Background
Chronic pelvic pain affects ∼15% of women, and is associated with significant societal cost and impact on women’s health. Identifying factors involved in chronic pelvic pain is challenging due to its multifactorial nature and confounding between potential factors. For example, while some women with endometriosis have chronic pelvic pain, there may be comorbid conditions that are implicated in the chronic pelvic pain rather than the endometriosis itself.
Objective
We sought to explore multifactorial variables independently associated with the severity of chronic pelvic pain in women.
Study Design
We used baseline cross-sectional data from an ongoing prospective cohort, collected from patient online questionnaires, physical examination, and physician review of medical records. Participants were recruited from a tertiary referral center for endometriosis and chronic pelvic pain in Vancouver, British Columbia, Canada, from December 2013 through April 2015. Exclusion criteria included menopausal status or age >50 years. Primary outcome was self-reported severity of chronic pelvic pain in the last 3 months (0-10 numeric rating scale). Potential associated factors ranged from known pain conditions assessed by standard diagnostic criteria, validated psychological questionnaires, musculoskeletal physical exam findings, as well as pain-related, reproductive, medical/surgical, familial, demographic, and behavioral characteristics. Mann-Whitney, Kruskal-Wallis, or Spearman test were used to identify variables with an association with the primary outcome ( P < .05), followed by multivariable linear regression to control for confounding and to identify independent associations with the primary outcome ( P < .05).
Results
Overall, 656 women were included (87% consent rate), of whom 55% were diagnosed with endometriosis. The following factors were independently associated with higher severity of chronic pelvic pain: abdominal wall pain ( P = .005), pelvic floor tenderness ( P = .004), painful bladder syndrome ( P = .019), higher score on Pain Catastrophizing Scale ( P < .001), adult sexual assault ( P = .043), higher body mass index ( P = .023), current smoking ( P = .049), and family history of chronic pain ( P = .038). Severity of chronic pelvic pain was similar between women with and without endometriosis.
Conclusion
Multifactorial variables independently associated with severity of chronic pelvic pain were identified, ranging from myofascial/musculoskeletal, urological, family history, and psycho-social factors. Continued research is required to validate these factors and to determine whether any are potentially modifiable for the management of chronic pelvic pain.
Introduction
Chronic pelvic pain is defined as continuous or intermittent pain in the lower abdomen or pelvis lasting for at least 3-6 months, which is differentiated from dysmenorrhea and pain during sexual intercourse. Chronic pelvic pain affects nearly 15% of women of reproductive age worldwide. There are significant costs of chronic pelvic pain due to physician visits, medication use, surgical interventions, emergency visits, and hospital admissions; outpatient management alone accounts for an estimated $881.5 million in annual costs based on data from 1994 in the United States.
A systematic review reported the association of noncyclical chronic pelvic pain with a variety of gynecologic, medical, and psycho-social factors. However, this review was limited by the absence of physical examination findings, since musculoskeletal contributors to chronic pelvic pain can only be assessed by clinician examination. As well, individual factors could not be assessed simultaneously to account for confounding and discern independent associations. For example, there has been greater recognition of the role of central sensitization in chronic pelvic pain, which conceptually refers to an increased sensitivity and amplification of nociceptive signaling in the central nervous system, and which could underlie a range of different conditions associated with chronic pelvic pain.
In December 2013, we initiated an ongoing prospective research cohort of women at a tertiary referral center for endometriosis and chronic pelvic pain. This cohort is highly phenotyped utilizing a multifactorial approach, to provide a simultaneous assessment of different factors potentially involved in chronic pelvic pain. The goal of the cohort is to identify factors that have an independent association with chronic pelvic pain at baseline and over the next 5 years.
In this exploratory study, we analyzed the baseline cross-sectional data from this prospective cohort to identify multifactorial variables associated with severity of chronic pelvic pain.
Materials and Methods
Study setting
This study involves analysis of cross-sectional baseline data from an ongoing prospective cohort of women at the Center for Pelvic Pain and Endometriosis at the BC Women’s Hospital and Health Center, a tertiary referral center for the province of British Columbia, Canada. After informed consent, participants in the prospective cohort were recruited from women newly referred or re-referred to the center who had a full assessment, defined as a complete history and examination, from December 2013 through April 2015. For patients seen as a new referral, this was the initial visit at the center; for patients seen as a re-referral, this was the return visit since the re-referral was made (although they had been seen before as a new referral, but before the study period). For re-referrals, the average time between the 2 visits was 33 months. The center specializes in the management of chronic pelvic pain and endometriosis through an interdisciplinary approach that includes pain education, pelvic floor physiotherapy, counseling, and medical/surgical management, including advanced excisional laparoscopic surgery. All data from this study were entered prospectively into an online Research Electronic Data Capture (REDCap) database. The study was approved by ethics board of the University of British Columbia (H11-02882).
Inclusion/exclusion criteria
The inclusion criteria for the study were: (1) patient completion of an online questionnaire, prior to the full assessment by a gynecologist; and (2) full assessment by a gynecologist with advanced training in pelvic pain, based on the patient questionnaires, review of medical records, and a standardized physical examination. The online questionnaire was completed on average 4 weeks before the full assessment. Physicians entered data in real time during the assessment. Exclusion criteria included patient age >50 years or postmenopausal status (including surgical menopause), since the natural history of endometriosis is to resolve at the time of menopause. Study flow chart is shown in the Figure .
Primary outcome
Primary outcome was patient-reported severity of chronic pelvic pain in the last 3 months from the online questionnaire, which was rated on a numeric rating scale from 0-10 as recommended for endometriosis research (where 0 represents “no pain” and 10 “worst pain imaginable”). In the questionnaire, patients were asked to specifically differentiate chronic pelvic pain from dysmenorrhea, deep or superficial dyspareunia, dyschezia, or back pain. The wording used for the questionnaire was decided upon after discussion and consensus between the gynecologists at the center ( Table 1 ). To validate the wording, we reviewed 50 charts and correlated the physician dictated report with the patient questionnaire response. We found a 98% concordance rate (49/50) between the physician diagnosis of chronic pelvic pain and the patient rating of chronic pelvic pain in the questionnaire. That is, when the physician diagnosed the presence of chronic pelvic pain, the patient had rated chronic pelvic pain between 1-10; and when the physician stated the absence of chronic pelvic pain, the patient had rated chronic pelvic pain as 0.
| Category | Wording | Rating a |
|---|---|---|
| Bleeding | In past 3 mo, about how many days have you had menstrual (vaginal) bleeding? | 0, 1–10, 10–20, 20–40, 40–60, 60–80, >80 |
| Dysmenorrhea | In past 3 mo, how painful were your menstrual cramps when bleeding? | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, No bleeding |
| Deep dyspareunia | If you are currently sexually active or have been sexually active in past, how painful was deep penetration during sexual activity? | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, Never sexually active |
| Superficial dyspareunia | If you are currently sexually active or have been sexually active in past, how painful was initial penetration (entry) during sexual activity? | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, Never sexually active |
| Dyschezia | In past 3 mo, how painful were bowel movements? | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 |
| Chronic pelvic pain | In past 3 mo, have you had pelvic pain when not bleeding, not during sexual activity, and not during bowel movements (yes/no)? If yes, how painful was this other pelvic pain? | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 |
| Follow-up questions | Participants were then asked to characterize this pelvic pain through additional questions. Questions involved location (eg, right, center, left) and characteristics (eg, burning, pins and needles). | Yes/no |
| Back pain | In past 3 mo, have you had back pain (yes/no)? If yes, how painful was this back pain? | 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 |
a Participants were asked to rate their pain from 0 (no pain) to 10 (worst pain imaginable).
Factors potentially associated with chronic pelvic pain
For pain conditions, the gynecologist classified endometriosis in 3 categories: endometriosis present (previous surgical diagnosis, current palpable nodule, or current endometrioma on endovaginal ultrasound); endometriosis absent (negative laparoscopy or no clinical suspicion by the gynecologist); and endometriosis suspected (no previous laparoscopy but clinically suspected by the gynecologist based on history and tenderness on endovaginal ultrasound-assisted pelvic exam). Irritable bowel syndrome was diagnosed based on the Rome III criteria. Painful bladder syndrome was diagnosed based on the clinical criteria of the American Urological Association or the International Continence Society. We also included physician impression of adhesions (based on history, eg, adhesions seen at prior surgery) and fibroids >4 cm on bedside endovaginal ultrasound.
For musculoskeletal physical exam findings, body mass index (BMI) was measured at the time of the gynecologist assessment. Pelvic floor (levator ani) tenderness was also assessed as we previously descsribed. The gynecologist also tested for abdominal wall pain (of which a common cause is myofascial trigger points) by using the Carnett test, where tenderness worsens with contraction of the abdominal wall. The Carnett test has a sensitivity of 78% and specificity of 88%, with an interobserver reliability 76%, for abdominal wall pain. The rationale behind the Carnett test is that a contracted abdominal wall (eg, through a sit-up) will prevent the examiner from palpating intraperitoneal structures, and thus palpation is primarily of the abdominal wall layer where myofascial trigger points are present. The Carnett test was performed separately in 3 locations in the lower abdomen: right lower quadrant, left lower quadrant, and the suprapubic area. Abdominal wall pain was considered present if the Carnett test was positive (ie, abdominal tenderness that is worse with abdominal wall contraction) in at least 1 location.
The validated psychological questionnaires were administered online before the gynecological assessment. We used the validated Patient Health Questionnaire (PHQ)-9 to assess severity of depression symptoms, and the validated Generalized Anxiety Disorder (GAD)-7 questionnaire to assess severity of anxiety symptoms. Pain catastrophizing was assessed using the validated Pain Catastrophizing Scale (PCS), which includes questions about magnification, rumination, and helplessness.
From the patient online questionnaire done prior to the gynecological assessment, we also obtained data from pain-related, reproductive, medical-surgical, familial, and demographic and behavioral variables. Pain-related factors included duration of pelvic pain, age when pelvic pain started, and number of laparoscopies and laparotomies for pelvic pain. Reproductive factors were age at menarche; overall severity of dysmenorrhea (0-10) over one’s lifetime; number of bleeding days in the last 3 months; current use of hormonal suppression; pregnancy history; history of intrauterine device use; tubal ligation; and previous diagnosis of chlamydia, gonorrhea, or pelvic inflammatory disease. Medical-surgical factors included fibromyalgia, history of ruptured appendix, and previous hernia surgery. Family history of chronic pain in mother, father, or other relative was also noted. Additionally, demographic and behavioral factors included re-referral vs new referral; age; ethnicity; caffeine intake (none, not daily, 1-3 cups/d, 4-6 cups/d, >6 cups/d); current smoking; alcohol (drinks/wk); current recreational drug use; currently working; marital status; education (some high school, graduated high school, some college, graduated 2-year college, graduated 4-year college, postgraduate); annual income (<$20,000, $20,000-40,000, $40,000-60,000, $60,000-80,000, $80,000-100,000, >$100,000); dietary patterns of eating red meat, dairy products, vegetables/fruits, fish, fiber, fatty foods, sugary foods (never, rarely, few times/wk, daily); exercise (rarely, 1-2 times/wk, 3-4 times/wk, daily); history of physical and sexual abuse in childhood (using selected questions from the Childhood Trauma Questionnaire : as a child, “someone in my family hit me so hard that it left me with bruises or marks” and “someone tried to touch me in a sexual way or tried to make me touch them”); and physical and sexual assault in adulthood (using questions designed in-house: as an adult, “someone hit me so hard it left bruises or marks” and “someone forced me to have sex when I did not want this”).
Data analysis and statistics
Statistical comparisons were performed between each factor and the primary outcome (severity of chronic pelvic pain 0-10). Mann-Whitney test was used for binary variables, Kruskal-Wallis test for categorical variables (≥3 categories), and Spearman rank correlation test for continuous/ordinal variables. Factors with a significant association with severity of chronic pelvic pain ( P < .05, n = 26) were then entered into a multivariable linear regression model, followed by backward elimination to obtain the final regression model. Binary and continuous/ordinal variables were entered directly into the model, while categorical variables (3 categories: yes, no, don’t know) were recoded into 2 binary variables with “no” as the reference. First, all variables were initially entered into the model, as we had sufficient sample size (N >600 or >20 cases per variable). Then sequential backward elimination was performed, where the variable with the smallest partial correlation and P value above a removal threshold ( P ≥ .05) was removed at each step. This was continued until all variables in the regression model had a P value below the removal threshold (ie, P < .05).
Data analysis was performed using software (SPSS, V22.0; IBM Corp, Armonk, NY). P value <.05 was considered statistically significant. Means were provided with SD, and medians provided with ranges. Observations with missing values were excluded from the regression model if the proportion of missing values was <5%. For depression (PHQ-9), anxiety (GAD-7), and PCS score, missing values (5.6%) were imputed for the regression model using multiple imputation procedures (proc MI, SAS software, Version 9.4; SAS Institute Inc, Cary, NC).
Results
A total of 656 women met the study criteria, with an 87% consent rate ( Figure ). Demographic characteristics are summarized in Table 2 .
| Variable | Mean ± SD, median [range], or % (x/n) |
|---|---|
| Age, y | 34.5 ± 7.6 |
| BMI, kg/m 2 | 25.3 ± 5.7 |
| Underweight, <18.5 | 4% (27/462) |
| Normal weight, 18.5–24.9 | 52% (342/642) |
| Overweight, 25.0–29.9 | 25% (160/642) |
| Obese, >30.0 | 18% (113/642) |
| Nulligravid | 49% (314/642) |
| Caucasian | 74% (472/641) |
| Chronic pelvic pain severity, range 0–10 | 5.8 ± 3.2 |
| New referral vs re-referral | 76% (498/656) vs 24% (158/656) |
| Irritable bowel syndrome | 53% (350/656) |
| Painful bowel syndrome | 43% (281/656) |
| Abdominal wall pain | 27% (179/656) |
| Endometriosis stage a | 57% (373/656) |
| I/II b | 38% (141/373) |
| III/IV b | 43% (162/373) |
| Unknown b , c | 9% (36/373) |
| Time between first surgical diagnosis of endometriosis and current study, y d | 6.6 ± 5.7 |
| No. of previous surgeries for pelvic pain | 1 [0–22] |
a See definition in “Materials and Methods”: previous surgical diagnosis, or current nodule or endometrioma
b Stage from those with previous surgical diagnosis
c Could not be determined from previous operative records
d Time between first laparoscopic ablation or excision of endometriosis, and full assessment in current study.
The following factors were significantly associated with increased severity of chronic pelvic pain ( P < .05) ( Tables 3 and 4 ): higher BMI, abdominal wall pain, pelvic floor tenderness, greater depression symptoms (PHQ-9), greater anxiety symptoms (GAD-7), higher PCS score, painful bladder syndrome, irritable bowel syndrome, physician impression of adhesions, greater number of laparoscopies, greater severity of dysmenorrhea, older age at menarche, history of pregnancy termination, history of cesarean delivery, fibromyalgia, hernia surgery, family history of chronic pain, lower education, lower income, no current employment, unmarried, current smoking, physical abuse in childhood, sexual abuse in childhood, physical abuse as an adult, and sexual assault as an adult. Among current smokers, the number of pack-years of smoking was also positively correlated with the severity of chronic pelvic pain (Spearman r = 0.24, n = 88, P = .022). In contrast, endometriosis was not associated with chronic pelvic pain severity ( Table 4 ). The distribution of each of the potential risk factors in women with varying severity of pain (in 3 categories: 0-3, 4-6, 7-10) is presented in Table 5 , demonstrating how characteristics differed by pain severity for those variables associated with the primary outcome.
| Factor | N | Spearman a | P |
|---|---|---|---|
| BMI | 642 | 0.13 | <.001 |
| Depression (PHQ-9) | 619 | 0.33 | <.001 |
| Catastrophizing (PCS) | 619 | 0.33 | <.001 |
| Anxiety (GAD-7) | 619 | 0.20 | <.001 |
| Laparoscopies, n | 656 | 0.13 | <.001 |
| Laparotomies, n | 656 | 0.05 | .17 |
| Age, y | 656 | –0.02 | .59 |
| Menarche, y | 639 | –0.1 | .014 |
| Bleeding d/3 mo, n | 636 | –0.08 | .06 |
| Age pain started, y | 646 | –0.06 | .15 |
| Duration since pain started | 656 | 0.05 | .22 |
| Severity of dysmenorrhea (scale 0–10) | 656 | 0.17 | <.001 |
| Education level b | 638 | –0.18 | <.001 |
| Income b | 638 | –0.09 | .032 |
| Exercise b | 638 | –0.03 | .45 |
| Caffeine, cups/d | 638 | 0.03 | .5 |
| Alcohol, drinks/wk | 633 | –0.05 | .2 |
| Red meat b | 638 | –0.03 | .51 |
| Dairy product b | 638 | 0.01 | .74 |
| Vegetables/fruits b | 637 | –0.03 | .51 |
| Fish b | 638 | –0.05 | .18 |
| Fiber b | 638 | –0.02 | .69 |
| Fatty food b | 638 | 0.05 | .19 |
| Sugary food b | 638 | –0.04 | .36 |
a Spearman correlation coefficient
| Variable | N | Mean CPP (0-10) | Statistic a | P |
|---|---|---|---|---|
| Endometriosis | ||||
| Yes | 373 | 5.8 ± 3.2 | 0.319 | .85 |
| Suspected b | 166 | 5.9 ± 3.0 | ||
| No | 117 | 5.8 ± 3.5 | ||
| Irritable bowel syndrome | ||||
| Yes | 350 | 6.2 ± 3.1 | 46,551.5 | .004 |
| No | 306 | 5.4 ± 3.3 | ||
| Painful bladder syndrome | ||||
| Yes | 281 | 6.6 ± 2.6 | 42,959.5 | <.001 |
| No | 375 | 5.2 ± 3.4 | ||
| Abdominal wall pain | ||||
| Yes | 179 | 7.0 ± 2.4 | 54,460.5 | <.001 |
| No | 477 | 5.4 ± 3.3 | ||
| Pelvic floor (levator ani) tenderness | ||||
| Yes | 193 | 6.9 ± 2.8 | 54,564 | <.001 |
| No | 444 | 5.4 ± 3.2 | ||
| Re-referral | ||||
| Yes | 158 | 5.9 ± 3.2 | 36,084.5 | .484 |
| No | 498 | 5.8 ± 3.2 | ||
| Adhesions (physician impression) | ||||
| Yes | 90 | 6.6 ± 2.8 | 29,488 | .015 |
| No | 566 | 5.7 ± 3.2 | ||
| Fibroids >4 cm | ||||
| Yes | 18 | 5.8 ± 2.7 | 5055.5 | .77 |
| No | 585 | 5.7 ± 3.2 | ||
| Ethnicity, Caucasian | ||||
| Yes | 472 | 6.0 ± 3.1 | 36,957.5 | .15 |
| No | 169 | 5.5 ± 3.4 | ||
| Hormonal suppression | ||||
| Yes | 237 | 5.6 ± 3.3 | 46,865.5 | .23 |
| No | 419 | 6.0 ± 3.1 | ||
| History of IUD use | ||||
| Yes | 96 | 6.2 ± 3.1 | 24,170 | .23 |
| No | 545 | 5.8 ± 3.2 | ||
| Tubal sterilization | ||||
| Yes | 21 | 6.6 ± 3.2 | 5321.5 | .15 |
| No | 620 | 5.8 ± 3.2 | ||
| Chlamydia | ||||
| Yes | 72 | 6.2 ± 3.1 | 18,982 | .31 |
| No | 569 | 5.8 ± 3.2 | ||
| Gonorrhea | ||||
| Yes | 7 | 5.0 ± 3.5 | 1839.5 | .43 |
| No | 634 | 5.8 ± 3.2 | ||
| Pelvic inflammatory disease | ||||
| Yes | 26 | 6.1 ± 3.1 | 7660 | .72 |
| No | 615 | 5.8 ± 3.2 | ||
| Fibromyalgia | ||||
| Yes | 36 | 7.1 ± 2.2 | 8465 | .024 |
| No | 605 | 5.8 ± 3.2 | ||
| Ruptured appendix | ||||
| Yes | 16 | 6.9 ± 2.3 | 4180.5 | .26 |
| No | 625 | 5.8 ± 3.2 | ||
| Hernia surgery | ||||
| Yes | 20 | 7.0 ± 2.6 | 4228 | .014 |
| No | 621 | 5.8 ± 3.2 | ||
| Fall on pelvis/tailbone | ||||
| Yes | 210 | 6.1 ± 3.0 | 42,077 | .17 |
| No | 429 | 5.7 ± 3.2 | ||
| Termination | ||||
| Yes | 98 | 6.4 ± 2.9 | 23,085 | .043 |
| No | 540 | 5.7 ± 3.2 | ||
| Miscarriage | ||||
| Yes | 136 | 6.1 ± 3.1 | 32,465 | .36 |
| No | 503 | 5.8 ± 3.2 | ||
| Cesarean delivery | ||||
| Yes | 111 | 6.4 ± 3.1 | 24,918.5 | .012 |
| No | 528 | 5.7 ± 3.2 | ||
| Vaginal delivery | ||||
| Yes | 164 | 6.1 ± 3.0 | 36,723.5 | .27 |
| No | 475 | 5.7 ± 3.2 | ||
| Family history of chronic pain | ||||
| Yes | 173 | 6.6 ± 2.8 | 16.285 | <.001 |
| No | 322 | 5.4 ± 3.3 | ||
| Do not know | 143 | 5.9 ± 3.1 | ||
| Working | ||||
| Yes | 464 | 5.6 ± 3.1 | 33,580.5 | .001 |
| No | 174 | 6.4 ± 3.1 | ||
| Married | ||||
| Yes | 288 | 5.6 ± 3.2 | 45,286.5 | .026 |
| No | 350 | 6.1 ± 3.1 | ||
| Smoking | ||||
| Yes | 89 | 7.2 ± 2.6 | 17,195 | <.001 |
| No | 550 | 5.6 ± 3.2 | ||
| Recreational drug use | ||||
| Yes | 67 | 6.5 ± 2.6 | 2.388 | .30 |
| No | 495 | 5.7 ± 3.2 | ||
| Do not know | 70 | 6.0 ± 3.3 | ||
| Physical abuse, child | ||||
| Yes | 84 | 6.8 ± 2.9 | 10.977 | .004 |
| No | 500 | 5.7 ± 3.2 | ||
| Do not know | 47 | 5.2 ± 3.0 | ||
| Sexual abuse, child | ||||
| Yes | 110 | 6.9 ± 2.6 | 15.967 | <.001 |
| No | 474 | 5.6 ± 3.2 | ||
| Do not know | 49 | 5.3 ± 3.0 | ||
| Physical abuse, adult | ||||
| Yes | 108 | 6.6 ± 3.0 | 11.96 | .003 |
| No | 474 | 5.7 ± 3.1 | ||
| Do not know | 48 | 5.4 ± 3.0 | ||
| Sexual assault, adult | ||||
| Yes | 86 | 7.1 ± 2.6 | 18.535 | <.001 |
| No | 499 | 5.6 ± 3.2 | ||
| Do not know | 48 | 5.4 ± 3.0 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
