Characterizing Chronic Pain Phenotypes in Multiple Sclerosis

A Nationwide Survey Study

Anna L. Kratz; Daniel Whibley; Kevin N. Alschuler; Dawn M. Ehde; David A. Williams; Daniel J. Clauw; Tiffany J. Braley


Pain. 2021;162(5):1426-1433. 

In This Article


Before study initiation, the protocol was submitted to the University of Michigan (UM) Medical Institutional Review Board; the study was deemed to meet federal and institutional criteria for exempt human subjects research. Survey data were collected through Qualtrics, an HIPAA compliant online research tool that allows for direct data entry by study participants and provides real-time data export and automated accrual report features.[48] Individuals who completed the survey had the option of entering a raffle to win one of twenty $100 gift cards.

Study Sample

Study inclusion criteria included a self-reported MS diagnosis and age 18 years or older. There were no other inclusion or exclusion criteria because the goal was to survey a diverse sample of adults with MS from across the United States. Participants were recruited through (1) an existing research registry at UM of people with medically documented MS, (2) posting of the study survey link on the UM research web site, and (3) through a National MS Society listserv email, which distributed the survey link nationwide to approximately 79,100 email addresses (~44,000 emails were opened). Data were collected between December 5, 2019, and January 13, 2020.


The online survey included an in-depth demographic questionnaire (eg, age, self-reported biological sex and gender, education level, income, marital status, employment status, and state of residence) as well the following validated self-report measures. To evaluate the accuracy of self-report of MS, survey items assessing source of MS diagnosis (eg, physician specialty), previous diagnostic workups, and current use of disease modifying therapy were also administered.

Neuropathic Pain

Neuropathic pain was assessed using the painDETECT questionnaire (PD-Q), a 13-item screening survey to determine the presence/severity of pain of neuropathic origin.[20] The PD-Q assesses current average and worst pain intensity over the past 4 weeks (rated on an 11-point numeric rating scale of 0–10) as well as the presence of neuropathic pain qualities (eg, burning sensation and tingling/prickling sensations; rated on a Likert scale from 0 [never] to 5 [very strongly]). Pain duration/pattern and radiation of pain are also assessed. The total score ranges from −1 to 38, with higher scores indicative of higher likelihood of neuropathic pain origin. Scores ≤12 indicate that a neuropathic component of pain is unlikely, scores between 13 and 18 are ambiguous, and scores ≥19 indicate that a neuropathic component of pain is likely.

Nociplastic Pain (Centralized Pain)

The degree of centrally enhanced pain processing was assessed using the American College of Rheumatology (ACR) 2011 Fibromyalgia (FM) Survey Criteria.[6,69,70] This survey includes the number of painful body regions using the Michigan Body Map (0–19) and related symptoms such as problems thinking, fatigue, and sleep difficulties (0–12). This continuously scaled metric (ranging between 0 and 31) can be used as a proxy index for central sensitization or can be used to indicate likely fibromyalgia with a cut-point of >13.[70] This survey has been previously used to quantify centralized pain in other clinical populations,[7,28,41] relates strongly to functional neuroimaging findings in nociplastic pain,[3,32] and is a robust predictor of both pain and disability.[67,68,71,72]

Pain Intensity

Pain intensity was assessed with the PROMIS Pain Intensity 3a,[10] a 3-item measure that assesses worst and average pain in the past 7 days as well as current pain. The item scores were summed, and the total scale score transformed into a normative T-score metric, with a mean = 50, SD = 10. Higher scores are indicative of higher pain intensity.

Chronic Overlapping Pain Conditions

Measures that screen for 3 common chronic overlapping pain conditions (migraine, temporomandibular disorders [TMDs], and pelvic pain)[37] were administered. The three-item ID Migraine was used to screen for presence of migraine.[34] Endorsement of at least 2 of the 3 symptoms (nausea, photophobia, and headache-related disability) has been shown to have excellent sensitivity, specificity, and predictive value relative to identifying migraine as well as excellent test–retest reliability.[34] Presence of likely TMD was assessed with a validated 3-item screening survey, the 3Q/TMD.[35] For this instrument, a score of 3 affirmative responses on 2 jaw pain items and 1 jaw dysfunction item indicates a conservative estimate of TMD risk. Presence of pelvic pain was assessed with a single yes/no item, "Do you have persistent or periodic pain in your pelvis (genitals, pubic, bladder, or perineum region)?".

Pain Medication Use and Associated Relief

A survey of pain medication use and relief was designed for this study. Current use of cannabinoids for medical purposes was collected. Respondents were also asked to endorse if they had used any of the following medications in the past month to manage any pain that was reported in the survey (selecting all that applied): nonsteroidal anti-inflammatory drugs (NSAIDs), opioids, anticonvulsants, selective serotonin reuptake inhibitors (SSRIs), serotonin-norepinephrine reuptake inhibitors, antispasmodics, steroids, and benzodiazepines. For each drug category, a list of example medications was provided on the survey. Respondents rated how much pain relief the medication provided on a 0 (no pain relief) to 10 (complete pain relief) numeric rating scale.

Data Analysis

Demographic and clinical characteristics of the sample were summarized as mean (SD) and/or median (interquartile range [IQR]) for continuous variables and frequency and proportion for categorical variables. χ 2 tests were used to compare frequency of chronic overlapping pain conditions and use of different pain medications by pain subtype group. One-way analysis of variance (ANOVA) was used to compare mean PROMIS pain intensity T-scores and reported pain relief from different pain medications by different pain subtype groups. In cases where the omnibus ANOVA test was significant, post hoc multiple comparison Tukey HSD tests were conducted to examine pairwise comparisons of pain intensity and pain relief ratings across the pain subtypes. Standardized effect sizes (Cohen's d) were calculated to further characterize statistically significant pairwise group differences.