Patients with postmenopausal bleeding (PMB) have a 10–15% chance of having endometrial carcinoma; they should therefore be seen within 2 weeks of referral. Cervical and vulval cancers remain important components of the differential diagnosis and can only be assessed by clinical examination.
There are well-developed and evidence-based strategies on how best to investigate women with PMB. These strategies are formulated in several guidelines.
To understand that the completion of investigations of patients with PMB should include test performance and patient characteristics.
To be able to evaluate the various investigations for women with PMB.
To be able to describe the appropriate management of women with PMB on tamoxifen, and unscheduled PMB in women undergoing hormone replacement therapy.
Incidental finding of thickened endometrium on transvaginal sonography (TVS) in asymptomatic individuals carries a management dilemma.
Approach to benign pathology: should subsequent evaluation for benign abnormalities be performed after malignancy has been ruled out?
Ongoing work and technology development will evolve our strategy in managing PMB.
Postmenopausal bleeding (PMB), defined as uterine bleeding occurring after at least 1 year of amenorrhoea, is a common clinical condition with an incidence of 10% immediately after menopause. Patients with PMB have a 10–15% chance of having endometrial carcinoma.[2-6] Therefore, the clinical approach to PMB requires prompt and effective evaluation to exclude cancer in the genital tract or precancerous lesions of the endometrium. However, vaginal atrophy, and benign focal lesions such as endometrial polyps and fibroids, are common. The prevalence of endometrial polyps and hyperplasia is estimated to be 40%.[2, 3]
Endometrial cancer is the most common gynaecological malignancy. Risk factors include obesity, unopposed estrogens, polycystic ovary syndrome, and nulliparity. Ninety percent of women with endometrial carcinoma present with vaginal bleeding. Unlike ovarian cancer, endometrial cancer often presents at an early stage when there is a possibility of curative treatment by hysterectomy; early, accurate and timely diagnosis is therefore important.
PMB is usually attributed to an intrauterine source, but may arise from the vulva, vagina, cervix, fallopian tubes, or it may be related to ovarian pathology. Haematometra may also result from cervical stenosis. The bleeding may originate from extragenital sites such as the urethra or bladder, and the rectum or bowel.
Patient specifications and the risk of endometrial cancer
The pre-test probability (before any test is done) of endometrial cancer in women with PMB is ~10%, but various clinical specifications may alter this proportion, which rises from 1% in women aged <50 years to almost 25% in women aged >80 years. The incidence of cancer is higher in women with PMB and obesity (18%) or with PMB and diabetes (21%). In obese women with diabetes the incidence may be as high as 29%.
Accuracy of transvaginal sonography for diagnosing endometrial cancer
Since it was introduced in the 1980s, transvaginal sonography (TVS) has become widely used in the evaluation of women with PMB. Before TVS, women with PMB underwent dilatation and curettage. The relatively non-invasive nature of TVS makes it more acceptable, especially to older women. The mean endometrial thickness in postmenopausal women is much thinner than in premenopausal women. The likelihood of important pathology (cancer) being present increases with increasing thickness of the endometrium (Figure 1).
The endometrial thickness cut-off
TVS can reliably assess thickness and morphology of the endometrium and can thus identify a group of women with PMB who have a thin endometrium (≤4 mm) and are therefore unlikely to have endometrial cancer (Figure 2). Endometrial sampling is therefore not recommended below this cut-off value.[6-9]
To date, four meta-analyses have been published; each has used different methods to determine the accuracy of TVS in diagnosing endometrial abnormalities in women with PMB.
The most cited meta-analysis by Smith-Bindman et al. included 5892 women from 35 prospective studies that compared endometrial thickness measured at TVS to presence or absence of endometrial carcinoma on histology. At ≤5 mm cut-off, the overall summary mean weighted estimates of the sensitivity for detecting endometrial cancer was 96% for a 39% false-positive rate. This would reduce a pre-test probability of 10% for endometrial cancer to a post-test probability of 1%. Therefore, expectant management (without the need for tissue sampling) is recommended for these women. There is only one study that looked at follow-up of women with PMB and an endometrial thickness of <4 mm. It showed that none of the women with the expectant management developed cancer over 1 year of follow-up.
The comprehensive systematic review of Gupta et al. included only best-quality studies (only four studies were included). For an endometrial thickness of ≤5 mm, pooling likelihood ratio (LR) was 0.16 for a negative test. This meant that in an individual with a negative test result, the post-test probability was 2.5% (compared with the previous meta-analysis finding of 1% post-test probability). The review concluded that, using the evidence from the best-quality studies, a negative result at ≤5 mm cut-off rules out endometrial pathology with a high degree of certainty.
In their meta-analysis, Tabor et al. included only studies from which they were able to extract the original data from the authors. The median endometrial thickness was calculated per study/centre, then pooled data for endometrial thickness were used with a sensitivity of 96% for a specificity of 50% and 4% false-negative rate. In their opinion, endometrial thickness measurement does not reduce the need for invasive diagnostic testing.
Limitations of these meta-analyses might be overcome by meta-analysis of individual patient data. Timmermans et al. conducted meta-analytic strategies whereby 79 primary investigators were contacted to obtain the individual patient data of their reported studies, of which 13 could provide data. Data on 2896 individuals, of whom 259 had cancer, were analysed. It was concluded that previous meta-analyses on endometrial thickness measurement have probably overestimated its diagnostic accuracy in the detection of carcinoma.
Meaningful assessment of the endometrium (thickness and morphology) by ultrasonography is not possible in all patients. In such cases or if bleeding persists despite negative initial evaluation, alternative methods are indicated.
Saline infusion sonography
Saline infusion sonography (SIS) involves the infusion of saline into the uterine cavity during ultrasound to separate the two walls of the endometrium, which allows their thicknesses to be measured. It also allows the evaluation of intracavitary lesions such as fibroids or polyps. There are no large studies evaluating the accuracy of SIS in the assessment of women with PMB. However, in their meta-analysis, de Kroon et al. conclude that SIS is accurate in the evaluation of the uterine cavity in pre- and postmenopausal women with abnormal uterine bleeding. The authors also conclude that the feasibility of SIS is high, although it is significantly better in premenopausal women compared with postmenopausal women. Therefore outpatient biopsy and hysteroscopy are still the methods of choice.
Accuracy of endometrial sampling methods
Patients with an increased endometrial thickness should undergo further invasive testing. Dilatation and curettage is now considered to be an outdated practice and is replaced by less invasive outpatient evaluation using endometrial biopsy devices and outpatient hysteroscopy-guided biopsies.
In the meta-analysis of Dijkhuizen et al. different endometrial biopsy devices were compared. In postmenopausal women endometrial sampling with both the Pipelle device (Pipelle de Cornier, Paris, France) and the Vabra device (Berkeley Medevices, Inc., Richmond, CA, USA) were very sensitive techniques for the detection of endometrial carcinoma, with detection rates of 99.6% and 97.1%, respectively.
The amount of tissue obtained by office sampling is sometimes insufficient for a histological diagnosis. In those cases, the clinician is in doubt whether or not to proceed with more invasive testing or to rely on the negative biopsy. In a prospective study performed by Van Doorn et al., four (6%) out of 66 patients with insufficient tissue at office endometrial sample were subsequently diagnosed with endometrial cancer (n =3) or atypical hyperplasia (n =1). This finding implies that women with an insufficient sample and an endometrial thickness of ≥5 mm should not be reassured.
It would appear from the controlled regression analysis by Bakour et al. that clinicians can be confident in reassuring women with an insufficient sample on outpatient endometrial biopsy, provided that the hysteroscopic and sonographic endometrial assessment is consistent with endometrial atrophy. This means that it is reasonable to reassure and discharge women with an insufficient endometrial sample with negative scan (≤4 mm) without the need to expose them to hysteroscopy and curettage. The need for reinvestigation on recurrence of symptoms should be borne in mind; for example, a small polyp in an atrophic endometrium may not be detected because the endometrial sampling does not yield enough cells.
Accuracy of hysteroscopy in detecting endometrial carcinoma
Compared with traditional methods such as curettage, hysteroscopy offers the possibility of visualising macroscopically focal abnormalities and taking directed biopsies. Outpatient hysteroscopy allows direct visualisation of the uterine cavity, which is particularly useful for excluding endometrial polyps or fibroids. With the development of smaller diameter hysteroscopic systems and the introduction of a ‘vaginoscopic’ approach, patient acceptance has improved and hysteroscopy nowadays can be performed in an outpatient setting without anaesthesia. Inpatient hysteroscopy is required only if the outpatient assessment is either inadequate or impossible to perform.
Clark et al. conducted a systematic quantitative review looking at the accuracy of hysteroscopy in the diagnosis of endometrial cancer and hyperplasia in women with abnormal uterine bleeding. The result of this systematic review revealed that for endometrial cancer a positive hysteroscopy had a pooled LR of 62.8 (95% confidence interval [CI] 52.8–74.6), whereas a negative hysteroscopy had a pooled LR of 0.15 (95% CI 0.13–0.18). The review concluded that the diagnostic accuracy of hysteroscopy is high for endometrial cancer but only moderate for endometrial disease defined as cancer and/or hyperplasia.
Diagnostic strategies for postmenopausal bleeding
The true clinical value of a test lies in the information obtained beyond what was already known from the history and examination. In the evaluation of diagnostic tests, Khan et al.[17, 18] built a stepwise four-multivariable approach to take into account the clinical context. The first model provides a valid estimate of the combined predictive value of the clinical history variables. The other three models determine the combined predictive added value of history and tests (ultrasonography or hysteroscopy, or ultrasonography and hysteroscopy combined). The predictive ability of the addition of both ultrasonography and hysteroscopy increased the receiver operating characteristic (ROC) area to 0.84 (ROC is a graphical plot of the true-positive rate versus false-positive rate). This represents significantly improved predictive ability from the clinical history plus ultrasonography model (0.82 versus 0.84, χ2 for improvement = 6.9, P =0.009) and from the clinical history plus hysteroscopy model (0.81 versus 0.84, χ2 for improvement = 5.4, P =0.02).[19, 20]
Similarly, to determine the most cost-effective testing strategy for diagnosing endometrial carcinoma in women with PMB, Clark et al. constructed a decision model and evaluated 12 different strategies for combination of tests. The strategy of TVS as initial investigation with a cut-off of 5 mm and endometrial biopsy were most cost-effective. Khan et al. proposed to evaluate tests using a multivariable approach and proposed the use of individual patient data meta-analysis in calculating the post-test probability of a disease for a different combination of test results (including patient characteristics and information from clinical history).
Individual patient characteristics including age, time since menopause, obesity, hypertension, and diabetes mellitus are known risk factors of endometrial carcinoma. However, current policy is based not on these risk factors but on endometrial thickness. The flow chart (Figure 3) is proposed as a possible algorithm for diagnostic pathways in women with PMB. This algorithm refers to the need for future research to develop a multivariable decision model in which calculation of the pretest probability of endometrial cancer is performed, based on the individual's characteristics.
Thickened endometrium in asymptomatic postmenopausal women: should it be investigated?
To date, screening tools for endometrial cancer have not been identified. Studies demonstrating a positive effect of performing routine endometrial cancer screening on overall mortality would possibly have to include hundreds of thousands of women based on the relatively low incidence of the disease, the high rate of symptomatic women in early stage disease, and the good prognosis of affected women. Such a study would be hard to justify.
TVS has been suggested as a screening test. The UK Collaborative Trial for Ovarian Cancer Screening (UKCTOCS), which aims to establish the impact of ovarian cancer screening on ovarian cancer mortality, is the world's largest collaborative screening trial, involves >200 000 UK women, and reports in 2015. Data from the assessment of endometrial thickness and morphology in the course of this trial have provided invaluable information on endometrial thickness in asymptomatic women and the risk of endometrial cancer in this population. The optimum endometrial thickness cut-off for endometrial cancer or atypical endometrial hyperplasia was 5.15 mm, with sensitivity of 80.5% (95% CI 72.7–86.8) and specificity of 86.2% (95% CI 85–86.6). When the analysis was restricted to the women with endometrial cancer or atypical endometrial hyperplasia who reported no symptoms of postmenopausal bleeding at the UKCTOCS scan before diagnosis, a cut-off of 5 mm achieved a sensitivity of 77.1% (95% CI 67.8–84.3) and specificity of 85.8% (95% CI 85.7–85.9). The logistic regression model identified 25% of the population as at high risk, and 39.5% of endometrial cancer or atypical endometrial hyperplasia cases were identified within this high-risk group; a cut-off at 6.75 mm achieved sensitivity of 84.3% (95% CI 71.4–93.0) and specificity of 89.9% (95% CI 89.3–90.5). These findings concluded that TVS screening for endometrial cancer has high sensitivity in postmenopausal women. There are complex and challenging issues around balancing the cost, positive health impact and unwanted consequences of screening that require careful study and analysis; however, these findings are of immediate value in the management of increased endometrial thickness in postmenopausal women undergoing pelvic scans for reasons other than vaginal bleeding.
Schmidt et al. proposed that hysteroscopy represents an easy, safe and effective method for the investigation of asymptomatic women with a thickened endometrium (>6 mm). The commonest pathology was endometrial polyps (74.3%).
Endometrial fluid detected by transvaginal sonography
Curcić et al. concluded that the presence of endometrial fluid detected by TVS is a good marker for pathological changes of the endometrium in postmenopausal women if the endometrial thickness is >4 mm. If the endometrial thickness is <4 mm, the presence of endometrial fluid is not an indication for further invasive investigation.
Risk of malignancy in endometrial polyps in asymptomatic postmenopausal women
In a multicentre study by Ferrazzi et al. on 1152 asymptomatic women and 770 women with PMB, only one case (0.1%) of stage 1, grade 1 endometrial carcinoma on a polyp with a mean diameter of 40 mm was observed in asymptomatic women. This prevalence was 10 times lower than in symptomatic patients (P <0.0001). The prevalence of atypical hyperplastic polyps was 1.2% in asymptomatic women (2.2% in symptomatic patients; P <0.005). At multivariate analysis, polyp diameter was the only variable significantly associated with an abnormal histology (cancer/atypical hyperplasia) in asymptomatic women (odds ratio for polyps with mean diameter >18 mm: 6.9; 95% CI 2.2–21.4). The authors concluded that follow-up and/or treatment of endometrial polyps incidentally diagnosed in asymptomatic postmenopausal patients could be safely restricted to a few select cases based on polyp diameter.
Risk of malignancy in endometrial polyps in women with PMB
Although endometrial polyps are a frequent finding in postmenopausal bleeding, high-quality evidence regarding the efficacy of hysteroscopic polypectomy in symptomatic women is lacking. One cohort study investigated the efficacy of treatment regarding recurrent bleeding, and found that the recurrence rate of postmenopausal bleeding in women with endometrial thickness >4 mm was 20%. There was no difference with respect to recurrence rate between patients with polyp removal, patients with a normal hysteroscopy, and patients with office endometrial sampling alone at the initial work-up. The question is therefore whether these should be diagnosed and removed at all.
Tamoxifen and PMB
Tamoxifen is a cornerstone in the treatment of breast cancer. Women receiving tamoxifen experience a three- to six-fold greater incidence of endometrial cancer due to its weak estrogenic effect on the endometrium. The presentation of PMB therefore requires urgent investigation in this group of women. Treatment beyond 5 years increases risk by at least four-fold.
There has been some debate about how closely women on tamoxifen should be monitored for the development of endometrial cancer. However, evidence mainly from observational studies indicates that periodic investigations are unlikely to be cost-effective. Therefore PMB should remain the primary trigger for investigation of women on tamoxifen.
Unscheduled bleeding on HRT
Unscheduled bleeding is the term used for breakthrough bleeding occurring in women on cyclical HRT or any bleeding in women on continuous combined HRT, although it can take up to 6 months for amenorrhoea to develop in the latter treatments. For sequential regimens, abnormal bleeding may:
be heavy or prolonged at the end of, or after, the progestogen phase
or occur at any time (breakthrough bleeding).
Bleeding on continuous combined regimens should be considered abnormal (requiring endometrial assessment) if it occurs:
after the first 6 months of treatment, or
after amenorrhoea has been established.
Endometrial carcinoma risk among women with endometrial hyperplasia
In their nested case-control study of endometrial hyperplasia (EH) progression, Lacey et al. summarised the 34-year experience of the included 138 cases, which were diagnosed with EH and then with carcinoma (between 1970 and 2003) at least 1 year (median 6.5 years) later, and 241 matching controls. Atypical hyperplasia (AH) significantly increased the relative risk of carcinoma (14; 95% CI 5–38). This risk justifies discussing the management of these cases case by case in the multidisciplinary gynaecology oncology meeting whereby hysterectomy is generally recommended for women with AH because of a high probability of underlying carcinoma. Treatment of endometrial hyperplasia without atypia in postmenopausal women with a levonorgestrel intrauterine device has been suggested to be an effective and safe alternative.
Role of three-dimensional ultrasound
It does not appear that the diagnostic performance of three-dimensional (3D) ultrasound imaging for discrimination between benign and malignant endometrium is superior to 2D ultrasound examination; and 3D power Doppler imaging added little to endometrial thickness or volume. 3D ultrasound imaging added little benefit for discrimination between benign and malignant endometrium in women with PMB and sonographic endometrial thickness of ≥4.5 mm.
This article has reviewed the modern management of women with postmenopausal bleeding. There are well-developed and evidence-based strategies on how best to investigate women with PMB. These strategies, formulated in several guidelines, are discussed and critically appraised.