Genitourinary May C. Lin, BA Cynthia

Barbara

#{149}

A. Stuenkel,

MD

B. Gosink, Patricia

#{149}

MD

Susan MD

I. Wolf,

#{149}

S. Braly,

MD,

Dolores

#{149}

Endometrial Thickness after Effect ofllormone Replacement’ Ultrasound (US) images of the pelvis were evaluated in 112 asymptomatic postmenopausal women to investigate the normal range of endometnial thickness (double-layer measurement) and the effect of hormone replacement on these measurements. Twenty-one patients (19%) had endometrial thickness greater than 0.8 cm. One patient, with an endometrial thickness of 2.5 cm, had endometrial

carcinoma.

Consideration

of the

known actions of estrogen and progestogen on the endometrium led the authors to believe that the clinical significance of an endometrium measuring more than 0.8 cm depends on the patient’s hormonal status. Among asymptomatic postmenopausal women with an endometrial thickness between 0.8 and 1.5 cm, those receiving unopposed estrogen or continuous estrogen and progestogen need to undergo dilatation and curettage (D&C) or biopsy and those re-

sequential

no hormones estrogen

or receiving and progestogen

should

be encouraged

to undergo

ceiving

D&C or biopsy. If a patient in one of the latter two groups is unwilling to undergo an invasive procedure, then US examination at 3-month intervals is acceptable. Any patient with endometrial thickness of at least 1.5 cm should undergo histologic diagnosis, regardless of symptoms or hormone status.

U

PhD

Marc R. Feldesman, MD

#{149}

H. Pretorius,

(US) scanning has been used in a number of studies to identify changes in endometnial thickness and texture during the normal menstrual cycle (1-5). Sonographic staging of the endometrium the

different

phases

of

the menstrual cycle has proved to be accurate when correlated with sameday biopsy specimens (6). In addition, US has been proposed as a useful screening method to evaluate infertility by identifying endometria that are inadequately responsive to hormonal stimulation (7). Although the sonographic appearance of the endometrium has been well studied and characterized in premenopausal women, there are no accepted criteria in the imaging literature for normal endometrial thickness in asymptomatic postmenopausal women. In particular, we know of no systematic attempt to examine endometrial thickness in women following various common hormone regimens. When prescribing hormone replacement for postmenopausal women, physicians are often faced with the problem of balancing estrogen and progestogen dosages to alleviate the distressing symptoms of menopause while avoiding the development of endometrial hyperplasia. Although conjugated estrogens are effective in suppressing vasomotor

I From the Departments of Radiology, Ultrasound Division (H-759) (M.C.L., B.B.G., D.H.P.) and Reproductive Medicine (C.A.S., P.S.B.), University of California, San Diego, Medical Center,

225 Dickinson St, San Diego, CA 92103; Department of Radiology, Kaiser Sunnyside Medical

Center, Clackamas, Ore (S.I.W.); ment of Anthropology, Portland Index

terms:

Hormones

#{149} Uterine

neoplasms,

diagnosis, 854.32 #{149} Uterine neoplasms, US studies, 854.32, 854.1298 #{149} Uterus, endometrium, 854.92 #{149} Uterus, US studies, 854.1298 Radiology

1991;

180:427-432

and DepartState University, Portland, Ore (M.R.F.). From the 1990 RSNA scientific assembly. Received December

1990; revision sion received

PhD

Menopause:

LTRASOUND

throughout

Radiology

requested January 3, 1991; reviMarch 19; accepted March 25. in part by a grant from the San Di-

Supported ego Radiology Research and Education Foundation. Address reprint requests to B.B.G. C RSNA, 1991

3,

symptoms (hot flashes) and promoting bone-mineral conservation, about 24% of women experience irregular breakthrough bleeding for the initial 16 months (8). In addition, it is now well established that unopposed exogenous estrogen increases a wornan’s risk for endometrial hyperplasia and carcinoma and that this effect is both dose- and duration-dependent (9). Progestogens have been found to effectively reduce estrogen-induced breakthrough bleeding and lower the risk of endometrial hyperplasia and carcinoma (9). These effects are now commonly accepted indications for the use of progestogens, and they are often prescribed along with estrogens for women who have not undergone hysterectomy. Progestogens may have unwelcome side effects of their own, however. Eighty percent to 90% of women experience regular monthly withdrawal bleeding, and many have symptoms of breast tenderness, bloating, edema, abdominal cramping, anxiety, irritability, and depression (8). Progestogens have also been found to adversely alter blood lipid composition, increasing the risk for arterial thromboembolic disease (9). Because of these undesired side effects, some women choose to discontinue the use of progestogens and are advised by their physicians to then discontinue unopposed estrogen as well. This discontinuation leaves them at risk for the problems of osteoporosis and myocardial infarction attendant on postmenopausal loss of estrogen. In those women who continue using exogenous hormones, the delicate task of balancing appropriate dosages of estrogen and progestogen may lead to breakthrough bleeding. Because vaginal bleeding may be a symptom of endornetrial hyperplasia

Abbreviation:

D&C

=

dilatation

and curettage.

427

or carcinoma, these women are then subjected to either endometrial biopsy or dilatation and curettage (D&C) for further evaluation. Some women require repeated invasive procedures before an appropriate hormonal balance is reached. It is therefore desirable to correlate hormone replacement regimens with the sonographic appearance of the endometrium.

Establishment

-

of the

normal range of endometrial widths for each hormone regimen would help limit unnecessary invasive procedures for suspected endometnial hyperplasia. It is the purpose of this study to investigate, by means of US, the normal thickness of the postmenopausal endometnum in asymptomatic women and to determine the effect, if any, of hormone replacement on this measurement.

.

-

AND

b.

Figure 1. Transabdominal Digital calipers were used considered normal in this

Table

One

from

were

the study.

These

part of a larger

atic postmenopausal in an

ongoing

of asymptom-

women

participating prevalence and

natural history of simple adnexal cysts. Women were included if they met the following criteria: (a) 50 years of age or older at the time of first examination and (b) clearly postmenopausal for at least 1 year, on the basis of (1) time elapsed since last menstrual since

first

levels

period,

documentation

of follicle

luteinizing

since

time

(ii)

of Patients

stimulating

the onset

symptoms

was

about

obtained

from

(a) time

dominal

underwent

and

Acoustic Acuson

scans

Imaging (Mountain

(Mitsubishi

International,

with

Radiology

#{149}

of onset

of

was dereplacehormone

(g) parity, or signifi-

pelvic

were

transabUS.

obtained

with

Technologies (Phoenix), View, Calif), or Hitachi

equipment 428

both

transvaginal

Transabdominal

estrogen

58 (52) 10 (9) 9(8)

and progestogen and

112)

=

progestogen

35(31)

are in parentheses.

Thickness

Hormone Group

(cm) by Hormone

Regimen

Mean

Thickness Standard Deviation

Median

Range

1

0.52

0.4

0.45

0.01-2.5

2 3 4

0.68 0.53 0.66

0.7 0.6 0.6

0.53 0.17 0.39

0.01-1.4 0.2-0.7 0.01-1.7

measurement

of 0.01 indicates

that

the endometrium

either

transducer. Transvaginal scans were obtamed with Acoustic Imaging 5.0-MHz or Hitachi 6.5-MHz transducers.

The anteroposterior dometnum

conditions.

The women

Sequential

Note-A

menopausal

menopause, (b) how menopause termined, (c) type of hormone ment regimen, (d) duration of regimen, (e) height, (J) weight, and (h) any other medications medical

4

(n

was not well visualized.

flashes). (biii) was applied

each woman after the nature of the procedures had been fully explained. Each woman completed a questionnaire provid-

cant

No hormones Unopposed estrogen Continuous estrogen

No. of Patients

hot

consent

ing information

(b) views. of 3 mm was hormones.

Regimen

Endometrial

and

The last criterion to two women whose menopausal symptoms were so severe that either a gynecologist or an endocrinologist had prescribed hormone replacement, including a progestogen, before the complete cessation of natural menses. Informed

in coronal (a) and longitudinal thickness. A measurement who was not using exogenous

elapsed

time

of convincing

(usually

by Hormone

1 2 3

Table 2 Endometrial

elapsed

hormone

or (iii)

endometrial woman

Regimen

Note-Percentages

of menopausal

hormone,

the

women

group of the

study

of the uterus

additional

woman, in whom satisfactory US views of the uterus could not be obtained, was excluded

to measure postmenopausal

Group

METHODS

hysterectomy.

US images

1

The study population was composed of 1 12 unselected, self-referred, asymptomatic postmenopausal women who had not undergone

-

a.

Distribution

PATIENTS

.

-.

Philadelphia)

a 2.25- or 3.5-MHz

was

diameter

measured

with

of the endigital

cali-

pers in the sagittal plane on the scanner display screen. In some instances, these measurements were not optimal and measurements performed on the hard copy with hand-held calipers were preferred. Although there was generally close correlalion between measurements obtained on the transabdominal and transvaginal scans,

when

discrepancies

did occur,

the largest-

diameter measurement was chosen. The distance between the boundaries separating the hyperechoic endometrium from the adjacent inner layer of myometrium was measured (Fig 1). This sonographic measurement consists of two closely apposed layers of endometrium and has been shown

to be

in excellent

agreement

with

measurements of gross specimens (10). The hypoechoic halo commonly seen surrounding the echogenic endometrium of the premenopausal woman was observed in only

a few

women

in

this

study.

It has

been previously suggested that this hypoechoic area correlates histologically with vascular layer

structures of the

in the

the endometrium

compact than

with

(10). Its inclusion

measurements

lead

would

to overestimation

endometrial

of true

thickness.

In several within

rather

itself

in endometrial

therefore

inner

myometrium

the

women,

fluid

endometrial

collections

cavity

were

noted.

In these cases, the anteroposterior diameter of the fluid collection was subtracted from the endometrial measurement. The fluid volume was calculated by using the formula for a prolate ellipsoid (1,4 length

Each

x width

x height)

woman

was

(11).

scheduled

for folAugust

1991

Figure 2. Abnormally thick endometrium in coronal (a) and US images of the uterus. The endometrial thickness measured quential estrogen and progesterone. A small amount of fluid dometrial canal. D&C performed on the 3rd day of progesterone early secretory changes in the endometrium.

Table

3 of Patients Regimen

Distribution

Hormone

Hormone

with

Group

Total(n=112) *N),

Endometrial

Thickness

No.

with

1.1-1.5

1(n=58) 2(n=10) 3(n=9) 4(n=35)

in parentheses

of Patients

cm

1

0 0 1

0 0 0

14

1

1

of total number

writing,

At the

time

of this

two-thirds of the women had at least one follow-up exami-

On the basis of a review

of literature

at the onset of this study, of 1.0 cm was arbitrarily

as the

upper

in whom

limit

than

1.0 cm,

with

endometrial

a meachosen

In all women

the endometnum

greater

ommended. Women

of normal.

biopsy

Total 7(12)* 3(30)

0 6(17) 16(14)

in each group.

made with

patient

had

been

mone

regimen.

measurement regardless

to correlate endomelength of time that

following

a specific

The thickest

endometrial

identified

in each

woman,

of examination

number

and

approach,

was

a

hor-

used

for this

or D&C was recRESULTS

were

grouped

and progestogen

(typically

conjugated

estrogens, 0.625 mg, and medroxyprogesterone-acetate, 2.5-5 mg daily), and 4 = sequential estrogen and progestogen (typconjugated

estrogens,

0.625-1.25

mg

for the first 25 days of each month, and medroxyprogesterone-acetate, 5-10 mg for the last 10-13 days of estrogen administration).

For purposes

Volume

cm

evaluation

into four hormonal classifications: I = no hormones, 2 = unopposed estrogen (conjugated estrogens administered orally or transdermally), 3 = continuous combined estrogen

women

> 1.0cm

mone regimen for at least 3 months at the time of the examination were included in the corresponding hormone group. No

imaging study.

measured

further

of patients

attempt was trial thickness

nation.

available surement

2.1-2.5

0

are percentages

1.0 cm by

Thickness

3 0 5

tional

year.

Endometrial

than

6

3 months for an addi-

more than undergone

Greater

cm

1.6-2.0

low-up US examinations every for 1 year and every 6 months

ically

longitudinal (b) transabdominal 1.1 cm in this woman using Seis also present within the enadministration revealed

who

180

of statistical had

been

Number

#{149}

analysis,

following 2

only

a hor-

Group 1 consisted of 58 women (52%); group 2, 10 (9%); group 3, nine (8%); and group 4, 35 (31%) (Table 1). The mean, median, standard deviation, and range of endometnial thickness for each hormone regimen are shown in Table 2. Statistical analysis showed no significant difference (P > .05) in mean endometrial thickness among any groups. There was also no statistically significant correlation (P > .05) between endometrial thickness and age, length of time since onset of menopause, height, weight, or parity.

The endometrium was not well visualized on some examinations due to uterine retroversion, myomas, or technical factors. Of a total of 260 exarninations obtained in 112 women, 52 (20%) resulted in a poorly visualized endometrium by means of either transabdominal or transvaginal scanning. Of these, 26 examinations were obtained with transabdominal scanning and eight with transvaginal scanning. In only 18 examinations (7%) was the endometnium not well visualized with both techniques. Only one woman had a poorly visualized endometnium on all three of her examinations. With the exception of those women following a sequential hormone regimen, variation among measurements at different examinations for individual women was less than 3 mm. The variation between the measurements obtained by means of transvaginal and those obtained by means of transabdominal scanning during the same examination was generally no more than 3 mm. Endometnal thickness greater than 1.0 cm (Fig 2) was seen in 16 women (14%) (Table 3): seven women (12%) in group 1, three (30%) in group 2, none in group 3, and six (17%) in group 4. Because some referring physicians were reluctant to perform an invasive procedure in women who were asymptomatic, only eight of these 16 women underwent endometrial biopsy or D&C as a result of the US findings. Six of these eight women had an endometrial thickness between 1.1 and 1.5 cm: Three specimens consisted of insufficient tissue for diagnosis (two in group 1, one in group 2), two revealed proliferative endometria (both in group 4), and one consisted of early secretory endometnum (group 4). The other two women who underwent biopsy had endometrial thicknesses greater than 1.5 cm. thickness) scribed

One

of these had biopsy as “disordered

(group 4, 1.7-cm findings deproliferative,”

considered by the pathologist to be “mildly abnormal but not premalignant.” The other (group 1, 2.5-cm thickness) proved to have endometrial carcinoma (Fig 3). There were no cases of histologically confirmed endometrial hyperplasia. Of the 17 women with a coexisting ovarian cyst, in only one did the cyst exceed 3 cm in diameter. In only six of these women was the endometrium thicker than 1.0 cm. None of these women underwent surgery for the cysts,

lowed

and

all continued

up with

to be fol-

US examination. Radiology

The 429

#{149}

c’is L

LONG -

________

IG ML

--

SAG

ML.

UT

Al’

-

25

Figure image

3.

Transabdominal

of the

uterus.

a thickness

longitudinal

The

US

endometrium,

of 2.5 cm, was grossly

with

abnormal.

Endometrial biopsy in this patient revealed endometrial carcinoma. This patient was using exogenous hormones and had no

symptoms

prior

not

a.

b.

Figure

4.

Transabdominal

24 mL failure

of endometrial of her normal

She subsequently

(a) and

fluid was withdrawal

underwent

transvaginal

(b) US images

seen in this patient. bleeding associated

surgical

dilatation

of the uterus.

One week with her

for cervical

earlier, sequential

stenosis.

she

Approximately

had experienced hormone regimen.

Bloody

fluid

was drained.

to diagnosis. 30

functional cysts

status,

remains

if any,

of these

small

undetermined.

25

Endometrial fluid was found in 10 women: seven not receiving any hormone replacement and three following either continuous or sequential estrogen and progestogen regimens. Presence of fluid was unrelated to endometrial

thickness.

The

volume

of

fluid in eight of the 10 women was no more than 1 mL. One woman (24-mL volume of fluid) proved to have cervical stenosis that was attributed to previous

obstetrical

trauma

(Fig

ci)

0

15

0 I.. .0

E z1o

4). An-

other woman ( < 1-mL volume of fluid) underwent a biopsy, which revealed benign endometrial cells and fluid. The remaining women are being followed up with US, and no change in the volume of endometrial fluid has thus far been identified.

There is considerable disagreement in the literature about the normal endometrial thickness after menopause (Table 4). Six US studies have reported upper normal limits of endometrial thickness (double layer) of 5-10 mm in postmenopausal women (2,10,12-15). Two small magnetic resonance (MR) studies reported endometrial thickness of 3 mm or less in postmenopausal women receiving no hormone replacement and of 4-6 mm in women receiving exogenous hormones (regimen unspecified) (16,17). (Endometrial

measurements

MR imaging

be consistently

ments 430

obtained Radiology

#{149}

have smaller

with

0.2

0.3

0.4

obtained

been

shown

than

measure-

to

US [18]). No corre-

0.5

0.6

0.7

0.8

0.9

1

1.1

1.2

1.3

Endometrial Figure 5. Distribution used to indicate that

DISCUSSION

with

0.01

1.4

1.5

1.6

Thickness

1.7

1.8

1.9

2

2.1

2.2

2.3

2.4

2.5

(cm)

of patients by endometrial thickness. A measurement the endometrium was not well visualized.

of 0.01 cm was

lation with specific hormone regimens was attempted in any of these studies. Nearly all of these studies dealt with small numbers of women, some consisting of both pre- and postmenopausal

their hormonal status or insufficient tissue for diagnosis. Although Goldstein et al (15) considered proliferative endometrium to be abnormal, most investi-

women

sult

and

symptoms

logic

some or

of women

known

with

abnormal

histo-

features.

Our

study

revealed

112 asymptomatic

dometrial In those 1.0 and

thickness

that

women

greater

16 (14%) had

than

of

an en-

1.0 cm.

with a thickness greater than no more than 1.5 cm, biopsy or

D&C showed only proliferative or secretory endometrium consistent with

gators

believe

that

it is the

expected

re-

of unopposed estrogen replacement and sequential hormone replacement the estrogenic phase (9). Although at the beginning of the study,

limit has

we chose of normal, indicated

choice.

Osmers

endometrial tomatic

1.0 cm as the upper

one recent investigator that 0.8 cm is a better et al (14)

thickness women

in

receiving

studied

in 283 no

the

asympex-

August

1991

Table

4 of Current

Summary

on Postnsenopausal

Literature

Width

Endometrial

Thickness

Endometrial

No. of Authors/Date

Modality*

et aV1986 (16) et al/1986 (10) et at11987 (17)

Demas Fleischer Hricak

Patients

MR US

Nasrietat/1989(12)

US

90

Malpam et aI/1990 (13) Goldstein et ai/1990 (15) Osmers et al/1990 (14)

US US US

11 30 386

Note.-NA *MR

=

=

not applicable, resonance

of Patients

Hormone

Group

with

0.9-1.0

1 (n = 58) 2(n=10) 3(n=9) 4(n=35)

with Endometrial

1.1-13

cm

of total

All 45 underwent

hyperplasia

en-

or “other

of their an endo0.8 cm had seriously abnormal results of D&C. In their publication, the authors did not indicate the actual measurements of the in the

15 abnormal

cases.

From personal correspondence first author, however, we have that five asymptomatic women proved endometrial carcinoma endometrial

two

thickness

had

with the learned with had an

of 8-10

an endometrial

mm

and

thickness

of

10-15mm.

Figure 5 shows that 81% (n = 91) of women in our study had an endometrial

thickness

86% ness

(n

no

more

than

0.8

cm,

and

96) had an endometrial thickno more than 1.0 cm. It is striking

receiving 2), three

endometnia

greater

(30%)

measured

1.0 cm, whereas

sured

more

than

five (50%)

0.8 cm (Table

versely,

in the group

of nine

receiving

continuous

combined

mea5). Con-

women estro-

gen and progestogen (group 3), none had an endometrium measuring greater than 0.8 cm. The numbers of women in groups

Volume

No

=

103)

No

cm

by Hormone

Thickness cm

>0.8cm

2.1-2.5

0 0 0 1

Total

2 and

180

3 are

small,

Number

#{149}

Abnormal

3 4-6 3 4-6 10 ND ND >5 >10