Abstract
Objective
Discrepancies between self-reported and actigraphy sleep measures are common, producing ambiguity about which are better predictors of experimental pain outcomes. The current study tested if pain intensity and situational pain catastrophizing following experimental pain were differentially predicted by self-reported or actigraphy sleep measures in patients with chronic temporomandibular disorder (TMJD) or healthy controls (HCs).
Methods
Forty patients with TMJD and 20 HCs completed self-report sleep measures (Pittsburgh Sleep Quality Index, PSQI; Insomnia Severity Index, ISI; PROMIS Sleep-Related Impairment [SRI] and Sleep Disruption [SD]), underwent an experimental pain induction consisting of four consecutive cold-water hand immersions, and provided pain intensity and situational pain catastrophizing ratings. Participants also wore an actigraphy watch and completed sleep diaries for seven days, which were averaged for actigraphic indices of total sleep time, sleep efficiency, wake after sleep onset, and self-reported sleep quality and restfulness.
Results
Individuals with TMJD reported higher pain intensity during experimental pain (M=65.81 vs. 47.77, p=.007) and self-reported worse sleep compared to HCs (all p’s< .02, Cohen’s D=0.73–1.25). No group differences emerged for actigraphy measures (all p’s> .05, Cohen’s D=0.05–0.53). Sleep variables did not interact with group to predict responses to experimental pain (all p’s>.05). Across groups, PROMIS-SRI predicted pain intensity (β=0.36, p=.008) and catastrophizing (β =0.36, p=.009) after controlling for multiple comparisons, smoking, medications, and age.
Conclusion
Self-reported sleep (but not actigraphy) measures differentiate patients with TMJD from HCs. Sleep-related interference may place people at particular risk for higher pain intensity and catastrophizing following experimental pain.
Introduction
Temporomandibular joint disorder (TMJD) affects 7–15% of the population and is associated with impaired sleep.1–7 Two common methods to measure sleep, self-report and actigraphy, are only moderately correlated8–11 and differentially predict chronic pain outcomes in some chronic pain conditions.8–13 However, little is known about how self-report and actigraphy sleep measures differ between patients with TMJD and healthy controls (HCs), or about how they predict responses to experimentally-induced pain. Examining responses to experimentally-induced pain is important for predicting clinical outcomes and categorizing patients into clinically-meaningful subgroups.14
The current study sought to compare self-report and actigraphy measures between individuals with TMJD and HCs, and to test whether self-report or actigraphy measures were stronger predictors of pain intensity or situational pain catastrophizing (defined as magnification, rumination, and perceived helplessness from pain)15 following experimental pain. We hypothesized that TMJD patients would report poorer sleep based on self-report but not actigraphy compared to HCs. Additionally, we hypothesized that poorer self-reported sleep (but not actigraphy) would be associated with higher pain intensity and situational pain catastrophizing following experimental pain.
Methods
Participants
The current study presents data from a parent study examining immunological activation following experimental pain in TMJD. The sample size was selected to adequately power the parent study. A post-hoc power analysis revealed the current study was 80% powered to detect regression effects of R2=.12 or larger. Participants were recruited from the general community including Cincinnati Children’s hospital and the surrounding area using print and social media.
Sixty-two participants (n=40 TMJD, n=22 HCs) met the following inclusion criteria: age 18–50; English-speaking; no cancer treatment in the past year; not diagnosed with or receiving treatment for diabetes, thyroid disorders, cardiovascular disease, hypertension, pulmonary disease, chronic obstructive pulmonary disease (COPD), neurological disorders, or psychiatric disorders requiring hospitalization in the past year; not pregnant; not using opioids; and no hospitalization/surgery within the past 6 months. TMJD patients met the additional criteria of having a confirmed myofascial, arthralgia, or mixed facial pain diagnosis, having pain five or more days in the last month, and having a history of facial pain for more than six months.
Procedures
Participants came into the lab for two visits, seven days apart. Between visits, they continuously wore an actigraphy watch (Actiwatch2, Respironics) on their non-dominant wrist and completed sleep diaries each morning. The study was approved by the IRB (IRB#2015–4992).
Visit 1. Participants provided consent, completed a diagnostic exam to determine TMJD/HC status,16 completed self-report questionnaires, and were given the actigraphy watch.
Visit 2. Participants completed additional self-report questionnaires before undergoing the experimental pain induction involving four subsequent 60-sec immersions of their non-dominant hand into 8°C water bath. They rested 30-sec between immersions. After the task, they completed a situational pain catastrophizing questionnaire. Participants were compensated $45.
Materials
Demographics included self-reported age, race, marital status, smoking status, medication use, pain duration and severity (TMJD only), and health history.
Experimental Pain Intensity was reported at 30- and 45-sec for each immersion, producing eight total ratings (α=0.98). A verbal numeric rating scale ranging from 0 (no pain) to 100 (most pain imaginable) was used. An average across all eight ratings was computed.
Situational Pain Catastrophizing15 was measured using six items (α=0.67). Each item was rated on a 1–5 scale, with higher scores indicating greater catastrophizing. Participants were asked to refer to the pain induction procedure when completing the questionnaire.
Self-Reported Sleep Measures
Pittsburgh Sleep Quality Index (PSQI)17 measured sleep quality and patterns in the last month using 19 items (α=0.77). A global score ranging from 0–21 was computed, with higher scores indicating poorer sleep quality.
Insomnia Severity Scale (ISI)18 gauged the severity, distress, and daytime impairment of insomnia in the last two weeks using seven items (α=0.88). Each was scored on a 0–4 scale. Higher scores indicated greater insomnia symptomatology.
Patient-Reported Outcomes Measurement Information System (PROMIS) Sleep-Related Impairment (SRI).19 Sleep impairment over the last seven days was measured with eight items (α=0.77). Each was assessed on a 1–5 scale. Positively-worded items were reverse scored and all items were summed so that higher scores indicated greater dysfunction.
PROMIS Sleep Disturbance (SD).19 Sleep disturbance over the last seven days was evaluated with eight items (α=0.77). Each item was assessed on a 1–5 scale. Positively-worded items were reverse scored and all items were summed so that higher scores indicated greater sleep disturbance.
Sleep Diary - Sleep Quality and Restfulness. Each morning between visits, participants completed a daily diary assessing sleep the previous night. Sleep quality and restfulness were self-reported each day using single items. For each item, participants used a 1–5 scale. Across the seven days,α=0.76 for sleep quality and α=0.82 for restfulness.
Actigraphy Measures
Actigraphy data were compared to sleep diary data at the second visit. Discrepancies were discussed with the participant and adjusted in the actigraphy report. The following variables were calculated using Respironics software algorithms:
Total Sleep Time measured the duration in minutes of sleep onset to wake time (α=0.47 across seven days).
Sleep Efficiency was calculated by dividing sleep time by the number of minutes in the rest interval (α=0.69).
Wakefulness after Sleep Onset (WASO) measured the minutes of wakefulness after falling asleep (α=0.67).
Data Analysis
Variables were checked for missingness, normality, and outliers using a criterion of +/− 4 SD. Descriptive statistics and t-tests compared TMJD and HC groups. Effect sizes were calculated using Cohen’s D. To predict pain intensity during experimental pain, separate linear regression models were run using each sleep measure as the unitary predictor. Models were tested with a main effect for group (TMJD vs. HC) and a groupXsleep-measure interaction term. Models were tested with and without the covariates of medication usage (1=participants using prescribed/over-the-counter medication [n=42], 0=no medication [n=18]), smoking status (1=any current smoking [n=11], 0=no current smoking [n=49]), and age. A Holm-Bonferroni correction was used to control for multiple comparisons (9 IVs × 2 DVs × 2 [with/without covariates]=36 total models).20
Results
Missing Data and Outliers
There were no missing data for the PSQI or ISI. One HC participant did not return for Visit 2, leaving 61 cases available for PROMIS, pain intensity, and situational pain catastrophizing variables. One watch malfunctioned, leaving 60 participants for actigraphy analyses. Of those, 58 provided complete data (3.7% missing actigraphy data across the entire sample). No daily diary data were missing.
Group Differences in Sleep and Pain
Table 1 provides descriptive statistics. The TMJD group reported poorer sleep than the HC group, but did not differ based on actigraphy. The TMJD group reported higher pain intensity during experimental pain (p=.007) and marginally higher situational pain catastrophizing (p=.09) than the HC group.
Table 1.
Differences in Self-Reported Sleep, Actigraphy, and Pain Outcomes
| Full Sample Mean (SD) | TMJD (n=40)1 Mean (SD) | HC (n=22)2 Mean (SD) | t | p | Cohen’s D | |
|---|---|---|---|---|---|---|
| Demographics and Pain Description | ||||||
| Age | 29.47 (6.58) | 30.40 (6.19) | 27.77 (7.07) | 1.52 | .29 | 0.40 |
| % Female | 85.5 | 87.5 | ||||
| Race | ||||||
| Caucasian | 80.6 | 90.0 | 63.6 | - | - | - |
| African American | 6.5 | 2.5 | 13.6 | - | - | - |
| Asian American | 4.8 | 2.5 | 9.1 | - | - | - |
| Mixed/Other | 8.1 | 5.0 | 13.6 | - | - | - |
| % Married | 35.5 | 47.5 | 13.6 | - | - | - |
| Pain Duration (in years) | - | 7.71 (5.71) | - | - | - | - |
| Pain Severity | ||||||
| Mild n, (%) | - | 16 (40.0%) | - | - | - | - |
| Moderate n, (%) | - | 23 (57.5%) | - | - | - | - |
| Severe n, (%) | - | 1 (2.5%) | - | - | - | - |
| Self-Report Measures | ||||||
| Sleep Surveys | ||||||
| PSQI (range) | 6.68 (3.62) | 7.90 (3.12) | 4.45 (3.46) | 4.00 | <.001 | 1.05 |
| ISI Total (0–28) | 6.92 (5.00) | 8.83 (4.36) | 3.45 (4.22) | 5.37 | <.001 | 1.25 |
| PROMIS SRI (range) | 17.03 (4.84) | 18.70 (4.93) | 13.86 (2.63) | 4.19 | <.001 | 1.22 |
| PROMIS SD (range) | 21.15 (5.14) | 22.43 (4.30) | 18.71 (5.82) | 2.83 | .006 | 0.73 |
| Sleep Diary | ||||||
| Sleep Quality (0–5) | 3.49 (0.55) | 3.32 (0.48) | 3.79 (0.56) | 3.49 | <.001 | 0.90 |
| Restfulness (0–5) | 3.04 (0.68) | 2.83 (0.59) | 3.43 (0.68) | 3.62 | <.001 | 0.94 |
| Actigraphy Measures | ||||||
| Total Sleep Time (minutes) | 414.70 (45.78) | 413.80 (44.01) | 416.36 (49.99) | 0.21 | .84 | 0.05 |
| Sleep Efficiency (%, 0–100) | 82.60 (5.58) | 82.81 (5.70) | 82.21 (5.45) | 0.36 | .69 | 0.11 |
| Wake After Sleep Onset (minutes) | 40.53 (14.23) | 37.98 (14.27) | 45.26 (13.21) | 1.93 | .058 | 0.53 |
| Pain Outcomes | ||||||
| Average Cold Pain Intensity (0–100) | 59.68 (25.13) | 65.81 (21.65) | 47.99 (27.61) | 2.77 | .007 | 0.72 |
| Situational Pain Catastrophizing (range) | 2.75 (1.24) | 2.95 (1.25) | 2.39 (1.16) | 1.71 | .09 | 0.46 |
Abbreviations: HC = Healthy controls; ISI = Insomnia Severity Index; PROMIS = Patient Reported Outcome Measurement Information System; PSQI = Pittsburgh Sleep Quality Index; TMJD = Temporomandibular joint disorder; SRI = Sleep-Related Impairment; SD = Sleep Disturbance.
Note:
One participant from the TMJD group did not complete actigraphy data, leaving 39 cases available for analysis of objective sleep measures in that group.
One participant in the HC group did not return for a second visit, leaving 21 cases available for analysis in that group on all measures except for PSQI and ISI which were collected at visit one.
Sleep Variables Predicting Experimental Pain Outcomes
Table 2 reveals that PSQI and PROMIS-SRI were significantly associated with pain intensity during experimental pain (Model 1), but only the PROMIS-SRI relationship remained significant after controlling for multiple comparisons and covariates (Model 2).
Table 2.
Predicting Pain Intensity and Situational Pain Catastrophizing from Self-Reported and Actigraphy Sleep Measures
| DV: Pain Intensity | DV: Situational Pain Catastrophizing | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Model1 | 1 (unadjusted) |
2 (adjusted) |
3 (unadjusted) |
4 (adjusted) |
||||||||
| β | t | p | β | t | p | β | t | p | β | t | p | |
| Self-Report Sleep Measures | ||||||||||||
| PSQI | 0.30 | 2.43 | .018 | 0.28 | 2.07 | .04 | 0.29 | 2.32 | .024 | 0.30 | 2.29 | .026 |
| ISI Total | 0.21 | 1.62 | .11 | 0.16 | 1.16 | .25 | 0.27 | 2.15 | .036 | 0.28 | 2.06 | .044 |
| Sleep Diary Sleep Quality | −0.05 | −0.35 | .73 | 0.02 | 0.15 | .88 | −0.04 | −0.03 | .98 | −0.01 | −0.08 | .93 |
| Sleep Diary Restfulness | −0.03 | −0.23 | .82 | 0.05 | 0.34 | .73 | 0.05 | 0.37 | .71 | 0.06 | 0.42 | .68 |
| PROMIS SRI | 0.37 | 3.04 | .004 | 0.36 | 2.74 | .008 | 0.36 | 2.97 | .004 | 0.36 | 2.71 | .009 |
| PROMIS SD | 0.06 | 0.44 | .66 | 0.01 | 0.11 | .92 | −0.02 | −0.17 | .87 | −0.02 | −0.11 | .91 |
| Actigraphy Sleep Measures | ||||||||||||
| Total Sleep Time | −0.16 | −1.21 | .23 | −0.16 | −1.17 | .25 | −0.03 | −0.25 | .80 | −0.01 | −0.07 | .95 |
| Sleep Efficiency | −0.19 | −1.51 | .14 | −0.19 | −1.45 | .15 | −0.24 | −1.86 | .068 | −0.21 | −1.65 | .11 |
| Wake After Sleep Onset | 0.004 | 0.03 | .98 | 0.01 | 0.10 | .92 | 0.05 | 0.38 | .71 | 0.05 | 0.38 | .71 |
Abbreviations: DV = Dependent Variable; ISI = Insomnia Severity Index; PROMIS = Patient Reported Outcome Measurement Information System; PSQI = Pittsburgh Sleep Quality Index; SRI = Sleep-Related Impairment; SD = Sleep Disturbance.
Note:
Each predictor was tested in a separate model. Model columns 1 and 3 show results of unadjusted analyses, where the predictor was the only independent variable in the model. Model columns 2 and 4 show these results controlling for medication use (Yes [1]/No [0]), smoking status (Yes [1]/No [0]), and age.
Italicized values are less than .05 but become nonsignificant after controlling for multiple comparisons using the Holm-Bonferroni correction. Bolded p-values are significant after controlling for multiple comparisons using the Holm-Bonferroni correction.
PSQI, ISI, and PROMIS-SRI were associated with situational pain catastrophizing following experimental pain, but only the PROMIS-SRI relationship remained significant after controlling for multiple comparisons and covariates (Models 3–4).
Diagnosis group did not interact with sleep measures to predict pain intensity or situational pain catastrophizing (all p’s>.05, results available upon request).
Discussion
Consistent with our first hypothesis, patients with TMJD self-reported poorer sleep than HCs on all sleep measures but did not differ from HCs on actigraphy measures. In partial support of our second hypothesis, self-reported sleep-related impairment predicted pain intensity and situational pain catastrophizing following experimental pain. The effects of sleep variables on experimental pain outcomes were similar between groups.
Each sleep measure included in the self-report battery represents a distinct aspect of sleep. The PSQI and PROMIS-SRI assess the impact of poor sleep,17,19 whereas the ISI and the PROMIS-SD assess the symptoms of poor sleep.18,19 Given that the PSQI and PROMIS-SRI had the largest effects on experimental pain outcomes, the impact of poor sleep may be more strongly associated with pain outcomes than the symptoms of poor sleep. Future work is needed to explicitly test this hypothesis.
Unexpectedly, no relationships were found between pain outcomes and self-reported sleep quality or restfulness from the daily diaries. Because diaries were completed between visits, they may be less strongly associated with experimental pain outcomes than measures completed the same day as the induction. Alternatively, single-items may not reliably capture between-person variance. This study is further limited by small sample size and the subclinical nature of the TMJD group (only 1/40 rated their jaw pain as “severe”). Findings may not generalize to other pain populations or to those with more severe TMJD. Method variance could account for why self-reported measures were stronger predictors of self-reported DVs than actigraphy measures. Data were collected by the same unblinded experimenter, introducing a possibility for bias.
Despite these limitations, the current study contributes to the literature by describing TMJD/HC differences in self-reported and actigraphy sleep measures and by examining the relationships between sleep measures and experimental pain outcomes in chronic pain. Results highlight the importance of assessing patient sleep perceptions even in the absence of objective sleep deficits.
Acknowledgment
This work was supported by the National Institute of Dental and Craniofacial Research of the National Institute of Health (R00DE022368).
References
- 1.Riley JL III, Benson MB, Gremillion HA, Myers CD, Robinson ME, Smith CL Jr, Waxenberg LB. Sleep disturbance in orofacial pain patients: pain-related or emotional distress?. Cranio 2001; 19(2): 106–113. [DOI] [PubMed] [Google Scholar]
- 2.Edwards RR, Grace E, Peterson S, Klick B, Haythornthwaite JA, Smith MT. Sleep continuity and architecture: associations with pain-inhibitory processes in patients with temporomandibular joint disorder. European J Pain 2009; 13(10): 1043–1047. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Sessle BJ, Adachi K, Yao D, Suzuki Y, Lavigne GJ. Orofacial pain and sleep. Contemp Oral Med 2017; 1–27. [Google Scholar]
- 4.Sommer I, Lavigne G, Ettlin DA. Review of self-reported instruments that measure sleep dysfunction in patients suffering from temporomandibular disorders and/or orofacial pain. Sleep Med 2015; 16(1): 27–38. [DOI] [PubMed] [Google Scholar]
- 5.Lei J, Liu MQ, Jin Yap AU, Fu KY. Sleep disturbance and psychologic distress: prevalence and risk indicators for temporomandibular disorders in a Chinese population. J Oral Facial Pain Headache 2014; 29(1): 24–30. [DOI] [PubMed] [Google Scholar]
- 6.Schmitter M, Kares-Vrincianu A, Kares H, Bermejo JL, Schindler HJ. Sleep-associated aspects of myofascial pain in the orofacial area among TMD patients and controls. Sleep Med 2015; 16(9): 1056–61. [DOI] [PubMed] [Google Scholar]
- 7.Dubrovsky B, Janal MN, Lavigne GJ, Sirois DA, Wigren PE, Nemelivsky L, Raphael KG. Depressive symptoms account for differences between self‐reported versus polysomnographic assessment of sleep quality in women with myofascial TMD. J Oral Rehab 2017; 44(12): 925–933. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Lauderdale DS, Knutson KL, Yan LL, Liu K, Rathouz PJ. Self-reported and measured sleep duration: how similar are they? Epidem 2008; 19(6), 838–845. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Okifuji A, Hare BD. Nightly analyses of subjective and objective (actigraphy) measures of sleep in fibromyalgia syndrome: what accounts for the discrepancy?. Clin J Pain 2011; 27(4), 289–293. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Stuifbergen AK, Phillips L, Carter P, Morrison J, Todd A. Subjective and objective sleep difficulties in women with fibromyalgia syndrome. J Amer Acad Nurse Pract 2010; 22(10): 548–556. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hughes JM, Song Y, Fung CH, Dzierzewski JM, Mitchell MN, Jouldjian S, Josephson KR, Alessi CA, Martin JL. Measuring sleep in vulnerable older adults: A comparison of subjective and objective sleep measures. Clinical Gerontol 2018; 41(2): 145–157. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Landis CA, Frey CA, Lentz MJ, Rothermel J, Buchwald D, Shaver JL. Self-reported sleep quality and fatigue correlates with actigraphy in midlife women with fibromyalgia. Nursing Res 2003; 52(3): 140–147. [DOI] [PubMed] [Google Scholar]
- 13.Lunde LH, Pallesen S, Krangnes L, Nordhus IH. Characteristics of sleep in older persons with chronic pain: a study based on actigraphy and self-reporting. Clin J Pain 2010; 26(2): 132–137. [DOI] [PubMed] [Google Scholar]
- 14.Edwards RR, Sarlani E, Wesselmann U, Fillingim RB. Quantitative assessment of experimental pain perception: multiple domains of clinical relevance. Pain 2005; 114(3): 315–319. [DOI] [PubMed] [Google Scholar]
- 15.Campbell CM, Kronfli T, Buenaver LF, Smith MT, Berna C, Haythornthwaite JA, Edwards RR. Situational versus dispositional measurement of catastrophizing: associations with pain responses in multiple samples. J Pain 2010; 11(5): 443–453. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Maixner W, Diatchenko L, Dubner R, Fillingim RB, Greenspan JD, Knott C, Slade GD. Orofacial pain prospective evaluation and risk assessment study–the OPPERA study. J Pain 2011; 12(11): 4–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Buysse DJ, Reynolds III CF, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res 1989, 28(2): 193–213. [DOI] [PubMed] [Google Scholar]
- 18.Morin CM. Insomnia: psychological assessment and management. New York: Guilford Press, 1993. [Google Scholar]
- 19.Yu L, Buysse DJ, Germain A, Moul DE, Stover A, Dodds NE, Pilkonis PA. Development of short forms from the PROMIS sleep disturbance and sleep-related impairment item banks. Behav Sleep Med 2014: 10(1); 6–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Aickin M, Gensler H. Adjusting for multiple testing when reporting research results: the Bonferroni vs Holm methods. Amer J Public Health 1996; 86(5): 726–728. [DOI] [PMC free article] [PubMed] [Google Scholar]