If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Clinical guidelines recommend that women with abnormal uterine bleeding with risk factors have an endometrial biopsy to exclude hyperplasia or cancer. Given the majority of endometrial cancer occurs in postmenopausal women, it has not been widely recognized that obesity is a significant risk factor for endometrial hyperplasia and cancer in young, symptomatic, premenopausal women.
We sought to evaluate the effect of body mass index on risk of endometrial hyperplasia or cancer in premenopausal women with abnormal uterine bleeding.
This was a retrospective cohort study in a single large urban secondary women’s health service. Participants were 916 premenopausal women referred for abnormal uterine bleeding of any cause and had an endometrial biopsy from 2008 through 2014. The primary outcome was complex endometrial hyperplasia (with or without atypia) or endometrial cancer.
Almost 5% of participants had complex endometrial hyperplasia or cancer. After adjusting for clinical and demographic factors, women with a measured body mass index ≥30 kg/m2 were 4 times more likely to develop complex hyperplasia or cancer (95% confidence interval, 1.36–11.74). Other risk factors were nulliparity (adjusted odds ratio, 3.08; 95% confidence interval, 1.43–6.64) and anemia (adjusted odds ratio, 2.23; 95% confidence interval, 1.14–4.35). Age, diabetes, and menstrual history were not significant.
Obesity is an important risk factor for complex endometrial hyperplasia or cancer in premenopausal women with abnormal uterine bleeding who had an endometrial biopsy in a secondary gynecology service. As over half of women with the outcome in this study were age <45 years, deciding to biopsy primarily based on age, as currently recommended in national guidelines, potentially misses many cases or delays diagnosis. Body mass index should be the first stratification in the decision to perform endometrial biopsy and/or to refer secondary gynecology services.
and frequently leads to an invasive diagnostic test such as endometrial biopsy or hysteroscopy. In a large study of premenopausal women with AUB referred for an endometrial biopsy, the prevalence of complex endometrial hyperplasia or cancer was 3.0%.
Endometrial hyperplasia is a precursor to endometrial cancer and is thought to result from persistent, prolonged, unopposed estrogenic stimulation of the endometrium, the most common cause being a succession of anovulatory cycles.
An important distinction in the evaluation of hyperplasia is whether or not nuclear atypia is present. Women with complex hyperplasia with atypia are at risk of progression to endometrial cancer and as many as 28% of cases will progress over time to cancer and up to 43% of cases have unrecognized concurrent carcinoma at the time of biopsy.
Commonly recognized risk factors for endometrial cancer include age, obesity, nulliparity, infertility, and late-onset menopause. Family history of hereditary nonpolyposis colorectal cancer is another risk factor. Diabetes and hypertension are frequently associated with endometrial cancer, while smoking and use of combined oral contraceptive pill are thought to be protective.
However, given that the majority of endometrial cancer occurs in postmenopausal women, it has not been widely recognized that obesity is a significant risk factor for endometrial hyperplasia and cancer in young, symptomatic, premenopausal women.
Developing a greater understanding of the leading risk factors for premenopausal women would lead to improved clinical pathways from primary to secondary care and improved targeting of invasive diagnostic testing. The objective of the current study was to evaluate the association of body mass index (BMI) and endometrial complex hyperplasia or cancer in premenopausal women with AUB who had an endometrial biopsy, adjusting for clinical and demographic factors. We hypothesized that obese women would be more likely to have complex hyperplasia or cancer compared to normal-weight women.
Materials and Methods
This was a retrospective cohort study including women age ≤55 years who had an endometrial biopsy at Auckland District Health Board from 2008 through 2014. Women were included if they had a history of AUB (clinical diagnosis by the referring doctor, including structural and nonstructural causes) and were not menopausal (either recorded as menopausal in the clinical record, or amenorrhea ≥6 months, or serum follicle-stimulating hormone level >20 IU/L). Women were excluded if they were known prior to 2008 to have had endometrial cancer. Women who were not residents of Auckland District Health Board were also excluded due to the referral pattern of surrounding district health boards to Auckland for tertiary cancer services. The primary study outcome was histologic diagnosis of endometrial biopsy of complex hyperplasia, complex atypical hyperplasia, or endometrial cancer, as stated in the pathology report.
Participants were identified from the hospital laboratory database of all inpatient and outpatient samples. The National Health Identifier was used to directly link with the hospital electronic medical records and clinical and demographic data were collected. The following clinical variables were collected: age, BMI, parity, self-reported use of hormone therapy, menstrual history, infertility (>12 months), medical history (smoking, diabetes, breast cancer, colorectal cancer), and family history (breast, colorectal, or endometrial cancer). Recent investigations (hemoglobin, pelvic ultrasound scan) and subsequent hysterectomy were also recorded.
Endometrial biopsy was performed by Pipelle (Pipelle De Cornier, Laboratoire CCD, Paris, France), sharp curettage, or both. If >1 biopsy was performed within 6 months, the one resulting in the most serious diagnosis was considered the final outcome. If a hysterectomy was performed within 6 months of the biopsy, the hysterectomy histology was considered the final outcome. The exception to this was in the setting where the biopsy reported hyperplasia, there was clear documentation that the patient was treated with progestogen in the interim, and the hysterectomy histology was normal. In that case, due to the known effectiveness of progestogen in treating endometrial hyperplasia,
then the original biopsy result was considered the final outcome.
Self-reported ethnicity was collected at hospital registration with a standard New Zealand (NZ) census question. The concept of ethnicity adopted by Statistics NZ is a social construct of group affiliation and identity. The present statistical standard for ethnicity states that ethnicity is the ethnic group or groups that people identify with or believe they belong to; thus, ethnicity is self-perceived and people can belong to >1 ethnic group.
Height and weight measurements used for calculating BMI were measured within 1 year of biopsy. Standard World Health Organization criteria were used to categorize BMI (normal 18.5–24.9 kg/m2, overweight 25–29.9 kg/m2, obese ≥30 kg/m2).
Deprivation centiles (1–10) were condensed into quintile scores from 1 (least deprived) to 5 (most deprived).
We determined the overall incidence of endometrial complex hyperplasia or cancer and compared crude rates across each clinical and demographic variable using simple logistic regression. Odds ratios (OR) and 95% confidence limits were estimated for each of these associations. In all cases, a P value of .05 was considered statistically significant. Multivariable analysis using logistic regression was performed to assess the independent association between each variable and the outcome. Included in the multivariable model were BMI and age (for a priori reasons) and those factors significantly associated with the outcome on univariate analysis.
As ethnicity and BMI are highly correlated in this population, only 1 of these variables was able to be retained in the multivariable model. We chose BMI as: (1) it meant a smaller number of degrees of freedom (2 vs 5), and (2) the use of BMI is more generalizable to other populations. Endometrial thickness was missing in one third of participants, hence we produced multivariable models both excluding and including endometrial thickness. Further sensitivity analysis was performed using the outcome of complex atypical hyperplasia or cancer. Data analysis was performed in software (SAS, Version 9.4; SAS Institute Inc, Cary, NC).
Based on a power of 80%, a level of significance of 5%, and an estimated prevalence of complex hyperplasia or cancer 3.0% and of obesity 30% (in the previous NZ study),
the current study with a sample of 840 subjects had the power to detect a relative risk of 1.89.
This study was approved on Oct. 5, 2012, by the University of Auckland Human Participants Ethics Committee (Ref. 8651).
The study cohort comprised 916 women who met the inclusion and exclusion criteria (Figure 1). Half of the women were obese. Obese women were more likely to be <40 years old, of Māori and Pacific ethnicity, and residing in areas of high deprivation. Characteristics of participants are shown in Table 1.
In the whole cohort, 85% of women had a Pipelle, 22% had sharp curettage, and 7% had both. Endometrial biopsy results are shown in Table 2. Of 840 women with sufficient tissue for diagnosis, 41 women (4.9%) were diagnosed with endometrial complex hyperplasia, complex atypical hyperplasia, or cancer. This is significantly higher than the incidence reported in the previous NZ study (P = .035).
Table 2Results of endometrial biopsy where tissue sample was sufficient for diagnosis (N = 840)
OR for the association of demographic and clinical factors with complex hyperplasia or cancer are shown in Table 3. Women of Indian and Pacific ethnicity had significantly higher odds of having complex hyperplasia or cancer compared to European women (OR, 3.72; 95% confidence intervals [CI], 1.07–13.00, and OR, 6.21; 95% CI, 2.11–18.32, respectively). A more in-depth investigation of the relationship of ethnicity to BMI found that we were unable to distinguish between Pacific ethnicity and BMI, due to the high prevalence of obesity in the Pacific sample (92.2%).
Table 3Demographic and clinical factors associated with endometrial complex hyperplasia or cancer
Cases, N, or mean (SD)
Controls, N, or mean (SD)
Odds ratio (95% confidence interval)
Age fitted as cubic term, y
χ2 = 7.29, P = .063
NZ deprivation score [missing = 244]
Body mass index
Duration of AUB, y
Cycle regularity [missing = 21]
Cycle length, d
Intermenstrual bleeding [missing = 24]
Current use of hormonal therapy
Diabetes [missing = 133]
Family history of breast cancer
Family history of colorectal cancer [missing = 92]
Numbers of women of “other” ethnicity, with personal history of breast or colorectal cancer, and with family history of endometrial cancer were too small to perform univariate analysis.
AUB, abnormal uterine bleeding; NZ, New Zealand.
Wise et al. BMI and endometrial pathology in premenopause. Am J Obstet Gynecol 2016.
Other factors associated with an increased risk of complex hyperplasia or cancer were nulliparity (OR, 2.51; 95% CI, 1.25–5.05), anemia (OR, 2.38; 95% CI, 1.25–4.56), and thickened endometrium on ultrasound scan (OR, 4.04; 95% CI, 1.69–9.65). We assessed the relationship of age using a generalized additive model and found the relationship to be cubic in nature with the highest risk association with the youngest women (mid-20s to mid-30s) and the oldest women (Figure 2). The overall association of age fitted in this form did not quite reach statistical significance (P = .063). Diabetes, hypertension, and menstrual history showed no effect.
The multivariable analyses are found in Table 4. After adjusting for age, anemia, and nulliparity, obese women had significantly higher odds of having complex hyperplasia or cancer compared to women with normal BMI (adjusted OR, 4.00; 95% CI, 1.36–11.74), while overweight women showed no increased risk. Nulliparous women were at increased risk (adjusted OR, 3.08; 95% CI, 1.43–6.64), as were women with anemia (adjusted OR, 2.23; 95% CI, 1.14–4.35). Age remained nonsignificant.
Table 4Multivariable model of risk factors for endometrial complex hyperplasia or cancer
In a sensitivity analysis including endometrial thickness in the model, the adjusted OR for the above variables did not change much, and women with endometrium thickness ≥12 mm had significantly higher odds of having complex hyperplasia or cancer (adjusted OR, 4.20; 95% CI, 1.58–11.15). In a sensitivity analysis including only complex atypical hyperplasia or cancer as the outcome, the point estimates of the OR did not differ in any meaningful way from those in Table 4 (data not shown).
This large retrospective cohort study has reported a 4.9% incidence of endometrial complex hyperplasia, complex atypical hyperplasia, or cancer in premenopausal women with AUB who underwent endometrial biopsy. Obese women were 4 times more likely than normal-weight women to have complex hyperplasia or cancer. Other significant risk factors were nulliparity, anemia, and thickened endometrium on ultrasound scan; age was not a significant risk factor.
The increased incidence of complex hyperplasia or cancer in our hospital from 3.0% (1995 through 1997)
Given the overall obesity rate in this study of 49%, and a univariate OR of 3.76, the population-attributable risk associated with obesity in this study is 48%. This shows the potential impact on complex hyperplasia or cancer that could be achieved by reducing obesity rates. These findings are also consistent with the increasing endometrial cancer rates from 2002 through 2012, from an age-standardized rate of 12.2 in 2002 to 16.2 per 100,000 in NZ,
In the only meta-analysis of endometrial cancer in women who were premenopausal at time of diagnosis, for the studies that reported on women with BMI ≥30 kg/m2 the pooled OR was 5.25 (95% CI, 4.00–6.90).
None of these reviews included endometrial hyperplasia as an outcome.
The main strength of this study is its cohort design. This is the more appropriate and more robust study design to answer the question of risk factors for complex hyperplasia or cancer in premenopausal women presenting with AUB, rather than case-control design of evaluating premenopausal women with and without endometrial pathology to explore the association with obesity. BMI was the primary variable of interest, and was measured around the time of the biopsy, in contrast to many other studies where self-reported BMI was used, which is not as accurate.
In addition, we were able to adjust for several demographic and clinical variables known to be associated with the outcome. To our knowledge, this is the first cohort study of premenopausal women with AUB with biopsy-proven endometrial hyperplasia or cancer, comparing obese to nonobese women.
Limitations of this study include that data were collected retrospectively from the electronic medical records of women referred to a secondary gynecology service at a single urban hospital. Ideally, a prospective cohort study of women presenting with AUB in the community setting could be performed to determine the true incidence of complex hyperplasia or cancer, however, this would be challenging given the large number of practices that would be involved. As with all observational research, there may be unmeasured and unknown confounders not included in the analysis that could contribute to the findings. In addition, some variables found to be nonsignificant may be due to small numbers, for example, young healthy women are not usually screened for diabetes. The diagnosis of atypical complex hyperplasia has low reproducibility,
and histology was not reviewed by a pathologist for the purposes of this study; however, this makes our findings more generalizable, and further strengthens our argument to include BMI to stratify risk. These findings may not be generalizable to all women with AUB, rather only those referred to a gynecologist and who had an endometrial biopsy; in our setting, about one third of women referred for AUB had an endometrial biopsy. These findings may also not be generalizable to populations with a different ethnic mix, given different possible genetic risks. However, we believe these findings are generalizable to a population of women in a country with high rates of obesity, and similar pattern of referral from primary to secondary services.
These findings could be used to inform current guidelines for the investigation of women with AUB. The US and Canadian guidelines on AUB recommend biopsy in women age >45 years or >40 years, respectively, or in younger women with risk factors for endometrial cancer, including obesity.
The 2007 National Institute for Health and Care Excellence guidelines recommend biopsy if indicated, for example, in cases of persistent intermenstrual bleeding, or in women aged ≥45 years who experience treatment failure or ineffective treatment.
All of these guidelines could now be updated to suggest that endometrial biopsy be performed in women with BMI ≥30 kg/m2 (or other clinical risk factors, eg, nulliparity or anemia).
Although age may be an important risk factor for endometrial cancer in perimenopausal and postmenopausal women, age does not appear in our study to be a risk factor in premenopausal women. Thus, we suggest that BMI be used as the primary deciding factor to further investigate AUB with an invasive test. In our study, the approach of using a BMI threshold, instead of age, would have reduced the number of women in our population needing biopsies by half. Another potential benefit is detecting endometrial complex hyperplasia or cancer in younger women at an earlier stage in disease, allowing for more conservative treatment.
The authors wish to acknowledge Steve Cooke for searching the Auckland District Health Board LabPlus database; Dr Lynn Sadler, perinatal epidemiologist, for searching the Auckland District Health Board Healthware database; Baljit Singh for developing the Access database; research assistants Josephine Rishworth, Nicola Arroll, and Nicole Wong for assisting with data collection; Avinesh Pillai statistician for data cleaning and analysis; and research assistant Catherine Coop for assisting with preparing the manuscript, references, and tables for submission. These individuals have consented to this acknowledgement of their contribution to this manuscript.
This study was supported by the Mercia Barnes Trust , Royal Australian and New Zealand College of Obstetricians and Gynecologists . The funding source had no role in study design, collection, analysis or interpretation of data, writing of the article, or decision to submit for publication. Researchers were independent from the funding source.
The authors report no conflict of interest.
Cite this article as: Wise MR, Gill P, Lensen S, et al. Body mass index trumps age in decision for endometrial biopsy: cohort study of symptomatic premenopausal women. Am J Obstet Gynecol 2016;215:598.e1-8.
We read with great interest the article by Wise et al1 comparing the incidence of complex endometrial hyperplasia (EH) or cancer in premenopausal women with abnormal uterine bleeding and body mass index (BMI) ≥30 vs <30 kg/m2. They showed that a BMI ≥30 strongly associated with EH or cancer, and concluded that “BMI should be the first stratification in the decision to perform endometrial biopsy.”1