Peripartum hysterectomy, a hysterectomy performed at the time of delivery or in the immediate postpartum period, is one of the most severe complications in obstetrics and is related to significant maternal mortality and morbidity.1–4 Typically reserved for situations in which severe obstetric hemorrhage fails to respond to conservative treatment, peripartum hysterectomy is associated with severe blood loss, risk of transfusion, intraoperative complications, and significant postoperative morbidity. It is important to estimate national incidence rates and trends for peripartum hysterectomy to inform obstetric practice and to assess risks and complications of pregnancy. Hospital-based retrospective case-reviews in the United States report incidence rates for peripartum hysterectomy ranging from 0.6 to 2.28 per 1,000 births.1–3,5–8 However, these studies are unable to provide reliable national incidence estimates because they were conducted in single institutions with small samples. Furthermore, their findings may be influenced by patient characteristics or practitioner practice patterns for hysterectomy at individual institutions.
Several studies examined pregnancy-related factors associated with risk for peripartum hysterectomy. Generally, these studies report a greater than 10-fold higher incidence of peripartum hysterectomy among women who have previously delivered by cesarean section than among those who have not.2,6–9 This finding deserves closer examination, given the increasing rate of cesarean deliveries in the United States, even among low-risk women.10 However, few studies examined the effect of previous cesarean deliveries within the context of the current mode of delivery. Another reported risk factor for peripartum hysterectomy is multiple births,5 the rate of which is also increasing in the United States.11 It is important to note, however, that other studies were small and limited in the ability to examine risk factors for the procedure while adequately controlling for potential confounding factors. The objectives of this study were to estimate the national incidence of peripartum hysterectomy, to describe recent national trends in incidence, and to examine risk factors for the procedure by using a large nationally representative database of inpatient hospitalizations.
MATERIALS AND METHODS
Data were from the Nationwide Inpatient Sample of the Healthcare Cost and Utilization Project for 1998 through 2003. The Healthcare Cost and Utilization Project is a family of health care databases and related software tools developed through a federal–state–industry partnership, sponsored by the Agency for Healthcare Research and Quality, in which state partners contribute hospital discharge data to the Healthcare Cost and Utilization Project. The Nationwide Inpatient Sample is the largest all-payer inpatient care database in the United States, containing data on approximately 7 million hospital stays from 800–1,000 hospitals per year to approximate a 20% stratified sample of U.S. hospitals.12 The number of states contributing data to the Healthcare Cost and Utilization Project increased from 22 in 1998 to 37 in 2003.
The Nationwide Inpatient Sample is a stratified probability sample of hospitals in the United States based on a sampling frame using five strata: location (rural or urban), hospital size (based on number of beds), region of the country, teaching status, and type of ownership. The universe of U.S. community hospitals includes all those open during any part of the calendar year and designated as community hospitals in the American Hospital Association Annual Survey of Hospitals.13 The American Hospital Association defines community hospitals as all nonfederal short-term (average length of stay less than 30 days) general and specialty hospitals whose facilities are open to the public. Data are retained for 100% of discharges for each sampled hospital. Inpatient-stay records in the Nationwide Inpatient Sample include information on patient characteristics, medical diagnoses, and surgical procedures. The Sample also contains hospital-level data from the American Hospital Association Annual Survey of Hospitals, including hospital region (northeast, midwest, south, or west), hospital location (rural or urban), hospital teaching status, and number of beds.
To maximize completeness of ascertaining delivery-related discharges, we used several sources of information in the Nationwide Inpatient Sample, including Diagnosis Related Groups (DRG); International Classification of Diseases, 9th revision, Clinical Modification (ICD-9-CM) codes; and Clinical Classification Software for ICD-9-CM codes. Clinical Classification Software is a tool developed as part of the Healthcare Cost and Utilization Project for clustering patient diagnoses and procedures into a manageable number of clinically meaningful categories.14 Delivery-related discharges were defined as those with any of the following codes: DRG delivery codes; ICD-9-CM codes for maternal outcome of delivery or normal delivery; ICD-9-CM codes for pregnancy or delivery complications with a fifth digit of 1 or 2 (indicates a delivery); and Clinical Classification Software codes for obstetric procedures indicating a delivery (episiotomy, cesarean delivery, forceps, vacuum, or breech delivery, artificial rupture of membranes to assist delivery, other procedures to assist delivery, repair of current obstetric laceration. We excluded a record if it included failed labor along with no other indication of delivery besides artificial rupture of membranes, other surgical inductions of labor, or medical induction of labor. Analyses were limited to delivery-related records of women aged 15 years or older. Peripartum hysterectomy was defined as a hysterectomy and delivery occurring during the same hospitalization. Hysterectomies were identified by using Clinical Classification Software.
We used SAS-callable SUDAAN version 9.1 (RTI, Research Triangle Park, NC) to account for the complex sampling design in Nationwide Inpatient Sample. Rates of peripartum hysterectomy were calculated per 1,000 deliveries overall and within subgroups of interest defined by age, delivery type, multiple births, insurance status, and hospital characteristics (region, location, teaching status, and size). Cesarean deliveries and repeat cesarean deliveries were identified by using Clinical Classification Software. Multiple births, defined as twin or higher-order gestation, were identified by ICD-9-CM code 651. Race was not examined because a large proportion of records did not have race information.
We used logistic regression to estimate adjusted odds ratios (ORs) as estimators of relative risk of peripartum hysterectomy by delivery, demographic, and hospital characteristics; 95% confidence intervals (CIs) were calculated for all OR point estimates. The risk of peripartum hysterectomy associated with a previous cesarean delivery varied by the mode of the current delivery (P for interaction<.001); therefore, type of delivery was classified according to current delivery type in combination with history of cesarean delivery. We used χ2 tests to compare proportions. Trends over time were assessed by entering year as a single term with equally spaced category scores into logistic regression models. The prevalence of selected conditions (identified by ICD-9-CM codes) was examined among women who underwent peripartum hysterectomies including placenta previa, uterine rupture, uterine atony, placental abruption, and placenta accreta. Because the Nationwide Inpatient Sample data do not contain personal identifiers, this research was determined by the Centers for Disease Control and Prevention to be exempt research not requiring institutional review board review.
From 1998 through 2003 an estimated 18,339 peripartum hysterectomies were performed in the United States, translating to a rate of 0.77 per 1,000 deliveries. The rate did not change significantly over the time period, from 0.73 per 1,000 deliveries in 1998 to 0.82 per 1,000 deliveries in 2003 (P for trend=0.18) (Table 1). The median length of stay for women who underwent peripartum hysterectomies was 4.3 days (4.3 days for both those with cesarean deliveries and for those with vaginal deliveries), compared with 1.7 days for women who delivered and did not undergo peripartum hysterectomies (2.8 days for those with cesarean deliveries and 1.5 days for those with vaginal deliveries).
The rate of peripartum hysterectomy increased with advancing maternal age, from 0.23 per 1,000 deliveries for women aged 15–24 years to 3.81 per 1,000 deliveries for women aged 40 or older (Table 2). Rates within age groups were stable during the period. The rate of peripartum hysterectomy varied by delivery type, with the lowest rate being for vaginal births without prior cesarean delivery (0.25 per 1,000 deliveries). In comparison, rates were higher among vaginal births after cesarean delivery (VBAC) 0.88 per 1,000 deliveries), primary cesarean deliveries (1.86 per 1,000 deliveries), and repeat cesarean deliveries (2.99 per 1,000 deliveries). Rates were also higher among multiple births (2.22 per 1,000 deliveries) compared with singleton births (0.75 per 1,000 deliveries). Among multiple births, there was a statistically significant increase in the rate over time from 1.28 per 1,000 deliveries in 1998 to 2.54 per 1,000 deliveries in 2003 (P for trend=.02). Rates tended to be higher in hospitals in the south than in other areas, in urban areas than in rural areas, in teaching hospitals than in nonteaching hospitals, and in large or medium-sized hospitals than in small hospitals. Rates within these hospital characteristics did not vary significantly over time.
The adjusted associations of peripartum hysterectomy with delivery, demographic, and hospital characteristics are shown in Table 3. The risk of peripartum hysterectomy was significantly greater among women who underwent repeat cesarean deliveries, compared with those who had vaginal deliveries without a previous cesarean delivery (OR 8.90, 95% CI 8.09–9.79). Women undergoing primary cesarean deliveries were also at significantly higher risk for peripartum hysterectomy than were women undergoing vaginal deliveries without a previous cesarean delivery (OR 6.54, 95% CI 5.95–7.18). Vaginal birth after cesarean was also associated with an increased risk, but of a smaller magnitude (OR 2.70, 95% CI 2.23–3.26). Multiple births were associated with an increased risk compared with singleton births (OR 1.41, 95% CI 1.16–1.71). Risk increased with increasing maternal age; compared with women aged 15 to 24, the OR for peripartum hysterectomy for women aged 40 or older was 11.83 (95% CI 10.19–13.73). Risk tended to be greater for women delivering in teaching hospitals than for women delivering in nonteaching hospitals, in hospitals located in the south, midwest, or west than in the northeast, or in large or medium hospitals than small hospitals. In addition, women with public insurance were at higher risk than those with private insurance.
The prevalence of selected conditions among women undergoing peripartum hysterectomies by delivery type is shown in Table 4. A diagnosis of hemorrhage was listed in 69.6% of peripartum hysterectomy discharge records. After hemorrhage, the most commonly listed condition overall was placenta previa, with a prevalence of 25.2%. The prevalence of the selected conditions varied by type of delivery. The prevalence of placenta previa was higher for those who underwent peripartum hysterectomies who delivered by cesarean section and was highest for women who underwent repeat cesarean deliveries (39.4%). Placenta accreta was also most prevalent among women with repeat cesarean deliveries (7.7%). Uterine rupture was most prevalent among women who underwent peripartum hysterectomy and a VBAC (25.8%). Both uterine atony and placental abruption were most prevalent among women who underwent a primary cesarean delivery. Approximately 16% of all peripartum hysterectomy discharge records contained none of the selected conditions.
Although some8,15 but not all7 studies report an increased risk for peripartum hysterectomy for women with previous cesarean deliveries, few examined previous cesarean deliveries within the context of the current mode of delivery. We found that the association between a previous cesarean delivery and peripartum hysterectomy varied significantly by the current mode of delivery, with the highest risk being associated with repeat cesarean deliveries. Similarly, a large study in the Netherlands found that both previous cesarean delivery and current cesarean delivery were associated with an increased risk for peripartum hysterectomy, with the highest risk being for those who had a repeat cesarean delivery.9 The most likely explanation for the increased risk for peripartum hysterectomy for women who underwent repeat cesarean deliveries is the association between cesarean delivery and an increased risk for placenta previa and placenta accreta in subsequent pregnancies, a risk that increases with the number of previous cesarean deliveries.9,16 Consistent with this, we found that among women who undergo peripartum hysterectomies, the prevalence of placenta previa and placenta accreta was highest for women who undergo repeat cesarean deliveries. Although the exact mechanism is unknown, it is hypothesized that uterine scarring from previous cesarean deliveries prevents normal implantation of the placenta.
This study found that women who underwent a primary cesarean delivery were also at a higher risk for peripartum hysterectomy than were those who underwent a vaginal birth without a previous cesarean delivery. This confirms the findings of other studies suggesting that cesarean delivery is associated with emergent peripartum hysterectomy2,6–9,15 even after women ineligible for vaginal delivery were excluded.7 Women who underwent VBAC were also at higher risk for peripartum hysterectomy than were those with a vaginal delivery without a previous cesarean delivery, although the magnitude of risk was not as large as that associated with repeat or primary cesarean delivery. The increased risk associated with VBAC is likely related to uterine rupture, which was most prevalent among VBAC peripartum hysterectomy cases in this study, consistent with reports suggesting VBAC is associated with an increased risk of uterine rupture.17 Some cases of uterine rupture occurring during a VBAC attempt may lead to an emergency cesarean delivery and would therefore be classified as a repeat cesarean delivery. We are unable to distinguish which peripartum hysterectomies among women undergoing repeat cesarean deliveries involved a VBAC attempt. Although a VBAC attempt may reduce the risk for peripartum hysterectomy in future pregnancies by possibly minimizing the need for additional cesarean deliveries, it may increase risk of hysterectomy in the current delivery because of the risk of uterine rupture.
Multiple factors may play a role in the increased risk for peripartum hysterectomy associated with cesarean delivery. The increased risk may not result from the procedure itself, but from the labor complications or other factors leading to the decision to perform a cesarean delivery. However, Kacmar et al7 found that cesarean delivery was associated with an increased risk for peripartum hysterectomy even after women with absolute indications for cesarean delivery before delivery (such placenta previa, 3 or more prior cesarean deliveries, or triplets or higher order multiple gestations) were excluded. Cesarean delivery may be associated with a higher risk for peripartum hysterectomy than vaginal delivery, because the uterus is more readily available for removal. With vaginal delivery, medical instead of surgical management of obstetric emergencies takes precedence because the uterus is not readily available for removal.7 Cesarean delivery may therefore increase risk for peripartum hysterectomy directly, or indirectly by increasing risk for indications for peripartum hysterectomy (placenta previa, placenta accrete, uterine rupture) in subsequent pregnancies.
We found that multiple births were associated with a moderately greater risk for peripartum hysterectomy than were singleton births. Similarly, a recent study also found a more pronounced increase in risk for higher-order gestations.5 We found that the rate of peripartum hysterectomy within multiple births increased significantly from 1998 to 2003. This may be a reflection of the increasing proportion of multiple births attributable to assisted reproductive technologies.18 A recent study of singleton pregnancies found that patients using in vitro fertilization were more likely to have placental abruption and placenta previa,19 which could in turn increase risk for peripartum hysterectomy.
To our knowledge, this is the first report to provide a population-based estimate of the incidence of peripartum hysterectomy in the United States (MEDLINE; 1966 to June 2006; English language; search terms: “hysterectomy,” “obstetric labor complications,” and “postpartum hemorrhage”). Our incidence estimate, 0.77 per 1,000 births during 1998–2003, is similar to those of some hospital-based retrospective case reviews,6,7 but is somewhat lower than estimated rates from other case reviews.1–3,5,8 We found that peripartum hysterectomy rates tended to be higher in large hospitals and in teaching hospitals, presumably because they are more likely to handle a greater proportion of complicated deliveries. Given that the case-review studies included cases only from a single hospital, they were probably influenced by the characteristics of the hospital's patient population as well as by practitioner practice patterns for hysterectomy. Furthermore, the cesarean delivery rate in a hospital may also influence the estimated incidence of peripartum hysterectomy. For example, the cesarean delivery rate in our study was 24.4%; whereas in a 5-year university hospital study, the cesarean delivery rate was higher (29.9%), as was the peripartum hysterectomy rate (2.7 per 1,000).2
We were limited in this analysis by several factors inherent in hospital discharge data. Our data depend on the accuracy of the diagnoses listed on the hospital discharge summary. There are diagnoses examined in this study, such as uterine rupture, for which ICD-9-CM codes lack specificity and consistency.20 In addition, without access to medical records, we were unable to classify precise indications for peripartum hysterectomy. However, our results are in agreement with recent studies suggesting abnormal placentation is emerging as the most common indication for peripartum hysterectomy, replacing uterine atony.1,4 This likely reflects both increases in the number of women delivering by cesarean as well as improvements in the treatment of uterine atony with uterotonic agents. We were unable to assess whether the peripartum hysterectomies in this study were emergent or elective, but recent studies suggest indications for the procedure are almost exclusively emergent complications.4,8 We also lacked information on reproductive history factors such as parity.
This report adds to the evidence suggesting that cesarean delivery is an important risk factor for peripartum hysterectomy in subsequent births. Risk is highest for those who underwent repeat cesarean deliveries. These findings deserve consideration in the decision to proceed with a primary cesarean delivery, especially when future pregnancies are desired. Furthermore, the larger magnitude of risk associated with repeat cesarean deliveries compared with VBAC should be considered in the management and counseling of women with previous cesarean deliveries within the context of future reproductive plans. We did not have information on the number of previous cesarean deliveries. Future studies should consider this factor, because it is likely that risk for peripartum hysterectomy increases with the number of previous cesarean deliveries. The risk for peripartum hysterectomy associated with cesarean delivery on maternal request may also be an important area for future research, given the growing interest and debate in that area.21 In light of the increasing rates of cesarean deliveries and multiple births in the United States, the rate of peripartum hysterectomy should continue to be monitored.
1. Kastner ES, Figueroa R, Garry D, Maulik D. Emergency peripartum hysterectomy: experience at a community teaching hospital. Obstet Gynecol 2002;99:971–5.
2. Bakshi S, Meyer BA. Indications for and outcomes of emergency peripartum hysterectomy: a five-year review. J Reprod Med 2000;45:733–7.
3. Zelop CM, Harlow BL, Frigoletto FD Jr, Safon LE, Saltzman DH. Emergency peripartum hysterectomy. Am J Obstet Gynecol 1993;168:1443–8.
4. Castaneda S, Karrison T, Cibils LA. Peripartum hysterectomy. J Perinat Med 2000;28:472–81.
5. Francois K, Ortiz J, Harris C, Foley MR, Elliott JP. Is peripartum hysterectomy more common in multiple gestations? Obstet Gynecol 2005;105:1369–72.
6. Forna F, Miles AM, Jamieson DJ. Emergency peripartum hysterectomy: a comparison of cesarean and postpartum hysterectomy. Am J Obstet Gynecol 2004;190:1440–4.
7. Kacmar J, Bhimani L, Boyd M, Shah-Hosseini R, Peipert J. Route of delivery as a risk factor for emergent peripartum hysterectomy: a case-control study. Obstet Gynecol 2003;102:141–45.
8. Stanco LM, Schrimmer DB, Paul RH, Mishell DR Jr. Emergency peripartum hysterectomy and associated risk factors. Am J Obstet Gynecol 1993;168:879–83.
9. Kwee A, Bots ML, Visser GH, Bruinse HW. Emergency peripartum hysterectomy: A prospective study in The Netherlands. Eur J Obstet Gynecol Reprod Biol 2006;124:187–92.
10. Menacker F. Trends in cesarean rates for first births and repeat cesarean rates for low-risk women: United States, 1990-2003. Natl Vital Stat Rep 2005;54:1–8.
11. Hoyert DL, Mathews TJ, Menacker F, Strobino DM, Guyer B. Annual summary of vital statistics: 2004. Pediatrics 2006;117:168–83.
12. Nationwide Inpatient Sample (NIS). Healthcare Cost and Utilization Project (HCUP). 1998-2003. Rockville, (MD): Agency for Healthcare Research and Quality. Available at: http://www.hcup-us.ahrq.gov/nisoverview.jsp
. Retrieved February 2, 2006.
13. American Hospital Association Hospital Statistics. Chicago (IL): Health Forum LLC; 1995. p. 6–2001.
15. Wingprawat S, Chittacharoen A, Suthutvoravut S. Risk factors for emergency peripartum cesarean hysterectomy. Int J Gynaecol Obstet 2005;90:136–7.
16. Miller DA, Chollet JA, Goodwin TM. Clinical risk factors for placenta previa–placenta accreta. Am J Obstet Gynecol 1997;177:210–4.
17. Macones GA, Peipert J, Nelson DB, Odibo A, Stevens EJ, Stamilio DM, et al. Maternal complications with vaginal birth after cesarean delivery: a multicenter study. Am J Obstet Gynecol 2005;193:1656–62.
18. Reynolds MA, Schieve LA, Martin JA, Jeng G, Macaluso M. Trends in multiple births conceived using assisted reproductive technology, United States, 1997-2000. Pediatrics 2003;111:1159–62.
19. Shevell T, Malone FD, Vidaver J, Porter TF, Luthy DA, Comstock CH, et al. Assisted reproductive technology and pregnancy outcome. Obstet Gynecol 2005;106:1039–45.
20. Centers for Disease Control and Prevention (CDC). Use of hospital discharge data to monitor uterine rupture—Massachusetts, 1990-1997. MMWR Morb Mortal Wkly Rep 2000;49:245–8.
21. Minkoff H, Chervenak FA. Elective primary cesarean delivery. N Engl J Med 2003;348: 946–50.