Anal sphincter trauma during vaginal delivery is believed to be the major cause of anal incontinence in women of childbearing years,1,2 with prevalence estimated at between 4% and 26% at 3–6 months postpartum.3–6 Symptoms of anal incontinence are strongly associated with disruption of the anal sphincter during vaginal delivery,7–9 but it is increasingly well-recognized that such disruption does not inevitably result in symptoms. Supporting this, a study of 134 women using anal endosonography to identify anal sphincter tears after vaginal delivery found that 17 (18%) of 97 women whose continence did not deteriorate postpartum had sonographic evidence of perineal trauma, including six women with anal sphincter tears.10 It is also recognized that many women present with anal incontinence later in life, several years after childbirth.11 Although some of these women will prove to have sustained anal sphincter damage previously, it is unclear to what extent these injuries underpin their current symptoms.
Longitudinal studies that follow the development of symptoms over time are needed to estimate the influence of sphincter tears on future symptoms, but these are inevitably time consuming and have usually relied on clinical detection of the tear after delivery. For example, Fornell and colleagues12 reported the subsequent decline in anal function over time of 82 women whose third-degree tear had been diagnosed by clinical examination, and Pollack and co-workers compared symptoms of incontinence at 5 years in women with and those without a clinically diagnosed tear.13 Similarly, Faltin and colleagues14 identified 445 women with a clinically diagnosed sphincter tear and compared their symptoms at 18 years with matched controls. However, it is well-recognized that clinical examination misses the majority of anal sphincter tears, and anal endosonography immediately after delivery is needed to estimate the true incidence: A study of 150 women with no evidence of injury on postpartum clinical examination found unsuspected anal sphincter tears revealed by anal endosonography in 28%.15
To estimate the influence of clinically silent anal sphincter tears on symptoms of anal incontinence over time, we performed anal endosonography after childbirth in a consecutive group of consenting women and assessed symptoms of anal incontinence at 5 and 10 years postpartum.
MATERIALS AND METHODS
The local Clinical Research Ethics Committee granted approval for the study. A female principal investigator prospectively recruited a cohort of consecutive primigravid women during their third trimester, inviting them to participate in the study during their routine antenatal clinic visit to the obstetric department of a large teaching hospital. The purpose of the study was explained, and verbal informed consent obtained; the Ethics Committee did not require written consent. Inclusion criteria were primiparity, low-risk singleton pregnancy, no apparent complications, no previous pelvic floor or anal surgery, and ability to give informed consent. Approximately 95% of women approached agreed to participate, and ultimately 156 consecutive consenting women were recruited. Recruitment lasted 10 months, from October 1996 to July 1997 inclusive.
Bowel habit was characterized using a questionnaire derived from the Jorge and Wexner16 20-point scoring system for anal incontinence. Two additional questions were asked: “Are you able to defer evacuation for at least five minutes?” (Possible responses: always, usually, sometimes, rarely, never; scores 0 to 4, respectively) and “Now that you are pregnant, have you noticed any change in your bowel habit from the time before you were pregnant?” (Possible responses: no, increased constipation, increased diarrhea, increased involuntary flatus, increased involuntary stool loss, decreased ability to defer evacuation for at least five minutes). The first question was added to the established scoring system to give a total possible score of 24 (with 0 representing full continence and 24 total incontinence). The second question was used to identify those women who believed their bowel habit had changed, irrespective of their questionnaire. A clinic nurse administered the questionnaire.
After routine antenatal examination and completion of the questionnaire, all consenting women underwent anal endosonography by the principal investigator using a B and K Medical ADI 2001 Panther ultrasound scanner fitted with a type 1850 anal endoprobe and type 6004 10 MHz transducer (B and K Medical, Gentofte, Denmark) using a standard technique, and images were taken at proximal, mid, and distal anal canal levels as per standard endosonographic practice.17–19 The proximal anal canal level was defined by visualization of the puborectalis sling. The mid canal level was defined by the completion of the external anal sphincter ring anteriorly in combination with maximum internal anal sphincter thickness. The distal canal was defined as the level immediately caudal to the termination of the internal anal sphincter. Sphincter integrity was noted, and the presence of external and/or internal sphincter defects, if any, were recorded on a study sheet. Sphincter defects were defined by discontinuity of the external and/or internal sphincter ring. Women with perineal tears that did not involve the anal sphincter apparatus were classified as having no tear. Images were stored using a personal computer fitted with a picture acquisition system (Medimage; Vepro, MHS Medizintechnik, Vienna, Austria). A single operator with personal experience of more than 1,000 anal scans performed the endosonography and was unaware of the questionnaire scoring.
After delivery, patients were re-examined at a 3-month postpartum pelvic floor clinic. Any woman who had a clinically detected third-degree anal sphincter tear at the time of delivery was excluded as was any woman who had a cesarean delivery because anal sphincter disruption is known not to occur in this group.20 Participants completed a questionnaire identical to the one administered antenatally. The previously completed questionnaire was not available for comparison. Women were asked about the duration of any change in bowel habit and whether it had resolved by the time of the clinic visit. Women then underwent a second endosonographic examination so that the presence of any sonographic anal sphincter defect after delivery could be estimated.
Five years after this examination, the principal investigator attempted to contact all participants by telephone, at which time the bowel habit questionnaire was completed again. If unsuccessful, attempts were made to contact participants by letter. Any intervening obstetric history was asked for, in particular whether there had been any subsequent vaginal deliveries.7 At 10 years participants were contacted again in a similar fashion by the same principal investigator and the questionnaire completed once more. Anal endosonography was again offered to any woman delivering vaginally subsequent to administration of the 5-year questionnaire.
Three outcome measures were examined: anal incontinence scores at 10 years, the change in incontinence scores from baseline to 10 years, and the change in incontinence scores from 5 years to 10 years. Participants were partitioned into those with incontinence scores of zero and those with positive scores. Logistic regression was then used to examine the effect of a sphincter tear on the 5- and 10-year scores. Analyses were performed unadjusted and adjusted for subsequent deliveries to estimate the effect of these on anal continence. Incontinence in patients with positive scores was examined with linear regression since there were more data points available, and scores both increased and decreased over time. The effect of a sphincter tear on outcome was examined separately for women with and those without postpartum changes in continence, and the relationship between change in incontinence and presence of a tear was estimated. However, there was insufficient data to examine the relationship between a sphincter tear and change in continence at 10 years using this method, and so the effect of tear at 10 years was examined in a separate analysis using Fisher exact test. There were 14 sphincter tears in our study, giving us 80% power at the 5% significance level to detect a change in continence score of 1.4.
Of the 156 women who were recruited initially, 18 delivered subsequently by cesarean, and four had a third-degree tear diagnosed clinically at the time of delivery. These 22 women were excluded, leaving 134 women for analysis, the demographics for whom are shown in Table 1.
None of these 134 women had evidence of anal sphincter trauma on anal endosonography predelivery nor on clinical examination after vaginal delivery. Anal continence had deteriorated compared with baseline in 37 (28%) women by the time of their 3-month assessment after vaginal delivery, 20 of whom (54%) had had an episiotomy cut. Eight (22%) of these 37 women had anal sphincter tears involving the external sphincter detected by anal endosonography. Anal continence did not deteriorate in 97 women, 48 (49%) of whom had had an episiotomy cut. Six (6%) of these 97 women had anal sphincter tears involving the external sphincter detected by anal endosonography. There were no isolated internal anal sphincter tears. Thus, anal sphincter tears were detected in 14 (10%) women overall, and six (43%) of these showed no clinical deterioration in anal continence as a result.
One hundred seven (80%) women could be contacted at both 5 and 10 years, including 11 of the 14 with a sphincter tear. Fifty-nine (55%) of these 107 women had a continence score greater than zero at 10 years, 23 of which had deteriorated since their baseline assessment (Table 2). When all 59 women were analyzed together, there was no evidence that a sphincter tear had a significant effect upon the incontinence scores at 10 years, both unadjusted and adjusted for subsequent deliveries. Similarly, when those whose score had deteriorated immediately after delivery and those whose score had not deteriorated were analyzed separately, there was no evidence that a sphincter tear significantly influenced continence scores at 10 years (Table 2).
The effect of a sphincter tear upon the change in score from baseline to 10 years was examined using linear regression. There was little evidence that a sphincter tear influenced the change in continence score from baseline to 10 years, either when all women were analyzed together or when the analysis was unadjusted and adjusted for subsequent vaginal deliveries (Table 3). However, the data indicated a significant relationship between a sphincter tear and those women whose continence had deteriorated after vaginal delivery, both unadjusted and adjusted for subsequent deliveries. Data adjusted for subsequent vaginal deliveries indicated that the change in continence scores was, on average, 2.8 times higher for those women with a sphincter tear, whose continence had deteriorated by the time of the initial 3-month assessment (Table 3).
The change in continence scores from 5 to 10 years was also examined. The results indicated that there was no effect of a sphincter tear upon the change in score from 5 to 10 years when all patients were examined together, both unadjusted and adjusted for a subsequent delivery in this period. Again, subset analysis revealed a significant relationship between a sphincter tear and those women whose continence deteriorated from baseline after the initial vaginal delivery, for both unadjusted analyses and analyses adjusted for subsequent deliveries: The adjusted results indicated that the change in continence scores from 5 to 10 years was 1.2 units higher for those women with a sphincter tear than for those without (Table 4). There was no effect of a tear on continence for those women whose continence did not deteriorate after the first vaginal delivery.
Clinical diagnosis of anal sphincter disruption after childbirth is uncommon. A prospective study of 2,883 deliveries found 95 tears via clinical examination of the perineum, an incidence of just 3.3%,2 but studies using anal endosonography have revealed a much higher incidence,7,8 suggesting that most anal sphincter injuries go unrecognized. For example, an immediate postpartum study of 150 apparently normal women found sphincter tears in 42 (28%) when endosonography followed clinical examination,15 and the present study identified only four women with a tear detected by clinical examination. Anal endosonography is now widely used to characterize anal sphincter morphology in patients who have symptoms of anal incontinence. Many patients present later in life, and some of these will be women who eventually prove to have an external anal sphincter tear sustained many years previously during vaginal delivery. It is possible that such sphincter tears become symptomatic later in life rather than at the time of delivery due to the cumulative effects of aging and menopause.11 It is also well established that sphincter function may deteriorate with time in the absence of a tear, and in these cases neurological damage is implicated, perhaps due to pudendal neuropathy sustained during pregnancy or childbirth.21,22 A study of 233 women found that sphincter defects could explain incontinence in only 45% at 8 weeks after delivery.23 It is, therefore, difficult to know how to advise a woman who has sustained a sphincter tear after childbirth, but who is currently asymptomatic because the contribution of the tear to future symptoms, if any, is uncertain. Longitudinal studies over time are necessary to answer such questions but are, by necessity, very time consuming.
We have used anal endosonography in a consecutive group of consenting women to estimate the true incidence of anal sphincter trauma after their first vaginal delivery. In doing so, we were able to identify a cohort of women who had tears but were asymptomatic. It was necessary to precisely characterize bowel function predelivery in all participants because community-based studies have established that young women have a small but significant prevalence of anal incontinence,24 and merely documenting symptoms after delivery would overestimate the proportion of women whose symptoms deteriorated as a consequence of delivery.10 Attempting to characterize predelivery bowel habit using postpartum questionnaires alone is unreliable because recollection of bowel habit is inaccurate,24 and studies specifically looking for predelivery anal incontinence have found an incidence as high as 8.6%.25 Supporting these data, we found that 19% of women complained of predelivery anal incontinence, with 11% declaring symptoms that dated from before their pregnancy. Symptoms were generally mild, and our figure may be explained by the sensitive and specific nature of the questionnaire we used. Vigorous characterization of predelivery bowel habit allowed us to confidently identify those women whose continence deteriorated after vaginal delivery, and our use of anal endosonography allowed us to confidently separate these into women with and those without clinically occult sphincter tears.
Overall, we found no evidence that a sphincter tear significantly influenced continence scores at 10 years or influenced the change from baseline scores over the 10-year follow-up period. However, subset analysis did reveal a significant relationship between a sphincter tear and those women whose continence had deteriorated immediately after vaginal delivery. This relationship was present for analyses unadjusted and adjusted for subsequent deliveries and was also apparent when the change in scores from 5 to 10 years was investigated. These data suggest that a sphincter tear combined with symptoms of anal incontinence commencing immediately after vaginal delivery is associated with sustained symptoms over the following decade. In many ways, these data are unsurprising and expected because the association between sphincter tears after vaginal delivery and incontinence is well established. For example, Nichols and colleagues9 investigated women with and those without clinically recognized anal sphincter trauma at 6 weeks postpartum and found a significant association between incontinence and sphincter tears apparent on anal endosonography.
Interested in the long-term consequences of anal sphincter injury, Faltin and colleagues14 investigated symptoms of anal incontinence 18 years after delivery, comparing 259 women who had a third-degree tear detected clinically at the time of delivery with 281 who did not. The authors found that 22 (8%) women in the control group had severe incontinence compared with 34 (13%) of those with a documented tear and concluded, therefore, that only a small proportion of incontinence could be directly attributed to anal sphincter laceration.14 However, recognizing that clinical examination alone will miss many sphincter tears and suspecting that the control group may have contained women with a clinically occult tear, the same authors used anal endosonography immediately after delivery to precisely characterize sphincter integrity and then subsequently assessed symptoms in 150 women via a questionnaire administered 3 months postpartum.15 There were 42 (28%) clinically unsuspected sphincter tears, but while sphincter tears were significantly associated with symptoms of anal incontinence, the authors also identified a group of women who had sustained sphincter laceration but who were asymptomatic; the authors hypothesized that these tears may become symptomatic in the future.15 Similarly, Damon and colleagues,26 while finding that sphincter tears diagnosed by ultrasonography were strongly associated with subsequent incontinence, also found that 50% of women who had a tear were asymptomatic when a questionnaire was administered at 6 years postpartum.
The present study extends these data by similarly using anal endosonography to characterize sphincter integrity precisely, identifying those women whose anal continence deteriorated after vaginal delivery, while also determining symptoms precisely pre- and postpartum and also at 5 and 10 years. Our data suggest that women who sustain a sphincter injury at the time of delivery, but whose continence does not deteriorate immediately postpartum, are not at increased risk of anal incontinence over and above similar women who have sustained no injury. We believe these data have immediate value for day-to-day clinical practice because they afford women who have sustained a clinically silent sphincter tear the reassurance that their continence is unlikely to deteriorate over the following decade, and clinicians can counsel them appropriately. Also, where local practice is to offer a sphincter repair, a more informed decision can be taken vis-à-vis whether such a repair should be performed in a woman with a known defect but who is asymptomatic.
Our study does have limitations. In particular, the study cohort was assessed 10 years postpartum, and so most women were not menopausal. The effects of the menopause on asymptomatic sphincter injuries thus remains to be determined,11 and it remains possible that the women harboring asymptomatic sphincter defects will become symptomatic at some point in the future. Also, a single investigator performed all sonographic examinations, so we were not able to estimate interobserver error for diagnosis of sphincter defects. However, the operator had very extensive clinical and research experience of performing anal endosonography, and inter- and intraobserver agreement for the procedure has previously been shown to be good.27
We considered the primary outcome the change in baseline score over 10 years (because this is more informative than simple occurrence at 10 years). Our data suggest that, given a standard deviation of 2 for the change in score from baseline to 10 years (calculated from the results) and assuming that the group without sphincter tear is four times the size of the tear group (again approximately the case from the data), with the given sample size we can detect a difference of about 1.4 units in the change of score between groups with 80% power at the 5% significance level. We believe that this is adequate since the questionnaire the employed used a 24-point scale.
In summary, we characterized the bowel habits of women, both before and after vaginal delivery, so that we could confidently identify those whose continence deteriorated subsequently, and we used anal endosonography to precisely characterize anal sphincter integrity. Although we found a strong association between anal sphincter defects and women whose continence deteriorated after childbirth, a small proportion of women developed sphincter defects that were asymptomatic. We found that obstetric anal sphincter defects in women initially asymptomatic after childbirth were unassociated with subsequent deterioration in anal continence over the following decade.
1. Donnelly V, Fynes M, Campbell D, Johnson H, O’Connell PR, O’Herlihy C. Obstetric events leading to anal sphincter damage. Obstet Gynecol 1998;92:955–61.
2. Samuelsson E, Ladfors L, Wennerholm UB, Gåreberg B, Nyberg K, Hagberg H. Anal sphincter tears: prospective study of obstetric risk factors. BJOG 2000;107:926–31.
3. MacArthur C, Bick DE, Keighley MR. Faecal incontinence after childbirth. Br J Obstet Gynaecol 1997;104:46–50.
4. Meyer S, Schreyer A, De Grandi P, Hohlfeld P. The effects of birth on urinary continence mechanisms and other pelvic floor characteristics. Obstet Gynecol 1998;92:613–8.
5. Zetterstrom JP, Lopez A, Anzen B, Dolk A, Norman M, Mellgren A. Anal incontinence after vaginal delivery: a prospective study in primiparous women. Br J Obstet Gynaecol 1999;106:324–30.
6. Groutz A, Fait G, Lessing JB, David MP, Wolman I, Jaffa A, et al. Incidence and obstetric risk factors of postpartum anal incontinence. Scand J Gastroenterol 1999;34:315–8.
7. Fines M, Donnelly V, Behan M, O’Connell O’Herlihy C. Effect of second vaginal delivery on anorectal physiology and faecal continence: a prospective study. Lancet 1999;354:983–6.
8. Sultan AH, Kamm MA, Hudson CN, Thomas JM, Bartram CI. Anal sphincter disruption during vaginal delivery. N Engl J Med 1993;329:1905–11.
9. Nichols CM, Nam M, Ramakrishnan V, Lamb EH, Currie N. Anal sphincter defects and bowel symptoms in women with and without recognized anal sphincter trauma. Am J Obstet Gynaecol 2006;194:1450–4.
10. Frudinger A, Halligan S, Bartram CI, Spencer J, Kamm MA, Winter R. Assessment of the predictive value of a bowel symptom questionnaire in identifying perianal and anal sphincter trauma after vaginal delivery. Dis Colon Rectum 2003;46:742–7.
11. Donnelly V, O’Connell PR, O’Herlihy C. The influence of oestrogen replacement on faecal incontinence in postmenopausal women. Br J Obstet Gynaecol 1997;104:311–5.
12. Fornell EU, Matthiesen L, Sjodahl R, Berg G. Obstetric anal sphincter injury ten years after: subjective and objective long term effects. BJOG 2005;112:312–6.
13. Pollack J, Nordenstam J, Brismar S, Lopez A, Altman D, Zetterstrom J. Anal incontinence after vaginal delivery: a five-year prospective cohort study. Obstet Gynaecol 2004;104:1397–1402.
14. Faltin DL, Otero M, Petignat P, Sangalli MR, Floris LA, Boulvain M, Irion O. Women’s health 18 years after rupture of the anal sphincter during childbirth: I. Fecal incontinence. Am J Obstet Gynaecol 2006;194:1255–9.
15. Faltin DL, Boulvain M, Irion O, Bretones S, Stan C, Weil A. Diagnosis of anal sphincter tears by postpartum endosonography to predict fecal incontinence. Obstet Gynecol 2000;95:643–7.
16. Jorge JM, Wexner SD. Etiology and management of faecal incontinence. Dis Colon Rectum 1993;36:77–97.
17. Frudinger A, Bartram CI, Halligan S, Kamm MA. Examination techniques for anal endosonography of the anal canal. Abdom Imaging 1998;23:301–3.
18. Frudinger A, Halligan S, Bartram CI, Price AB, Kamm MA, Winter R. Female anal sphincter: age-related differences in asymptomatic volunteers with high frequency endoanal US. Radiology 2002;224:417–23.
19. Williams AB, Bartram CI, Halligan S, Spencer JA, Nicholls RJ, Kmiot WA. Anal sphincter damage after vaginal delivery using three-dimensional endosonography. Obstet Gynecol 2001;97:770–5.
20. Sultan AH, Kamm MA, Hudson CN, Bartram CI. Effect of pregnancy on anal sphincter morphology and function. Int J Colorect Dis 1993;8:206–9.
21. Vaizey CJ, Kamm MA, Bartram CI. Primary degeneration of the internal anal sphincter as a cause of passive faecal incontinence. Lancet 1997;349:612–5.
22. Briel JW, Stoker J, Rociu E, Lameris JS, Hop WC, Schouten WR. External anal sphincter atrophy on endoanal magnetic resonance imaging adversely affects continence after sphincteroplasty. Br J Surg 1999;86:1322–7.
23. Abramowitz L, Sobhani I, Ganansia R, Vuagnat A, Benifla JL, Darai E, et al. Are sphincter defects the cause of anal incontinence after vaginal delivery? Results of a prospective study. Dis Colon Rectum 2000;43:590–8.
24. Drossman DA, Li Z, Andruzzi E, Temple RD, Talley NJ, Thompson WG, et al. U.S. householder survey of functional gastrointestinal disorders. Prevalence, sociodemography and health impact. Dig Dis Sci 1993;38:1569–80.
25. Hojberg KE, Salvig JD, Winslow NA, Bek KM, Laurberg S, Secher NJ. Flatus and faecal incontinence: prevalence and risk factors at 16 weeks of gestation. BJOG 2000;107:1097–103.
26. Damon H, Bretones S, Henry L, Mellier G, Mion F. Long-term consequences of first vaginal delivery-induced anal sphincter defect. Dis Colon Rectum 2005;48:1772–6.
27. Gold DM, Halligan S, Kmiot WA, Bartram CI. Intraobserver and interobserver agreement in anal endosonography. Br J Surg 1999;86:371–5.