Intraperitoneal levonorgestrel-releasing intrauterine devicefollowing uterine perforation: the role of progestins inadhesion formation

Ronit Haimov-Kochman1,3, Victoria Doviner2, Hagay Amsalem1, Diane Prus2,Amiram Adoni1 and Yuval Lavy1

1Department of Obstetrics and Gynecology and 2Department of Pathology, Hadassah University Hospital, Jerusalem, Israel

3To whom correspondence should be addressed at: Department of Obstetrics and Gynecology, Hadassah University Hospital,

Mount Scopus, P.O.B 24035, il-91240, Jerusalem, Israel. E-mail: rkochman@hotmail.com

BACKGROUND: Intrauterine contraception is a widely used, highly effective means of birth control. Uterine

perforation is a serious, albeit rare, complication of intrauterine device (IUD) use. Although uterine perforation by

levonorgestrel-releasing (20 mg/day) intrauterine system (LNG-IUS) has already been reported, the peritoneal

adhesion potential of this IUD is unknown. METHODS: The medical ®les of all patients diagnosed with an

intra-peritoneal IUD between the years 1990±2002 at Hadassah Medical Center were reviewed. Histopathological

study of peritoneal adhesion tissue adjacent to levonorgestrel medicated IUD was conducted in one case. RESULTS:

Eight cases of dislocated IUDs were found. Four cases used LNG-IUS and four other cases used copper-IUD.

Laparoscopy for IUD removal disclosed mild local peritoneal adhesions between omentum and pelvic organs in all

cases. No difference was noted in the appearance of the peritoneum in the presence of either a copper-IUD or

LNG-IUS. Histological examination of peritoneal tissue encasing the levonorgestrel-intrauterine system revealed

loose connective tissue with aggregates of submesothelial cells with a pseudo-decidual change. Immunohistochemical

staining for progesterone receptor was negative. CONCLUSIONS: The peritoneal adhesions potential of LNG-IUS

is low, similar to that of the copper-bearing IUD.

Key words: adhesions/intrauterine device/levonorgestrel/perforation

Introduction

Intrauterine contraception is a widely used, highly effective

means of birth control. Uterine perforation is a serious, albeit

rare, potential complication of intrauterine device (IUD) use.

Copper-bearing IUD is known to cause local peritoneal

adhesions (Adoni and Ben Chetrit, 1991). Therefore its

removal from the peritoneal cavity is recommended once

perforation is diagnosed. Levonorgestrel-releasing (20 mg/day)

intrauterine system (LNG-IUS) (Mirenaâ, Schering AG,

Germany) was introduced to the market in Israel in 1998.

Although uterine perforation by this form of IUD has already

been reported (Andersson et al., 1998), the peritoneal adhesion

potential of the levonorgestrel-releasing intrauterine device is

unknown.

Materials and methods

The medical ®les of all patients diagnosed with an IUD in the

peritoneal cavity between the years 1990±2002 at Hadassah Medical

Center were reviewed. In case number 4 (Table I) peritoneal adhesions

encasing the IUD were separated, formalin ®xed and paraf®n

embedded. Sections of 4 mm thickness were stained with haema-

toxylin and eosin. Immunohistochemistry was performed using

antibodies to vimentin (clone Vim3B4; Dako, Glostrup, Denmark),

smooth muscle actin (clone 1A4; Dako), CD-68 (clone KP1; Dako),

caldesmon (clone h-CD; Dako), calretinin ([PAD:DC8]; Zymed, San

Francisco, CA, USA), cytokeratin (clone LP34; Dako), desmin (clone

NCL-DE-R-11; Ventana Medical Systems Inc., Harvard, MA, USA),

estrogen receptor (clone 6F11; Ventana Medical Systems Inc.) and

progesterone receptor (clone 1A6; Ventana Medical Systems Inc.).

Results

Eight cases of dislocated IUDs were treated at Hadassah

Medical Center between the years 1990±2002 (Table I). Seven

of them were inserted within 3 months post-partum. Six of the

patients were breast-feeding at the time of IUD placement. In

four cases LNG-IUS (Mirena) was used and in four other cases

copper-IUD was inserted. The main reasons for investigation

of the IUD localization were mild abdominal pain and irregular

uterine bleeding. One patient was found to be 8 weeks

pregnant. The means for investigation were a transvaginal

ultrasonogram and an antero±posterior radiograph of the

pelvis. Diagnosis of dislocated IUD was made 12 days to 7

Human Reproduction Vol.18, No.5 pp. 990±993, 2003 DOI: 10.1093/humrep/deg203

990 ã European Society of Human Reproduction and Embryology

by guest on Novem

ber 14, 2015http://hum

rep.oxfordjournals.org/D

ownloaded from

months following insertion. Laparoscopy was performed

uneventfully in all patients. On laparoscopy, the perforation

site was noted in only four cases, when a relatively short time

had elapsed from insertion to laparoscopic removal. The IUD

was disclosed in ®ve cases encased in the omentum. Mild

peritoneal adhesions were reported between omentum and

pelvic organs in all eight cases. Lysis of peritoneal adhesions

was undertaken in only one case with a history of ®ve previous

Caesarean sections. No difference was noted in the appearance

of the peritoneum in the presence of either a copper-IUD or a

levonorgestrel-releasing intrauterine system.

Histological examination of peritoneal tissue adjacent to

LNG-IUS (in case number 4), revealed richly vascularized

loose connective tissue with calci®cations and mild chronic

in¯ammation admixed with a few foreign body-type giant

cells. Aggregates of swollen cells with eosinophilic cytoplasm

and vesicular nucleus were found embedded in the tissue.

These cells stained with vimentin only and were identi®ed as

sub-mesothelial stromal cells of the peritoneal cavity that

underwent pseudo-decidual changes (Figure 1). Immuno-

histochemical staining for estrogen receptor was positive, but

staining for progesterone receptor was negative.

Discussion

The most serious potential complication of IUD use is uterine

perforation, which has been reported to happen in 0±1.3 events

per 1000 insertions (Andersson et al., 1998; Markovitch et al.,

2002). Uterine perforation by a copper-bearing IUD, although

usually asymptomatic, could rarely entail severe morbidity

such as bowel obstruction and infection. This type of IUD is

known to cause local peritoneal adhesions. The process of

adhesion formation in reaction to copper-bearing IUD is self-

limited and stops after a short period of time (Adoni and

Ben Chetrit, 1991). The recommended treatment of a mis-

placed copper-bearing IUD is its removal from the abdominal

cavity by laparoscopy (Silva and Larson, 2000), although

recently conservative management of mislocated IUDs has also

been suggested (Markovitch et al., 2002). Laparotomy is

considered to be by far too dangerous a procedure for the

management of an ectopic asymptomatic copper-IUD (Adoni

and Ben Chetrit, 1991).

Levonorgestrel-releasing (20 mg/day) intrauterine system

(Mirena) is a relatively new form of contraception. Perforation

of the uterus during its insertion has been reported among other

IUD types in a Swedish survey (Andersson et al., 1998);

however, the abdominal cavity in its presence was not

described. The adhesion formation potential of LNG-IUD

remained largely unknown. Mirena, similar to copper-bearing

IUDs, consists of a plain plastic T-shaped frame. The

development of peritoneal adhesions in response to a non-

irritating plastic-made foreign body was described by

Echenberg and Ledger (1968). This process involves encase-

ment of the device in delicate peritoneal bands. The plastic

skeleton of Mirena carries a cylindric progestogen reservoir

that contains 52 mg levonorgestrel and is covered by a

polydimethylsiloxane membrane which regulates the release of

levonorgestrel. Local release of levonorgestrel by Mirena

results in very high tissue concentrations, ranging from

470±1500 ng/g wet weight. The endometrial changes seen in

the presence of LNG-IUD are: endometrial gland atrophy,

stromal decidualization, thickened arterial walls and endome-

trial capillary thrombosis. An in¯ammatory reaction involving

neutrophils, lymphocytes, plasma cells and macrophages is

described (Zhu et al., 1989), and focal stromal necrosis may

also occur (Silverberg et al., 1986). The local effect of high

levonorgestrel concentrations on the peritoneum involves a

pseudo-decidual change of swollen cells with rich eosinophilic

Table I. Data of eight patients with uterine perforation by an IUD

Patientno.

Age Lysis ofadhesions

Presenceofperitonealadhesions

Previousoperations

IUDlocationin theperitoneum

Perforationsite

Time ofremoval sinceinsertion indays/months

MainSymptom

Breastfeedingstatus

Time ofinsertionpost-partumin days

Year ofIUDinsertion

IUDtype

CS P G

1 32 + Mild CSx5 Rt gutter Broadligament

3 months Irregularbleeding

+ 75 2001 Mirena 5 5 6

2 32 ± Mild Laparoscopiccholecystectomy

Omentum Uterinefundus

12 days Pain uponinsertion.Irregularbleeding

+ 38 2000 Mirena 0 4 4

3 37 ± Mild Appendectomy Omentum Not seen 7 months None + 60 1999 Mirena 0 4 64 33 ± Mild Splenectomy.

CSx1Pouch ofDouglas

Not seen 2 months Irregularbleeding

± 60 2002 Mirena 1 2 2

5 27 ± Mild Ovariancystectomy.Appendectomy

Omentum Not seen 5 months 8 weekpregnancy

+ 87 1996 CopperIUD

0 2 2

6 33 ± Mild None Omentum Uterinefundus

2 months Mild pelvicpain

+ 80 1997 CopperIUD

0 6 6

7 33 ± None CSx1 Pouch ofDouglas

Uterinefundus

2 days Mild pelvicpain

+ 38 1995 CopperIUD

1 5 8

8 39 ± None None Omentum Not seen 4 months Mild pelvicpain

± 100< 1994 CopperIUD

0 3 3

G = gravida; P = para; CS = Caesarean sections; IUD = intrauterine device.

Peritoneal adhesion potential of intrauterine device

991

by guest on Novem

ber 14, 2015http://hum

rep.oxfordjournals.org/D

ownloaded from

cytoplasm identi®ed as sub-mesothelial cells by immuno-

histochemical studies. So far, the pseudo-decidual change,

also known as ectopic peritoneal decidua, was described as

developing solely in relation to pregnancy, when it was found

at sites such as the sub-mesothelial stroma of the uterus, the

uterine ligaments, adnexae, appendix and omentum (Clement

et al., 1999). No evidence of tissue necrosis was demonstrated

in response to high local levonorgestrel concentrations. A high

rate of progesterone receptivity was disclosed when pelvic

adhesion tissue was evaluated for sex steroid receptors

(Wiczyk et al., 1998). In our study negative staining for

progesterone receptor in the connective tissue in proximity to

levonogestrel-medicated IUD could indicate down-regulation

of the receptor in reaction to local high levels of the

progestogen.

The issue of the effect of progestogens on peritoneal

adhesions is controversial. Progesterone has well-established

anti-in¯ammatory and immunosuppressive properties that

could play a role in the prevention of peritoneal adhesion

formation. Intra-peritoneal instillation of progesterone was

shown to prevent the development of adhesions in pelvic

surgery (Maurer and Bonaventura, 1983). However, other

studies using locally or parenterally administered progestogens

failed to support this claim (Blauer and Collins, 1988; Con®no

et al., 1988). Recently, it was argued that pre-operative use of

medroxyprogesterone acetate prevents post-operative adhe-

sions development (Ustun et al., 1998; Sagol et al., 1999;

Baysal, 2001). Our data demonstrated that high local concen-

trations of progestins neither affect nor dramatically prevent

adhesion formation.

Our results show a lack of difference in the peritoneal cavity

appearance in the presence of either a copper-IUD or a LNG-

IUS. Although no treatment policy should be based on small

case-series reports, in the infrequent event of uterine perfor-

ation by IUD, much is to be learnt from each report. Our

presented cases, in accordance with what has been previously

reported regarding perforating copper-bearing IUDs (Adoni

and Ben Chetrit, 1991; Silva and Larson, 2000; Markovitch

et al., 2002), disclosed only local and delicate peritoneal

adhesions in the presence of both copper-bearing IUDs and

LNG-IUS. Bands of adhesions that could cause bowel

obstruction did not develop. Therefore, the same treatment

mode of a suggested laparoscopy for IUD removal for

symptomatic, medico-legal or psychological reasons, could

®t both copper-bearing and levonorgestrel-medicated IUDs.

From the point of view of peritoneal adhesion formation, given

the risk associated with laparotomy, it may be safer to leave a

perforated LNG-IUS in place than to remove it by laparotomy.

However, our advice is based only on our limited experience

and further data is needed, probably through a multi-centre

study.

References

Adoni, A. and Ben Chetrit, A. (1991) The management of intrauterine devicesfollowing uterine perforation. Contraception, 43, 77±81.

Andersson, K., Ryde-Blomqvist, E., Lindell, K., Odlind, V. and Milsom, I.(1998) Perforations with intrauterine devices. Report from a Swedishsurvey. Contraception, 57, 251±255.

Baysal, B. (2001) Comparison of the resorbable barrier interceed (TC7) andpreoperative use of medroxyprogesterone acetate in postoperative adhesionprevention. Clin. Exp. Obstet. Gynecol., 28, 126±127.

Figure 1. Peritoneal adhesion tissue adjacent to levonorgestrel-releasing IUD. Aggregates of sub-mesothelial stromal cells showingpseudo-decidual changes (white arrows). Note calci®cations (black arrows) between the cells (haematoxylin and eosin, original magni®cation3200).

R.Haimov-Kochman et al.

992

by guest on Novem

ber 14, 2015http://hum

rep.oxfordjournals.org/D

ownloaded from

Blauer, K.L. and Collins, R.L. (1988) The effect of intraperitonealprogesterone on postoperative adhesion formation in rabbits. Fertil.Steril., 49, 144±149.

Clement, P.B., Young, R.H. and Scully, R.E. (1999) The Peritoneum. InSternberg, S.C. (ed) Diagnostic Surgical Pathology, Vol. 2, 3rd edn.Lippincott Williams and Wilkins, USA. p. 2437.

Con®no, E., Friberg, J., Vermesh, M., Thomas, W. and Gleicher, N. (1988)Effects of progesterone on postoperative adhesion formation inhysterectomized rabbits. Int. J. Fertil., 33, 139±142.

Echenberg, R. and Ledger, W.J. (1968) Peritoneal response to polyethyleneforeign bodies. Obstet. Gynecol., 31, 795±798.

Markovitch, O., Klein, Z., Gidoni, Y., Holzinger, M. and Beyth, Y. (2002)Extrauterine mislocated IUD: is surgical removal mandatory?Contraception, 66, 105±108.

Maurer, J.H. and Bonaventura, L.M. (1983) The effect of aqueousprogesterone on operative adhesion formation. Fertil. Steril., 39, 485±489.

Sagol, S., Ozsener, S., Dincer, O. Yilmaz, H. and Karadadas, N. (1999) Theeffect of medroxyprogesterone acetate and heparin in the prevention ofpostsurgical adhesion formation in the rat uterine model. J. Obstet. Gynecol.Res., 25, 287±293.

Silva, P.D. and Larson, K.M. (2000) Laparoscopic removal of a perforated

intrauterine device from the perirectal fat. J. Soc. Laparosc. Surg., 4,159±162.

Silverberg, S.G., Haukkamaa, M., Arko, H., Nilsson, C.G. and Luukkainen, T.(1986) Endometrial morphology during long-term use of levonorgestrel-releasing intrauterine devices. Int. J. Gynecol. Pathol., 5, 235±241.

Ustun, C., Yanik, F.F., Kocak, I., Canbaz, M.A. and Cayli, R. (1998) Effectsof Ringer's lactate, medroxyprogesterone acetate, gonadotropin-releasinghormone analogue and its diluent on the prevention of postsurgical adhesionformation in rat models. Gynecol. Obstet. Invest., 46, 202±205.

Wiczyk, H.P., Grow, D.R., Adams, L.A., O'Shea, D.L. and Reece, M.T.(1998) Pelvic adhesions contain sex steroid receptors and produceangiogenesis growth factors. Fertil. Steril., 69, 511±516.

Zhu, P.D., Luo, H., Xu, R.H., Cheng, J., Wu, S.C., Chen, J.H., Wu, M.Z. andWang, X.P. (1989) The effect of intrauterine devices, the stainless steel ring,the copper T220, and releasing levonorgestrel, on the bleeding pro®le andthe morphological structure of the human endometrium±a comparativestudy of three IUDs. A morphometric study of 96 cases. Contraception, 40,425±438.

Submitted on September 13, 2002; resubmitted on December 12, 2002;accepted on January 14, 2003

Peritoneal adhesion potential of intrauterine device

993

by guest on Novem

ber 14, 2015http://hum

rep.oxfordjournals.org/D

ownloaded from