- Womb cancer
- What a risk factor is
- Being overweight
- Diet and alcohol
- Having had children
- Menstrual history
- Thickened womb lining
- Polycystic ovary syndrome
- Family History
- Previous cancer
- Hormone replacement therapy (HRT)
- The contraceptive pill
- Physical activity
- Other possible causes
- Reducing your risk
- Bleeding after menopause: It’s not normal
- Why you shouldn’t ignore postmenopausal bleeding
- Causes of postmenopausal bleeding
- Why you should see a gynecologic oncologist
- Make your health a priority
- Postmenopausal AUB: Rule out endometrial cancer first
- A Risk-Scoring Model for the Prediction of Endometrial Cancer among Symptomatic Postmenopausal Women with Endometrial Thickness > 4 mm
- The significance of sonographically thickened endometrium in asymptomatic postmenopausal women
- Endometrial thickness
- Abnormal uterine bleeding in peri- and postmenopausal women
Womb cancer is the 4th most common cancer in women in the UK. We don’t know what causes most womb cancers. But there are some factors that can increase your risk of developing it.
What a risk factor is
Anything that increases your risk of getting a disease is called a risk factor. Different cancers have different risk factors.
Having a risk factor does not necessarily mean that you will develop cancer. Also, not having any risk factors does not mean that you definitely won’t get cancer.
Some factors lower your risk of cancer and are known as protection factors.
The risk of womb cancer increases with age. Most women diagnosed with womb cancer have had their menopause. And almost three quarters of cases of womb cancer are in women aged 40 to 74.
Just over 1 in 100 cases (1%) are diagnosed in women under 40. Women who have a particular gene fault called Lynch syndrome are more likely to develop endometrial cancer at a younger age than the general population. Lynch syndrome used to be called hereditary nonpolyposis colorectal cancer.
Oestrogen is a female hormone. Before the menopause the ovaries make most of the oestrogen a woman needs. Along with another female hormone called progesterone it regulates womens reproductive cycle.
After the menopause the ovary stops producing hormones. But the body continues to make a small amount of oestrogen. Fat cells also make oestrogen.
Oestrogen causes the cells in the womb to divide increasing the risk of mistakes being made. So anything that increases the amount of oestrogen in your body increases your risk of womb cancer.
Being overweight or obese is the biggest preventable risk factor of womb cancer. Women who are very overweight (obese) are 2.5 times more likely to develeop womb cancer than women of a healthy weight.
A Cancer Research UK study published in 2011 found that being overweight or obese causes around a third of womb cancers in the UK each year.
Higher levels of oestrogen
Overweight women have higher levels of oestrogen. Fat cells covert hormones into a type of oestrogen. So the more body fat you have, generally the more oestrogen you produce. When more oestrogen is produced, the lining of the womb builds up. When more lining (endometrial) cells are produced, there is a greater chance of one of them becoming cancerous.
Another reason for the higher risk of womb cancer in overweight women may be related to Insulin.
Insulin helps the body to unlock and use the energy in food that we eat. People who are overweight can sometimes become resistant to insulin. This means that although the body produces insulin, the insulin doesn’t work as well as usual. To make up for this, the body makes too much.
Some research has shown higher levels of insulin are linked to an increased risk of womb cancer. The extra insulin tells cells to divide more quickly in the womb lining, increasing the chance of cancer cells being produced.
Diet and alcohol
Studies have looked at whether diet could affect womb cancer risk. At the moment there are no convincing dietary factors that directly increase or decrease your womb cancer risk. But a healthy diet helps you keep a healthy weight, which in turn reduces the risk of womb cancer.
Coffee has also been linked to a reduced womb cancer risk. But overall the evidence is not strong.
An analysis of studies hasn’t shown a link between drinking alcohol and the risk of womb cancer. But alcohol increases the risk of many other types of cancer.
Having had children
Studies show having children lowers womb cancer risk by around a third. The risk decreases with the more children a women has.
Oestrogen levels are low and progesterone levels are high during pregnancy. During the menstrual cycle, there is oestrogen in the body without progesterone. This is called unopposed oestrogen.
Unopposed oestrogen increases womb cancer risk. So anything that stops this (such as pregnancy) lowers the risk of womb cancer.
Some factors linked with periods (menstruation) can increase your risk of womb cancer because they cause higher levels of oestrogen. These include:
- starting your period at a young age
- a late menopause
Thickened womb lining
Endometrial hyperplasia is a non cancerous (benign) condition where the lining of the womb becomes thicker. You have a higher risk of developing womb cancer if you have this thickening, especially if the extra lining cells are abnormal.
Symptoms of endometrial hyperplasia are heavy periods, bleeding between periods, and bleeding after menopause.
Polycystic ovary syndrome
Women with polycystic ovary syndrome (PCOS) have a hormone imbalance which may cause very irregular periods. Women with PCOS have an increased risk of womb cancer compared to women who don’t have PCOS. The cause of the increase is unclear but it may be due to the hormone imbalance.
Polycystic ovary syndrome is also linked with insulin resistance, being overweight and type 2 diabetes. These are risk factors for womb cancer.
Research has shown that daughters of women with womb cancer have double the risk of women in the general population.
If you have several close relatives on the same side of the family who have had bowel cancer or womb cancer you may be at increased risk of womb cancer.
Lynch syndrome is an inherited faulty gene linked with an increased risk of some cancers, including bowel cancer and womb cancer. Out of every 100 women who carry this rare gene fault, 40 to 60 will develop womb cancer at some point in their lives.
You are at a slightly increased risk of getting womb cancer if you have had breast or ovarian cancer in the past. This may be because of shared lifestyle or environmental factors. For example, the risk of womb cancer in breast cancer survivors is higher in those who are overweight. The increase in risk of womb cancer may also be due to treatment for cancer, such as tamoxifen.
Tamoxifen is a hormone therapy for some types of breast cancer. It can increase womb cancer risk, as it is thought to have a similar effect to oestrogen on the womb. This is a rare side effect of taking it.
And the benefits of taking tamoxifen as part of your treatment for breast cancer outweigh the small risk of womb cancer.
If you are taking tamoxifen, tell your doctor if you have:
- unexpected vaginal bleeding
- vaginal bleeding after your periods have stopped
Hormone replacement therapy (HRT)
There are different types of hormone replacement therapy (HRT). Oestrogen only HRT increases the risk of womb cancer. Because of this, doctors normally only prescribe oestrogen only HRT for women who have had their womb removed (a hysterectomy).
Combined HRT contains the hormones oestrogen and progesterone. There is evidence that the progesterone part can counteract the cancer causing effects of the oestrogen part. But it depens on the type of combined HRT used.
The contraceptive pill
The combined pill, the most common type of birth control pill, is linked with a reduced risk of womb cancer. These protective effects are bigger the longer a woman takes the combined pill for. They can continue for decades after she stops taking it.
Using a non hormonal intrauterine device (IUD or coil) has also been linked with a decreased risk of womb cancer.
Several studies show a higher risk of womb cancer in women with diabetes, for both Type 1 and Type 2. This link may be due to being overweight and we need more research to find out about why it increases risk.
The world cancer research fund has listed physical activity as probably being protective against womb cancer.
This link may partly be because women who are more active have a lower body weight. Being physically active also helps to control hormones in the body, such as oestrogen and insulin.
Aspirin may reduce the risk of womb cancer in women who are very overweight (obese). But more studies are needed to confirm this link.
At the moment there are no recommendations to take aspirin to reduce womb cancer risk.
Other possible causes
Stories about potential causes of cancer are often in the media and it isn’t always clear which ideas are supported by evidence. There might be things you have heard of that we haven’t included here. This is because either there is no evidence about them or it is less clear.
Reducing your risk
There are ways you can reduce your risk of cancer.
Bleeding after menopause: It’s not normal
Too often I see women with advanced endometrial cancer (uterine cancer) who tell me they experienced postmenopausal bleeding for years but didn’t think anything of it. This shows we need to do a better job educating our patients about what to expect after menopause.
Women need to know postmenopausal bleeding is never normal, and it may be an early symptom of endometrial cancer. Any bleeding, even spotting, should trigger a visit to your doctor as soon as possible. Don’t wait to make an appointment until after the holidays or even next week. Do it today.
Why you shouldn’t ignore postmenopausal bleeding
A woman is considered to be in menopause after 12 consecutive months without a period. You may experience irregular bleeding leading up to menopause, a stage known as perimenopause. But once you’re in menopause, all vaginal bleeding should stop.
There are benign causes of postmenopausal bleeding. For 10 percent of women, however, the cause is endometrial cancer.
Early diagnosis offers the best chance to beat endometrial cancer. I urge women to treat postmenopausal bleeding as cancer until proven to be something else. I don’t say this to scare people, but a healthy amount of worry in this situation is warranted.
Causes of postmenopausal bleeding
A variety of conditions can cause postmenopausal bleeding, including hormone therapy, infection, or the use of medications such as blood thinners.
Some of the most common causes of postmenopausal bleeding are:
- Cancer: Postmenopausal bleeding is a common symptom of endometrial cancer, but it also can be caused by cervical and vulvar cancer.
- Endometrial atrophy: The tissue that lines the uterus can become very thin after menopause. As the lining thins, bleeding may occur.
- Endometrial hyperplasia: Sometimes the lining of the uterus becomes thick, usually because of too much estrogen and too little progesterone. This is considered a precursor to endometrial cancer.
- Fibroids: These growths develop in the uterine muscle tissue.
- Polyps: These usually noncancerous growths can develop in the lining of the uterus.
To diagnose the cause of your postmenopausal bleeding, the doctor will perform a physical exam and may order one or more of the following tests:
- Endometrial biopsy: A thin tube is inserted into the uterus and a tiny sample of the lining is removed. The sample is sent to the lab for examination.
- Transvaginal ultrasound: An imaging device is inserted into the vagina to examine the pelvic organs.
- Hysteroscopy: An instrument with a light and camera at the end, called a hysteroscope, is inserted into the vagina and through the cervix to examine the uterus.
- Dilation and curettage (D&C): After enlarging the cervix, tissue is scraped from the lining of the uterus to be examined in a lab.
Treatment for postmenopausal bleeding depends on its cause. For example, if you have polyps, simply removing them may correct the problem. But if the diagnosis is cancer, you should seek care from a gynecologic oncologist.
Why you should see a gynecologic oncologist
When postmenopausal bleeding is diagnosed as endometrial cancer, most cases can be cured with a hysterectomy. However, because endometrial cancer can spread into the lymph nodes, many patients also should have a lymph node dissection at the time of hysterectomy. Gynecologic oncologists are specifically trained to perform this procedure when it is indicated.
If only a hysterectomy is performed and it turns out the lymph nodes are at risk, we’re left with difficult decisions. Should the patient start radiation therapy, or should she go back into the operating room to perform the lymph node dissection? Seeing a gynecologic oncologist immediately after diagnosis can avoid these complications, simplifying care and improving the chance of survival.
It’s not always easy to travel to a gynecologic oncologist’s office. Dallas-Fort Worth residents are lucky in this respect, as there are a number of us in the area. I have patients who come from several hours away because we’re the closest available clinic. While making the trip to see a gynecologic oncologist may be inconvenient, it’s important for your care.
Make your health a priority
Women are known to focus on their families first and put their own health second. But you can’t care for loved ones if you’re not healthy yourself. Listen to your body. Alert your doctor to any changes or abnormal issues – such as postmenopausal bleeding – as soon as possible.
Don’t stop seeing your general gynecologist for an annual exam when you hit menopause. Just because your reproductive years have ended doesn’t mean those body parts go away! Your cancer risk increases as you age, and your gynecologist can screen for the disease and help you manage any conditions caused by hormone changes.
If you’re experiencing postmenopausal bleeding or have any concerns about your gynecologic health, request an appointment online or by calling 214-645-8300.
Postmenopausal AUB: Rule out endometrial cancer first
- Uterine bleeding or spotting after the initiation of hormone replacement therapy (HRT) is not unusual.
- Endometrial evaluation is called for in women not taking HRT who develop uterine bleeding after more than 1 year of amenorrhea. It also is indicated in postmenopausal women on HRT for more than 6 months with persistent uterine bleeding, and previously amenorrheic women on HRT who begin bleeding without apparent cause.
- Screening asymptomatic women for endometrial cancer through transvaginal ultrasound or endometrial biopsy is not recommended.
- If endometrial thickness is greater than 4 mm, sonohysterography should be performed.
Approximately 1 of every 8 post-menopausal women who present with abnormal uterine bleeding (AUB) will be diagnosed with endometrial cancer, making this one of the most troubling symptoms clinicians encounter in gynecologic practice. Because a wide variety of pathophysiologic problems can cause AUB in post-menopausal women, these patients require prompt evaluation. At a minimum, this should include a clinical history and physical examination, as well as endometrial sampling or evaluation by ultrasound—especially for women not taking hormone replacement therapy (HRT). This process enables clinicians to detect endometrial cancer at an early stage, before it has spread beyond the uterus. Early detection is associated with an expected survival rate of 90%.1
Over the past decade, the evaluation and treatment of postmenopausal bleeding has evolved significantly, thanks to the availability of office-based transvaginal ultrasound (TVUS) and saline-infusion sonohysterography, as well as the advent of endometrial ablation (TABLE 1). Here, I discuss diagnostic testing and review conditions that can lead to postmenopausal bleeding. Once the etiology is identified, the choice of therapeutic intervention usually is self-evident.
Tests for evaluating postmenopausal AUB
- Endometrial biopsy
- Transvaginal ultrasound
- Saline-infusion sonohysterography
- Hysteroscopy and directed endometrial biopsy
Our understanding of the pathophysiology of AUB is in its infancy. Here’s what we do know: During a woman’s reproductive years, her endometrium constantly remodels itself under the influence of estrogen, which stimulates cellular growth, and progesterone, which antagonizes the estrogen’s growth effects. When a woman is anovulatory, however, the endometrium can be stimulated by continual estrogen exposure, resulting in endometrial proliferation. This leads to endometrial instability and, in turn, uterine bleeding. Several cellular signaling molecules—such as cytokines, growth factors, and matrix metalloproteinases—are involved. Once menopause occurs, estrogen and progesterone are no longer produced by the ovaries; nor are they produced in any appreciable amounts by the liver and fat. The endometrium regresses to some degree, and no further bleeding should occur. When bleeding does resume, therefore, abnormal pathologies must be investigated.
Endometrial hyperplasia and cancer
Endometrial cancer is the most common gynecologic cancer in the United States, affecting 21 of every 100,000 women, according to the National Cancer Institute’s (NCI) Surveillance, Epidemiology, and End Results (SEER) Program.2 Fortunately, because AUB is an early harbinger of the disease, endometrial cancer generally has a favorable prognosis: When all endometrial cancer cases are considered, the 5-year survival rate is 84%, though the prognosis for individual patients depends on a variety of factors, including the initial stage of the endometrial cancer as well as the cellular differentiation.
Over the past 2 decades, the rate of endometrial cancer has decreased by more than 26%.2 This may be because many women are entering menopause who previously used combination oral contraceptives (OCs) containing progestin, which is associated with a reduced risk of endometrial hyperplasia and cancer. The increased use of progestin in HRT regimens in recent years also likely plays a role.
Risk factors for endometrial cancer are conditions typically associated with chronic elevations of endogenous estrogen levels or increased estrogen action at the level of the endometrium. These include obesity, history of chronic anovulation, diabetes mellitus, estrogen-secreting tumors, exogenous estrogen unopposed by progesterone or progestin, tamoxifen use, and a family history of Lynch type II syndrome (hereditary nonpolyposis colorectal, ovarian, or endometrial cancer). Since endometrial cancer may occur in the absence of risk factors, however, they should not be the sole means of identifying patients.
Since the cost-effectiveness of endometrial biopsy and transvaginal ultrasound are comparable, the choice of modality rests with the physician.
Diagnostic tests. Endometrial evaluation is called for when any menopausal woman not taking HRT develops uterine bleeding after more than 1 year of amenorrhea. It also is indicated in any postmenopausal woman on HRT for 6 months or more with persistent uterine bleeding, and any previously amenorrheic woman on HRT who begins bleeding without apparent cause.
The measurement of endometrial thickness by TVUS is now almost standard in the evaluation of postmenopausal bleeding. However, screening asymptomatic women for endometrial cancer with TVUS is not recommended. Although the test is very specific (the number of false positives is exceedingly small when the endometrial thickness is set at less than 5 mm), it isn’t sensitive. Many women without endometrial cancer will have an endometrial thickness of 5 mm or more.
A Risk-Scoring Model for the Prediction of Endometrial Cancer among Symptomatic Postmenopausal Women with Endometrial Thickness > 4 mm
Objective. To develop and test a risk-scoring model for the prediction of endometrial cancer among symptomatic postmenopausal women at risk of intrauterine malignancy. Methods. We prospectively studied 624 postmenopausal women with vaginal bleeding and endometrial thickness > 4 mm undergoing diagnostic hysteroscopy. Patient characteristics and endometrial assessment of women with or without endometrial cancer were compared. Then, a risk-scoring model, including the best predictors of endometrial cancer, was tested. Univariate, multivariate, and ROC curve analysis were performed. Finally, a split-sampling internal validation was also performed. Results. The best predictors of endometrial cancer were recurrent vaginal bleeding (odds ratio ), the presence of hypertension endometrial thickness > 8 mm , and age > 65 years . These variables were used to create a risk-scoring model (RHEA risk-model) for the prediction of intrauterine malignancy, with an area under the curve of 0.878 (95% CI 0.842 to 0.908; ). At the best cut-off value (score ≥ 4), sensitivity and specificity were 87.5% and 80.1%, respectively. Conclusion. Among symptomatic postmenopausal women with endometrial thickness > 4 mm, a risk-scoring model including patient characteristics and endometrial thickness showed a moderate diagnostic accuracy in discriminating women with or without endometrial cancer. Based on this model, a decision algorithm was developed for the management of such a population.
It is known that about 90–95% of postmenopausal women with endometrial cancer report a vaginal bleeding experience , whereas about 10% of symptomatic postmenopausal women reveal an intrauterine malignancy . So, a postmenopausal vaginal bleeding is a sign that should not be underestimated. In this regard, a good clinical practice provides, as first diagnostic step, a transvaginal ultrasound in order to discriminate a woman at high or low risk of malignancy.
Usually, an endometrial thickness ≤ 4 mm is a cut-off value for which a conservative management should be adopted. Indeed, in the latter case the posttest probability of having an endometrial cancer drops from 10% to 0.8% . Conversely, among symptomatic postmenopausal women with endometrial thickness > 4 mm, there is an increased risk of cancer . In these cases, further examinations are needed and, usually, an endometrial sampling or an outpatient hysteroscopy should be performed. However, approximately 80–90% of these examinations will not reveal a cancer in a population considered at risk of malignancy . This apparent “inappropriateness” is justified by the fact that our goal is to miss the lowest number of women with cancer.
Despite keeping this important objective in mind, one wonders if there are clinical variables that can improve the diagnostic performance of our procedures. Several studies including patient characteristics or sonographic features were performed in order to test their clinical usefulness. Some authors included, as study participants, all postmenopausal women with vaginal bleeding, whereas other authors included only symptomatic postmenopausal women with an endometrial thickness at risk of intrauterine malignancy . The majority of these studies showed fair outcomes with an improvement of diagnostic performance in detecting endometrial cancers. However, to date, these models are not yet validated externally, for which endometrial thickness remains the most important feature to be evaluated in these cases. It is likely that endometrial thickness assessment, along with further predictive factors, could provide better results in the prediction of intrauterine malignancy among high-risk women.
In this regard, the aim of the present study was to create and test a risk-scoring model, including endometrial assessment and patient characteristics, among symptomatic postmenopausal women with endometrial thickness > 4 mm, and furthermore, to develop a decision algorithm for the management of such a population.
2. Materials and Methods
This prospective observational study included 624 symptomatic postmenopausal women with endometrial thickness > 4 mm undergoing diagnostic hysteroscopy. The present study was performed at Cesare Magati Hospital, Division of Obstetrics and Gynecology, Scandiano, and University Hospital, Institute of Obstetrics and Gynecology, Modena, Italy, from March 2008 to November 2013. Our Institutional Review Board approved this study and each woman gave an informed consent.
Each postmenopausal woman with vaginal bleeding was subjected to transvaginal ultrasound. The latter examination was performed using a 5–9 MHz vaginal transducer and the thickest part of the anteroposterior bilayer endometrial thickness was measured in the sagittal plane. Furthermore, endometrial echogenicity was evaluated and defined according to the IETA terms (uniform or nonuniform) .
Based on our Protocol which suggests further evaluations in all cases with an endometrial thickness > 4 mm, we recruited only those women then subjected to diagnostic hysteroscopy. We excluded all symptomatic postmenopausal women with a vaginal bleeding arising from a cervical or vaginal or vulvar disease, as well as all vaginal bleedings due to hormone replacement therapy (HRT). Conversely, all postmenopausal women under HRT with unscheduled vaginal bleeding were included in the study. Postmenopausal status was defined as the absence of menstruation for at least 12 months after the age of 40 years, where any pathological condition of amenorrhea was excluded.
All eligible women, after transvaginal ultrasound, filled out a questionnaire for their medical history including age; age at menarche; age at menopause; time since menopause; body mass index (BMI = weight (kg)/height2 (m2)); parity; presence of hypertension or diabetes; HRT, anticoagulant, or tamoxifen use; history of breast cancer; recurrent vaginal bleeding or single episode; endometrial thickness; and echogenicity. Based on previous studies, recurrent vaginal bleeding was defined as any bleeding that lasted seven or more days, or two or more separate episodes of vaginal bleeding over the last year .
All symptomatic postmenopausal women with endometrial thickness > 4 mm were subjected to diagnostic outpatient hysteroscopy in vaginoscopy with a saline solution as distension medium and narrow instrumental diameters. The latter examination was performed by an experienced hysteroscopist who was blinded to the ultrasound findings. Each woman was subjected to an endometrial sampling which we considered our reference standard. Based on our previous study , a Vabra endometrial sampling was performed in women without any intrauterine lesion; a targeted biopsy along with random biopsies of each uterine wall was performed in women with suspected premalignant or malignant lesion; intrauterine lesion resection was performed in women with polyps or myomas; all women with an atypical endometrial hyperplasia (AEH), as well as all women with an intrauterine malignancy, underwent a hysterectomy which represented our reference standard as definitive histological finding.
The Kolmogorov-Smirnov test was used as test for normal distribution. Nonparametric Mann-Whitney test was performed to compare continuous variables with nonnormal distribution. Categorical variables were evaluated by χ2 analysis or Fisher’s exact test where appropriate. Variables that showed significant differences in univariate analysis () were the candidate predictor variables for the stepwise logistic regression analysis including both forward and backward selections. In order to create a parsimonious model, we used an entrance and exit value of 0.05/0.05. Then, to test the goodness of fit for the logistic regression model, the Hosmer-Lemeshow test was performed considering the fact that a large value of Chi-squared (with small value < 0.05) indicates poor fit.
In order to overcome some limitations of the stepwise method, such as variable selection, uncertainty about the variables, and overfitting, and based on our sample size (624 women), we performed a split-sampling internal validation . We divided our cohort into two, trying to maintain the same number of endometrial cancers in the two halves of our sample, and developed the model on one half (training sample) and tested it on the other (validation sample). We evaluated whether the stepwise regression of the training sample produced the same subset of predictors produced by the regression model of the full dataset . Then, we compared the coefficient of determination () between the training and validation sample ( for the 50% training sample— for the 50% validation sample). If the shrinkage was 2% (0.02) or less, validation was considered successful . If so, we derived the final prediction model from the full derivation sample . The coefficient of determination of the training and validation sample was obtained by multiple regression analysis.
Receiver operating characteristic (ROC) curve analysis was used to determine the optimal cut-off value of predictive continuous variables associated with endometrial cancer. According to the predictive odds ratio of each variable obtained in the multivariate analysis, a score for each significant predictive factor was assigned. Then, a ROC curve analysis was performed identifying the score as the variable under study. For each score, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (LR+), and negative likelihood ratio (LR−) were reported. After considering our disease prevalence (all cases of endometrial cancer) as the pretest probability for endometrial cancer, the likelihood ratio was used to calculate the posttest odds from the pretest odds of disease: posttest odds = pretest odds × likelihood ratio. The relation between odds and probability is odds = and = odds/(). Using these equations, we could calculate the posttest probability of disease from the pretest probability of disease .
Statistical analyses were performed with MedCalc (MedCalc Software, Mariakerke, Belgium). A value of less than 0.05 was considered to be statistically significant.
We enrolled 648 symptomatic postmenopausal women with endometrial thickness > 4 mm referred to diagnostic hysteroscopy. 24 women were excluded from this prospective study because a cervical canal stenosis made impracticable an outpatient hysteroscopy for intolerable pain. So, 624 participants were included for our statistical analysis.
Histological examination revealed the presence of 157 (25.2%) women with endometrial atrophy, 275 (44.1%) cases of endometrial polyps, 58 (9.3%) women with submucosal myomas, 62 (9.9%) cases of endometrial hyperplasia (15 cases of complex hyperplasia with atypia, 9 cases of simple hyperplasia with atypia, 22 cases of complex hyperplasia without atypia, and 16 cases of simple hyperplasia without atypia), and 72 women (11.5%) with endometrial cancer.
Patient characteristics showed no significant differences with regard to age at menarche, age at menopause, BMI, parity, diabetes, tamoxifen and anticoagulant use, and breast cancer history (Table 1). Conversely, significant differences were present with regard to age (), time since menopause (), HRT use (), recurrent vaginal bleeding (), presence of hypertension (), endometrial echogenicity (), and endometrial thickness () (Table 1).
Table 1 Univariate analysis comparing clinical variables and endometrial assessment between women with ( ) or without ( ) endometrial cancer.
The seven variables that showed significant difference in univariate analysis were included in multivariate analysis (age, time since menopause, HRT use, recurrent vaginal bleeding, presence of hypertension, endometrial echogenicity, and endometrial thickness). Then, stepwise logistic regression analysis showed the significant predictive variables associated with endometrial cancer (acronym, RHEA): R for recurrent vaginal bleeding (, confidence interval 1.32–6.66, ); H for the presence of hypertension (, confidence interval 1.10–4.50, ); E for endometrial thickness (, confidence interval 1.18–1.45, , criterion > 8 mm); and A for age (, confidence interval 1.07–1.15, , criterion > 65 years) (Table 2). To test the goodness of fit for the logistic regression model, the Hosmer-Lemeshow test was performed and showed a value of 0.218.
Table 2 Multivariate analysis showing clinical and endometrial variables associated with intrauterine malignancy.
A split-sampling internal validation was performed. The same predictors of the full dataset (recurrent vaginal bleeding, age, endometrial thickness, and hypertension) were produced after the stepwise regression of the training sample. Then, a multiple regression analysis was performed to obtain the coefficient of determination () for the training and validation sample. The shrinkage between training and validation sample () was 0.017 (≤2%), and validation was considered successful. We based our interpretation on the model that included all cases.
Table 3 Sensitivity, specificity, PPV, NPV, LR+, LR−, pre-, and posttest probability for each score of our risk-scoring model.
ROC curve associated with the risk-scoring model. The area under the curve was 0.878 (95% CI to 0.908; ).
According to the accuracy of diagnostic systems , the present study showed that a risk-scoring model, including recurrent vaginal bleeding, endometrial thickness > 8 mm, presence of hypertension, and age > 65 years, called RHEA, provided a moderate diagnostic accuracy for the prediction of intrauterine malignancies among symptomatic postmenopausal women at risk of endometrial cancer. At a cut-off score ≥ 4, we obtained a posttest probability of 1.9%, as percentage of missed cancers, and a posttest probability of 35.1%, as percentage of having cancer, from a pretest probability of 11.5%.
Strengths and Weaknesses of the Study. We performed a prospective assessment of our women which allowed us to standardize any type of examination, so as to have more reliable data. Furthermore, all our women had a definitive histological diagnosis with an optimal reference standard. Conversely, it is true that some patient characteristics were collected retrospectively, with clinical questions to our women about past events (e.g., recurrent vaginal bleeding).
We chose symptomatic postmenopausal women with endometrial thickness > 4 mm because women with a lower endometrial thickness have a very low incidence of cancer and, usually, do not perform further examinations in our centers. So, in order not to include in our sample women without a histological diagnosis as reference standard, we selected only women then subjected to hysteroscopy.
In a previous study, including symptomatic postmenopausal women, Bruchim et al. showed that an endometrial thickness of 5–9 mm revealed a cancer in 10% of cases only. For an endometrial thickness > 9 mm, the percentage of cancer reached 18% . These results are in line with the best cut-off value of our prediction model, where an endometrial thickness ≥ 9 mm was one of the predictors associated with endometrial cancer.
In a very interesting study, Opolskiene et al. compared different prediction models for endometrial cancer among postmenopausal women with vaginal bleeding and endometrial thickness ≥ 4.5 mm . They reached the conclusion that, adding endometrial thickness and power Doppler information to patient characteristics, the diagnostic performance of prediction models increased significantly. Concerning this latter study, if we consider only their prediction model including endometrial thickness and clinical variables, we can note that the AUC of their model was similar to that of our risk-scoring model (0.82 and 0.87, resp.). Also Opmeer et al. showed that, taking into account patient characteristics (age, time since menopause, BMI, and diabetes) and endometrial thickness, the appropriateness of their procedures improved significantly. In the latter case, the AUC of their model reached a value of 0.90 .
There was a previous study showing a risk-scoring model (Norwich DEFAB) for endometrial cancer including patient characteristics and endometrial thickness . The authors included a very large sample size (3047 postmenopausal women), recruiting all symptomatic postmenopausal women with the assumption that all women with endometrial thickness < 5 mm did not have an intrauterine cancer. Despite the presence of several differences compared to our study, such as studied population, sample size, and disease prevalence, there are many similarities between their results and ours. In this regard, also their best predictor for endometrial cancer was a recurrent vaginal bleeding () to which a score of 4 was assigned. The best cut-off value concerning the women’s age was similar to ours, with a higher risk of cancer for women over 64 years of age (score = 1). For both models, endometrial thickness was a fair predictor of intrauterine malignancy, but at different cut-off values (≥14 mm versus ≥9 mm, resp.). Conversely, in our univariate analysis diabetes and BMI, which were significant predictive factors for intrauterine malignancy in Burbos’s study, there was no statistical difference between women with or without endometrial cancer. Similar results were shown by Opolskiene et al. in their study, where there was no difference in terms of BMI and diabetes in univariate analysis between women with and without cancer . Conversely, as reported by other authors on the same topic , our results showed as a good predictor of intrauterine malignancy the presence of hypertension, to which a score of 2 was assigned.
Based on their model and results, Burbos et al. proposed, as discriminatory cut-off point, a score ≥3 which showed a LR+ of 1.64 and LR− of 0.36. Based on this value, the authors proposed a helpful algorithm with several management options for symptomatic postmenopausal women .
Our risk-scoring model, at the best cut-off score (≥4), showed a fair LR− (0.16) with a posttest probability for endometrial cancer of 1.9%. Given the fact that our first objective should be to decrease the number of missed cancer, this score value had a good diagnostic yield for that purpose. A score ≥ 4 means that, at least, there is a woman with an endometrial thickness ≥ 9 mm and recurrent vaginal bleeding. In that case we recommend performing an outpatient hysteroscopy or sonohysterography, because the posttest probability for cancer was 35.1% (). The latter finding, from a statistical point of view, showed that our model decreased also the false positives, given the pretest probability for endometrial cancer of 11.5%. The issue is much more controversial when the score is less than 4. As mentioned previously, the probability of having cancer was low but present (1.9%) and, according to the algorithm proposed by Burbos et al., we suggested some management options (Figure 2). If endometrial thickness is 5–8 mm, without the presence of recurrent vaginal bleeding (the strongest predictor of endometrial cancer), an outpatient endometrial sampling should be performed and, if negative, no further evaluation should be made. If endometrial thickness is > 8 mm, an outpatient endometrial sampling should be performed and, if negative, a close follow-up with further ultrasound or endometrial sampling could be proposed; an outpatient hysteroscopy or sonohysterography could be also performed. The same management should be adopted for women with endometrial thickness of 5–8 mm and recurrent vaginal bleeding.
Flow-chart showing a decision algorithm for the management of symptomatic postmenopausal women with endometrial thickness > 4 mm.
This clinical approach makes possible a risk assessment focusing on a more comprehensive clinical evaluation, rather than on endometrial evaluation alone. In this regard, for example, a hypertensive woman of 70 years of age with recurrent vaginal bleeding and endometrial thickness of 4 mm should perform a diagnostic hysteroscopy because, despite her endometrial thickness, she would be more at risk for an intrauterine malignancy.
Adding some patient characteristics to endometrial thickness, we built a risk-scoring model (RHEA risk-model) with a moderate diagnostic accuracy in detecting intrauterine malignancies among symptomatic postmenopausal women with endometrial thickness > 4 mm.
However, we want to emphasize that, at present, our results are not generalizable and further studies of external validation are mandatory.
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
The significance of sonographically thickened endometrium in asymptomatic postmenopausal women
The current study aimed to investigate the clinical significance of histological and hysteroscopic evaluations of increased endometrial stripe thickness detected by TVUSG in postmenopausal women without any symptoms of vaginal bleeding.
After menopause, atrophic changes occur in the endometrium due to estrogen deficiency. Therefore, for cases of increased endometrial thickness detected on TVUSG in the postmenopausal period, several endometrial thickness cutoff values have been considered to indicate the need for further evaluation . Although the significance of TVUSG and hysteroscopy in postmenopausal women with bleeding has been evaluated in several studies , data on asymptomatic postmenopausal women with thickened endometrium are limited.
In our study, 168 patients (63.1%) had endometrial polyps. Similarly, Kim et al. reported that endometrial polyps were the most frequently detected focal intrauterine lesions in asymptomatic postmenopausal women. Interestingly, while 57.1% of the patients in our study had an endometrial thickness of 6–10 mm, only 26.3% showed a thickness of 11–15 mm.
While Schmidt et al. and Lev-Sagie et al. reported that no malignancy was detected concomitant to endometrial polyps, an evaluation of 560 asymptomatic women with an endometrial thickness >5 mm on transvaginal ultrasonography showed a cancer development incidence of 0.1% based on the presence of a polyp , and this ratio was as high as 2.7% in symptomatic women. In a recent study, 438 women were evaluated, and the incidence of polypoid carcinoma in symptomatic women was found to be 10%, whereas it was as low as 0.9% in asymptomatic women .
In Antunes et al.’s study , 78.5% of the 475 cases presented with endometrial polyps, while 2.7% were found to be carcinomatous polyps. In our study, we reported only 1 case (1.1% of adenocarcinoma that was identified based on an endometrial polyp within an endometrial stripe thickness of 11–15 mm.
Although Schmidt et al. reported that the prevalence of premalignant and malignant lesions was low in endometrial polyps, a 5.3 times increased prevalence was demonstrated in women aged over 60 years in their study. As a result, we recommend that polyps detected in postmenopausal women should be removed and examined histologically. However, it may be technically difficult to completely remove polypoid structures using D&C, in which case hysteroscopy should be performed.
In our study, intrauterine pathologies were detected in 132 (49.6%) patients with an endometrial thickness of 6–10 mm. Osmers et al. proposed an endometrial thickness of 8 mm as a cutoff value for the diagnosis of pathological endometrial changes, with a sensitivity of 81% and specificity of 89% for the diagnosis of endometrial pathologies. Endometrial cancer was found in 3.5% of patients in the symptom-free group.
In contrast, a large cohort of postmenopausal women was evaluated in the United Kingdom Collaborative Trial of Ovarian Cancer Screening study (UKCTOCS). When an endometrial thickness cutoff value of 5 mm was used for the diagnosis of endometrial cancer or atypical hyperplasia, it showed a sensitivity of 80.5% and a specificity of 85.7% . Meanwhile, no endometrial cancer was reported in postmenopausal women without vaginal bleeding with an endometrial thickness of 5 mm or above in a study by Worley et al. . In our study, the optimal cutoff value for premalignant and malignant lesions was found to be 10.5 mm, with 77% sensitivity and 62% specificity (Fig.1).
In a study by Schmidt et al. , intrauterine pathologies were detected in 127 (41.7%) patients who had endometrial thickness measurements of 6–10 mm, while endometrial cancer was identified in 4% of the patients and 4 of the 5 carcinoma cases occurred in patients with an endometrial thickness less than 10 mm.
In the study by Saatli et al. of 530 asymptomatic cases, 5 adenocarcinoma cases (0.9%) were detected using a cutoff value of 5 mm, and it was concluded that 1 case of cancer was identified for every 106 cases examined.
In Breijer et al.’s meta-analysis , the prevalence of endometrial carcinoma and atypical endometrial hyperplasia in asymptomatic women who were not receiving hormone replacement therapy was found to be 0.62% and 0.59%, respectively, and they proposed the use of endometrial thickness as a predictor for premalignant and malignant lesions in women with vaginal bleeding. In a study by Gambacciani et al. , only 1 case (0.7%) of endometrial cancer was detected among 148 asymptomatic postmenopausal women who underwent hysteroscopy based on the endometrial thickness measured via ultrasonography.
In our study, a total of 266 cases were examined, and endometrial adenocarcinoma was detected in 8 (3%) cases; in 4 of the 8 detected endometrial cancer cases (1.5%), the endometrial thickness was between 6 and 10 mm. There were 4 cases of endometrial hyperplasia with atypia that showed endometrial thickness greater than 16 mm, and endometrial hyperplasia was found in 1.5% (n=4) of cases.
Conversely, Yasa et al. found that 11 patients with an endometrial thickness of 11 mm or less had endometrial hyperplasia and endometrial cancer on histological examination.
In a study by Smith-Bindman et al. , a theoretical model was developed, according to which the risk of endometrial cancer was found to be 6.7% in asymptomatic postmenopausal women when the endometrial thickness was greater than 11 mm, with a very low incidence of 0.002% among women with endometrial thickness less than 11 mm. Therefore, using a cutoff below 10 or 11 mm would lead to biopsies in approximately 1% of normal postmenopausal women.
In the present study, our results are in agreement with those discussed above in that the endometrial thicknesses observed in the endometrial cancer cases were reported to be 6–10 mm (n=4), 11–15 mm (n=2), and >20 mm (n=2). These data highlight the clinical importance of the possibility of detecting a cancer case originating from a thinner endometrium.
In the present study, hysteroscopy was limited in the diagnosis of endometrial polyps in some cases. In 40 patients, areas identified as endometrial hyperplasia during hysteroscopy were histopathologically diagnosed as endometrial polyps. In addition, although their hysteroscopic appearance was suggestive of endometrial polyps in 4 cases, histologic examination diagnosed them as submucous myomas, which did not match the final pathology (Table 2). In line with our study, Widrich et al. also showed that not every structure with a polypoid appearance met the pathologic criteria for a polyp on hysteroscopy. In our study, while 6 cases were hysteroscopically described as endometrial hyperplasia, histological examination diagnosed them as endometrial carcinoma. Therefore, it is very important to combine hysteroscopy and histopathology to obtain a final diagnosis .
One of the limitations of our study is its retrospective design, which limited our ability to analyze the causes and effects of endometrial hyperplasia. Therefore, larger prospective trials are warranted, and an individualized assessment based on patient characteristics and risk factors for endometrial pathologies should be conducted in asymptomatic postmenopausal women.
In conclusion, although we determined an optimal cutoff value for premalignant and malignant lesions of 10.5 mm, our current study also revealed that 8 endometrial cancer cases were detected, 2 on a polypoid basis, in patients with an endometrial thickness of 10 mm or less. Therefore, every patient whose endometrial thickness falls below this cutoff value should still be thoroughly evaluated according to their particular conditions. D&C and hysteroscopy are endometrial sampling techniques used to diagnose endometrial abnormalities. Although D&C is considered the gold standard for endometrial sampling, hysteroscopy has the advantages of achieving a panoramic view of the uterine cavity and hysteroscopy-guided biopsy.
Endometrial thickness is a commonly measured parameter on routine gynecological ultrasound and MRI. The appearance, as well as the thickness of the endometrium, will depend on whether the patient is of reproductive age or postmenopausal and, if of reproductive age, at what point in the menstrual cycle they are examined.
The endometrium should be measured in the long axis or sagittal plane, ideally on transvaginal scanning, with the entirety of the endometrial lining through to the endocervical canal in view. 10 The measurement is of the thickest echogenic area from one basal endometrial interface across the endometrial canal to the other basal surface. Care should be taken not to include hypoechoic myometrium or intrauterine fluid in this measurement.
The normal endometrium changes in appearance as well as in thickness throughout the menstrual cycle:
- in the menstrual and early proliferative phase it is a thin, brightly echogenic stripe comprising of the basal layer (figure 1); minimal fluid can be appreciated endovaginally within the endometrium in the menstrual phase
- in the late proliferative phase it develops a trilaminar appearance: outer echogenic basal layer, middle hypoechoic functional layer, and an inner echogenic stripe at the central interface
- in the secretory phase it is at its thickest, up to 16 mm 10, and becomes uniformly echogenic, as the functional layer becomes edematous and isoechoic to the basal layer (figure 2); there is through transmission and posterior acoustic enhancement noted
The postmenopausal endometrium should be smooth and homogeneous.
Normal range of endometrial thickness
The designation of normal limits of endometrial thickness rests on determining at which thickness the risk of endometrial carcinoma is significantly increased.
Whilst quantitative assessment is important, endometrial morphology and the presence of risk factors for endometrial malignancy should also be taken into account when deciding whether or not endometrial sampling is indicated.
Commonly accepted endovaginal ultrasound values are as follows:
In premenopausal patients, there is significant variation at different stages of the menstrual cycle.
- during menstruation: 2-4 mm 1,4
- early proliferative phase (day 6-14): 5-7 mm
- late proliferative / preovulatory phase: up to 11 mm
- secretory phase: 7-16 mm
- following dilatation and curettage or spontaneous abortion: <5 mm, if it is thicker consider retained products of conception
Please note that these measurements are a guide only, as endometrial thickness may be variable from individual to individual.
The postmenopausal endometrial thickness is typically less than 5 mm in a postmenopausal woman, but different thickness cut-offs for further evaluation have been suggested.
- vaginal bleeding (and not on tamoxifen):
- suggested upper limit of normal is <5 mm 5
- the risk of carcinoma is ~7% if the endometrium is >5 mm and 0.07% if the endometrium is <5 mm 8
- on hormonal replacement therapy: upper limit is 5 mm
- no history of vaginal bleeding:
- the acceptable range of endometrial thickness is less well established in this group, cut-off values of 8-11 mm have been suggested
- the risk of carcinoma is ~7% if the endometrium is >11 mm, and 0.002% if the endometrium is <11 mm 8
- if on tamoxifen 3: <5 mm (although ~50% of those receiving tamoxifen have been reported to have a thickness of >8 mm 7)
If a woman is not experiencing bleeding, and the endometrium is thickened, the guidelines are less clear. Either a repeat transvaginal ultrasound or a referral to a gynecologist is reasonable.
Endometrial thickness is well assessed on MRI. Measurement should be taken at a mid-sagittal slice, similar to the ultrasound assessment plane.
- T2: normal endometrium is homogeneously hyperintense regardless of the phase of the menstrual cycle or menopausal status and well outlined by the low signal myometrial junctional zone
- tamoxifen-associated endometrial changes
- endometrial reflectivity grading
- endometrial hyperplasia
- endometrial atrophy
- abnormally thickened endometrium
Abnormal uterine bleeding in peri- and postmenopausal women
Updated: January 6, 2020Published: January, 2011
When should you see a clinician about excessive or unexpected bleeding?
Abnormal uterine bleeding (AUB) is a common problem for women of all ages, accounting for up to one-third of gynecologic office visits. The two main types are heavy bleeding that occurs at an appropriate or expected time, such as a heavy menstrual period (menorrhagia), and any type of bleeding that occurs unexpectedly (metrorrhagia). The absence of regular menstrual periods for several months (amenorrhea) is also considered an abnormal bleeding pattern. AUB can be tricky to identify, because what’s normal depends on a woman’s reproductive age.
To continue reading this article, you must login.
Subscribe to Harvard Health Online for immediate access to health news and information from Harvard Medical School.
- Research health conditions
- Check your symptoms
- Prepare for a doctor’s visit or test
- Find the best treatments and procedures for you
- Explore options for better nutrition and exercise
Learn more about the many benefits and features of joining Harvard Health Online “