Endometrial cancer is the most common malignancy arising from the female reproductive tract in the Western world. In 2012, an estimated 47,130 new cases and 8,010 deaths will be attributed to this disease in the United States.1 Also, the incidence and death rate of endometrial cancer have increased in the past several decades.1,2 The identification of women who are at increased risk for endometrial cancer may allow early diagnosis, prevention, and a reduction in the considerable societal burden imposed by endometrial cancer. In fact, investigators have shown that application of endometrial cancer screening and preventive strategies to the entire population of women is not cost-effective, whereas targeting patients at high risk may be an effective option.3–5
A history of benign breast biopsy results is a well-known risk factor for development of breast cancer and is routinely used for risk stratification.6 Twu and Chen7 and Gull et al8 reported a higher risk of endometrial cancer in patients undergoing endometrial biopsy (with a benign diagnosis) when compared with the general population. However, there are no studies that identify and quantify risk or protective factors for subsequent development of endometrial cancer after a benign result on an index endometrial biopsy or dilation and curettage (D&C) in a comprehensive manner. The absence of such studies was confirmed on PubMed and Ovid search engines. We undertook the present study to identify risk factors for subsequent development of endometrial cancer after a benign result on index endometrial biopsy or D&C, using a partial nested case-control study design.
MATERIALS AND METHODS
This study was approved by the Mayo Clinic and Olmsted Medical Center Institutional Review Boards. The Rochester Epidemiology Project was used to identify patients with endometrial cancer. The project is a unique research infrastructure (R01 AG034676) that was established in 1966 and has been used to conduct population-based studies among residents of Olmsted County, Minnesota. As part of these resources, the medical records are linked and indexed, and persons in the Olmsted County population with specific diseases may be identified through retrieval of records with pertinent diagnostic codes.9
With these resources and careful review of the medical records, we identified 370 patients who had received a diagnosis of endometrial cancer between January 1, 1970, and December 31, 2008, while they were residents of Olmsted County, and who had not denied access to their medical records for research purposes.
Of the 370 patients initially identified with endometrial cancer, 101 (27%) had a benign endometrial biopsy or D&C result before diagnosis of endometrial cancer. From these 101 patients, we excluded 11 patients who had simple or complex atypical hyperplasia in the endometrial biopsy or D&C (because of intrinsic high risk of endometrial cancer), leaving 90 (24.5%) patients for the analysis. Their first benign endometrial biopsy or D&C is herein referred to as the index endometrial biopsy or D&C.
Using institutional databases, we identified all patients with an endometrial biopsy or D&C performed at Mayo Clinic between January 1, 1985 (electronic data not available before 1985), and December 31, 2008, while they were Olmsted County residents, who were aged 40 to 85 years and had not denied access to their medical records for research purposes. These patients were cross-referenced with the institution’s medical index to exclude any patient who ever received a diagnosis of endometrial cancer or atypical hyperplasia. The remaining pool of control group participants with no endometrial cancer but with a benign endometrial biopsy or D&C result consisted of 7,994 patients. For each of the 90 endometrial cancer cases with benign endometrial biopsy or D&C results, two matched control cases with at least as much follow-up after the index endometrial biopsy or D&C as the matched case were randomly identified from the control pool. The matching criteria included benign compared with hyperplasia diagnosis on the endometrial biopsy or D&C, age at endometrial biopsy or D&C (within ±8 years), and date of the endometrial biopsy or D&C (within ±3 years if the endometrial biopsy or D&C was performed after 1985).
For this study, the following definition was used for personal history of hereditary nonpolyposis colorectal cancer–related malignancy: a personal history of cancer of the pancreas, colon, rectum, ovary, small bowel, stomach, biliary tract, or brain, and transitional cancers of the urinary tract. Family history of hereditary nonpolyposis colorectal cancer–related malignancy was defined as presence of at least one first-degree relative with cancer of the pancreas, colon, rectum, endometrium, ovary, small bowel, stomach, biliary tract, or brain, and transitional cancers of the urinary tract. This definition satisfied the need of using an easily assessable and simple variable at the time of the index endometrial biopsy or D&C. Also, the inaccuracy of family history in the identification of patients with hereditary nonpolyposis colorectal cancer has been demonstrated before.10
The following definition was used to categorize a polyp: combined epithelial stromal sessile proliferation consisting of glands and fibrotic stroma with large and thick-walled arteries coated by an epithelium surface. The endometrial biopsy was classified as benign nonspecified when it was defined as benign by the pathologist but was not better characterized and did not fit in the other pathologic categories.
Data were summarized using standard descriptive methods, frequency, and percentages for categorical variables, and mean and standard deviation or median and range for continuous variables. For each factor of interest, a separate conditional logistic regression model was fit to evaluate the association between the factor and the case-control status. In addition, a multivariable conditional logistic model was fit using stepwise and backward variable selection methods. Associations were summarized using the odds ratio (OR) and the corresponding 95% confidence interval (CI). All calculated P values were two-sided and P<.05 was considered statistically significant. Statistical analyses were performed using the SAS 9.2 software package.
The final analysis included 90 case group participants and 172 control group participants (eight case group participants had only one control group participant each). Demographic data of the patients are summarized in Table 1. Mean ages at the time of the index endometrial biopsy or D&C were 51.8 (standard deviation 11.0, range 17.9–81.8) and 51.9 (standard deviation 10.4, range 21.5–80.1) years, respectively, among case and control group participants. The time interval between the index endometrial biopsy or D&C and the diagnosis of endometrial cancer among case group participants ranged from 0 to 23.3 years, with a median duration of 6.7 years. Matched control group participants were selected to have at least as much follow-up as the corresponding case group participants, with an overall median of 13.8 years from the date of the index endometrial biopsy or D&C to the date of last follow-up or hysterectomy.
Among the 262 patients, the index endometrial biopsy or D&C was performed as part of the work-up of known nonendometrial disease (eg, uterine fibroids, cervical disease) in eight patients (3%); by comparison, in 246 patients (94%), the endometrial biopsy or D&C was performed because of known endometrial diagnoses documented before the endometrial biopsy or D&C (eg, endometritis) or symptoms (bleeding) directly referable to the endometrium itself. Two other patients (1%) had an endometrial biopsy or D&C for screening, and no indication was documented by the clinician for six patients. The method of endometrial sampling for the index endometrial biopsy or D&C was available for 247 patients: an endometrial biopsy in 86 (35%) and a D&C in 161 (65%). Patients who underwent D&C at the time of their index endometrial biopsy or D&C were 2.1-times more likely to later have endometrial cancer than patients who underwent an endometrial biopsy (OR 2.10, 95% CI 1.08–4.12; P=.03).
The final pathologic diagnosis of the endometrial biopsy or D&C was endometrial hyperplasia without atypia in 16 (19%) of the 85 case group participants and 37 (22%) of the 170 control group participants with detailed histologic information. Histologic diagnoses are summarized in Figure 1. Of the 85 case group participants, 19 (22%) had polyps on endometrial biopsy or D&C, compared with only 8 (5%) of the 170 control group participants (P<.001). In accordance with the study definition, all case group participants and control group participants with hyperplasia at the time of the index endometrial biopsy or D&C had a diagnosis of hyperplasia without atypia.
The histologic subtype of the endometrial cancer that developed after the index endometrial biopsy or D&C was endometrioid (78; 87%), serous (2; 2%), mixed (2; 2%), and unknown (8; 9%). The stages of these endometrial cancers were as follows: stage I (81; 90%); stage II (3; 3%); stage III (5; 6%); and stage IV (1; 1%).
Colorectal cancer, within case group participants and control group participants, was the predominant cancer in personal and family history of hereditary nonpolyposis colorectal cancer–related malignancy. Colon cancer represents 100% (18 patients) of the cancers related to hereditary nonpolyposis colorectal cancer in the personal history and 66% (29 patients) in the family history.
On univariable analysis, patient weight and body mass index (BMI, calculated as weight (kg)/[height (m)]2), nulliparous status, personal history of hereditary nonpolyposis colorectal cancer–related malignancy (100% colorectal cancer), use of unopposed estrogen therapy, D&C, endometrial polyp, and lack of oral contraceptive pill (OCP) use before or at the time of the benign endometrial biopsy or D&C were each identified as being significantly associated with subsequent development of endometrial cancer (Table 1).
Use of tamoxifen before or at the time of index endometrial biopsy or D&C was almost significantly associated with endometrial cancer development (P=.08). Among the 172 control group participants, one had a history of tamoxifen use, and this patient did not have endometrial polyps present on histologic evaluation. Of the 170 control group participants (one had missing information) who did not have a history of tamoxifen use, eight (5%) had polyps present on histologic evaluation. Among the 90 case group participants, four had a history of tamoxifen use and one (25%) of these had endometrial polyps. Of the 83 case group participants (three had missing information) who did not have a history of tamoxifen use, 17 (21%) had polyps present on histologic evaluation.
The following four variables were identified as independently associated with subsequent development of endometrial cancer on the basis of a multivariable conditional logistic regression model using variable selection methods: OCP use before or at endometrial biopsy or D&C (protective factor) (OR 0.18, 95% CI 0.08–0.45; P<.001); presence of polyp on histologic evaluation (OR 4.12, 95% CI 1.40–12.17; P=.01); personal history of hereditary nonpolyposis colorectal cancer–related malignancy (100% colorectal cancer) (OR 4.44, 95% CI 1.02–19.31; P<.05); and BMI of 35 or more (OR 3.40, 95% CI 1.18–9.78; P<.03).
We stratified patients on the basis of the number of the four independent predictors (BMI 35 or more, no OCP use, endometrial polyps, and personal history of hereditary nonpolyposis colorectal cancer–related malignancy). Overall, 86 patients had one risk factor, 39 patients had two risk factors, and 7 had three risk factors. Patients with at least one risk factor had an 8.12-times higher risk of endometrial cancer than women without risk factors. The presence of two or more risk factors increased the risk of cancer by 17.87-times (Table 2). Based on the Surveillance, Epidemiology, and End Results (SEER) rates from 2007 to 2009, there is a 2.6% lifetime risk of endometrial cancer in United States women.11 Assuming this lifetime risk estimate is applicable to the cohort of women with benign endometrial biopsy or D&C with none of the four aforementioned risk factors, ORs of 8.12 and 17.87 for one and two or more of the four aforementioned risk factors confer a lifetime risk of approximately 18% and 32%, respectively.
In this study, we evaluated risk factors associated with endometrial cancer development after benign endometrial biopsy or D&C. The endometrial biopsy or D&C gives pathologic information about the endometrium and also is a surrogate for identifying women at increased risk for endometrial cancer. This notion is supported by our data depicting that approximately 25% of women with endometrial cancer in our population had already undergone an endometrial biopsy or D&C in the past showing benign findings (excluding atypical hyperplasia). The cohort of women who undergo endometrial biopsy or D&C is inherently a high-risk subgroup compared with the rest of the population. These women have had multiple-layer filtering in terms of their risk stratification by virtue of their history and examination, with an increased probability of endometrial cancer development.
In our population, the median time to endometrial cancer was 6.7 years after an index endometrial biopsy or D&C. This long period provides a potential window of opportunity in which surveillance programs can target the identification of women who might have had a benign endometrial biopsy or D&C and might have one or more risk factors identified in our study (eg, no OCP use, personal history of hereditary nonpolyposis colorectal cancer–related malignancy [colorectal cancer], presence of polyp on endometrial biopsy or D&C, morbid obesity). Although some of the risk factors found to be of significance in our study had been already implicated for the development of endometrial cancer,12,13 the uniqueness of our findings is that we identified those risk factors at a time when several steps can be undertaken to address endometrial cancer prevention. We suggest that after such women at high risk are identified, a closer follow-up may be performed to address screening and preventive actions. Steps that can be taken include risk modification (such as weight loss), routine serial pelvic ultrasonographic scans, targeted endometrial sampling, administration of OCPs, or prophylactic hysterectomy. However, we lack prospective demonstration of the cost-effectiveness of these suggestions.
Unfortunately, our case-control design and the lack of abstracted clinical information on the total Olmsted County population who had a benign endometrial biopsy between 1970 and 2008 do not allow the precise estimation of the risk of development of endometrial cancer after benign endometrial biopsy or D&C in our population. For this reason, we opted to use the SEER database for estimating a baseline risk. According to the SEER database, the lifetime risk of endometrial cancer in United States women is 2.6%.11 Similarly, it has been reported that the overall risk of development of endometrial cancer after a benign endometrial sample for postmenopausal bleeding is 2.7%,8 and this is increased to 3.5% in the presence of an endometrial polyp.14 These estimations have been utilized for baseline risk of endometrial cancer in patients with benign endometrial biopsy or D&C results and no risk factors. However, lifetime risks may change according to the age of the patient and presence of risk factors, and our percentages are only an approximation based on general population data.
In hereditary nonpolyposis colorectal cancer syndrome, the lifetime risk of endometrial cancer may be as high as 60%.15 Our observation that personal history of hereditary nonpolyposis colorectal cancer–related malignancy (colorectal cancer) is a strong risk factor for endometrial cancer developing after a benign endometrial biopsy or D&C is consistent with the aforementioned data.
The presence of an endometrial polyp already has been described as a risk factor for endometrial cancer.16,17 We demonstrated that the presence of an endometrial polyp in an index endometrial biopsy or D&C increases the risk of endometrial cancer development by 4.2-times compared with other benign diagnoses.
Our study was not designed to identify any association between tamoxifen and endometrial polyps. However, despite the small numbers, results of our statistical analysis seem to favor a direct association between endometrial polyp and future endometrial cancer, independent of tamoxifen use.
We observed that OCP use before the benign endometrial biopsy or D&C was associated with more than six times lower likelihood of endometrial cancer. Similar findings in the general population were reported by other investigators.18,19 Estimated protection with use of OCPs ranged from 20% with 1 year of use to 80% with 10 years of use. Interestingly, the protective role of OCPs may not be observed in patients with a personal history of hereditary nonpolyposis colorectal cancer–related malignancy, because hereditary nonpolyposis colorectal cancer–related endometrial cancer usually occurs independently of hormonal stimulation.20
Patients with morbid obesity at the time of benign endometrial biopsy or D&C are at increased risk for endometrial cancer. It is known that obese patients have more bioavailable estrogens; therefore, they are more vulnerable to endometrial cancer type I.21,22 Because obesity is a modifiable risk factor, preventive strategies and lifestyle changes may have an important effect in this population.
In our study, we excluded patients with atypical hyperplasia for the following reasons. First, 40–50% of these patients have concomitant endometrial cancer23,24 at the time of diagnosis of atypical hyperplasia. Second, atypical hyperplasia is a well-recognized risk factor for and precursor of endometrial cancer and has been extensively studied.25 Third, most women (80%)26 with a diagnosis of atypical hyperplasia may undergo hysterectomy as a primary and preventive treatment.
In contrast to studies that report equal or better diagnostic efficacy of D&C,27 we observed that D&C was more likely to be associated with subsequent endometrial cancer when compared with Pipelle biopsy. This divergent observation may be related to the time course of our study, which included early years when office endometrial biopsy was not routinely performed. In later years, selection bias of patients at high risk undergoing D&C may explain the discrepancy.
The limitations of our study include its retrospective nature, the lack of detailed information about progesterone use and history of infertility, the long observation time period, and the reliance on data collected from a geographically limited area with relatively homogeneous populations that have reasonably easy access to medical care. Also, control group participants and case group participants were selected from slightly different time periods because of the unavailability of endometrial biopsy or D&C electronic data before 1985. However, we report a long follow-up period, our data collection was comprehensive for the population undergoing study, and we monitored all residents of Olmsted County. Hence, the selection bias in our study is likely minimal. Moreover, the resources from the Rochester Epidemiology Project provide accurate patient history and a foundation for population-based studies with comprehensive disease, follow-up, and outcome information. Although, as seen in previous studies, the interpretation of an endometrial biopsy or D&C28 may vary depending on pathologists, our study is strengthened by a robust central pathology review by specialized gynecologic pathologists.
In summary, we have shown that approximately one fourth of women with endometrial cancer had a previous benign endometrial biopsy or D&C result. Considering the cohort of women who have benign endometrial biopsy or D&C findings, we observed that personal history of colorectal cancer, presence of a benign endometrial polyp in the sample, and morbid obesity are strong risk factors for future development of endometrial cancer, whereas use of OCPs is protective. These data can guide clinicians and patients for efficient and targeted use of diagnostic or preventive strategies for endometrial cancer.
1. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA Cancer J Clin 2012;62:10–29.
2. Weiss NS, Szekely DR, Austin DF. Increasing incidence of endometrial cancer in the United States. N Engl J Med 1976;294:1259–62.
3. Havrilesky LJ, Maxwell GL, Myers ER. Cost–effectiveness analysis of annual screening strategies for endometrial cancer. Am J Obstet Gynecol 2009;200:640.e1–8.
4. Kwon JS, Lu KH. Cost-effectiveness analysis of endometrial cancer prevention strategies for obese women. Obstet Gynecol 2008;112:56–63.
5. Leslie KK, Thiel KW, Yang S. Endometrial cancer: potential treatment and prevention with progestin-containing intrauterine devices. Obstet Gynecol 2012;119:419–20.
6. Hartmann LC, Sellers TA, Frost MH, Lingle WL, Degnim AC, Ghosh K, et al.. Benign breast disease and the risk of breast cancer. N Engl J Med 2005;353:229–37.
7. Twu NF, Chen SS. Five-year follow-up of patients with recurrent postmenopausal bleeding. Zhonghua Yi Xue Za Zhi (Taipei) 2000;63:628–33.
8. Gull B, Karlsson B, Milsom I, Granberg S. Can ultrasound replace dilation and curettage? A longitudinal evaluation of postmenopausal bleeding and transvaginal sonographic measurement of the endometrium as predictors of endometrial cancer. Am J Obstet Gynecol 2003;188:401–8.
9. Melton LJ III. History of the Rochester Epidemiology Project. Mayo Clin Proc 1996;71:266–74.
10. Hampel H, Frankel WL, Martin E, Arnold M, Khanduja K, Kuebler P, et al.. Screening for the Lynch syndrome (hereditary nonpolyposis colorectal cancer). N Engl J Med 2005;352:1851–60.
11. Howlader N, Noone AM, Krapcho M, Neyman N, Aminou R, Altekruse SF, et al. (editors). SEER cancer statistics review, 1975-2009 (vintage 2009 populations). Avaiable at: http://seer.cancer.gov/csr/1975_2009_pops09/
. Retrieved April 2012.
12. Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen M. Body-mass index and incidence of cancer: a systematic review and meta-analysis of prospective observational studies. Lancet 2008;371:569–78.
13. Parazzini F, Negri E, La Vecchia C, Benzi G, Chiaffarino F, Polatti A, et al.. Role of reproductive factors on the risk of endometrial cancer. Int J Cancer 1998;76:784–6.
14. Armenia CS. Sequential relationship between endometrial polyps and carcinoma of the endometrium. Obstet Gynecol 1967;30:524–9.
15. Aarnio M, Sankila R, Pukkala E, Salovaara R, Aaltonen LA, de la Chapelle A, et al.. Cancer risk in mutation carriers of DNA-mismatch-repair genes. Int J Cancer 1999;81:214–8.
16. Cohen I. Endometrial polyps in pre- and postmenopausal women: factors associated with malignancy. Maturitas 2008;59:99–100.
17. Lev-Sagie A, Hamani Y, Imbar T, Hurwitz A, Lavy Y. The significance of intrauterine lesions detected by ultrasound in asymptomatic postmenopausal patients. BJOG 2005;112:379–81.
18. Kiley J, Hammond C. Combined oral contraceptives: a comprehensive review. Clin Obstet Gynecol 2007;50:868–77.
19. Schlesselman JJ. Risk of endometrial cancer in relation to use of combined oral contraceptives: a practitioner’s guide to meta-analysis. Hum Reprod 1997;12:1851–63.
20. Zighelboim I, Goodfellow PJ, Gao F, Gibb RK, Powell MA, Rader JS, et al.. Microsatellite instability and epigenetic inactivation of MLH1 and outcome of patients with endometrial carcinomas of the endometrioid type. J Clin Oncol 2007;25:2042–8.
21. Kulie T, Slattengren A, Redmer J, Counts H, Eglash A, Schrager S. Obesity and women’s health: an evidence-based review. J Am Board Fam Med 2011;24:75–85.
22. Kaaks R, Lukanova A, Kurzer MS. Obesity, endogenous hormones, and endometrial cancer risk: a synthetic review. Cancer Epidemiol Biomarkers Prev 2002;11:1531–43.
23. Trimble CL, Kauderer J, Zaino R, Silverberg S, Lim PC, Burke JJ II, et al.. Concurrent endometrial carcinoma in women with a biopsy diagnosis of atypical endometrial hyperplasia: a Gynecologic Oncology Group study. Cancer 2006;106:812–9.
24. Silverberg SG. Problems in the differential diagnosis of endometrial hyperplasia and carcinoma. Mod Pathol 2000;13:309–27.
25. Lacey JV Jr, Ioffe OB, Ronnett BM, Rush BB, Richesson DA, Chatterjee N, et al.. Endometrial carcinoma risk among women diagnosed with endometrial hyperplasia: the 34-year experience in a large health plan. Br J Cancer 2008;98:45–53.
26. Pennant S, Manek S, Kehoe S. Endometrial atypical hyperplasia and subsequent diagnosis of endometrial cancer: a retrospective audit and literature review. J Obstet Gynaecol 2008;28:632–3.
27. Leitao MM Jr, Kehoe S, Barakat RR, Alektiar K, Gattoc LP, Rabbitt C, et al.. Comparison of D&C and office endometrial biopsy accuracy in patients with FIGO grade 1 endometrial adenocarcinoma. Gynecol Oncol 2009;113:105–8.
28. Zaino RJ, Kauderer J, Trimble CL, Silverberg SG, Curtin JP, Lim PC, et al.. Reproducibility of the diagnosis of atypical endometrial hyperplasia: a Gynecologic Oncology Group study. Cancer 2006;106:804–11.