Pelvic Floor Electromyography in Men with Chronic Pelvic Pain Syndrome

Jan 24, 2006

The objectives of this study were 1. To identify differences between men with Chronic Pelvic Pain Syndrome (CPPS), compared with pain-free men, in surface electromyography (sEMG)/Biofeedback (BFB) readings of pelvic floor muscles and 2.


The objectives of this study were 1. To identify differences between men with Chronic Pelvic Pain Syndrome (CPPS), compared with pain-free men, in surface electromyography (sEMG)/Biofeedback (BFB) readings of pelvic floor muscles and 2. To determine which pelvic floor muscle sEMG readings may have differential diagnostic and treatment selection value by accurately predicting group membership, CPPS v Normal

Twenty-one men with CPPS and 21 healthy men without pelvic pain underwent a standardized sEMG examination by a licensed physical therapist.  

On group difference measures men with CPPS showed significantly greater sEMG instability in preliminary resting baseline.  Three sEMG measures reliably categorized CPPS vs. Normals with CPPS showing greater preliminary resting baseline hypertonicity and instability with lowered voluntary endurance contraction amplitude.

CPPS patients manifest pelvic floor muscle instability compared to normals.  Prebaseline resting hypertonicity and instability along with endurance contraction weakness reliably predicts subject membership in the CPPS vs. normal group. Pelvic floor muscle sEMG may be a valuable screening tool to identify patients with CPPS who may benefit from therapies aimed at correcting pelvic floor muscle dysfunction.

Therapies aimed at improving pelvic floor muscle functioning, such as biofeedback, have been associated with improvement in chronic prostatitis/chronic pelvic pain syndrome (CPPS) symptoms1. However, there have been no studies comparing pelvic floor muscle by surface electromyography (SEMG) in men with and without CPPS to determine if pretreatment pelvic floor muscle status can differentiate these groups and therefore permit better treatment selection criteria.

MethodsStudy participants
Twenty-one men with CPPS and 21 controls were evaluated from November 2000 until August 2001 as part of a comprehensive study assessing men with CPPS. The study was approved by the institutional review board and all subjects signed University of Washington Human Subject Committee approved consents. Other findings from this study were published previously. Men aged 18-65 with CPPS were identified at the prostatitis clinic at The University of Washington and approached for participation by the attending urologist. All patients that were approached agreed to participate. Inclusion criteria for patients were pelvic pain at least 3 months in duration with no identified specific pathological condition to explain symptoms. Controls were healthy volunteers without pelvic pain or history of any other urologic disease, as confirmed by a National Institutes of Health Chronic Prostatitis Symptom Index score of zero Exclusion criteria for patients and controls included active urinary tract infection or infection localized to the prostate on a 4-glass urine sample, positive cultures for Chlamydia trachomatis or Neisseria gonorrhoeae, genitourinary malignancy, suicidal ideation or psychosis, postoperative pain, pain from another source in the genitourinary tract (e.g. renal calculi) and a history of radiation therapy or genitourinary tuberculosis. Controls were primarily University of Washington students and employees who were recruited by advertisements and were paid $50 for the visit. Pain patients were not paid but received free evaluation and were offered a discussion of the results at the end of the study. Only patients meeting inclusion and exclusion criteria were approached.

Procedures and Measures
Subjects provided demographic information via questionnaires. All subjects were evaluated by a physical therapist (DCH) who was experienced treating patients with pelvic floor pain and dysfunction, and trained in sEMG/BFB. The evaluator was blinded to the subject’s pain status. This was accomplished by having the subject checked into the room by another clinician, who advised them not to disclose their status to the evaluator. Although pain patient’s behavior can sometimes reveal their status, sEMG readings could not be altered in any way by the evaluator and the verbal cueing was scripted to be the same for all subjects. Surface electromyography/biofeedback (sEMG/BFB) evaluation was done as part of the overall physical therapy evaluation, described previously. Subjects were positioned semi-reclined in the supine position, with pillows flexing the knees and with slight external rotation of the thighs. Because the purpose of the evaluation was to gather data regarding pelvic floor functioning, and not to serve as an educational tool for the subject, the subject was not able to see the computer screen display. The subjects were verbally cued using a standardized format during the five and a half minute protocol.

The Glazer protocol is defined by a sequence of voluntary pelvic floor muscle actions and statistically defined measurements of the pelvic floor muscle sEMG during these actions. The sequence of actions are derived from pelvic floor muscle functions (sexual, sphinteric, and support)5 and include: 1. Prebaseline rest, 2. Rapid flicking, phasic, contractions, 3. Ten second holding, tonic, contractions, 4. A sixty second holding, endurance, contraction, and 5. Postbaseline rest6. The statistical measurements of the pelvic floor sEMG during these actions assess basic myoelectric characteristics including sEMG: 1. Amplitude, 2. Stability, 3. Muscle activation and recovery times, 7. Pelvic floor muscle sEMG reliability and validity have been demonstrated in previous studies.8Statistical Analysis:Table 1 shows statistical analyses of pelvic floor muscle sEMG readings including independent sample between group mean differences and logistic regression analysis. Independent sample differences assess which sEMG readings show overall between group mean differences, CPPS v normals. Beyond the existence of any overall group mean differences we wanted to know which, if any, pelvic floor muscle sEMG readings could accurately assign each study subject to their group membership, CPPS v normals. To determine which sEMG readings reliably predict group membership of each individual subject in this study, CPPS v normal, we conducted a logistic regression analysis as shown in Tables 2 and 3.

The statistic used to test significance of group mean differences was determined by the distribution of the data. Independent t-tests were used where data is normally distributed in both groups (AVPEAK1, WR5WKM, WR5WKCV, RESTM, and END60SM). Data was normalized by a natural log (LN) transformation where this transformation normalized data from both groups, and then the data was subject to independent t-tests (PREBLM, PREBLCV, RESTCV, POSTBLM, and POSTBLCV). For the remaining data group mean differences were tested with nonparametric statistics (Mann-Whitney test). These measures include AVONSET1, AVONSET2, AVREL2, AB_THR, END60SCV, and DECL. In order to examine the ability of the measures to predict group membership patient vs. control, we used logistic regression analysis employing the backward stepwise method to eliminate non-significant predictors.

ResultsTable 1 shows statistical results of between group differences and logistic regression analyses for each sEMG variable. The mean coefficient of variation (PREBLCV) of the 60 second initial resting baseline sEMG reading differed significantly between the CPPS group and normals (p values =.021).

The mean amplitude of the sEMG signal represents the total amount of electrical activity detected in the muscle which, at rest, reflects the tone of the muscle and during contraction reflects the contractile force of the muscle. The coefficient of variation is a statistic derived from the sEMG signal standard deviation divided by the mean amplitude of the signal for the period specified and reflects the degree of stability in the sEMG signal10. Therefore, these results indicate statistically significant between group differences in initial resting stability with CPPS patients manifesting greater instability. (Figure 1)

Significant mean group differences between variables do not tell us which variables can reliably assign an individual to one status or another, that is CPPS v normals. In addition to finding mean group pelvic floor muscle sEMG differences it is important to determine if any sEMG readings can reliably predict group membership, CPPS v normals. The analysis conducted is a logistic regression equation employing the backward stepwise method to eliminate non-significant predictors.

Table 2 shows the sEMG variables found to be significant in the logistic regression equation and Table 3 shows the resulting classification. The logistic regression analysis identified three statistically significant sEMG variables which can predict group membership, CPP v Normal. These variables include 1.) 60 second preliminary baseline sEMG mean amplitude, PREBLM (Wald=5.76, p=.016), 2.) 60 second preliminary baseline coefficient of variation, PREBLCV, (Wald=3.83, p=.050), and 3.) 60 second endurance contraction amplitude mean (END60SM, Wald=5.10, p=.024).

As shown in Table 3, the logistic regression model accurately predicts group membership in 18 of the 21, (85.71%) controls and 14 of the 21, (66.67%) CPPS patients for a combined correct group membership assignment of 76.19% of study participants. Demographic information was analyzed to see if it impacted our results from logistics regression. Marital status, education and employment were not statistically different. Race and sexual orientation showed some statistically significant difference. However, since we had too many missing values for sexual orientation (7 controls and 3 CPPS subjects) the difference in sexual orientation is doubtful.

Overall group differences indicate CPPS sufferers show greater pelvic floor muscle sEMG resting instability than their asymptomatic counterparts. Our findings also indicate that the probability of CPPS group membership are significantly higher for men with higher resting preliminary baseline mean amplitude, PREBLM, higher resting preliminary baseline variability, PREBCV, and lower amplitude 60 second endurance contractions, END60SM. This reflects greater resting hypertonicity and instability in the pubococcygeus muscle at rest and a reduced endurance contractile capacity in this muscle.

This is the first study to describe differences in pelvic floor sEMG between men with CPPS and controls. If we accept the premise that this idiopathic chronic pain disorder parallels similar lower urogenital tract idiopathic painful conditions in women, e.g. vulvodynia, 11-14 we would expect that normalization of the sEMG in confirmed pelvic floor dysfunction would lead to symptomatic improvement Further research is needed to test this hypothesis.

The objective nature of the sEMG recordings and standardized protocol allows us to consider the data itself as fairly reliable. These sEMG findings, in combination with our musculoskeletal physical exam findings of the earlier published report on this comprehensive study add even more evidence to the argument that men with CPPS have differences in pelvic muscle function compared to men without CPPS that could be important in the evaluation and treatment of these men.

In summary, this study shows that CPPS sufferers as compared to Normals manifest preliminary resting baseline instability. Initial resting baseline hypertonicity, instability and decreased endurance contractile capacity statistically significantly categorize men as more likely to suffer from CPPS. These findings indicate that pelvic floor muscle status likely plays a role in at least some subset of CPPS sufferers and that pelvic floor muscle sEMG evaluation can help identify this population as those who may benefit from pelvic floor muscle rehabilitation.

Table 1. Mean sEMG Readings of CPPS and Control groups




Control Mean



t-Test sig

Logistic Regression sig





P= .016





P= .050







































































PREBLM  Prebaseline mean 60-second  rest amplitude   PREBLCV Prebaseline stability  AVPEAK1 Average flick peak

AVONSET1 Average flick  activation time WR5WKM 10-second contract mean amplitude WR5WKCV 10-second  contract mean stability  RESTM 10-second rest mean amplitude RESTCV 10-second rest stability AVONSET22 10-second contract mean onset time AVREL2 10-second contract mean release time AB_THR Percentage of 60 sec endurance contraction spent above 80% of mean

10- second contraction amplitude END60SM 60-second endurance contraction mean amplitude END60SCV 60-second endurance contraction stability DECL Amplitude decline during 60-second endurance contraction POSTBLM  Postbaseline mean 60-second rest amplitude POSTBLCV  Postbaseline stability

 Table 2.  sEMG variables in the logistic regression equation
































Table 3. Classification of study participants by sEMG characteristics in the logistic regression equation.

Classification Table

Observed status

Predicted Status

Correct (%)











Overall Percentage


*The cut value is .500


Figure 1
Intra anal surface electromyographic tracings of normal male (right)  showing low amplitude (Amplitude=4.3uv) stable (Coefficient of Variance=.13) prebaseline mean resting tone and Chronic Pelvic Pain male (left) showing elevated mean prebaseline resting amplitude (Amplitude=9.8uv) with instability (Coefficient of Variance=.34).



Supported by the Paul G. Allen Foundation for Medical Research to Richard E. Berger MD


1. Clemens, J. Q., Nadler, R. B., Schaeffer, A. J., Belani, J., Albaugh, J., Bushman, W. 2000. Biofeedback, pelvic floor re-education, and bladder training for male chronic pelvic pain syndrome. Urology, 56: 951, 2000

2. Hetrick, D. C., Ciol, M. A., Rothman, I., Turner, J.A., Frest, M., Berger, R.E. 2000. Musculoskeletal dysfunction in men with chronic pelvic pain syndrome type III: a case-control study. J Urol, 170: 828, 2003

3. Litwin, M. S., McNaughton-Collins, M., Fowler, F. J., Jr., Nickel, J.c., calhoun, E.A., Pontari, M.A., Alexander, r.B., Farar, J.T., Oleary, M.P. 1999. The National Institutes of Health chronic prostatitis symptom index: development and validation of a new outcome measure. Chronic Prostatitis Collaborative Research Network. J Urol, 162: 369, 1999

4. Glazer, H. I., Rodke, G., Swencionis, C., Hertz, R., Young, A.W. 1995. Treatment of vulvar vestibulitis syndrome with electromyographic biofeedback of pelvic floor musculature. J Reprod Med, 40: 283, 1995

5. Kegel, A.H. 1950. Active exercise of the pubococcygeus muscle: Physiologic basis and therapeutic implications. Progress in Gynecology, 778-792

6. Glazer, H.I., Marinoff, S.M. and Sleight, I, J. 2002. The Web-enabled Glazer Surface Electromyographic Protocol for the Remote, Real-time Assessment and Rehabilitation of Pelvic Floor Dysfunction in Vulvovaginal Pain Disorders. J. Reprod 2002

7. Basmajian, John V. and De Luca, C., 1985. Muscles ALive: Their Functions Revealed by Electromyography, 5th Ed, Williams and Wilkins, Baltimore, Md. 1985.

8. Romanzi, L., Polaneczky, M., Glazer, H.I., 1999. A Simple Test of Pelvic Muscle During Pelvic Examination: Correlation to Surface Electromyography, Neurourology and Urodynamics, 1999; 18:603-612

9. Glazer, H.I., Romanzi, L., and Polaneczky, M., 1999. Pelvic Floor Muscle Surface Electromyography; Reliability and Clinical Predictive Validity, J. Reprod. Med., 1999; 44:779-782

10. Christodoulou, C., Bugmann, G. 1997. Coefficient of variation (CV) vs Mean interspike interval (ISI) curves: what do they tell us about the brain? Neural Computation, 9(1997) 985-1000

11. McKay, E., Kaufman, R. H., Doctor, U., Berkova, Z., Glazer, H., Redko, V. 2001. Treating vulvar vestibulitis with electromyographic biofeedback of pelvic floor musculature. J Reprod Med, 46: 337, 2001

12. Bergeron, S., Binik, Y. M., Khalife, S., Pagidas, K., Glazer, H.I. 2001. Vulvar vestibulitis syndrome: reliability of diagnosis and evaluation of current diagnostic criteria. Obstet Gynecol, 98: 45, 2001

13. White, G., Jantos, M., Glazer, H. 1997. Establishing the diagnosis of vulvar vestibulitis. J Reprod Med, 42: 157, 1997

14. Glazer, H. I., Jantos, M., Hartmann, E. H., Swencionis, c. 1998. Electromyographic comparisons of the pelvic floor in women with dysesthetic vulvodynia and asymptomatic women. J Reprod Med, 43: 959, 1998