Association of Central Sensitization Inventory Scores With Pain Outcomes After Endometriosis Surgery

Key Points Question Are greater Central Sensitization Inventory scores associated with worse pain outcomes after endometriosis surgery? Findings In this longitudinal presurgery and postsurgery cohort study of 239 people with endometriosis, higher Central Sensitization Inventory scores at baseline were statistically significantly associated with higher chronic pelvic pain scores at follow-up. Meaning These findings suggest that the Central Sensitization Inventory may be used preoperatively by gynecologists to counsel patients on expected postoperative pain outcomes and identify people who may have persistent pain after surgery.


Introduction
3][4][5] Surgery is a common treatment for endometriosis, including fertilitysparing surgery (eg, excision of lesions) and hysterectomy (with or without oophorectomy, ideally combined with excision). 6Potential reasons for proceeding to surgery in this population include persistent pain or adverse effects with hormonal therapy, trying to conceive, or patient choice.However, surgery is not adequate for complete resolution of symptoms in some people. 7,8operation rates after conservative fertility-sparing surgery can reach 50% within 5 years, suggesting that pain mechanisms beyond the endometriosis lesions themselves may be involved. 9 addition, recurrent pain and reoperation can occur in patients undergoing hysterectomy. 10Pain mechanisms in endometriosis are multifactorial. 11 note, central sensitization has been increasingly recognized as a crucial component in the pathogenesis of endometriosis-associated pain, occurring alongside the peripheral nociceptive contributors on a continuum. 113][14][15][16] An increasing body of evidence for central sensitization in endometriosis involves the use of neurophysiologic testing, such as quantitative sensory testing (QST), and magnetic resonance imaging (MRI).8][19] Brain MRI studies have identified regional changes in brain morphology and function, such as thalamic gray matter volume reduction and alterations in functional connectivity in endometriosis-associated CPP. 20,21Such changes in brain function and morphologic mechanisms may contribute to the development of chronic pain and associated comorbidities, such as mood disorders and cognitive impairment. 20,21A limitation of QST and MRI is that they are less feasible for implementation into clinical practice.
3][24][25] This questionnaire assesses systemic pain and pain-related symptoms and mental health but is not a neurophysiological measure and thus not a direct marker of central sensitization. 26A previous study 27 found that the number of CSSs was significantly correlated with increasing pain scores and CSI scores in endometriosis.Although the CSI is not a direct or exclusive measure of central sensitization, 26,[28][29][30] it is able to discriminate between patients with CSSs, such as fibromyalgia, and healthy controls. 25,31,32,34 Endometriosis costs approximately $16 000 annually per patient in the US, 35 and at a tertiary center, patients with pelvic pain reported up to 20 operations before referral, with a possible explanation being unrecognized and untreated central sensitization and overreliance on surgery. 36us, an important clinical problem in the management of endometriosis is the ability to identify individuals who may not benefit from surgery in order to prevent unnecessary additional operations.
Currently, there is a lack of clinical tools to counsel patients with endometriosis on their probable In this prospective, longitudinal study, the association between CSI and persistent pain after endometriosis surgery was explored, with a primary outcome of CPP and secondary outcomes of dysmenorrhea, deep dyspareunia, dyschezia, and back pain.We hypothesized that higher baseline CSI scores would be associated with worse pain outcomes.In a corollary analysis, we also measured the change in CSI score after surgery to assess the association of surgery with symptoms of central sensitization.

Participants
Prospective, longitudinal data from a patient registry, the Endometriosis Pelvic Pain Interdisciplinary Cohort (EPPIC), were analyzed at a tertiary center for endometriosis and pelvic pain.The EPPIC registry has been described previously. 27,36After patients provide consent to EPPIC, the following data sources are entered prospectively and in real time into the registry: ( withdrawn consent.Follow-up questionnaires are sent 1 and 2 years after baseline, although there is variability in when participants actually complete the questionnaires.We used the most recently completed follow-up questionnaire available.Our predetermined goal was to follow up patients through approximately 18 months after surgery.The study cohort is presented in Figure 1.

Measures
The primary outcome was CPP at follow-up.The secondary outcomes were deep dyspareunia, dysmenorrhea, dyschezia, and back pain.All the pain scores were rated on an 11-point numeric rating scale (0-10, with 0 indicating no pain and 10 indicating worst pain imaginable), and because assumptions of linear regression were not met (normality), we then categorized the pain scores as none to mild (scores, 0-3), moderate (scores, 4-6), and severe (scores, 7-10) for ordinal regression.

JAMA Network Open | Obstetrics and Gynecology
In the registry, follow-up questionnaires are sent at specific intervals (eg, 1 year and 2 years after baseline), but there is variability when individuals complete the questionnaire: for each participant, we used the most recent follow-up questionnaire available (ie, longest duration of follow-up).
The main variable of interest was the CSI score at baseline measured from 0 to 100, consisting of 25 self-reported questions rated from 0 to 4 (never, rarely, sometimes, often, and always, respectively). 25

Chronic Pelvic Pain
Higher baseline CSI scores were significantly associated with higher CPP scores at follow-up (odds ratio [OR], 1.02; 95% CI, 1.00-1.03;P = .02),controlling for baseline CPP (Table 1).Of note, although patients with higher CSI scores showed improvement in CPP on average, the degree of improvement was less than for patients with lower CSI scores; that is, patients with higher CSI scores at baseline were more likely to still have persistently elevated pain scores at follow-up (eFigure in Supplement 1).

JAMA Network Open | Obstetrics and Gynecology
Older age and hysterectomy (vs conservative surgery) were associated with less severe CPP scores at follow-up.Stage of endometriosis was not significantly associated with CPP at follow-up.In the multivariable model, higher CSI scores at baseline remained significantly associated with a more severe CPP score at follow-up (OR, 1.02; 95% CI, 1.00-1.03;P = .03).We repeated the analyses after removing the 8 people with bilateral salpingo-oophorectomy, and the results were similar (eTable 4 in Supplement 1).We also considered the potential impact of the duration of follow-up (surgery to follow-up questionnaire).The duration of follow-up was not associated with CPP at follow-up and did not change the association with CSI (eTable 5 in Supplement 1).

Other Pain Scores
We excluded individuals without menses when evaluating dysmenorrhea and those not sexually active when evaluating deep dyspareunia (eTable 3 in Supplement 1).Higher CSI scores at baseline were significantly associated with higher scores for deep dyspareunia (OR, 1.03; 95% CI, 1.01-1.04;P = .004),dyschezia (OR, 1.03; 95% CI, 1.01-1.04;P < .001),and back pain (OR, 1.02; 95% CI, 1.00-1.03;P = .02)at follow-up, controlling for baseline pain scores (Table 2).The associations between higher CSI scores and more severe deep dyspareunia, dyschezia, and back pain at follow-up remained significant in the multivariable regression models (Table 2).In addition, although CSI scores were not significantly associated with dysmenorrhea on the univariable model, CSI was significantly associated in the multivariable model adjusted for surgery group (Table 2).This analysis was followed up with a subanalysis of people who underwent conservative surgery only, and we confirmed that CSI was significantly associated with dysmenorrhea at follow-up (eTable 6 in Supplement 1).See eTable 7 in Supplement 1 for multivariable models before backward elimination.

CSI and CSS
There was a statistically significant, but clinically small, improvement in the CSI score at baseline compared with the CSI score at follow-up after surgery (mean [SD], from 43.8 [18.2] to 41.7 [18.9]; effect size, 0.13) (eTable 3 in Supplement 1).We identified a significant association between the number of CSSs and pelvic pain-related comorbidities at baseline and increasing CSI scores at both baseline (r = 0.71, P < .001)and follow-up (r = 0.45, P < .001)(Figure 2).As illustrated in Figure 2, CSI scores remained persistently high on average after surgery, particularly in patients with high baseline CSI scores related to a multiple coexistent CSS.

Discussion
In this prospective study of 239 patients with endometriosis undergoing surgery at a tertiary center, significant reductions of moderate effect size in pain scores were observed on average but with some patients experiencing persistent pain.We found that the baseline CSI score was associated with higher pain scores at follow-up after surgery (both conservative surgery or hysterectomy), including for CPP, deep dyspareunia, dyschezia, and back pain.Furthermore, there was a statistically significant but clinically small reduction in the CSI score itself after surgery; however, CSI scores remained persistently high, especially in those with multiple CSSs.
Our findings suggest that patients with endometriosis with a higher degree of symptom severity (possibly indirectly related to central sensitization) as measured by the CSI at baseline had worse pain outcomes after surgery compared with those with a lesser degree of symptom severity.
This finding is in line with our hypothesis because surgical treatment of peripheral pain generators (eg, endometriosis lesions) may not directly treat central nervous system pain generators (eg, central sensitization).Similar to our study, in people who underwent spinal surgery, higher preoperative CSI scores were associated with higher postoperative scores on the Pain Disability Questionnaire. 46rthermore, Roh et al 47 found that centrally mediated symptoms measured by the CSI were associated with poor functional outcome at 3 months after carpal tunnel surgery, although this a Excluding those not sexually active ever at baseline and/or not sexually active for the last 3 months at follow-up.
b Each row is a univariable analysis that is adjusted for baseline pain scores.

Follow-up Baseline
A greater number of CSSs and pelvic pain-related comorbidities was associated with high CSI scores at both baseline (r = .71,P < .001)and follow-up (r = .45,P < .001).association diminished at 12 months.A prospective, observational study by As-Sanie et al 48 found that the odds of persistent pelvic pain at 6 months after hysterectomy for benign indications increased by 27% for every 1-point increase in central sensitization pain on the Fibromyalgia Survey.
Together, these studies 27,46 indicate that a higher degree of centralized pain may be associated with persistent chronic pain after surgery.
Although pain scores significantly decreased on average with moderate effect size after surgery in our population, there was a statistically significant, but small effect size, change in CSI scores from baseline to follow-up after surgery.This finding is consistent with studies 16,49 using QST, which found statistically significant but relatively small changes in pain thresholds after endometriosis surgery.Of particular concern were individuals with high CSI scores and multiple CSS comorbidities for whom CSI scores remained persistently high after surgery.These observations suggest that surgery may have little effect on many symptoms of central sensitization, despite improvements in pain.It may be that systemic symptoms of central sensitization, once developed, have a tendency to persist over time despite improvements in other pain-related parameters.However, it remains possible that with longer-term follow-up there may be additional gradual reduction in CSI scores after endometriosis surgery.

Strengths and Limitations
This study has several strengths, including access to a large patient population from a tertiary center with a 71.0%follow-up rate.This study includes a prospective registry with validated clinical questionnaires (eg, CSI), physical examination, and surgical findings.Endometriosis surgery was also performed by endometriosis specialist surgeons, thereby reducing the possibility of the CPP variable being influenced by significant residual disease.Of note, stage of endometriosis was not associated with the follow-up pain outcomes in our analyses.This finding suggests that inadequate excision and recurrence of disease were likely not factors, even in the high-stage group, because higher pain at follow-up was not found in the people with high-stage disease.
This study also has limitations, including the fact that the CSI mainly measures indirect downstream symptoms, such as pain intensity and sleep disturbances, and that neurophysiological markers of central sensitization were not analyzed.To conclude that the CSI is a useful tool to identify pain sensitization and ultimately alter treatment decisions, the CSI should be correlated with surrogate measures of central sensitization, such as QST, in individuals with endometriosis.In addition, the study setting was a referral center for endometriosis-associated chronic pain, and research is needed to validate the findings in other settings, such as with community gynecologists.
This study is also limited by its variability in time between the baseline visit (ie, CSI completion) and index surgery, as well as between index surgery and completion of the follow-up questionnaire (ie, pain outcomes).

Conclusions
In this cohort study of 239 patients with endometriosis, the CSI at baseline was significantly associated with persistent CPP after endometriosis surgery.These findings suggest that the CSI may be a useful tool to assist clinicians in phenotyping endometriosis-associated pain and in counseling patients on the expected outcomes after surgery.
Association of Central Sensitization Inventory Scores With Endometriosis Surgery Pain Outcomes JAMA Network Open.2023;6(2):e230780.doi:10.1001/jamanetworkopen.2023.0780(Reprinted) February 27, 2023 3/12 Downloaded From: https://jamanetwork.com/ on 09/24/2023 Each of the questions targets a different component of downstream symptoms, such as feeling depressed or having low energy, muscle stiffness, poor sleep, and light sensitivity.A higher score indicates greater symptom severity that may be indirectly related to conditions with underlying central sensitization.The CSI is in the public domain and is freely available.37Othervariablesconsidered included demographic characteristics (age, martial status, highest level of education, race, parity, and endometriosis stage), type of surgery (conservative [fertilitysparing] or hysterectomy with or without bilateral salpingo-oophorectomy), and number of CSSs.Patient self-reported race from the EPPIC data registry was included as a variable in this study because there is potential clinical variation by race in endometriosis.As previously reported, we assessed 10 CSSs and pelvic pain-related comorbidities: (1) fibromyalgia, (2) chronic fatigue meet the normality assumption of linear regression when the 0 to 10 values were used for the primary and secondary outcomes.The assumption of proportional odds was met for all ordinal logistic regression models (full likelihood ratio test P > .05and/or by running separate binomial logistic regressions on cumulative dichotomous-dependent variables).Significant variables on univariable ordinal regressions were put into a multivariable model, followed by backward elimination.Bivariate associations (2-tailed) were performed using an unpaired, independent t test or paired-sample t test (continuous variables) or a χ 2 test, Fisher exact test, or Wilcoxon signed rank test (categorical variables).Effect sizes were calculated as follows: (z divided by the square root of B) for the Wilcoxon signed rank test and (mean divided by SD) for the paired-sample t test Cohen d.SPSS Statistics, version 24 (IBM Corp) was used for analysis.Significance was set at a 2-sided P < .05.

Table 1 .
Ordinal Regression Between Chronic Pelvic Pain Score at Follow-up, Controlling for Baseline Chronic Pelvic Pain Score a Each row is a univariable analysis that is adjusted for baseline pain scores.b c Initial multivariable model with CSI, age, and type of surgery, adjusting for baseline pain scores; after backward elimination, type of surgery fell out of the model.The multivariable regression model results before backward elimination are presented in eTable 7 in Supplement 1.

Table 2 .
Ordinal Regression Between Secondary Pain Outcomes at Follow-up (Ordinal) and Variables of Interest, Controlling for Baseline Pain Scores

Table 2 .
Ordinal Regression Between Secondary Pain Outcomes at Follow-up (Ordinal) and Variables of Interest, Controlling for Baseline Pain Scores (continued) Initial multivariable model after backward elimination, adjusting for baseline pain scores.The multivariable regression models before backward elimination are presented in eTable 7 in Supplement 1.Excluding those without menses at baseline and/or follow-up (eg, after hysterectomy) (n = 131).
d e