Abnormal uterine bleeding and uterine pathology in patients on tamoxifen therapy

INTRODUCTION — Tamoxifen is a nonsteroidal selective estrogen receptor modulator (SERM) that is used primarily for adjuvant treatment of estrogen receptor-positive breast cancer in premenopausal and some postmenopausal patients. It is also used for chemoprevention in selected patients at increased risk for breast cancer.

Tamoxifen is associated with an increased risk of uterine pathology. Benign endometrial polyps are the most common type of pathology. Other types of uterine pathology (eg, endometrial hyperplasia and carcinoma, uterine sarcoma, uterine carcinosarcoma) may also occur and the risk depends on many factors, including menopausal status and whether tamoxifen is being used for chemoprevention or treatment of breast cancer.

Uterine pathology in patients on tamoxifen is reviewed here. Use of tamoxifen for breast cancer therapy or breast cancer prevention and management of other adverse effects associated with tamoxifen are reviewed separately.

●(See "Adjuvant endocrine and targeted therapy for postmenopausal women with hormone receptor-positive breast cancer".)

●(See "Selective estrogen receptor modulators and aromatase inhibitors for breast cancer prevention".)

●(See "Managing the side effects of tamoxifen and aromatase inhibitors".)

ENDOMETRIAL EFFECTS — Tamoxifen is a selective estrogen receptor modulator (SERM) with mixed estrogen antagonist and agonist properties. In the breast, tamoxifen is antiestrogenic; it binds to estrogen receptors and blocks tumor proliferation [1,2]. Its mechanism of action on the endometrium, however, is more complex.

In premenopausal patients, the estrogen antagonist effects of tamoxifen in the endometrium predominate because endogenous estrogen levels are high (typically >20 pg/mL). The effect of tamoxifen on hypothalamic-pituitary-ovarian axis in such patients is similar to clomiphene: tamoxifen partially blocks estradiol's negative feedback loop and therefore up-regulates gonadotropin-releasing hormone agonist secretion from the hypothalamus. This results in an increase in follicle-stimulating hormone and, more notably, estradiol levels. However, the mechanisms of action are complex and data in humans are limited [3].

In postmenopausal patients, the estrogen agonist effects of tamoxifen in the endometrium can predominate because endogenous estrogen levels are low (typically <20 pg/mL) [4]. Thus, tamoxifen stimulates endometrial proliferation by promoting [5]:

●Cell proliferation through mitogen-activated protein (MAP) kinase pathways, c-MYC, and insulin-like growth factor 1 (IGF1) pathways.

●Cell proliferation and invasion through alterations in estrogen receptor-alpha and the membrane-associated estrogen receptor G protein-coupled receptor 30 (GPR30).

●Cell migration through extracellular signal-regulated kinase (ERK) and Src signaling.

●Deoxyribonucleic acid (DNA) damage through DNA adduct formation.

●Upregulation of the prosurvival unfolded protein response (UPR) pathway.

However, the identification of all tamoxifen targets in the endometrium is likely incomplete and the effects of long-term tamoxifen exposure on molecular changes remain unclear [5].

It is also well recognized that KRAS mutations (an oncogene) occur at a high rate (40 to 75 percent) in the endometrium of both pre- and postmenopausal patients on tamoxifen [6,7]. However, after cessation of tamoxifen, KRAS mutations appear to resolve and remain absent even in long-term follow-up [7]. (See 'Risk after discontinuation of tamoxifen' below.)

RISK OF UTERINE PATHOLOGY — Patients on tamoxifen therapy are at an increased risk of developing uterine pathology (eg, endometrial polyps, hyperplasia, carcinoma); within four years, this risk is up to 67 percent [8-12]. Benign endometrial polyps are the most common type of pathology.

Other types of uterine pathology (eg, fibroids, adenomyosis) and cervical neoplasia do not appear to be associated with tamoxifen, and tamoxifen does not need to be discontinued in patients with these pathologies.

Endometrial polyps — Endometrial polyps are the most common type of endometrial pathology associated with tamoxifen use [11,13-16]. Rates of polyp formation differ depending on whether tamoxifen is being used for chemoprevention or treatment of breast cancer and may be higher in patients with a history of endometrial polyps. Representative studies are as follows:

●In the National Surgical Adjuvant Breast and Bowel Project P-2 (NSABP P-2, also referred to as the STAR) trial including 4693 postmenopausal patients with an intact uterus who were treated with tamoxifen for chemoprevention of breast cancer, the average annual rate of endometrial polyps was 21.1 per 1000 females [13]. Higher rates (30 to 60 percent) have been reported in postmenopausal patients using tamoxifen for treatment of breast cancer; however, evaluation for endometrial polyps prior to tamoxifen exposure was not performed [17].

●In a retrospective study including over 78,000 premenopausal patients (mean age 42 years) with breast cancer in Korea, those treated with versus without tamoxifen had higher rates of endometrial polyps (20.1 versus 5.5 cases per 1000 person-years, hazard ratio [HR] 3.9, 95% CI 3.7-4.2) at six years of follow-up [18]. However, this study had several limitations including that patients with breast cancer receiving tamoxifen in Korea receive annual gynecologic screening which may help diagnose uterine diseases more easily than those who are not treated with tamoxifen and not receiving annual screening.

●In a prospective study of over 500 postmenopausal patients with breast cancer receiving tamoxifen, the rate of de novo polyp development was higher in those with a polyp removed prior to starting tamoxifen compared with those in whom a baseline polyp was ruled out (17.6 versus 12.9 percent) [19]. Importantly, those with a polyp removed prior to starting tamoxifen also had higher rates of developing atypical endometrial lesions (11.7 versus 0.7 percent) while on tamoxifen therapy.

Polyps associated with tamoxifen may be large (>2 cm), multiple, or show molecular alterations [20-23]. Patients with a history of an endometrial polyp prior to tamoxifen therapy appear to have a higher rate of recurrent endometrial polyps while on tamoxifen [19,24].

Most endometrial polyps are benign. However, some polyps will undergo malignant transformation, which is more common in post- compared with premenopausal patients on tamoxifen [20,25-27]. This rate is also higher than those not on tamoxifen (see "Endometrial polyps", section on 'Risk of malignancy'). Polyp size and duration of tamoxifen use do not appear to be associated with malignant transformation.

Endometrial polyps are discussed in more detail separately. (See "Endometrial polyps".)

Endometrial hyperplasia — Postmenopausal patients on tamoxifen have an increased risk of endometrial hyperplasia. Patients who are premenopausal before starting tamoxifen for adjuvant breast cancer therapy also may be at an increased risk of endometrial hyperplasia. However, other breast cancer treatment-associated factors, including those impacting ovarian function, may contribute to the risk of endometrial hyperplasia in this complex group of patients.

In the NSABP P-2 trial including postmenopausal patients on tamoxifen discussed above (see 'Endometrial polyps' above), the average annual rate of endometrial hyperplasia was 4.4 per 1000 females; among those diagnosed with hyperplasia, the majority (82.5 percent) were without atypia [13]. In a subsequent retrospective study including over 78,000 premenopausal patients (mean age 42 years) with breast cancer in Korea, those treated with versus without tamoxifen had higher rates of endometrial hyperplasia (13.5 versus 2.1, HR 5.6, 95% CI 5.1-6.1) at six years of follow-up [18]. Some limitations to this study are discussed above. (See 'Endometrial polyps' above.)

Endometrial hyperplasia is discussed in detail separately. (See "Endometrial hyperplasia: Clinical features, diagnosis, and differential diagnosis".)

Endometrial carcinoma

Post- versus premenopausal patients

●Postmenopausal patients – Postmenopausal patients on tamoxifen have an increased risk of endometrial carcinoma (EC); this risk is well established and based on data from randomized trials:

•In the NSABP P-1 trial including 13,338 patients at high risk for breast cancer, patients ≥50 years receiving tamoxifen prophylaxis (20 mg/day for five years) compared with placebo had a higher rate of EC at five years (3.05 versus 0.76 per 1000 females; risk ratio [RR] 4.01, 95% CI 1.7-10.9) and at seven years (RR 5.33, 95% CI 2.47-13.17) [28,29].

•In the NSABP P-2 trial including postmenopausal patients on tamoxifen for chemoprevention of breast cancer discussed above (see 'Endometrial polyps' above), the average annual rate of EC was 2.3 per 1000 females, respectively [13].

Higher rates of EC are reported in postmenopausal patients with abnormal bleeding. In a cohort study of patients with breast cancer treated with tamoxifen and undergoing hysteroscopy, the overall rate of EC in patients with postmenopausal bleeding (51 patients) was 7.8 percent [12].

●Premenopausal patients – Premenopausal patients using tamoxifen do not appear to be at the same increased risk of EC as postmenopausal patients.

In the NSABP P-1 trial discussed above, patients <50 years receiving tamoxifen prophylaxis compared with placebo had a similar risk of EC at both five and seven years; however, the number of cases of EC in premenopausal patients in these studies were low, and there was insufficient statistical power to detect such a difference [28,29]. A subsequent meta-analysis that pooled data from NSABP P-1 with another randomized trial of tamoxifen for breast cancer prevention and evaluated the risk of EC at five years found similar results [30]. Similarly, in the cohort study of patients with breast cancer treated with tamoxifen and undergoing hysteroscopy, there were no cases of EC in premenopausal patients with abnormal uterine bleeding (AUB; 19 patients) [12].

By contrast, in the large retrospective study including premenopausal patients with breast cancer in Korea discussed above (see 'Endometrial hyperplasia' above), tamoxifen was associated with higher rates of EC (2 versus 0.45 cases per 1000 person-years, HR 3.8, 95% CI 3-4.7) at six years of follow-up [18]. However, the tamoxifen cohort had higher body mass index and higher rates of diabetes, both of which are independent risk factors for EC; rates of polycystic ovarian syndrome, which is also an independent risk factor of EC, were lower in the tamoxifen group.

Duration, dose, and type of therapy

●Duration – The risk of EC increases with increased duration of tamoxifen therapy, which is clinically important as some patients are candidates for extending tamoxifen treatment from 5 to 10 years. Representative studies include the following:

•In the Adjuvant Tamoxifen: Longer Against Shorter (ATLAS) trial including 12,894 patients with early breast cancer, those who completed 10 compared with 4 years of tamoxifen had higher rates of EC (RR 1.74, 95% CI 1.3-2.34) [31]. Absolute cumulative risk of EC during years 5 to 14 compared with those who stopped tamoxifen after 5 years was 3.1 and 1.6 percent, respectively.

•In a randomized trial of 4610 postmenopausal patients with breast cancer (mean age 63 years) receiving adjuvant tamoxifen, those receiving treatment for five compared with two years had higher rates of EC (70 versus 42 cancers, HR 1.65, 95% CI 1.12-2.42) [32].

•In a case-control study including 299 patients in the Netherlands with breast cancer and a subsequent diagnosis of EC, the odds of EC increased 2.3-fold with ≥2 years of tamoxifen and 6.6-fold with ≥6 years [33].

•In the cohort study including 262 postmenopausal patient discussed above (see 'Post- versus premenopausal patients' above), the rate of EC increased with duration of therapy (<3 years: 1.3 percent versus >3 years: 11.7 percent) [12].

●Dose – It is unclear whether the dose of tamoxifen impacts EC risk.

In an analysis of individual patient data from participants in randomized trials including 37,000 female patients from 55 trials, the risk of EC in patients treated with adjuvant tamoxifen 20 or 30 to 40 mg/day compared with placebo were similar (ratio of incidence rates: 2.7 and 2.4, respectively) [34]. These risks are similar to those found in breast cancer chemoprevention trials using tamoxifen 20 mg/day in which the increased risk of EC in postmenopausal patients was approximately two- to threefold [29,35]. By contrast, in the randomized trial of postmenopausal patients with breast cancer described above, the risk of EC trended higher in those receiving 40 compared with 20 mg of tamoxifen (HR 1.44, 95% CI 0.96-2.61), but this was not statistically significant [32]. While these studies include doses of 30 and 40 mg/day, the standard dose for breast cancer treatment is 20 mg/day and the clinical benefit of higher doses has not been demonstrated. (See "Ductal carcinoma in situ: Treatment and prognosis" and "Ductal carcinoma in situ: Treatment and prognosis", section on 'Dose'.)

Increasing cumulative dose of tamoxifen, however, may increase the risk of EC. In a case-control study of 324 patients with breast cancer, an increasing cumulative dose was associated with an increased risk of EC; for each cumulative dose range the risks compared with no tamoxifen were: ≤7500 mg (odds ratio [OR] 0.94), 7501 to 15,000 mg (OR 1.37), 15,001 to 30,000 (OR 1.86), and >30,000 (OR 3.3) [36]. The cumulative dose may reflect either duration or daily dose.

●Type of SERM – Tamoxifen has also been compared with other SERMs. In the NSABP P-2 trial, those receiving tamoxifen (20 mg/day for five years) compared with raloxifene had a higher rate of EC (2.25 versus 1.23 per 1000 females) during the 81 month (median) follow-up period [13].

Risk after discontinuation of tamoxifen — Data are limited regarding the risk of EC after tamoxifen therapy is discontinued. Both clinical and molecular data suggest that the risk of EC decreases after cessation of tamoxifen and is similar to patients who have not been treated with tamoxifen by approximately two years [37,38].

In the International Breast Cancer Intervention Study-1 (IBIS-1) of 7154 patients with an increased risk of developing breast cancer, those treated with tamoxifen (20 mg) chemoprevention compared with placebo had a nearly fourfold risk (OR 3.76) of EC during the five-year treatment period [39]. After discontinuation of tamoxifen, the risk of EC was similar in the tamoxifen and placebo arms during the subsequent 5- and 16-year surveillance periods. Similarly, KRAS mutations, which occur at a high rate in the endometria of patients on tamoxifen, appear to resolve and remain absent after cessation of tamoxifen, even in long-term follow-up [6,7]. (See 'Endometrial effects' above.)

Studies have used ultrasound determination of endometrial thickness as a surrogate for endometrial pathology after discontinuation of tamoxifen therapy. In one cohort study of 153 patients followed with ultrasound measurement of endometrial thickness after discontinuing tamoxifen, no patients developed EC. Endometrial thickness decreased gradually over time, as did the diagnosis of any endometrial pathology. Those who had persistent thickening of the endometrium underwent hysteroscopic assessment and, in the overall cohort, 14 percent had benign endometrial pathology, 77 percent of which were benign endometrial polyps [37]. In another study of patients who switched to an aromatase inhibitor (AI) due to thickened endometrium found while on tamoxifen, endometrial thickness progressively decreased after 3, 6, and 12 months of AI administration. However, after 24 months of AI use, no further reduction in endometrial thickness was observed [38]. As there was no control group, it is unknown whether the endometrial thickness decrease was secondary to AI use or to stopping tamoxifen alone.

Prognosis — EC prognosis appears to be worse for patients treated with tamoxifen compared with no tamoxifen [15,33,40].

While early studies suggested that EC in the setting of tamoxifen was not histologically different from EC arising in patients not on tamoxifen [41,42], additional data suggest that tamoxifen-associated ECs are more likely to have poor prognostic features (eg, nonendometrioid histology, estrogen and progesterone receptor negative, p53 positive) [25,33,40,43,44]. Longer duration and higher cumulative doses of tamoxifen are also associated with worse prognosis.

In a retrospective study and follow-up data from 641 patients with breast cancer who subsequently developed uterine cancer, tamoxifen use for ≥2 years compared with no tamoxifen was associated with an increased risk of nonendometrioid histologies (clear cell and serous: 15.9 versus 8.7 percent; carcinosarcoma: 8.4 versus 3.7 percent) and stage III or IV disease (20 versus 11.3 percent) [33,40]. Tamoxifen use was also associated with decreased uterine cancer-specific survival at three years (82 versus 93 percent).

Uterine sarcoma and carcinosarcoma — Tamoxifen appears to be associated with an increased risk of uterine sarcoma and carcinosarcoma (previously termed malignant mixed Müllerian tumor [MMMT]) [10,33,45-48], leiomyosarcomas [46], adenosarcomas [49], and endometrial stromal sarcomas [50]. MMMT is no longer classified as a sarcoma but rather a histology of EC; however, studies performed before this reclassification often include carcinosarcoma cases in reported rates of uterine sarcoma.

While the absolute risk of such cancers is low [51,52], given the association between tamoxifen and the increased risk of uterine sarcoma, the US Food and Drug Administration issued a black box warning regarding the use of tamoxifen [53], and American College of Obstetricians and Gynecologists (ACOG) advises that patients taking tamoxifen be advised of the risk of uterine sarcoma, along with other risks [24].

Sarcomas typically present two to five years following the start of therapy and often present at an advanced stage [48,54]. There also may be a persistent risk of uterine sarcoma after completion of tamoxifen therapy, as sarcomas have been reported 4 to 20 years after tamoxifen has been completed [50,52].

Representative studies regarding tamoxifen and the risk of uterine sarcoma include:

●In the NSABP P-1 randomized trial discussed above, uterine sarcoma occurred more frequently in patients on tamoxifen therapy compared with placebo (17/100,000 person-years versus no cases) [10].

●In data derived from the Surveillance, Epidemiology, and End Results (SEER) database of the National Cancer Institute including over 39,000 patients with breast cancer treated with tamoxifen between 1980 and 2000, compared with the general SEER population, the relative risk of a subsequent uterine corpus cancer (observed to expected [O/E] ratio) was 2.2 and was higher for carcinosarcomas (referred to in this study as MMMTs) than for other endometrial adenocarcinomas (O/E 4.6 and 2.1, respectively) [55]. The relative risk of carcinosarcoma rose further in patients surviving five years, while the risk of adenocarcinoma did not change appreciably. However, because of the rarity of carcinosarcoma, this translated into a smaller excess absolute risk for carcinosarcoma than for uterine adenocarcinomas (an additional 1.4 versus 8.4 cancers per 10,000 patients per year, respectively). Prognosis was poor among patients diagnosed with carcinosarcoma, with 25 deaths among the 34 diagnosed patients.

●In the case-control study in the Netherlands comparing 299 patients with breast cancer who subsequently developed EC discussed above (see 'Duration, dose, and type of therapy' above), the risk of uterine sarcoma and carcinosarcoma increased with ≥2 years of tamoxifen therapy (15.4 percent on tamoxifen versus 2.9 percent in nonusers) [33].

The ATLAS trial did not separately report uterine sarcoma risk in the setting of 10 years of adjuvant tamoxifen therapy [31]. (See 'Duration, dose, and type of therapy' above.)

CLINICAL PRESENTATION — Patients on tamoxifen who develop uterine pathology may be asymptomatic or present with AUB.

AUB occurs in over 50 percent of premenopausal patients [8,56] and up to 25 percent of postmenopausal patients [13]. AUB may include amenorrhea (secondary to tamoxifen or chemotherapy), irregular menses, heavy menses, or intermenstrual bleeding; resumption of menses after a period of chemotherapy-induced amenorrhea may also occur [3,4,8,56].

Asymptomatic patients typically present after an incidental finding (eg, thickened endometrium, endometrial polyp) on pelvic imaging that was performed for another indication (eg, pelvic pain).

INITIAL EVALUATION — The unique aspects in evaluation of uterine pathology in patients on tamoxifen are presented here. The evaluation of AUB in the general population is presented separately. (See "Abnormal uterine bleeding in nonpregnant reproductive-age patients: Terminology, evaluation, and approach to diagnosis" and "Approach to the patient with postmenopausal uterine bleeding".)

Symptomatic patients — Evaluation of symptomatic patients includes a detailed medical history, determination of menopausal status, and pelvic ultrasound/endometrial sampling.

Medical history — The medical history should include:

●A menstrual history, including the bleeding pattern before and after breast cancer treatment.

Some patients may have amenorrhea or irregular menses following chemotherapy. Alkylating agents (eg, cyclophosphamide) are often used in breast cancer, and these are potent causes of ovarian failure. The likelihood of ovarian insufficiency following chemotherapy depends upon the agent, dose, and patient age (table 1).

●The quantity and quality of bleeding; patients should be asked about bloody discharge, spotting, staining, or leukorrhea [24].

●Risk factors, other than tamoxifen, for endometrial carcinoma (EC), uterine sarcoma, and uterine carcinosarcoma, which may include obesity, a long-term history of ovulatory dysfunction, increasing age, nulliparity, pelvic radiation, and some hereditary conditions. This is discussed in detail separately. (See "Endometrial carcinoma: Epidemiology, risk factors, and prevention", section on 'Risk factors' and "Clinical features, diagnosis, staging, and treatment of uterine carcinosarcoma", section on 'Epidemiology and risk factors'.)

Determining menopausal status — Because bleeding patterns and ultrasound findings differ for pre- and postmenopausal patients, defining menopausal status is important. Our clinical approach to determining menopausal status presented in the algorithm (algorithm 1); if there is uncertainty about menopausal status, the patient should be evaluated as postmenopausal.

Unlike the general population, menstrual history is not a reliable method of determining menopausal status in patients on tamoxifen [56]. (See "Clinical manifestations and diagnosis of menopause".)

In addition, approximately 20 percent of breast cancer cases in the United States are diagnosed between the ages of 45 to 54 years, and at these ages, many patients will be in the menopausal transition and have irregular menses (figure 1) [57]. The approach to determining menopausal status prior to initiating endocrine therapy may also differ somewhat from determining menopausal status in patients on tamoxifen with AUB because the threshold for clinical decision-making and the hormonal milieu are different. (See "Adjuvant endocrine therapy for premenopausal women with hormone receptor-positive breast cancer", section on 'Definition of menopause'.)

Additional laboratory testing for endocrine abnormalities or other etiologies of AUB should be performed as appropriate. (See "Abnormal uterine bleeding in nonpregnant reproductive-age patients: Terminology, evaluation, and approach to diagnosis".)

Pelvic ultrasound and endometrial sampling — Patients on tamoxifen therapy with AUB require evaluation for uterine pathology, typically with transvaginal ultrasound (TVUS) and endometrial biopsy. Saline infusion sonography (SIS), hysteroscopy, and dilation and curettage (D&C) may be used for selected patients. This is discussed below and in more detail separately. (See "Overview of the evaluation of the endometrium for malignant or premalignant disease", section on 'Methods of evaluation'.)

Evaluation differs for pre- and postmenopausal patients:

●Postmenopausal patients – Either endometrial biopsy or TVUS can be used as the first diagnostic step for evaluating the endometrium. The choice of procedure is often driven by the clinician's specialty, available resources, and patient-specific factors (eg, parity).

While endometrial biopsy and TVUS are both reasonable initial tests, both have limitations. For example, endometrial biopsy samples only a portion of the endometrium and is most reliable when at least one-half of the endometrium is affected by disease [58]. Thus, focal pathology can be missed with blind endometrial biopsy. Similarly, the diagnostic efficacy of TVUS is lower in selected patients (eg, patients with uterine fibroids, presence of serous and clear cell endometrial cancers) and there are no high-quality data supporting the use of TVUS in postmenopausal patients on tamoxifen. These issues are discussed in detail separately. (See "Overview of the evaluation of the endometrium for malignant or premalignant disease", section on 'Postmenopausal patients with bleeding' and "Overview of the evaluation of the endometrium for malignant or premalignant disease", section on 'Methods of evaluation'.)

Our approach to postmenopausal patients on tamoxifen with AUB (including red, pink, or brown spotting, staining, or discharge) depends on which test (ie, endometrial biopsy or TVUS) was initially performed, and is detailed in the following algorithms (algorithm 2 and algorithm 3).

●Premenopausal patients – The initial evaluation of premenopausal patients depends on the bleeding pattern, since AUB is defined broadly, and menstrual irregularities and amenorrhea are common among premenopausal patients on tamoxifen. (See 'Clinical presentation' above.)

•For patients on tamoxifen with new onset of amenorrhea or oligomenorrhea, menopausal status should be evaluated (see 'Determining menopausal status' above). If testing shows that the patient is perimenopausal, we do not perform an endometrial biopsy unless there are other risk factors for endometrial cancer (table 2).

•For patients with heavy menstrual bleeding, intermenstrual bleeding, or frequent irregular menses, we perform an office endometrial biopsy; we also perform TVUS if structural uterine pathology is suspected. TVUS measurement of endometrial thickness is not a reliable test for evaluating for endometrial hyperplasia or carcinoma in premenopausal patients since endometrial thickness varies with normal hormonal fluctuations in this population. (See "Overview of the evaluation of the endometrium for malignant or premalignant disease", section on 'Premenopausal patients with abnormal bleeding'.)

-If the biopsy is benign and there is no endometrial pathology on ultrasound, no further evaluation is needed.

-If the biopsy is benign and there is endometrial pathology on ultrasound, either SIS or office hysteroscopy is performed to distinguish global from focal findings.

Incidental findings on pelvic ultrasound — Tamoxifen may result in an unusual ultrasound appearance of the endometrium [59]. This may be reported as endometrial thickening (>4 mm) or cystic changes in the endometrium or myometrium. Originally, the ultrasound appearance associated with tamoxifen was postulated to be microcystic changes that represented glandular cystic atrophy in the basalis of the endometrium, proximal myometrium, or even within polyps [60]. Further evaluation of these changes with SIS typically reveals these changes to be associated with atrophic endometrium covering microcystic areas. Other selective estrogen receptor modulators (SERMs) are also associated with similar endometrial hydrodynamics. These changes make it virtually impossible to measure endometrial thickness without sonohysterography [61,62]. The presence of these findings alone in a patient without vaginal bleeding does not require further endometrial assessment.

A uterine mass consistent with a leiomyoma is also a common finding on pelvic imaging. However, it is often not possible to distinguish a benign leiomyoma from a sarcoma based on imaging alone, though features suggestive of sarcoma (eg, central necrosis, irregular tumor borders, color Doppler findings of irregular vessel distribution, low impedance to flow, and high peak systolic velocity) may be present. This is discussed in detail separately. (See "Uterine fibroids (leiomyomas): Differentiating fibroids from uterine sarcomas", section on 'Imaging'.)

SUBSEQUENT EVALUATION FOR RECURRENT OR PERSISTENT BLEEDING — All patients with bleeding symptoms that persist or recur require further evaluation; we perform a hysteroscopy with dilation and curettage (D&C) regardless of previous evaluations.

Evaluation of other causes of AUB (table 3) should also be performed, as needed. (See "Abnormal uterine bleeding in nonpregnant reproductive-age patients: Terminology, evaluation, and approach to diagnosis" and "Approach to the patient with postmenopausal uterine bleeding".)

MANAGEMENT

Patients with uterine pathology

Patients with polyps: Polypectomy — For pre- and postmenopausal patients on tamoxifen, endometrial polyps are removed, even if the patient is asymptomatic. While most endometrial polyps are benign, some polyps will undergo malignant transformation, and this risk is higher than for those not on tamoxifen. This is a different approach than for patients who are not on tamoxifen. (See "Endometrial polyps", section on 'Patients without bleeding or other indications for removal'.)

Hysteroscopic polypectomy is the treatment of choice for most patients with endometrial polyps; hysteroscopy allows for visualization of the polyp and confirmation of complete removal. Hysterectomy is not performed for benign polyps. (See "Endometrial polyps", section on 'Hysteroscopic polypectomy'.)

Patients with endometrial hyperplasia — Endometrial hyperplasia may progress to, or coexist with, endometrial carcinoma (EC). (See "Endometrial hyperplasia: Clinical features, diagnosis, and differential diagnosis", section on 'Natural history'.)

Postmenopausal patients — Postmenopausal patients on tamoxifen who are diagnosed with endometrial hyperplasia (with or without atypia) are typically treated with hysterectomy. While progestin therapy may be used as an alternative for selected patients not on tamoxifen, the role of progestins in patients on tamoxifen is less clear. (See "Endometrial hyperplasia: Management and prognosis".)

Premenopausal patients — Premenopausal patients on tamoxifen who are diagnosed with endometrial hyperplasia with atypia, particularly those who have completed childbearing, are typically treated with hysterectomy. (See "Endometrial hyperplasia: Management and prognosis", section on 'Preferred treatment: Hysterectomy'.)

The management of endometrial hyperplasia without atypia in premenopausal patients represents a more complex clinical scenario and depends on the reason for tamoxifen use and desire for future fertility.

●On tamoxifen for chemoprevention – For patients on tamoxifen for chemoprevention with endometrial hyperplasia without atypia, management options include insertion of the levonorgestrel-releasing intrauterine device (LNG IUD), of which the 52 mg LNG (LNG 52; Mirena, Liletta) is the most studied; discontinuation of tamoxifen; or hysterectomy. This decision is made in collaboration with the patient's medical oncologist. For those who are medically managed, careful follow-up is needed to assess for disease persistence, progression, or recurrence. This is discussed in detail separately. (See "Endometrial hyperplasia: Management and prognosis", section on 'Endometrial surveillance'.)

●On tamoxifen for adjuvant breast cancer therapy – For premenopausal patients on tamoxifen for adjuvant breast cancer therapy with endometrial hyperplasia, hysterectomy is typically performed, even in the setting of hyperplasia without atypia. However, expectant management (with close observation and serial endometrial biopsies) of endometrial hyperplasia without atypia may be a reasonable alternative in well-counseled patients who desire future fertility.

There is a lack of high-quality data regarding the safety of the LNG 52 in progesterone receptor-positive breast cancer and the LNG 52 labeling lists breast cancer as a contraindication for its use [63]. As such, utilization of the LNG IUD in patients receiving tamoxifen for adjuvant breast cancer therapy is not the standard of care and should be individualized based on a discussion with the patient and the breast oncologist (see "Approach to the patient following treatment for breast cancer", section on 'Contraception after breast cancer'). Data regarding the LNG 52 in patients on tamoxifen are limited but include the following:

•In meta-analyses including four randomized trials and 543 pre- and postmenopausal patients with breast cancer on tamoxifen, the LNG 52 reduced the risk of some endometrial pathologies but increased the rate of AUB symptoms [64]. Patients using the LNG 52 compared with controls had lower rates of endometrial polyps (odds ratio [OR] 0.22, 95% CI 0.13-0.39) and endometrial hyperplasia (OR 0.13, 95% CI 0.03-0.67) over 24 to 60 months of follow-up. However, the number of events of endometrial hyperplasia was low (six cases), and none of the trials were adequately powered to detect a difference in endometrial cancer rates. The rate of AUB symptoms was higher in patients using the LNG 52, although the odds compared with no IUD decreased from 12 to 24 months of follow-up (12 months: sevenfold; 24 months: threefold), and by 60 months of follow-up, no cases of AUB were reported in either group. The numbers of events of breast cancer recurrence (18 patients) and breast cancer-related deaths (16 patients) were similar between the IUD and control groups; however, there was insufficient power to detect a difference in these outcomes.

•In a retrospective study of patients with breast cancer, those with an LNG 52 in place at the time of breast cancer diagnosis and who continued its use (38 patients) compared with nonusers had a threefold increase in the risk of breast cancer recurrence (adjusted hazard ratio [HR], 3.39, 95% CI 1.01-11.35) [65]. There was no increase in recurrence risk in patients who had the LNG 52 placed after breast cancer diagnosis; however, the study lacked sufficient power to detect a difference. In a subsequent case-control study including almost 200 patients with breast cancer (ages 50 to 62 years), the risk of a new breast cancer diagnosis was 1.5 times higher in those who used the LNG IUD [66].

Patients with malignancy: Hysterectomy

●Endometrial carcinoma – Patients on tamoxifen with EC are typically managed with hysterectomy and bilateral salpingo-oophorectomy with appropriate staging and adjuvant therapy where indicated. This is the same approach to patients not on tamoxifen. (See "Overview of resectable endometrial carcinoma".)

While selected patients not on tamoxifen who desire fertility preservation may be candidates for progestin therapy (with deferral of surgical staging until after completion of childbearing), the safety of progestins in patients with progesterone receptor-positive breast cancer is not well established. This is discussed in detail above. (See 'Patients with endometrial hyperplasia' above.)

●Uterine sarcoma or carcinosarcoma – Patients with uterine sarcoma or carcinosarcomas are managed with hysterectomy and bilateral salpingo-oophorectomy with appropriate staging and adjuvant therapy. In our practice, we have a low threshold for surgical evaluation when a suspicious myometrial or endometrial mass is identified in a symptomatic patient on tamoxifen. (See "Treatment and prognosis of uterine leiomyosarcoma" and "Clinical features, diagnosis, staging, and treatment of uterine carcinosarcoma", section on 'Diagnosis' and "Uterine fibroids (leiomyomas): Differentiating fibroids from uterine sarcomas".)

Symptomatic patients without uterine pathology — For symptomatic patients in whom no pathology was identified during evaluation, management options are limited:

●Hysterectomy is a definitive treatment option for patients with recurrent or persistent bleeding on tamoxifen therapy. In postmenopausal patients, persistent or recurrent symptoms should raise concern for a missed diagnosis of uterine cancer. As an example, in a retrospective study of 286 patients (not on tamoxifen) with peri- or postmenopausal bleeding and an initial endometrial biopsy that was negative or had insufficient tissue, four patients (2 percent) developed uterine malignancy (two had stage I adenocarcinoma, one had stage IV adenocarcinoma, and one had stage I high-grade stromal sarcoma) within two years [67].

●The LNG 52 may be used for management of bleeding symptoms in selected patients (eg, premenopausal patients on tamoxifen for chemoprevention). For premenopausal patients with breast cancer, however, the use of the LNG 52 is more limited as it may increase the risk of breast cancer recurrence. (See 'Patients with uterine pathology' above and "Approach to the patient following treatment for breast cancer", section on 'Contraception after breast cancer'.)

●We suggest not performing endometrial ablation in patients on tamoxifen. Residual endometrium remains after endometrial ablation and prior endometrial ablation may interfere with subsequent evaluation of the endometrium. In addition, while endometrial ablation does not appear to increase the risk of EC [68,69], this has not been well studied in the context of tamoxifen use. (See "Overview of endometrial ablation", section on 'Relative contraindications'.)

As noted above, all patients with bleeding symptoms that persist or recur require further evaluation. (See 'Subsequent evaluation for recurrent or persistent bleeding' above.)

ROLE OF SCREENING — All patients who undergo treatment with tamoxifen should undergo routine gynecologic examinations and be counseled to notify their provider about any AUB (including pink or brown staining or spotting, bloody discharge, or any other change in color or volume of bleeding) [24].

The role of routine screening (with ultrasound and endometrial biopsy) during or prior to tamoxifen use, however, is limited:

●During tamoxifen use – The American College of Obstetricians and Gynecologists (ACOG) recommends against screening asymptomatic patients on tamoxifen for endometrial cancer (EC) [24]. This is similar to the general population, in whom screening of asymptomatic patients for EC is also not performed. (See "Endometrial carcinoma: Clinical features, diagnosis, prognosis, and screening", section on 'Screening'.)

The benefits of screening for EC in patients on tamoxifen have not been found to outweigh the risks and burdens [70-74]. In one cohort study of 304 asymptomatic patients with breast cancer receiving adjuvant tamoxifen therapy and undergoing yearly transvaginal ultrasound (TVUS) for six years, the sensitivity and specificity for EC (using an endometrial thickness of >9 mm) were low: 63 and 60 percent, respectively [70]. Similarly, in a prospective study of 159 asymptomatic patients on tamoxifen undergoing screening endometrial biopsies, no cases of EC were diagnosed during the five-year surveillance period; an average of 5.8 endometrial biopsies were performed per patient [74]. Screening led to an increase in operative procedures (14 patients underwent dilation and curettage [D&C], three patients underwent hysterectomy) despite benign final pathology. Other studies including asymptomatic premenopausal patients treated with tamoxifen have also reported no cases of EC detected with routine screening (ultrasound and/or biopsy) [16,56].

●Prior to tamoxifen use – Screening for uterine pathology prior to initiation of tamoxifen therapy in postmenopausal patients has been proposed, and this strategy may benefit from further study. Implementation would require collaboration between medical oncology and gynecology clinicians.

Some data suggest that as many as 17 percent of newly diagnosed postmenopausal patients with breast cancer have a benign endometrial polyp prior to initiating tamoxifen therapy [19,75]. In a prospective study of 510 postmenopausal patients with breast cancer, TVUS prior to initiation of tamoxifen identified uterine pathology (defined as an endometrial thickness >5 mm) in 16.7 percent of participants; all cases were removed at time of outpatient hysteroscopy and identified as endometrial polyps [19]. Patients with versus without pretreatment uterine pathology had a higher rate of atypical endometrial lesions during tamoxifen therapy (11.7 versus 0.7 percent person-years).

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Abnormal uterine bleeding" and "Society guideline links: Uterine cancer".)

SUMMARY AND RECOMMENDATIONS

●Risk of uterine pathology – Tamoxifen is associated with an increased risk of uterine pathology; benign endometrial polyps are the most common type of endometrial pathology. Postmenopausal patients on tamoxifen have an increased risk of endometrial hyperplasia and carcinoma. Premenopausal patients using tamoxifen do not appear to be at the same increased risk of endometrial carcinoma (EC) as postmenopausal patients. Tamoxifen also appears to be associated with an increased risk of uterine sarcomas and carcinosarcomas, but the absolute risk of such cancers is low. (See 'Risk of uterine pathology' above.)

●Clinical presentation – Uterine pathology in patients on tamoxifen therapy usually presents with abnormal uterine bleeding (AUB) or an incidental finding of uterine pathology on pelvic imaging. (See 'Clinical presentation' above.)

●Evaluation – In addition to the medical history and determination of menopausal status (algorithm 1), patients are initially evaluated as follows (see 'Initial evaluation' above):

•Postmenopausal patients – For postmenopausal patients with any bleeding, spotting, staining, or bloody vaginal discharge, evaluation of the endometrium for EC is required. Either office endometrial biopsy or transvaginal ultrasound (TVUS) can be used as the initial diagnostic test (algorithm 2 and algorithm 3). The choice of procedure is often driven by the clinician's specialty, available resources, and patient-specific factors (eg, parity) and is discussed in detail separately. Both tests have limitations which make either procedure a reasonable first test. (See 'Pelvic ultrasound and endometrial sampling' above and "Overview of the evaluation of the endometrium for malignant or premalignant disease".)

•Premenopausal patients – For premenopausal patients with heavy menstrual bleeding, intermenstrual bleeding, or frequent irregular menses, we perform an office endometrial biopsy; we also perform TVUS if structural uterine pathology is suspected. If the biopsy is benign and endometrial pathology is demonstrated on ultrasound, either office hysteroscopy or saline infusion sonography (SIS) is performed to distinguish global from focal findings. For patients with new onset of amenorrhea or oligomenorrhea, we do not perform an endometrial biopsy unless there are other risk factors for endometrial cancer (table 2). (See 'Pelvic ultrasound and endometrial sampling' above.)

•Patients with persistent or recurrent bleeding – All patients with bleeding symptoms that persist or recur require further evaluation; we perform hysteroscopy with dilation and curettage (D&C) regardless of previous evaluations. (See 'Pelvic ultrasound and endometrial sampling' above.)

●Management of patients with uterine pathology (see 'Patients with uterine pathology' above)

•Endometrial polyps – For pre- and postmenopausal patients on tamoxifen, endometrial polyps are removed, even if the patient is asymptomatic. While most endometrial polyps are benign, some polyps will undergo malignant transformation, and this risk is higher than those not on tamoxifen. This is discussed in detail separately. (See "Endometrial polyps", section on 'Patients without bleeding or other indications for removal'.)

•Endometrial carcinoma – Most patients with EC are managed with hysterectomy. This is discussed in detail separately. (See 'Patients with uterine pathology' above and "Endometrial carcinoma: Staging and surgical treatment".)

•Endometrial hyperplasia

-Postmenopausal patients – Postmenopausal patients on tamoxifen with endometrial hyperplasia are typically managed with hysterectomy. This is discussed in detail separately. (See 'Postmenopausal patients' above and "Endometrial hyperplasia: Management and prognosis".)

-Premenopausal patients – Premenopausal patients on tamoxifen with endometrial hyperplasia with atypia, particularly those who have completed childbearing, are typically managed with hysterectomy. (See 'Premenopausal patients' above and "Endometrial hyperplasia: Management and prognosis", section on 'Preferred treatment: Hysterectomy'.)

The management of endometrial hyperplasia without atypia in premenopausal patients on tamoxifen represents a more complex clinical scenario and depends on the reason for tamoxifen use and desire for future fertility.

For premenopausal patients with endometrial hyperplasia without atypia who are on tamoxifen for chemoprevention, options may include insertion of a levonorgestrel-releasing intrauterine device (LNG IUD), discontinuation of tamoxifen, or hysterectomy. These decisions are made in collaboration with the patient's oncologist. (See 'Premenopausal patients' above and "Endometrial hyperplasia: Management and prognosis".)

For most premenopausal patients with endometrial hyperplasia who are on tamoxifen for adjuvant breast cancer therapy, we suggest hysterectomy rather than the LNG IUD (Grade 2C). The safety of progestins in patients with progesterone receptor-positive breast cancer is uncertain and rates of breast cancer recurrence may be increased. (See 'Premenopausal patients' above.)

●Symptomatic patients without uterine pathology – For symptomatic patients in whom no pathology was identified during evaluation, management options are limited and may include expectant management, LNG IUD, or hysterectomy. Endometrial ablation is typically not performed in patients on tamoxifen. (See 'Symptomatic patients without uterine pathology' above and "Overview of endometrial ablation", section on 'Relative contraindications'.)

●Role of screening – We do not screen asymptomatic patients on tamoxifen for endometrial cancer. This is consistent with guidelines from the American College of Obstetricians and Gynecologists (ACOG) and is similar to the general population, in whom screening of asymptomatic patients is also not performed. (See 'Role of screening' above and "Endometrial carcinoma: Clinical features, diagnosis, prognosis, and screening", section on 'Screening'.)