Tu, Frank F. MD, MPH1,2; Fitzgerald, Colleen M. MD2; Kuiken, Todd MD, PhD2; Farrell, Todd2; Norman, Harden R. MD2
The underlying pathophysiology of pelvic pain conditions remains poorly characterized. Localization of pelvic pain symptoms to specific pain generators is a challenge, due to the convergence of afferent input from both somatic and visceral tissues in central representations of the peripheral nervous system.1 Given the health care effect of pelvic pain (a 3-month prevalence approaching 24% among reproductive-age women, with the U.S. financial burden exceeding $3 billion), effective, mechanism-based treatment is desirable.2,3 Although much attention has been paid to intraabdominal processes, such as endometriosis, experimental evidence suggests that central aberrations in pain processing (enhanced pain sensitivity) are common in pelvic pain.4,5
Traditional experimental measures used to evaluate pain processing include determination of pressure-pain thresholds, where an examiner presses with increasing force on a target area of tissue until the patient reports pain. These measures, obtained using hand-held algometers, have been classically described for fibromyalgia assessment, which is characterized by heightened pain sensitivity to pressure.6 Recently, such techniques have demonstrated that vulvodynia sufferers display both enhanced vulvar and distant site (lower extremity) pain sensitivity, but a related technique failed to demonstrate enhancement of pain sensitivity in selected pelvic dermatomes of women with painful bladder syndrome.7,8 Thus, it remains unclear whether peripheral pain sensitivity in pelvic and/or abdominal dermatomes is a consistent feature of pelvic pain syndromes.
We recently have validated use of an internal pelvic algometer designed to measure vaginal and pelvic floor pain sensitivity (Tu F, Fitzgerald C, Kuiken T, Farrell T, Harden R. Pelvic floor pressure-pain thresholds: a pilot study [abstract]. J Pain 2006;7:S30). Because the pelvic floor has been frequently described as dysfunctional in women with pelvic pain, its intrinsic pain sensitivity should be formally characterized. Pelvic floor pain has been described as a discrete condition, labeled historically as levator ani syndrome or pelvic floor tension myalgia, but diagnostic criteria for these conditions have never been formally validated, and the sensation of pain may simply reflect central neurologic aberrations in nociception rather than peripheral somatic dysfunction.9 Pressure-pain threshold measurement of the pelvic floor may therefore be useful in both defining objective criteria for a putative pelvic floor pain syndrome and for characterizing a state of regional neurologic dysfunction shared by seemingly disparate conditions, such as painful bladder syndrome, endometriosis, vulvodynia, and irritable bowel syndrome. The present study compares pressure-pain thresholds between women with chronic pelvic pain and healthy controls to estimate the difference in relative pelvic floor pain sensitivities between pain patients and pain-free controls.
MATERIALS AND METHODS
To test our primary hypothesis, that pelvic floor pain thresholds are lower (reflecting enhanced pain sensitivity) among women with chronic pelvic pain compared with healthy controls, we conducted a prospective, comparative observational study. Pain patients were defined as women with a history of pelvic pain for at least 3 months' duration. Pain had to be localized to the area bounded superiorly by the umbilicus and inferiorly by the inguinal ligament. Symptoms could not solely arise from the skin (as in vulvodynia or incisional neuropathic pain), only involve the hip or back, or only occur with menstrual flow (isolated dysmenorrhea). These women were all identified by prescreening otherwise unselected patients presenting for care to a female chronic pelvic pain clinic. Controls could not have an active history of pelvic pain and were recruited from a general obstetrician–gynecologist's office when presenting for their annual examination. All participants were required to be between the ages of 18 years and 55 years, nonpregnant, without active genital tract infection (ie, vaginitis, cystitis, or pelvic inflammatory disease), and could not have a history of genital tract malignancy. Women received $10 for participating in the study. Approval for the study was obtained from the Evanston Northwestern Healthcare Institutional Review Board.
After consenting, participants filled out brief medical history questionnaires (including demographics, menstrual status, date of last period, and queries for pelvic pain-related symptoms and diagnoses). Women with pain also completed the McGill Pain Questionnaire-short form, a 20-item instrument widely used to assess both affective and sensory components of pain intensity.10 The subsequent experimental pain measures were performed on cases during their regular clinical evaluation, before their scheduled pelvic examination, and on controls immediately before and separate from their scheduled evaluation with their own physician.
This study used a previously validated pressure algometer designed for intravaginal examinations, similar to those used in experimental pain assessment in fibromyalgia. The design was based on a previously published study of normal vaginal pressure-pain thresholds.11 Briefly, the algometer consists of a commercially available force sensor (Load Cell Central, Monroeton, PA) fitted under a flat, stainless steel metal cap (1 cm2 in surface area). The metal cap is glued to the end of a plastic thimble, cast to fit the distal tip of the specific examiner's right index finger; the small profile minimizes examinee discomfort from the device itself. The force signal was fed into a laptop computer through an analog-to-digital converter. This tool allows for instantaneous measurement of the pressure level an examiner applies to a specific area of tissue. Pain sensitivity was assessed as the pressure-pain detection threshold (the amount of pressure, in kg/cm2, at which an examinee indicates that the applied pressure has become painful). A computer program tailored to process these measurements (LabView, National Instruments, Austin, TX) continuously displays the amount of applied pressure recorded from the strain gauge and marks the time a participant indicates they first feel pain (by squeezing an attached hand-held trigger). For each site being palpated, the same investigator performed exams wearing the algometer under a clean vinyl examination glove. Pressure was steadily increased at 0.5 kg/cm2/s. This is the same rate chosen for our validation study, because it allows for patient discrimination and response when transitioning between pressure and pain. The maximal pressure tolerated by this system (limited by the strain gauge's inherent material properties) was approximately 3.7 kg/cm2 (Tu et al. Pelvic floor pressure-pain thresholds: a pilot study [abstract]. J Pain 2006;7:S30).
A set of 10 intravaginal measures was performed on each examinee in the lithotomy position. With the introitus oriented to a virtual clock face, the pubococcygeus muscle was measured at 2 and 10 o'clock, the iliococcygeus and coccygeus at 5 and 7 o'clock, and the obturator internus at 3 and 9 o'clock. Two vaginal comparison sites were also measured corresponding to 12 and 6 o'clock (anterior vagina and posterior vagina). All measures were conducted with the investigator's right index finger inserted 5 cm proximal to the introitus, with the exception of the coccygeus, which was measured at the same clock face orientation as the iliococcygeus, but at 10 cm proximal. All pain threshold measurements were separated from one another by a 30-second break. The entire complement of tests was repeated, after a 2-minute break, to derive mean measures of pain threshold for each site—an approach demonstrated in our preliminary pilot studies to provide appropriate retest reliability.
The primary null hypothesis was that average pressure-pain detection thresholds at the pelvic floor would be no different in women with pelvic pain compared with healthy controls. Preliminary power calculations, assuming a parametric test and based on our pilot studies of healthy controls (unpublished observations), showed that a sample size of 24 controls and 12 cases would yield sufficient power to demonstrate a difference of 0.5 kg/cm2 (1.4 kg/cm2 in controls compared with 0.9 kg/cm2 in cases) at an alpha of 0.05 and a power of 80%. The 2:1 ratio of controls to cases was chosen to reflect the ease of finding controls for such studies compared with cases, and we also planned to recruit additional participants to account for the possibility that we might be overestimating the degree of difference between pain patients and controls. All analyses were conducted using Stata 8.0 (StataCorp LP, College Station, TX). Normality of continuous variables was evaluated with the Shapiro-Wilks statistic. Univariable statistics were calculated for demographic variables and for pain thresholds for both groups. Comparison of continuous variables between groups was conducted using Student t test for normal data (after assessing for equal variance with the F test for homogeneity) and Wilcoxon rank sum test for nonnormal data. Categorical data were compared using Fisher exact test. P values for the statistical association between pain thresholds and pain status reflect results of analysis of covariance controlling for the confounding influence of age.7 Correlation between continuous variables was assessed using Pearson's and Spearman's correlation coefficients.
A total of 16 women with chronic pelvic pain and 30 healthy controls completed the study between February and May of 2006. Two of the 16 women with chronic pelvic pain were not able to be fully evaluated. One reported immediate, uncomfortable tonic pain on insertion of the algometer. The other declined to have a second assessment of her pelvic floor after the rest period. Consequently, both women's measures were excluded in the comparisons. We also could not measure pressure-pain detection threshold at the left coccygeus site in two controls and the right coccygeus in one control due to anatomic constraints posed by a prominent adjacent iliococcygeus.
The typical participant in both groups was reproductive age and white. Controls were older on average, and 27% were menopausal compared with 0% of the women with chronic pelvic pain (P=.03). No significant difference was noted in which menstrual phase cycling women were examined in (of these, 43% of controls and 50% of women with chronic pelvic pain were in the luteal phase). Women with pelvic pain were more likely to report a history of interstitial cystitis or painful bladder syndrome, irritable bowel syndrome, endometriosis, or dyspareunia, as well as moderate-to-severe dysmenorrhea. (Table 1). Among women with chronic pelvic pain, the median (with 25th–75th percentile) total, sensory and affective McGill pain inventory scores were 23.5 (12–30), 14 (12–22), and 4.5 (3–9), respectively.
Significantly lower pressure-pain detection threshold values were noted at all eight examination sites among women with pelvic pain when compared with healthy controls. (Table 2). All genital site pressure-pain detection threshold results (pelvic floor and vaginal wall) were correlated with one another (r=0.62–0.91), all P<.001). We therefore collapsed the individual pelvic floor pressure-pain detection threshold results into right- and left-sided mean measures to assess associations between overall pain threshold and subjective pain experience in the pain cohort (McGill scores). Right- and left-sided pelvic floor measures were negatively correlated with McGill sensory subscores (r=–0.57 and –0.67, P=.03 and .009, respectively) as was posterior vaginal wall pressure-pain detection threshold (r=–0.55, P=.04). None of the affective scores were correlated with pressure-pain detection threshold at any site.
Because a quarter of the controls were menopausal compared with none of the women with pelvic pain, we also repeated the analyses excluding all eight of those women, which did not affect the findings (data not shown). Among the women with pelvic pain, we also looked to see whether the presence of particular diagnoses or specific pain complaints was associated with pressure-pain detection threshold values from specific locations in the pelvic floor or vagina. The presence of common pelvic pain subdiagnoses (including dyspareunia, dyschezia, dysuria, irritable bowel syndrome, painful bladder syndrome, or endometriosis), were not significantly associated with pressure-pain detection thresholds at any site, although the power to detect small differences in this study are limited.
This study confirmed our hypothesis that both pelvic floor and vaginal pressure-pain thresholds are lowered in women with chronic pelvic pain compared with healthy controls. This conclusion appears to be applicable to the eight pelvic floor sites assessed and to the two vaginal locations as well. These results are consistent with the observation that pelvic floor pain is more commonly identified by physical therapists on examination in both male and females with pelvic pain to the extent that they conduct their examination focusing on the pelvic floor rather than on the viscera (unpublished observations).12 We also noted correlations between the vaginal and pelvic floor pain thresholds, suggesting that future studies may only need to employ a subset of these measures to characterize pelvic floor and vaginal pain sensitivity. Because the coccygeus measurements may not be uniformly obtainable by virtue of that muscle's posterior and distal location, the observed correlations will allow for exclusion of this assessment from future studies of pelvic floor pressure-pain detection threshold. Additionally, the observation that the two vaginal (nonmuscle) sites also display a lower pain threshold suggests that peripheral somatic pain sensitivity may underlie many pelvic pain states rather than an isolated myofascial pain disorder, such as levator ani syndrome or trigger points. Composite pelvic floor pain thresholds were negatively associated with McGill sensory pain scores in women with pelvic pain, as were the posterior vaginal wall measures, but not the affective pain subscores. These findings are consistent with provocative pain studies in which experimental somatic pain stimulation is less associated with affective, emotional aspects of pain experience, compared with visceral causes.13
The nature of our case ascertainment from an existing pelvic pain clinical practice may have introduced unintended bias in this study. However, because the manufacturing of the algometer to fit the pelvic pain clinician was required along with pretrial training to use the instrument appropriately, we were unable to mask the patient identity from the examiner in this study. Future trials will employ masked performance of these pressure-pain detection threshold evaluations, preferably by a third-party clinician, to avoid any potential examiner-patient interaction bias. One suitable alternative would be to provide algometer training to physical therapists familiar with pelvic floor anatomy and examination. Removing any overt gynecologic perspective from the evaluation paradigm would add another layer of objectivity to pressure-pain detection threshold assessment.
Our data are consistent with many other studies of pelvic pain syndromes, including vulvodynia, painful bladder syndrome, and endometriosis, which suggest both regional and generalized pain sensitivity as characteristic factors.5,7,14 Although Fitzgerald and colleagues8,15 did not find differences in current perception thresholds between painful bladder syndrome and healthy controls, their chosen measure may only reflect changes present after direct peripheral nerve damage rather than enhanced states of central or peripheral pain sensitivity, thought to be prevalent among chronic pain patients.
We also recognize that our study design limited our attention exclusively to the somatic components of chronic pelvic pain (exclusive of a more traditional visceral focus), resulting in an artificial construct. For example, many of our pain patients had histories of moderate-to-severe dysmenorrhea, making it difficult to determine whether their pain intensity is more closely related to enhanced somatic pain sensitivity (ie, the lower pressure-pain detection threshold values observed) or visceral pain sensitivity (ie, abnormalities of their uterine, parametrial, peritoneal, or other structures). Our study does not allow conclusions to be drawn on this subject. The dilemma of sorting out relative somatic and visceral contributions to pelvic pain is further obscured by the anatomic convergence of somatic and visceral pain generators in the lumbosacral dorsal horn such that central pain perception is “anatomically mixed” by default. Perhaps larger studies designed to address individual somatic and visceral contributions as well as somatic–visceral interactions will facilitate our understanding of these two related elements.
1. Berkley KJ, Guilbaud G, Benoist JM, Gautron M. Responses of neurons in and near the thalamic ventrobasal complex of the rat to stimulation of uterus, cervix, vagina, colon, and skin. J Neurophysiol 1993;69:557–68.
2. Zondervan KT, Yudkin PL, Vessey MP, Jenkinson CP, Dawes MG, Barlow DH, et al. The community prevalence of chronic pelvic pain in women and associated illness behaviour. Br J Gen Pract 2001;51:541–7.
3. Mathias SD, Kuppermann M, Liberman RF, Lipschutz RC, Steege JF. Chronic pelvic pain: prevalence, health-related quality of life, and economic correlates. Obstet Gynecol 1996;87:321–7.
4. Bajaj P, Madsen H, Arendt-Nielsen L. Endometriosis is associated with central sensitization: a psychophysical controlled study. J Pain 2003;4:372–80.
5. Ness TJ, Powell-Boone T, Cannon R, Lloyd LK, Fillingim RB. Psychophysical evidence of hypersensitivity in subjects with interstitial cystitis. J Urol 2005;173:1983–7.
6. Fischer AA. Reliability of the pressure algometer as a measure of myofascial trigger point sensitivity. Pain 1987;28:411–4.
7. Giesecke J, Reed BD, Haefner HK, Giesecke T, Clauw DJ, Gracely RH. Quantitative sensory testing in vulvodynia patients and increased peripheral pressure pain sensitivity. Obstet Gynecol 2004;104:126–33.
8. Fitzgerald MP, Koch D, Senka J. Visceral and cutaneous sensory testing in patients with painful bladder syndrome. Neurourol Urodyn 2005;24:627–32.
9. Tu FF, As-Sanie S, Steege JF. Musculoskeletal causes of chronic pelvic pain: a systematic review of diagnosis: part I. Obstet Gynecol Surv 2005;60:379–85.
10. Melzack R. The short-form McGill Pain Questionnaire. Pain 1987;30:191–7.
11. Baguley SD, Curnow JS, Morrison GD, Barron LF. Vaginal algometer: development and application of a device to monitor vaginal wall pressure pain threshold. Physiol Meas 2003;24:833–6.
12. Hetrick DC, Ciol MA, Rothman I, Turner JA, Frest M, Berger RE. Musculoskeletal dysfunction in men with chronic pelvic pain syndrome type III: a case–control study. J Urol 2003;170:828–31.
13. Strigo IA, Bushnell MC, Boivin M, Duncan GH. Psychophysical analysis of visceral and cutaneous pain in human subjects. Pain 2002;97:235–46.
14. Giesecke T, Gracely RH, Grant MA, Nachemson A, Petzke F, Williams DA, et al. Evidence of augmented central pain processing in idiopathic chronic low back pain. Arthritis Rheum 2004;50:613–23.
15. Yamashita T, Kanaya K, Sekine M, Takebayashi T, Kawaguchi S, Katahira G. A quantitative analysis of sensory function in lumbar radiculopathy using current perception threshold testing. Spine 2002;27:1567–70.