identified clusters, revealing a significant association. Significance levels were set at p < 0.05 for all statistical tests.

3. Results

3.1. Clinical Characteristics and Pain‐Related Outcomes

Out of a total of 773 subjects, 662 were included in the study. Exclusions comprised 2 individuals with leukemia, 1 with a mandible fracture, 3 undergoing orthodontic extractions, 2 with rheumatic diseases, and 5 on medication for psychiatric or neurological conditions. The remaining exclusions were due to incomplete questionnaires or missing data.

Table 1 details the clinical characteristics and pain‐related outcomes in the cohort of 662 adolescent patients with TMD. Among these, 64% (n = 424) were female, with an average age of 16.86 years (SD 1.83). The TMD classification categorized patients as follows: 43.4% (n = 282) with internal derangement, 22.9% (n = 152) with joint pain, 28.3% (n = 187) with muscle pain, and 5.4% (n = 36) with combined joint and muscle pain. The mean symptom duration was 12.83 months (SD 18.00). Pain severity (mean 2.35, SD 1.89) was higher than pain interference (mean 2.14, SD 1.99) as measured by the BPI (p = 0.046).

TABLE 1.

Clinical features and pain‐related outcome measures of adolescent temporomandibular disorders (n = 662).

Variables Mean ± SD Median (IQR)
Clinical features Sex (female), n (%) 424 (64%)
Age (years) 16.86 ± 1.83 17 [16.00–18.00]
TMD classification, n (%) 1:287 (43.4), 2:152 (22.9), 3:187 (28.3), 4:36 (5.4)
Pain duration (month) 12.83 ± 18.00 6 [1.50–12.00]
BPI Pain severity* 2.35 ± 1.89 2 [0.75–3.75]
Pain interference 2.14 ± 1.99 1.71 [0.43–3.43]
PCS Magnification 3.00 ± 2.78 2 [1.00–5.00]
Rumination 3.40 ± 3.93 2 [0.00–6.00]
Helplessness* 4.53 ± 4.22 4 [1.00–6.00]
Total score 10.94 ± 10.00 8 [3.00–16.00]
SCL‐90R Somatization* 45.61 ± 8.31 43 [40.00–49.00]
Obsessive‐compulsive 40.88 ± 9.76 39 [34.00–46.00]
Interpersonal sensitivity 40.45 ± 9.92 37 [34.00–45.00]
Depression 40.74 ± 8.36 38 [35.00–44.00]
Anxiety 42.10 ± 7.43 40 [37.00–45.00]
Hostility 43.01 ± 7.52 40 [38.00–45.00]
Phobic anxiety* 44.15 ± 6.32 41 [41.00–45.00]
Paranoid ideation 41.67 ± 7.71 38 [37.00–43.00]
Psychoticism 42.19 ± 6.89 39 [39.00–43.00]
Global Severity Index 40.58 ± 8.25 38 [35.00–44.00]
PSQI C1: Subjective sleep quality 1.11 ± 0.82 1 [1.00–2.00]
C2: Sleep latency 0.95 ± 1.04 1 [0.00–2.00]
C3: Sleep duration 0.95 ± 0.92 1 [0.00–1.00]
C4: Habitual sleep efficiency 1.07 ± 1.39 0 [0.00–3.00]
C5: Sleep disturbances 0.95 ± 0.61 1 [1.00–1.00]
C6: Use of sleeping medication 0.09 ± 0.47 0 [0.00–0.00]
C7: Daytime dysfunction* 1.58 ± 0.99 2 [1.00–2.00]
Total score 6.70 ± 3.28 6 [4.00–9.00]
Open in a new tab

Note: The TMD classification categorized TMD into four groups: group 1 with internal derangement, group 2 with joint pain, group 3 with muscle pain, and group 4 with combined joint and muscle pain. An independent t‐test for BPI and ANOVA for PCS, SCL‐90R, and PSQI were performed for group comparison. Tukey's HSD was performed for post hoc comparison following ANOVA.

Abbreviations: BPI, brief pain inventory; PCS, Pain Catastrophizing Scale; PSQI, Pittsburgh Sleep Quality Index; SCL‐90R, symptom checklist‐90‐revised.

*

Indicates p < 0.05.

PCS had a total mean score of 10.94 (SD 10.00), with subscale means for magnification, rumination, and helplessness at 3.00 (SD 2.78), 3.40 (SD 3.93), and 4.53 (SD 4.22), respectively. Post hoc comparisons indicated that the mean score for helplessness was significantly higher than the other subdimensions of the PCS (p < 0.05).

The GSI of the SCL‐90R had a mean score of 40.58 (SD = 8.25). Among the subdimensions of the SCL‐90R, somatization and phobic anxiety exhibited significantly higher scores compared to other dimensions (p < 0.05), indicating more pronounced symptoms in these areas among adolescent TMD patients. The global score of the PSQI revealed a mean score of 6.70 (SD = 3.28). Among the seven subdimensions of the PSQI, the daytime dysfunction component (C7) recorded the highest scores (p < 0.05).

3.2. Factors Influencing Pain Interference in Adolescent TMD Patients

Table 2 summarizes the linear regression analysis aimed at identifying factors influencing pain interference in adolescent patients with TMD. The revealed results demonstrated that pain severity (Estimate = 0.610, p < 0.0001), the total PCS score (Estimate = 0.049, p < 0.0001), the GSI score (Estimate = 0.023, p < 0.0001), and the global PSQI score (Estimate = 0.073, p < 0.0001) were significant contributors to increased pain interference. Conversely, variables such as sex, age, symptom duration, and TMD classification did not show significant effects. The model exhibited a strong fit, with an adjusted R 2 of 64.68% (p < 0.0001).

TABLE 2.

Factors influencing pain interference in adolescent patients with temporomandibular disorders.

Estimate SE t‐value p
Intercept −1.580 0.487 −3.247 0.001
Sex −0.068 0.096 −0.706 0.480
Age 0.027 0.026 1.043 0.297
Symptom duration −0.0002 0.0003 −0.084 0.933
TMD classification −0.048 0.050 −0.958 0.338
Pain severity*** 0.610 0.030 20.95 < 0.0001
PCS*** 0.049 0.006 8.59 < 0.0001
GSI*** 0.023 0.006 3.64 < 0.0001
PSQI*** 0.073 0.016 4.65 < 0.0001
Open in a new tab

Note: Linear regression analysis was conducted for the statistical analysis. The TMD classification categorized TMD into four groups: group 1 with internal derangement, group 2 with joint pain, group 3 with muscle pain, and group 4 with combined joint and muscle pain. F‐statistic = 152.3, p‐value < 0.0001, adjusted R 2 = 0.6468.

Abbreviations: GSI, Global Severity Index; PCS, Pain Catastrophizing Scale; PSQI: Pittsburgh Sleep Quality Index; SE: standard error.

***

Indicates p < 0.0001.

3.3. Comparison of Pain‐Related Outcome Measures Across the Three Clusters Identified Through Unsupervised K‐Means Clustering

The significant pain‐related factors from the regression analysis were further explored using an unsupervised K‐means clustering algorithm, which grouped the adolescent TMD patients into three distinct clusters, as depicted in Figure 1. In the cluster plot, Dim1 and Dim2, which account for 53.2% and 19.8% of the variance, respectively, represent principal components summarizing the primary patterns of variation in the dataset. Significant differences in patient‐reported pain‐related outcomes were observed across clusters (Table 3 and Figure 2).

FIGURE 1.

FIGURE 1

Open in a new tab

Cluster plot showing the three clusters of adolescent patients with TMD.

TABLE 3.

Comparison of pain‐related outcome measures among the three clusters.

Cluster 1 (n = 70) Cluster 2 (n = 423) Cluster 3 (n = 169) p TukeyHSD
Pain severity 3.21 ± 1.94 1.76 ± 1.58 3.47 ± 1.94 < 0.0001 1, 3 > 2
Pain interference 3.82 ± 2.46 1.38 ± 1.37 3.33 ± 2.08 1, 3 > 2
PCS 20.06 ± 11.60 5.01 ± 3.91 22.00 ± 7.34 3 > 1 > 2
GSI 59.01 ± 8.51 37.56 ± 4.47 40.49 ± 4.85 1 > 3 > 2
PSQI 9.97 ± 2.85 5.93 ± 2.88 7.28 ± 3.44 1 > 3 > 2
Open in a new tab

Note: Three clusters were identified using unsupervised K‐means clustering. Within‐cluster sum of squares: Cluster 1 = 15448.58, Cluster 2 = 20238.31, and Cluster 3 = 16358.58. Between_SS/total_SS = 57.8% where SS indicates sum of squares. Group differences between these clusters for the included variables were tested using ANOVA, with Tukey's HSD conducted for posthoc comparisons.

Abbreviations: GSI, Global Severity Index; PCS, Pain Catastrophizing Scale; PSQI, Pittsburgh Sleep Quality Index.

FIGURE 2.

FIGURE 2

Open in a new tab

Comparison of pain‐related outcome measures across clusters of adolescent patients with TMD. Boxplots depict differences across three identified clusters in (a) pain severity, (b) pain interference, (c) pain catastrophizing, (d) psychological distress (Global Severity Index), and (e) subjective sleep quality (Pittsburgh Sleep Quality Index, PSQI). Statistical comparisons indicate: (a, b) Cluster1, Cluster3 > Cluster2, (c) Cluster3 > Cluster1 > Cluster2, and (d, e) Cluster1 > Cluster3 > Cluster2.

Cluster 1 included 70 patients (10.6%) with the highest levels of pain severity (mean = 3.21, SD = 1.94), pain interference (mean = 3.82, SD = 2.46), and PCS scores (mean = 20.06, SD = 7.34). This group also had the highest GSI (mean = 59.01, SD = 8.51) and PSQI (mean = 9.97, SD = 2.85) scores, indicating severe psychological distress and poor sleep quality.

Cluster 2, the largest group with 423 patients (63.9%), exhibited the lowest scores across all measured variables. Pain severity (mean = 1.76, SD = 1.58) and pain interference (mean = 1.38, SD = 1.37) were significantly milder, and PCS (mean = 5.01, SD = 3.91), GSI (mean = 37.56, SD = 4.47), and PSQI (mean = 5.93, SD = 2.88) scores were notably lower than those in the other clusters.

Cluster 3 comprised 169 patients (25.5%) and was characterized by high pain severity (mean = 3.47, SD = 1.94) and pain interference (mean = 3.33, SD = 2.08), similar to Cluster 1. The PCS score (mean = 22.00, SD = 7.34) was the highest among all clusters, reflecting a very high level of pain catastrophizing cognition, while the GSI (mean = 40.49, SD = 4.85) and PSQI (mean = 7.28, SD = 3.44) scores were lower compared to those in Cluster 1.

The within‐cluster sum of squares values 15,448.58 for Cluster 1, 20,238.32 for Cluster 2, and 16,358.58 for Cluster 3 along with the between‐cluster/total sum of squares ratio of 57.8% indicate strong cluster distinctiveness.

Table 4 presents the regression coefficients from linear regression analyses conducted for each of the three clusters of adolescent TMD patients to predict pain interference. The results highlight the varying influences of pain severity, pain catastrophizing (PCS), GSI, and PSQI on pain interference across the clusters. Cluster 1 displayed the highest coefficient for pain severity (ß = 0.755238, p < 0.001) and significant contributions from PCS (ß = 0.047218, p < 0.05) and subjective sleep quality (ß = 0.154022, p < 0.05). This model explained a substantial proportion of variance (R 2 = 0.6416, p < 0.0001). GSI did not significantly influence pain interference in this cluster.

TABLE 4.

Regression coefficients for predicting pain interference in the three clusters.

Cluster 1 (n = 70) Cluster 2 (n = 423) Cluster 3 (n = 169) Order
Intercept −1.59489 −1.42391 −1.71198
Pain severity 0.755238*** 0.56468*** 0.61846*** 1 > 3 > 2
PCS 0.047218* 0.03463** 0.05873* 3 > 1 > 2
GSI 0.008537 0.03434** 0.02713
PSQI 0.154022* 0.0585*** 0.0698* 1 > 3 > 2
R2 0.6416 (p < 0.0001) 0.5418 (p < 0.001) 0.446 (p < 0.001)
Open in a new tab

Note: Linear regression analyses were conducted for each cluster to predict pain interference. For significant parameters, the order of the coefficients was displayed. Asterisks indicate the level of significance, with ***p < 0.001, **p < 0.01, *p < 0.05.

Abbreviations: GSI, Global Severity Index; PCS, Pain Catastrophizing Scale; PSQI, Pittsburgh Sleep Quality Index.

Cluster 2 showed statistically significant coefficients for all five pain‐related outcomes, with lower coefficients for pain severity (ß = 0.56468, p < 0.001), PCS (ß = 0.03463, p < 0.01), GSI (ß = 0.03434, p < 0.01), and PSQI (ß = 0.0585, p < 0.001). This model explained 54.18% of the variance in pain interference (p < 0.001). Cluster 3 demonstrated moderate influences from pain severity and the highest impact from PCS (ß = 0.05873, p < 0.05). Sleep quality's effect (ß = 0.0698, p < 0.05) was also significant. However, GSI did not significantly influence pain interference in this cluster. The model's R 2 was 0.446 (p < 0.001).

3.4. Multinomial Logistic Regression Analysis for Predicting the Conventional TMD Classification

The multinomial logistic regression analysis presented in Table 5 was conducted to evaluate the validity of conventional TMD classifications compared to the Axis II‐based clusters identified in this study. This analysis aims to determine how various pain‐related and psychological variables influence the likelihood of patients being classified into different TMD groups, providing a quantitative comparison between conventional classifications and empirical phenotypes derived from patient‐reported outcomes.

TABLE 5.

Multinomial logistic regression coefficients for predicting conventional TMD classification.

TMD group 2 TMD group 3 TMD group 4
Intercept −1.354 −1.589 −3.998
Pain severity 0.305*** 0.398*** 0.378***
Pain interference −0.110 −0.044 −0.207
PCS 0.010 0.010 0.023
GSI 0.011 0.011 0.022
PSQI −0.042 −0.034 0.048
Open in a new tab

Note: The TMD classification categorized TMD into four groups: group 1 with disc displacement (n = 287), group 2 with joint pain (n = 152), group 3 with muscle pain (n = 187), and group 4 with combined joint and muscle pain (n = 36). TMD group 1 serves as the reference group for comparison.

Abbreviations: GSI, Global Severity Index; PCS, Pain Catastrophizing Scale; PI, pain interference; PS, pain severity; PSQI, Pittsburgh Sleep Quality Index.

***

p < 0.0001.

In the regression analysis, pain severity was the only variable that showed statistically significant differences across all three groups (Group 2, Group 3, and Group 4) compared to the reference group (Group 1). The odds of being classified into Group 2, Group 3, and Group 4, relative to Group 1, increased by 35.7% (odds ratio = 1.357), 48.9% (odds ratio = 1.489), and 45.9% (odds ratio = 1.459), respectively. Other factors, including pain interference, PCS, GSI, and PSQI, did not significantly predict the likelihood of classification into these groups.

3.5. Distribution of Conventional TMD Diagnostic Classification Across Three Clusters

Figure 3 displays a mosaic plot illustrating the distribution of TMD diagnostic classifications across the different clusters. In Cluster 1, TMD Group 3 (muscle pain) constituted 43.1%, TMD Group 1 (internal derangement) 33.8%, TMD Group 2 (joint pain) 24.6%, and TMD Group 4 (combined joint and muscle pain) 6.1%. In Cluster 2, TMD Group 1 accounted for 51.1%, TMD Group 3 for 26.4%, TMD Group 2 for 23.9%, and TMD Group 4 for 4.0%. In Cluster 3, TMD Group 1 comprised 31.6%, TMD Group 3 28.0%, TMD Group 2 21.1%, and TMD Group 4 8.4%. The Chi‐square test (χ 2 = 20.009, p = 0.003) and Fisher's exact test (p = 0.003) demonstrated a significant association between the clusters and TMD diagnostic categories.

FIGURE 3.

FIGURE 3

Open in a new tab

Mosaic plot showing the TMD diagnostic distribution across clusters. The TMD classification categorized TMD into four groups: group 1 with disc displacement, group 2 with joint pain, group 3 with muscle pain, and group 4 with combined joint and muscle pain. In the mosaic plot, each cell presents the number of individuals corresponding to the specific category. The axis labeled 1, 2, 3, and 4 corresponds to the conventional categories used in TMD diagnosis. χ 2 = 20.009, df = 6, p‐value = 0.003. Fisher's test, p‐value = 0.003.

4. Discussion

This cross‐sectional, comprehensive study investigated the multifaceted nature of TMD in adolescents through detailed pain‐related outcome measures and psychological assessments. Key findings reveal significant diversity in symptom presentation among adolescent TMD patients, which conventional diagnostic methods often fail to capture.

Unsupervised K‐means clustering was applied to objectively classify 662 adolescent TMD patients into three distinct clusters based on their pain and psychological profiles. These clusters exhibited significant differences in pain severity, pain interference, and psychosocial distress indicators, including PCS, GSI from SCL‐90R, and PSQI.

Cluster 1, the “High Impact Group,” comprises 70 patients, approximately 10.6% of the cohort. This group is characterized by the highest levels of pain severity and interference, severe psychological distress, elevated PCS, and poor sleep quality, reflecting a profound impact on daily life and well‐being.

Cluster 2, the “Mild Symptoms Cluster,” is the largest group, including 432 patients (63.9%). This cluster exhibits substantially lower scores across all measured variables, indicating a significantly milder TMD experience. The name “Mild Symptoms” aptly conveys the less debilitating nature of TMD in this group, with minimal impact on daily life.

Cluster 3, the “High Catastrophizing Cluster,” includes 169 patients, approximately 25.5% of the cohort. This group shares high pain severity and interference with Cluster 1 but is distinguished by the highest PCS levels and moderate psychological distress (GSI) and PSQI. The cohort's overall mean pain severity (2.35) and interference (2.14) were mild in this study, consistent with findings by Ujin Yap and colleagues, who demonstrated age‐related differences in TMD presentations. Their research highlights that younger patients report less pain‐related TMD compared to middle‐aged and older adults (Ujin Yap et al. 2021). However, even mild pain levels are associated with diverse manifestations, including PCS, psychological distress, and sleep disturbances.

PCS and sleep quality, though not included in the DC/TMD Axis II for adults, are strongly recommended by expert consensus as critical assessment tools for adolescent TMD (Ekberg et al. 2023). Both PCS and poor subjective sleep quality are among the strongest predictors for the onset and persistence of TMD pain in experimental and clinical settings (Fillingim et al. 2011, 2013; Willassen et al. 2020; Kim and Kim 2023; Häggman‐Henrikson et al. 2021). A systematic review highlighted that high catastrophizing levels correlate with severe pain and poor treatment outcomes (Häggman‐Henrikson et al. 2020). Even mild catastrophizing significantly influences pain intensity and jaw functional limitations in community samples (Häggman‐Henrikson et al. 2021). Aligning with prior research (Januzzi et al. 2023; Kim and Kim 2021; Craner et al. 2016), the helplessness subcategory of the PCS showed the highest scores among the three dimensions of the PCS in the present study. This component was also significantly linked to sleep disturbances in a large adult TMD study involving 1488 participants (Kim and Kim 2021). Studies report a bidirectional relationship between poor subjective sleep quality and TMD pain severity in adults (Dreweck et al. 2020; Roithmann et al. 2021). In a study of 690 adolescents (Fernandes et al. 2022), poor subjective sleep quality characterized by frequent awakenings and daytime fatigue was associated with a higher likelihood of painful TMD symptoms. Additionally, TMD pain intensity showed a negative correlation with sleep quality. A 2021 systematic review by Scott et al. found that adolescents with sleep disturbances face increased odds of developing mental disorders, including mood disorders, with odds ratios ranging from 1.39 to 1.88 (Scott et al. 2021).

These findings emphasize the urgent need for proactive interventions in high‐risk clusters such as Cluster 1 and Cluster 3, characterized by elevated PCS and pain scores. Specifically, cognitive‐behavioral therapy targeting the helplessness aspect of catastrophizing is critical. For Cluster 3, which exhibits high catastrophizing but comparatively milder psychological distress and sleep disturbances than Cluster 1, early intervention can prevent progression to more severe states. Addressing sleep disturbances in Cluster 1, where pain and psychological distress are pronounced, is equally crucial. Managing these factors effectively should be a core component of treatment strategies to reduce the progression and burden of TMD in these vulnerable adolescent groups. Given the elevated somatization and phobic anxiety levels observed in Cluster 1, interventions must also focus on these domains to provide comprehensive care.

While numerous studies have employed cluster analysis within chronic pain populations (Bäckryd et al. 2018; Gilam et al. 2021; Larsson et al. 2017; Dueñas et al. 2015; Hall‐Lord et al. 1999; Alter et al. 2021; Little et al. 2024), research specifically focusing on TMD patients remains limited (Bair et al. 2016; Pimenta e Silva Machado et al. 2012; Zhu et al. 2022; Velly et al. 2002; Gaynor et al. 2021). Among these, the OPPERA study stands out as a notable large‐scale investigation that categorized patients based on diverse biopsychosocial dimensions.

Comparing this study with the Orofacial Pain: Prospective Evaluation and Risk Assessment (OPPERA) research, which included both chronic TMD and TMD‐free cases, the OPPERA study identified three clusters by integrating pain sensitivity and psychological distress measures (Bair et al. 2016). The Adaptive Cluster in the OPPERA study, characterized by low pain sensitivity and minimal psychological distress, is similar to the Mild Symptoms Cluster in the current study. Conversely, the Global Symptom Cluster in the OPPERA study, marked by high pain sensitivity and significant psychological distress, aligns closely with the High Impact Group identified in this study. The Global Symptom Cluster in the OPPERA study not only demonstrated a higher likelihood of chronic TMD pain but also indicated the highest risk for the initial onset of TMD, underscoring the vulnerability of these individuals to developing persistent TMD conditions. This characterization parallels the High Impact Group in the current study, highlighting the critical importance of early and active intervention for high‐risk adolescents. Timely recognition and treatment of severe cases are vital to preventing the progression of TMD into chronic conditions during adolescence, a period of significant physical and psychological development. Proactive measures can significantly reduce long‐term impacts and enhance the quality of life for these young patients.

Unlike the OPPERA study, which identified a pain sensitive cluster characterized by high pain sensitivity but without extensive psychological distress, the present study did not identify a comparable group among adolescents. Instead, it revealed a high catastrophizing group, marked by high pain, elevated catastrophizing, and moderate psychological distress. This difference may be attributed to the adolescent cohort in the current study, where pain severity and interference were less pronounced compared to the OPPERA cohort, which included chronic pain patients. The presence of the High Catastrophizing Group in this study underscores the significant role psychological factors play in shaping the TMD experience during adolescence. This developmental period involves profound emotional and psychological changes, which may influence pain perception and reporting.

The efficacy of the clustering in this study was validated by the substantial proportion of total variance explained by between‐cluster variability, approaching nearly 60%. This high percentage highlights the distinctiveness and relevance of the cluster divisions, strengthening the validity and rigor of the study's findings.

Additionally, multinomial logistic regression analysis was employed to evaluate the clustering approach against conventional TMD classifications. Table 5 reveals that while pain severity significantly predicts the likelihood of classification into pain‐related TMD groups—Group 2 (joint pain), Group 3 (muscle pain), and Group 4 (combined joint and muscle pain)—with respective odds ratios of 1.357, 1.489, and 1.459 (all p < 0.0001) compared to Group 1 (Internal Derangement Group), other factors like pain interference, PCS, psychological distress, and sleep disturbances show less differentiation among these groups.

The mosaic plot in Figure 3 visually depicts the TMD diagnosis distribution across clusters, offering practical insights. Cluster 1 (High Impact Group) comprised 3.32% of the cohort from TMD Group 1, while Cluster 3 (High Catastrophizing Group) included 9.06% from the same group. Conversely, Cluster 2 (Mild Symptom Cluster)—predominantly consisting of TMD Group 1 patients (30.97%)–also included individuals from pain‐related TMD Groups 2, 3, and 4. This distribution indicates that even within the same diagnostic category of internal derangement, some patients experience heightened pain, emotional distress, PCS, and sleep disturbances, leading to more severe pain experiences and restrictions in daily activities. It also demonstrates that patients classified in pain‐related TMD groups (Groups 2, 3, and 4) can also exhibit lower levels of pain and Axis II‐related symptoms, underscoring the diverse manifestations of TMD across individuals.

These findings highlight the limitations of conventional diagnostic approaches that prioritize physical symptoms and anatomical locations. They suggest that integrating a comprehensive, biopsychosocial Axis II‐based framework can better capture the complex dynamics of TMD in diverse patient subgroups. By incorporating Axis II‐related risk factors alongside traditional diagnostic criteria, clinicians can gain deeper insights into individual patient profiles, enhancing targeted therapeutic strategies and improving treatment outcomes.

One of the strengths of this study is its use of a comprehensive, multidimensional approach to classify adolescent patients with TMD into distinct clusters based on Axis II factors. Notably, the study developed practical Axis II‐based TMD phenotypes using only five validated pain‐related variables. These variables are critical dimensions for accurately assessing adolescent TMD patients (Ekberg et al. 2023), thereby enhancing the clinical relevance and applicability of the findings. However, the study has certain limitations. First, its retrospective design restricts the ability to establish causality between the identified clusters and TMD treatment outcomes. Second, the sample was confined to patients from a single dental hospital, potentially limiting the findings' generalizability to the broader adolescent population with TMD. Additionally, while referencing the DC/TMD criteria, the study used questionnaires designed for adults due to the unavailability of adolescent‐specific tools in the clinical setting of the author. This is a significant limitation, especially considering recent advancements in the field. Rongo et al. have developed specific diagnostic criteria for children and adolescents—Axis I and Axis II—respectively—that address unique aspects of TMD in younger populations (Rongo et al. 2021, 2022). Recognizing these issues, the applicability of this study to adolescents may be limited, and future research should prioritize using these newly established criteria to ensure diagnostic accuracy and relevance in pediatric and adolescent cohorts.

To address these limitations, future research should prioritize conducting prospective studies to establish causality, broadening the sample size across multiple regions to improve representativeness, and employing assessment tools validated specifically for adolescents to enhance the accuracy and reliability of findings. Furthermore, according to the ACCTION‐American Pain Society Pain Taxonomy (Dworkin et al. 2016), chronic pain evaluation should adopt a multidimensional framework. This includes core symptoms, common features, medical and psychiatric comorbidities, neurobiological and psychosocial mechanisms, and functional consequences, as well as risk and protective factors. Importantly, a study testing a clustering algorithm with only four variables—pressure pain threshold, depression, somatization, and anxiety—achieved over 87% cluster reliability in classifying chronic pain groups (Gaynor et al. 2021). These findings underscore the necessity of identifying the critical dimensions essential for focusing on the core aspects of TMD assessment and management while considering clinical feasibility.

5. Conclusion

This study underscores the significance of a multidimensional approach to understanding TMD in adolescents. Using unsupervised clustering based on Axis II criteria, it successfully identified three distinct phenotypes within the adolescent TMD cohort: high impact, mild symptoms, and high catastrophizing. These phenotypes highlight substantial variations in pain severity, psychological distress, and other related symptoms, revealing the diverse manifestations of TMD often overlooked in conventional diagnostic categories.

Identifying these phenotypes allows for a more precise understanding of the underlying Axis II characteristics, which is vital for tailoring treatment strategies to the unique needs of each patient. By integrating conventional diagnostic methods with Axis II‐based clustering, this study advances a holistic understanding of adolescent TMD. Given the complexities of chronic TMD treatment, early intervention during adolescence is crucial. This refined perspective is especially important for improving therapeutic outcomes and enhancing the quality of life for adolescents with TMD.

Author Contributions

Hye Kyoung Kim: conceptualization, investigation, funding acquisition, writing – original draft, methodology, validation, visualization, writing – review and editing, software, formal analysis, project administration, resources, supervision, data curation.

Conflicts of Interest

The author declares no conflicts of interest.

Acknowledgements

This research was supported by a grant from Dankook University (No. R202401227).

Funding: This work was supported by Dankook University.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

References

  1. Alter, B. J. , Anderson N. P., Gillman A. G., Yin Q., Jeong J. H., and Wasan A. D.. 2021. “Hierarchical Clustering by Patient‐Reported Pain Distribution Alone Identifies Distinct Chronic Pain Subgroups Differing by Pain Intensity, Quality, and Clinical Outcomes.” PLoS One 16, no. 8: e0254862. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bäckryd, E. , Persson E. B., Larsson A. I., Fischer M. R., and Gerdle B.. 2018. “Chronic Pain Patients Can Be Classified Into Four Groups: Clustering‐Based Discriminant Analysis of Psychometric Data From 4665 Patients Referred to a Multidisciplinary Pain Centre (a SQRP Study).” PLoS One 13, no. 2: e0192623. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bair, E. , Gaynor S., Slade G. D., et al. 2016. “Identification of Clusters of Individuals Relevant to Temporomandibular Disorders and Other Chronic Pain Conditions: The OPPERA Study.” Pain 157, no. 6: 1266–1278. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bertoli, F. M. P. , Bruzamolin C. D., Pizzatto E., Losso E. M., Brancher J. A., and de Souza J. F.. 2018. “Prevalence of Diagnosed Temporomandibular Disorders: A Cross‐Sectional Study in Brazilian Adolescents.” PLoS One 13, no. 2: e0192254. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Buysse, D. J. , Reynolds C. F. 3rd, Monk T. H., Berman S. R., and Kupfer D. J.. 1989. “The Pittsburgh Sleep Quality Index: A New Instrument for Psychiatric Practice and Research.” Psychiatry Research 28, no. 2: 193–213. [DOI] [PubMed] [Google Scholar]
  6. Christidis, N. , Lindström Ndanshau E., Sandberg A., and Tsilingaridis G.. 2019. “Prevalence and Treatment Strategies Regarding Temporomandibular Disorders in Children and Adolescents‐A Systematic Review.” Journal of Oral Rehabilitation 46, no. 3: 291–301. [DOI] [PubMed] [Google Scholar]
  7. Cleeland, C. S. , and Ryan K. M.. 1994. “Pain Assessment: Global Use of the Brief Pain Inventory.” Annals of the Academy of Medicine, Singapore 23, no. 2: 129–138. [PubMed] [Google Scholar]
  8. Craner, J. R. , Gilliam W. P., and Sperry J. A.. 2016. “Rumination, Magnification, and Helplessness: How Do Different Aspects of Pain Catastrophizing Relate to Pain Severity and Functioning?” Clinical Journal of Pain 32, no. 12: 1028–1035. [DOI] [PubMed] [Google Scholar]
  9. da Silva, C. G. , Pachêco‐Pereira C., Porporatti A. L., et al. 2016. “Prevalence of Clinical Signs of Intra‐Articular Temporomandibular Disorders in Children and Adolescents: A Systematic Review and Meta‐Analysis.” Journal of the American Dental Association (1939) 147, no. 1: 10–18. [DOI] [PubMed] [Google Scholar]
  10. de Paiva Bertoli, F. M. , Bruzamolin C. D., de Almeida Kranz G. O., Losso E. M., Brancher J. A., and de Souza J. F.. 2018. “Anxiety and Malocclusion Are Associated With Temporomandibular Disorders in Adolescents Diagnosed by RDC/TMD. A Cross‐Sectional Study.” Journal of Oral Rehabilitation 45, no. 10: 747–755. [DOI] [PubMed] [Google Scholar]
  11. Derogatis, L. R. 1977. SCL‐90R: Administration, Scoring and Procedures. Clinical Psychometric Research. [Google Scholar]
  12. Dreweck, F. D. S. , Soares S., Duarte J., Conti P. C. R., de Luca Canto G., and Luís Porporatti A.. 2020. “Association Between Painful Temporomandibular Disorders and Sleep Quality: A Systematic Review.” Journal of Oral Rehabilitation 47, no. 8: 1041–1051. [DOI] [PubMed] [Google Scholar]
  13. Dueñas, M. , Salazar A., Ojeda B., et al. 2015. “A Nationwide Study of Chronic Pain Prevalence in the General Spanish Population: Identifying Clinical Subgroups Through Cluster Analysis.” Pain Medicine 16, no. 4: 811–822. [DOI] [PubMed] [Google Scholar]
  14. Durá, E. , Andreu Y., Galdón M. J., et al. 2006. “Psychological Assessment of Patients With Temporomandibular Disorders: Confirmatory Analysis of the Dimensional Structure of the Brief Symptoms Inventory 18.” Journal of Psychosomatic Research 60, no. 4: 365–370. [DOI] [PubMed] [Google Scholar]
  15. Dworkin, R. H. , Bruehl S., Fillingim R. B., Loeser J. D., Terman G. W., and Turk D. C.. 2016. “Multidimensional Diagnostic Criteria for Chronic Pain: Introduction to the ACTTION‐American Pain Society Pain Taxonomy (AAPT).” Journal of Pain 17, no. 9: T1–T9. [DOI] [PubMed] [Google Scholar]
  16. Ekberg, E. , Nilsson I. M., Michelotti A., et al. 2023. “Diagnostic Criteria for Temporomandibular Disorders‐INfORM Recommendations: Comprehensive and Short‐Form Adaptations for Adolescents.” Journal of Oral Rehabilitation 50, no. 11: 1167–1180. [DOI] [PubMed] [Google Scholar]
  17. Fernandes, G. , van Selms M. K. A., Lobbezoo F., et al. 2022. “Subjective Sleep Complaints Were Associated With Painful Temporomandibular Disorders in Adolescents: The Epidor‐Adolescere Study.” Journal of Oral Rehabilitation 49, no. 9: 849–859. [DOI] [PubMed] [Google Scholar]
  18. Fillingim, R. B. , Ohrbach R., Greenspan J. D., et al. 2013. “Psychological Factors Associated With Development of TMD: The OPPERA Prospective Cohort Study.” Journal of Pain 14, no. 12: T75–T90. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Fillingim, R. B. , Ohrbach R., Greenspan J. D., et al. 2011. “Potential Psychosocial Risk Factors for Chronic TMD: Descriptive Data and Empirically Identified Domains From the OPPERA Case‐Control Study.” Journal of Pain 12, no. 11: T46–T60. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Gatchel, R. J. , Stowell A. W., Wildenstein L., Riggs R., and Ellis E. 3rd. 2006. “Efficacy of an Early Intervention for Patients With Acute Temporomandibular Disorder‐Related Pain: A One‐Year Outcome Study.” Journal of the American Dental Association (1939) 137, no. 3: 339–347. [DOI] [PubMed] [Google Scholar]
  21. Gaynor, S. M. , Bortsov A., Bair E., et al. 2021. “Phenotypic Profile Clustering Pragmatically Identifies Diagnostically and Mechanistically Informative Subgroups of Chronic Pain Patients.” Pain 162, no. 5: 1528–1538. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Gilam, G. , Cramer E. M., Webber K. A. 2nd, Ziadni M. S., Kao M. C., and Mackey S. C.. 2021. “Classifying Chronic Pain Using Multidimensional Pain‐Agnostic Symptom Assessments and Clustering Analysis.” Science Advances 7, no. 37: eabj0320. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Häggman‐Henrikson, B. , Bechara C., Pishdari B., Visscher C. M., and Ekberg E.. 2020. “Impact of Catastrophizing in Patients With Temporomandibular Disorders‐A Systematic Review.” Journal of Oral & Facial Pain and Headache 34, no. 4: 379–397. [DOI] [PubMed] [Google Scholar]
  24. Häggman‐Henrikson, B. , Visscher M. C., Wänman A., Ljótsson B., Peck C., and Lövgren A.. 2021. “Even Mild Catastrophic Thinking Is Related to Pain Intensity in Individuals With Painful Temporomandibular Disorders.” Journal of Oral Rehabilitation 48, no. 11: 1193–1200. [DOI] [PubMed] [Google Scholar]
  25. Hall‐Lord, M. L. , Larsson G., and Steen B.. 1999. “Chronic Pain and Distress in Older People: A Cluster Analysis.” International Journal of Nursing Practice 5, no. 2: 78–85. [DOI] [PubMed] [Google Scholar]
  26. Haythornthwaite, J. A. , Campbell C. M., and Edwards R. R.. 2024. “When Thinking About Pain Contributes to Suffering: The Example of Pain Catastrophizing.” Pain 165, no. 11S: S68–S75. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Januzzi, M. S. , Neto C. M., Moreno A., et al. 2023. “Relationship Between Self‐Reported Pain, Pain Threshold, Pain Catastrophization and Quality of Life in Patients With TMD.” Journal of Clinical and Experimental Dentistry 15, no. 1: e23–e31. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Jaworska, N. , and MacQueen G.. 2015. “Adolescence as a Unique Developmental Period.” Journal of Psychiatry & Neuroscience 40, no. 5: 291–293. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Kim, H. K. , and Kim M. E.. 2021. “Phenotyping 1488 Patients With Painful Temporomandibular Disorders and Its Relevance to Subjective Sleep Quality: A Key Step for Stratified Medicine.” Cranio 39, no. 6: 491–501. 10.1080/08869634.2019.1682750. [DOI] [PubMed] [Google Scholar]
  30. Kim, H. K. , and Kim M. E.. 2023. “Principal Component Analysis of the Biopsychosocial Features Relevant to Temporomandibular Disorders.” Oral Diseases 29, no. 7: 2917–2927. [DOI] [PubMed] [Google Scholar]
  31. Larsson, B. , Gerdle B., Bernfort L., Levin L. Å., and Dragioti E.. 2017. “Distinctive Subgroups Derived by Cluster Analysis Based on Pain and Psychological Symptoms in Swedish Older Adults With Chronic Pain—A Population Study (PainS65+).” BMC Geriatrics 17, no. 1: 200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Little, C. L. , Schultz D. M., House T., Dixon W. G., and McBeth J.. 2024. “Identifying Weekly Trajectories of Pain Severity Using Daily Data From an mHealth Study: Cluster Analysis.” JMIR mHealth and uHealth 12: e48582. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Manfredini, D. , Marini M., Pavan C., Pavan L., and Guarda‐Nardini L.. 2009. “Psychosocial Profiles of Painful TMD Patients.” Journal of Oral Rehabilitation 36, no. 3: 193–198. [DOI] [PubMed] [Google Scholar]
  34. Minervini, G. , Franco R., Marrapodi M. M., Fiorillo L., Cervino G., and Cicciù M.. 2023. “Prevalence of Temporomandibular Disorders in Children and Adolescents Evaluated With Diagnostic Criteria for Temporomandibular Disorders: A Systematic Review With Meta‐Analysis.” Journal of Oral Rehabilitation 50, no. 6: 522–530. [DOI] [PubMed] [Google Scholar]
  35. Nilsson, I. M. , and List T.. 2020. “Does Adolescent Self‐Reported TMD Pain Persist Into Early Adulthood? A Longitudinal Study.” Acta Odontologica Scandinavica 78, no. 5: 377–383. [DOI] [PubMed] [Google Scholar]
  36. Pigozzi, L. B. , Pereira D. D., Pattussi M. P., et al. 2021. “Quality of Life in Young and Middle Age Adult Temporomandibular Disorders Patients and Asymptomatic Subjects: A Systematic Review and Meta‐Analysis.” Health and Quality of Life Outcomes 19, no. 1: 83. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Pimenta e Silva Machado, L. , de Macedo Nery M. B., de Góis Nery C., and Leles C. R.. 2012. “Profiling the Clinical Presentation of Diagnostic Characteristics of a Sample of Symptomatic TMD Patients.” BMC Oral Health 12: 26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Preti, A. , Carta M. G., and Petretto D. R.. 2019. “Factor Structure Models of the SCL‐90‐R: Replicability Across Community Samples of Adolescents.” Psychiatry Research 272: 491–498. [DOI] [PubMed] [Google Scholar]
  39. Rentsch, M. , Zumbrunn Wojczyńska A., Gallo L. M., and Colombo V.. 2023. “Prevalence of Temporomandibular Disorders Based on a Shortened Symptom Questionnaire of the Diagnostic Criteria for Temporomandibular Disorders and Its Screening Reliability for Children and Adolescents Aged 7‐14 Years.” Journal of Clinical Medicine 12, no. 12: 4109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Roithmann, C. C. , Silva C. A. G. D., Pattussi M. P., and Grossi M. L.. 2021. “Subjective Sleep Quality and Temporomandibular Disorders: Systematic Literature Review and Meta‐Analysis.” Journal of Oral Rehabilitation 48, no. 12: 1380–1394. [DOI] [PubMed] [Google Scholar]
  41. Rongo, R. , Ekberg E., Nilsson I. M., et al. 2021. “Diagnostic Criteria for Temporomandibular Disorders (DC/TMD) for Children and Adolescents: An International Delphi Study‐Part 1‐Development of Axis I.” Journal of Oral Rehabilitation 48, no. 7: 836–845. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Rongo, R. , Ekberg E., Nilsson I. M., et al. 2022. “Diagnostic Criteria for Temporomandibular Disorders in Children and Adolescents: An International Delphi Study‐Part 2‐Development of Axis II.” Journal of Oral Rehabilitation 49, no. 5: 541–552. [DOI] [PubMed] [Google Scholar]
  43. Schiffman, E. , Ohrbach R., Truelove E., et al. 2014. “Diagnostic Criteria for Temporomandibular Disorders (DC/TMD) for Clinical and Research Applications: Recommendations of the International RDC/TMD Consortium Network* and Orofacial Pain Special Interest Group†.” Journal of Oral & Facial Pain and Headache 28, no. 1: 6–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Scott, J. , Kallestad H., Vedaa O., Sivertsen B., and Etain B.. 2021. “Sleep Disturbances and First Onset of Major Mental Disorders in Adolescence and Early Adulthood: A Systematic Review and Meta‐Analysis.” Sleep Medicine Reviews 57: 101429. [DOI] [PubMed] [Google Scholar]
  45. Sullivan, M. J. L. , Bishop S. R., and Pivik J.. 1995. “The Pain Catastrophizing Scale: Development and Validation.” Psychological Assessment 7, no. 4: 524–532. [Google Scholar]
  46. Ujin Yap, A. , Cao Y., Zhang M. J., Lei J., and Fu K. Y.. 2021. “Age‐Related Differences in Diagnostic Categories, Psychological States and Oral Health‐Related Quality of Life of Adult Temporomandibular Disorder Patients.” Journal of Oral Rehabilitation 48, no. 4: 361–368. [DOI] [PubMed] [Google Scholar]
  47. Valesan, L. F. , Da‐Cas C. D., Réus J. C., et al. 2021. “Prevalence of Temporomandibular Joint Disorders: A Systematic Review and Meta‐Analysis.” Clinical Oral Investigations 25, no. 2: 441–453. [DOI] [PubMed] [Google Scholar]
  48. Velly, A. M. , Philippe P., and Gornitsky M.. 2002. “Heterogeneity of Temporomandibular Disorders: Cluster and Case‐Control Analyses.” Journal of Oral Rehabilitation 29, no. 10: 969–979. [DOI] [PubMed] [Google Scholar]
  49. Willassen, L. , Johansson A. A., Kvinnsland S., Staniszewski K., Berge T., and Rosén A.. 2020. “Catastrophizing has a Better Prediction for TMD Than Other Psychometric and Experimental Pain Variables.” Pain Research & Management 12: 7893023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. World Health Organization . 2024. “Adolescent Health.”
  51. Zhu, R. , Zheng Y. H., Zhang Z. H., Fan P. D., Wang J., and Xiong X.. 2022. “Development of a New Category System for the Profile Morphology of Temporomandibular Disorders Patients Based on Cephalograms Using Cluster Analysis.” Frontiers in Public Health 10: 1045815. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

Articles from Oral Diseases are provided here courtesy of Wiley

RESOURCES