Does a second delivery increase the risk of anal incontinence?

British Journal of Obstetrics and Gynaecology July 2001, Vol. 108, pp. 684±688

Does a second delivery increase the risk of anal incontinence? Daniel L. Faltin a,*, Michel R. Sangalli a,c, Bruno Roche b, Lucia Floris a, Michel Boulvain a, Antoine Weil a Objective To evaluate the prevalence of anal incontinence and anal sphincter defects after a ®rst vaginal delivery and assess the effect of a second delivery. Design Prospective cohort study using postal questionnaires assessing incontinence to ¯atus and stools at three and thirty months postnatally and anal endosonography at three months following delivery. Setting Recruitment was from the antenatal clinic at the University Hospitals of Geneva, Switzerland. Population One hundred women with a vaginal delivery of their ®rst child. Main outcome measures Prevalence of anal incontinence and anal sphincter defects. Results Anal incontinence was reported by 16/92 (17%) of women at three months after delivery and by 11/77 (14%) at 30 months. At that time, 5/54 (9%) with no further delivery reported incontinence, compared with 6/ 23 (26%) of those who had had another delivery (RR 2.8, 95% CI 1.0-8.3). Anal sphincter defects were diagnosed by endosonography in 46/87 (53%) women and were associated with reported incontinence at both three months (RR 1.9; 95% CI 1.4-2.6) and 30 months (RR 1.9; 95% CI 1.3-2.8) after delivery. The prevalence of anal incontinence at 30 months was highest (5/13, 39%) among those in whom a sphincter defect was diagnosed by endosonography after their ®rst delivery and with a second delivery. Conclusion Anal incontinence after childbirth is associated with defects of the anal sphincter diagnosed by endosonography. Subsequent deliveries increase the risk of incontinence, particularly among women with a sphincter defect diagnosed after the ®rst delivery.

INTRODUCTION Anal incontinence, the involuntary loss of ¯atus or faeces, is reported by up to 40% of women after vaginal delivery 1,2. Tears of the anal sphincter are diagnosed clinically in as many as 24% of women at the time of delivery and are associated with subsequent anal incontinence 3±6. Postpartum sonographic studies of the anal sphincter have shown that clinically occult anal sphincter tears are frequent with an actual prevalence of sphincter defects of between 35% and 41% 7±12. Clinically undiagnosed tears are associated with subsequent anal incontinence in up to 50% of affected women 7±12. The natural history of these clinically undiagnosed anal sphincter tears remains unknown due to short term follow up. The above-mentioned studies, with a follow up of less than six months, suggest that improvement in symptoms can occur some time after injury. However, an increased prevalence of anal incontinence and sphinc-

a

Department of Gynaecology and Obstetrics, University Hospitals of Geneva, Switzerland b Department of Surgery, University Hospitals of Geneva, Switzerland c Fetal Medicine Unit, Department of Obstetrics and Gynaecology, The Canberra Hospital, Australia * Correspondence: Dr D. L. Faltin, Department of Gynaecology and Obstetrics, University Hospitals of Geneva, 1211 Geneva 14, Switzerland. q RCOG 2001 British Journal of Obstetrics and Gynaecology PII: S03 06-5456(00)0018 5-6

ter defects has been reported in multiparous women 8,13. In a prospective study of 59 women after two deliveries, the risk of anal incontinence was higher in those with symptoms or an anal sphincter defect diagnosed by endosonography after the ®rst delivery 14. No data were provided on the outcome of women with only one delivery. Our objective was to evaluate the prevalence of anal incontinence and anal sphincter defects after a ®rst vaginal delivery and to assess the effect of a second delivery. METHODS We recruited 107 women attending the antenatal clinic of our university hospital maternity between May 1995 and April 1996. Eligible women were nulliparous, over 37 weeks of gestation, with no symptoms of anal incontinence, and had no history of previous anorectal surgery nor known in¯ammatory bowel disease. Deliveries were performed by midwives and residents under the supervision of senior registrars. A policy of restrictive episiotomy use was applied. When performed, episiotomies were mostly median, while mediolateral episiotomies were usually performed for operative vaginal deliveries. Forceps or vacuum delivery was performed according to the operator's preferences. Women who were delivered by caesarean section were not included in the study as the probability of anal sphincter rupture and anal incontinence is low in this group 8,12. Obstetric characteristics of the participants were compared with those from eligiwww.bjog-elsevier.com

SECOND DELIVERY INCREASE AND ANAL INCONTINENCE 685

ble women not participating in the study. Baseline data of non-participants were extracted from the computerised hospital database. Three months after delivery, participants were sent a questionnaire to evaluate anal incontinence. If no response was obtained after one month, a reminder letter was sent. Non-respondents were contacted one month later by telephone to encourage them to return the form. Women were asked about any incontinence to ¯atus, liquid or solid stools since the delivery and if these symptoms were still present. A positive response to any of the questions with at least monthly events was considered as indicating anal incontinence. These questions have been previously used in patients treated for anal incontinence and found to be reliable, valid and sensitive to change 15,16. Copies of the questionnaire are available from the authors. All participants were given an appointment three months after delivery with a coloproctologist, who evaluated the tone of the anal sphincter by digital examination and performed an anal endosonography (Bruel & Kjaer, Nñrum, Denmark; system 3535 endosonic probe Type 1850 set at 10 MHz). Anal sphincter defect was de®ned as a gap in the hyperechogenic ring of the external anal sphincter or in the hypoechogenic ring of the internal anal sphincter 17±20. Two years later, participants were sent a second questionnaire. At that time, we also assessed the effect of anal incontinence on quality of life by the Short Form Incontinence Impact Questionnaire which evaluates the effect of incontinence using the following seven items: household chores; physical recreation; entertainment activities; travelling more than 30 minutes from home; social activities; emotions (nervousness, depression); and feeling frustrated 21. Each item was scored on a 10 cm long visual analogue scale, anchored with the words ªnot at allº and ªvery muchº. All scores were pooled to obtain a mean impact. Although primarily designed for urinary incontinence, this questionnaire has good face validity to measure the impact of anal incontinence. Women were also asked if they had had subsequent deliveries, and, if this was the case, we collected the obstetric data. Differences between groups were assessed by unpaired t test for continuous variables and by Fisher's exact test for categorical variables. A strati®ed analysis was carried out to evaluate the effect of a second delivery on anal continence according to the presence or absence of anal sphincter defect after the ®rst delivery. Relative risks (RR) and their 95% con®dence interval (95% CI) were calculated. A sample size of 97 women was required to obtain a precision of 10% in the estimate of the prevalence of anal sphincter defects observed by endosonography. The statistical analysis was performed with SPSS (Statistical Package for Social Science, Chicago, Illinois, USA) and Epi Info (Centers for Disease Control and Prevention, Atlanta, Georgia, USA). q RCOG 2001 Br J Obstet Gynaecol 108, pp. 684±688

The study protocol was approved by the institutional ethics committee. All participants gave written informed consent. RESULTS Of 107 women recruited, seven had a caesarean section and were excluded from follow up. Among the remaining participating 100 women, 66 had epidural analgesia, 58 assisted vaginal deliveries (36 forceps; 18 vacuum; 4 forceps after failed vacuum extraction). Fourteen women had an anal sphincter tear (third- or fourth-degree perineal tear) diagnosed clinically and immediately sutured. The characteristics of participating and eligible women are reported in Table 1. Eighty-seven participants had an endoanal ultrasound three months after delivery. Anal sphincter defects were diagnosed in 46 women (53%; 95% CI 42-64), of which 10 (21.7%) involved only the external sphincter, 11 (23.9%) only the internal sphincter, and 25 (54.3%) both the internal and external sphincters. Three months after delivery, 92 women returned the ®rst questionnaire. Anal incontinence during the three months following delivery was reported by 24 (26%; 95% CI 17-36%). All these women had had episodes of ¯atus incontinence, 18 reported daily occurrences. Five women reported having had incontinence to faeces, one with daily occurrences. Three months after delivery, persistent incontinence to ¯atus was reported by 16 women (17%; 95% CI 10-27), ®ve of whom had daily symptoms. Three women (3.3%) were incontinent to liquid or solid stools, one daily. Seventy-seven women returned the second questionnaire after a median interval of 30 months (25 th centile 25 months; 75 th centile 33 months) since the ®rst delivery. Eleven women (14%; 95% CI 7-24) reported anal incontinence. Nine women were incontinent to ¯atus, two of them reporting daily episodes. Six women (8%; 95% CI 3-16) were incontinent to liquid or solid stools, none of them reporting daily episodes. One woman underwent Table 1. Characteristics of participating and eligible women. Means of continuous variables were compared by unpaired t-test and proportions of categorical variables with Fisher's exact test. Values are given as n (%) or mean [SD]. Participants n ˆ 100 Maternal age at ®rst delivery (years) Birthweight (g) Epidural analgesia Vacuum or forceps Third or fourth degree perineal tear

Non-participants n ˆ 1020

29.8 [4.2]

29.4 [6.6]

3413 [377] 66 (66.0) 58 (58.0) 14 (14.1)

3312 [512] 492 (48.2) 419 (41.1) 90 (8.8)

P 0.55 0.06 ,0.001 0.001 0.13

686 D. L. FALTIN ET AL. Table 2. Maternal characteristics and obstetric variables at the ®rst delivery among continent and incontinent women. Means of continuous variables were compared by unpaired t-test and proportions of categorical variables with Fisher's exact test. Values are given as n (%) or mean [SD]. Anal incontinence 3 months after delivery

Age at ®rst delivery (years) Birthweight (g) Epidural analgesia Vacuum or forceps Intact perineum Episiotomy Third or fourth-degree perineal tear

30 months after delivery

Incontinent (n ˆ 16)

Continent (n ˆ 76)

P

Incontinent (n ˆ 11)

Continent (n ˆ 66)

P

30.7 [5.4] 3493 [336] 11 (68.8) 12 (75.0) 0 16 (100) 7 (43.8)

29.7 [3.9] 3418 [361] 53 (69.7) 42 (55.3) 6 (7.9) 53 (69.7) 7 (9.2)

0.49 0.43 1.00 0.17 0.59 0.009 0.002

32.0 [4.3] 3384 [335] 8 (72.7) 9 (81.8) 0 46 (69.7) 4 (36.4)

29.8 [4.3] 3462 [360] 47 (71.2) 39 (59.1) 5 (7.6) 11 (100) 7 (10.6)

0.14 0.49 1.00 0.19 1.00 0.057 0.045

surgical repair of the anal sphincter after the three months' assessment, but remained incontinent to ¯atus. Among the 11 women reporting anal incontinence, the mean score on the visual analogue scale of the Short Form Incontinence Impact Questionnaire was 2.0 cm (standard deviation 1.7cm, range 0-6.1 cm) and nine of these women (82%; 95% CI 48-98) placed a mark beyond 1.0 cm for at least one item of the questionnaire. Twentytwo women had had a second delivery and one a third, with a median time interval of 9 months (25 th centile 3 months; 75 th centile 13 months) before responding to the second questionnaire. All were spontaneous vaginal deliveries. Two women had had an episiotomy and none a clinically diagnosed sphincter tear. The proportion of anal incontinence was 5/54 (9%; 95% CI 3-20) among women who had no subsequent delivery and 6/23 (26%; 95% CI 10-48) among those who delivered again in the interval (RR 2.8; 95% CI 1.0-8.3). Clinically diagnosed anal sphincter tears were associated with anal incontinence at three months (RR 4.3; 95% CI 1.9-9.7) and 30 months after delivery (RR 3.4; 95% CI 1.2-9.8). The association between delivery variables and subsequent incontinence is reported in Table 2. Sphincter defects diagnosed after endosonography three months after delivery were associated with anal incontinence whether transient shortly after delivery (RR 2.0; 95% CI 1.4-2.8), three months after delivery (RR 1.9; 95% CI 1.4-2.6), or thirty months after delivery (RR 1.9; 95% CI 1.3-2.8). Among the 12 women having

sustained a clinically diagnosed anal sphincter tear studied by endosonography, 10 (83%; 95% CI 52-98) had a persisting anal sphincter defect. The main predictor of incontinence was a defect of the external anal sphincter. The effect of defects diagnosed by endosonography on anal incontinence 3 months and 30 months after delivery is shown in Table 3. After restricting the analysis to the women with no clinical diagnosis of anal sphincter rupture after delivery, we found that 36/75 (48%, 95% CI 36-60) had an anal sphincter defect on endosonography (i.e. a clinically occult anal sphincter tear). Of these, 7/36 (19%, 95% CI 8-36) reported anal incontinence three months after delivery and 6/29 (21%, 95% CI 8-40) 30 months after delivery. When the anal sphincter was found intact both clinically and by endosonography, 2/39 (5%, 95% CI 117) women reported anal incontinence 3 months after delivery and 1/37 (3%, 95% CI 0-14) 30 months after delivery. Clinically occult anal sphincter tears were associated with anal incontinence three months after delivery (RR 3.8, 95% CI 0.8-17.1) and 30 months after delivery (RR 7.7, 95% CI 1-60.1). The prevalence of anal incontinence 30 months after delivery was highest (5/13, 39%, 95% CI 14-68) in women with an anal sphincter defect diagnosed by endosonography after their ®rst delivery and who delivered again in the interval. For those with an anal sphincter defect diagnosed by endosonography after their ®rst delivery, the relative risk of anal incontinence with a second

Table 3. Anal incontinence and anal sphincter defects by endosonography after the ®rst delivery. ASD ˆ anal sphincter defect. 3 months after delivery

Intact anal sphincters Isolated internal ASD Isolated external ASD External & internal ASD

30 months after delivery

[n/n] (%)

Relative risk (95%CI)

[n/n] (%)

RR (95%CI)

2/41 (5) 1/11 (9) 3/10 (30) 9/25 (36)

1 (Reference) 1.9 (0.2-18.7) 6.2 (1.2-32.0) 7.4 (1.7-31.4)

2/39 (5) 0/7 (0) 3/9 (33) 6/22 (27)

1 (Reference) ± 6.5 (1.3-33.4) 5.3 (1.2-24.1)

q RCOG 2001 Br J Obstet Gynaecol 108, pp. 684±688

SECOND DELIVERY INCREASE AND ANAL INCONTINENCE 687 Table 4. Interaction of anal sphincter defects (ASD) by endosonography and a second delivery as risk factors for anal incontinence. ASD

2nd delivery

No No Yes Yes

No Yes No Yes

Incontinent n/n (%)

RR (95% CI)

1/29 (3) 1/10 (10) 4/25 (16) 5/13 (38)

1 (Reference) 2.9 (0.2-42.2) 4.6 (0.6-38.9) 11.2 (1.4-86.2)

delivery was 2.4 (95% CI 0.8-7.5) when compared with those women without a subsequent delivery. Table 4 shows the interaction between a second delivery and an anal sphincter defect diagnosed by endosonography after the ®rst delivery on the risk of anal incontinence. DISCUSSION We have found a high proportion of women reporting anal incontinence after childbirth, in agreement with previously published studies 1±14. Defects of the anal sphincter diagnosed by endoanal ultrasound were associated with anal incontinence, particularly those involving the external anal sphincter. Thirty months after delivery, the proportion of women reporting incontinence remained high, especially when a sphincter defect was diagnosed after the ®rst delivery. Most women reporting incontinence 30 months after delivery reported an impact of the symptoms on their daily activities and mood. One woman had already elected to have a surgical repair of the anal sphincter. A second vaginal delivery appeared to be an independent risk factor for anal incontinence. Although we made every effort to recruit women with no previous anorectal disease, the possibility of a bias in the recruitment can be raised, especially as concerns the differences between participating and non-participating women in the proportion of epidural analgesia and assisted vaginal deliveries. Possible factors biasing the recruitment were language and socio-economic status. This last factor could also explain the high proportion of epidural analgesia, which is associated with a higher proportion of assisted vaginal deliveries and the resulting increased risk of anal sphincter tear and incontinence. Women were recruited before their ®rst delivery, and we believe that a bias in the recruitment could in¯uence the prevalence of symptoms but not confound signi®cantly the association between a tear of the anal sphincter after the ®rst delivery and incontinence after the next delivery. The proportion of women reporting incontinence decreased over time among women who had no further delivery. In those who delivered again, the interval between the questionnaire and the last delivery was shorter. The greater prevalence of incontinence among multiparous women could re¯ect the short interval since the last delivery rather than additional damage q RCOG 2001 Br J Obstet Gynaecol 108, pp. 684±688

sustained during the second delivery. However, transient incontinence after delivery is a risk factor for later incontinence and therefore a marker of underlying damage to the continence mechanism 14. This is con®rmed by the strong association with anal sphincter defects. Multiparity is associated with an increased prevalence of anal incontinence 8,13. Several mechanisms could explain this association. An anal sphincter tear during the second delivery can damage a previously intact sphincter or increase the severity of a previous tear. Women sustaining a clinically overt anal sphincter tear (third- or fourth-degree perineal laceration) during their ®rst delivery are at risk of repeated damage during a second vaginal delivery and of subsequent anal incontinence 22±24. The second delivery could also repeat the stretching produced on other structures participating in the continence mechanism, such as the levator ani muscles and the pudendal nerve. A prolonged pudendal nerve terminal motor latency time has been demonstrated after a second delivery, and nerve conduction alterations have also been shown to worsen with time 14,25. Therefore, women with an asymptomatic anal sphincter tear after the ®rst delivery could become incontinent after the second delivery due to neurological damage. Long term studies may provide more information on this issue and on the cumulative effect of ageing and menopause. The assessment and repair of anal sphincter damage after childbirth should be improved. Knowledge of the anatomy of the perineum has been found to be insuf®cient among midwives and trainee doctors 26. Further studies are required to evaluate whether primary suture of clinically occult sphincter tears diagnosed by ultrasound will improve functional outcomes 27. Alternative techniques for surgical repair of anal sphincter by overlap repair are being currently evaluated 28. Caesarean delivery has been advocated for women at high risk of severe perineal damage 29,30. Symptomatic women may bene®t from physiotherapy and surgery followed by caesarean section for their subsequent deliveries. However, the management of asymptomatic women with an anal sphincter tear diagnosed by endosonography is controversial. Some women with extensive sphincter defects or low sphincter pressures on manometry appear to have a particularly high risk of developing anal incontinence after a second delivery 14. The bene®t of preventive strategies should be assessed. CONCLUSION Anal incontinence after childbirth is associated with defects of the anal sphincter diagnosed by endosonography. Although the overall prevalence of anal incontinence decreases with time, a signi®cant number of women remain affected. A subsequent vaginal delivery increases the risk of anal incontinence, particularly

688 D. L. FALTIN ET AL.

among those with an anal sphincter defect diagnosed after the ®rst delivery. Acknowledgements The authors would like to thank Mrs J. Ruche for secretarial help and Mrs R. Sudan for editorial assistance. This project was supported by the Swiss National Science Foundation, Grant No. 32-55907.98. References 1. MacArthur C, Bick DE, Keighley MR. Faecal incontinence after childbirth. Br J Obstet Gynaecol 1997;104:46±50. 2. Fornell EK, Berg G, Hallbook O, Matthiesen LS, Sjodahl R. Clinical consequences of anal sphincter rupture during vaginal delivery. J Am Coll Surg 1996;183:553±558. 3. Thacker SB, Banta HD. Bene®ts and risks of episiotomy: an interpretative review of the English language literature, 1860-1980. Obstet Gynecol Surv 1983;38:322±338. 4. Crawford LA, Quint EH, Pearl ML, DeLancey JO. Incontinence following rupture of the anal sphincter during delivery. Obstet Gynecol 1993;82:527±531. 5. Sorensen M, Tetzschner T, Rasmussen OO, Bjarnesen J, Christiansen J. Sphincter rupture in childbirth. Br J Surg 1993;80:392±394. 6. Haadem K, Ohrlander S, Lingman G. Long-term ailments due to anal sphincter rupture caused by delivery: a hidden problem. Eur J Obstet Gynecol Reprod Biol 1988;27:27±32. 7. Burnett SJ, Spence-Jones C, Speakman CT, Kamm MA, Hudson CN, Bartram CI. Unsuspected sphincter damage following childbirth revealed by anal endosonography. Br J Radiol 1991;64:225±227. 8. Sultan AH, Kamm MA, Hudson CN, Thomas JM, Bartram CI. Analsphincter disruption during vaginal delivery. N Engl J Med 1993;329:1905±1911. 9. Campbell DM, Behan M, Donnelly VS, O'Herlihy C, O'Connell PR. Endosonographic assessment of postpartum anal sphincter injury using a 120 degree sector scanner. Clin Radiol 1996;51:559±561. 10. Sandridge DA, Thorp Jr JM, Roddenberry P, Kuller J, Wild J. Vaginal delivery is associated with occult disruption of the anal sphincter mechanism. Am J Perinatol 1997;14:527±533. 11. Rieger N, Schloithe A, Saccone G, Wattchow D. A prospective study of anal sphincter injury due to childbirth. Scand J Gastroenterol 1998;33:950±955. 12. Zetterstrom J, Mellgren A, Jensen LL, et al. Effect of delivery on anal sphincter morphology and function. Dis Colon Rectum 1999;42:1253± 1260. 13. Ryhammer AM, Bek KM, Laurberg S. Multiple vaginal deliveries increase the risk of permanent incontinence of ¯atus urine in normal premenopausal women. Dis Colon Rectum 1995;38:1206±1209.

14. Fynes M, Donnelly V, Behan M, O'Connell PR, O'Herlihy C. Effect of second vaginal delivery on anorectal physiology and faecal continence: a prospective study. Lancet 1999;354:983±986. 15. Sangalli MR, Marti MC. Results of sphincter repair in postobstetric fecal incontinence. J Am Coll Surg 1994;179:583±586. 16. Osterberg A, Graf W, Karlbom U, Pahlman L. Evaluation of a questionnaire in the assessment of patients with faecal incontinence and constipation. Scand J Gastroenterol 1996;31:575±580. 17. Sultan AH, Nicholls RJ, Kamm MA, Hudson CN, Beynon J, Bartram CI. Anal endosonography and correlation with in vitro and in vivo anatomy. Br J Surg 1993;80:508±511. 18. Deen KI, Kumar D, Williams JG, Olliff J, Keighley MR. Anal sphincter defects. Correlation between endoanal ultrasound and surgery. Ann Surg 1993;218:201±205. 19. Sultan AH, Kamm MA, Talbot IC, Nicholls RJ, Bartram CI. Anal endosonography for identifying external sphincter defects con®rmed histologically. Br J Surg 1994;81:463±465. 20. Sentovich SM, Blatchford GJ, Rivela LJ, Lin K, Thorson AG, Christensen MA. Diagnosing anal sphincter injury with transanal ultrasound and manometry. Dis Colon Rectum 1997;40:1430±1434. 21. Uebersax JS, Wyman JF, Shumaker SA, McClish DK, Fantl JA. Short forms to assess life quality and symptom distress for urinary incontinence in women: the Incontinence Impact Questionnaire and the Urogenital Distress Inventory. Continence Program for Women Research Group. Neurourol Urodyn 1995;14:131±139. 22. Bek KM, Laurberg S. Risks of anal incontinence from subsequent vaginal delivery after a complete obstetric anal sphincter tear. Br J Obstet Gynaecol 1992;99:724±726. 23. Peleg D, Kennedy CM, Merrill D, Zlatnik FJ. Risk of repetition of a severe perineal laceration. Obstet Gynecol 1999;93:1021±1024. 24. Sangalli MR, Floris L, Faltin D, Weil A. Anal incontinence in women with third or fourth degree perineal tears and subsequent vaginal deliveries. Aust N Z J Obstet Gynaecol 2000;40:244±248. 25. Snooks SJ, Swash M, Mathers SE, Henry MM. Effect of vaginal delivery on the pelvic ¯oor: a 5-year follow-up. Br J Surg 1990;77:1358± 1360. 26. Sultan AH, Kamm MA, Hudson CN. Obstetric perineal tears: an audit of training. J Obstet Gynecol 1995;15:19±23. 27. 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±764. 28. Sultan AH, Monga AK, Kumar D, Stanton SL. Primary repair of obstetric anal sphincter rupture using the overlap technique. Br J Obstet Gynaecol 1999;106:318±323. 29. Sultan AH, Stanton SL. Preserving the pelvic ¯oor and perineum during childbirth±elective caesarean section? Br J Obstet Gynaecol 1996;103:731±734. 30. Handa VL, Harris TA, Ostergard DR. Protecting the pelvic ¯oor: obstetric management to prevent incontinence and pelvic organ prolapse. Obstet Gynecol 1996;88:470±478. Accepted 22 February 2001

q RCOG 2001 Br J Obstet Gynaecol 108, pp. 684±688