Postmenopausal bleeding

Introduction

Endometrial cancer is the most common gynaecological cancer and causes one in 10 of all cancer cases and one in 20 of all deaths from cancer in European women. ¹ Over the last decade, most countries have observed stable incidence, declines in mortality and small improvements in five-year survival. ² However, in the UK both incidence and mortality rates have increased over the last decade with the greatest increases in incidence observed in postmenopausal women. ^2–4 It has been postulated that this may be due to temporal changes in reproductive behaviour, prior use of unopposed estrogen hormone replacement therapy and/or increases in the prevalence of obesity. ⁴ It has been estimated that, in the UK, mortality rates from endometrial cancer are projected to increase over the next couple of decades. ⁵

A temporal relationship between obesity and increased risk of endometrial cancer has been consistently observed in a number of cohort studies, ⁶ a recent review of which suggests a 1.5 times increase in risk of endometrial cancer with a 5 kg/m² increase in body mass index (BMI). ⁷ The association is observed with higher BMI in middle age and in early adulthood and also gain in BMI: the latter appears to be the strongest BMI-related predictor. ^8,9 The increase in risk associated with BMI has been observed for both premenopausal and postmenopausal cancers, but appears to be greater for those occurring postmenopause. ¹⁰

Postmenopausal bleeding (PMB) is the most common presenting symptom of endometrial cancer. ¹¹ PMB is defined as any vaginal bleeding occurring at least 12 months after the last menstrual period. ¹² Cancer is the underlying aetiology in up to 10% of women referred with this symptom. ^13,14 Most cases of endometrial cancer presenting with PMB will be early stage disease, which is amenable to treatment. PMB may also be the presenting symptom for other gynaecological cancers. Other conditions commonly found on assessment of PMB include atrophy of the female genital tract as a result of low serum estrogen levels, and endometrial or cervical polyps.

In this article, we discuss the organization of clinical services for the assessment of women with PMB, provide an overview of current diagnostic modalities, review the literature on women's perspectives and highlight the special groups who are at increased risk of endometrial cancer. This review will not cover the management or sequelae of endometrial cancer.

Impact on clinical services

In the UK, PMB is a common reason for consulting primary care doctors. ^15,16 Although the risk of endometrial cancer increases with age, ¹⁷ primary care consultation rates for this symptom are highest in younger women: incident consultation rates of 8.3/1000 for women aged 50–59 years, 4.7/1000 for women aged 60–69 years and 2.2/1000 for women aged 70 years and over have been reported. ¹⁵

While it was traditionally taught that all cases of PMB should be referred for specialist investigation urgently to exclude cancer, the increased use of hormone replacement therapy in this age group, with the attendant problem of unscheduled bleeding, has led some bodies to produce specific guidance about urgent referral for further assessment. ^12,18 National Institute for Clinical Excellence guidelines advise that women with PMB who are not on hormone replacement therapy or who have persistent or unexplained bleeding after six weeks cessation of hormone therapy should be urgently referred for specialist assessment. They should be investigated within two weeks of referral to exclude the presence of gynaecological malignancy. ¹⁸ An analysis of primary care data found that only 40% of women presenting to primary care with PMB were referred for specialist opinion and that the proportion of women referred decreased in those aged over 75 years, although the risk of cancer was greater in these age groups. ¹⁶ A study of referrals to one gynaecologist observed that many women who were subsequently diagnosed with endometrial cancer had not been referred urgently. ¹⁹

PMB accounts for about 5% of referrals to gynaecology outpatient services. In a unit of our size, serving a local population of 390,000, we expect to see about 40 patients presenting with this condition each month. There, therefore, needs to be a system in place to ensure these women can be seen within two weeks of referral. There is no consistent method of achieving this target throughout the UK; some units reserve a number of clinic appointments for these urgent referrals, other units run dedicated ‘one-stop’ clinics for the investigation of PMB. Studies have demonstrated that one-stop investigation of PMB reduces waiting times and theatre costs, and provides quick reassurance, diagnosis and initiation of management, compared with traditional outpatient clinics. ^20,21 The most cost-effective method of achieving the two-week target within any specialist service will be dependent on the number of women referred with this symptom.

Investigation of PMB also impacts on scan, outpatient hysteroscopy and day case services. It may be more cost-effective for larger units to consider a ‘one-stop’ service, which would have less resource implications both for the service provider and the patient. To provide a ‘one-stop’ service, facilities for all outpatient investigation modalities need to be available at the initial visit. Unit protocols and care pathways need to reflect the selected first-line investigations and triage pathways for women needing further investigation. Information needs to be provided to patients prior to first attendance, so that they are aware of the investigations likely to be performed.

Review of methods of diagnostic evaluation

The aim of examination and investigation of women presenting with PMB is first and foremost to correctly identify those women with malignancy, and also to identify benign conditions that require treatment. Clinical examination should include abdominal palpation, speculum examination and bimanual examination. It is reasonable to expect cervical causes (for example cervical cancer) to be identified by this examination. Indeed, the risk of cervical cancer has substantially declined with the introduction of the population-based cervical screening programme.

Modalities for the investigation of PMB include transvaginal scanning (TVS), endometrial biopsy (EB), hysteroscopy and more recently saline infusion hysterosonography (SIS). The sensitivity, patient acceptability and cost-effectiveness of each method should be considered when investigation protocols are devised.

Hysteroscopy is considered to be the gold standard for investigating lesions within the uterine cavity and visually assessing the endometrium. With the introduction of small diameter hysteroscopes, this investigation can now be performed in an outpatient setting for the vast majority of women; therefore, avoiding the need for a general anaesthetic, time off work and the resulting domestic inconvenience. A retrospective observational study reported 98.2% of endometrial cancers to be correctly identified by outpatient hysteroscopy. ²² Hysteroscopy has also been reported to be superior to scanning alone in the detection of endometrial polyps. ²³ A further advantage of hysteroscopy is that endometrial lesions can be directly biopsied and some endometrial polyps removed during the procedure, avoiding the need for a second visit. Disadvantages of outpatient hysteroscopy include patient morbidity (Critchley et al. ²³ reported minor adverse events, for example shock) and possible failure secondary to pain or cervical stenosis (Alexopoulos et al. ²⁴ reported a 2.8% failure rate mainly due to these conditions, Sahla et al. ²⁵ reported slightly higher figures, 3–8%). Hysteroscopy is also a more expensive investigation tool when compared with TVS and EB alone. Clark et al. ²⁶ constructed a decision analytic model to determine the cost-effectiveness of initial diagnostic strategies in the investigation of PMB for endometrial cancer. They concluded that hysteroscopy alone was not cost-effective.

TVS can be used as a modality in the investigation of PMB. The double layer of endometrium is measured and reported as an endometrial thickness. This method has the added benefit of being able to visualize the ovaries and therefore exclude ovarian pathology as a cause of PMB. It is unlikely that endometrial pathology exists if the endometrium is thin; recent guidelines published by the American College of Obstetricians and Gynecologists recommend that EB is not required when the endometrial thickness is less than or equal to 4 mm. ²⁷ A meta-analysis of 35 studies, involving 5892 women, reported a 96% (95% CI: 93–98%) sensitivity to diagnose endometrial cancer, using a 4 mm cut-off; the specificity was 53% (95% CI: 51–55%). ²⁸ Benign endometrial pathology, for example endometrial polyps, can also give the appearance of a thickened endometrium on scan, hence the relatively low specificity of this investigation. More recently, Gull et al. ²⁹ reported a 100% negative predictive value based on zero cases of endometrial cancer in 394 women investigated for PMB with an endometrial thickness of less than or equal to 4 mm.

EB is a method of obtaining a sample of endometrium to obtain a histological diagnosis. It can be performed blind with endometrial samplers, for example Pipelle sampler, Tao brush, or as a directed biopsy at the time of hysteroscopy. As a single investigation, the failure rate is high in the postmenopausal age group, presumably as a result of both cervical stenosis and a high prevalence of endometrial atrophy. Critchley et al. ²³ reported an adequate EB obtained by the Pipelle sampler in 43% of these high-risk women and by the Tao brush in 72%. They also reported a 10% incidence of minor adverse events (shock or patient distress).

SIS is less widely used than the other investigation modalities discussed. Rogerson et al. ³⁰ conducted a prospective parallel blinded comparison of this procedure with outpatient hysteroscopy (OPH). They reported a much higher failure rate of SIS (34%) compared with OPH (11%) in the postmenopausal age group. The proportion of cases in which there was agreement between SIS and OPH findings in this study was 0.66 (95% CI: 0.50–0.81), indicating poor validity of SIS accepting OPH as the gold standard, although the two cases of endometrial malignancy were correctly identified by both investigation modalities. SIS was associated with lower pain scores than OPH. The authors concluded that further research should focus on refining SIS in order to reduce failure rates and increase accuracy.

Most authors agree that the first-line investigation of PMB should be TVS with an endometrial thickness cut-off of 4 mm. ^27,31 It would seem reasonable to coordinate the TVS appointment with the initial consultation. This would enable women to be appropriately triaged at the first visit. Ultrasonographic assessment of endometrial thickness and the ovaries, combined with clinical examination of the lower genital tract will enable those women who are unlikely to have gynaecological pathology to be reassured at a single visit. Present evidence suggests that women with an endometrial thickness greater than 4 mm, and women in whom meaningful assessment of the endometrium by ultrasonography is not possible, should have further assessment in the form of EB with or without hysteroscopy. ²⁷

Recurrence of symptoms

Critchley et al. ²³ have reported that 16% of women still have symptoms 10 months after investigation of PMB. There is no evidence to direct recommendations on when women should be re-investigated for recurrent or persistent PMB. However, re-investigation should be considered in view of the small false-negative rate associated with all methods of diagnosis. ¹² Committee opinion of the American College of Obstetricians and Gynecologists recommend additional assessment when bleeding persists despite negative initial evaluation. ²⁷ It would therefore be reasonable to perform hysteroscopy and EB on women who have only been investigated by TVS when they re-present with PMB. It is not as clear-cut when to re-investigate women who have already undergone all modalities of investigation with negative results, although some authorities recommend re-investigation after a period of six months. ¹² The possible causes of persistent or recurrent PMB include missed cancers, ongoing estrogen production and genital tract atrophy. Occasionally haematuria and rectal bleeding may present as suspected PMB. Cystoscopy to exclude bladder tumours and sigmoidoscopy to exclude large bowel tumours may be indicated if the site of the bleeding is unclear. ³²

Women's perspectives

There have been very few studies that have examined women's knowledge, attitudes or concerns about PMB or its assessment. This is in contrast to the wealth of research on the psychological impact of the assessment of a cervical smear abnormality, which has informed national guidelines on information and support within the cervical screening programme.

There is some evidence that postmenopausal women have poor knowledge about endometrial cancer. A German study of women attending gynaecological clinics observed that only 10% of those aged 50 and over thought endometrial cancer was the most common gynaecological cancer and the vast majority thought that genetic factors and previous history of breast and gynaecological cancer were the most important risk factors. ³³ Nearly two-thirds felt they should be better informed. Studies have demonstrated limited knowledge about the increased risk of endometrial cancer with obesity. ^34,35

In a study of women referred for assessment of PMB, we identified that 12/55 (22%) women waited more that three months before consulting their general practitioner about their symptoms and those who delayed presentation were less anxious and stressed than those who had not. ³⁶ We also confirmed high levels of anxiety in postmenopausal women attending for assessment, ³⁶ initially repeated by Gupta et al. ³⁷ Studies suggest that postmenopausal women find transvaginal ultrasound an acceptable procedure, but about one in six will find hysteroscopy markedly unpleasant. ²³ However, in a small study of women with postmenopausal women who had undergone hysteroscopy, the vast majority stated that they would prefer to undergo more invasive treatment to rule out cancer than accept a small risk of missed diagnosis with expectant management. ³⁸

Special circumstances

Some groups of women are believed to have a higher incidence of endometrial cancer than the general population. These include breast cancer survivors being treated with tamoxifen and women from families with a genetic predisposition to this disease.

It is well recognized that tamoxifen has an effect on the endometrium, which may be pathological; focal cystic hyperplasia, atrophy, polyps, hyperplasia and carcinoma have all been diagnosed in women presenting with PMB while taking tamoxifen. ³⁹ Tamoxifen can produce subendometrial oedema, making it difficult to obtain an accurate endometrial thickness measurement by TVS. Dijkhuizen et al. ⁴⁰ demonstrated that 30% of postmenopausal breast cancer survivors taking tamoxifen had evidence of an endometrial abnormality on scan, compared with 6% in the reference group (postmenopausal breast cancer survivors not taking tamoxifen). Women on tamoxifen had a significantly thicker endometrium, a larger uterine volume and more uterine fluid. Nearly a quarter of the group taking tamoxifen had a suspicious ‘swiss cheese’ appearance to the endometrium on scan, but this was not associated with intracavity pathology. This high false-positive rate brings into question the value of ultrasonography in women on tamoxifen who present with PMB. It has therefore been suggested that hysteroscopy and EB should be the first-line investigation in this group of women. ⁴¹

The risk of endometrial cancer is also increased in certain families, for example those with hereditary non-polyposis colorectal cancer (HNPCC). Although there is no prospective evidence for screening for gynaecological disease in HNPCC, current recommendations include the use of ultrasound and EB. Wood et al. ⁴² looked at anxiety levels in women from HNPCC families undergoing gynaecological screening in the form of TVS, hysteroscopy and EB. They concluded that screening for gynaecological malignancy in HNPCC does not appear to be associated with any psychological morbidity. The evidence regarding optimum frequency of screening is lacking; therefore, it would appear prudent to investigate PMB in this high-risk population outwith any screening programme.

Conclusion

PMB is a symptom of possible gynaecological malignancy. Units should have protocols available for the investigation of these women, where TVS should be the initial investigation. OPH and EB are the gold standard diagnostic modalities for further investigation of those women with an increased endometrial thickness on scan.

Further research is needed on women's understanding about PMB, knowledge of the risk factors for endometrial cancer and experiences of referral for and assessment of PMB to inform awareness raising and development of patient pathways and support strategies.

Competing interests

None declared.