In some women, parturition causes damage to the structure and function of pelvic organs.1,2 Clinical interest has focused on factors in labor and delivery that might predispose women to pelvic floor injury.3 Structural changes include damage to normal pelvic supports, perineal body, and external anal sphincter.4–6 Pelvic nerve studies found prolonged terminal motor latency and evidence of postinjury reinervation.7–10 Peripartum factors that increase the risk of postpartum urinary incontinence are increased prepregnancy body mass index, increased parity, vaginal delivery, prolonged length of the second stage of labor, and episiotomy.11,12 Cesarean delivery protects against urinary incontinence.11,12 Although clinical data suggest that mode of delivery is of paramount importance and, in particular, that cesarean delivery reduces rates of postpartum urinary incontinence, neurophysiologic data suggest that, once labor has progressed to the second stage, cesarean is no longer protective.6,10
The nulliparous pelvis represents the best available clinical model of normal function. Despite their presumed normality, there is a significant incidence of urinary incontinence in nulliparas.13 This prospective study was designed to identify women with preconception incontinence because we speculated that this group might be inherently at increased risk for more severe sequelae from childbirth. We believe that a woman who is continent before pregnancy is the best clinical model of normal pelvic function in which to assess the risk of incontinence associated with labor and delivery factors. We also speculated that the period during which the head was deeply engaged in the pelvis in the second stage of labor might be a variable associated more closely with postpartum incontinence.
Materials and Methods
This prospective survey was conducted in a tertiary-care obstetric hospital with approximately 5000 deliveries each year. Between January 1996 and December 1998, healthy, pregnant nulliparas were invited to enroll. These women were recruited during routine prenatal visits or prenatal classes. Inclusion criteria were nulliparity, no history of urinary tract abnormalities or pelvic surgery, no significant medical illness, and no medication that would alter urinary tract function. The protocol was approved by the institutional ethical review board and involved completion of a simple questionnaire that inquired about symptoms of urinary, fecal, and flatal incontinence. The questionnaire had three sections. Participants chose answers that best described their usual experiences. Each section began with a basic question: “Do you accidentally lose urine from your bladder (flatus or stool from the rectum)?” Participants who answered yes to the first question in each section were asked to answer three questions designed to determine the frequency, severity as reflected by precipitating factors, and impact of incontinence on normal daily activities. A research nurse clarified and completed the questionnaire with each participant.
The questionnaire was 100% sensitive when it was given to a group of 20 patients with the diagnosis of genuine stress incontinence. It is not possible to test specificity because there is no identifiable group of women who are 100% continent of urine. To simplify the analysis of the answers from the questionnaire, women who indicated that they experienced daily episodes of urinary incontinence were classified as having clinically significant incontinence. Severity of urinary incontinence was classified as mild (precipitated only by vigorous exercise), moderate (precipitated by a strong cough or sneeze), or severe (precipitated by daily activities). Data from the labor and delivery of each participant were abstracted from hospital charts. Each participant was mailed a duplicate of the antepartum questionnaire to be completed at 6 weeks' and 6 months' postpartum. On the 6-month questionnaire, patients were asked additional questions about performance of Kegel exercises before, during, and after pregnancy. When answers were omitted, patients were contacted by telephone to complete the questionnaires.
Our power analysis, based on predicted urinary incontinence rates of 5% for the cesarean group and 15% for the vaginal delivery group with an α of .05 and β of .2, indicated that 452 women must complete the study. The calculation assumed a ratio of three vaginal deliveries to one cesarean delivery. Considering current population mobility patterns, we anticipated a loss to follow-up rate of approximately 30%. Based on these calculations, 590 women were required. The dependent variable was the incidence of urinary incontinence. Independent variables included nominal and interval variables. Nominal variables included mode of delivery and performance of Kegel exercises before, during, or after pregnancy. Two ordinal categorical variables were condition of the perineum at delivery and type of repair. The perineum was described as intact or suffering a first- (perineal skin), second- (perineal skin and perineal body), third- (extending through the external anal sphincter), or fourth- (including the anal mucosa) degree laceration. The six repair categories ranged from repair of a simple first-degree perineal laceration to layered repair of a fourth-degree laceration. The six delivery categories were spontaneous vaginal delivery, instrumental assisted vaginal delivery by forceps or vacuum, elective cesarean before labor onset, cesarean in the first stage of labor, and cesarean in the second stage of labor. Chi-square and Fisher exact tests calculated the effects of the nominal variables.
Interval variables consisted of the duration of these labor components (reported in minutes): total labor = onset of active labor to delivery of the infant; second stage of labor = full cervical dilatation until delivery; active second stage = onset of pushing by the mother to delivery; and deeply engaged labor = when the occiput was more than 2 cm below the level of the ischial spines to delivery. Other interval variables were the infant's birth weight (g) and head circumference (cm) and maternal age at the time of completion of the 6-month questionnaire and maternal body mass index. This analysis was done using analysis of variance or, where the distribution was not normal, the Kruskal–Wallis test. Results are reported as relative risk (RR) with confidence intervals (CI). For logistic regression analysis, spontaneous vaginal delivery was used as the reference variable. Significant variables found in univariate analysis were entered into a logistic regression model to test for independent significance. Statistical analysis used Epi Info 6.04 (Centers for Disease Control and Prevention, Atlanta, GA) and SAS 6.12 (SAS Institute, Cary, NC).
Between January 1996 and December 1998, 690 women were enrolled in the study. Ninety-five women did not deliver at our hospital. Ninety-four were presumed to have experienced spontaneous abortion or to have moved from our province. One woman delivered at a peripheral hospital. A total of 595 women delivered at our hospital. There were 147 (25%) cesareans, 333 (56%) spontaneous vaginal deliveries, and 115 (19%) instrumental deliveries. Women with one stillbirth and one neonatal death were withdrawn from the study. Median age was 28 years (range 15–48) and median birth weight was 3489 g. A total of 495 (83.5%) women had epidural anesthetics. Thirty-four patients were lost to follow-up. Thus, 559 patients completed the 6-week questionnaire. Between completion of the 6-week and 6-month questionnaires, 75 patients were lost to follow-up, resulting in a study group of 484 primiparas.
Labor and delivery characteristics of women who were continent and those who were incontinent of urine before pregnancy among the 593 women who delivered at our hospital were compared and no significant difference was found (Table 1).
Urinary incontinence rates at 6 months were 11 of 115 (10%) cesareans, 50 of 233 (22%) spontaneous vaginal deliveries, and 24 of 74 (33%) forceps deliveries. The risk by mode of delivery in women who were continent before pregnancy is shown in Table 2. Spontaneous vaginal delivery was associated with increased risk of urinary incontinence at both follow-up intervals when compared with cesarean delivery (RR 2.1; CI 1.1, 3.7). Forceps delivery further increased the risk of urinary incontinence when compared with the cesarean (RR 3.1; CI 1.7, 5.9) and the spontaneous vaginal delivery (RR 1.5; CI 1.0, 2.3) groups at both follow-up intervals. Risk of incontinence with vacuum delivery was not increased compared with vaginal delivery but was increased compared with cesarean delivery at 6 months (RR 3.5; CI 1.3, 9.1). There was no difference between the cesarean subgroups. Elective cesarean delivery, both elective and after labor onset (RR 0.6; CI 0.3, 1.0) protected against urinary incontinence when compared with spontaneous vaginal delivery. Univariate analysis of obstetric variables at the 6-month follow-up is shown in Table 3. On univariate analysis, the duration of passive second stage was significant and the duration of epidural anesthesia approached statistical significance. These factors, along with forceps delivery, were used in the multivariable analysis. The analysis did not show independent significance of any of these factors (Table 4). The maximum rescaled R2 was .05. Among patients complaining of urinary incontinence at 6 months, the incidence of clinically significant incontinence was 13.5% and classified as severe in 13 (14.8%) of that group. Only three (3.4%) indicated that their urinary incontinence was limiting their activities. Risk of urinary incontinence was not affected by performance of Kegel exercises before (RR 1.06, CI 0.94, 1.2) or during (RR 0.9; CI 0.6, 1.5) pregnancy.
Among the 80 patients experiencing urinary incontinence before pregnancy, 11 (15%) experienced clinically significant postpartum incontinence. At the 6-week and 6-month follow-ups, only 59% and 60%, respectively, of these women still complained of urinary incontinence. To identify risk factors in the subgroup of women with persistent incontinence, we examined the same obstetric variables used in Table 2 and found no association between any of these variables and persistent urinary incontinence. Forceps delivery rates were the same and there was no difference in the frequency and timing of Kegel exercise performance.
Forty-eight women (60%) experienced antepartum urinary and flatal incontinence (double incontinence) and two (2.5%) had triple incontinence (urinary, flatal, and fecal). Postpartum incidence of urinary incontinence in the different risk groups is shown in Table 5. Incontinence rates are indicated as any incontinence or clinically significant incontinence (daily loss of urine). Patients who complained of urinary incontinence alone (group 3) or urinary and flatal incontinence (group 4) before conception had significantly higher rates of urinary incontinence postpartum at 6 weeks and 6 months. Table 6 shows the risk of incontinence by group. At 6 weeks, women in groups 3 (RR 2.7; CI 1.5, 4.9) and 4 (RR 2.6; CI 1.2, 5.5) were at increased risk for clinically significant incontinence. At 6 months, groups 3 (RR 2.3; CI 1.8, 3.0) and 4 (RR 2.8; CI 2.1, 3.5) were still at increased risk for urinary incontinence, but risk of clinically significant incontinence increased only in group 3 (RR 2.9; CI 1.0, 8.1).
This report represents one of a small number of prospective studies on the effects of parturition on the pelvic floor.5,12,14 By identifying women with preconception incontinence, we could investigate the importance of parturition risk factors for urinary incontinence in a group of women who were continent before pregnancy and were presumed to have normal antepartum pelvic function. This study design also permitted us to examine the proposition that women who complain of preconception incontinence are at higher risk of postpartum urinary incontinence.
One of the central tenets that guide obstetric practice is that cesarean delivery should be avoided as long as the safety of mother and child is protected.15 Advocates of this policy cite the maternal risks associated with cesarean delivery. Vaginal birth after previous cesarean delivery is promoted as one way to reduce cesarean rates. Advocates of these policies are guided by a concern for the immediate peripartum well-being of mother and infant. There is a growing body of evidence finding that vaginal birth has a detrimental effect on the pelvic floor.1,3 This information has prompted women to question the wisdom of the precept that vaginal is always better than cesarean birth.16
Clinicians who counsel pregnant women must be prepared to answer the following questions: Does the method of delivery make a difference to short- and long-term maternal pelvic morbidity? If so, can cesarean delivery prevent this morbidity? Independent of the mode of delivery, is there evidence to support the contention that management of parturition can be modified to reduce long-term pelvic floor morbidity? Is it possible to identify women who, before onset of labor and delivery, are at increased risk of pelvic floor morbidity from parturition?
In this study, cesarean delivery performed at all stages of labor on continent nulliparas clearly reduced the rate of urinary incontinence in that group. This finding concurs with evidence from other large clinical studies.11,12 Wilson et al found that cesarean, performed in the first and second stages of labor, was clearly most beneficial to primiparas.11 Only three of 58 primiparas who underwent cesarean delivery developed postpartum stress urinary incontinence. Viktrup et al found that, among women who were continent before and during pregnancy, none of 35 women who underwent cesarean delivery experienced postpartum incontinence.12 Meyer et al found only one woman who experienced postpartum urinary incontinence in 33 who underwent cesarean.5 Most cesarean deliveries in their study were elective and none were performed for midpelvis arrest. Although cesarean delivery reduces the rate of postpartum urinary incontinence, pelvic neurophysiologic injury can occur.6,10 This injury might be cumulative because Wilson et al found that the protective effect of cesarean is lost after three cesareans.11
In our study, spontaneous vaginal delivery increased the risk of postpartum incontinence. In univariate analysis, forceps delivery significantly increased the risk of postpartum urinary incontinence compared with spontaneous vaginal delivery. This significant association was not found with logistic regression analysis. Unfortunately, the R2 of .05 indicates that the logistic regression model did not explain much of the variance. The purported mechanisms of forceps injury to the pelvis include neuropathy and damage to pelvic support structures. Vacuum-assisted delivery did not increase the risk of incontinence when compared with spontaneous vaginal delivery but did compared with cesarean. Based on this evidence, it is reasonable to state that cesarean delivery reduces the risk of urinary incontinence when compared with spontaneous and instrumental vaginal deliveries, and this effect might be a consequence of decreased neurophysiologic morbidity.
We did not identify obstetric factors that increased the risk of postpartum urinary incontinence, which agrees with results of other large clinical studies11,12 and those using multivariable analysis.11 The neurophysiologic studies that implicated obstetric factors such as duration of the second stage of labor and birth weight did not use multivariable analysis, so their conclusions are questionable.6,10 The current medical literature does not implicate obstetric factors as independent risk factors for postpartum urinary incontinence.
The most important clinical decision during parturition involves the method by which delivery is done. Despite the limitations of the statistical methodology used in the neurophysiologic studies, which make the attribution of incontinence risk to specific obstetric factors questionable, those studies clearly showed that structural and functional damage occurs to the pelvis during the process of labor and delivery, regardless of the method of delivery.7,10 This damage appears to have a cumulative effect with subsequent deliveries and to confer a long-term risk for urinary incontinence. In their longitudinal study, Snooks et al found that women who recover normal function over the short term might experience recurrence of incontinence symptoms over the long term.8 Until better study designs identify specific obstetric factors that are independently responsible for the neurophysiologic damage sustained during parturition, guidelines for modifying labor management cannot be developed.
There are no accepted clinical criteria to identify women at increased risk for incontinence after parturition. No association was found between markers of collagen weakness, such as abdominal stria and joint hypermobility, and risk of urinary incontinence.14 Antenatal bladder neck hypermobility and maternal history of urinary incontinence were risk factors for postpartum incontinence.17,18 In this study, history of antepartum incontinence correlated with a significantly increased risk of postpartum incontinence. If these findings are confirmed, antepartum symptoms of urinary or anal incontinence might identify a woman who has inherent pelvic dysfunction, and this woman should be counseled that she is at increased risk of developing more severe urinary incontinence as a result of pelvic damage at the time of parturition.
We and other investigators found that some women who complained of antepartum urinary incontinence found that this symptom resolved postpartum. Without a more precise elucidation of the cause of antepartum incontinence, it is difficult to offer any explanation for this apparent therapeutic effect of labor and delivery.
Our follow-up was limited to 6 months. After a follow-up of 3 months, Viktrup et al found that symptoms had resolved in most of their patients.12 We chose a 6-month follow-up to provide sufficient time for recovery of the pelvic structures and minimize the risk of loss to follow-up. The rate of urinary incontinence at 6 months was the same as at 6 weeks. Viktrup et al reported a 7% de novo incontinence rate and their persistent incontinence rate was only 4%.12 There are two possible explanations for the differences between the rates of incontinence reported by those authors and Wilson et al11 (34%) and this study. Longer follow-up (Viktrup et al followed patients for 12 months) allows more time for pelvic function to return to normal. There were no forcep deliveries in the series by Viktrup et al. Longer-term follow-up might resolve this issue. Only 3–4% of women in our study indicated that their urinary incontinence interfered with their daily activities. Such a low incidence of significant incontinence might not warrant much clinical concern, but evidence suggests that symptoms may recur over time8 and that future parturition results in increasing rates of significant incontinence.11 Over the course of a woman's life, her obstetric experience might be the most significant risk factor for subsequent urinary incontinence and pelvic prolapse. Efforts to ameliorate the impact of labor and delivery on the pelvis should continue.
If cesarean delivery is proved to minimize the neurophysiologic morbidity resulting from labor and delivery, high-risk women might choose this mode of delivery, especially when instrumental intervention is necessary for a vaginal delivery.
1. Sultan AH, Monga AK, Stanton SL. The pelvic floor sequelae of childbirth. Br J Hosp Med 1996;55:575–9.
2. Deindl FM, Vodusek DB, Hesse U, Schüssler B. Pelvic floor activity patterns: Comparison of nulliparous continent and parous urinary stress incontinence women. A kinesiological EMG study. Br J Urol 1994;73:413–7.
3. Handa VL, Harris TA, Ostergard DR. Protecting the pelvic floor: Obstetric management to prevent incontinence and pelvic organ prolapse. Obstet Gynecol 1996;88:470–8.
4. Sultan AH, Kamm MA, Hudson CN, Chir M, Thomas JM. Bartram CI. Analsphincter disruption during vaginal delivery. N Engl J Med 1993;329:1905–11.
5. Meyer S, Schreyer A, DeGandi P, Hohlfeld P. The effects of birth on urinary incontinence mechanisms and other pelvic floor characteristics. Obstet Gynecol 1998;92:613–8.
6. Sultan AH, Kamm MA, Hudson CN. Pudendal nerve damage during labour: Prospective study before and after childbirth. Br J Obstet Gynaecol 1994;101:22–8.
7. Snooks SJ, Henry MM, Swash M. Faecal incontinence due to external anal sphincter division in childbirth is associated with damage to the innervation of the pelvic floor musculature: A double pathology. Br J Obstet Gynaecol 1985;92:824–8.
8. Snooks SJ, Swash M, Mathers SE, Henry MM. Effect of vaginal delivery on the pelvic floor: A 5-year follow-up. Br J Surg 1990;77:1358–60.
9. Snooks JJ, Swash M, Setchell M, Henry MM. Injury to innervation of the pelvic floor sphincter musculature in childbirth. Lancet 1984;2:546–50.
10. Allen RE, Hosker GL, Smith ARB, Warrell DW. Pelvic floor damage and childbirth: A neurophysiologic study. Br J Obstet Gynaecol 1990;97:770–9.
11. Wilson PD, Herbison R, Herbison GP. Obstetric practice and the prevalence of urinary incontinence three months after delivery. BMJ 1996;103:154–61.
12. Viktrup L, Lose G. Epidural anesthesia during labour and stress incontinence after delivery. Obstet Gynecol 1993;82:984–6.
13. Thomas TM, Plymat KR, Blannin J, Meade TW. Prevalence of urinary incontinence. BMJ 1980;281:1243–5.
14. Chaliha C, Veena K, Stanton SL, Monga A, Sultan AH. Antenatal prediction of postpartum urinary and fecal incontinence. Obstet Gynecol 1999;94:689–94.
15. Paterson-Brown S, Nicholas MF. Caesarean section: Every woman's right to choose? Curr Opin Obstet Gynecol 1997;9:351–5.
16. Al-Mufti R, McCarthy A, Fisk NM. Survey of obstetrician's personal preference and discretionary practice. Eur J Obstet Gynecol Reprod Biol 1997;73:1–4.
17. Skoner MM, Thompson WD, Caron VA. Factors associated with stress urinary incontinence in women. Nurs Res 1994;43:301–6.
© 2001 The American College of Obstetricians and Gynecologists
18. King JK, Freeman RM. Is antenatal bladder neck mobility a risk factor for postpartum stress incontinence. Br J Obstet Gynecol 1998;105:1300–7.