Objective
Despite recommendation for ovarian conservation in low-risk, premenopausal women, bilateral oophorectomy is often performed. The purpose of this study was to investigate factors associated with removal of normal ovaries at the time of hysterectomy for benign indication in women age <51 years.
Study Design
Demographics, indication for surgery, adnexal pathology, and surgical approach were analyzed for hysterectomies from a voluntary, statewide surgical quality collaborative. Cases were excluded if the surgical indication was cancer, pelvic mass, or obstetric, or if age was >50 years. Cases were categorized according to pathology of the adnexal specimen as cancer, benign findings, normal ovary, or no ovarian specimen. Variables including demographics, medical comorbidities, and surgical characteristics were analyzed to identify characteristics associated with oophorectomy at the time of hysterectomy. A logistic regression model was then developed to identify factors independently associated with removal of normal ovaries.
Results
A total of 6789 subjects were included. Oophorectomy was performed in 44.2% of women (n = 3002). In all, 23.1% (n = 1565) had normal ovaries on pathology. Incidental ovarian cancer was found in 0.2% (n = 12), and benign pathology was found in 21% (n = 1425). Removal of normal ovaries was less likely when the surgical approach was vaginal (18%) as opposed to laparoscopic (23.1%) or abdominal (26.0%). With adjustment, abdominal (odds ratio [OR], 1.73; 95% confidence interval [CI], 1.43–2.09]) and laparoscopic (OR, 1.27; 95% CI, 1.08–1.50) approach showed significantly higher odds of normal ovary removal compared to vaginal hysterectomy. Age 46-50 years was also significantly associated (OR, 1.78; 95% CI, 1.53–2.07). Surgical indications associated with increased oophorectomy with normal resultant pathology were family history of cancer (OR, 3.09; 95% CI, 1.94–4.94), endometrial hyperplasia (OR, 2.36; 95% CI, 1.38–4.01), endometriosis (OR, 2.01; 95% CI, 1.30–3.09), and cervical dysplasia (OR, 1.91; 95% CI, 1.12–3.28).
Conclusion
Removal of histologically normal ovaries is performed in nearly 1 of every 4 women age <51 years undergoing hysterectomy for benign indications. Factors associated include age closer to menopause, surgical approach, and certain indications for hysterectomy. Reducing the rate of elective oophorectomy in low-risk, premenopausal women may be a target for quality improvement efforts. Future work should continue to evaluate this practice, associated factors, physician counseling, and patient decision-making.
Hysterectomy is the most common major gynecologic surgery done in the United States, with about 433,000 performed in 2010. Elective oophorectomy is performed in nearly 40% of women undergoing hysterectomy for benign disease. The decision to remove ovaries at the time of hysterectomy can be an especially contentious topic among surgeons, and a confusing decision for patients.
The debate about oophorectomy at the time of hysterectomy centers on several topics. Bilateral salpingo-oophorectomy (BSO) is the most definitive prevention for ovarian cancer, which carries a lifetime risk of 1.4%. There also exists the potential for reoperation in women whose ovaries are retained at the time of hysterectomy, at rates of 2.8-9.2%. These risks must be weighed against the potential adverse effects of premenopausal oophorectomy. Studies suggest that premenopausal oophorectomy has negative cardiovascular effects. Earlier age of menopause has been associated with diabetes, and oophorectomy has been linked to faster rates of bone mineral loss and carotid thickening. There is also the risk of sexual and mood dysfunction in women, and negative effects on cognitive function. The American Congress of Obstetricians and Gynecologists (ACOG), in the practice bulletin reaffirmed in 2014, states that strong consideration should be made for retaining normal ovaries in premenopausal women who are not at increased genetic risk of ovarian cancer.
We sought to investigate current practice of oophorectomy at the time of hysterectomy. Our objective was to determine which patient and operative characteristics are associated with the removal of histologically normal ovaries at the time of hysterectomy for benign disease in women ≤50 years of age.
Materials and Methods
Data were obtained from the Michigan Surgical Quality Collaborative, a group of 52 hospitals from academic and community settings throughout the state. Patients with all types of insurance are included, and Blue Cross Blue Shield of Michigan/Blue Care Network provides funding for the collaborative. Data are abstracted from charts by specially trained, dedicated nurse abstractors. Patient characteristics, surgical indication, intraoperative processes of care, and surgical pathology are routinely collected. To reduce sampling error, a standardized data collection methodology is employed that uses only the first 25 cases of an 8-day cycle (alternating on different days of the week for each cycle). Routine validation of the data is maintained by scheduled site visits, conference calls, and internal audits. The University of Michigan Institutional Review Board granted “not regulated” status to this study (HUM00073978).
Data including demographics, indication for hysterectomy, medical comorbidities, adnexal pathology, and surgical approach were analyzed for women who underwent hysterectomy from Jan. 2, 2013, through July 2, 2014. Women age ≤50 years who underwent hysterectomy for the following indications were included: prolapse; family history of breast, ovary, or other cancer syndrome involving the reproductive organs; cervical dysplasia; endometrial hyperplasia with or without atypia; abnormal uterine bleeding (AUB) and/or fibroids; and endometriosis, including endometriosis along with pelvic pain, abnormal bleeding, and/or fibroids. Cases were excluded if age was >50 years, or if the surgical indication was an obstetrical event, gynecologic cancer, or pelvic mass. This age cutoff was chosen because the average age of menopause, according to the National Institute on Aging, is 51 years, and in accordance with the ACOG statement that strong consideration should be given to retaining normal ovaries in premenopausal women who are not at increased genetic risk of ovarian cancer.
Cases were categorized with findings abstracted from the pathology report. The following groups were identified: (1) incidental ovarian cancer; (2) benign pathology, which included mature teratoma or dermoid, fibroma, endometrioma or endometriosis, adhesive disease, tubo-ovarian abscess, or other cystic mass; (3) normal ovary, which included small simple/functional cysts, no pathologic abnormality, or other histologic diagnoses such as corpora albicans or tubal Walthard cell rests; (4) no ovarian specimen, indicating ovarian preservation; or (5) insufficient data for categorization. Cases were categorized as “no ovarian specimen” if the uterine specimen was completely described but no ovarian findings were described in the pathology report. Ovarian pathology was considered missing or insufficient if the ovaries were not described in the pathology report and the uterine specimen was incompletely described, ie, there were not comments on both myometrium and endometrium. Patients were classified based on the most significant pathologic diagnosis on either ovary (malignant > benign > normal). For example, if endometriosis was reported on 1 ovary but the other ovary was normal, the patient was included only in the benign pathology group.
The following variables were assessed using bivariate analyses to identify characteristics significantly associated with oophorectomy at the time of hysterectomy: age (<40, 40-45, 46-50 years), parity, medical comorbidities including diabetes and cardiovascular disease, surgical indication, and surgical approach. Patients with coronary artery disease, with hypertension, on statin therapy, and/or on beta-blocker therapy were included in the cardiovascular disease group. The diabetes group included types 1 and 2, either diet or medically managed. Indications for hysterectomy were abstracted from the operative report and categorized as described within the inclusion criteria. The surgical approach for hysterectomy was categorized as abdominal, laparoscopic (which included robotic-assisted techniques), or vaginal (which included laparoscopic-assisted vaginal hysterectomy [LAVH] based on the standard grouping by the National Hospital Safety Network). Bivariate analyses were done with analysis of variance for continuous measures and χ 2 analysis for tabulation of categorical variables.
A logistic regression model was used to identify factors that were independently associated with removal of normal ovaries at the time of hysterectomy. Factors that were statistically significant in bivariate analyses or were plausibly related to the decision to proceed with ovarian removal were included in the multivariate model. For route of hysterectomy, vaginal was used as the referent group, and for surgical indication, prolapse was the referent group, as these groups had the lowest removal rate in bivariate analysis. Age <40 years was the referent age group, again because this group had the lowest removal rate. All statistical analyses were generated using software (STATA v13; StataCorp, College Station, TX) and P < .05 was considered significant for all analyses.
Results
Data for 7597 hysterectomies in women age ≤50 years were available for analysis. There were 436 excluded for an indication of cancer, 223 for pelvic mass, and 15 for obstetric indications, leaving 6923 hysterectomies meeting our inclusion criteria. Of these, 134 cases (1.9%) had insufficient pathologic data for analysis. Oophorectomy with any pathologic diagnosis (incidental cancer, benign pathology, and normal ovarian pathology) was found in 44.2% of women (n = 3002) at the time of hysterectomy. Normal ovarian pathology was found in 23.1% (n = 1565). Incidental ovarian cancer was found in 0.2% (n = 12), and benign pathology was found in 21% (n = 1425). Ovaries were preserved in 55.8% (n = 3787). Bivariate analysis of demographic and medical comorbidities is shown in Table 1 . Women aged 46-50 years had the highest rate of ovarian removal. Median parity was 2, and the interquartile ranges were the same in 3 of 4 groups, with ovarian cancer being statistically significantly different but with no clear clinical significance. The presence of diabetes and patient body mass index were not significantly different across the pathologic groups. The rate of cardiovascular disease was lowest among those with incidental ovarian cancer, and present in about one quarter of women in the other groups.
| Patient characteristic | No oophorectomy (n = 3787) | Oophorectomy (n = 3002) | P value | ||
|---|---|---|---|---|---|
| Normal ovaries (n = 1565) | Benign ovarian pathology (n = 1425) | Incidental ovarian cancer (n = 12) | |||
| Age in y, % (n) | |||||
| <40 (n = 2123) | 59.4 (1315) | 20.8 (461) | 19.7 (435) | 0.09 (2) | < .001 |
| 40–45 (n = 2692) | 59.2 (1594) | 20.7 (558) | 19.9 (538) | 0.07 (2) | < .001 |
| 46–50 (n = 1884) | 46.6 (878) | 28.9 (546) | 23.9 (452) | 0.4 (8) | < .001 |
| Parity, median (interquartile range) | 2 (1–3) | 2 (1–3) | 2 (1–3) | 2 (2–2) | < .001 |
| BMI in kg/m 2 , mean ± SD | 30.6 ± 7.6 | 30.9 ± 7.8 | 30.9 ± 7.7 | 32.9 ± 8.9 | .534 |
| Diabetes mellitus, % (n) | 5.8 (218) | 7.2 (112) | 6.8 (97) | 8.3 (1) | .209 |
| Cardiovascular disease, % (n) | 22.6 (855) | 24.9 (390) | 26.3 (375) | 8.3 (1) | .013 |
The most common surgical indications were AUB/fibroids at 40% and endometriosis at 45%. The most common route of hysterectomy was laparoscopic (which included robotic) at nearly 60%, with abdominal and vaginal/LAVH each comprising 20%. Table 2 shows the bivariate analyses for surgical details, with breakdown by pathology groups. The rate of ovarian preservation was higher when the route was vaginal/LAVH (64.9%, n = 862), compared to laparoscopic (54.6%, n = 2220) or abdominal (50.3%, n = 705), P < .001. Similarly, prolapse as surgical indication had the highest rate of ovarian preservation, while family history had the lowest.
| Variable | No oophorectomy, % (n) | Oophorectomy | ||
|---|---|---|---|---|
| Normal ovaries, % (n) | Benign ovarian pathology, % (n) | Incidental ovarian cancer, % (n) | ||
| Surgical indication | ||||
| Prolapse (n = 186) | 72.0 (134) | 13.9 (26) | 13.9 (26) | 0 (0) |
| Family history (n = 412) | 36.7 (151) | 35.7 (147) | 27.1 (112) | 0.5 (2) |
| Cervical dysplasia (n = 195) | 56.9 (111) | 24.1 (47) | 18.5 (36) | 0.5 (1) |
| Endometrial hyperplasia (n = 180) | 51.7 (93) | 30.6 (55) | 17.8 (32) | 0 (0) |
| AUB/fibroids (n = 2716) | 65.1 (1767) | 16.9 (460) | 17.9 (485) | 0.2 (4) |
| Endometriosis (n = 3068) | 49.6 (1522) | 26.8 (822) | 23.4 (719) | 0.2 (5) |
| Surgical approach | ||||
| Vaginal/LAVH (n = 1327) | 64.9 (862) | 18.0 (239) | 16.8 (223) | 0.2 (3) |
| Laparoscopic (n = 4061) | 54.6 (2220) | 23.1 (937) | 22.1 (898) | 0.2 (6) |
| Abdominal (n = 1401) | 50.3 (705) | 27.8 (389) | 21.7 (304) | 0.2 (3) |
Table 3 shows the multivariate regression analysis used to identify independent associations with removal of histologically normal ovaries. Age 46-50 years was associated with removal of normal ovaries, although body mass index, diabetes, and cardiovascular disease were not. For route of hysterectomy, both abdominal (odds ratio [OR], 1.73; 95% confidence interval [CI], 1.43–2.09) and laparoscopic (OR, 1.27; 95% CI, 1.08–1.50) approach showed significantly higher odds of removal of normal ovaries compared to vaginal hysterectomy. Surgical indications associated with increased oophorectomy with normal resultant pathology were family history of cancer (OR, 3.09; 95% CI, 1.94–4.94), followed by endometrial hyperplasia (OR, 2.36; 95% CI, 1.38–4.01), endometriosis (OR, 2.01; 95% CI, 1.30–3.09), and cervical dysplasia (OR, 1.91; 95% CI, 1.12–3.28). Uterine fibroids/AUB were not significantly associated with removal of normal ovaries (OR, 1.12; 95% CI, 0.73–1.73). In analyzing the location of endometriosis, with no endometriosis as the referent group, endometriosis located surrounding the bowel or bladder was associated with the highest risk of removal of normal ovaries (OR, 2.39; 95% CI, 1.68–3.40).
