- Email: [email protected]
Neuropathic injury to the levator ani occurs in 1 in 4 primiparous women
American Journal of Obstetrics and Gynecology (2006) 195, 1851–6
www.ajog.org
Neuropathic injury to the levator ani occurs in 1 in 4 primiparous women A. C. Weidner, MD,a M. G. Jamison, PhD,b V. Branham, MD,a M. M. South, MD,a K. M. Borawski, MD,c A. A. Romero, MDa Division of Urogynecology,a Department of Obstetrics and Gynecology, Division of Clinical and Epidemiological Research,b Department of Surgery/Urology,c Duke University Medical Center, Durham, NC Received for publication January 16, 2006; revised May 11, 2006; accepted June 17, 2006
KEY WORDS Delivery Pelvic floor Neuropathy Neuromuscular function
Objective: We measured levator ani neuromuscular function before and after first delivery to identify the location, timing, and mechanism of injury. Study design: Fifty-eight primiparous women underwent electromyographic examination of the levator ani antepartum at 6 weeks and 6 months after the delivery. Antepartum turns/amplitude data were pooled to create a normal range. We calculated each woman’s percentage of outliers from this range and assessed relationships between delivery and extent of injury. Results: At 6 weeks, 14 of 58 women (24.1%) had neuropathy, with 9 of those 14 women recovering by 6 months. At 6 months, 17 of 58 women (29.3%) were neuropathic, which included 12 new injuries. Women who had elective cesarean delivery had virtually no injury, but all other modes of delivery had similar injury rates. Conclusion: Obstetric delivery is associated frequently with electromyographic evidence of neuropathic injury to the levator ani. The entire levator complex is at risk, and cesarean delivery while in labor is not protective. Ó 2006 Mosby, Inc. All rights reserved.
Increasing awareness of the negative impact of pelvic floor dysfunction on the lives of adult women1 and controversy regarding the relative role of vaginal parity in these conditions2,3 have prompted multiple studies that have investigated the mechanisms of maternal pelvic floor injury in both human4 and animal models.5 Other studies have compared groups of nulliparous women to women of mixed parity or used more indirect techniques such as pudendal nerve conduction studies6 or surface Supported by National Institutes of Health grant HD38661-05. Presented at the Thirty-Second Annual Meeting of the Society of Gynecologic Surgeons, April 3-5, 2006, Tucson, AZ. Reprints not available from the authors. 0002-9378/$ - see front matter Ó 2006 Mosby, Inc. All rights reserved. doi:10.1016/j.ajog.2006.06.062
electromyography.7 Lack of precision in the method of measurement and wide variability in baseline pelvic floor muscle function makes the comparison of cases difficult.8 We used a more sensitive diagnostic technique (concentric needle electromyography) to monitor localized changes in female pelvic floor neuromuscular function after a single obstetric delivery.
Material and methods Primiparous women were recruited after full Institutional Review Board approval was obtained. Exclusion criteria were a history of pelvic surgery, pre-existing
1852 pelvic floor symptoms, diabetes mellitus or neuromuscular disorder, and planned cesarean delivery. At 28 to 34 weeks of gestation, the women underwent baseline electromyography assessment of the levator ani muscles.9 A pelvic organ prolapse quantitation10 and a clinical assessment of voluntary pelvic floor contraction were performed.11 A concentric electromyography needle electrode was placed under the vaginal epithelium at 4 sites, 2 on each side, to sample the puborectalis and iliococcygeus muscles.9 The puborectalis muscle was identified 2 cm medially and 2 cm caudad from the ischial spine (left lateral and right lateral). The iliococcygeus muscle was located medial and closer to the vaginal apex (left medial and right medial). The needle was inserted at a depth that yielded crisp motor unit action potentials with maximal amplitude with the muscle at rest. This electromyography signal was filtered (5 Hz-10 kHz) and digitally recorded at each site while the women remained at rest, with partial contraction, and with maximal effort. Labor and delivery were managed by the principles of active management of labor.12 Anesthesia was chosen according to the woman’s preference. Epidural anesthesia was administered per protocol with a ropivacaine/ fentanyl-loading dose and patient-controlled pump for maintenance. We recorded time of labor benchmarks, use of oxytocin augmentation, time of epidural administration, epidural dosing, minutes spent pushing, neonatal weight and head circumference, and mode of delivery. We defined 4 modes of delivery: spontaneous vaginal delivery (SVD), operative vaginal delivery (OVD) that involved vaginal delivery after placement of forceps and/or vacuum, cesarean delivery performed after the onset of labor, and elective cesarean delivery before the onset of labor. After delivery, the women returned for repeat assessments.
Electromyography analysis Analysis of digital electromyography files was performed on a Synergy 2 channel electromyograph (Oxford Instruments Medical Systems, Hawthorne, NY) with turns/ amplitude analysis.13 We used the Synergy 8.2 software that measures the amplitude in microvolts of the signal between successive changes in direction of the electrical signal (amplitude) and the number of turns per second in the signal (turns). Nonoverlapping 500 msec epochs at each site were sampled 6 times each at resting, moderate contraction, and maximum contraction. A previous report described how the normal turns versus amplitude plot on the XY axis is used to generate a normal scatter plot (cloud).9 For this study, a normal cloud for each of the 4 muscle sites at the antepartum visit was generated with the use of a regression model. These clouds represented pooled data from all of the women. Individual muscle sites with R10% of postpartum data points
Weidner et al outside the margins of the normal cloud for the corresponding site were considered injured. We divided the women into 4 groups according to observed patterns of injury and recovery: Group A women had injury at both 6 weeks and 6 months; group B women had injury at 6 weeks but had recovered at 6 months; group C women were normal at 6 weeks, but injury was evident at 6 months, and group D women had no injury at either visit. Tests for incidence and location of injury on electromyography were determined with the 1-way c2 test, and differences among the groups of women were determined with 2-way c2 tests and analysis of variance. Logistic regression models that predicted injury from obstetric variables were estimated. Analysis was conducted with SPSS software (version 12.0; SPSS Inc, Chicago, IL).
Results Ninety-six women were recruited from 2001 to 2004. Nineteen women withdrew after the first visit, citing pain from the electromyography. One woman had an emergent preterm cesarean delivery. Seven women were examined at the 6-week postpartum visit, but not at the 6-month visit. Because the objective of this analysis was to assess injury and recovery over the full study period, those 7 women were excluded, which left 69 women. Quality control criteria that were used in repeated measures analysis forced exclusion of 11 more women because of missing or otherwise incomplete data at muscle sites, which left 58 women. Forty of the women were white; 11 of the women were black; 2 of the women were Hispanic; 4 of the women were Asian, and 1 woman was Indian. Women were 29.6 G 5.6 years of age at delivery with a body mass index of 25.7 G 5.8 kg/m2 at the 6-month postpartum visit. Assessments of baseline muscle function are indicated in the Figure. Muscle performance was not significantly different between sites before delivery. The subject represented by red dots had R10% of abnormal data points for the left lateral and right lateral muscle sites, but normal electromyography (EMG) for the left and right medial sites. This subject was also representative of the pattern that we noted in all cases of muscle injury, namely, deflection of data points above the normal cloud to a region higher on the Y axis, indicating a neuropathic injury mechanism, rather than a myopathic injury mechanism.14 The mode of obstetric delivery in these 58 primigravid women is given in Table I. In the cesarean in labor group, 7 women had epidural analgesia placed, but only 2 women progressed enough in labor to commence pushing. Three women had elective cesarean delivery before the onset of labor because of breech presentation or fetal macrosomia. We pooled levator electromyography data from all sites and determined the number of women
Weidner et al
1853
Figure Normal cloud pattern (turns/second vs log [amplitude]) for four muscle sites in 58 subjects (black dots). Superimposed red dots show the distribution of EMG data for one 6 week postpartum subject with neuropathic injury at the left and right lateral sites. Amplitude is shown in microvolts.
with abnormal electromyography at 6 weeks and 6 months after delivery. Any site with R10% of points outside normal range defined the woman’s condition to be abnormal. The frequency of injury by electromyography criteria across all women was 24.1% at 6 weeks and 29.3% at 6 months, but the elective cesarean group notably had no injury at 6 weeks, and only 1 woman had 10.3% abnormal data points at 1 site at 6 months. Table II compares women by injury group. Most women did not have electromyography evidence of muscle injury (group D) at either postpartum visit. However, most of the women who met the criteria for injury did so at only 1 of the 2 postpartum visits (groups B and C). A smaller fraction of women showed injury at both visits (group A). No significant differences existed between groups A and D with regard to any obstetric, neonatal, or maternal physical characteristics that were tested. However, when groups A through C were compared with group D, those women with no injury had a
Table I Percentage of patients with levator ani injury at 6 weeks and 6 months after delivery by mode of delivery Mode of delivery
N
6 Weeks (%)
6 Months (%)
SVD OVD Cesarean delivery while in labor Elective cesarean without labor Overall
36 8 11
22.2 25.0 36.4
30.6 25.0 27.3
3
0
33.3
58
24.1*
29.3y
* P = .50. y P = .77.
significantly longer duration of epidural analgesia before delivery. Although not statistically significant, group A had the shortest mean duration of the third stage of labor and the shortest duration of epidural analgesia.
1854 Table II
Weidner et al Relationship of obstetric, neonatal, and maternal characteristics to levator injury and recovery Group*
Variable Labor Oxytocin administration (%) Mean duration epidural analgesia (min) Mean duration second-stage labor (min) Neonate characteristics Neonatal head circumference (cm) Mean neonatal weight (g) Mode of delivery SVD (%) OVD (%) Cesarean delivery while in labor (%) Elective cesarean delivery (%) Maternal characteristics Postpartum mean pelvic floor contraction strength (0-9) 6-Month postpartum pelvic organ prolapse quantitation stage R2 (%) 6-Month postpartum maternal BMI O30 kg/m2 (%)
P valuey
P valuez
N
A (n = 5)
B (n = 9)
C (n = 12)
D (n = 32)
58 47 58
60.0 310 35
66.7 334 48
36.4 326 67
65.6 484 69
.26 .16 .52
.23 .02 .30
58 58
34.8 3363
34.3 3208
34.5 3473
34.5 3327
.95 .80
.94 .83
36 8 11 3
8.3 0 18.2 0
13.9 25.0 18.3 0
22.2 25.0 9.1 33.3
55.6 50.0 54.5 66.7
.89
.97
58
7.4
6.2
7.5
6.5
0.45
0.36
58
60.0
33.3
8.3
31.3
.39
.72
58
40.0
22.2
8.3
21.9
.51
.36
* Group A, Injury at 6 weeks and 6 months; Group B, injury at 6 weeks and normal at 6 months; Group C, normal at 6 weeks and injury at 6 months; Group D, no injury. y Compares all groups. z Compares groups A, B, and C versus group D.
To investigate the role of oxytocin administration, epidural administration, and the duration of the third stage, we constructed a logistic regression model using injury as the dependent variable (Table III). Significant independent predictors of levator injury were shorter duration of epidural analgesia and OVD. One interesting capacity of this study design was to detect the location of levator ani muscle injury as it traverses the bony pelvis (Table IV). Injury occurred at all sites of the levator ani that were tested. Women who have SVD appear to have a greater frequency of injury to the lateral levator sites (ie, puborectalis muscle). Cesarean delivery while in labor did not protect against injury at multiple levator sites. The only site that met the criteria for injury in 1 of the 3 women who had chosen elective cesarean delivery was the right medial levator.
Comment This study was designed to apply the highest standards of diagnostic precision for neuromuscular function8 and to establish definitions of normal and abnormal electromyography14 to pelvic floor muscles in a longitudinal fashion that enabled each woman to serve as her own control. This was intended to minimize bias that could have been introduced by the assumption that every woman has the same baseline pelvic floor function.
Our objective was to identify the mechanism of maternal levator ani injury during delivery, to locate the most frequent sites of injury, and to monitor the recovery from acute injury during the postpartum period. We found that approximately 1 in 4 women have a detectable neuropathic injury. However, our findings indicated that most women with neuropathy at 6 weeks had normal muscle at 6 months. Furthermore, there were some women with neuropathy at 6 months who did not meet the criteria for injury at 6 weeks. This temporal evolution of injury and recovery bears further study. Acute neuropathic injury causes loss of motor units, and those action potentials that are then generated are fewer and generally of low amplitude. Data on our XY turns/amplitude plot from a muscle with this type of injury therefore would be clustered close to the origin of the graph and might not have sufficient outliers from the normal cloud to be considered abnormal. Recovery from focal injury then results in permanent motor unit loss, atrophy, low or absent amplitude; if reinnervation occurs, those larger, more complex motor units generate action potentials of abnormally high amplitude. We postulate that this might explain our observation of several women who showed evidence of injury only at 6 months. At 6 weeks, these particular women had a wider variability of muscle function, which suggests variable degrees of focal injury but, when turns and amplitude were analyzed together, did not meet our injury threshold. At 6 months, the signal in these women had
Weidner et al Table III
1855
Logistic regression model that predicted injury in women who received oxytocin
Independent model variables Duration of epidural analgesia (min) Baby weight (g)z Mode of delivery (%) Cesarean delivery while in labor OVD SVD (reference group)
z
Any injury (groups A, B, and C)*
No injury (group D)y
D
P value
264.1 G 73.2 3446 G 195
498.8 G 49.5 3274 G 128
C234.7 172
.04 .17
37.5 57.1 31.6
62.5 42.9 68.4
C5.9x C25.5 d
.25 .06 d
* N = 13. y N = 20. z Data given as mean G SE. x Two linear contrasts SVD–cesarean delivery while in labor (C5.9%) and SVD–OVD (C25.5%).
Table IV
Percentage of women with levator ani injury at 6 weeks after delivery by muscle site and mode of delivery
Variable
N
Right lateral (%)
Right medial (%)
Left medial (%)
Left lateral (%)
P value
SVD OVD Cesarean delivery while in labor Elective cesarean delivery
31 8 11
35.5 2.5 18.2
19.4 25.0 27.3
22.6 12.5 9.1
32.3 12.5 36.4
.62 .80 .38
3
0
33.3
0
0
d
low turns and high amplitude and was abnormal with the use of our definition of R10% outliers. If this explanation is true, then the actual prevalence of injury at the 6-weeks postpartum period actually would be higher than our data suggest. Because this was our first study in an obstetric population, we could not predict frequency of injury according to any obstetric variable. We are unable therefore to comment about differences that were observed with regard to mode of delivery or neonatal weight, for example. However, the striking absence of injury that was seen in women who had elective cesarean deliveries stands in contrast to the frequency of injury that was observed for all other modes of delivery. We would have anticipated that women who had a cesarean delivery while in labor would have preserved levator function at lateral sites, because simulations of labor have demonstrated that the puborectalis undergoes the most stretch during actual distension by the fetal head, late in the third stage.4 However, cesarean delivery while in labor is not protective to the pelvic floor and shows injury rates similar to SVD or OVD. We also would not have predicted our finding that women who had levator injury had a significantly shorter duration of epidural analgesia (an indirect correlate measure of the duration of the second and third stages of labor), which is the opposite of that reported by retrospective human15 and prospective animal studies16 (Table III). This bears further study.
Other mechanisms of muscle injury (such as actual disruption of levator insertions into the endopelvic fascia) have been demonstrated with SVD.17 Acutely and at 6 weeks, muscular avulsion would result classically in electromyography of diminished maximal amplitude, with spontaneous activity and fibrillations. Chronically, a muscle that has been separated from its insertion loses tone and motor units atrophy; therefore, the 6-month electromyography findings in such an injury would be similar those of myopathy. Because our abnormal observations at 6 months were uniformly of increased amplitude, we postulate a neuropathic mechanism of injury and recovery for the levator, rather than muscle avulsion. The muscle avulsion that was imaged in previous studies may represent progressive neurogenic atrophy that results in muscle thinning and loss. We intend to test this hypothesis in future studies. Similarly, this study does not allow us to completely rule out that there could be multiple coexisting mechanisms of injury such as muscle fiber injury, proximal nerve injury, or demyelination; however, the uniformly neurogenic nature of those women whose test results were abnormal supports our contention that the injury is primarily axonal neuropathy with some recovery. We do not know the relationship of an injury to the pelvic floor to the later development of pelvic floor dysfunction. How cumulative is the neuropathic change with subsequent deliveries? How closely does the
1856 re-innervated muscle mimic normal muscle? How valid is electromyography injury compared with measures of pelvic floor symptoms and quality of life? We intend to address these and other questions in future reports. This study was limited by its relatively small sample. Our longitudinal design required a repeated measures statistical model, which forced us to exclude 11 women. Because 1 limitation of electromyography is that no data are available when muscle is absent, it is possible that these women actually had more severe neuropathic injury to the point of focal muscle loss. Their exclusion could have led us to report a lower injury rate than actually occurred. A study in progress that correlates these neuromuscular findings to morphologic changes on pelvic magnetic resonance imaging in the same women is intended to address that possibility.
Acknowledgments We thank Elizabeth C. Coats, RN, and Jean H. Maynor, RN, for their coordination of this study and Donald B. Sanders, MD, for his generous mentorship and guidance.
References 1. MacLennan AH, Taylor AW, Wilson DH, Wilson D. The prevalence of pelvic floor disorders and their relationship to gender, age, parity and mode of delivery. BJOG 2000;107:1460-70. 2. Buchsbaum GM, Chin M, Glantz C, Guzick D. Prevalence of urinary incontinence and associated risk factors in a cohort of nuns. Obstet Gynecol 2002;100:226-9. 3. Hannah ME, Hannah WJ, Hodnett ED, Chalmers B, Kung R, Willan A, et al. Outcomes at 3 months after planned cesarean vs planned vaginal delivery for breech presentation at term: the international randomized Term Breech Trial. JAMA 2002;287:1822-31. 4. Lien KC, Mooney B, DeLancey JO, Ashton-Miller JA. Levator ani muscle stretch induced by simulated vaginal birth. Obstet Gynecol 2004;103:31-40.
Weidner et al 5. Damaser MS, Broxton-King C, Ferguson C, Kim FJ, Kerns JM. Functional and neuroanatomical effects of vaginal distention and pudendal nerve crush in the female rat. J Urol 2003;170: 1027-31. 6. Kiff ES, Swash M. Slowed conduction in the pudendal nerves in idiopathic (neurogenic) faecal incontinence. Br J Surg 1984;71: 614-6. 7. Deindl FM, Vodusek DB, Hesse U, Schussler B. Pelvic floor activity patterns: comparison of nulliparous continent and parous urinary stress incontinent women: a kinesiological EMG study. Br J Urol 1994;73:413-7. 8. Gregory WT, Lou JS, Stuyvesant A, Clark AL. Quantitative electromyography of the anal sphincter after uncomplicated vaginal delivery. Obstet Gynecol 2004;104:327-35. 9. Weidner AC, Sanders DB, Nandedkar SD, Bump RC. Quantitative electromyographic analysis of levator ani and external anal sphincter muscles of nulliparous women. Am J Obstet Gynecol 2000;183:1249-56. 10. Bump RC, Mattiasson A, Bo K, Brubaker LP, DeLancey JO, Klarskov P, et al. The standardization of terminology of female pelvic organ prolapse and pelvic floor dysfunction. Am J Obstet Gynecol 1996;175:10-7. 11. Brink CA, Sampselle CM, Wells TJ, Diokno AC, Gillis GL. A digital test for pelvic muscle strength in older women with urinary incontinence. Nurs Res 1989;38:196-9. 12. Hansen SL, Clark SL, Foster JC. Active pushing versus passive fetal descent in the second stage of labor: a randomized controlled trial. Obstet Gynecol 2002;99:29-34. 13. Nandedkar SD, Sanders DB, Stalberg EV. Automatic analysis of the electromyographic interference pattern: part I, development of quantitative features. Muscle Nerve 1986;9:431-9. 14. Nandedkar SD, Sanders DB, Stalberg EV. Automatic analysis of the electromyographic interference pattern: part II, findings in control subjects and in some neuromuscular diseases. Muscle Nerve 1986;9:491-500. 15. Fritel X, Fauconnier A, Levet C, Benifla JL. Stress urinary incontinence 4 years after the first delivery: a retrospective cohort survey. Acta Obstet Gynecol Scand 2004;83:941-5. 16. Cannon TW, Wojcik EM, Ferguson CL, Saraga S, Thomas C, Damaser MS. Effects of vaginal distension on urethral anatomy and function. BJU Int 2002;90:403-7. 17. DeLancey JO, Kearney R, Chou Q, Speights S, Binno S. The appearance of levator ani muscle abnormalities in magnetic resonance images after vaginal delivery. Obstet Gynecol 2003;101: 46-53.
www.ajog.org
Neuropathic injury to the levator ani occurs in 1 in 4 primiparous women A. C. Weidner, MD,a M. G. Jamison, PhD,b V. Branham, MD,a M. M. South, MD,a K. M. Borawski, MD,c A. A. Romero, MDa Division of Urogynecology,a Department of Obstetrics and Gynecology, Division of Clinical and Epidemiological Research,b Department of Surgery/Urology,c Duke University Medical Center, Durham, NC Received for publication January 16, 2006; revised May 11, 2006; accepted June 17, 2006
KEY WORDS Delivery Pelvic floor Neuropathy Neuromuscular function
Objective: We measured levator ani neuromuscular function before and after first delivery to identify the location, timing, and mechanism of injury. Study design: Fifty-eight primiparous women underwent electromyographic examination of the levator ani antepartum at 6 weeks and 6 months after the delivery. Antepartum turns/amplitude data were pooled to create a normal range. We calculated each woman’s percentage of outliers from this range and assessed relationships between delivery and extent of injury. Results: At 6 weeks, 14 of 58 women (24.1%) had neuropathy, with 9 of those 14 women recovering by 6 months. At 6 months, 17 of 58 women (29.3%) were neuropathic, which included 12 new injuries. Women who had elective cesarean delivery had virtually no injury, but all other modes of delivery had similar injury rates. Conclusion: Obstetric delivery is associated frequently with electromyographic evidence of neuropathic injury to the levator ani. The entire levator complex is at risk, and cesarean delivery while in labor is not protective. Ó 2006 Mosby, Inc. All rights reserved.
Increasing awareness of the negative impact of pelvic floor dysfunction on the lives of adult women1 and controversy regarding the relative role of vaginal parity in these conditions2,3 have prompted multiple studies that have investigated the mechanisms of maternal pelvic floor injury in both human4 and animal models.5 Other studies have compared groups of nulliparous women to women of mixed parity or used more indirect techniques such as pudendal nerve conduction studies6 or surface Supported by National Institutes of Health grant HD38661-05. Presented at the Thirty-Second Annual Meeting of the Society of Gynecologic Surgeons, April 3-5, 2006, Tucson, AZ. Reprints not available from the authors. 0002-9378/$ - see front matter Ó 2006 Mosby, Inc. All rights reserved. doi:10.1016/j.ajog.2006.06.062
electromyography.7 Lack of precision in the method of measurement and wide variability in baseline pelvic floor muscle function makes the comparison of cases difficult.8 We used a more sensitive diagnostic technique (concentric needle electromyography) to monitor localized changes in female pelvic floor neuromuscular function after a single obstetric delivery.
Material and methods Primiparous women were recruited after full Institutional Review Board approval was obtained. Exclusion criteria were a history of pelvic surgery, pre-existing
1852 pelvic floor symptoms, diabetes mellitus or neuromuscular disorder, and planned cesarean delivery. At 28 to 34 weeks of gestation, the women underwent baseline electromyography assessment of the levator ani muscles.9 A pelvic organ prolapse quantitation10 and a clinical assessment of voluntary pelvic floor contraction were performed.11 A concentric electromyography needle electrode was placed under the vaginal epithelium at 4 sites, 2 on each side, to sample the puborectalis and iliococcygeus muscles.9 The puborectalis muscle was identified 2 cm medially and 2 cm caudad from the ischial spine (left lateral and right lateral). The iliococcygeus muscle was located medial and closer to the vaginal apex (left medial and right medial). The needle was inserted at a depth that yielded crisp motor unit action potentials with maximal amplitude with the muscle at rest. This electromyography signal was filtered (5 Hz-10 kHz) and digitally recorded at each site while the women remained at rest, with partial contraction, and with maximal effort. Labor and delivery were managed by the principles of active management of labor.12 Anesthesia was chosen according to the woman’s preference. Epidural anesthesia was administered per protocol with a ropivacaine/ fentanyl-loading dose and patient-controlled pump for maintenance. We recorded time of labor benchmarks, use of oxytocin augmentation, time of epidural administration, epidural dosing, minutes spent pushing, neonatal weight and head circumference, and mode of delivery. We defined 4 modes of delivery: spontaneous vaginal delivery (SVD), operative vaginal delivery (OVD) that involved vaginal delivery after placement of forceps and/or vacuum, cesarean delivery performed after the onset of labor, and elective cesarean delivery before the onset of labor. After delivery, the women returned for repeat assessments.
Electromyography analysis Analysis of digital electromyography files was performed on a Synergy 2 channel electromyograph (Oxford Instruments Medical Systems, Hawthorne, NY) with turns/ amplitude analysis.13 We used the Synergy 8.2 software that measures the amplitude in microvolts of the signal between successive changes in direction of the electrical signal (amplitude) and the number of turns per second in the signal (turns). Nonoverlapping 500 msec epochs at each site were sampled 6 times each at resting, moderate contraction, and maximum contraction. A previous report described how the normal turns versus amplitude plot on the XY axis is used to generate a normal scatter plot (cloud).9 For this study, a normal cloud for each of the 4 muscle sites at the antepartum visit was generated with the use of a regression model. These clouds represented pooled data from all of the women. Individual muscle sites with R10% of postpartum data points
Weidner et al outside the margins of the normal cloud for the corresponding site were considered injured. We divided the women into 4 groups according to observed patterns of injury and recovery: Group A women had injury at both 6 weeks and 6 months; group B women had injury at 6 weeks but had recovered at 6 months; group C women were normal at 6 weeks, but injury was evident at 6 months, and group D women had no injury at either visit. Tests for incidence and location of injury on electromyography were determined with the 1-way c2 test, and differences among the groups of women were determined with 2-way c2 tests and analysis of variance. Logistic regression models that predicted injury from obstetric variables were estimated. Analysis was conducted with SPSS software (version 12.0; SPSS Inc, Chicago, IL).
Results Ninety-six women were recruited from 2001 to 2004. Nineteen women withdrew after the first visit, citing pain from the electromyography. One woman had an emergent preterm cesarean delivery. Seven women were examined at the 6-week postpartum visit, but not at the 6-month visit. Because the objective of this analysis was to assess injury and recovery over the full study period, those 7 women were excluded, which left 69 women. Quality control criteria that were used in repeated measures analysis forced exclusion of 11 more women because of missing or otherwise incomplete data at muscle sites, which left 58 women. Forty of the women were white; 11 of the women were black; 2 of the women were Hispanic; 4 of the women were Asian, and 1 woman was Indian. Women were 29.6 G 5.6 years of age at delivery with a body mass index of 25.7 G 5.8 kg/m2 at the 6-month postpartum visit. Assessments of baseline muscle function are indicated in the Figure. Muscle performance was not significantly different between sites before delivery. The subject represented by red dots had R10% of abnormal data points for the left lateral and right lateral muscle sites, but normal electromyography (EMG) for the left and right medial sites. This subject was also representative of the pattern that we noted in all cases of muscle injury, namely, deflection of data points above the normal cloud to a region higher on the Y axis, indicating a neuropathic injury mechanism, rather than a myopathic injury mechanism.14 The mode of obstetric delivery in these 58 primigravid women is given in Table I. In the cesarean in labor group, 7 women had epidural analgesia placed, but only 2 women progressed enough in labor to commence pushing. Three women had elective cesarean delivery before the onset of labor because of breech presentation or fetal macrosomia. We pooled levator electromyography data from all sites and determined the number of women
Weidner et al
1853
Figure Normal cloud pattern (turns/second vs log [amplitude]) for four muscle sites in 58 subjects (black dots). Superimposed red dots show the distribution of EMG data for one 6 week postpartum subject with neuropathic injury at the left and right lateral sites. Amplitude is shown in microvolts.
with abnormal electromyography at 6 weeks and 6 months after delivery. Any site with R10% of points outside normal range defined the woman’s condition to be abnormal. The frequency of injury by electromyography criteria across all women was 24.1% at 6 weeks and 29.3% at 6 months, but the elective cesarean group notably had no injury at 6 weeks, and only 1 woman had 10.3% abnormal data points at 1 site at 6 months. Table II compares women by injury group. Most women did not have electromyography evidence of muscle injury (group D) at either postpartum visit. However, most of the women who met the criteria for injury did so at only 1 of the 2 postpartum visits (groups B and C). A smaller fraction of women showed injury at both visits (group A). No significant differences existed between groups A and D with regard to any obstetric, neonatal, or maternal physical characteristics that were tested. However, when groups A through C were compared with group D, those women with no injury had a
Table I Percentage of patients with levator ani injury at 6 weeks and 6 months after delivery by mode of delivery Mode of delivery
N
6 Weeks (%)
6 Months (%)
SVD OVD Cesarean delivery while in labor Elective cesarean without labor Overall
36 8 11
22.2 25.0 36.4
30.6 25.0 27.3
3
0
33.3
58
24.1*
29.3y
* P = .50. y P = .77.
significantly longer duration of epidural analgesia before delivery. Although not statistically significant, group A had the shortest mean duration of the third stage of labor and the shortest duration of epidural analgesia.
1854 Table II
Weidner et al Relationship of obstetric, neonatal, and maternal characteristics to levator injury and recovery Group*
Variable Labor Oxytocin administration (%) Mean duration epidural analgesia (min) Mean duration second-stage labor (min) Neonate characteristics Neonatal head circumference (cm) Mean neonatal weight (g) Mode of delivery SVD (%) OVD (%) Cesarean delivery while in labor (%) Elective cesarean delivery (%) Maternal characteristics Postpartum mean pelvic floor contraction strength (0-9) 6-Month postpartum pelvic organ prolapse quantitation stage R2 (%) 6-Month postpartum maternal BMI O30 kg/m2 (%)
P valuey
P valuez
N
A (n = 5)
B (n = 9)
C (n = 12)
D (n = 32)
58 47 58
60.0 310 35
66.7 334 48
36.4 326 67
65.6 484 69
.26 .16 .52
.23 .02 .30
58 58
34.8 3363
34.3 3208
34.5 3473
34.5 3327
.95 .80
.94 .83
36 8 11 3
8.3 0 18.2 0
13.9 25.0 18.3 0
22.2 25.0 9.1 33.3
55.6 50.0 54.5 66.7
.89
.97
58
7.4
6.2
7.5
6.5
0.45
0.36
58
60.0
33.3
8.3
31.3
.39
.72
58
40.0
22.2
8.3
21.9
.51
.36
* Group A, Injury at 6 weeks and 6 months; Group B, injury at 6 weeks and normal at 6 months; Group C, normal at 6 weeks and injury at 6 months; Group D, no injury. y Compares all groups. z Compares groups A, B, and C versus group D.
To investigate the role of oxytocin administration, epidural administration, and the duration of the third stage, we constructed a logistic regression model using injury as the dependent variable (Table III). Significant independent predictors of levator injury were shorter duration of epidural analgesia and OVD. One interesting capacity of this study design was to detect the location of levator ani muscle injury as it traverses the bony pelvis (Table IV). Injury occurred at all sites of the levator ani that were tested. Women who have SVD appear to have a greater frequency of injury to the lateral levator sites (ie, puborectalis muscle). Cesarean delivery while in labor did not protect against injury at multiple levator sites. The only site that met the criteria for injury in 1 of the 3 women who had chosen elective cesarean delivery was the right medial levator.
Comment This study was designed to apply the highest standards of diagnostic precision for neuromuscular function8 and to establish definitions of normal and abnormal electromyography14 to pelvic floor muscles in a longitudinal fashion that enabled each woman to serve as her own control. This was intended to minimize bias that could have been introduced by the assumption that every woman has the same baseline pelvic floor function.
Our objective was to identify the mechanism of maternal levator ani injury during delivery, to locate the most frequent sites of injury, and to monitor the recovery from acute injury during the postpartum period. We found that approximately 1 in 4 women have a detectable neuropathic injury. However, our findings indicated that most women with neuropathy at 6 weeks had normal muscle at 6 months. Furthermore, there were some women with neuropathy at 6 months who did not meet the criteria for injury at 6 weeks. This temporal evolution of injury and recovery bears further study. Acute neuropathic injury causes loss of motor units, and those action potentials that are then generated are fewer and generally of low amplitude. Data on our XY turns/amplitude plot from a muscle with this type of injury therefore would be clustered close to the origin of the graph and might not have sufficient outliers from the normal cloud to be considered abnormal. Recovery from focal injury then results in permanent motor unit loss, atrophy, low or absent amplitude; if reinnervation occurs, those larger, more complex motor units generate action potentials of abnormally high amplitude. We postulate that this might explain our observation of several women who showed evidence of injury only at 6 months. At 6 weeks, these particular women had a wider variability of muscle function, which suggests variable degrees of focal injury but, when turns and amplitude were analyzed together, did not meet our injury threshold. At 6 months, the signal in these women had
Weidner et al Table III
1855
Logistic regression model that predicted injury in women who received oxytocin
Independent model variables Duration of epidural analgesia (min) Baby weight (g)z Mode of delivery (%) Cesarean delivery while in labor OVD SVD (reference group)
z
Any injury (groups A, B, and C)*
No injury (group D)y
D
P value
264.1 G 73.2 3446 G 195
498.8 G 49.5 3274 G 128
C234.7 172
.04 .17
37.5 57.1 31.6
62.5 42.9 68.4
C5.9x C25.5 d
.25 .06 d
* N = 13. y N = 20. z Data given as mean G SE. x Two linear contrasts SVD–cesarean delivery while in labor (C5.9%) and SVD–OVD (C25.5%).
Table IV
Percentage of women with levator ani injury at 6 weeks after delivery by muscle site and mode of delivery
Variable
N
Right lateral (%)
Right medial (%)
Left medial (%)
Left lateral (%)
P value
SVD OVD Cesarean delivery while in labor Elective cesarean delivery
31 8 11
35.5 2.5 18.2
19.4 25.0 27.3
22.6 12.5 9.1
32.3 12.5 36.4
.62 .80 .38
3
0
33.3
0
0
d
low turns and high amplitude and was abnormal with the use of our definition of R10% outliers. If this explanation is true, then the actual prevalence of injury at the 6-weeks postpartum period actually would be higher than our data suggest. Because this was our first study in an obstetric population, we could not predict frequency of injury according to any obstetric variable. We are unable therefore to comment about differences that were observed with regard to mode of delivery or neonatal weight, for example. However, the striking absence of injury that was seen in women who had elective cesarean deliveries stands in contrast to the frequency of injury that was observed for all other modes of delivery. We would have anticipated that women who had a cesarean delivery while in labor would have preserved levator function at lateral sites, because simulations of labor have demonstrated that the puborectalis undergoes the most stretch during actual distension by the fetal head, late in the third stage.4 However, cesarean delivery while in labor is not protective to the pelvic floor and shows injury rates similar to SVD or OVD. We also would not have predicted our finding that women who had levator injury had a significantly shorter duration of epidural analgesia (an indirect correlate measure of the duration of the second and third stages of labor), which is the opposite of that reported by retrospective human15 and prospective animal studies16 (Table III). This bears further study.
Other mechanisms of muscle injury (such as actual disruption of levator insertions into the endopelvic fascia) have been demonstrated with SVD.17 Acutely and at 6 weeks, muscular avulsion would result classically in electromyography of diminished maximal amplitude, with spontaneous activity and fibrillations. Chronically, a muscle that has been separated from its insertion loses tone and motor units atrophy; therefore, the 6-month electromyography findings in such an injury would be similar those of myopathy. Because our abnormal observations at 6 months were uniformly of increased amplitude, we postulate a neuropathic mechanism of injury and recovery for the levator, rather than muscle avulsion. The muscle avulsion that was imaged in previous studies may represent progressive neurogenic atrophy that results in muscle thinning and loss. We intend to test this hypothesis in future studies. Similarly, this study does not allow us to completely rule out that there could be multiple coexisting mechanisms of injury such as muscle fiber injury, proximal nerve injury, or demyelination; however, the uniformly neurogenic nature of those women whose test results were abnormal supports our contention that the injury is primarily axonal neuropathy with some recovery. We do not know the relationship of an injury to the pelvic floor to the later development of pelvic floor dysfunction. How cumulative is the neuropathic change with subsequent deliveries? How closely does the
1856 re-innervated muscle mimic normal muscle? How valid is electromyography injury compared with measures of pelvic floor symptoms and quality of life? We intend to address these and other questions in future reports. This study was limited by its relatively small sample. Our longitudinal design required a repeated measures statistical model, which forced us to exclude 11 women. Because 1 limitation of electromyography is that no data are available when muscle is absent, it is possible that these women actually had more severe neuropathic injury to the point of focal muscle loss. Their exclusion could have led us to report a lower injury rate than actually occurred. A study in progress that correlates these neuromuscular findings to morphologic changes on pelvic magnetic resonance imaging in the same women is intended to address that possibility.
Acknowledgments We thank Elizabeth C. Coats, RN, and Jean H. Maynor, RN, for their coordination of this study and Donald B. Sanders, MD, for his generous mentorship and guidance.
References 1. MacLennan AH, Taylor AW, Wilson DH, Wilson D. The prevalence of pelvic floor disorders and their relationship to gender, age, parity and mode of delivery. BJOG 2000;107:1460-70. 2. Buchsbaum GM, Chin M, Glantz C, Guzick D. Prevalence of urinary incontinence and associated risk factors in a cohort of nuns. Obstet Gynecol 2002;100:226-9. 3. Hannah ME, Hannah WJ, Hodnett ED, Chalmers B, Kung R, Willan A, et al. Outcomes at 3 months after planned cesarean vs planned vaginal delivery for breech presentation at term: the international randomized Term Breech Trial. JAMA 2002;287:1822-31. 4. Lien KC, Mooney B, DeLancey JO, Ashton-Miller JA. Levator ani muscle stretch induced by simulated vaginal birth. Obstet Gynecol 2004;103:31-40.
Weidner et al 5. Damaser MS, Broxton-King C, Ferguson C, Kim FJ, Kerns JM. Functional and neuroanatomical effects of vaginal distention and pudendal nerve crush in the female rat. J Urol 2003;170: 1027-31. 6. Kiff ES, Swash M. Slowed conduction in the pudendal nerves in idiopathic (neurogenic) faecal incontinence. Br J Surg 1984;71: 614-6. 7. Deindl FM, Vodusek DB, Hesse U, Schussler B. Pelvic floor activity patterns: comparison of nulliparous continent and parous urinary stress incontinent women: a kinesiological EMG study. Br J Urol 1994;73:413-7. 8. Gregory WT, Lou JS, Stuyvesant A, Clark AL. Quantitative electromyography of the anal sphincter after uncomplicated vaginal delivery. Obstet Gynecol 2004;104:327-35. 9. Weidner AC, Sanders DB, Nandedkar SD, Bump RC. Quantitative electromyographic analysis of levator ani and external anal sphincter muscles of nulliparous women. Am J Obstet Gynecol 2000;183:1249-56. 10. Bump RC, Mattiasson A, Bo K, Brubaker LP, DeLancey JO, Klarskov P, et al. The standardization of terminology of female pelvic organ prolapse and pelvic floor dysfunction. Am J Obstet Gynecol 1996;175:10-7. 11. Brink CA, Sampselle CM, Wells TJ, Diokno AC, Gillis GL. A digital test for pelvic muscle strength in older women with urinary incontinence. Nurs Res 1989;38:196-9. 12. Hansen SL, Clark SL, Foster JC. Active pushing versus passive fetal descent in the second stage of labor: a randomized controlled trial. Obstet Gynecol 2002;99:29-34. 13. Nandedkar SD, Sanders DB, Stalberg EV. Automatic analysis of the electromyographic interference pattern: part I, development of quantitative features. Muscle Nerve 1986;9:431-9. 14. Nandedkar SD, Sanders DB, Stalberg EV. Automatic analysis of the electromyographic interference pattern: part II, findings in control subjects and in some neuromuscular diseases. Muscle Nerve 1986;9:491-500. 15. Fritel X, Fauconnier A, Levet C, Benifla JL. Stress urinary incontinence 4 years after the first delivery: a retrospective cohort survey. Acta Obstet Gynecol Scand 2004;83:941-5. 16. Cannon TW, Wojcik EM, Ferguson CL, Saraga S, Thomas C, Damaser MS. Effects of vaginal distension on urethral anatomy and function. BJU Int 2002;90:403-7. 17. DeLancey JO, Kearney R, Chou Q, Speights S, Binno S. The appearance of levator ani muscle abnormalities in magnetic resonance images after vaginal delivery. Obstet Gynecol 2003;101: 46-53.