International Urogynecology Journal

, Volume 18, Issue 12, pp 1399–1404

Decreased rate of obstetrical anal sphincter laceration is associated with change in obstetric practice

Authors

    • Department of Obstetrics and Gynecology, Keck School of MedicineUniversity of Southern California
    • Kaiser Permanente Baldwin Park Medical Center
  • Begüm Özel
    • Department of Obstetrics and Gynecology, Keck School of MedicineUniversity of Southern California
  • Nicole M. Gatto
    • Department of Preventive Medicine, Keck School of MedicineUniversity of Southern California
  • Lisa Korst
    • Department of Obstetrics and Gynecology, Keck School of MedicineUniversity of Southern California
  • Daniel R. MishellJr
    • Department of Obstetrics and Gynecology, Keck School of MedicineUniversity of Southern California
  • David A. Miller
    • Department of Obstetrics and Gynecology, Keck School of MedicineUniversity of Southern California
Original Article

DOI: 10.1007/s00192-007-0353-5

Cite this article as:
Minaglia, S.M., Özel, B., Gatto, N.M. et al. Int Urogynecol J (2007) 18: 1399. doi:10.1007/s00192-007-0353-5

Abstract

A study was conducted to describe the rate of obstetrical anal sphincter laceration in a large cohort of women and to identify the characteristics associated with this complication. Data from all vaginal deliveries occurring between January 1996 and December 2004 at one institution were used to compare women with and without anal sphincter lacerations. Among 16,667 vaginal deliveries, 1,703 (10.2%) anal sphincter lacerations occurred. Regression models suggested that episiotomy (OR 1.36; 95% CI 1.16, 1.58), vacuum delivery (OR 3.19; 95% CI 2.69, 3.79), and forceps delivery (OR 2.79; 95% CI 1.94, 4.02) were each associated with the increased risk of anal sphincter laceration. Year of delivery was associated with a decreased risk of anal sphincter laceration (OR 0.94; 95% CI 0.92, 0.96) with the rate of laceration decreasing from 11.2% to 7.9% during the study period. Episiotomy and operative vaginal delivery are significant, modifiable risk factors. Changes in obstetric practice may have contributed to the dramatic reduction in anal sphincter laceration during the study period.

Keywords

Perineal lacerationObstetrical anal sphincter lacerationOperative vaginal deliveryEpisiotomy

Introduction

Obstetrical anal sphincter lacerations (third-degree or fourth-degree perineal lacerations) are the leading cause of fecal incontinence in young women [1, 2]. An estimated 20–50% of women have experienced anal incontinence symptoms after primary repair of anal sphincter lacerations [38]. Other long-term complications may include perineal pain, dyspareunia, and recto-vaginal fistula. A number of observational studies suggest that modifiable risk factors, such as operative vaginal delivery and episiotomy, may increase the risk of obstetrical anal sphincter laceration [915].

Changes in obstetric practice such as decreased use of operative vaginal delivery and episiotomy appear associated with a decline in the rate of anal sphincter laceration [11, 12, 16]. The objective of the current study was to determine the rate of obstetrical anal sphincter laceration in a large cohort of women undergoing vaginal delivery and to identify characteristics associated with this complication.

Materials and methods

Study population

Women with vaginal deliveries occurring between January 4, 1996 and December 8, 2004 at one urban university medical center were selected for the study. Women were excluded if they had a vaginal twin delivery. Cesarean deliveries were counted for descriptive purposes but were not included in the study sample.

Practitioners at this institution included obstetrics and gynecology residents, midwives, generalist obstetricians, and maternal fetal medicine fellows and attending physicians. Medical students and interns performed the majority of deliveries and were supervised by midwives, attending physicians or resident physicians. Third-degree perineal lacerations were defined as an obstetrical laceration involving any portion of the anal sphincter complex. Fourth-degree perineal lacerations were defined as an obstetrical laceration involving any portion of the rectal mucosa.

Midline episiotomies were nearly always performed during the study period, although mediolateral episiotomies may have been utilized in some cases. During the study period, medical students and interns performing vaginal deliveries were generally instructed not to perform episiotomy unless expeditious delivery was necessary for fetal indications.

A computerized database with recorded information for all deliveries occurring at >20 weeks gestation at the institution was available and was used to obtain data for each woman with a delivery meeting the inclusion criteria during the study period. The Institutional Review Board at the Los Angeles County Medical Center, University of Southern California approved the study.

Statistical analysis

Rates of anal sphincter lacerations and vaginal and cesarean deliveries were calculated for each year during the time period. T tests and chi-squared tests were used to identify the factors associated with anal sphincter laceration including maternal age, gestational age, birth weight, vacuum or forceps delivery, episiotomy, estimated blood loss at delivery, presence of shoulder dystocia, use of epidural anesthesia, and use of oxytocin. Because some women had more than one delivery at the medical center, generalized estimating equations (GEE) with robust standard error estimates, an identity link function, and exchangeable working correlation matrix were used to model the association between anal sphincter lacerations and the factors identified in univariate analyses. A backward stepwise elimination approach was employed to build a regression model for anal sphincter lacerations with significant independent predictors of anal lacerations. A p value of <0.05 was used as the criterion to retain a factor in the model given the large sample size included in the analysis. Odds ratios associated with each factor were estimated from the models. Potential confounding factors such as maternal age were retained in the final models if they changed the effect estimates (beta coefficients) by more than 10%. The variables birth weight and estimated blood loss were centered on their means in order for the odds ratios associated with these factors to be clinically meaningful.

All analyses were performed using the Intercooled STATA 8.0 (Stata, College Station, TX).

Results

During the 9-year time period, there were 16,667 vaginal births among 14,949 women; 1,718 women had multiple deliveries at the medical center. Mean age (SD) was 27.1 (6.5) years, and the women were 87.7% Hispanic, 6.1% African American, 3.2% Asian Pacific Islander, 2.7% Caucasian, and 0.03% other or unknown. Additional demographic and obstetrical characteristics of the 14,949 women in the study are presented in Table 1.
Table 1

Study population demographic and obstetric characteristics (N = 14,949 women and 16,667 deliveries)

Variable

Case (laceration) N = 1,703

Control (no laceration) N = 14,964

p value

Maternal characteristics

Age (years)

25.5 ± 6.2

27.3 ± 6.5

<0.0001

Gravidity

2.2 ± 1.6

3.2 ± 2.1

<0.0001

Parity

1.0 ± 1.2

1.7 ± 1.7

<0.0001

Obstetric procedure used

 Episiotomy

  Yes

316 (18.6%)

1,108 (7.4%)

<0.0001

  No

1,387 (81.4%)

13,856 (92.6%)

 Vacuum

  Yes

276 (16.2%)

537 (3.6%)

<0.0001

  No

1,427 (83.8%)

14,427 (96.4%)

 Forceps

  Yes

46 (2.7%)

157 (1.1%)

<0.0001

  No

1,657 (97.3%)

14,807 (98.9%)

 Epidural anesthesia

  Yes

699 (41.1%)

5,283 (35.3%)

<0.0001

  No

1,004 (58.9%)

9,681 (64.7%)

 Oxytocin

  Yes

736 (43.2%)

5,851 (39.1%)

0.001

  No

967 (56.8%)

9,113 (60.9%)

Obstetric complications

 Shoulder dystocia

  Yes

51 (3.0%)

181 (1.2%)

<0.0001

  No

1,652 (97.0%)

14,783 (98.8%)

 Estimated blood loss (cc)

367 ± 252

304 ± 205

<0.0001

Fetal characteristics

 Birth weight (g)

3,354 ± 496

3,248 ± 651

<0.0001

 Estimated gestational age (weeks)

38.8 ± 1.9

38.2 ± 2.9

<0.0001

Values are expressed as the mean±standard deviation (SD) or frequency (%)

There were 1,508 third-degree perineal lacerations and 135 fourth-degree perineal lacerations for a total of 1,703 (10.2%) perineal lacerations involving the anal sphincter: 857 (17.5%) occurred among 4,891 primiparas and 846 (7.2%) occurred among 11,776 multiparas. Women with a third-degree or fourth-degree anal sphincter laceration were significantly younger and had significantly fewer pregnancies and deliveries than women without a laceration during vaginal delivery.

All factors analyzed were individually significant predictors of anal sphincter lacerations at p < 0.001. Women with an anal sphincter laceration were significantly more likely to have a fetus of greater gestational age and birth weight than women without a laceration. Women with an anal sphincter laceration were also more likely to have had more blood loss and experienced shoulder dystocia during delivery than women without a laceration. Episiotomy, vacuum, forceps, epidural anesthesia and oxytocin were all more common among women who had lacerations compared to those without lacerations.

Gravidity and parity were highly significantly correlated (r = 0.91, p < 0.0001) and the decision was made to include parity but not gravidity in the model selection process. Maternal age was not a significant predictor of anal lacerations when other factors were included in the model, nor was there evidence that it was a confounder, and thus it was not included in the final model. With the exception of oxytocin use during labor, all other factors were significant independent predictors of anal sphincter lacerations after adjusting for the effect of the other variables.

Results from regression models suggested that women who had an episiotomy during delivery were almost 40% more likely to have an anal sphincter laceration than women without an episiotomy (OR 1.36; 95% CI 1.16, 1.58) (Table 2). Use of vacuum or forceps approximately tripled a woman’s risk for a laceration (OR 3.19; 95% CI 2.69, 3.79 and OR 2.79; 95% CI 1.94, 4.02, respectively) compared with lack of use of these instruments. Presence of shoulder dystocia at delivery doubled the risk of laceration (OR 2.03; 95% CI 1.44, 2.86). Increasing parity and use of epidural anesthesia both demonstrated a significant protective effect. Anal sphincter lacerations were also significantly related to blood loss. A laceration was associated with a 6% increase in the risk of blood loss of 100 ml. For every week increase in gestation age beyond 38.3 weeks, the risk of laceration increased by 4% and for every 500 g of birth weight greater than 3,259 g, the risk of laceration increased by 15%.
Table 2

Anal sphincter laceration odds ratios and 95% confidence intervals from regression models of the association with obstetric factors among women with vaginal deliveries

Variable

OR

95% CI

Episiotomy

 No

1.00

 Yes

1.36

1.16–1.58

Vacuum

 No

1.00

 Yes

3.19

2.69–3.79

Epidural anesthesia

 No

1.00

 Yes

0.86

0.76–0.96

Forceps

 No

1.00

 Yes

2.79

1.93–4.02

Shoulder dystocia

 No

1.00

 Yes

2.03

1.44–2.86

Estimated blood lossa

1.06

1.02–1.09

Estimated gestational ageb

1.03

1.01–1.06

Birth weightc

1.15

1.08–1.22

Parity (number of previous deliveries)

 None

1.00

 

 1

0.57

0.50–0.66

 2

0.41

0.35–0.48

 3 or more

0.18

0.15–0.22

Each factor in the regression model is adjusted for the effects of all other factors in the table.

OR: odds ratios, 95% CI: 95% confidence interval.

aPer 100 ml blood loss.

bPer week gestational age.

cPer 500 g birth weight.

Year of delivery was also associated with a decreased risk of anal sphincter laceration with each year after 1996 conferring a 6% reduction in risk (OR 0.94; 95% CI 0.92, 0.96) and the rate of anal sphincter laceration decreasing from 11.2% to 7.9% during the study period (Fig. 1). At the same time, the episiotomy rate decreased from 9% (95% CI 8.4, 10.3) to 8% (95% CI 6.4, 10.0), the vacuum-assisted vaginal delivery rate decreased from 5.1% (95% CI 4.3, 5.8) to 2.9% (95% CI 1.8, 4.1), and the forceps-assisted vaginal delivery rate decreased from 1.7% (95% CI 1.3, 2.1) to 0.0%. Decreases in the rates of vacuum-assisted and forceps-assisted vaginal delivery were statistically significant at p < 0.05. The cesarean section rate increased from 18.2% to 32.3% during the study period.
https://static-content.springer.com/image/art%3A10.1007%2Fs00192-007-0353-5/MediaObjects/192_2007_353_Fig1_HTML.gif
Fig. 1

Trend in obstetrical anal sphincter laceration, episiotomy, vacuum-assisted delivery, and forceps-assisted delivery between 1996 and 2004

Discussion

The rate of obstetrical anal sphincter laceration at this institution decreased from 11.2% to 7.9% during the study period. Handa et al. reported an overall rate of 5.85% among over 2 million vaginal deliveries in California during the period between 1992 and 1997. This rate had decreased from 6.35% in 1992 to 5.43% in 1997 [9]. Dandolu et al. reported a rate of 7.3% of third-degree and fourth-degree perineal tears among over 250,000 deliveries in Pennsylvania during the period between 1990 and 1991 [12]. The current rate of anal sphincter laceration of 7.9% is consistent with these reported rates. Other recently published rates have ranged from a low of 1.94% [17] to a high of 17% [10].

Most risk factors for anal sphincter laceration are not modifiable, these include primiparity [9, 13, 15], race [9, 13, 15], perineal anatomy [18], fetal position [17], and birth weight [9, 1315]. Potentially modifiable risk factors for obstetrical anal sphincter injury, such as operative vaginal delivery and episiotomy, have been established [915]. Reports indicate that restrictive use of episiotomy and declining use of operative vaginal delivery are associated with a decline in the rate of anal sphincter laceration [11, 12, 16]. Hence, long-term complications after primary repair, such as anal incontinence, recto-vaginal fistula, perineal pain, and dyspareunia may potentially decrease as well.

During the study period, the rate of episiotomy decreased from 9% to 8%. Overall, women who had an episiotomy during delivery were almost 40% more likely to have an anal sphincter laceration compared to women without an episiotomy. A comparison of risk with mediolateral vs midline episiotomy could not be performed due to the method of data collection during the study period. It is estimated that the majority of episiotomies performed during the study period were midline because this was the standard of practice at this institution.

The impact of episiotomy on the rate of anal sphincter laceration has been the subject of controversy. Handa et al. reported a decrease in the likelihood of third-degree lacerations when episiotomy was used [9]. This study, however, was a large population-based retrospective study in which both midline and mediolateral episiotomies were included together in the analysis. Dandolu et al. also reported that episiotomy was protective to the perineum, although the data could not be analyzed based on the type of episiotomy [12].

In a systematic review of mediolateral and midline episiotomy trials comparing restrictive and liberal use of episiotomy, Eason et al. concluded that “although mediolateral episiotomy does not protect the anal sphincter, median episiotomy clearly puts it in greater peril” [19]. Legino et al. reported an abrupt rise in third-degree lacerations when episiotomy technique changed from mediolateral to median at one hospital [10]. Clemons et al. later reported a decrease in anal sphincter lacerations associated with restrictive episiotomy use in an institution that typically used midline episiotomy [11].

There is no consensus as to what constitutes an ideal rate of episiotomy use during uncomplicated births. Systematic review suggests that a rate below 30% may be optimal [19]. Weber and Meyn reported that the rate of episiotomy in the United States decreased from 65.3% in 1979 to 38.6% in 1997 [20]. The CDC National Hospital Discharge Survey from 2002 reveals that the rate fell to 26.9% that year. Our data show that an episiotomy rate as low as 8% can be reasonably achieved when episiotomy is performed only for fetal indications. Ultimately, an approach that minimizes maternal trauma without jeopardizing neonatal outcome is necessary.

Both vacuum and forceps delivery have previously been identified as independent risk factors for anal sphincter laceration in retrospective studies [9, 12, 14, 15]. At this institution, the rate of vacuum delivery decreased to 42% while the rate of forceps delivery decreased to 88% during the study period. Eason et al. reviewed several randomized trials comparing maternal trauma after forceps delivery vs vacuum extraction and concluded that more women had anal sphincter injury with forceps delivery than with vacuum delivery [19]. These findings may have partially contributed to the greater, more disproportionate drop in forceps deliveries vs vacuum deliveries at our institution. Fear of fetal injury and a decline in resident education of forceps deliveries may have also contributed to this disproportionate drop.

An increase in the overall rate of cesarean section from 18.2% in 1996 to 32.3% in 2004 accompanied the overall decline in the use of operative vaginal delivery. For comparison, the overall cesarean section rate in the United States was 27.5% in 2003 [21]. This rate had risen over 30% since 1996. Furthermore, the overall operative vaginal delivery rate in the United States decreased from 9.5% in 1994 to 5.6% in 2003.

We found a protective effect of epidural anesthesia. This finding is in contrast to previous studies showing an increased risk associated with epidural use. Donnelly et al. found that epidural analgesia, used in 58% of vaginal deliveries, prolonged the second stage of labor (OR 7.7; 95% CI 4.0, 14.7) and was associated with the increased risk of anal sphincter injury (OR 2.1; 95% CI 1.1, 4.0) [22]. Poen et al. also found an increased risk of anal sphincter injury in women with epidural anesthesia [14] and Williams et al. found a tendency toward more epidural use (OR 1.64; CI 0.97–2.75) in women with anal sphincter injuries [23]. It has been suggested that prolongation of the second stage of labor may be associated with increased anal sphincter injury in women with epidural anesthesia [22].

Janni et al. suggested that the increase of maternal morbidity in patients with prolonged labor is partially attributed to a higher rate of operative procedures in these patients [24]. The presence of adequate anesthesia may also make it more likely for the obstetrician to elect to proceed with operative delivery rather than expectant management in these women [25]. The increased rate of cesarean section and the decreased rate of operative vaginal delivery at our institution may reflect an increasing tendency to use cesarean section for arrest of descent at complete cervical dilation. Before the study period, these patients may have been more likely to undergo an operative vaginal delivery and/or have had an episiotomy to facilitate vaginal delivery.

These results agree with prior studies that indicate an association between anal sphincter lacerations and increased birth weight [9, 1315]. This association is likely due to increased mechanical stress during delivery of a large fetus. Other fetal factors that have been associated with anal sphincter laceration include the presence of shoulder dystocia at delivery and occiput posterior position [17]. In this study, shoulder dystocia was a risk factor for anal sphincter laceration with an adjusted odds ratio of 2.03. Fetal position could not be evaluated because the database used for the study did not contain this information.

An additional limitation to the study is that the effect of race on the rate of anal sphincter laceration could not be evaluated. The database used for this study was not linked to the hospital database containing racial and ethnic demographic information at the time of this report. It is unlikely that this information will generate significant conclusions given the racial homogeneity of the study population.

Changes in obstetric practice such as increased cesarean delivery and decreased episiotomy and operative vaginal delivery may have contributed to the dramatic reduction in anal sphincter laceration during the study period. Modifiable risk factors should be minimized whenever possible to prevent long-term complications of obstetrical anal sphincter injury. Physicians should include a discussion of the risk of anal sphincter laceration when counseling women for non-emergent operative vaginal delivery. The assumption that vaginal birth, when it is achieved with instrumental assistance, is preferable to cesarean delivery does not take into consideration the implications of operative vaginal delivery for anal sphincter injury. Judicious use of episiotomy, restrictive use of operative vaginal delivery when the sole indication is prolonged second stage of labor, and offering cesarean delivery as an alternative option to women with significant risk factors are all potential obstetrical interventions that may decrease the rate of obstetrical anal sphincter injury.

Copyright information

© International Urogynecology Journal 2007