Episiotomy is one of the most commonly performed surgical procedures in women in the United States. More than 1.4 million episiotomies were performed in 1995 for women aged 15–44 years, for a rate of almost 1200 procedures per 100,000 female population.1 Population-based rates with a large denominator are useful for comparing rates among different operations, especially those that are less common. The rate of episiotomy is extraordinarily high, even compared with other commonly performed procedures in women, such as hysterectomy (434 per 100,000 female population) and cholecystectomy (244.6 per 100,000 female population).1
There is no consensus as to what constitutes an appropriate or ideal rate of episiotomy at otherwise uncomplicated vaginal births. Some studies have concluded that episiotomy rates over 20% cannot be justified, as this rate has been achieved in some studies when episiotomy use is restricted to specific indications.2,3 However, this is not evidence that this rate is optimal; many studies, particularly with midwife practitioners, have achieved even lower rates with good maternal outcomes.4–6 Different rates of episiotomy are not well explained by differences in the patient population but are largely due to differences in providers and their beliefs about the benefits of episiotomy.7–9
The use of episiotomy has been studied in single institutions,6,10,11 on a regional basis,12 and on a national basis in the United States.13,14 Some studies have documented impressive decreases in use of episiotomy, especially at single institutions. Comparisons regarding the use of episiotomy at a local or regional level are best understood using the number of vaginal deliveries as the denominator; 100 vaginal deliveries is chosen as a convenient reference point. For example, the rate of episiotomy dropped from 86.8% to 10.4% from 1976 to 1994 at the University of California in San Francisco,6 and from 69.6% to 19.4% from 1983 to 2000 at Thomas Jefferson University Hospital in Philadelphia.11 Based on National Hospital Discharge Survey data, the rate of episiotomy decreased slightly, from 64.0 per 100 deliveries in 1980 to 61.9 per 100 deliveries in 1987 (change not statistically significant).12 A more recent study demonstrated a 39% decrease in episiotomies in the United States14; however, the rate in 1998 was still far higher than current evidence suggests is appropriate. In that study, analyses were not performed to calculate age-adjusted rates based on population data.
The objective of this study was to describe time trends in the use of episiotomy in the United States from 1979 to 1997 and to describe demographic characteristics associated with episiotomy use.
MATERIALS AND METHODS
The data used in this study were from the National Hospital Discharge Survey, a survey that has been conducted continuously by the National Center for Health Statistics since 1965. This study used data from 1979 to 1997. National Hospital Discharge Survey data were collected from a sample of inpatient records acquired from a national probability sample of hospitals. Only general hospitals with an average length of stay of fewer than 30 days for all patients were included. Federal, military, and Department of Veterans Affairs hospitals were excluded; hospital units of institutions, such as prison hospitals, and hospitals with fewer than six beds were also excluded. The database provides patient characteristics (age, sex, and race), expected source of payment for hospitalization (grouped into three categories of private, government [includes Medicare and Medicaid], and other insurance), up to seven diagnoses and up to four procedures based on International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes, hospital size by number of beds, and geographic region (Northeast, Midwest, South, and West). The overall error rate for medical coding and data entry was 4.3% and, for demographic coding and data entry, the error rate was 1.4%.1
Using ICD-9-CM diagnosis codes, a population of women was identified with singleton, vertex presentation, live-born vaginal deliveries. Breech vaginal deliveries, cesarean deliveries, multiple gestations, and preterm deliveries were excluded. Using ICD-9-CM procedure codes, this population was then divided into two groups based on whether episiotomy was performed (code 73.6 for episiotomy with spontaneous delivery and codes 72.1, 72.21, 72.31, and 72.71 for episiotomy with operative vaginal delivery). We calculated the population-based rate of episiotomy per year, using the 1990 census population of women aged 15–44 years as the denominator. We also calculated the rate of episiotomy per 100 vaginal births. We then examined the time trend of rate of episiotomy by year and tested for change using the score test for linear trend.15 Patient and demographic characteristics were compared between women who did and did not undergo episiotomy, using the χ2 test for categorical variables and using the Student t test for continuous variables. Multiple logistic regression was used to examine variables for independent association with episiotomy. Weighted methods were used for statistical testing and estimation to account for the survey's sample design. All statistical analyses were performed using Stata statistical software 7.0 (Stata, College Station, TX). A P value of .05 or less was accepted as statistically significant.
All differences noted were statistically significant with P< .001, unless otherwise specified. The database represented an estimated 73,253,000 deliveries over the 19 years from 1979 to 1997, of which 54,085,000 (74%) were vaginal deliveries. During this time period, the age-adjusted rate for all deliveries stayed relatively stable, varying slightly from year to year, from a low of 62.6 (in 1986) to a high of 69.8 (in 1993) deliveries per 1000 women aged 15–44 (P = 0.1). The number of episiotomies performed per year ranged from a high of 2,015,000 in 1981 to a low of 1,128,000 in 1997. The age-adjusted rate of episiotomies for all vaginal deliveries per 1000 women aged 15–44 years ranged from a high of 32.7 in 1979 to a low of 18.7 in 1997 (Table 1). As shown in Table 1 and Figure 1, the rate of episiotomy per 100 vaginal deliveries decreased significantly over time. The rate with spontaneous vaginal deliveries decreased from 60.1% in 1979 to 32.8% in 1997.
Episiotomy with operative deliveries decreased over time from 87.0% in 1979 to 70.8% in 1997. The rate of episiotomy with operative vaginal deliveries stayed relatively level, in the range of 84.0% to 89.9% through most of the period under study (1979–1992) but then decreased in the last 5 years (1993–1997) to a low of 70.8% in 1997 (Figure 1). Overall, episiotomy was performed with 89.1% of forceps deliveries, 76.4% of vacuum deliveries, and in 88.1% of deliveries in which both forceps and vacuum were used.
By univariable and multivariable analyses, the characteristics of women who underwent episiotomy differed from those women in whom episiotomy was not performed. Table 2 presents data on the overall proportions of characteristics, the unadjusted odds ratios, and odds ratios adjusted for year of delivery, age, race, marital status, geographic region, insurance status, and hospital size. Another model was developed, in which year of delivery was not included and the adjusted odds ratios were similar to those in Table 2 (data not shown). Women who delivered with episiotomy were slightly younger (25.7 ± 5.5 years) than women without episiotomy (26.2 ± 5.7 years). An increased rate of episiotomy occurred in women of white race (compared with black), married (compared with single), with private insurance (compared with government insurance). Women living in the West had the lowest rate of episiotomy compared with the other geographic regions.
Anal sphincter laceration occurred in 8.3% of women (on average, 1 in 12) with episiotomy, compared with 3.8% of women (1 in 26) delivered without episiotomy. This reflects an excess of 1,546,577 anal sphincter lacerations in women with episiotomy, or an average of about 81,400 anal sphincter lacerations per year. The association of anal sphincter laceration did not change with time. Other complications in women with and without episiotomy during the index hospitalization are listed in Table 3. Women with episiotomy had a slightly longer hospital length of stay on average (2.6 ± 2.4 days), compared with women without episiotomy (2.3 ± 2.9 days).
Recent reviews have conclusively determined that the routine use of episiotomy should be abandoned and that perineal trauma is decreased when episiotomy is not performed.16,17 However, the optimal rate of episiotomy for maximizing maternal and fetal well-being is not known. Similar to rates of cesarean delivery, the “right” rate has not been identified by scientific evidence and is subject to change based on nonmedical factors. The strongest evidence-based support of the optimal rate of episiotomy that can be identified at present relies on studies that report good maternal and neonatal out comes with very low rates of episiotomy, ten or fewer per 100 vaginal deliveries.4–6 On that basis, even though the “right” rate of episiotomy is unknown, it is appropriate to conclude that the episiotomy rate of 39 per 100 vaginal deliveries in 1997 in the United States is still far higher than is supported by scientific evidence.
It is a positive finding of this study that episiotomy rates in the United States have decreased slowly but steadily over the past two decades. However, when considered in the context of operations performed on women, the rate of episiotomy (1050.3 per 100,000 female population) still dwarfs that of the next most common procedures, such as cesarean delivery (584.5 per 100,000 female population) and hysterectomy (434.0 per 100,000 female population).1 By traditional obstetric practice, women are not necessarily informed of the specific risks and benefits associated with performing episiotomy, and rarely is written informed consent obtained, somehow abrogating the standard set for every other surgical procedure. The view that episiotomy is not a significant operation has contributed to the serious underestimation of its risks and long-term adverse consequences.
One of the most serious consequences of midline episiotomy is its association with a greatly increased risk of anal sphincter damage.18–20 Even when anal sphincter damage is recognized at delivery, current methods of surgical repair are inadequate. Persistent anal sphincter defects are present in up to 85% of women who sustain anal sphincter laceration and repair at the time of vaginal delivery.21,22 After sphincter damage and repair, 42%–54% of women still experience symptoms of gas or fecal incontinence after delivery.22,23 Especially because surgical repair cannot restore normal anatomy and function, it is critically important to prevent the initial damage at vaginal delivery. Reducing the rate of episiotomy could avoid tens of thousands of instances of anal sphincter damage at vaginal delivery each year.
Our study found several demographic variables that predicted a higher rate of episiotomy in women of younger age, white race, who are married, with private insurance. The association of younger age with episiotomy may reflect an interaction between younger age at first birth and a higher rate of episiotomy at first birth, as other studies have reported5,8; however, parity was not available in this database. Other studies have found that the rate of episiotomy does not depend on clinical factors, but rather on physician preference. Our findings likely reflect the practice patterns of providers who care for women with these characteristics, rather than any specific biologic or pathophysiologic reason why these women would be at increased risk for episiotomy.
There are potential limitations in using a database such as the National Hospital Discharge Survey. Clinical information is only available as coded by diagnoses and procedures, so it was not possible to examine many maternal and fetal factors (such as fetal weight) that may be associated with an increased risk of episiotomy. The data reflect only the index hospitalization for delivery, and therefore complications associated with episiotomy are underestimated. Data on outpatient management, readmission to the hospital, and long-term complications are not available. Misclassification may occur if errors in coding or data entry are made; however, as noted in the Material and Methods section, the error rate is relatively low for this database. The database represents a sample of hospital discharges in the United States; the sampling procedure is designed to be representative of the population as a whole. Estimates based on small numbers of cases (such as some of the complications) have low reliability and should be interpreted with caution.
1. Gillum BS, Graves EJ, Wood E. National hospital discharge survey. Vital Health Stat 13 1998;133:1–51.
2. Argentine Episiotomy Trial Collaborative Group. Routine versus selective episiotomy: A randomised controlled trial. Lancet 1993;342:1517–8.
3. Henriksen TB, Bek KM, Hedegaard M, Secher NJ. Episiotomy and perineal lesions in spontaneous vaginal deliveries. Br J Obstet Gynaecol 1992;99:950–4.
4. Blanchette H. Comparison of obstetric outcome of a primary-care access clinic staffed by certified nurse-midwives and a private practice group of obstetricians in the same community. Am J Obstet Gynecol 1995;172:164–71.
5. Albers LL, Anderson D, Cragin L, Daniels SM, Hunter C, Sedler KD, et al. Factors related to perineal trauma in childbirth. J Nurse Midwifery 1996;41:269–76.
6. Bansal RK, Tan WM, Ecker JL, Bishop JT, Kilpatrick SJ. Is there a benefit to episiotomy at spontaneous vaginal delivery? A natural experiment. Am J Obstet Gynecol 1996;175:897–901.
7. Klein MC, Kaczorowski J, Robbins JM, Gauthier RJ, Jorgensen SH, Joshi AK. Physicians' beliefs and behaviour during a randomized controlled trial of episiotomy: Consequences for women in their care. Can Med Assoc J 1995;153:769–79.
8. Hueston WJ. Factors associated with the use of episiotomy during vaginal delivery. Obstet Gynecol 1996;87:1001–5.
9. Garcia FAR, Miller HB, Huggins GR, Gordon TA. Effect of academic affiliation and obstetric volume on clinical outcome and cost of childbirth. Obstet Gynecol 2001;97:567–76.
10. Ecker JL, Tan WM, Bansal RK, Bishop JT, Kilpatrick SJ. Is there a benefit to episiotomy at operative vaginal delivery? Observations over ten years in a stable population. Am J Obstet Gynecol 1997;176:411–4.
11. Goldberg J, Holtz D, Hyslop T, Tolosa JE. Has the use of routine episiotomy decreased? Examination of episiotomy rates from 1983 to 2000. Obstet Gynecol 2002;99:395–400.
12. Learman LA. Regional differences in operative obstetrics: A look to the South. Obstet Gynecol 1998;92:514–9.
13. Kozak L. Surgical and nonsurgical procedures associated with hospital delivery in the United States: 1980–1987. Birth 1989;16:209–13.
14. Weeks JD, Kozak LJ. Trends in the use of episiotomy in the United States: 1980–1998. Birth 2001;28:152–60.
15. Clayton D, Hills M. Statistical models in epidemiology Oxford: Oxford University Press, 1993.
16. Eason E, Labrecque M, Wells G, Feldman P. Preventing perineal trauma during childbirth: A systematic review. Obstet Gynecol 2000;95:464–71.
17. Carroli G, Belizan J. Episiotomy for vaginal birth. Cochrane Database Syst Rev 2000;1:1–9.
18. Shiono P, Klebanoff MA, Carey JC. Midline episiotomies: More harm than good? Obstet Gynecol 1990;75:765–70.
19. Klein MC, Janssen PA, MacWilliam L, Kaczorowski J, Johnson B. Determinants of vaginal-perineal integrity and pelvic floor functioning in childbirth. Am J Obstet Gynecol 1997;176:403–10.
20. Kammerer-Doak DN, Wesol AB, Rogers RG, Dominguez CE, Dorin MH. A prospective cohort study of women after primary repair of obstetric anal sphincter laceration. Am J Obstet Gynecol 1999;181:1317–23.
21. Fitzpatrick M, Behan M, O'Connell PR, O'Herlihy C. A randomized clinical trial comparing primary overlap with approximation repair of third-degree obstetric tears. Am J Obstet Gynecol 2000;183:1220–4.
22. Sorensen SM, Bondesen H, Istre O, Vilmann P. Perineal rupture following vaginal delivery. Long-term consequences. Acta Obstet Gynecol Scand 1988;67:315–8.
23. Zetterstrom J, Lopez A, Anzen B, Norman M, Holmstrom B, Mellgren A. Anal sphincter tears at vaginal delivery: Risk factors and clinical outcome of primary repair. Obstet Gynecol 1999;94:21–8.