The effect of pregnancy and mode of delivery on the prevalence of urinary and fecal incontinence

Transcription

1 American Journal of Obstetrics and Gynecology (2005) 193, The effect of pregnancy and mode of delivery on the prevalence of urinary and fecal incontinence Vikki McKinnie, MD, a Steven E. Swift, MD, a, * Wei Wang, MSPH, b Patrick Woodman, DO, c Amy O Boyle, MD, c Margie Kahn, MD, d Michael Valley, MD, e Deirdre Bland, MD, f Joe Schaffer, MD g Division of Benign Gynecology, Department of Obstetrics & Gynecology, a and Department of Biometry and Epidemiology, b Medical University of South Carolina, Charleston, SC; Division of Urogynecology and Pelvic Reconstructive Surgery, Department of Obstetrics & Gynecology, c Madigan Army Medical Center, Tacoma, WA; Division of Gynecology, Department of Obstetrics & Gynecology, and Division of Urology, Department of Urology, d University of Texas Medical Branch-Galveston, Galveston, TX; Department of Obstetrics & Gynecology, e Health System Minnesota, Shakopee, MN; Department of Obstetrics & Gynecology, Section on Gynecology, f Bowman Gray Medical Center, Winston-Salem, NC; and Division of Urogynecology and Reconstructive Pelvic Surgery, Department of Obstetrics & Gynecology, g University of Texas Southwestern Medical Center, Dallas, TX Received for publication November 18, 2004; revised March 16, 2005; accepted March 25, 2005 KEY WORDS Urinary and fecal incontinence: mode of delivery Objective: The purpose of this study was to determine the relative effects of pregnancy and mode of delivery on the prevalence of urinary and fecal incontinence. Study design: This was a prospective, observational multicenter study of women presenting to 6 gynecology clinics. Demographic data collected included: height, weight, gravidity, parity, and number of vaginal deliveries. Patients were diagnosed with incontinence by questionnaire. Standard univariate logistic regression analyses were performed to determine the contribution of pregnancy, mode of delivery, and on the prevalence of urinary and fecal incontinence. Results: One thousand and four women were enrolled over an 18-month period. Two hundred and thirty-seven and 128 subjects had urinary and fecal incontinence, respectively. Odds ratio (95% CI) calculated for the prevalence of urinary incontinence by pregnancy and mode of delivery were: any term pregnancy vs no term pregnancy was 2.46 ( ), any term pregnancy but no vaginal deliveries (cesarean section only) vs no term pregnancy was 1.95 ( ), any term pregnancy and at least 1 vaginal delivery vs no term pregnancy was 2.53 ( ), and any term pregnancy but no vaginal delivery (cesarean section only) vs any term pregnancy, and at least 1 vaginal delivery was 1.30 ( ). Odds ratio (95% CI) calculated for the prevalence of fecal incontinence by pregnancy and mode of delivery were: any term pregnancy vs no term pregnancy was 2.26 ( ), any term pregnancy but no vaginal deliveries (cesarean section Presented at the Sixty-Seventh Annual Meeting of the South Atlantic Association of Obstetricians and Gynecologists, January 22-25, 2005, White Sulphur Springs, WVa. * Reprint requests: Steven Swift, MD, Department of Obstetrics and Gynecology, Medical University of South Carolina, 96 Jonathon Lucas St, Suite 634, Charleston, SC swifts@musc.edu /$ - see front matter Ó 2005 Mosby, Inc. All rights reserved. doi: /j.ajog

2 McKinnie et al 513 only) vs no term pregnancy was 1.13 ( ), any term pregnancy and at least 1 vaginal delivery vs no term pregnancy was 2.41 ( ), and any term pregnancy but no vaginal deliveries (cesarean section only) vs any term pregnancy, and at least 1 vaginal delivery was 2.15 ( ). and age did not impact these results. Conclusion: Pregnancy increases the risk of urinary and fecal incontinence. Cesarean section does not decrease the risk of urinary or fecal incontinence compared to pregnancy with a vaginal delivery. Ó 2005 Mosby, Inc. All rights reserved. Since the turn of the last century, investigators have hypothesized that vaginal delivery of an infant can result in incontinence. This observation was first put forth by Howard Kelly MD in his landmark 1914 article, in which he wrote, There is a type of urinary incontinence in women, without manifest injury to the bladder and having no relationship to fistula, which most frequently comes on following childbirth. 1 However, it was not until the mid 1980s that investigators elucidated a possible mechanism. They demonstrated that urinary and fecal incontinence were secondary to a pudendal nerve injury, and further demonstrated that vaginal delivery could produce a similar pudendal nerve injury, which could be avoided by cesarean section. 2-5 These investigators demonstrated that there was a stretching and compression injury of the pudendal nerve that was likely due to the fetal vertex passing through the birth canal. This neuropathy can then lead to a weakening of the pelvic floor musculature that eventually results in incontinence of urine and or stool. Although this provided a biologically plausible mechanism to explain how the vaginal delivery of an infant could predispose to the development of incontinence, no prospective clinical trials with long-term follow-up have documented this relationship. More importantly, no clinical trials have been done to determine if cesarean section is protective against the development of incontinence. Two large epidemiologic studies have reported on the relationship between incontinence and aspects of pregnancy and mode of delivery. 6,7 While both studies agreed that the risk of urinary incontinence is higher among parous women compared with nulliparous women, they disagreed about the protective effect of cesarean delivery. In response to this dilemma and the evolving controversy, we sought to determine the impact of pregnancy and mode of delivery on the prevalence of urinary and fecal incontinence, employing a previously developed database of gynecology clinic patients. Material and methods Data were collected from 1004 women enrolled over an 18-month period as part of the Pelvic Organ Support Study Project, a multicenter study involving 6 outpatient gynecology clinics in the US. 8 The study was approved by local IRBs at each site, and before participation each subject signed an informed consent. Demographic data were collected and included; age, height, and weight. In addition, detailed information regarding the subjects obstetric history was collected and included; gravity, parity, number of vaginal deliveries, and weight of largest infant delivered vaginally. Subjects also completed a questionnaire involving 7 genitourinary symptoms. Two questions queried subjects specifically on the presence and bothersome nature of symptoms of either urinary or fecal incontinence. Subjects were asked if they have uncontrollable loss of urine, and if they have uncontrollable loss of stool or gas. If subjects responded yes or sometimes to the presence of the symptom they were asked a follow-up question that queried if the symptom bothered them. If they responded yes to the follow-up bothersome question they were considered incontinent for the purpose of this study. This then represents a symptom of incontinence that is bothersome to the subject and one which they would complain about. This is consistent with the International Continence Society Standardization of Terminology Committee s definition of urinary incontinence, and the Second International Consultation on Incontinence s definition of fecal incontinence. 9,10 Odds ratios (OR) and confidence intervals (CI) for having either urinary or fecal incontinence were then determined for the following groups: women who had a term pregnancy vs women who had never had a term pregnancy (P R1 vs. P = 0); women who had a term pregnancy and no vaginal delivery (cesarean section only) vs women who had never had a term pregnancy (P R1 and vaginal delivery = 0 vs P = 0); women who had a term pregnancy and at least 1 vaginal delivery vs women who never had a term pregnancy (P R1 and vaginal delivery R1 vs P = 0); women who had term pregnancy but no vaginal delivery (cesarean section only) vs women who had a term pregnancy, and at least 1 vaginal delivery (P R1 and vaginal delivery R1 vs P R1 and vaginal delivery = 0). Univariate logistical regression analysis was performed to evaluate the effect of maternal body mass index (), age, and increasing weight of a vaginally delivered infant on the risk of both urinary and fecal incontinence. In addition, for the risk factor parity, a multiple logistic regression model was rerun with covariating

3 514 McKinnie et al Table I Effect of pregnancy and mode of delivery on the prevalence of urinary incontinence Risk factor for urinary incontinence No. with disease* OR and 95% CI P = 0 (n = 174) P R1 (n = 811) vs P = ( ) P R1 and vaginal delivery = ( ) (n = 91) vs P =0 P R1 and vaginal delivery R ( ) (n = 713) vs P =0 P R1 and vaginal delivery R1 vs. P R1 and vaginal delivery = ( ) * No. with disease refers to the number of subjects with urinary incontinence designated by the (n =) in that category. Table III Univariate logistic regression analysis results for and age on urinary and fecal incontinence Risk factor for urinary incontinence OR (95% CI) Overweight 1.60 ( ) Obese 2.55 ( ) Age (per year) 1.04 ( ) Risk factor for fecal incontinence OR (95% CI) Overweight 1.08 ( ) Obese 1.20 ( ) Age (per year) 1.05 ( ) Table II Effect of pregnancy and mode of delivery on the prevalence of fecal incontinence Risk factor for urinary incontinence and age to obtain adjusted ORs. Statistical significance was considered if P!.05. Descriptive statistics are summarized as counts and percentages for categorical variables; mean and standard deviations for continuous variables. Standard univariate unadjusted logistic regression analyses were performed to calculate ORs along with 95% CIs for all risk factors, independently. Results No. with disease* OR and 95% CI P = 0 (n = 174) P R1 (n = 810) vs P = ( ) P R1 and vaginal delivery = ( ) (n = 91) vs P =0 P R1 and vaginal delivery R ( ) (n = 713) vs P =0 P R1 and vaginal delivery R1 vs. P R1 and vaginal delivery = ( ) * No. with disease refers to the number of subjects with fecal incontinence designated by the (n =) in that category. The mean age of subjects was 42.7 G 13.9 (G SD) years. Racial distribution was 43% Caucasian, 24% African American, 29% Hispanic, 2% Asian, 2% other. Of the 1004 subjects in the study, 978 had complete data on gravidity (G), parity (P), and normal spontaneous vaginal delivery (NSVD), and fully answered the questionnaire. They are the basis of this report. There were 174 nulliparous women (P = 0), 91 parous women with Table IV Multiple logistic regression with age C and P for urinary and fecal incontinence Risk factor for urinary incontinence OR (95% CI) Age (per year) 1.04 ( ) Overweight 1.48 ( ) Obese 2.42 ( ) P P =0vsPO ( ) P O 0 and NVD = 0 vs 1.26 ( ) P O 0 and NVD O0 Risk factor for fecal incontinence OR (95% CI) Age (per year) 1.05 ( ) Overweight 0.99 ( ) Obese 1.10 ( ) P P =0vsPO ( ) P O 0 and NVD = 0 vs P O ( ) and NVD O 0 no vaginal deliveries (P R1 and vaginal delivery = 0), and 712 parous women with at least 1 vaginal delivery (P R1 and vaginal delivery R1). Two hundred and thirty-seven, or 23%, of this population reported bothersome urinary incontinence, and 128, or 13%, reported bothersome anal incontinence. The ORs as determined by univariate analysis for the various pregnancy and delivery routes demonstrate that for both urinary and fecal incontinence the risk is greater in parous versus nulliparous women (Tables I and II). However, those subjects delivered by cesarean section exclusively did not have a statistically significant different risk of fecal or urinary incontinence than those subjects that had at

4 McKinnie et al 515 least 1 vaginal delivery of a term infant. This suggests that cesarean section does not provide protection from incontinence, but that nulliparity has a protective effect. A univariate logistic regression analysis on and age and incontinence revealed that only obese women are at increased risk for urinary incontinence, and overweight and obese women are not at any increased risk for fecal incontinence in this population. In addition, increasing age increases the risk of both fecal and urinary incontinence (Table III). A secondary multiple logistic regression analysis on the effect of and age on the data regarding pregnancy and mode of delivery failed to demonstrate any increased risk for increasing age or. Subject s had essentially no effect on the OR (2.30 and 2.29) for urinary and fecal incontinence in the P = 0 vs P R1 (Table IV). We further evaluated the effect of increasing weight of a vaginally delivered infant on the risk of urinary and fecal incontinence. For each additional 16 ounces of infant weight delivered vaginally, the OR (95% CI) for urinary incontinence and fecal incontinence increased by 1.13 ( ) and 1.10 ( ), respectively. However, women who are obese women are at increased risk of urinary incontinence, but they do not incur an additional risk from a pregnancy. Comment There is a developing consensus of opinion in obstetrics and gynecology that pregnancy and a vaginal delivery contributes to the eventual development of urinary and fecal incontinence. This relationship has biologic plausibility and stems from research that established a relationship between vaginal delivery, pudendal neuropathy, and urinary and fecal incontinence. Also, there is a growing controversy that cesarean delivery protects against this injury, and should be offered to patients who are concerned about protecting the function of their pelvic floor. However, while the data are suggestive, there is no prospective randomized trial with long-term follow-up to conclusively establish the protective nature of elective cesarean section. To date, there has been only 1 prospective randomized trial that was performed in subjects with breech presentation comparing planned cesarean to vaginal delivery. Their outcomes of fecal and urinary incontinence were measured at 3 months postdelivery. 11 These investigators found no difference in fecal incontinence, and only a small statistically significant difference in urinary incontinence. However, the short follow-up makes these data of limited value. Because of the brewing controversy, and the lack of data, several investigators have recently called for prospective long-term trials to be done, but the scope of such a trial may preclude its feasibility. 12,13 Therefore, it may be some time before we have conclusive type 1 evidence, and so we must look to observational, epidemiologic studies to help us decide how to address this issue. Should cesarean delivery be offered electively to patients to prevent pelvic floor injury? Two large epidemiologic studies have attempted to answer this question. However, the studies reached different conclusions regarding the protective nature of cesarean delivery. The Norwegian Epidemiology of Incontinence in the County of Nord-Trøndelag (EPI- NCONT) study was a large population-based questionnaire study involving over 15,000 women that compared 2 groups of women and their risk of urinary incontinence: women who had only vaginal deliveries and women who had only cesarean deliveries. 6 They did not include women who had had both, and this study found that cesarean delivery was protective against developing urinary incontinence. The South Australian Health Omnibus Survey was a population-based interview study involving over 3000 women that evaluated the relationship between pregnancy/mode of delivery and the prevalence of both urinary and fecal incontinence. 7 In contrast, this study compared women who had cesarean deliveries only with women who had any vaginal delivery, including women with only vaginal deliveries and women with both vaginal and cesarean deliveries. In addition, they analyzed women who had labored and then had a cesarean delivery vs women who had an unlabored cesarean delivery. They demonstrated no protective effect of cesarean delivery, either with or without labor, against the eventual development of urinary or fecal incontinence. Both of these studies agreed and demonstrated a greater risk of incontinence in parous over nulliparous women. The differences in their conclusions are most likely related to the populations they chose to compare. Also, the definitions of incontinence used in these 2 studies were different, and neither complied strictly with the ICS definition of incontinence. The use of nonstandard and differing definitions may also explain the differences in prevalence of urinary incontinence in the 2 populations, 20.7% and 35.3% in the Norwegian and South Australia studies, respectively. So it may have been that the populations studied and the incontinence definitions used contributed to the differences in results. The present study, although not the gold standard, prospective, randomized clinical trial, does provide additional and unique information. In this study, we used standard definitions for defining urinary and fecal incontinence. We only included those women who reported bothersome symptoms of incontinence as having incontinence, consistent with the present standard definitions. 9,10 The reported prevalence rates in this data for fecal and urinary incontinence are consistent with previous reports. 9,14,15 In addition, our population had a diverse racial and geographic distribution. The previously mentioned studies did not report

5 516 McKinnie et al on racial distribution and were very restricted in their geography. Similar to the South Australian Health Omnibus, we compared a cesarean only group with those who had any vaginal delivery. We felt this technique of evaluating data would more accurately reflect what would occur with a clinical policy of allowing elective cesarean sections. Therefore, it is not surprising that our results agree with theirs, and suggest that cesarean delivery is not protective against developing urinary or fecal incontinence (Tables I and II). Our data are also consistent with both of the previously mentioned studies, suggesting that pregnancy provides the greatest risk for developing urinary and fecal incontinence. If the data on fecal incontinence are evaluated closely, it appears that, while the confidence intervals cross one, suggesting no difference, the OR is suggestive of a trend. In those women delivered by cesarean only vs those delivered by the vaginal only or combined vaginal and cesarean delivery, the OR is O2 and the lower limit of the CI is There were only 7 subjects who had a term pregnancy without a vaginal delivery (P R1 and vaginal delivery = 0) and fecal incontinence (Table II). Therefore, the lack of statistical difference may be related to low numbers in this group, and we can only postulate that a larger cohort of subjects with fecal incontinence and no term vaginal deliveries may have revealed a different finding. There is another apparent contradiction in our results. When comparing subjects with no term pregnancies with the group who were delivered by cesarean section only, there is no increased risk of either fecal or urinary incontinence, and when comparing women with no term pregnancy with those with any vaginal delivery, there is and increased risk of both fecal and urinary incontinence. However, when comparing the cesarean section only group with the group with 1 or more vaginal deliveries, there is no difference. It is difficult to explain these differences easily, it may be an issue with numbers or it may point to the complex interaction between pregnancy and pelvic floor disorders. We took into consideration in our analysis. Obese patients tended to have more urinary incontinence, but did not appear to affect fecal incontinence (Table III). When considering in a multiple logistic regression analysis, there was no impact on developing urinary or fecal incontinence (Table IV). However, fetal weight of the largest infant delivered vaginally did appear to affect both urinary and fecal incontinence, with a roughly 10% increase with each 1-pound increase in fetal weight. We did not record the weight of infants delivered by cesarean section, so it is difficult to determine if fetal weight would have a similar impact in those patients delivered by cesarean section. The controversy still remains about whether or not cesarean section should be offered electively as a mode of delivery to protect patients from pelvic floor injury and dysfunction. Our study does not provide a solution to this complex issue. However, it does provide support that cesarean delivery itself does not lessen the risk of developing urinary or fecal incontinence when compared with vaginal delivery. Pregnancy in and of itself increases the risk of developing incontinence regardless of the mode of delivery. Therefore, our data cannot recommend a policy of offering elective cesarean sections to all subjects. However, there is probably a cohort of women who would benefit from elective cesarean sections, and it will be the job of future investigators to determine who they are so that a rational policy of elective cesarean sections can be advocated. References 1. Kelly HA, Dunn WM. Urinary incontinence in women, without manifest injury to the bladder. Surg Gynecol Obstet 1914;18: Snooks SJ, Swash M, Henry MM, Setchell M. Risk factors in childbirth causing damage to the pelvic floor innervation. Int J Colorectal Dis 1986;1: Snooks SJ, Swash M, Setchell M, Henry MM. Injury to innervation of pelvic floor sphincter musculature in childbirth. Lancet 1984;8: Allen RE, Hosker GL, Smith ARB, Warrell DW. Pelvic floor damage and childbirth: a neurophysiological study. BJOG 1990;97: Smith ARB, Hosker GL, Warrell DW. The role of partial denervation of the pelvic floor in the aetiology of genital prolapse and stress incontinence of urine. A neurophysiological approach. BJOG 1989;96: Rotveit G, Daltveit AK, Hannestad YS, Hunskaar S. Urinary incontinence after vaginal delivery or cesarean section. N Engl J Med 2003;348: 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: Swift SE, Woodman P, O Boyle A, Kahn M, Valley M, Bland D, et al. Pelvic Organ Support Study; the distribution, clinical definition and epidemiologic condition of pelvic organ support defects. Am J Obstet Gynecol 2005;192: Abrams P, Cardozo L, Fall M, Griffiths D, Rosier P, Ulmsten U, et al. The standardization of terminology in lower urinary tract function: report from the Standardisation Sub-committee of the International Continence Society. Neurouro Urodyn 2002;21: Norton C, Christiansen J. Anal incontinence. In: Abrams P, Cardozo L, Khoury S, Wein A, editors. Incontinence: 2nd international consultation on incontinence. 2nd ed. Plymouth: Pub. Health Publications LTD 2002, Dist. Plymbridge Distrubtors, Ltd; p Hannah ME, Hannah WJ, Hodnett ED, Chalmers B, Kung R, Willan A, et al. Outcomes at three months after planned cesarean cesarean versus planned vaginal delivery for breech presentation at term. JAMA 2002;287: Ecker J. Once a pregnancy, always a cesarean? Rationale and feasibility of a randomized controlled trial. Am J Obstet Gynecol 2004;190: Wax J, Cartin A, Pinette MG, Blackstone J. Patient choice cesarean: an evidence-based review. Obstet Gynecol Survey 2004;59:

6 McKinnie et al Samuelsson E, Arne V, Tibblen G. A population study of urinary incontinence and nocturia among women aged years. Acta Obstet Gynecol Scand 1997;76: Jackson S, Hull T. Fecal incontinence in women. Obstet Gynecl Survey 1998;53: Discussion JOHN R. PARTRIDGE, MD, FACOG. Dr McKinnie s paper The effect of pregnancy and mode of delivery on the prevalence of urinary and fecal incontinence is a valuable contribution to the evolving data on the issue of the relative safety and maternal after-effects of cesarean vs vaginal delivery. This is a relatively large multicenter prospective study of 1004 women conducted over 18 months. The data gathered showed that patients managed exclusively by cesarean did not have statistically significantly lower fecal or urinary incontinence compared with those who had had at least 1 vaginal delivery at term. The authors related their data to that of 2 large epidemiologic studies from Norway and Australia, which disagreed with each other on the protective effect of cesarean. Pointing out that they, like the Australian investigators, had differentiated patients who had only had cesareans vs others who had also had vaginal births, Dr McKinnie noted that her findings agreed with those in the Australian paper. McKinnie et al also noted a correlation of increasing risk of fecal and urinary incontinence with increasing size of the infant delivered, and increased urinary incontinence in obese patients, but with no additional increase in those patients attributable to vaginal birth. This new study agrees with other recent publications but disagrees with others. Lal s 2003 report noted anal incontinence in 2 out of 104 primiparous women who had had emergency cesarean with no previous labor. Out of 80 who had had scheduled cesareans, 2 (3%) had severe anal symptoms vs just 1 (1%) delivered vaginally. They concluded pregnancy itself has the capacity to injure pelvic floor function. 1 Dr McKinnie s study contrasts with a recent report by Groutz et al, in which a prospective study of 363 women 1 year after their first delivery found significantly lower urinary incontinence in those who had delivered by elective scheduled cesarean than by other modalities. 2 A 2003 paper by McKenna et al turned to computer decision analysis modeling of the approximately 4 million US births in 2000 to evaluate the effect of subsequent vaginal delivery in women who had had previous repairs of rectal injury after previous vaginal delivery, and who had attained continence after that previous repair. They noted subsequent vaginal delivery provoked a high incidence of anal incontinence in these women, a sequela that was largely prevented by management of the new pregnancy with elective cesarean. 3 A small 2002 German study by Faridi concluded that in 71 patients studied there was no alteration of anal pressure, rectal sensation, or pudendal nerve latency among the patients delivered by cesarean vs a 4.8 incidence of anal incontinence in those delivered vaginally. 4 A 2000 French report by Abramowitz et al of 233 patients found new anal sphincter defects in 16.7% of patients delivered vaginally, and none in those delivered by cesarean. However, incontinence was present in under half of those with demonstrated sphincter defects, pointing to the multifactorial nature of anal incontinence. 5 Not addressed in Dr McKinnie s article are a number of points that might shed additional light, for instance, was there any apparent impact of maternal age? Pollack et al identified increasing maternal age as a risk factor for anal incontinence as a sequela of vaginal delivery, with twice that risk for a 30-year-old as for a 20-year-old woman. 6 Another aspect that would be interesting to know is whether the stage of labor at which cesarean was done made a difference in pelvic floor injury. In 1998, Fynes et al published data from Ireland that showed that anal sphincter injury was not prevented by cesarean performed in late labor (at 8 cm dilation or further). And they differentiated 2 groups with rectal injury: muscular injury in patients who delivered vaginally vs nerve injury in those who had cesarean late in labor. 7 Increasingly we hear calls for offering cesareans on demand, with one principal argument in favor of such a policy being to protect the pelvic floor and its function. Dr McKinnie s article provides important additional data that must be balanced in weighing this important issue. References 1. Lal M, H Mann C, Callender R, Radley S. Does cesarean delivery prevent anal incontinence? Obstet Gynecol 2003;101: Groutz A, Rimon E, Peled S, Gold R, Pouzner D, Lessing JB, et al. Cesarean section: does it really prevent the development of postpartum stress urinary incontinence? A prospective study of 363 women one year after their first delivery. Neurourol Urodyn 2004;23: McKenna D, Ester J, Fischer J. Elective cesarean delivery for women with a previous anal sphincter rupture. Am J Obstet Gynecol 2003;189: Faridi A, Willis S, Schelzig P, Siggelkow W, Schumpelick V, Rath W. Anal sphincter injury during vaginal deliverydan argument for cesarean section on request? J Perinat Med 2002;30: Abramowitz L, Sobhani I, Ganansia R, Vuagnat A, Benifla J, Darai E, et al. Are sphincter defects the cause of anal incontinence after vaginal delivery? Results of a prospective study. Dis Colon Rectum 2000;43: Pollack J, Nordenstam J, Brismar S, Lopez A, Altman D, Zetterstrom J. Anal incontinence after vaginal delivery: a five year prospective cohort study. Obstet Gynecol 2004;104: Fynes M, Donnelly V, O Connell P, O Herlihy C. Cesarean delivery and anal sphincter injury. Obstet Gynecol 1998;92: