Peripartum hysterectomy can be a life-saving procedure for women with obstetric hemorrhage. In the United States, it is estimated that peripartum hysterectomies are performed in approximately 0.08% of all deliveries.1,2 Cesarean delivery is the most important risk factor for peripartum hysterectomy; those women who undergo abdominal delivery are more than six times more likely to require hysterectomy than are patients who undergo vaginal delivery.1 The risk of peripartum hysterectomy increases with the number of prior cesarean deliveries.3
Among women who undergo peripartum hysterectomy, the most common indication for the procedure is obstetric hemorrhage. In most series, placenta accreta and uterine atony are cited as the most frequent inciting factors for hysterectomy.4–8 A large study from the United Kingdom noted that more than half of the peripartum hysterectomies performed were for uterine atony, whereas 38% were secondary to placenta accreta.4 Other reported indications for peripartum hysterectomy include uterine rupture, extension of a uterine incision, leiomyoma, infection, genital lacerations, and cervical cancer.4–8
When peripartum hysterectomy is required, it is often an emergent situation, frequently in the background of substantial surgical bleeding. These factors, along with the large size of the gravid uterus, lead to substantial perioperative morbidity and mortality for the procedure. Most of the previously reported data on peripartum hysterectomy are from small series of patients. We performed a population-based analysis to examine the morbidity and mortality of peripartum hysterectomy in comparison to nonobstetric hysterectomy.
MATERIALS AND METHODS
Data from the Nationwide Inpatient Sample were used. The Nationwide Inpatient Sample is a national database maintained by the Agency for Healthcare Research and Quality. The Nationwide Inpatient Sample contains a random sample of approximately 20% of discharges from hospitals within the United States. The sampling frame for the Nationwide Inpatient Sample includes nonfederal, general, and specialty-specific hospitals throughout the United States. Sampled hospitals include both academic and community facilities. The sampling scheme represents approximately 90% of hospitals in United States. The Nationwide Inpatient Sample is the largest all-payer inpatient care database; in 2007, Nationwide Inpatient Sample included 8 million hospital stays from 40 states.9 We analyzed data from 1998 to 2007. The study was approved by the Columbia University Institutional Review Board.
International Classification of Diseases, 9th Revision, Clinical Modification procedure and diagnosis codes were used to identify all women younger than 50 years of age who underwent a peripartum or nonobstetric hysterectomy. The peripartum hysterectomy cohort included patients who underwent cesarean delivery (74.x, 669.70, 669.71) in combination with either a total abdominal hysterectomy (68.4x) or subtotal abdominal hysterectomy (68.3x). The nonobstetric hysterectomy cohort included women younger than 50 years of age who underwent either a total abdominal (68.4x) or subtotal abdominal (68.3x) hysterectomy. Patients with a concomitant diagnosis of an invasive malignancy were excluded from the analysis. Patients who underwent a vaginal or laparoscopic hysterectomy were excluded.
We examined outcomes consistent with prior studies and known complications of hysterectomy. Complications were classified into the following groups: 1) intraoperative complications (bladder injury, ureteral injury, intestinal injury, vascular injury, other operative injury), 2) perioperative surgical complications (reoperation, postoperative hemorrhage, wound complication, venous thromboembolism), and 3) postoperative medical complications (cardiovascular, pulmonary, gastrointestinal, renal, infectious). We calculated the percentages of patients who required blood transfusion. Length of stay was calculated by subtracting the admission day from the date of discharge. Perioperative death was defined as death during the hospitalization in which the patient underwent hysterectomy.
Control variables included demographic, clinical, and hospital characteristics. Age was stratified as younger than 25, 25–34, and 35–50 years. Race was categorized as white, African American, Hispanic, and other. Each patient's insurance status, household income, and year of diagnosis were noted. The hospitals in which patients were treated were characterized based on location (urban, rural), region of the country (northeast, midwest, west, south), size (small, medium, large), and teaching status (teaching, nonteaching). The surgical diagnosis of patients who underwent peripartum hysterectomy was classified into the following: placenta accreta (666.0x, 667.0x), uterine atony (666.1x), uterine rupture (665.0x, 665.1x), hysterotomy extension (665.3x, 665.4x), delayed hemorrhage (666.2x), and leiomyoma (218.x, 219.x). Surgical indications were not mutually exclusive and patients may have had multiple codes. Risk adjustment was performed using the Deyo index.10
Univariable analysis was done to provide descriptive information comparing women who underwent peripartum and nonobstetric hysterectomy. Categorical variables were compared using chi square analysis and Student's t test used to compare continuous variables.
Logistic regression was used to determine independent predictors of operative complications and mortality. We developed models for each of the individual complications. All of the predictor variables described were included in the models. Separate logistic regression models were constructed for women who underwent peripartum hysterectomy to compare total and subtotal hysterectomy and to compare hysterectomy for placenta accreta and uterine atony. Information on race was missing from 29% of patients. To account for the influence of these patients, all analyses were performed with patients missing race data as a separate category. All analyses were performed with SAS 9.1.3 (SAS Institute Inc, Cary, NC).
A total of 4,967 women who underwent peripartum hysterectomy and 578,179 patients who had a nonobstetric hysterectomy were identified. The demographic and clinical characteristics of the cohort are displayed in Table 1. Those who underwent peripartum hysterectomy were younger, more often nonwhite, more likely to have noncommercial insurance, and more often reported a low household income (P<.001 for all). Compared with patients who underwent nonobstetric hysterectomy, women who had a peripartum hysterectomy were more often treated at urban hospitals (92% compared with 84%) and at large facilities (67% compared with 59%) (P<.001). Thirty-four percent of peripartum hysterectomies were subtotal compared with 9% of nonobstetric procedures (P<.001). The most common indications for peripartum hysterectomy were placenta accreta (36%) and uterine atony (31%).
Morbidity and mortality are shown in Table 2. Both bladder (9% compared with 1%) and ureteral (0.7% compared with 0.1%) injuries were more common for peripartum hysterectomy (P<.001). In adjusted logistic regression models bladder, injuries were nine times more common (odds ratio [OR] 9.02; 95% confidence interval [CI] 8.10–10.05) and ureteral injuries five times more common (OR 5.04; 95% CI 3.47–7.32) for peripartum hysterectomy. There were no difference in the rates of intestinal or vascular injuries between peripartum and nonobstetric hysterectomy. Rates of reoperation (4% compared with 0.5%), postoperative hemorrhage (5% compared with 2%), wound complications (10% compared with 3%), and venous thromboembolism (1% compared with 0.7%) were all higher in women who underwent peripartum hysterectomy. Likewise, perioperative cardiovascular, pulmonary, gastrointestinal, renal, and infections morbidities were higher for peripartum hysterectomy (P<.001 for all).
Forty-six percent (46%) of women who underwent peripartum hysterectomy received a transfusion compared with 4% of nonobstetric hysterectomies (P<.001). The mean length of stay for women who underwent peripartum hysterectomy was 8.7 days compared with 2.9 days for nonobstetric hysterectomy (P<.001). The perioperative mortality for peripartum hysterectomy was 1% compared with 0.04% for nonobstetric hysterectomy (P<.001). In multivariable analysis, the odds ratio for death for peripartum compared with nonobstetric hysterectomy was 14.4 (95% CI 9.84–20.98).
Table 3 compares the morbidity and mortality of peripartum hysterectomy for placenta accreta and uterine atony. Women with placenta accreta were more likely to experience bladder (15% compared with 5%, P<.001; OR 3.61; 95% CI 2.74–4.76), ureteral (1% compared with 0.3%, P<.02; OR 2.87; 95% CI 1.03–7.97), and other intraoperative injuries (13% compared with 7%, P<.001; OR 2.01; 95% CI 1.57–2.58). In contrast, patients undergoing hysterectomy for placenta accreta were less likely to require reoperation (2% compared with 6%, P<.001), experience a postoperative hemorrhage (3% compared with 6%, P<.001), or have a wound complication (7% compared with 10%, P=.01) than those with uterine atony. The rate of transfusion was lower for placenta accreta than for uterine atony (49% compared with 55%, P<.001). Perioperative mortality was 0.6% for hysterectomy for placenta accreta and 0.8% for hysterectomy for atony (P=.34).
Outcomes of total and subtotal hysterectomy were then compared (Table 4). Intraoperative injuries, including bladder damage (10% compared with 7%, P<.001; OR 1.42; 95% CI 1.14–1.77) and other operative injuries (10% compared with 8%, P<.001; OR 1.30; 95% CI 1.06–1.61), were more common for total hysterectomy. In contrast, reoperation was required more often in patients who underwent subtotal peripartum hysterectomy (5% compared with 4%, P=.02, OR 0.69, 95% CI 0.52–0.93). Perioperative mortality was 0.8% for total compared with 1.4% for subtotal peripartum hysterectomy (P=.04, OR 0.53, 95% CI 0.29–0.94).
Peripartum hysterectomy is accompanied by substantial morbidity and mortality. Compared with nonobstetric hysterectomy, the procedure is associated with increased rates of both intraoperative and postoperative complications. The mortality of peripartum hysterectomy is more than 25 times that of hysterectomy performed outside of pregnancy.
Our nationwide estimates of intraoperative injuries for peripartum and nonobstetric hysterectomy are in line with those reported in prior small studies.3–7,11–15 Among women who underwent peripartum hysterectomy, bladder damage, ureteral injuries, and other intraoperative injuries were all more common than in patients who underwent nonobstetric hysterectomy. We noted bladder complications in 9% of peripartum hysterectomies and ureteral injuries in 0.7%. The reported rates of bladder and ureteral injuries for peripartum hysterectomy range from 6% to 29% and 0% to 7%, respectively.3–7,11 In a large series of patients who underwent abdominal hysterectomy outside of pregnancy, cystotomies were encountered in 0.5% and ureteral injuries in 0.2%.12 Although we did not find differences in the incidence of major vascular or intestinal injuries, the rates of reoperation, postoperative hemorrhage, wound complications, and venous thromboembolic events were all higher in the peripartum hysterectomy cohort. Compared with nonobstetric hysterectomy, women who underwent a peripartum procedure were nearly eight times more likely to require surgical reexploration and almost three times as likely to develop a wound complication.
We also noted that patients who underwent a peripartum hysterectomy were at significantly higher risk for the development of postoperative medical complications. Cardiovascular, pulmonary, gastrointestinal, renal, and infectious complications were all greater in the women who underwent peripartum hysterectomy. Infectious morbidity was seen in more than 12% of peripartum patients, more than three times the rated noted in women who underwent nonobstetric hysterectomy. In a series of more than 37,000 women who underwent nonobstetric hysterectomy for benign indications, 1% of patients experienced severe postoperative complications, whereas 8% reported other, nonsevere postoperative complications.13 It is difficult to determine postoperative morbidity from previous reports of women who underwent peripartum hysterectomy. Given the often emergent nature of peripartum hysterectomy and the difficulty of the surgery, the morbidity reported is not surprising.
We found that the patterns of complications for peripartum hysterectomy differed between women with placenta accreta and those with uterine atony. As would be expected, bladder and ureteral injuries were more frequent in women with placenta accreta. Somewhat surprisingly, we noted that reoperation, postoperative hemorrhage, and wound complications were more common in women with uterine atony. Likewise, both cardiovascular and pulmonary complications were seen more often in cases of uterine atony. Why these complications are more common in women with uterine atony and without placenta accreta is not intuitively clear.
The decision to perform a subtotal compared with total hysterectomy is often debated in gynecology.16–19 A Cochrane review found no evidence for a difference in the rates of incontinence, constipation, or sexual function between total and subtotal hysterectomy.16 However, subtotal hysterectomy has been associated with shorter lengths of stay and reduced intraoperative bleeding.16 Small studies of women undergoing peripartum hysterectomy have found no difference in morbidity and operative times between the two procedures.4,5,20 We noted that subtotal hysterectomy was performed more frequently in women who underwent a peripartum procedure (34% for peripartum hysterectomy compared with 9% for nonobstetric hysterectomy). Although intraoperative injuries are higher for total hysterectomy, rates of reoperation were greater when subtotal peripartum hysterectomy was undertaken. These complications should be weighed and the choice between subtotal and total hysterectomy individualized in patients.
Although our analysis benefits from the inclusion of a large number of cases and control subjects, several limitations must be acknowledged. The primary purpose of administrative data is billing. As such, it is often difficult to extract detailed clinical information on outcomes. Although we attempted to meticulously select codes for the majority of procedures and complications, it is likely that some women experienced complications that were not recorded.21 It is impossible to distinguish between preexisting diseases and complications. However, given the young age of the women we included, it is unlikely that a significant number of patients had major medical comorbidities. Given the blood loss associated with cesarean delivery, it is likely that a portion of the women who underwent peripartum hysterectomy were transfused before the hysterectomy. Likewise, some of the perioperative complications encountered by women who underwent peripartum hysterectomy may have been the result of the pregnancy and not the hysterectomy. Lastly, like with any observational study, we cannot exclude the possibility that unmeasured biases affected our findings.
Our data provide important national estimates of the morbidity of peripartum hysterectomy and we believe this will prove useful in counseling patients at risk for requiring the procedure. A number of patient safety initiatives, including standardized protocols and obstetric rapid response teams, have been developed to decrease the mortality of obstetric hemorrhage.22–25 Obstetricians should be aware of the morbidity associated with peripartum hysterectomy and ensure that hospital resources and surgical assistants are available for patients who may undergo the procedure.
Peripartum hysterectomy is associated with substantial morbidity and mortality. Compared with nonobstetric hysterectomy, intraoperative, perioperative, and postoperative complications are more frequent. Among women who undergo peripartum hysterectomy, the morbidity associated with the procedure differs based on the indication (accreta compared with atony) and type (total compared with subtotal) of procedure performed. Patients at risk for requiring a peripartum hysterectomy should be counseled regarding the sequelae of the procedure.
1. Whiteman MK, Kuklina E, Hillis SD, Jamieson DJ, Meikle SF, Posner SF, et al. Incidence and determinants of peripartum hysterectomy. Obstet Gynecol 2006;108:1486–92.
2. Stanco LM, Schrimmer DB, Paul RH, Mishell DR Jr. Emergency peripartum hysterectomy and associated risk factors. Am J Obstet Gynecol 1993;168:879–83.
3. Silver RM, Landon MB, Rouse DJ, Leveno KJ, Spong CY, Thom EA, et al. Maternal morbidity associated with multiple repeat cesarean deliveries. Obstet Gynecol 2006;107:1226–32.
4. Knight M; UKOSS. Peripartum hysterectomy in the UK: management and outcomes of the associated haemorrhage. BJOG 2007;114:1380–7.
5. Kastner ES, Figueroa R, Garry D, Maulik D. Emergency peripartum hysterectomy: experience at a community teaching hospital. Obstet Gynecol 2002;99:971–5.
6. 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.
7. Shellhaas CS, Gilbert S, Landon MB, Varner MW, Leveno KJ, Hauth JC, et al. The frequency and complication rates of hysterectomy accompanying cesarean delivery. Obstet Gynecol 2009;114:224–9.
8. Glaze S, Ekwalanga P, Roberts G, Lange I, Birch C, Rosengarten A, et al. Peripartum hysterectomy: 1999 to 2006. Obstet Gynecol 2008;111:732–8.
9. Nationwide Inpatient Sample. Available at: http://www.hcup-us.ahrq.gov/nisoverview.jsp
. Retrieved November 6, 2009.
10. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45:613–9.
11. Eller AG, Porter TF, Soisson P, Silver RM. Optimal management strategies for placenta accreta. BJOG 2009;116:648–54.
12. Makinen J, Johansson J, Tomás C, Tomás E, Heinonen PK, Laatikainen T, et al. Morbidity of 10 110 hysterectomies by type of approach. Hum Reprod 2001;16:1473–8.
13. McPherson K, Metcalfe MA, Herbert A, Maresh M, Casbard A, Hargreaves J, et al. Severe complications of hysterectomy: the VALUE study. BJOG 2004;111:688–94.
14. Garry R, Fountain J, Mason S, Hawe J, Napp V, Abbott J, et al. The eVALuate study: two parallel randomised trials, one comparing laparoscopic with abdominal hysterectomy, the other comparing laparoscopic with vaginal hysterectomy. BMJ 2004;328:129.
15. Kafy S, Huang JY, Al-Sunaidi M, Wiener D, Tulandi T. Audit of morbidity and mortality rates of 1792 hysterectomies. J Minim Invasive Gynecol 2006;13:55–9.
16. Lethaby A, Ivanova V, Johnson NP. Total versus subtotal hysterectomy for benign gynaecological conditions. Cochrane Database Syst Rev 2006;2:CD004993.
17. Thakar R, Ayers S, Clarkson P, Stanton S, Manyonda I. Outcomes after total versus subtotal abdominal hysterectomy. N Engl J Med 2002;347:1318–25.
18. Gimbel H, Zobbe V, Andersen BM, Filtenborg T, Gluud C, Tabor A. Randomised controlled trial of total compared with subtotal hysterectomy with one-year follow up results. BJOG 2003;110:1088–98.
19. Kuppermann M, Summitt RL Jr, Varner RE, McNeeley SG, Goodman-Gruen D, Learman LA, et al. Sexual functioning after total compared with supracervical hysterectomy: a randomized trial. Obstet Gynecol 2005;105:1309–18.
20. Yucel O, Ozdemir I, Yucel N, Somunkiran A. Emergency peripartum hysterectomy: a 9-year review. Arch Gynecol Obstet 2006;274:84–7.
21. Heisler CA, Melton LJ 3rd, Weaver AL, Gebhart JB. Determining perioperative complications associated with vaginal hysterectomy: code classification versus chart review. J Am Coll Surg 2009;209:119–22.
22. Skupski DW, Lowenwirt IP, Weinbaum FI, Brodsky D, Danek M, Eglinton GS. Improving hospital systems for the care of women with major obstetric hemorrhage. Obstet Gynecol 2006;107:977–83.
23. Gosman GG, Baldisseri MR, Stein KL, Nelson TA, Pedaline SH, Waters JH, et al. Introduction of an obstetric-specific medical emergency team for obstetric crises: implementation and experience. Am J Obstet Gynecol 2008;198:367.e1–7.
24. Homer C, Clements V, McDonnell N, Peek M, Sullivan E. Maternal mortality: what can we learn from stories of postpartum haemorrhage? Women Birth 2009;22:97–104.
25. Cameron CA, Roberts CL, Bell J, Fischer W. Getting an evidence-based post-partum haemorrhage policy into practice. Aust N Z J Obstet Gynaecol 2007;47:169–75.