Each year approximately 600,000 hysterectomies are performed in the United States, making it one of the most common surgical procedures in women.1–3 Although there are several indications for hysterectomy, the majority are performed for benign diseases of the uterus.1,4 At the time of hysterectomy, gynecologic surgeons and patients are often faced with the decision of whether to perform a concomitant prophylactic oophorectomy.
The major benefits of prophylactic oophorectomy are the reduction in the risk of ovarian cancer and the need for oophorectomy in the future.5–7 However, oophorectomy is associated with a number of sequelae. Among premenopausal women, oophorectomy results in surgical menopause and all of its attendant side effects, including mood changes, hot flushes, and vaginal dryness. In addition, several studies now have suggested that oophorectomy increases the long-term risks of coronary heart disease, bone fractures, neurologic changes, and possibly mortality.8–14 Furthermore, the potential benefits of ovarian conservation may not be limited to premenopausal women. Postmenopausal ovaries continue to produce androgens that are peripherally converted to estrogen. There are emerging data that suggest that ovarian preservation may be beneficial even up to 65 years of age.15
Both the American College of Obstetricians and Gynecologists and the Society of Gynecologic Oncology have released guidelines to help inform the decision of whether to perform oophorectomy.16,17 In large part, these guidelines recommend individualized risk assessment based on a women's personal and family history.16,17 Given the uncertainty surrounding the performance of oophorectomy, we performed a population-based analysis to estimate the rates of oophorectomy in women undergoing hysterectomy for benign indications. Additionally, we estimated the influence of patient, physician, and hospital characteristics and between-hospital variation on the decision to undertake prophylactic oophorectomy.
MATERIALS AND METHODS
The Perspective database (Premier) was analyzed. Perspective comprehensively collects demographic and clinical data for inpatient admissions from more than 600 acute-care hospitals across the United States. In 2006, Perspective recorded data of approximately 5.5 million discharges, which represents approximately 15% of all hospitalizations in the United States.18 Perspective has been validated and utilized in a number of outcomes studies.18,19 Study exemption was obtained from the Columbia University Institutional Review Board.
Women 18–64 years of age who underwent hysterectomy from 2000 to 2010 were analyzed. Patients who underwent abdominal (International Classification of Diseases, 9th Revision, Clinical Modification [ICD-9-CM], 68.3, 68.39, 68.4, 68.49, 68.9), vaginal (ICD-9-CM 68.5, 68.59), and laparoscopically assisted (ICD-9-CM 68.31, 68.41, 68.51) hysterectomy were included. For both abdominal and laparoscopic hysterectomy the procedure was further classified as either a total or a subtotal (supracervical) hysterectomy. Patients who underwent surgery or had a code for a gynecologic cancer (180--184.9x) were excluded. Patients were grouped based on whether they underwent oophorectomy at the time of hysterectomy. Prophylactic oophorectomy was considered to have been performed in patients who had a concomitant code for a bilateral oophorectomy or bilateral salpingo-oophorectomy (ICD-9-CM 65.5x and 68.6x). Patients with codes for unilateral oophorectomy were classified as not having undergone a prophylactic oophorectomy.
For the primary analysis, age was stratified by 5-year increments into the following categories: younger than 40, 40–44, 45–49, 50–54, 55–59, and 60–64 years. Other clinical and demographic factors analyzed included year of surgery (2000–2010), race (white, black, and other), marital status (married, single, and unknown), and insurance status (commercial, Medicare, Medicaid, uninsured, and unknown). Based on ICD-9-CM coding, the indications for surgery for each patient were identified and included leiomyomata, endometriosis, abnormal bleeding, benign ovarian neoplasms, and pelvic organ prolapse. These categories were not mutually exclusive. The performance of concomitant gynecologic procedures, including anterior colporrhaphy, posterior colporrhaphy, and incontinence surgery, also was noted. Risk adjustment for medical comorbidities was performed using the Charlson comorbidity index.20 The ICD-9-CM coding to define the Charlson index described by Deyo et al21 was used.
The hospital in which each patient underwent surgery was classified based on location (metropolitan and nonmetropolitan), region of the country (Northeast, Midwest, West, and South), size (fewer than 400 beds, 400–600, beds and more than 600 beds), and teaching status (teaching and nonteaching). Hospitals were further classified based on the case mix of patients who underwent hysterectomy. Hospital characteristics analyzed included percentage of black patients (less than 10%, 10–25%, and more than 25%) and the percentage of Medicaid or uninsured patients (less than 5%, 5–15%, and more than 15%).
Surgical volume for hospitals and surgeons was determined. Because not all surgeons and hospitals contributed procedures for the entire study period, we calculated annualized procedure volumes. The annualized volume was estimated by dividing the total number of patients who underwent a procedure by the number of years a given surgeon or hospital contributed at least one procedure. The volumes were then divided to create three approximately equal tertiles of surgeon (low, less than 14.3 procedures; intermediate, 14.3–28.4 procedures; and high, more than 28.4 procedures per year) and hospital volume (low, less than 220.5; intermediate, 220.5–380.2; and high, more than 380.2 procedures per year).22,23
The association between clinical characteristics, hospital and physician factors, and ovarian conservation were compared using χ2 tests. Multivariable mixed-effects log-Binomial regression models were used to examine the association between patient, procedural, physician, and hospital characteristics and ovarian conservation while controlling for other predictor variables. To account for hospital-level clustering, these models included a random intercept term for the hospital in which the hysterectomy was performed. To examine whether the associations for the entire population differed by patient age, we performed several sensitivity analyses in which the cohort was limited by age (age younger than 45 years, 45–49 years, and 50–54 years of age). To estimate risk ratios (RRs) and 95% confidence intervals (CIs), we fit log-binomial regression models as previously reported.24
To examine the contributions of individual physician, procedural, physician, and hospital characteristics to ovarian conservation, we explored the between-hospital variation using a stepwise approach, similar to the method previously used by Haymart et al.25 Initially, we developed a regression model with no covariates (null model); this model only included the hospital-specific random intercept term. Subsequently, four separate models were developed to examine patient (clinical and demographic factors), procedural (type of hysterectomy, concomitant procedures), physician (volume), and hospital (location, region, size, teaching status, patient mix factors) characteristics. From these models we calculated the amount of between-hospital variation in ovarian conservation that could be attributed to these factors. We then developed a fully adjusted model incorporating all observed patient, procedural, physician, and hospital characteristics. From this model, we estimated the residual intraclass correlation coefficient, which represents the variance attributable to the hospital after adjustment for all other measured characteristics, using the linear threshold model method to partition the variance with a dichotomous outcome.25,26
To further characterize between-hospital variation, we examined hospital-specific rates of ovarian conservation.25 To sample a homogenous population that would be candidates for ovarian conservation, we limited this analysis to patient 45–49 years of age and excluded patients who underwent a vaginal hysterectomy or those with a code for endometriosis or an ovarian neoplasm. The hospital-specific rate of ovarian conservation was calculated using the generalized linear mixed-model described, correcting for patient, procedural, physician, and hospital characteristics. Hospital-specific rates of ovarian conservation were estimated using empirical Bayes predictions for each hospital with a patient matching the selection criteria. Empirical Bayes predictions shrink hospital rates toward the average.25,27 These rates then were plotted by hospital, each ranked based on the hospital's predicted risk. The overall hospital mean and CIs were calculated and the number of hospitals with a rate of ovarian conservation statistically different from the overall group mean was determined. Separate analyses were performed for the entire study period (2000–2010) and for the more recent years of study (2007–2010).
All analyses were performed with SAS 9.2. The generalized linear mixed models were implemented specifically using the GLIMMIX procedure.
A total of 752,045 women who underwent hysterectomy, including 403,073 (53.6%) who had ovarian conservation and 348,972 (46.4%) who underwent bilateral oophorectomy, were identified (Table 1). When stratified by age, the rates of ovarian conservation were as follows: 74.3% for women aged younger than 40 years; 62.7% for women aged 40–44 years; 40.8% for women aged 45–49 years; 25.2% for women aged 50–54 years; 25.5% for women aged 55–59 years; and 31% for women aged 60–64 years. The rate of ovarian conservation increased with more recent year of surgery for all age groups but was greatest for women 45–49 years of age (Fig. 1).
In a multivariable model, age and year of surgery showed the strongest association with ovarian conservation (Table 2). The RR for ovarian conservation for treatment in 2010 was 2.33 (95% CI 2.21–2.46) compared with patients who underwent surgery in 2000 based on a multivariable log-binomial regression model. Ovarian conservation was more common in black than in white women (RR 1.58, 95% CI 1.55–1.60), in single than in married women (RR 1.03, 95% CI 1.01–1.04), and in Medicaid recipients compared with commercial insurance recipients (RR 1.02, 95% CI 1.00–1.05). In contrast, the uninsured compared with those with commercial insurance (RR 0.89, 95% CI 0.86–0.92), residents in nonmetropolitan compared with metropolitan areas (RR 0.80, 95% CI 0.71–0.92), residents outside of the eastern United States, and patients with medical comorbidities were less likely to have ovarian conservation. Compared with women undergoing abdominal hysterectomy, ovarian conservation was more common for all other forms of hysterectomy. Although hospital volume had no direct effect on ovarian conservation, the likelihood of ovarian conservation decreased with physician surgical volume.
In a series of sensitivity analyses stratified by age, these findings were largely unchanged (Table 2). For women younger than 45 years of age, those 45–49 years of age, and those 50–54 years of age, ovarian conservation increased with time. Patients with medical comorbidities, residents outside of the eastern United States, and women who underwent operation on at nonmetropolitan centers were less likely to have ovarian conservation. Likewise, for all age groups, the route of hysterectomy was an important predictor of ovarian conservation; women who underwent abdominal surgery were more likely to undergo oophorectomy.
In a series of models to distinguish the influence of various factors on ovarian conservation, we noted that procedural factors including the indication for surgery and route of hysterectomy had the greatest influence on the decision to retain the ovaries; these factors accounted for 32% of the between-hospital variation in ovarian conservation (Table 3). Hospital characteristics accounted for 10% of the between-hospital variation in ovarian conservation, whereas patient factors and physician characteristics accounted for 5% and 3% of the between-hospital variation in ovarian conservation, respectively. A final model including all of these measured characteristics accounted for 46% of the total variation in the rate of ovarian conservation. These findings suggest that there is substantial residual between-hospital variation (54%) in the rate of ovarian conservation that cannot be explained by observed characteristics.
To further examine the between-hospital variation in the rate of ovarian conservation, we selected women 45–49 years of age after excluding patients who underwent vaginal hysterectomy or those with endometriosis or an ovarian mass. From 2,000 to 10,578 hospitals treated 68,022 women who met these criteria (Fig. 2A). Across these hospitals the mean rate of ovarian conservation was 37.2% (95% CI 34.4–40.1%). Among these hospitals, 15.6% of the centers had a rate of ovarian conservation that was statistically significantly different from the mean rate of conservation (P<.001 based on two-sample binomial test for proportions). When the analysis was limited to 27,520 patients treated at 422 hospitals from 2007 to 2010 after similar exclusions, the mean rate of ovarian conservation across the centers was 47.7% (95% CI 43.5–51.9%) (Fig. 2B). During this time period, 7.8% of the hospitals had a mean ovarian conservation rate that was statistically significantly different than the mean rate (P<.001 based on two-sample binomial test for proportions).
Our findings suggest that the rate of ovarian conservation during hysterectomy for benign indications in women younger than 50 years of age is increasing. Change in the rate of oophorectomy in women 50–65 years of age has been more modest, and the majority of these women still undergo ovarian removal. Although demographic and clinical factors influence the decision to perform oophorectomy, there appears to be substantial between-hospital variation in performance of oophorectomy that is not explained by measurable patient, physician, or hospital characteristics.
The increasing impetus for ovarian conservation stems predominantly from data suggesting that oophorectomy may increase the long-term risk of coronary heart disease and possibly mortality.8,10,15,16 The Nurse's Health Study, a prospective observational study, found that after 24 years of follow-up the risk of coronary heart disease was increased by 17%, whereas total mortality was 12% higher in women who underwent concomitant oophorectomy during hysterectomy.10 A cohort study from Olmsted County, Minnesota, similarly noted increased mortality in women younger than 45 of age who underwent oophorectomy.8 A decision analysis found that the benefits of ovarian conservation outweighed the risks up to age 64.15 However, not all studies have found that ovarian conservation is beneficial.28,29 Data from the Women's Health Initiative, which included older women than did the Nurse's Health Study, found no beneficial effect of ovarian conservation and, likewise, a recent meta-analysis concluded that the available evidence was inconclusive.28,29
Our findings of ovarian conservation in 54% of women are in line with previous studies that have reported ovarian conservation in 46–53% of women.1,2,30,31 Age and year of diagnosis have the strongest association with oophorectomy; younger women and those who underwent operation more recently are less likely to undergo oophorectomy.1,2,30 We noted that the absolute rate of ovarian preservation increased by more than 28% in women 45–49 years of age and by more than 18% in those 50–54 years of age. In contrast, whereas ovarian conservation increased in women 55–64 years of age, it was at a slower rate; similar findings were noted in a study by Novetsky et al.30 Surveys of gynecologists support these findings; many respondents are reticent to recommend ovarian preservation in older women.32,33
A number of patient, procedural, and hospital factors were associated with oophorectomy. Like other studies, we noted that women who underwent vaginal and laparoscopic hysterectomy were less likely to undergo oophorectomy than those who had an abdominal procedure.1,30 This is perhaps driven by not only technical considerations but also the underlying indication for the procedure.1,30 Data describing the influence of race and insurance status on ovarian conservation have been variable.1,30 We found that ovarian conservation was 58% more common in black women and slightly more common among Medicaid recipients. There were also strong regional differences with women in the eastern United States being more likely to have ovarian preservation at the time of hysterectomy.
In addition to patient-related and disease-related characteristics, hospital factors have a strong influence on individual care.34–36 Identifiable hospital characteristics, such as teaching status, size, and location, play an important role in the allocation of care.34,35 However, even after accounting for these quantifiable traits, there appears to be substantial variation, so-called between-hospital variation, in how a given hospital delivers care. One report examining radioactive iodine administration for thyroid cancer noted that even after accounting for patient and hospital characteristics, 29% of the variance in rates of administration were attributable to unexplained hospital characteristics and between-hospital variability.25 A large part of this between-hospital variation likely is attributable to customary institutional practices. We noted substantial between-hospital variation for ovarian conservation.
Our findings should be interpreted in the context of a number of limitations. Using administrative data, we lack information on various important clinical characteristics such as family history, surgical history, including adnexal procedures, and underlying pathology that undoubtedly influenced the decision to perform oophorectomy. We were unable to capture patient and physician preferences that informed the decision to perform oophorectomy. We cannot exclude the possibility that performance of oophorectomy was misclassified in a small number of patients in our analysis. Finally, although our dataset sampled hospitals from across the country, we cannot exclude the possibility that our findings could be different in other centers.
Given the lack of randomized data and the conflicting observational studies of the effects of oophorectomy, there is currently no standardized algorithm to guide performance of oophorectomy. Current recommendations by professional societies endorse individualized risk assessment in women undergoing hysterectomy for benign indications.16,17 Our findings demonstrating the increasing rate of ovarian conservation suggest that the potential benefits of ovarian preservation are being increasingly recognized. Despite increased rates of ovarian conservation, we noted significant disparities in utilization. In addition to patient, physician, and hospital characteristics, there is substantial between-hospital variation in the allocation of ovarian conservation.
1. Jacoby VL, Vittinghoff E, Nakagawa S, Jackson R, Richter HE, Chan J, et al.. Factors associated with undergoing bilateral salpingo-oophorectomy at the time of hysterectomy for benign conditions. Obstet Gynecol 2009;113:1259–67.
2. Whiteman MK, Hillis SD, Jamieson DJ, Morrow B, Podgornik MN, Brett KM, et al.. Inpatient hysterectomy surveillance in the United States, 2000-2004. Am J Obstet Gynecol 2008;198:34 e1–7.
3. Wu JM, Wechter ME, Geller EJ, Nguyen TV, Visco AG. Hysterectomy rates in the United States, 2003. Obstet Gynecol 2007;110:1091–5.
4. Jacoby VL, Autry A, Jacobson G, Domush R, Nakagawa S, Jacoby A. Nationwide use of laparoscopic hysterectomy compared with abdominal and vaginal approaches. Obstet Gynecol 2009;114:1041–8.
5. Baloglu A, Bezircioglu I, Cetinkaya B, Karci L, Bicer M. Development of secondary ovarian lesions after hysterectomy without oophorectomy versus unilateral oophorectomy for benign conditions: a retrospective analysis of patients during a nine-year period of observation. Clin Exp Obstet Gynecol 2010;37:299–302.
6. Holub Z, Jandourek M, Jabor A, Kliment L, Wagnerova M. Does hysterectomy without salpingo-oophorectomy influence the reoperation rate for adnexal pathology? A retrospective study. Clin Exp Obstet Gynecol 2000;27:109–12.
7. Plockinger B, Kolbl H. Development of ovarian pathology after hysterectomy without oophorectomy. J Am Coll Surg 1994;178:581–5.
8. Rocca WA, Grossardt BR, de Andrade M, Malkasian GD, Melton LJ III. Survival patterns after oophorectomy in premenopausal women: a population-based cohort study. Lancet Oncol 2006;7:821–8.
9. Rivera CM, Grossardt BR, Rhodes DJ, Rocca WA. Increased cardiovascular mortality after early bilateral oophorectomy. Menopause 2009;16:15–23.
10. Parker WH, Broder MS, Chang E, Feskanich D, Farquhar C, Liu Z, et al.. Ovarian conservation at the time of hysterectomy and long-term health outcomes in the nurses' health study. Obstet Gynecol 2009;113:1027–37.
11. Colditz GA, Willett WC, Stampfer MJ, Rosner B, Speizer FE, Hennekens CH. Menopause and the risk of coronary heart disease in women. N Engl J Med 1987;316:1105–10.
12. Rocca WA, Grossardt BR, Geda YE, Gostout BS, Bower JH, Maraganore DM, et al.. Long-term risk of depressive and anxiety symptoms after early bilateral oophorectomy. Menopause 2008;15:1050–9.
13. Rocca WA, Grossardt BR, Shuster LT. Oophorectomy, menopause, estrogen, and cognitive aging: the timing hypothesis. Neurodegener Dis 2010;7:163–6.
14. Rocca WA, Grossardt BR, Shuster LT. Oophorectomy, menopause, estrogen treatment, and cognitive aging: clinical evidence for a window of opportunity. Brain Res 2011;1379:188–98.
15. Parker WH, Broder MS, Liu Z, Shoupe D, Farquhar C, Berek JS. Ovarian conservation at the time of hysterectomy for benign disease. Obstet Gynecol 2005;106:219–26.
16. Berek JS, Chalas E, Edelson M, Moore DH, Burke WM, Cliby WA, et al.. Prophylactic and risk-reducing bilateral salpingo-oophorectomy: recommendations based on risk of ovarian cancer. Obstet Gynecol 2010;116:733–43.
17. Prophylactic oophorectomy. Clinical management guidelines for obstetrician-gynecologists (replaces ACOG Technical Bulletin No. 111). ACOG Practice Bulletin No. 7. American College of Obstetricians and Gynecologists. Int J Gynaecol Obstet 1999;67:193–9.
18. Lindenauer PK, Pekow PS, Lahti MC, Lee Y, Benjamin EM, Rothberg MB. Association of corticosteroid dose and route of administration with risk of treatment failure in acute exacerbation of chronic obstructive pulmonary disease. JAMA 2010;303:2359–67.
19. Lindenauer PK, Pekow P, Wang K, Mamidi DK, Gutierrez B, Benjamin EM. Perioperative beta-blocker therapy and mortality after major noncardiac surgery. N Engl J Med 2005;353:349–61.
20. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83.
21. 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.
22. Hollenbeck BK, Wei Y, Birkmeyer JD. Volume, process of care, and operative mortality for cystectomy for bladder cancer. Urology 2007;69:871–5.
23. Birkmeyer JD, Sun Y, Wong SL, Stukel TA. Hospital volume and late survival after cancer surgery. Ann Surg 2007;245:777–83.
24. Spiegelman D, Hertzmark E. Easy SAS calculations for risk or prevalence ratios and differences. Am J Epidemiol 2005;162:199–200.
25. Haymart MR, Banerjee M, Stewart AK, Koenig RJ, Birkmeyer JD, Griggs JJ. Use of radioactive iodine for thyroid cancer. JAMA 2011;306:721–8.
26. Merlo J, Chaix B, Ohlsson H, Beckman A, Johnell K, Hjerpe P, et al.. A brief conceptual tutorial of multilevel analysis in social epidemiology: using measures of clustering in multilevel logistic regression to investigate contextual phenomena. J Epidemiol Community Health 2006;60:290–7.
27. Birkmeyer NJ, Dimick JB, Share D, Hawasli A, English WJ, Genaw J, et al.. Hospital complication rates with bariatric surgery in Michigan. JAMA 2010;304:435–42.
28. Jacoby VL, Grady D, Sawaya GF. Oophorectomy as a risk factor for coronary heart disease. Am J Obstet Gynecol 2009;200:140 e1–9.
29. Jacoby VL, Grady D, Wactawski-Wende J, Manson JE, Allison MA, Kuppermann M, et al.. Oophorectomy vs ovarian conservation with hysterectomy: cardiovascular disease, hip fracture, and cancer in the Women's Health Initiative Observational Study. Arch Intern Med 2011;171:760–8.
30. Novetsky AP, Boyd LR, Curtin JP. Trends in bilateral oophorectomy at the time of hysterectomy for benign disease. Obstet Gynecol 2011;118:1280–6.
31. Lowder JL, Oliphant SS, Ghetti C, Burrows LJ, Meyn LA, Balk J. Prophylactic bilateral oophorectomy or removal of remaining ovary at the time of hysterectomy in the United States, 1979-2004. Am J Obstet Gynecol 2010;202:538 e1–9.
32. Brand AH. The RANZCOG College Statement on prophylactic oophorectomy in older women undergoing hysterectomy for benign disease: is the evidence sufficient to change practice? Aust N Z J Obstet Gynaecol 2011;51:296–300.
33. Plusquin C, Fastrez M, Vandromme J, Rozenberg S. Factors affecting gynaecologists' decision to perform prophylactic oophorectomy concomitantly with hysterectomy: a Belgian survey. Maturitas 2011;70:391–4.
34. Allison JJ, Kiefe CI, Weissman NW, Person SD, Rousculp M, Canto JG, et al.. Relationship of hospital teaching status with quality of care and mortality for Medicare patients with acute MI. JAMA 2000;284:1256–62.
35. Joynt KE, Harris Y, Orav EJ, Jha AK. Quality of care and patient outcomes in critical access rural hospitals. JAMA 2011;306:45–52.
© 2013 The American College of Obstetricians and Gynecologists
36. Werner RM, Goldman LE, Dudley RA. Comparison of change in quality of care between safety-net and non-safety-net hospitals. JAMA 2008;299:2180–7.