CLIMACTERIC 2011;14:622–632

The efficacy of levonorgestrel intrauterine

systems for endometrial protection:
a systematic review
Y.-L. Wan and C. Holland

Academic Unit of Obstetrics & Gynaecology, University of Manchester School of Cancer, St Mary’s Hospital, Manchester, UK

Key words: ENDOMETRIAL NEOPLASM, INTRAUTERINE DEVICE, MIRENA, INTRAUTERINE SYSTEM, CHEMOPREVENTION,
LEVONORGESTREL, HYPERPLASIA, ENDOMETRIAL POLYPS
Climacteric Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14

ABSTRACT
Background Oral progestogens are commonly used for endometrial protection in women at higher risk of
developing endometrial abnormality. Long-term intrauterine progestogens may offer an attractive alternative
to oral therapy.
Objective To review the evidence regarding the efficacy of intrauterine levonorgestrel-releasing systems
(LNG-IUS) in preventing endometrial pathology in high-risk women.
Method Searches were made of the Cochrane Central Register of Controlled Trials, UK National Research
Register (NRR) Archive, Current Controlled Trials, MEDLINE, EMBASE and CINAHL. The selection cri-
For personal use only.

teria were randomized, controlled trials (RCTs) comparing LNG-IUS with no treatment, placebo or other
hormonal therapy in adult females. Where no RCTs were available, prospective cohort studies were analyzed.
Data was extracted using a standardized data collection form. Meta-analysis was performed using RevMan
software.
Results There were six RCTs that investigated LNG-IUS in women using estrogen replacement therapy
(ERT). LNG-IUS was at least as effective as other routes of progestogen administration. Only two studies
investigated LNG-IUS as treatment for endometrial hyperplasia. Hyperplasia without atypia regressed in all
women treated with LNG-IUS. In three studies of LNG-IUS in tamoxifen users, LNG-IUS was associated
with reduced risk of endometrial polyps (Peto odds ratio (OR) 0.28; 95% confidence interval (CI) 0.15–0.55)
and hyperplasia (Peto OR 0.14; 95% CI 0.02–0.80).
Conclusions LNG-IUS counters endometrial proliferation and causes regression of and prevents endometrial
hyperplasia in selected groups of women. There is, however, insufficient evidence to recommend LNG-IUS
as the treatment of choice for hyperplasia and no evidence to adequately support its use as chemoprevention
in women with hereditary non-polyposis colorectal cancer syndrome or obesity.

INTRODUCTION including a longer premenopausal exposure to estrogens. This

Endometrial cancer is the most common gynecological cancer
in affluent developed countries and accounts for one in every
18 new female cancers1. The incidence of endometrial cancer
has been increasing in the last few decades in postmenopausal
women and approximately 288 000 new cancers of the uter-
ine corpus are diagnosed each year2.
A hyperestrogenic milieu appears to be the strongest risk
factor for type I (hormone-dependent) endometrial cancer,
is suggested by the association of an early menarche and late
menopause with increased risk of endometrial cancer3,4. In
further support of this hypothesis, exogenous hormones also
affect lifetime risk5. Other exogenous steroid compounds such
as tamoxifen exert estrogenic effects on the endometrium,
conferring a two- to three-fold increase in risk of developing
endometrial cancer6. Furthermore, diet and lifestyle also have
a significant impact on risk. Women with a body mass index
(BMI) greater than 30 kg/m2 have three times the risk of

Correspondence: Dr C. Holland, Academic Unit of Obstetrics & Gynaecology, University of Manchester School of Cancer, St Mary's Hospital, Oxford
Road, Manchester M13 9WL, UK

REVIEW Received 04-01-2011

© 2011 International Menopause Society Revised 20-02-2011
DOI: 10.3109/13697137.2011.579650 Accepted 02-04-2011
 Levonorgestrel intrauterine systems for endometrial protection
developing endometrial cancer compared with non-obese age-
matched controls7. This is an important consideration given the
current concerns about the incidence of obesity in the general
population. Obese women have higher circulating levels of
endogenous estrogens, thought to be derived from the conver-
sion of androgens to estrogens in peripheral adipose tissue8.
The predominant theory describing the relationship of
estrogen and progesterone to endometrial cancer risk is known
as the unopposed estrogen hypothesis9,10. Under this hypoth-
esis, cancer risk is increased in women with high plasma levels
of bioavailable estrogens and/or low plasma progesterone.
Mechanistically, estrogens have a mitogenic effect and this is
seen physiologically with cyclical proliferation of the endome-
trium in response to increasing estrogen levels. Progestogens
have an antagonistic effect on this proliferation and induce
secretory change11. In an environment of unopposed estrogen
or reduced progestogen, endometrial proliferation occurs
Climacteric Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14
unchecked, increasing the chances of accumulation of genetic
abnormalities that may lead on to the development of malig-
nancy. Endometrial cancer is thought to develop via a number
of potentially reversible stepwise precursor stages12. Endome-
trial hyperplasia of varying degrees confers increased risk of
cancer. Atypical endometrial hyperplasia is associated with the
highest risk of endometrial cancer, being 14 times more likely
to progress than hyperplasia without cellular atypia (relative
risk 14.2, 95% confidence interval (CI) 5.3–38.0)13. Concur-
rent cancer may be found in up to half of those women with
For personal use only.
atypical hyperplasia14. In non-atypical hyperplasia, the risk of
both concurrent and subsequent cancer is thought to be much
lower but large cohort studies have not been performed14.
A further group of women at higher risk of developing endo-
metrial malignancy are those with hereditary non-polyposis
colorectal cancer syndrome (HNPCC) in which germline muta-
tions in DNA repair genes lead to an increased risk of develop-
ing endometrial cancer in addition to several other cancers15.
Women with HNPCC have a 40–60% lifetime risk of endo-
metrial cancer16–18 although this inherited predisposition
accounts for less than 2% of all endometrial cancer cases.
Currently, preventative interventions for women at high
risk of developing endometrial cancer are a contentious issue,
even for those at highest risk. For example, in women with
HNPCC, the effectiveness of screening and chemoprevention
remains unproven and risk-reducing surgery remains the only
proven method of prevention for those younger women who
have completed their families18. Similarly, hysterectomy is cur-
rently the preferred treatment for women with atypical endo-
metrial hyperplasia. Although tamoxifen use is a risk factor
for endometrial cancer, screening of women using tamoxifen
appears to be neither cost effective nor likely to reduce mortal-
ity from endometrial cancer19.
Progestogens, by virtue of their protective effect on the
endometrium, may offer an attractive second option for
women at higher risk of developing endometrial cancer who
wish to retain fertility or avoid surgery. There is limited evi-
dence to suggest that high-dose oral progestogens may be an
effective treatment of atypical hyperplasia and even early can-
cers confined to the endometrium20. Similarly, when used in
Climacteric
Wan and Holland
conjunction with estrogen replacement therapy, oral proges-
togens provide protection against the development of endo-
metrial cancer21. The levonorgestrel-releasing intrauterine
system (LNG-IUS) offers a route for local administration of
progestogen, delivering 20 μg of levonorgestrel daily. The
most commonly used device is the Mirena® (Bayer Pharma
AG, Berlin, Germany). This system is currently used in the
contraceptive setting but has found wider uses in patients with
menorrhagia. Being less user-dependent, it seems an attractive
option for chemoprevention where long-term compliance can
be problematic. If effective, the LNG-IUS could be particularly
helpful for surgically unfit women with atypical hyperplasia,
women with HNPCC and those taking tamoxifen.
OBJECTIVES
This systematic review analyzes the current evidence for the use
of LNG-IUS in women at high risk of endometrial cancer. In
this review, high-risk women were defined as women using
estrogen replacement therapy (ERT), women using tamoxifen,
women with a diagnosis of endometrial hyperplasia, women
with HNPCC and women with a BMI ⬎ 30 kg/m2. The primary
outcome measure was incidence of endometrial cancer. Second-
ary outcomes assessed were endometrial proliferative activity,
development of atypical hyperplasia and adverse events.
METHODS
A systematic review with a prospective protocol was under-
taken (the protocol is available from the authors on request).
Identification of sources
Searches were performed of the Cochrane Central Register of
Controlled Trials, MEDLINE (1950 to July 2009), EMBASE
(1988 to July 2009), CINAHL (1982 to July 2009). Ongoing
trials were identified by searching the UK National Research
Register (NRR) Archive (to September 2007), and Current
Controlled Trials (accessed July 2009). The reference lists of
articles chosen for inclusion were also hand-searched.
Eligibility criteria
Randomized, controlled trials (RCTs) comparing LNG-IUS with
no treatment, placebo or other hormonal therapy in adult (⬎ 18
years old) females were identified. Where no RCTs were identi-
fied for a particular risk group under study, prospective cohort
studies were identified. Studies that were included tested inter-
ventions that were initiated prior to a diagnosis of primary endo-
metrial cancer. In these studies, the development of endometrial
cancer, endometrial hyperplasia or endometrial polyp formation
was either their primary or secondary endpoint. Additionally,
studies were included where measurement of endometrial
thickness and histological evaluation of the endometrium were
623
 Levonorgestrel intrauterine systems for endometrial protection
performed. Studies that examined the use of LNG-IUS in the
treatment of early invasive endometrial cancers were excluded.
In addition, studies comparing different doses of intrauterine
levonorgestrel in reduction of endometrial cancer risk without
a control group were also excluded.
Search strategy
The following search strategy was used to search MEDLINE,
EMBASE and CINAHL via the NHS Evidence portal.
(1) exp PROGESTERONE OR exp PROGESTINS OR mirena.
ti.ab OR exp INTRAUTERINE DEVICES, MEDICATED
OR exp LEVONORGESTREL OR (levonorgestrel AND
releasing).ti,ab
(2) exp ENDOMETRIAL NEOPLASMS OR exp ENDOME-
TRIAL HYPERPLASIA
Climacteric Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14
(3) obesity.ti, ab OR expOBESITY OR obese.ti,ab OR over-
weight.ti,ab OR over-weight.ti,ab OR BMI.ti,ab OR (body
AND mass AND index).ti,ab
(4) HNPCC.ti,ab OR exp COLORECTAL NEOPLASMS,
HEREDITARY NONPOLYPOSIS/
(5) exp TAMOXIFEN/
(6) exp HORMONE REPLACEMENT THERAPY/ OR (hor-
mone AND replacement).ti,ab OR exp ESTROGEN
REPLACEMENT THERAPY/
For personal use only.
Each of searches (2)–(6) was combined with search (1) using
the AND operator. A simplified search was applied to the UK
National Research Register (NRR) Archive and the Current
Controlled Trials register.
Figure 1 Study selection process for systematic review of intrauterine progestogens to prevent endometrial cancer
624
Wan and Holland
Study selection and data extraction
One reviewer (Y.L.W.) scanned the titles and abstracts of each
record retrieved from the electronic searches. Abstracts that
did not meet the inclusion criteria were rejected. All accepted
abstracts and abstracts which could not be rejected with cer-
tainty were inspected in full by the two reviewers (Y.L.W. and
C.M.H.) independently. Both reviewers independently used
an inclusion criteria form to assess the trial’s eligibility for
inclusion and then confirmed eligibility of selected studies
together. The references from each eligible report were hand-
searched for further studies of relevance. Data from the
selected trials was extracted using a standardized data collec-
tion form and included details of the study population, trial
design, interventions and outcomes. The risk of bias was
assessed using the Cochrane bias assessment tool. Data were
combined using Peto odds ratios22. Meta-analyses of the stud-
ies investigating the hormone replacement therapy (HRT) and
tamoxifen subgroups were performed using a fixed effects
model in the RevMan program. The magnitude of statistical
heterogeneity was assessed using the I2 statistic. Funnel plots
were used to assess the risk of bias across studies.
RESULTS
Eleven studies were identified for inclusion in this review (see
Figure 1). On initial review of 254 study titles and abstracts, 195
reports were excluded as not meeting the eligibility criteria; 59
reports were reviewed in full. On detailed analysis, 44 of these
Climacteric
 Levonorgestrel intrauterine systems for endometrial protection
did not meet the inclusion criteria. The remaining 15 articles
described studies examining the effect of LNG-IUS in women
using HRT, in women with endometrial hyperplasia and women
using tamoxifen. From the 15 studies identified, a further four
were excluded; one, although having comparison groups, com-
pared different doses of intrauterine levonorgestrel with each
other but had no control group. Another study was excluded
because the intervention evaluated was Progestasert®, a proges-
terone-releasing intrauterine system no longer on the market23.
A third excluded study was a further report of a study already
included in the review. The final study was excluded as it was a
quasi-randomized, controlled trial, assigning treatments accord-
ing to the order of enrolment24. Unfortunately, randomized,
controlled trials investigating the use of intrauterine levonor-
gestrel as a chemopreventative agent in women with HNPCC
and women with a BMI ⬎ 35 kg/m2 have failed to recruit suffi-
cient numbers of participants25,26. No other suitable studies
Climacteric Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14
related to women in these groups were identified for inclusion.
LNG-IUS and endometrial cancer risk reduction in
women using ERT
Six studies investigated the effect of LNG-IUS in women
using ERT, four of which were in postmenopausal women
For personal use only.
Figure 2 Forest plot of pooled odds ratios in (a) endometrial proliferation and (b) study withdrawal in postmenopausal women using the levonorgestrel-
releasing intrauterine system (LNG-IUS) and estrogen therapy compared to other combinations of estrogen and progestogen
Climacteric
Wan and Holland
(see Figure 2) and two in perimenopausal women (see
Table 1). Data synthesis was not possible due to the high
number of zero cell counts. In total, the six studies included
578 women with study sizes ranging from 19 to 200 par-
ticipants. No new cancers or cases of endometrial hyperpla-
sia were seen during follow-up. The effects of LNG-IUS
on endometrial proliferation and compliance are shown in
Figure 2.
In postmenopausal women using LNG-IUS, histological
assessment showed profound progestogenic effects and endo-
metrial atrophy in all studies27–29. Only one study showed a
significant difference in proliferative activity when compared
to an oral progestogen (medroxyprogesterone acetate, MPA)30.
Endometrial thickness did not differ between the LNG-IUS or
other progestogen therapy groups27–30.
Results in perimenopausal women are similar, with histo-
logical evidence of endometrial atrophy. There was no differ-
ence in the protection offered against hyperplasia between
the use of oral progestogen or the use of LNG-IUS31. Simi-
larly, cyclical oral levonorgestrel and LNG-IUS appeared
equally efficacious in preventing endometrial proliferation32.
LNG-IUS was associated with fewer bleeding days after
12 months of therapy in the majority of regimens27,30–32,36.
In the remaining studies, no differences in bleeding patterns
625
 Climacteric Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14
For personal use only.

626
Table 1 Summary of comparative, randomized, controlled trials using the levonorgestrel-releasing intrauterine system (LNG-IUS) to prevent endometrial pathology in postmenopausal
and perimenopausal women using hormone replacement therapy (HRT)

Proliferative effects seen Withdrew from

Number of
Follow-up hyperplasia/ LNG-IUS Comparison LNG-IUS Conventional
Study IUS type ⫹ estrogen Comparison groups n (months) cancers group group arm HRT arm
Postmenopausal women
Raudaskoski 29 estradiol continuous oral 40 6, 12 0 0/15 2/17 3 pre-insertion 3
patch ⫹ Levonova estradiol and 0/15 0/17 2 post-insertion
(20 µg release) norethindrone

Suhonen 27 oral estradiol ⫹ oral estradiol ⫹ 19 6, 12 0 0/9 no information 1 2

LNG-IUS (20 µg) subdermal provided 1 excluded due
levonorgestrel to IUS
Levonorgestrel intrauterine systems for endometrial protection

expulsion

Antoniou 28 estradiol vaginal estradiol 56 12 0 0/28 0/28 0 0

patch ⫹ LNG-IUS ring ⫹ progesterone
(20 µg release) suppositories

Raudaskoski30 LNG-IUS group: oral oral estradiol ⫹ 109 6, 12 0 0/55 19/50 1 4

estradiol ⫹ LNG- MPA (day 1–14) 0/55 18/47
IUS (20 µg)

Total 220 8/113 (7%) 9/111 (8%)

Perimenopausal women
Andersson32 oral estradiol ⫹ LNG- cyclic oral 40 12 0 0/18 0/19 2 1
IUS (20 µg) estradiol ⫹ oral
levonorgestrel
Boon31 oral estradiol 2 cyclical oral 200 13, 26 0 0/82 0/71 18 29
mg ⫹ LNG-IUS estradiol and 0/76 0/63 6 8
20 µg release) norethisterone

Total 240 26/120 (22%) 38/120 (32%)

MPA, medroxyprogesterone acetate

Wan and Holland

Climacteric
 Levonorgestrel intrauterine systems for endometrial protection Wan and Holland

were seen at 12 months28,29. As a result, continuation and

Control
arm
Withdrawal from

n/a
acceptability were high30,31,36.

0
Table 2 Summary of studies of levonorgestrel-releasing intrauterine system (LNG-IUS, Mirena) in women with endometrial hyperplasia (total number of women studied ⫽ 315)

study

LNG-IUS

LNG-IUS:
LNG-IUS and endometrial cancer risk reduction in

months

month
22 by
arm women with endometrial hyperplasia
0

108
0 at 6
Two non-randomized, prospective studies investigated the
Comparison

effect of LNG-IUS in women with endometrial hyperplasia

Number of cancers

group

compared with oral progestogen therapy (see Table 2). None

0

0
of the 315 participants in these two studies developed invasive
cancer.
Vereide and colleagues described the effect of LNG-IUS on
group
Study

women with both atypical and non-atypical endometrial

0

0
hyperplasia and compared this to the outcomes of women
treated with 10 mg medroxyprogesterone acetate daily for 10
oral progestin

oral progestin
Comparison

observation

observation
days per month37. At 3 months, nearly half of the systemically
52/107

53/107
group

46/85

51/85
17/31
Climacteric Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14

treated patients had persisting hyperplasia compared to none

of the LNG-IUS-treated patients. When histological scores
% Resolution

were compared between the two treatment groups, LNG-IUS

appeared superior to oral treatment (p ⫽ 0.018). This differ-
LNG-IUS group

44/44 LNG-IUS
14/22 LNG-IUS

ence was more marked in women whose pretreatment histol-

ogy indicated a likely high risk of progression (p ⫽ 0.002 for
26/26

66/66

removed

LNG-IUS, p ⫽ 0.05 for oral progestogens).

in situ

In a later study by the same group, a multicenter, prospec-

tive trial design was used to evaluate long-term effects of
MPA, medroxyprogesterone acetate; pre, premenopausal; post, postmenopausal; n/a, not available

LNG-IUS compared to oral MPA (10 mg daily for 3 months)

For personal use only.

follow-up
Follow-up

or surveillance alone38. In this study, a morphometric scoring

(months)

(58 –108
long-term

weeks)

system (the D-score) was used to determine the likely risk of

3

progression to endometrial malignancy. Most women with a

D-score of ⬎ 1 had non-atypical hyperplasia using conven-
tional histopathological interpretation. LNG-IUS was found
26 (prospective

31 (historical

to be superior to MPA and surveillance alone in inducing

controls)
cohort)

histological regression at all risk levels at 6 months. In long-

258
n

term follow-up, those patients who had had their LNG-IUS

removed had similar rates of resolution to those in the oral
progestogen or surveillance groups; 100% of women who
retained the LNG-IUS had regression. Of this latter group,
Comparison

progesterone

31 of 44 women had a D-score of ⬎ 1 that is, low risk of

cyclical oral

cyclical oral
group

surveillance

progression to significant endometrial abnormality. Only four

women had a score indicative of severe atypical hyperplasia.
MPA

only
or

LNG-IUS and endometrial cancer risk reduction in

Menopausal

pre and post

pre and post

women using tamoxifen

state

Three randomized, controlled trials investigated the use of

LNG-IUS in women using tamoxifen 20 mg per 24 h com-
Type of study

pared to surveillance alone (see Table 3); 354 women were

prospective

prospective

included in these studies. No cancers or cases of hyperplasia

multicenter

were seen in the LNG-IUS groups at long-term follow-up.

open
trial

trial

When considered together, the use of LNG-IUS reduced the

open

risk of endometrial polyp formation (Peto odds ratio (OR)

0.28 95% CI 0.15–0.55) and endometrial hyperplasia (Peto
Vereide37

OR 0.14; 95% CI 0.02–0.80) (Figure 3). Sonographic endo-

Orbo38
Study

metrial thickness did not differ between the treatment and

control groups during long-term follow-up. Overall, a greater

Climacteric 627
 Levonorgestrel intrauterine systems for endometrial protection Wan and Holland

Table 3 Summary of comparative, randomized, controlled trials using the levonorgestrel-releasing intrauterine system (LNG-IUS) to prevent endometrial pathology in tamoxifen
proportion of women with LNG-IUS complained of abnormal

Comparison arm
vaginal bleeding/spotting at 6 months compared with the con-

11/182 (6%)
Withdrawal from study
trol groups, but, by 12 months, this difference did not reach

7
4

0
significance in any of the studies. By the end of the first year
of follow-up, up to 5% of patients in the study groups had
withdrawn due to bleeding. The potential difference in abnor-
mal bleeding was not translated into a significant difference

Control LNG-IUS arm

19/172 (11%)
in withdrawal rates in those using the LNG-IUS compared to
those under observation alone (Peto OR 2.02; 95% CI 0.94–

9
10

0
4.36) (Figure 3).

DISCUSSION
4/72
4/52
9/58
endometrium
Proliferative

LNG-IUS and endometrial cancer risk reduction in

women using ERT
LNG-IUS

4/70
0/47
0/18

In the UK in 2005, the LNG-IUS was licensed for endometrial

Climacteric Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14

protection for women using ERT who retained their uterus,

although it is not licensed for this indication in the US or
Control
hyperplasia/cancers

4/72*
1/52*

Canada42. Compared with conventional oral and vaginal pro-

0/58
Number of

gestogen preparations, the LNG-IUS appears to be more effec-

tive at opposing estrogenic effects and inducing pseudode-
LNG-IUS

cidual stromal reactions in women using ERT. Histological

0/70
0/47
0/55

evidence of endometrial quiescence and significant progesto-

genic activity at 12 months of therapy suggests that adequate
endometrial protection may be provided by LNG-IUS in
(months) LNG-IUS Control
endometrial polyps
For personal use only.

14/72
8/52
9/58

women using estrogen replacement. Overall, the risk of endo-

Number of

metrial cancer in women taking combined estrogen/progesto-

gen therapy with a minimum of 10 days progestogen is low33
and, for those taking continuous combined treatment, the
4/70
3/47
1/55

incidence of endometrial hyperplasia is less than 1% per

year34. Therefore, the published studies are inadequately pow-
ered to detect any significant difference in endometrial cancer
Follow-up

or endometrial hyperplasia as the primary outcomes. Endo-

12, 60
12
36

metrial thickness appears not to correlate with changes in

endometrial histology due to the LNG-IUS and therefore
endometrial ultrasound may not be the best tool for investi-
Number in
LNG-IUS

gating abnormal bleeding in women with an IUS in

group

113
99

354
142

situ24,27,28,30,35,41.
As well as significant clinical heterogeneity across the
included studies, no two studies investigated like-for-like
surveillance alone

regimens. Additionally, bias assessment revealed significant

Comparison

variation in allocation concealment, treatment washout peri-

Pre, premenopausal; post, postmenopausal
group

surveillance
surveillance

ods and blinding across the studies. Similarly, there are

reported differences in the histological appearances that are
taken as confirmation of adequate endometrial suppression.
The lowest effective dose of LNG-IUS has yet to be estab-
lished, although a smaller device delivering 10 μg/24 h was
*, Non-atypical hyperplasia
Menopausal

pre and post

developed for ease of insertion in postmenopausal women

state

and was associated with a favorable serum lipid profile43.

post
post

Trials involving a variety of lower-dose LNG-IUS have all

shown adequate endometrial suppression but varying amen-
orrhea rates30,44. Notably, in studies using 20 μg LNG-IUS,
Gardner35,39

there was no greater withdrawal rate due to side-effects.

Kesim41
Chan40
Study

Despite these limitations, it seems that the LNG-IUS is at least

Total
users

as effective as oral and vaginal progestogens in preventing

628 Climacteric
 Levonorgestrel intrauterine systems for endometrial protection Wan and Holland
Climacteric Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14
For personal use only.

Figure 3 Forest plot of pooled odds ratios in (a) endometrial polyps formation, (b) endometrial hyperplasia, (c) endometrial proliferation, (d) study
withdrawal in women using the levonorgestrel-releasing intrauterine system (LNG-IUS) and tamoxifen compared to tamoxifen alone

endometrial proliferation in women using estrogen replace-
ment and is therefore likely also to be as effective in prevent-
ing endometrial pathology in this group. It is important to
note that there have been concerns regarding adverse effects
of systemically absorbed levonorgestrel from the LNG-IUS.
In particular, a large case–control study from Finland has
recently reported the unexpected finding that sole use of the
LNG-IUS is associated with a significantly elevated risk of
developing breast cancer and that this risk was heightened by
the concurrent use of estradiol45. This further suggests that
continued development towards a lower-dose LNG-IUS
would be beneficial.

Climacteric 629
 Levonorgestrel intrauterine systems for endometrial protection
LNG-IUS and endometrial cancer risk reduction in
women with endometrial hyperplasia
The results of this systematic review suggest that the LNG-IUS
is effective in inducing regression in women with hyperplasia
of low malignant potential. This conclusion is limited, how-
ever, by a high degree of selection bias in these cohorts and
by non-random loss of patients to follow-up and missing data.
Only two studies were identified. One of these had very small
sample sizes and only very short-term follow-up. Further-
more, patients in the oral therapy groups were treated with
relatively low doses of MPA for only 3 months. Women at
high risk of progression to cancer were offered hysterectomy
in these studies. There are inherent problems in investigating
the use of a medical therapy for women with hyperplasia of
high malignant potential, within a robust and ethically accept-
able study design, given the concern about occult malignancy
Climacteric Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14
and malignant progression. A sufficiently powered, random-
ized study could only be reasonably considered as a multi-
center trial and women in such a study would need to be
deemed unfit for surgery by agreed criteria prior to being
randomized to either high-dose oral progestogen or LNG-
IUS.
Currently, therefore, there is insufficient evidence to recom-
mend the LNG-IUS as an equivalent treatment to high-dose
oral progestogens for non-atypical hyperplasia. Similarly,
there is insufficient evidence to support LNG-IUS either as a
For personal use only.
safe alternative to hysterectomy for atypical hyperplasia in
women fit enough for surgery or to high-dose oral progesto-
gens in those unfit for surgery.
LNG-IUS and endometrial cancer risk reduction in
women using tamoxifen
This review extends the findings of the recent Cochrane sys-
tematic review investigating the efficacy of LNG-IUS in women
using tamoxifen46 by including a further randomized, con-
trolled trial41 and the long-term follow-up data from the RCT
by Gardner and colleagues35. The meta-analysis confirms that
not only is endometrial polyp formation reduced at long-term
follow-up but additionally suggests that endometrial hyper-
plasia is also reduced. However, meta-analysis with respect to
development of hyperplasia is limited due to the relative rarity
of the event and thus the odds ratio calculation is weighted
heavily towards a single study. In the trials reviewed here and
in the Cochrane review, the risk of bias is low but none of the
trials were sufficiently powered to detect a significant differ-
ence in development of cancer. The effect of LNG-IUS in
reducing the incidence of cancer in these women is therefore
unknown. Regarding the reporting of unwanted side-effects,
it must be noted that the definition of abnormal vaginal bleed-
ing varies with each study and must be considered in the
cultural context of the country in which the study was
conducted. Additionally, the trials by Gardner and
colleagues35,39 and Kesim and colleagues41 were conducted in
630
Wan and Holland
postmenopausal patients whilst that by Chan and colleagues40
included both pre- and postmenopausal women.
Limitations
Despite a thorough search strategy, unpublished studies may
have been missed leading to publication bias. Funnel plots
examining the existence of bias are not meaningful due to the
small number of studies in each subgroup analysis. Addition-
ally, as no studies were adequately powered or of sufficient
duration to detect differences in development of endometrial
cancer, the secondary outcomes of endometrial proliferative
activity and hyperplasia are considered as surrogate endpoints.
Although there is a correlation between endometrial activity
and hyperplasia and development of cancer, this is not an
absolute relationship. Thus, definitive conclusions as to the
efficacy of LNG-IUS in endometrial cancer prevention are not
possible. In this review, meta-analysis was performed within
the estrogen replacement and tamoxifen subgroups as the
studies in each of these groups were of sufficiently high quality
and had similar endpoints. Furthermore, the three trials of
women using LNG-IUS in conjunction with tamoxifen showed
no statistical evidence of significant heterogeneity in the endo-
metrial outcomes. However, due to the relative rarity of endo-
metrial activity in postmenopausal women using ERT, zero
counts in both arms of trials considering LNG-IUS for endo-
metrial protection made relative effect calculations problem-
atic. Additionally, differences in the formulations and dosing
schedules of estrogens and progestogens, compliance and
washout periods may have influenced results further.
CONCLUSION
No studies have adequately demonstrated a beneficial effect
of the LNG-IUS as prevention of endometrial cancer in the
following high-risk groups: women using ERT, women using
tamoxifen, and women with endometrial hyperplasia. In par-
ticular, there is insufficient evidence to recommend LNG-IUS
as treatment of non-atypical or atypical endometrial hyper-
plasia compared to high-dose oral progestogens or hysterec-
tomy, respectively. LNG-IUS is, however, effective in reducing
unwanted benign endometrial effects (e.g. polyps) in women
taking tamoxifen. Overall, the LNG-IUS is superior to other
routes of progestogen administration with respect to compli-
ance rates, and potentially fewer systemic side-effects47,48. The
most common reason for removal of LNG-IUS in the trials
reviewed was unacceptable vaginal bleeding, which is worry-
ing for this group of women who tend to be postmenopausal
or under surveillance for potential endometrial pathology. We
propose that multicenter, randomized studies of LNG-IUS vs.
standardized doses of oral progestogen in specific groups of
high-risk women are a high priority in determining the true
value of LNG-IUS as chemoprevention in these groups. We
believe that this is important given the increasing prevalence
of risk factors for endometrioid endometrial cancer.
Climacteric
 Levonorgestrel intrauterine systems for endometrial protection
ACKNOWLEDGEMENTS
Many thanks to Dr Xin Shi, Biostatistics Group, University
of Manchester and Dr Sofia Dias, University of Bristol for
providing statistical advice. Many thanks also to Dr Mourad
Seif, University of Manchester for helpful comments on the
manuscript. Y.L.W. performed the searches, data selection,
data collection and analysis and wrote the final manuscript.
References
1. Bray F, Dos Santos Silva I, Moller H, Weiderpass E. Endometrial
cancer incidence trends in Europe: underlying determinants and
prospects for prevention. Cancer Epidemiol, Biomarkers Preven-
tion 2005;14:1132–42
Climacteric Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14
2. Ferlay JSH, Bray F, Forman D, Mathers C, Parkin DM. GLOBO-
CAN 2008, Cancer Incidence and Mortality Worldwide. IARC
CancerBase No 10, 2010 [cited; Available from: http://globocan.
iarc.fr/]
3. Elwood JM, Cole P, Rothman KJ, Kaplan SD. Epidemiology of
endometrial cancer. J Natl Cancer Instit 1977;59:1055–60
4. La Vecchia C, Franceschi S, Decarli A, Gallus G, Tognoni
G. Risk factors for endometrial cancer at different ages. J Natl
Cancer Instit 1984;73:667–71
5. Cogliano V, Grosse Y, Baan R, Straif K, Secretan B, El
Ghissassi F. Carcinogenicity of combined oestrogen-progesta-
For personal use only.
gen contraceptives and menopausal treatment. Lancet Oncol
2005;6:552–3
6. Bergman L, Beelen ML, Gallee MP, Hollema H, Benraadt
J, van Leeuwen FE. Risk and prognosis of endometrial cancer af-
ter tamoxifen for breast cancer. Comprehensive Cancer Centres’
ALERT Group. Assessment of Liver and Endometrial Cancer
Risk following Tamoxifen. Lancet 2000;356:881–7
7. Humphrey MM, Apte SM. The use of minimally invasive
surgery for endometrial cancer. Cancer Control 2009;16:30–7
8. Fader AN, Arriba LN, Frasure HE, von Gruenigen VE. En-
dometrial cancer and obesity: epidemiology, biomarkers, preven-
tion and survivorship. Gynecol Oncol 2009;114:121–7
9. Siiteri PK. Steroid hormones and endometrial cancer. Can-
cer Res 1978;38:4360–6
10. Key TJ, Pike MC. The dose-effect relationship between ‘unop-
posed’ oestrogens and endometrial mitotic rate: its central role in
explaining and predicting endometrial cancer risk. Br J Cancer
1988;57:205–12
11. Good RG, Moyer DL. Estrogen-progesterone relationships in the
development of secretory endometrium. Fertil Steril 1968;19:
37–49
12. Sherman ME. Theories of endometrial carcinogenesis: a multidis-
ciplinary approach. Mod Pathol 2000;13:295–308
13. Lacey JV Jr, Ioffe OB, Ronnett BM, et al. Endometrial carcinoma
risk among women diagnosed with endometrial hyperplasia: the
34-year experience in a large health plan. Br J Cancer 2008;98:
45–53
14. Lacey JV Jr, Chia VM. Endometrial hyperplasia and the risk of
progression to carcinoma. Maturitas 2009;63:39–44
15. Gruber SB, Thompson WD. A population-based study of endome-
trial cancer and familial risk in younger women. Cancer and
Steroid Hormone Study Group. Cancer Epidemiol Biomarkers
Prev 1996;5:411–17
Climacteric
Wan and Holland
C.M.H. conceived the review, selected the studies and co-
authored the final manuscript.
Conflict of interest The authors report no confl icts of
interest. The authors alone are responsible for the content
and writing of the paper.
Source of funding Y.L.W. is a NIHR Academic Clinical
Fellow.
16. Aarnio M, Mecklin JP, Aaltonen LA, Nystrom-Lahti M, Jarvinen
HJ. Life-time risk of different cancers in hereditary non-polyposis
colorectal cancer (HNPCC) syndrome. Int J Cancer 1995;
64:430–3
17. Vasen HF, Wijnen JT, Menko FH, et al. Cancer risk in families
with hereditary nonpolyposis colorectal cancer diagnosed by
mutation analysis. Gastroenterology 1996;110:1020–7
18. Manchanda R, Menon U, Michaelson-Cohen R, Beller U, Jacobs
I. Hereditary non-polyposis colorectal cancer or Lynch syndrome:
the gynecological perspective. Curr Opin Obstet Gynecol
2009;21:31–8
19. Neven P, Vergote I. Should tamoxifen users be screened for
endometrial lesions? Lancet 1998;351:155–7
20. Gadducci A, Spirito N, Baroni E, Tana R, Genazzani AR. The
fertility-sparing treatment in patients with endometrial atypical
hyperplasia and early endometrial cancer: a debated therapeutic
option. Gynecol Endocrinol 2009;25:683–91
21. Lethaby A, Farquhar C, Sarkis A, Roberts H, Jepson R, Barlow
D. Hormone replacement therapy in postmenopausal women:
endometrial hyperplasia and irregular bleeding. Cochrane data-
base of systematic reviews (Online) 2000(2):CD000402
22. Peto odds ratio method. Cochrane Handbook for Systematic
Reviews of Interventions, Version 5.0.2 [updated September
2009]. Chapter 9.4.4.2. [cited; Available from: http://www.
cochrane-handbook.org/]
23. Volpe A, Botticelli A, Abrate M, et al. An intrauterine progester-
one contraceptive system (52 mg) used in pre- and peri-menopau-
sal patients with endometrial hyperplasia. Maturitas 1982;
4:73–9
24. Suvanto-Luukkonen E, Malinen H, Sundstrom H, Penttinen J,
Kauppila A. Endometrial morphology during hormone replace-
ment therapy with estradiol gel combined to levonorgestrel-re-
leasing intrauterine device or natural progesterone. Acta Obstet
Gynecol Scand 1998;77:758–63
25. Intrauterine levonorgestrel and observation or observation alone
in preventing atypical endometrial hyperplasia and endometrial
cancer in women with hereditary non-polyposis colorectal cancer
or Lynch Syndrome. National Institute of Health: Clinical Trials.
gov [cited 17th July 2009]; Available from: http://clinicaltrials.
gov/ct2/show/record/NCT00566644
26. A randomized, controlled, comparative study of a levonorgestrel
intrauterine system for the prevention of endometrial cancer in
patients aged 40–50 with BMI greater than 35. National Instititue
of Health: Clinical Trials.gov. [cited 17th July 2009]; Available
from: http://clinicaltrials.gov/ct2/show/record/NCT00161226
27. Suhonen SP, Holmstrom T, Allonen HO, Lahteenmaki P. Intrau-
terine and subdermal progestin administration in postmenopausal
hormone replacement therapy. Fertil Steril 1995;63:336–42
631
 Levonorgestrel intrauterine systems for endometrial protection
28. Antoniou G, Kalogirou D, Karakitsos P, Antoniou D, Kalogirou
O, Giannikos L. Transdermal estrogen with a levonorgestrel-re-
leasing intrauterine device for climacteric complaints versus
estradiol-releasing vaginal ring with a vaginal progesterone sup-
pository: clinical and endometrial responses. Maturitas 1997;
26:103–11
29. Raudaskoski TH, Lahti EI, Kauppila AJ, Apaja-Sarkkinen MA,
Laatikainen TJ. Transdermal estrogen with a levonorgestrel-re-
leasing intrauterine device for climacteric complaints: clinical and
endometrial responses. Am J Obstet Gynecol 1995;172:114–19
30. Raudaskoski T, Tapanainen J, Tomas E, et al. Intrauterine 10 μg
and 20 μg levonorgestrel systems in postmenopausal women
receiving oral oestrogen replacement therapy: clinical, endome-
trial and metabolic response. BJOG 2002;109:136–44
31. Boon J, Scholten PC, Oldenhave A, Heintz AP. Continuous intra-
uterine compared with cyclic oral progestin administration in
perimenopausal HRT. Maturitas 2003;46:69–77
32. Andersson K, Mattsson LA, Rybo G, Stadberg E. Intrauterine
release of levonorgestrel – a new way of adding progestogen in
hormone replacement therapy. Obstet Gynecol 1992;79:963–7
Climacteric Downloaded from informahealthcare.com by Universitaet Zuerich on 12/26/14
33. Archer DF. The effect of the duration of progestin use on the
occurrence of endometrial cancer in postmenopausal women.
Menopause 2001;8:245–51
34. Riphagen FE. Intrauterine application of progestins in hormone
replacement therapy: a review. Climacteric 2000;3:199–211
35. Gardner FJ, Konje JC, Bell SC, et al. Prevention of tamoxifen
induced endometrial polyps using a levonorgestrel releasing intra-
uterine system long-term follow-up of a randomised control trial.
Gynecol Oncol 2009;114:452–6
36. Suvanto-Luukkonen E, Sundstrom H, Penttinen J, Laara E, Pram-
ila S, Kauppila A. Percutaneous estradiol gel with an intrauterine
For personal use only.
levonorgestrel releasing device or natural progesterone in hor-
mone replacement therapy. Maturitas 1997;26:211–17
37. Vereide AB, Arnes M, Straume B, Maltau JM, Orbo A. Nuclear
morphometric changes and therapy monitoring in patients with
endometrial hyperplasia: a study comparing effects of intrauter-
ine levonorgestrel and systemic medroxyprogesterone. Gynecol
Oncol 2003;91:526–33
38. Orbo A, Arnes M, Hancke C, Vereide AB, Pettersen I, Larsen K.
Treatment results of endometrial hyperplasia after prospective
632
Wan and Holland
D-score classification: a follow-up study comparing effect of
LNG-IUD and oral progestins versus observation only. Gynecol
Oncol 2008;111:68–73
39. Gardner FJ, Konje JC, Abrams KR, et al. Endometrial protection
from tamoxifen-stimulated changes by a levonorgestrel-releasing
intrauterine system: a randomised controlled trial. Lancet
2000;356:1711–17
40. Chan SS, Tam WH, Yeo W, et al. A randomised controlled trial
of prophylactic levonorgestrel intrauterine system in tamoxifen-
treated women. BJOG 2007;114:1510–15
41. Kesim MD, Aydin Y, Atis A, Mandiraci G. Long-term effects of
the levonorgestrel-releasing intrauterine system on serum lipids
and the endometrium in breast cancer patients taking tamoxifen.
Climacteric 2008;11:252–7
42. Varma R, Sinha D, Gupta JK. Non-contraceptive uses of lev-
onorgestrel-releasing hormone system (LNG-IUS) – a systematic
enquiry and overview. Eur J Obstet Gynecol Reprod Biol
2006;125:9–28
43. Sturdee DW, Rantala ML, Colau JC, Zahradnik HP, Riphagen
FE. The acceptability of a small intrauterine progestogen-releas-
ing system for continuous combined hormone therapy in early
postmenopausal women. Climacteric 2004;7:404–11
44. Wollter-Svensson LO, Stadberg E, Andersson K, Mattsson LA,
Odlind V, Persson I. Intrauterine administration of levonorgestrel
5 and 10 μg/24 hours in perimenopausal hormone replacement
therapy. A randomized clinical study during one year. Acta Obstet
Gynecol Scand 1997;76:449–54
45. Lyytinen HK, Dyba T, Ylikorkala O, Pukkala EI. A case-control
study on hormone therapy as a risk factor for breast cancer in
Finland: Intrauterine system carries a risk as well. Int J Cancer
2010;126:483–9
46. Chin J, Konje JC, Hickey M. Levonorgestrel intrauterine system
for endometrial protection in women with breast cancer on adju-
vant tamoxifen. Cochrane Database of Systematic Reviews
(Online) 2009(4):CD007245
47. Luukkainen T, Toivonen J. Levonorgestrel-releasing IUD as a
method of contraception with therapeutic properties. Contracep-
tion 1995;52:269–76
48. Backman T. Benefit-risk assessment of the levonorgestrel intrau-
terine system in contraception. Drug Saf 2004;27:1185–204
Climacteric