Association Between Nociplastic Pain and Pain Severity and Impact in Women With Chronic Pelvic Pain

This prospective cross-sectional study included all consecutive new patients who presented to an academic referral clinic for chronic pelvic pain and endometriosis between January 2013 and December 2015. Patients are typically referred from a generalist obstetrician-gynecologist or a primary care provider. Patients travel from across a large encatchment area, encompassing diverse socioeconomic, population density, and geographic locations over a 3 state region. Prior to their first clinic appointment, all new patients complete a detailed questionnaire, which includes questions regarding demographic information, self-reported race and ethnicity, medical and surgical history, current and prior treatments utilized for pelvic pain symptoms, dysmenorrhea, and pain exacerbating factors, as well as multiple validated self-report measures of pain characteristics, psychological symptoms, functional status. Finally, patients undergo a comprehensive history and standardized physical exam on the day of the first appointment. Of note, this database only collects information from the first clinic visit and does not currently prospectively follow patients at subsequent visits. The primary predictor of interest was the American College of Rheumatology 2011 FM Survey Score, which is the sum the widespread pain index (WPI) and the symptoms severity (SS) scale. WPI assesses pain in 19 body areas (score 0−19). SS assesses fatigue, sleep dysfunction, depression, headache, and memory/cognitive symptoms (score 0−12). 37 This measure has demonstrated good reliability, convergent and discriminant validity, with a score of 13 or greater on the FM Survey Score demonstrating high sensitivity and specificity for clinical diagnosis of fibromyalgia. 37 However, previous work has shown that nociplastic pain is more accurately considered a continuum rather than a cutoff, demonstrated by the finding that even sub-fibromyalgia scores positively correlate with symptoms and treatment response in a variety of pain conditions. 5 , 9 , 12 , 16 Higher scores on this measure are associated with replicable patterns of pain-promoting brain activity on functional neuroimaging and quantitative sensory testing, even when the index pain condition differs. 7 , 26 Relative to history of endometriosis, we used a 2-question screening instrument that asks about prior diagnoses of endometriosis and whether or not this diagnosis was surgically confirmed, a method previously shown to strongly correlate with examination of medical records. 24 Given the tertiary referral nature of this clinic, the majority of patients with surgically confirmed endometriosis had undergone surgery at other facilities. We obtained operative notes whenever possible but were unable to reliably determine stage of endometriosis given the variability in reporting of operative findings and inability to obtain operative notes for every patient. Therefore, we treated history of endometriosis as a binary outcome in this analysis. Given the extensive data indicating poor correlation between stage of endometriosis and symptom severity, 34 , 35 we deemed this an acceptable categorization strategy. Measures related to pain included the Brief Pain Inventory (BPI) pain severity and pain interference scores. 17 Pain severity scores range from 0–10 and pain interference scores from 0 to 10, with increasing scores indicated more severity or interference, respectively. We also collected self-report data regarding degree of dysmenorrhea (0−10), dysuria (0−10), pain with full bladder (0−10), dyschezia (0−10), and deep dyspareunia (0−10). Patients also indicated the number of days per month that they experienced bothersome pain (0−30) and number of days during a 3-month period that they missed activities such as work or school due to pain. Measures related to functional status included Patient-Reported Outcomes Measurement Information System (PROMIS) 10 Physical Function short form (range 10−50, higher scores indicate better function), PROMIS Fatigue short form 7a (range 7−35, higher scores indicate more fatigue), and PROMIS Sleep Disturbance 8b (range 8 −40, higher scores indicating greater degree of disturbance). Measures related to psychological symptoms included PROMIS Anxiety short form 7a (range 7−35, higher scores indicating worse anxiety) and PROMIS depression short form 8a (range 8−40, higher scores indicating worse depression). We also included average pain score on palpation of pelvic floor muscles. All new patients undergo a standardized physical exam at their new patient visit, during which we palpate bilateral anterior levator ani (pubococcygeus), posterior levator ani (iliococcygeus), and obturator muscles transvaginally with patient in dorsal lithotomy position using approximately 2 kg of pressure. Patients are asked to rate their pain severity upon palpation of each muscle group on a scale of 0 to 10. We then averaged the scores across all 6 muscle groups (range 0−10, increasing score indicating greater pain severity). This pelvic floor examination technique has been described and utilized for assessment of pelvic myofascial pain in several chronic pelvic pain conditions. 14 , 31 , 39 We excluded patients who had missing questionnaires and incomplete FM Survey Score assessments. Given our interest in whether endometriosis might predict many of these outcomes, we also excluded patients who had not answered the questions related to history of endometriosis. We did include patients who had missing data related to isolated pain or functional outcome measures, such as dysmenorrhea or dyspareunia, given that the questionnaire had not included an option to indicate that the patient was amenorrheic or not sexually active, for example. We indicated the final number of patients that contributed to these specific descriptive analyses in the tables. For descriptive analyses, we divided our patient sample into tertiles based on FM Survey Scores. These were compared by 1-way analysis of variance and chi-square contingency analyses for continuous and categorical variables respectively. General Linear Models were used to assess the association of total FM Survey Scores with pain and functional outcomes. All models included the following covariates: age, BMI, presence of surgically confirmed endometriosis, history of any laparoscopy/laparotomy for pelvic pain, and current use of opioid medication. Notably, we did not include sleep, fatigue, or depression as covariates in the general lineal models as these symptoms are included in the FM Survey Score. The strength of the association between FM Survey Scores and pain/function outcomes compared to other predictors was assessed by eta-squared values. While chronic pain is highly prevalent, a smaller subset of patients experience persistent and debilitating interference with their daily lives − this concept was formalized using the National Health Interview Survey by Pitcher et al. 29 High-impact pain was defined as pain experienced on most days or every day that rendered them unable to perform one or more daily activities. Here we adopt an exploratory version of this concept by defining high-impact pain as pain occurring at least a majority of days (defined as fifteen or more days in a 30-day period) and major interference on the Brief Pain Inventory (defined as 8 or above on a 10-point scale) with one or more of the following activities: general activity, walking ability, normal work (outside the home or housework), or relations with other people. We then used a binary logistic regression model with high impact pain status as the outcome of interest. The same set of covariates were used as above. The primary predictor of interest was FM Survey Scores. This study was approved by the Institutional Review Board of the University of Michigan (IRB#HUM00003797). Requirement for written consent was waived by the IRB as the questionnaire data is part of the standard clinical record and all data entered into the database was deidentified, and therefore posed minimal risk to subjects. Reporting of study findings is consistent with STROBE guidelines. Data analyses were performed using SPSS version 28.0 and R version 3.6.

A total of 401 new patients were evaluated in clinic between January 2013 to December 2015. Of these, 98 were excluded due to missing data (57 had incomplete responses for FM Survey Score and 41 had incomplete responses for endometriosis history). Therefore, 303 women were included in this analysis ( Fig 1 ). Tertiles of FM Survey Scores resulted in the following groups: 0 to 9 were categorized as Low FM (n = 112), scores 10 to 12 were categorized as moderate FM (n = 81), and scores 13 to 31 categorized as high FM (n = 110). There were no significant differences across demographic or clinical characteristics, including history of endometriosis ( Table 1 ). Average scores for each FM Survey Score subscale (WPI and SS) are presented in Table 2 . As shown in Fig 2 , distribution of pain sites differed widely across the 3 groups, with a high proportion of patients in the high FM group reporting chronic pain distant from the pelvis. As expected, unadjusted comparisons between the groups showed differences on pain and functional outcomes (all P < .05), with the exception of pain during urination ( P = .39). After controlling for covariates using general linear models, higher FM Survey Scores were independently associated with greater pain severity ( Table 3 ), pain interference ( Supplementary Table 1 ), and pain frequency ( Supplementary Table 2 ), as well as greater pelvic floor myofascial pain on examination ( Supplementary Table 3 ) (all P < .01). Patient age was inversely correlated with pain severity, pain interference, and pelvic floor myofascial pain. The only other factor that surpassed the predictive power of the FM score on poor pain outcomes was current opioid use, which was the strongest predictor. By eta-squared values, FM Survey Scores were consistently the second strongest predictor of outcomes after opioid use. By our definition, 95 or approximately 31% of the sample had high impact pain. In the binary logistic regression model, higher FM Survey Scores were independently associated with the likelihood of high impact pain when controlling for all covariates. Specifically, each 1-point increase on the 31-point FM Survey Score was associated with a roughly 7% increase in the likelihood of high impact pain (OR: 1.076; 95% CI= 1.017, 1.139; P = .011). All model parameters are shown in Table 4 .

In this prospective cross-sectional study, patients with higher levels of nociplastic pain symptoms experienced worse pelvic pain symptoms across a range of pain morbidity measures. These measures include more severe or intense pain, more widespread foci of pain, more frequent pain, pain that causes greater interference with daily activities, as well as greater pain/tenderness during pelvic myofascial exam. These relationships were independent of the presence of endometriosis, history of surgical procedures for pelvic pain, and BMI. Our findings are consistent with a recent study demonstrating a positive association between increased levels of nociplastic pain and higher severity of pelvic pain, dysmenorrhea, dyschezia, and dyspareunia in an endometriosis population. 27 Our findings build upon this earlier analysis by expanding to a more broad CPP population, including those with and without endometriosis, examining association with high impact pain, and evaluating whether a prior diagnosis of endometriosis influences this relationship. Additionally, we have utilized a different measurement tool for nociplastic pain that has been extensively studied in a variety of chronic pain conditions and has previously been associated with persistent pain after hysterectomy, postsurgical opioid use, and correlates with neurobiological features of nociplastic pain on functional neuroimaging and quantitative sensory testing. Given the inconsistent relationships between endometriosis and pelvic pain, it is perhaps unsurprising that we did not find a clear relationship between endometriosis and pain or functional outcomes in this analysis. In fact, a history of endometriosis was associated with less pelvic pain severity and interference in multivariable models, after controlling for nociplastic pain. A likely interpretation is that patients with a nociceptive component to their pain (ie, endometriosis) gain some benefit from the available treatment strategies (eg, hormonal suppression, surgical excision) that are largely focused on nociceptive input. Conversely, use of opioids was a powerful correlate of pain and functional outcomes, with patients using opioids faring worse on all measures of pain morbidity. Apart from FM Survey Score, use of opioids demonstrated the strongest association with pain morbidity. It is worth noting that this database did not collect details of current opioid use in a standardized fashion, such as daily oral morphine equivalents (OME), and relied on patient reporting of current opioid use as a binary indicator. Regardless, this strong cross-sectional relationship between opioid use and poorer pain outcomes is seen often, and normally attributed to “confound by indication”, ie, individuals with the worst pain are given opioids. However, chronic opioid use may also lead to a state of opioid-induced hyperalgesia and long-term outcomes tend to be very poor, with patients who chronically use opioids reporting a greater degree of sleep dysfunction, depression, functional outcomes such as disability, and higher pain severity and interference compared to similar patients who do not use opioids. 13 In fact, one of the only long term RCTs comparing opioid to nonopioid therapy for chronic pain recently showed that the group receiving opioids had statistically worse pain intensity after 1 year compared to those randomized to receive NSAIDs and acetaminophen. 19 The fact that this sample displayed high levels of nociplastic symptoms may also explain why patients with nociplastic pain have decreased responsiveness to opioids, as previous work has shown deficiencies in the endogenous opioid system of FM patients. 15 , 30 At the very least, use of opioids did not improve pain or functional status in our patient population and may be associated with worse pain and functional outcomes, which further supports recommendations against routine use of opioids for management of chronic pain conditions. We explored the emerging concept of high-impact pain by modeling the likelihood of an individual experiencing major life disruption because of pain. Those with high-impact pain are more likely to suffer from mental health conditions and cognitive impairment, struggle more with self-care, and use more health care. 29 We found that every 1-point increase on the 0 to 31 FM Survey Score was associated with a 7% increase in the likelihood of high-impact pain. This suggests FM Survey Score may be useful for identifying the most vulnerable patients who may benefit from more regular monitoring or aggressive multimodal treatment strategies. Elevated FM Survey Scores are associated with greater sensitivity to experimental pain in multiple body regions, which is taken to indicate global pain sensitivity governed by the central nervous system. Here we show that this sensitivity extends to the pelvic floor, as patients with higher FM Survey Scores showed more myofascial pain on pelvic exam, linking local pelvic sensitization to a much broader process of sensitization across the entire body. This is consistent with several recent studies that demonstrated an association between nociplastic pain and pelvic myofascial pain in women with chronic pelvic pain and endometriosis. 28 , 31 , 33 Whether pelvic floor myofascial pain is a cause, consequence, or simply a biomarker of nociplastic pain remains unknown and likely varies across patients. We found an inverse relationship between patient age and pain severity, pain interference, and pelvic floor myofascial pain. Data regarding relationship between age and pelvic pain symptoms is scant, but several retrospective studies have demonstrated similar findings of increased risk for presence and severity of pelvic pain in younger women. 1 , 21 Another retrospective study found that women reported a decrease in nonpelvic pain symptoms such as headache, osteoarthritis, and back pain after the menopausal transition. 23 Additional research replicating this finding in a prospective, longitudinal cohort is warranted, and well as possible physiologic and psychological factors that may influence this relationship. Strengths of this study include detailed assessment of pain and other symptoms with a broad range of validated measures. However, this tertiary referral population may not be generalizable to other clinical settings. Notably, this cohort is largely white and non-Hispanic, which further limits generalizability. We believe that the fact that the referral population is more homogeneous than that of our immediate surrounding geographic region, although is reflective of our larger statewide catchment area, may highlight the disparity in access to specialty care for pelvic pain and endometriosis, which could contribute to disparate rates of endometriosis and worse outcomes related to endometriosis in patients of color. An important limitation is that approximately 30% of the patients in this study had not previously undergone surgery for pain, and therefore it is possible that some of these patients could have endometriosis that had not yet been diagnosed. We attempted to account for this limitation by including prior procedure for pain in our multivariable model, even though there was no difference in this covariate between the FM tertiles in unadjusted analyses. This covariate was only significantly associated with frequency of pain, and prior surgery was associated with a greater number of painful days per month. Similarly, the majority of patients with surgically confirmed endometriosis had undergone surgery at other facilities and we were unable to reliably determine stage of endometriosis given the inconsistent availability of operative reports. Therefore, history of endometriosis was considered a binary outcome in this analysis. In addition, this database includes only cross-sectional data from the initial clinic visit and did not prospectively follow patients, so we are unable to include information on rate or operative findings from subsequent surgery that patients underwent as part of their care in this referral clinic, so it is possible that we may have underestimated prevalence of endometriosis in this population. Our assessment of pelvic myofascial pain is performed by a core group of clinicians who were trained in a standard pelvic floor myofascial examination technique by a single senior physician, but we cannot rule out some variability in assessment given that examinations have been performed by multiple providers. Finally, this cross-sectional study demonstrates an important association between FM Survey Score and pain morbidity, but future longitudinal studies are required to determine definitive conclusions regarding pain outcomes. Prior research has demonstrated that patients with CPP display clinical, psychophysical, and neuroimaging findings consistent with nociplastic pain. Here we demonstrate that greater nociplastic pain as reflected by FM Survey Score is associated with multiple measures of pain and dysfunction even when controlling for clinical characteristics assumed to have a strong relationship with these measures. Together, these findings raise the possibility that a simple screening tool for nociplastic pain might provide clinically actionable information without the need for deep neurobiological phenotyping and may inform development of personalized management strategies for CPP.