Need for withdrawal bleeding

The benefits of estrogen replacement therapy in estrogen-deficient women have been widely documented. Such treatment relieves vasomotor symptoms, improves mental functions, prevents bone loss and reduces the risk of cardiovascular disease. Administered to women with an intact uterus, unopposed estrogen can stimulate endometrial growth, resulting in a high incidence of endometrial hyperplasia and endometrial carcinoma; this can lead, in turn, to a high incidence of gynecological intervention. It has been established that the administration of a progestogen for 10-12 days every 28 days may protect the endometrium against hyperplasia and diminish the risk for developing endometrial carcinoma. The addition of a progestogen in a cyclical manner results in the reinitiation of cyclical bleeding, which for the majority of postmenopausal women is not a welcome event, is regarded as a burden that serves little purpose, and becomes tolerable only as a price to be paid for the relief of symptoms. To promote long-term continuation with postmenopausal hormone replacement therapy (hormone replacement therapy), the regulation of menstrual episodes, in terms of cycle length, duration and amount of bleeding, becomes an essential clinical skill to facilitate the care of postmenopausal women. The alternative would be the administration of an amenorrheic regimen, but this is not a straightforward solution.

Relative safety of hormone replacement therapy preparations

Down-regulation of estrogen receptor expression is the intended action in the uterus of continuous progestogen administration, but given the ubiquitous distribution of the estrogen receptors throughout the body, the issue of long-term safety adds further concerns about the suitability of available combinations and their potential attenuation of bone sparing, or cardiovascular benefits.

The long-term safety of one type of continuous combined hormone replacement therapy (ccHRT) regimen has been addressed in two reported randomized placebo controlled trials in the primary (Women’s Health Initiative, WHI) and secondary (Heart and Estrogen/progestin Replacement Study, HERS) prevention of ischemic heart disease, which used conjugated equine estrogens and medroxyprogesterone acetate. Unopposed by a progestogen, conjugated equine estrogens have been in use for about 50 years in hysterectomized women, as well as in those with an intact uterus, and the majority of observational studies point to its preventive effects against ischemic heart disease. Indeed, the estrogen-placebo-controlled arm of the Women’s Health Initiative study has been allowed to continue, since no increased incidence of coronary events or breast cancer has been noted. The adverse coronary artery constrictive effect of medroxyprogesterone acetate has been elegantly described in primates. Sequential hormone replacement therapy regimens have not yet been subjected to a randomized controlled trial for the primary or secondary prevention of ischemic heart disease. However, shorter tissue exposure to the progestogen may be rectified by the subsequent, longer, estrogen-only phase.

Regularity of the natural menstrual cycle and bleeding intervals

Endometrial growth, differentiation and subsequent shedding represent a highly integrated cascade of events in response to cyclical ovarian function. Therefore, natural menstruation is an expression of failure of potential implantation during the previous cycle, and represents in itself the start of another series of preparative steps in anticipation of hormonal changes in the following ovarian cycle. For the woman, menstruation is perceived evidence of potential fertility and health, and by inference is a feature of the young female. Consequently, menstruation is a significant event in a woman’s life, and, irrespective of their sociocultural background, most women are clearly aware of their menstrual rhythm, especially the ‘regularity’ of the event.

The published literature maintain that only 13-16% of women have a range of cycle length of less than 6 days, i.e. the difference between the longest and the shortest cycle recorded. These studies confirm the wide variability of the natural cycle in the individual woman. Furthermore, these studies show a steady decline in average cycle length from the youngest to the oldest age groups, being highest for women aged 15-19 years and lowest for 20-39-year-old women. This notion forms the basis for education and counselling of women using hormone replacement therapy, when variability of bleeding intervals raises health concerns in these women.

Assessment of bleeding patterns

Menstrual dates and patterns, as collected through interviews or questionnaires, are subject to problems of recall, and such data are less accurate the longer is the interval between the occurrence of the event and its reporting. Therefore, for more accurate data collection, prospective recording is favored, and is best achieved by conscientious noting of bleeding episodes or spotting (blood loss that does not require sanitary protection) in a menstrual diary, as well as the noting of bleed-free days. Analysis of menstrual data can be based on the conventional perception of ’28-day’ menstrual cycles, and this is useful when evaluating menstrual rhythms in conjunction with a 28-day gonadal steroid regimen.

The clinically important parameters of menstrual assessment are centered on the number of bleeding/spotting episodes, mean lengths of episodes and range of lengths of bleed-free intervals; hence, the following definition have been introduced:

  1.  (1) Bleeding episode: bleeding per vaginam for 1 or more days with at least 2 bleed-free days before and after;
  2. (2) Progestogen-associated bleed (PAB): bleeding per vaginam starting between day 8 of progestogen administration of one cycle and day 7 after its discontinuation, inclusive; all other bleeding is defined as intermenstrual bleeding (1MB);
  3. (3) Bleeding interval: the number of days between the first days of bleeding of two consecutive PABs.

It is interesting to note that sequential hormone replacement therapy regimens have been marketed as 28-day treatment cycles, when one might be able to argue for a calendar month-based cyclical progestogen administration, instead. To reduce the frequency of bleeding episodes, a regimen was devised to administer the progestogen once every 3 months, but the amount of blood loss was excessive, and these women frequently experienced progestogenic adverse effects due to the high dose of the progestogen used. In 15% of endometrial samples obtained at the end of the estrogen phase, endometrial hyperplasia was detected, but all were reversed after the progestogen phase. The long-term endometrial safety of such regimens, therefore, remains questionable.

Interpretation of menstrual diaries

In clinical studies involving postmenopausal women treated with a sequential combined hormone replacement therapy regimen, for example, the summary statistics of mean and standard deviation used for the description of the day of onset of bleeding, its duration and the amount of blood loss can give an idea of the overall efficacy of that regimen. However, summary statistics tend to be dominated by data obtained from women with frequent, short cycles, and may give a falsely optimistic impression of the efficacy of a given regimen in the management of the individual woman.

A menstrual diary completed by the woman represents individual responses to a particular treatment, and can be used for comparison with the natural physiological cycles or with other treatment regimens without forcing the data into an artificial method of analysis, the summary statistics, to which neither the woman nor the clinician is able to relate. Analysis of individual women’s diaries and stratifying them by the mean day of onset of bleeding has identified two groups of women: early bleeders and late bleeders. Early bleeders, whose mean day of onset of bleeding commenced before the end of the progestogen phase, had a wide variability of the day of onset of bleeding (i.e. poor predictability), with longer and heavier episodes of withdrawal bleeding compared with those women whose mean day of onset of bleeding occurred after the end of the progestogen phase, or late bleeders. This method was prospectively applied to a dose-ranging study of a new progestogen, trimegestone, in which the bleeding characteristics of early and late bleeders were confirmed and the dose of progestogen was the main determinant for being an early or a late bleeder. Indeed, the clinical applicability of this method of evaluation of the individual woman’s experience of bleeding was further documented when the dose of trimegestone was changed, and in comparative studies with other progestogens. Therefore, in clinical practice, we recommend analysis of the individual woman’s diary, collected during a treatment phase, to adjust the dose of treatment or mode of administration of hormone replacement therapy regimens.

Severity of bleeding

The amount of blood loss is difficult to document, since it is difficult to evaluate between individual women, in whom the perceived amount of loss depends on myriad factors related to personal hygiene and social and cultural attributes. A method for measurement of blood loss by alkaline elution of hematin from collected sanitary towels has been established. The improved accuracy of measurement of blood loss is countered by the laborious nature of the technique, which requires stringent precautions against transmission of blood-borne diseases and is certainly not practical for large-scale studies. A surrogate measure was suggested by Higham and colleagues, based on a pictorial chart. The woman completes her menstrual diary on a daily basis, and refers to the degree of soiling of sanitary towels or tampons. However, the reproducibility of this method has recently been questioned when applied to a group of women; nonetheless, withinsubject reproducibility is much more important in the management of the individual woman.

Intermenstrual bleeding

In analysing menstrual diaries, problems arise when trying to include episodes of spotting during the main menstrual phase, or the occurrence of truncated events, i.e. bleeding or spotting that occurs 1 or 2 days before or after the main menstrual event. Statistically presented summaries of menstrual data, particularly those collected over long observation periods, may not be seriously affected by such episodes in naturally menstruating women who are not using contraceptives. In these women, they are largely accepted as part of the physiological phenomenon, but not in postmenopausal women using hormone replacement therapy.

Assessment of bleeding patterns in amenorrheic regimens

Since the re-establishment of uterine bleeding dissuades women from long-term use of hormone replacement therapy, an amenorrheic regimen was devised by providing continuously administered estrogen and a progestogen (ccHRT). The expectation in this case is that the progestogen will act significantly only on the uterus to induce atrophic endometrium that will not be shed.

The induction of an amenorrheic state is the most important feature for the increasing popularity of this type of hormone replacement therapy regimen. However, as many as 50% of women who commence ccHRT will experience bleeding episodes during the first 3-6 months, although for those who continue the treatment, amenorrhea is achieved in about 70-90%, with high continuation rates. Other than the bleeding problems, women discontinue ccHRT owing to the adverse effects of progestogens, such as bloatedness, fluid retention and dysphoria.

A reference-period method of observation based on multiples of 30 days has been suggested, since menstruation occurs as a ‘monthly event. The reference-period method is particularly useful when assessing bleeding events associated with an intrauterine contraceptive device, long-acting injectable contraceptive steroids and the progestogen-only pill, on the ‘mini pill. Therefore, this method is more applicable for the evaluation of bleeding episodes associated with ccHRT regimens. With this method, the bleeding events are assessed as they happen during the observation period, but many such events will overlap the boundaries of the reference period. Regardless of the length of the reference period, the recording must start with the commencement of treatment, which permits uniformity of assessment between subjects. Data obtained through the reference-period method may not be easily assimilated in a clinical situation, particularly from the woman’s viewpoint, since amenorrhea is the goal of ccHRT and any bleeding episode represents, in principle, treatment failure.

We recommend the use of 90-day referenceperiod diaries for practical reasons. First, it is implicit in this practice that bleeding events, should they occur with ccHRT, are assessed over a longer period and that the woman should not panic when such episodes occur. Second, this approach enables the woman who bleeds initially with ccHRT to evaluate the change in the pattern of bleeding during the second 3-6 months, and reassure herself of the progressive subsidence of these episodes; thus, in the majority of instances it helps to promote perseverance with treatment. Third, such an approach will reduce the frequency of unnecessary clinic visits. Finally, these diaries may also prompt uterine investigation should bleeding persist or become heavier, particularly after the first 3 months of treatment. In clinical studies, the 90-day reference period simplifies the assessment of bleeding episodes, and reduces the frequency of statistical handling of truncated bleeding events that overlap the boundaries of the reference periods.

Histology of the endometrium

Changes in endometrial histology in response to the combined oral contraceptive pill are frequently described as ‘retarded’, and similarly the endometria of postmenopausal women treated with hormone replacement therapy. These descriptions are largely related to poorer glandular development and stromal changes, compared with the luteal-phase endometrium of the natural cycle. This can be explained by the fact that the ovaries produce many intermediary steroid metabolites, all of which may contribute to gonadal steroid receptor activation. Consequently, these metabolites may influence the enzyme systems responsible for metabolic clearance of these steroids and their cognate receptors from the cell machinery. The metabolic handling and bioavailability of the constituting steroids used in combined oral contraceptive and in hormone replacement therapy preparations result in different intermediary metabolites and, as such, may not be adequate to sustain the physiological growth and integrity of the endometrium, as is the case with the natural cycle. Another collective description commonly used by histopathologists is that of ‘dyssynchronous endometrium’, which denotes the out-of-phase changes in this tissue. This was well illustrated by Good and Moyer’s approach to the adequacy of estrogen or progestogen contained in a particular treatment regimen, and was further documented in a mathematical model using a linear discriminant analysis technique. Different progestogens, for example, induce different vascular patterns in the endometrium, with androgenic types resulting in smaller and a higher number of vascular spaces compared with the non-androgenic trimegestone or progesterone.

Factors that influence bleeding patterns with hormone replacement therapy

Ethnicity and geographical location

These may play an important role in the pattern of menstruation. European women tend to have more bleeding/spotting days, while women from Latin America have shorter bleeding episodes and longer bleed-free intervals. These features of menstrual loss have been observed in those using oral contraceptive pills, as well as in those menstruating spontaneously. Furthermore, only 25% of European women treated with depot medroxyprogesterone acetate experienced amenorrhea by their fourth injection, compared with 72% of women in North Africa . However, there are no comparative data on the influence of ethnicity or geographical location on menstrual events associated with the use of hormone replacement therapy.

Nature of the hormones used

Compared with the combined oral contraceptive pill, orally administered hormone replacement therapy consists of a relatively low dose of estrogen and a high dose of progestogen. One of the main pathophysiological processes responsible for unscheduled bleeding in postmenopausal women taking hormone replacement therapy is the bio-availability of the estrogen used. In particular, the susceptibility of estradiol for inactivation by 17β-hydroxysteroid dehydrogenase and subsequent metabolic clearance to the weaker estrone and its conjugates is higher than that of the resistant ethinylestradiol, or of its 3-methyl ether, mestranol. In addition, the bioavailable dose of the sequential progestogen to the uterus may be inadequate to sustain endometrial stability until the end of the progestogen phase.

Increasing the dose of the progestogen may overcome this problem in some cases, but this is counterbalanced by a higher incidence of bloatedness, fluid retention and mastalgia. Furthermore, one should be mindful of the potential metabolic effect of long-term administration of high doses of progestogens, and therefore such regimens should be reviewed regularly. Alternatively, a higher dose of ethinylestradiol (20-30 µg) m ay be required, but there is general concern about prescribing higher doses of estrogen to older women.

Various progestogens have been used in combination with estrogen. Norethisterone is one of the most widely used progestogens in cyclical sequential regimens. However, in 50% of instances, withdrawal bleeding occurs before the end of the progestogen phase of treatment Z. Medroxyprogesterone acetate has been used in preference to norethisterone because of its lesser androgenic adverse effect; however, it results in a less favorable bleeding pattern compared with other progestogens. Women taking medroxyprogesterone acetate bleed earlier in the progestogen phase, compared with levonorgestrel (LNG) or desogestrel.

Endometrial polyps

Endometrial polyps are localized outgrowths of endometrial tissue covered by epithelium, and contain a variable amount of glands, stroma and blood vessels. They are encountered most commonly during the fifth decade of life. In premenopausal women they may present clinically as intermenstrual bleeding or as heavy periods, but sometimes the clinical presentation is that of excessive vaginal discharge.

Endometrial polyps vary in size, from 1 mm to a large fleshy mass that may protrude through the cervix. They can be broad-based and sessile, or attached to the endometrium by a slender stalk. They are usually solitary, but multiple polyps can be found as well. The glandular element of a polyp is characteristically out of phase with the rest of endometrial development, and secretory changes in these glands are seldom found. Occasionally, hyperplasia or early carcinoma may develop exclusively in the polyp. There are indications for an increased incidence of endometrial hyperplasia and carcinoma in women found to have endometrial polyps, but this may be due to the increased propensity of such endometria to develop neoplasia rather than an inherent feature of the polyps.

In a prospective cohort study involving 248 women who underwent out-patient hysteroscopy and endometrial biopsy, 62 had endometrial polyps. These were compared with 186 endometrial samples that did not have polyps, taken as control. The incidence of hyperplasia in polyps was 11.3% compared with 4.3% in control samples, while the incidence of carcinoma was identical in both groups (3.2%).

Removal of the polyp with endometrial curettage may be all that is required, but a check hysteroscopy is necessary, since the consistency of these polyps may be so pliable that they are missed by the curettage. Where hyperplastic changes are found localized to the polyp in premenopausal women, then polypectomy with curettage may be adequate. However, in postmenopausal women hysterectomy will be the appropriate choice if atypia is also present.

Submucous fibroids

The presence of submucous fibroids has been linked to the development of abnormal uterine bleeding. We documented a two-fold increase in the risk of having submucous fibroids in premenopausal women who presented with heavy and irregular menses, and a three-fold increase for the presence of these fibroids in postmenopausal women who presented with heavy and irregular bleeding with hormone replacement therapy, compared with asymptomatic controls. Furthermore, the presence of submucous fibroids has been associated with heavy and prolonged bleeding in addition to a higher incidence of 1MB in postmenopausal women taking hormone replacement therapy.

The identification of polyps and submucous fibroids may help to predict a woman’s response to sex steroid therapy, and aids the planning of local surgical management if deemed necessary. Moreover, 87% of postmenopausal women with heavy and unscheduled bleeding with hormone replacement therapy, who had their submucous fibroids removed by hysteroscopic laser myomectomy, improved and experienced regular bleeding when they resumed their hormone replacement therapy. The result of this intervention, however, was less successful in premenopausal women, and adds further evidence that abnormal uterine bleeding in the presence of submucous fibroids in this group of women forms only part of the abnormal hormonal responsiveness of the endometrium. In women who present with spontaneous postmenopausal bleeding, the presence of submucous fibroids or polyps was not found to be more frequent than in controls. It has long been recognized that the development of the endometrium overlying submucuos fibroids is less advanced, and shows a mixed picture of proliferative, secretory and poorly developed patches compared with the rest of the adjacent endometrium. Whether such endometrial defects are due to the altered vascular pattern that supplies these fibroids, or due to paracrine factors induced as a result of the development of the fibroids, is not known.

Poor compliance with hormone replacement therapy

Irregular use of hormone replacement therapy regimens is an important issue that deserves thorough consideration in the clinical assessment of women presenting with heavy or unscheduled bleeding, since missing tablets, for example, may alter the bio-availability of gonadal steroids, render the endometrium unstable and cause bleeding. Where progestogenic adverse effects are experienced, women try to stop using part or the whole of the progestogen phase of the hormone replacement therapy regimen, to avoid these symptoms. It is estimated that about 30% of women who are prescribed sequential combined hormone replacement therapy regimens stop taking the progestogen component owing to symptoms of fluid retention, bloatedness, mastalgia and mood swings; consequently, the endometrial changes will be similar to those on unopposed estrogen. In the Postmenopausal Estrogen/Progestin Interventions trial, 62.5% of women with an intact uterus who used unopposed estrogen for 3 years developed abnormal bleeding or had an abnormal Papanicolaou test, which necessitated further investigation. Eighty per cent of unopposed estrogen users had endometrial abnormalities, of which 80.5% were simple hyperplasia, 18.8% atypical hyperplasia and 0.7% endometrial carcinoma.

Endometrial neoplasia

The incidence of endometrial carcinoma varies with race and geographical location. Adjusted for age, the rate of endometrial cancer in the USA is reported to be from 17.6 to 22.2 per 100 000 in white women and 10.4 for Afro-Caribbean women, while that among Chinese women is 10-15 times lower. Endometrial cancer is associated with one of the highest survival rates for all cancer sites; 74% of uterine cancer cases are diagnosed at a localized stage, and a great majority of women under the age of 50 have their cancer diagnosed at an early stage.

Hormonal treatment and endometrial cancer

Chronic unopposed estrogen administration induces a proliferative endometrium, which progresses to a well-differentiated endometrial adenocarcinoma in a significant number of women. Such progress is thought to pass through several intermediary stages of proliferation of increasing complexity, collectively termed ‘endometrial hyperplasia. The duration of use of unopposed estrogen in postmenopausal women remains the strongest predictor for the development of endometrial cancer. The relative risk increases exponentially from 1.4 (95% confidence interval (CI) 1.0-1.8) for less than 1 year of use to 9.5 (95% CI 7.4-12.3) among users for 10 years or more.

The effect of unopposed estrogen use on the risk of endometrial cancer is also directly related to the dose of estrogen”, although a consistent dose-response relationship is lacking. Cessation of therapy results in a rapid decline of the risk within about 2 years, but may remain higher than in the non-user population for up to 10-15 years. The cyclical addition of progestogen for 10-14 days each month to continuously administered estrogen has been shown to confer reduction in the risk of endometrial hyperplasia and cancer. However, this addition of a progestogen does not eliminate the increased risk: 1.9 (95% CI 0.9-3.8), compared with 6.5 (95% CI 3.1-13.3) for unopposed estrogen users.

Indeed, Weiderpass and colleagues showed that sequential regimens with the progestogen being administered for fewer than 16 days in a 28-day treatment cycle were associated with increased risk of endometrial cancer: odds ratio (OR) 2.9 (95% CI 1.8-4.6) for 5 or more years of use. In the same report, users of ccHRT enjoyed significant protection against endometrial hyperplasia and cancer: OR 0.2 (95% CI 1.1-1.8) for 5 or more years of use. The increased risk of endometrial cancer was dependent on the type and duration (days per treatment cycle) of the progestogen used. For example, 19-nortestosterone derivatives conferred higher protection than the progesterone-derived progestins.

Investigation of abnormal bleeding with hormone replacement therapy

The clinical principles that govern the management of abnormal bleeding associated with hormone replacement therapy are those of the management of spontaneous postmenopausal bleeding.

A detailed clinical history to include systems review and drug history is essential to exclude conditions that may be the underlying cause of bleeding, or which may influence subsequent management. A general physical examination including the breasts is required to establish normality, in particular a thorough abdominal examination to detect areas of tenderness, possible masses or ascites. Pelvic examination must be systematically conducted to exclude local vulvar, vaginal or cervical lesions. A cervical smear for cytological assessment is obtained, since the detection of abnormal squamous or glandular cells will be a clear indication for colposcopic assessment of the cervix and endometrial evaluation in this group of women. A bimanual examination is performed to assess overall uterine size, mobility of the uterus, characteristics of the supportive tissue and existence of adnexal masses.

Ultrasound scan

Pelvic organ imaging using ultrasound is a fundamental investigative tool in modern gynecologic practice. Ultrasound scanning of the pelvic organs helps to define the uterine silhouette, uterine cervix and maximum endometrial thickness, corresponding to the apposed endometrial surfaces, in the sagittal plane. It may also identify the contrasting echoes of uterine fibroids or any fluid collection within the endometrial cavity. This imaging technology is invaluable in excluding ovarian cystic lesions.

Attempts at color Doppler imaging in conjunction with vaginal ultrasound scan to evaluate endometrial neovascularization did not fulfil the initial promise of adding further information obtained by ultrasound scan alone, and the technique did not discriminate between benign and malignant disease.

Saline instillation sonohysterography

The instillation of a contrast medium into the uterine cavity using normal saline helps to separate the anterior and posterior walls of the endometrial cavity” and delineate the presence of endometrial polyps, and sometimes submucous fibroids, as they distort the endometrial surface. Recently, a prospective comparison of the accuracy of the saline instillation sonohysterography (SIS) technique with that of panoramic hysteroscopy has documented its shortcomings.


Earlier reports described the use of the hysteroscope under general anesthesia; however, it is widely accepted that out-patient hysteroscopy with or without local anesthesia is a very well-tolerated procedure, and in experienced hands is just as accurate. Moreover, it is easier to repeat when required, for example, in cases of recurrence of postmenopausal bleeding.

The use of CO2 gas or fluid distension for panoramic hysteroscopy offers a direct visual examination of the whole of the endometrial cavity. It helps to identify endometrial abnormalities such as polyps and submucous fibroids, and regional differences in endometrial thickness and its background color, as well as the endometrial vascular pattern. The outer diameter of the instruments in common use ranges from 3.5 to 6 mm and, therefore, the need for cervical dilatation is limited. The disadvantages of using a gas distension medium include bubble formation, but, by waiting for 30-60 s without manipulating the hysteroscope, these bubbles will burst, and the view of the cavity will become clearer. Note that the escape of gas through the Fallopian tubes may cause diaphragmatic irritation with a typical shoulder-tip referred pain. Alternatively, and especially in the event of bleeding, distension of the uterus with normal saline can be used.

Endometrial sampling

Having assessed the uterine cavity by ultrasound scan and/or by office-based hysteroscopy, an endometrial biopsy is essential to establish a definitive histologic diagnosis. The endometrial biopsy is obtained using a pipelle de Cornier sampler or the Vabra aspirator. Should there be a localized lesion deemed not to have been adequately sampled, hysteroscopically directed biopsy with a biopsy forceps or with the LENS device (Leicester endometrial needle sampler) will help.

Choice of method of investigation

The choice of method(s) of investigating the endometrium should be based on the objectives of that clinical encounter. If the question is purely whether the patient has endometrial neoplasia, then an adequate endometrial biopsy should suffice. If clinical evaluation of postmenopausal bleeding is required, then the clinician is obliged to provide a thorough clinical assessment and full investigation of the reproductive system.

The number of menopausal women is increasing, with the ‘baby boomers’ entering their sixth decade of life, and the problem of functional bleeding disorders may be encountered more often with the increase in hormone replacement therapy use. These functional bleeding disorders mandate effective methods in identifying endometrial polyps and submucous fibroids, since their removal is highly successful in rectifying abnormal bleeding while using hormone replacement therapy. Moreover, appropriate and detailed evaluation of the endometrial histology may guide corresponding adjustment in the dose of estrogen or progestogen (see above).

The enthusiasm for ultrasound evaluation of the endometrium prompted many investigators to test how safe it would be to forgo the need for endometrial biopsy. Cutoff points in endometrial thickness have been tested, and the 5-mm limit has been widely adopted. A prerequisite of a clear symmetrical endometrial echo of less than 5 mm encouraged Briley and Lindsell to omit endometrial biopsy.

Many reports, however, have shown that cancer cannot be excluded when endometrial thickness is below 5 mm, and have suggested 4- and 3-mm cut-off points instead. Schramm and colleagues showed that the median endometrial thickness among uterine cancer patients was 4 mm, as opposed to 5.5 mm for those with an atrophic endometrium, in a prospective evaluation of transvaginal ultrasound (TVU) compared with dilatation and curettage (D&C) as an independent procedure in 195 women presenting with spontaneous postmenopausal bleeding. Endometrial cancer was detected in 29 women (15%). By adopting a cut-off point of >4 mm for endometrial thickness, the sensitivity and specificity of TVU was 62 and 50%, respectively. Similar concerns were reported from a series of 54 women who presented with spontaneous postmenopausal bleeding, in which median endometrial thickness as measured by TVU was 5, 8.6 and 6 mm for histologically documented endometrial atrophy, hyperplasia and adenocarcinoma, respectively. Moreover, of the nine malignant samples, three cases had an endometrial thickness of 3 mm.

The SIS technique is claimed to provide accurate diagnosis of atrophic endometrium and endometrial polyps with a sensitivity of 87.8 and 89.6%, and specificity of 90.7 and 95%, respectively. However, in this study, SIS diagnosed only 40% of endometrial cancers, confirming the fundamental requirement for histologic examination of the endometrium in all cases. Soares and colleagues equated the accuracy of SIS to that of the ‘gold standard’ hysteroscopy and endometrial biopsy in the diagnosis of benign lesions, i.e. polyps, submucous fibroids and endometrial hyperplasia in premenopausal infertile women, but it failed to diagnose accurately intrauterine adhesions. A more cautious approach was suggested by O’Connell and co-workers, who concluded that an office-based SIS with endometrial biopsy would be reliable enough in the management of postmenopausal bleeding when no endometrial abnormality was detected. Where intrauterine luminal masses are suspected, then the patient should be referred for hysteroscopy and fractional curettage.

In a series of 1398 women, 43% of whom were postmenopausal, diagnostic hysteroscopy and subsequent curettage confirmed that, in addition to being a basic tool in gynecological practice, the sensitivity, specificity and global diagnostic precision of hysteroscopic evaluation of the endometrium was 100, 99.4 and 99.5%, respectively. A similar experience was recorded in 980 women presenting with menstrual disorders, and in another series of 106 women who presented with postmenopausal bleeding, confirming the accuracy, safety and simplicity of out-patient hysteroscopy and endometrial sampling.

Management of irregular bleeding with HRT

The occurrence of an unscheduled or abnormal bleeding pattern in women using a sequential combined hormone replacement regimen is an indication for ultrasound assessment of the pelvic organs, and for endometrial evaluation by hysteroscopy and curettage. In women using ccHRT, the occurrence of irregular bleeding during the first 3 months of treatment is quite common. Full investigations will be required if such bleeding lasts for more than 6 months, or earlier in women with higher risks of developing endometrial neoplasia such as those with obesity, diabetes mellitus, nulliparity or a family history of endometrial or colonic carcinoma. In a population-based study, the incidence of adenomatous and atypical hyperplasia was estimated at 44 per 100 000 per year, and one of the characteristics of these women was overall obesity, rather than specific regional fat distribution. Despite the identification of risk factors, investigation of the endometrium cannot be safely withheld from women who do not have these risks. Where malignant disease is excluded, and in the absence of endometrial polyps or submucous fibroids, adjustment of the hormone replacement therapy regimen is attempted. A common approach is change of the progestogen to a more androgenic type, but this may cause bloatedness, fluid retention and mood swings. The use of a levonorgestrel-laiden intrauterine device (Mirena ® coil) has been advocated, but this is not without drawbacks. Prolonged bleeding and spotting can be anticipated for up to 1 year, and the systemic absorption of levonorgestrel may result in androgenic adverse effects. Furthermore, long-term safety as a result of its systemic absorption with regard to the cardiovascular system and bone resorption has not been evaluated. Another approach involves reducing the dose of estrogen, but this may risk a recurrence of vasomotor symptoms or may not be adequate to control bone loss. A non-oral route of administration is generally associated with less bleeding, and this may be related to lesser amounts of estrone generated. In women treated with ccHRT a lower-dose formulation may be adequate. However, the view of the present authors is to offer a change to a sequential combined hormone replacement therapy regimen as the treatment of choice. This may improve endometrial thickness and thus result in more predictable withdrawal bleeding, which may be acceptable to the woman especially if these bleeding episodes are short and light.

Bleeding patterns and hormone replacement therapy: Conclusions

Abnormal bleeding in women treated with hormone replacement therapy is an important cause of discontinuation of treatment. A woman’s ability to predict the day of onset of bleeding in the following cycle, for example, helps her to plan various life events, and increases her confidence to continue with treatment in the long term. Understanding the bleeding phenomenon and accurate prediction of which woman is likely to have an abnormal bleeding pattern may enhance the clinical management of these women, and enable the clinician to adjust the hormone replacement therapy regimen appropriately. In postmenopausal women presenting with abnormal or unscheduled bleeding while receiving gonadal steroid treatment, thorough endometrial evaluation is mandatory. In the absence of neoplasia or structural abnormalities of the endometrial cavity, tailoring the treatment regimen to the individual woman’s situation facilitates the administration of very effective health-promoting therapeutics. Such endeavor should be guided by the clinical presentation, information gleaned from menstrual diaries and detailed evaluation of the functional aspects of endometrial histology. Incomplete assessment can lead to inappropriate treatment, which eventually frustrates patients and prompts many of them to opt for cessation of treatment or for probably an unnecessary hysterectomy.


Selections from the book: “The Management of the Menopause” (2003).

Share →