Postmenopausal vaginal bleeding in women using hormone replacement therapy

Abstract

Objective

To estimate the risk of endometrial cancer in postmenopausal women presenting with vaginal bleeding using estrogen–progestogen hormone replacement therapy (HRT) regimens and to assess if the duration of HRT use has an effect on the risk of diagnosing endometrial cancer.

Study design

Cross-sectional study of consecutive women presenting with postmenopausal vaginal bleeding at a gynaecological oncology centre in the UK.

Main outcome measures

Endometrial cancer diagnosis.

Results

Over a 62-month period, 4847 women were investigated for postmenopausal vaginal bleeding. The majority of women (4097, 84.5%) did not use any HRT preparation at the time of initial referral and 750 (15.5%) women were using combined HRT preparations. A total of 298 (6.1%) women were diagnosed with endometrial carcinoma. Women using HRT preparations were significantly less likely to be diagnosed with endometrial cancer compared with women not using HRT (adjusted odds ratio = 0.229, 95% CI 0.116–0.452; P < 0.0001). The longer duration of HRT use did increase the risk of diagnosing endometrial cancer in women presenting with postmenopausal vaginal bleeding, but this was not statistically significant.

Conclusions

Postmenopausal women presenting with vaginal bleeding and using combined HRT preparations have significantly lower risk of being diagnosed with endometrial cancer when compared with women not using HRT.

Keywords

Postmenopausal bleeding hormone replacement therapy endometrial cancer

Introduction

According to national recommendations, any episodes of vaginal bleeding in postmenopausal women other than the monthly withdrawal bleeding with cyclical combined hormone replacement therapy (HRT) should be investigated in order to exclude malignancy.¹ The most common type of malignancy in this group of women is endometrial cancer; the reported risk in literature varies from 1% to 24%.^2–5

Type I endometrial carcinoma is estrogen-related and associated with the presence of risk factors such as obesity, diabetes, nulliparity and unopposed estrogen stimulation.⁶ The risk of developing endometrial carcinoma in women using estrogen-only HRT has been shown to vary between 2% and 15%.^7,8 Available data suggest that the risk of endometrial cancer in postmenopausal women using HRT is largely reduced by addition of progestogens in either continuous or cyclical regimens.^9–12

Up to 90% of women using cyclical combined HRT preparations will experience monthly withdrawal bleeding.¹³ In contrast, continuous administration of progestogens will lead to endometrial atrophy and amenorrhoea.¹⁴ However, this phase of amenorrhoea may be preceded by an initial period of irregular vaginal bleeding. This fact is often disregarded and women using continuous combined HRT preparations for less than six months are referred for investigation of postmenopausal vaginal bleeding.

Studies on assessing the performance of ultrasonography in predicting endometrial malignancy in women with postmenopausal vaginal bleeding often excluded individuals using HRT or when included in the analysis, the overall numbers were too small to draw any significant conclusions regarding the outcome of investigation.^15–17

The aim of this study is to estimate the risk of endometrial cancer in postmenopausal women presenting with vaginal bleeding using estrogen–progestogen HRT regimens. In addition, we aim to assess if the duration of HRT use has an effect on the risk of diagnosing endometrial cancer.

Methods

This study of consecutive postmenopausal women presenting with vaginal bleeding was conducted at a gynaecological oncology centre in the UK, between February 2006 and April 2011. Women were diagnosed postmenopausal after at least 12 months of amenorrhoea. Excluded from the study were premenopausal women, asymptomatic women with an incidental finding of increased endometrial thickness on imaging, asymptomatic women with abnormal endometrial cytology found on cervical smear and women with a history of hysterectomy.

All women underwent transvaginal ultrasound scanning as the initial investigation tool to evaluate the endometrium. We used greyscale ultrasonography to measure the double-wall endometrial thickness in an anteroposterior dimension, in the sagittal plane from one basalis layer to the other. In keeping with departmental guidelines, when the endometrial thickness measured less than 5 mm no further investigations were performed as evidence suggests a low probability of cancer below this threshold.

Women found to have an endometrial thickness equal to or greater than 5 mm had endometrial sampling performed using an endometrial Pipelle device (Pipelle de Cornier; Laboratoire CCD, Paris, France). Endometrial biopsy was also performed in cases where the endometrial thickness was not clearly visualized on transvaginal ultrasonography. Hysteroscopic evaluation of the endometrium with biopsy was performed if Pipelle biopsy was not possible or did not yield sufficient tissue for histological diagnosis. In spite of benign histology on Pipelle^® biopsy, hysteroscopy was also performed in cases where endometrial thickness was greater than 10 mm, due to possibility of an endometrial polyp.

Data regarding the following characteristics were recorded for all women: age at presentation, body mass index (BMI) calculated as weight (kg)/(height [m])², use of HRT, presence of hypertension and diabetes, previous history of breast cancer, use of tamoxifen at presentation, amount of blood lost and frequency of the episodes of vaginal bleeding. The above characteristics, endometrial thickness measurement and the histology results were collected and recorded prospectively in an electronic database.

Ethical approval for the use of the postmenopausal clinic database was granted by the National Research Ethics Service Committee South Central – Oxford C on 29 July 2011 (reference number: 11/SC/0285). The local Research and Development study number is 2011O&G06L (120-08-11).

The distributions of continuous variables were not symmetric. To test for normality, the Shapiro–Wilk W test was used, as was the q–q plot to investigate normality graphically (results not shown). There was no evidence to suggest that data were normally distributed, hence in the descriptive statistics for continuous variables, we report median and interquartile range. To test any differences we used a non-parametric Wilcoxon rank sum (Mann-Whitney) test. Chi-squared test was used after checking the expected assumptions. An adjusted logistic regression was carried out to investigate the odds of HRT controlling for clinical characteristics and cancer diagnosis. All analyses were done using STATA software, version 11.2 SE (Stata Corporation, College Station, TX, USA).

Results

The median age and BMI of women using HRT were significantly lower than the values for women not using HRT preparations (P < 0.0001). The differences on the above characteristics, although statistically significant, have no clinical relevance. In addition, a smaller percentage of women using HRT compared with women not using any HRT preparations were previously diagnosed with diabetes or hypertension as shown in Table 1.

Table 1

Clinical characteristics of the individuals in the study

Characteristics	HRT use	No HRT	P value
	N = 750	N = 4097
Age (years)	58 (54, 62)	59 (54, 68)	<0.0001*
BMI^†	27 (24, 30)	28 (25, 33)	<0.0001*
Diabetes	24 (3.28%, 2.1–4.7%)	260 (6.3%, 5.6–7.1%)	0.001^‡
Hypertension	154 (21%, 18–24%)	1109 (27%, 26–28%)	<0.0001^‡
Frequency of bleeding ^†
Single episode	422 (57%, 53–60%)	2022 (50%, 48–51%)	<0.0001^‡
Recurrent	323 (43%, 40–47%)	2062 (50%, 49–52%)
Amount of bleeding ^†
Spotting	110 (15%, 12–18%)	718 (18%, 17–19%)	0.117^‡
Light	471 (64%, 60–67%)	2495 (62%, 61–64%)
Heavy	159 (21%, 19–25%)	798 (20%, 19–21%)
Endometrial thickness	5.2 (3.5, 7.9)	4.6 (2.90, 8.8)	0.0024*

HRT, hormone replacement therapy

Values are median (interquartile range), number (percentage, 95% CI)

*Two-sample Wilcoxon rank sum test (Mann-Whitney test)

^†Percentages or medians worked on less numbers from the overall due to missing values

^‡Chi-squared test

A total of 298 (6.1%) women were diagnosed with endometrial carcinoma during the study period. Type I endometrial cancer accounted for 87.9% of all the cases. As shown in Table 2, women using HRT preparations were significantly less likely to be diagnosed with type I endometrial cancer (P < 0.0001). Significant difference was also observed in the incidence of type II endometrial cancer cases diagnosed in women using HRT compared with women not using HRT. However, only one case of type II endometrial cancer was diagnosed in women using HRT preparations, which makes it difficult to draw firm conclusions. The overall risk of diagnosing endometrial cancer in women using HRT preparations was significantly lower compared with women not using HRT.

Table 2

Diagnosis of endometrial cancer in the group of women using and those not using HRT preparations at the time of referral

Endometrial cancer	HRT use (N = 750)	No HRT (N = 4097)	P value
Type I	9 (1.2%, 0.6–2.3%)	253 (6.2%, 6.5–9.6%)	<0.0001*
Type II	1 (0.13%, 0.003–0.7%)	35 (0.9%, 0.6–1.2%)	0.035*

*Chi-squared test

On a multivariate analysis shown in Table 3, when adjusting for other clinical characteristics such as age, BMI, bleeding patterns, diabetes and hypertension, women using HRT were less likely to be diagnosed with endometrial cancer (odds ratio = 0.229, 95% CI 0.116–0.452; P < 0.0001).

Table 3

The effect of using HRT on the odds ratio for endometrial cancer adjusted for age, BMI, bleeding patterns, hypertension and diabetes

Variables	Odds ratio	P value	95% CI
HRT
No	1	1	1
Yes	0.229	<0.0001	(0.116, 0.452)
Age	1.052	<0.0001	(1.039, 1.065)
BMI	1.050	<0.0001	(1.034, 1.066)
Bleeding frequency
Single	1	1	1
Recurrent	5.310	<0.0001	(3.847, 7.332)
Hypertension
No	1	1	1
Yes	1.368	0.0220	(1.046, 1.789)
Diabetes
No	1	1	1
Yes	1.486	0.049	(1.001, 2.204)
Amount of bleeding
Spotting	1	1	1
Light	0.890	0.473	(0.647, 1.223)
Heavy	1.119	0.589	(0.743, 1.685)

The risk of diagnosing endometrial cancer (per 1000 women) in relation to duration of HRT use is graphically presented in Figure 1. Due to small number of endometrial cancer cases in women using HRT, three-year intervals were used. The risk of diagnosing endometrial cancer in women presenting with postmenopausal vaginal bleeding appears to increase with longer duration of HRT use, but this is not statistically significant (P = 0.420). The risk of diagnosing endometrial cancer was higher for women who were unable to provide information about the duration of HRT use. However, the risk was not significantly different even when compared with women using HRT for 1–3 years.

Figure 1

The risk of diagnosing endometrial cancer in relation to duration of HRT use (3-year intervals: 1–3 years, 4–6 years, >6 years and unknown duration). [#]=number of individuals with cancer; (#)=total number of women using HRT in each category

As shown in Table 4, the majority of women (49%) using HRT preparations were found to have normal endometrial histology. In 40% of women using HRT, the endometrial thickness measured less than 5 mm and the symptoms in this group of women were attributed to genital tract atrophy. The incidence of benign endometrial polyps in women using HRT was significantly lower than in women not using HRT (P < 0.0001).

Table 4

Outcome of investigation in women using and those not using HRT preparations at the time of referral

Outcome of investigations	HRT use	No HRT
	N (%, 95% CI)	N (%, 95% CI)	P value
Atrophy (ET <5 mm)	300 (40.0%, 36.4–43.6)	1920 (46.8%, 45.3–48.4)	=0.0001*
Normal histology	370 (49%, 46–53%)	1316 (32%, 31–34%)	<0.0001*
Endometrial polyps	51 (6.8%, 5.1–8.8%)	472 (12%, 11–13%)	<0.0001*
Endometrial hyperplasia	15 (2.0%, 1.1–3.3%)	76 (1.9%, 1.5–2.3%)	0.788*
Endometrial cancer	10 (1.3%, 0.6–2.4)	288 (7.0%, 6.3–7.8)	<0.0001*
Other	4 (0.5%, 0.1–1.3)	25 (0.6%, 0.4–0.9)	0.802

Overall P value <0.0001*

*Chi-squared test

Discussion

This study shows that the likelihood of diagnosing endometrial cancer in postmenopausal women when presenting abnormal vaginal bleeding and using combined HRT preparations is significantly lower when compared with women not using HRT. In this study, we did not distinguish between women using continuous combined and cyclical HRT preparations. However, the majority of women in the study used continuous combined preparations and the endometrial atrophy induced by the daily administration of progestogens in this group is likely to account for the lower incidence of cancer in the overall population of women using HRT. Similarly, a lower incidence of diagnosing endometrial polyps was observed in women using HRT. This is likely to be related to the antiproliferative effect of the progestogens in the endometrium.¹⁸

In our study, we used transvaginal ultrasonography to select the patients who require endometrial biopsy. Lin et al. ¹⁹ showed that the mean endometrial thickness measurement in asymptomatic postmenopausal women was 5.3 mm in the group using continuous combined HRT and 6.6 mm in the group using cyclical combined preparations. Similarly, Levine et al.²⁰ found that endometrial thickness measurement was greater in the group of postmenopausal women using cyclical compared with continuous HRT (8.3 versus 6.2 mm).Both these studies did not report any cases of endometrial cancer among women using combined HRT preparations.^19,20 In a study of 327 postmenopausal women (including 46 women with abnormal vaginal bleeding) using estrogen only or combined HRT preparations, Holbert et al.²¹ reported one case of endometrial cancer in a patient with abnormal vaginal bleeding using estrogen-only HRT.

Langer et al.²² reported a 99% negative predictive value of transvaginal ultrasonography and use of an endometrial thickness threshold of 5 mm for detecting serious endometrial pathology in asymptomatic postmenopausal women using estrogen-only or combined HRT preparations. However, no cases of endometrial cancer were found in women with endometrial thickness measurement of less than 5 mm, including cases in women using estrogen-only preparations. Langer et al. reported no cases of endometrial cancer in women using combined HRT preparations.

Our study did not distinguish between women using continuous and cyclical HRT preparations. In addition, we investigated all women in the same way, using an endometrial thickness of 5 mm on transvaginal ultrasonography as a threshold to perform Pipelle biopsy. This practice would be in line with the results of the studies mentioned previously.^19–22 However, the majority of the postmenopausal women included in these studies were asymptomatic and endometrial thickness thresholds for women with vaginal bleeding using HRT are not well known.

It is common practice for the primary care practitioners in the UK to stop HRT in women presenting with postmenopausal bleeding prior to referring them to secondary care for further investigation. Although there is no evidence to support this practice, it leads to increased anxiety in women and is likely to enhance the negative opinion about HRT among patients. Based on the findings of our study, we suggest that it is not necessary to advise women to stop combined HRT preparations prior to investigation of abnormal vaginal bleeding. If transvaginal ultrasonography is used as the initial tool to investigate women with postmenopausal vaginal bleeding, it is likely that women using cyclical combined preparations will require further evaluation by endometrial biopsy due to variation in endometrial thickness during the hormonal cycle.

In conclusion, we report significantly lower incidence of endometrial cancer in symptomatic postmenopausal women using combined HRT preparations when compared with women not using HRT. Despite the inherent limitations with observational research and the bias generated by confounding factors, the results of this large study can be used to guide clinicians when investigating women with postmenopausal vaginal bleeding. Women using combined HRT preparations can be investigated on a less urgent basis depending on the available resources.

Competing interests

None declared.

References

National Institute for Health and Clinical Excellence. Referral Guidelines for Suspected Cancer. NICE Clinical Guidelines 27, 2005

Alberico

, Conoscenti

, Veglio

, Bogatti

, Di Bonito

, Mandruzzato

. A clinical and epidemiological study of 245 postmenopausal metrorrhagia patients. Clin Exp Obstet Gynecol 1989;16:113–21

Choo

, Mak

, Hsu

, Wong

, Ma

. Postmenopausal uterine bleeding of nonorganic cause. Obstet Gynecol 1985;66:225–8

Gredmark

, Kvint

, Havel

, Mattsson

. Histopathological findings in women with postmenopausal bleeding. Br J Obstet Gynaecol 1995;102:133–6

Lidor

, Ismajovich

, Confino

, David

. Histopathological findings in 226 women with post-menopausal uterine bleeding. Acta Obstet Gynecol Scand 1986;65:41–3

Bokhman

. Two pathogenetic types of endometrial carcinoma. Gynecol Oncol 1983;15:10–7

Henderson

. The cancer question: an overview of recent epidemiologic and retrospective data. Am J Obstet Gynecol 1989;161:1859–64

Persson

, Adami

, Bergkvist

, Risk of endometrial cancer after treatment with oestrogens alone or in conjunction with progestogens: results of a prospective study. BMJ 1989;298:147–51

Lacey

Jr , Leitzmann

, Chang

, Endometrial cancer and menopausal hormone therapy in the National Institutes of Health-AARP Diet and Health Study cohort. Cancer 2007;109:1303–11

10.

Judd

, Mebane-Sims

, Legault

, Effects of hormone replacement therapy on endometrial histology in postmenopausal women. The Postmenopausal Estrogen/Progestin Interventions (PEPI) Trial. The Writing Group for the PEPI Trial. JAMA 1996;275:370–75

11.

Anderson

, Judd

, Kaunitz

, Effects of estrogen plus progestin on gynecologic cancers and associated diagnostic procedures: the Women's Health Initiative randomized trial. JAMA 2003;290:1739–48

12.

Rossouw

, Anderson

, Prentice

, Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results from the Women's Health Initiative randomized controlled trial. JAMA 2002;288:321–33

13.

Archer

, Pickar

, Bottiglioni

. Bleeding patterns in postmenopausal women taking continuous combined or sequential regimens of conjugated estrogens with medroxyprogesterone acetate. Menopause Study Group. Obstet Gynecol 1994;83:686–92

14.

Udoff

, Langenberg

, Adashi

. Combined continuous hormone replacement therapy: a critical review. Obstet Gynecol 1995;86:306–16

15.

van Doorn

, Dijkhuizen

, Kruitwagen

, Heintz

, Kooi

, Mol

. Accuracy of transvaginal ultrasonography in diabetic or obese women with postmenopausal bleeding. Obstet Gynecol 2004;104:571–78

16.

Ferrazzi

, Torri

, Trio

, Zannoni

, Filiberto

, Dordoni

. Sonographic endometrial thickness: a useful test to predict atrophy in patients with postmenopausal bleeding. An Italian multicenter study. Ultrasound Obstet Gynecol 1996;7:315–21

17.

Tabor

, Watt

, Wald

. Endometrial thickness as a test for endometrial cancer in women with postmenopausal vaginal bleeding. Obstet Gynecol 2002;99:663–70

18.

Chan

, Tam

, Yeo

, A randomised controlled trial of prophylactic levonorgestrel intrauterine system in tamoxifen-treated women. Br J Obstet Gynaecol 2007;114:1510–15

19.

Lin

, Gosink

, Wolf

, Endometrial thickness after menopause: effect of hormone replacement. Radiology 1991;180:427–32

20.

Levine

, Gosink

, Johnson

. Change in endometrial thickness in postmenopausal women undergoing hormone replacement therapy. Radiology 1995;197:603–8

21.

Holbert

. Transvaginal ultrasonographic measurement of endometrial thickness in postmenopausal women receiving estrogen replacement therapy. Am J Obstet Gynecol 1997;176:1334–8

22.

Langer

, Pierce

, O'Hanlan

, Transvaginal ultrasonography compared with endometrial biopsy for the detection of endometrial disease. Postmenopausal Estrogen/Progestin Interventions Trial. N Engl J Med 1997;337:1792–98