Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Objective: To determine risk factors for obstetric anal with endoanal sonography.1,4 Additionally, anal incon-
sphincter tears and to evaluate symptomatic outcome of tinence is seldom spontaneously mentioned by patients,
primary repair. and therefore there is a risk that these problems remain
Methods: Obstetric-procedure, maternal, and fetal data
undetected.5
were registered in 845 consecutive vaginally delivered
The incidence of clinically detected anal sphincter
women. Risk factors for anal sphincter tears were calculated
by multiple logistic regression. All 808 Swedish-speaking
tears at delivery most often is reported to be less than
women who delivered vaginally were included in a ques- 3%.6 –10 During recent years, the incidence of tears has
tionnaire study regarding anal incontinence in relation to increased in Sweden.11 The reason for this is not clear,
the delivery. Questionnaires were distributed within the but it has been thought that altered obstetric routines,
first few days postpartum, and at 5 and 9 months postpar- such as upright maternal delivery positions, could be
tum. one contributing factor. At our institution, the fre-
Results: Six percent of the women had a clinically detected quency of upright delivery positions has become in-
sphincter tear at delivery. Sphincter tears were associated creasingly popular, and today approximately 60% of all
with nulliparity (odds ratio [OR] 9.8, 95% confidence interval deliveries are performed in these positions. Gardeberg
[CI] 3.6, 26.2), postmaturity (OR 2.5, 95% CI 1.0, 6.2), fundal
et al12 have reported a seven-fold increased risk for
pressure (OR 4.6 95% CI 2.3, 7.9), midline episiotomy (OR 5.5
sphincter tears in upright delivery positions without
95% CI 1.4, 18.7), and fetal weight in intervals of 250 g (OR 1.3
95% CI 1.1, 1.6). Fifty-four percent of women with repaired
support of the pelvic floor.
sphincter tears suffered from fecal or gas incontinence or The aims of the present study were to identify risk
both at 5 months and 41% at 9 months. Most of the symp- factors for obstetric anal sphincter tears, to evaluate the
toms were infrequent and mild. association between sphincter tears and upright deliv-
Conclusion: Several risk factors for sphincter tear were ery positions, and to study the symptomatic outcome of
identified. Sphincter tear at vaginal delivery is a serious primary sphincter repair.
complication, and it is frequently associated with anal in-
continence. Special attention should be directed toward risk
factors for this complication. Symptoms of anal incontinence
should explicitly be sought at follow-up after delivery. Materials and Methods
(Obstet Gynecol 1999;94:21– 8. © 1999 by The American
During a 10-week period, April 1 through June 9, 1995,
College of Obstetricians and Gynecologists.)
at Danderyd Hospital, all deliveries were studied and
analyzed with respect to risk factors for development of
perineal and anal sphincter tears. During the same
Anal incontinence after childbirth may be due to injury
period all Swedish-speaking, vaginally delivered
to the anal sphincter or its innervation, or both.1–3
women were asked to participate in a prospective
Recent studies have demonstrated a significant inci-
questionnaire study regarding incontinence symptoms
dence of sphincter injuries after delivery, and the ma-
that could be linked to childbirth.
jority of these injuries are occult and only detectable
Deliveries in the present study were performed ac-
cording to Swedish obstetric routines in which most
From the Divisions of Obstetrics & Gynaecology and Surgery, uncomplicated deliveries are handled by midwives.
Karolinska Institutet, Danderyd Hospital, Stockholm, Sweden. Obstetricians were called upon when necessary to as-
Gestational age (d) 286 6 8 281 6 10 Fetal weight (g) 3740 6 413 3610 6 791
Cx opening on admission (cm) 462 463 Fetal length (cm) 51 6 2 50 6 2
Duration of labor (h) 10.8 6 5.3 7.6 6 4.8 Fetal head circumference (cm) 35 6 1 35 6 1
(n 5 42) (n 5 480) Fetal presentation
Duration of first stage of labor (h) 9.2 6 4.8 6.6 6 4.4 Vertex occipito-anterior 54 (100) 774 (98)
(n 5 42) (n 5 480) Vertex occipito-posterior 0 12 (1.5)
Duration of second stage (h) 1.5 6 1.3 0.9 6 0.9 Breech 0 5 (0.5)
(n 5 781)
Data are expressed as mean 6 standard deviation or n (%).
Delivery positions
Kneeling 17 (31) 441 (56)
Standing 3 (6) 61 (8)
Squatting 0 2 (0.2) than 12 hours was defined as pathologic, and the data
Sitting 25 (46) 220 (28) were dichotomized at this limit. Duration of second
Lithotomy 7 (13) 25 (3) stage of labor more than 1 hour was defined as patho-
Lateral 2 (4) 42 (5) logic, and the data were dichotomized at this limit. Fetal
Upright position* 20 (37) 504 (64)
weight was treated as a continuous variable, presented
Induction of labor 3 (6) 66 (8)
Prostaglandin 2 (4) 47 in changes of 250 g.
Oxytocin 1 (2) 15 The association between interventions (one or more
Amniotomy 0 4 of fundal pressure, instrumental delivery, and episiot-
Augmentation of labor omy) and delivery positions was tested in a x2 analysis.
,30 mL/h 4 (7) 39
McNemars test was used when analyzing changes of
30 – 60 mL/h 9 (17) 125
.60 mL/h 28 (52) 28 symptoms over time. The frequency of anal inconti-
Instrumental delivery 12 (22) 33 nence at 9 months after delivery was compared with the
Vacuum extractor 12 31 frequency before delivery and at 5 months after deliv-
Forceps 0 2 ery.
External fundal pressure 30 (57) 115 (15)
The study was approved by the Local Ethics Com-
(n 5 53)
Episiotomy 6 (11) 9 (1) mittee at Karolinska Hospital, Karolinska Institutet,
Midline 6 7 Stockholm. All subjects were informed by oral and
Mediolateral 0 2 written sources and gave their consent by filling out the
Analgesia questionnaires.
N2O 46 (85) 580 (74)
Epidural 19 (35) 144 (18)
Paracervical block 8 (15) 93 (12) Results
Pudendal block 2 (4) 17 (2)
Morphine 0 9 (1) Clinically diagnosed anal sphincter tears (third- or
TNS 9 (17) 87 (11) fourth-degree tears) were registered in 6% (54 of 845) of
Acupuncture 2 (4) 6 (1)
the women. Of the third-degree tears, 18% (nine of 54)
Cx 5 cervix; N2O 5 nitrous oxide; TNS 5 transcutaneous nerve involved the whole muscle and 82% (42 of 54) partial
stimulation.
Data are expressed as mean 6 standard deviation or n (%). tears of the muscle.
* Upright position was defined as one of kneeling, standing, or All clinically diagnosed sphincter tears were repaired
squatting delivery positions. primarily. Eighty-five percent (46 of 54) of these repairs
were done in the delivery room, and 6% (three of 54) in
the operation theater; in 9% (five of 54) it was unclear in
active, walk and be in an upright position, but this was which of these locations the repair was done.
not documented in our protocol. Fetal characteristics All repairs were performed with absorbably sutures
are found in Table 3. (polyglycolic acid or polygalactin). In 80% (43 of 54) the
Maternal, fetal, and obstetric risk factors for sphincter sphincter was repaired with figure-of-eight sutures, in
tears were analyzed using logistic regression in a uni- 2% (one of 54) interrupted sutures were used, and in the
variate model.14 Using a 5% significance level, variables remaining women the sutured method was unclear (not
were analyzed by using multivariate logistic regression specified). Antibiotics were given to 4% (two of 54) of
in a forward stepwise procedure.14 the women postoperatively, and 78% (42 of 54) received
Gestational age at delivery was dichotomized at the laxatives for 3–5 days postoperatively.
postmaturity limit (294 days). Duration of labor more In the univariate analysis, anal sphincter tears were
VOL. 94, NO. 1, JULY 1999 Zetterström et al Obstetric Anal Sphincter Tears 23
Table 4. Factors Associated With Anal Sphincter Tears at Table 5. Preexisting Anal Incontinence in Relation to
Delivery Vaginal Parity
Univariate Multivariate Preceding vaginal Frequency of anal incontinence
analysis OR analysis OR parity before pregnancy
Risk factor (95% CI) (95% CI)
0 para 7% (22/301)
Nulliparity 12.4 (4.9, 31.5) 9.8 (3.6, 26.2) I para 9% (21/222)
Gestational age .294 d 3.3 (1.4, 7.4) 2.5 (1.0, 6.2) II para 19% (15/77)
Duration of first stage of labor 1.1 (1.1, 1.2) $ III para 5% (1/20)
Second stage of labor .1 h 2.6 (1.5, 4.5)
Duration of labor .12 h 2.6 (1.5, 4.5)
Oxytocin augmentation 4.1 (2.2, 7.8)
Epidural anesthesia 2.3 (1.3, 4.1) parts of the sphincter. Before pregnancy, 2% (one of 46)
Sitting position 2.2 (1.3, 3.9) had symptoms of fecal incontinence and an additional
Lithotomy position 4.6 (1.9, 11.1) 13% (6 of 46) had gas incontinence only. At 5 months
Kneeling position 0.4 (0.2, 0.7)
after primary sphincter repair, 4% (two of 46) had
Upright position 0.3 (0.2, 0.6)
Fundal pressure 7.6 (4.3, 13.6) 4.6 (2.3, 7.9)
symptoms of fecal incontinence and an additional 50%
Use of instruments 6.5 (3.2, 13.6) (23 of 46) had symptoms of gas incontinence only. At 9
Midline episiotomy 14.0 (4.5, 43.2) 5.5 (1.4, 18.7) months after the primary repair, 2% (one of 46) had
Fetal weight (in steps of 250 g, 1.1 (1.0, 1.3) 1.3 (1.1, 1.6) symptoms of fecal incontinence and an additional 39%
continuous)
(18 of 46) of gas incontinence only (Table 6). Of women
OR 5 odds ratio; CI 5 confidence interval. having symptoms before pregnancy, three had under-
Variables presented in “chronologic” order in relation to labor.
gone one previous vaginal delivery, one had two pre-
vious deliveries, and three were nulliparous.
significantly associated with nulliparity, postmaturity, Of the 789 women without a clinically detected
duration of first stage of labor, pathologic duration of sphincter tear, 574 completed all questionnaires to be
second stage of labor, pathologic duration of labor, included in the questionnaire study. Twenty-three per-
oxytocin augmentation, and epidural anesthesia (Table cent (132 of 574) of the women had no tear at all, 77%
4). Maternal birth positions were significantly associ- (441 of 574) had first- or second-degree tears, and 0.2%
ated with sphincter tears in the sitting and lithotomy (one of 574) had a tear of unknown degree (they did not
positions, but inversely significantly associated in the have a third- or fourth-degree tear and no primary
kneeling position (Table 4). Of interventions toward the repair was performed). Before pregnancy, 1% (five of
end of labor, external fundal pressure, instrumental 574) of women had symptoms of fecal incontinence and
delivery, and midline episiotomy were associated with an additional 8% (48 of 574) of gas incontinence only. At
sphincter tears. Fetal weight also was associated with 5 months after delivery, 2% (nine of 574) had fecal
sphincter tears (Table 4). Remaining maternal and fetal incontinence, and an additional 22% (126 of 574) had
factors or different types of interventions during labor symptoms of gas incontinence only. At 9 months after
were not associated with anal sphincter tears in univar-
iate analyses. Table 6. Symptoms in Women With Clinically Detected
When analyzing the above identified risk factors in a Anal Sphincter Tears
multivariate model, nulliparity, postmaturity, fundal
5 mo 9 mo
pressure, midline episiotomy, and birth weight were Before pregnancy* postpartum† postpartum
independently associated with sphincter tears (Table 4). Symptoms (%) (n 5 46) (%) (n 5 46) (%) (n 5 46)
Interventions, defined as one or more of fundal pres-
No symptoms 85 46 59
sure, instrumental delivery, and episiotomy, were rela- Fecal incontinence‡
tively less used in association with the upright positions ,1/wk 2 0 2
(kneeling, standing, or squatting) than in nonupright .1/wk 0 4 0
position (lithotomy or sitting) (P , .001). Daily 0 0 0
Gas incontinence§
The frequency of preexisting symptoms of anal in-
,1/wk 9 22 28
continence in relation to parity are presented in Table 5. .1/wk 2 24 9
Forty-six of the 54 women with a clinically detected Daily 2 4 2
sphincter injury completed all questionnaires to be * Change of anal incontinence before pregnancy and 9 months after
included in the questionnaire study. Of these women, delivery, P 5 .006.
†
4% (two of 46) had fourth-degree tears, 15% (seven of Change of anal incontinence from 5 to 9 months after delivery, P 5
.11.
46) had third-degree tears involving the complete ‡
Some of these women were incontinent also of gas.
sphincter, and 80% (37 of 46) had tears involving just §
None of these women were incontinent of feces.
VOL. 94, NO. 1, JULY 1999 Zetterström et al Obstetric Anal Sphincter Tears 25
Table 8. Outcome of Primary Repair in Previous Studies
Number of
evaluated Incidence of fecal Incidence of gas Incidence of unspecified
Study sphincter tears incontinence incontinence anal incontinence
Another possible problem might stem from the anat- may be the low number of instrumental deliveries.
omy of the birth canal. The axis of the birth canal is However, any intervention that substantially acceler-
slightly J-shaped with the bottom of the “J” represented ates the last part of the second stage of labor could be
by the rectum, anal canal, and the perineum. The main harmful to the tissues of the pelvic floor.
part of the applied power will thus be directed into this In a previous study by Sultan et al,1 35% of nullipa-
area. rous and 4% of parous women sustained an occult
Midline episiotomy is known to be closely connected sphincter injury at delivery. The clinical detection rate
to a risk for development of third- and fourth-degree of sphincter tears in the same study was 3% of all
obstetric tears.25,29 –31 In spite of this knowledge, mid- vaginal deliveries. This raises an important problem.
line episiotomy is still frequently used, with a reported Most delivery institutions do not have access to endo-
incidence of 50% in the United States.32 Prevailing anal ultrasonography needed to assess the sphincters
arguments in favor of this technique are better healing immediately postpartum. Thus, the majority of obstetric
conditions and less postoperative complaints.29 At our practitioners must rely on their clinical experience and
institution, where the overall incidence of episiotomy judgment and therefore be aware of the risk of unde-
was low (2%), all but two were midline, the association tected sphincter tears at clinical examination.
with sphincter tears was strong in the multivariate The symptomatic outcome of primary repair in the
analysis. Even though our sample was small, we believe present study must be considered as reasonable, even
that midline episiotomy preferably should not be used. though 41% of treated women reported symptoms. The
Mediolateral episiotomy was found by Poen et al23 to majority of these women reported minor and infrequent
have a decreased risk for sphincter tears among nullip- bouts of gas incontinence, and only one of the 46
arous but not among parous women. In our study, only woman experienced fecal incontinence at 9 months
two women had a mediolateral episiotomy and thus no (Table 5). Our frequency of incontinence after primary
conclusions can be made concerning the mediolateral repair is lower than in several previous studies (Table
technique. Both these women were nulliparous, and 8). We are not sure about the reason. It is noteworthy
none had a sphincter tear. that several of the sphincter injuries in the present study
We found an increased risk for sphincter tears with were minor. However, when comparing the symptom-
increased fetal weight. We have chosen to analyze this atic outcome in women with partial and complete
variable as a continuous variable in steps of 250 g, sphincter tears, numbers were too small to find any
which gave us an odds ratio of 1.3. Analyzed in steps of significant difference.
500 or 1000 g, the odds ratio was 2.1 and 4.4, respec- The frequency of anal incontinence was more than
tively. Our finding is in accordance with several previ- twice as high in the sphincter injury group compared
ous studies,20,22,23,30,33 and is quite understandable, with the non–sphincter injury group at 5 months (54%
given that increased size of the fetus might predispose compared with 23%). At 9 months, there was still a
for sphincter injuries. difference, but injury group showed improvement (41%
Instrumental delivery is known to increase risk for of the injury group was incontinent compared with 24%
sphincter tears, and this risk is more pronounced with of the noninjury group). The frequency of anal inconti-
forceps compared with vacuum delivery.20,34 In the nence did not change between 5 and 9 months (Tables 6
present study, all but two instrumental deliveries were and 7). However, the injury group showed a tendency
by vacuum extraction. We found an association be- to less severe symptoms (Table 6). This improvement
tween instrumental delivery and sphincter tears using between 5 and 9 months is in contrast with the findings
univariate but not multivariate analysis. The reason reported by Haadem et al,36 who did not find any
VOL. 94, NO. 1, JULY 1999 Zetterström et al Obstetric Anal Sphincter Tears 27
28. World Health Organization. Maternal and newborn health/safe rupture and controls. Eur J Obstet Gynecol Reprod Biol 1990;35:7–
motherhood unit, family and reproductive health. Care in normal 13.
birth: A practical guide. Geneva, Switzerland, 1996. 37. Haadem K, Ling L, Ferno M, Graffner H. Estrogen receptors in the
29. Coats PM, Chan KK, Wilkins M, Beard RJ. A comparison between external anal sphincter. Am J Obstet Gynecol 1991;164:609 –10.
midline and mediolateral episiotomies. Br J Obstet Gynaecol 38. Schellart RP, Schouten WR, Huikeshoven FJ. Anorectal manome-
1980;87:408 –12. try before, during and after estrogen replacement therapy. Int
30. Shiono P, Klebanoff MA, Carey JC. Midline episiotomies: More Urogynecol J Pelvic Floor Dysfunct 1996;7:77– 80.
harm than good? Obstet Gynecol 1990;75:765–70.
31. Labrecque M, Baillargeon L, Dallaire M, Tremblay A, Pinault JJ,
Gingras S. Association between median episiotomy and severe
Address reprint requests to:
perineal lacerations in primiparous women. Can Med Assoc J
1997;156:797– 802.
Jan Zetterström, MD
32. National Center for Health Statistics. Vital and Health Statistics. Division of Obstetrics and Gynecology
Detailed diagnosis and procedures, national hospital discharge Karolinska Institutet
survey, 1994. Hyattsville, Maryland, 1997. Danderyd Hospital
33. Klein MC, Janssen PA, MacWilliam L, Kaczorowski J, Johnson B. S-182 88 Danderyd
Determinants of vaginal–perineal integrity and pelvic floor func- Sweden
tioning in childbirth. Am J Obstet Gynecol 1997;176:403–10.
34. Sultan AH, Kamm MA, Bartram CI, Hudson CN. Anal sphincter
trauma during instrumental delivery. Int J Gynaecol Obstet 1993;
43:263–70. Received September 1, 1998.
35. Nielsen MB, Hauge C, Rasmussen OO, Pedersen JF, Christiansen J. Received in revised form December 21, 1998.
Anal endosonographic findings in the follow-up of primarily Accepted January 7, 1999.
sutured sphincteric ruptures. Br J Surg 1992;79:104 – 6.
36. Haadem K, Dahlstrom JA, Lingman G. Anal sphincter function Copyright © 1999 by The American College of Obstetricians and
after delivery: A prospective study in women with sphincter Gynecologists. Published by Elsevier Science Inc.