Uterine perforation caused by intrauterine devices: clinical course and treatment
Advanced Access publication on March
Uterine perforation caused by intrauterine devices: clinical course and treatment
Janina Kaislasuo 2
Satu Suhonen 1
Mika Gissler 0 3
Pekka La¨ hteenma¨ ki 2
Oskari Heikinheimo 2
0 National Institute for Health and Welfare , P.O. Box 30, 00271 Helsinki , Finland
1 Centralized Family Planning, Department of Social Services and Health Care, City of Helsinki , P.O. Box 6440, 00099 Helsinki , Finland
2 Department of Obstetrics and Gynaecology/Ka ̈tilo ̈ opisto Hospital, University of Helsinki and Helsinki University Central Hospital , P.O. Box 610, SF-00029 HUS Helsinki , Finland
3 Nordic School of Public Health , Box 12133, 402 42 Gothenburg , Sweden
study questions: What are the symptoms of uterine perforation caused by modern copper intrauterine devices (Cu-IUDs) and the levonorgestrel-releasing intrauterine system (LNG-IUS); how is perforation detected and what are the findings in abdominal surgery? summary answer: Symptoms are mostly mild and 30% of women are asymptomatic. Surgical findings are mainly minimal; no visceral complications were found in this study. However, adhesions as well as pregnancies seem to be more common among women using Cu-IUDs. what is known already: Prior studies and case reports have suggested that uterine perforation by modern IUDs/IUSs is rarely serious. study design, size, duration: A retrospective study of 75 patients (54 LNG-IUS and 21 Cu-IUD) treated surgically for uterine perforation between 1996 and 2009. participants/materials, setting, methods: The patients treated for uterine perforation by an IUD/IUS at clinics of the Helsinki and Uusimaa Hospital District were identified using the National Care Register for Health Institutions in Finland. The clinical data were collected from individual patient records. main results and the role of chance: The majority of patients (n ¼ 53; 71%) had mild symptoms of abnormal bleeding or abdominal pain or both, in combination with missing IUD/IUS threads. Asymptomatic patients (n ¼ 22; 29%) were examined because of missing threads or pregnancy. Failure to remove the IUD/IUS by pulling visible threads was the reason for referral in seven women (9%) requesting removal of the device. Eleven women (15%) were pregnant. Misplaced IUDs/IUSs were localized by a combination of vaginal ultrasonography (US) and X-ray, hysteroscopy or curettage. Only after this were patients treated by means of laparoscopy. The majority (n ¼ 44; 65%) of the 68 intra-abdominal devices were located in the omentum, the remaining 24 (35%) around the uterus. Partial perforation or myometrial embedding was diagnosed in all seven cases (9%) with visible threads, but unsuccessful removal by pulling. During laparoscopy, filmy adhesions were found in 21 patients (30%). Pregnancy (33 versus 7%, P ¼ 0.009) and intra-abdominal adhesions (58 versus 20%, P ¼ 0.002) were significantly more common in the Cu-IUD group. Infections were rare; one non-specific acute abdominal infection, later found to be unrelated to the IUD, led to laparoscopy and in four cases the IUD was surrounded by pus, but there were no symptoms of infection. limitations, reasons for caution: The study setting revealed only surgically treated symptomatic patients and asymptomatic women attending regular follow-up. Women not treated, but only followed or not attending follow-up, were not identified, excluding the possibility to analyse missed undiagnosed perforations, or conservative follow-up as a treatment option. wider implications of the findings: As surgical findings are minimal, asymptomatic women may need no treatment at all. An alternative form of contraception is, however, important as pregnancies do occur. If a woman plans a pregnancy, a misplaced LNG-IUS should be removed, as it may act as a contraceptive. study funding/competing interest(s): Helsinki University Central Hospital research funds. O.H. has lectured and designed educational events with Bayer AG and MSD, and serves occasionally on scientific advisory boards for these companies. S.S. has
lectured in educational events organized by MSD Finland (part of Merck & Co. Inc.) and served on the advisory board for contraception of
this company. The other authors have no conflicts of interest to declare.
Globally, intrauterine devices (IUDs) are a major form of contraception.
With 14% of women worldwide using an IUD, it is the second most
popular form of contraception after female sterilization (World
Population Bureau, 2008). In addition to efficient contraception, the
levonorgestrel-releasing intrauterine system (LNG-IUS) is a treatment
of choice for menorrhagia as well as dysmenorrhoea and as part of
hormone replacement therapy (Andersson et al., 1992; Milsom, 2007;
Heikinheimo and Gemzell-Daniselsson, 2012). Cu-IUDs are also still
widely used and are of interest to women wishing to use non-hormonal
In Finland, 23% of fertile women use IUDs (Olivieira da Silva et al.,
2011). After the introduction of the LNG-IUS in the 1990s, all IUD use
has increased significantly in Finland. Currently 85% of inserted IUDs in
Finland are LNG-IUSs. The remaining 15% of inserted devices are
copper IUDs (Cu-IUDs; Bayer AG, Berlin, Germany, 2010).
A rare complication of IUD contraception is uterine perforation.
Known associated factors are the post-partum period and lactation
as well as an inadequate insertion technique (Heinemann et al.,
2011; Kaislasuo et al., 2012). Previously reported perforation
incidence rates are 0.4 – 2.2/1000 insertions with Cu-IUDs and 0.68 –
2.6/1000 insertions with the LNG-IUS (Andersson et al., 1998;
Caliskan et al., 2003; van Maudenhoven et al., 2006; Heinemann
et al., 2011). Our own recent study found an incidence rate of
0.4/1000 insertions for both types of devices (Kaislasuo et al., 2012).
The accepted treatment for IUD-associated perforations has been
abdominal surgery, initially via laparotomy and, as surgical techniques
have developed, via laparoscopy. Case reports show that misplaced
IUDs have caused bowel perforations and adhesions leading to
peritonitis (Zakin et al., 1981; Gill et al., 2012). However, as many
investigators have found minimal or no adhesions with both modern
non-irritating polyethylene-framed Cu-IUDs and the LNG-IUS, the
need to remove intra-abdominal IUDs in asymptomatic cases has
repeatedly been questioned (Adoni and Ben Chetrit, 1991; Markovitch
et al., 2002; Haimov-Kohman et al., 2003b). Rare bowel complications
or widespread adhesions do cause symptoms and these need to be
treated. In contrast to life-threatening symptoms described in case
reports (Zakin et al., 1981), larger studies concerning IUD-associated
perforations indicate that the majority of perforations are either
asymptomatic or associated with mild symptoms such as abnormal
bleeding or mild pain or both, combined with missing threads or
unintended pregnancy (Heinemann et al., 2011; van Grootheest
et al., 2011).
The purpose of this retrospective population-based study was to
examine the clinical course and treatment of patients with uterine
perforation caused by modern Cu-IUDs and the LNG-IUS by
assessing symptoms, method of detection, type of surgical treatment and
Materials and Methods
Patients were identified in the National Care Register for Health
Institutions in Finland, later called the Hospital Register. The register contains
ICD-10 (International Classification of Diseases, 2010) and operation
codes on all patients treated at Finnish hospitals. As there is no specific
ICD-10 code for IUD/IUS-associated perforation, women with potential
uterine perforation were identified in the Hospital Register by using the
operation codes of the Nordic Medico-Statistical Committees
(NOMESCO) Classification of Surgical Procedures, available since 1997,
for removal of an intra-abdominal or intrauterine foreign object
(JAL10-JAL11, JAL20-JAL22), combined with the ICD-10 codes
concerning gynaecological procedures related to IUD/IUS complications (T83.3,
T19.3), insertion and follow-up (Z30). As ICD-10 codes were introduced
in Finland in 1996, women treated from 1 January 1996 to 31 December
2009 were included in the analysis. Combining diagnostic and operation
codes gave us 3909 patients. By excluding codes unrelated to possible
perforation, as well as male patients, 370 patients with probable surgically
treated uterine perforation were found nationwide. Records of all 108
patients treated at the clinics of the hospital district of Helsinki and
Uusimaa were examined by J.K., resulting in identification of 78 patients
with surgically treated uterine perforations and 30 patients treated for
other IUD/IUS-related reasons. As the type of IUD could not be
determined in three cases, 75 patients were included in the study. All data
used in the study were derived from hospital records covering referral
for examination and treatment, background information on the patient
as well as reports on findings and treatment.
The Ethics Committee of the Hospital District of Helsinki and Uusimaa
gave a positive statement concerning the study. The Ministry of Social
Affairs and Health and the National Institute for Health and Welfare
gave permission to use the nationwide register and medical record data
for the research. Final permission for the study was granted by the
Hospital District of Helsinki and Uusimaa.
Data analysis was performed using the Predictive Analysis Software
(PASW18; SPSS Inc., Chicago, IL, USA). The subjects were analysed
both as one group as well as in separate groups by type of device
(LNG-IUS versus Cu-IUD) using the Mann – Whitney, x 2 and Fisher’s
exact tests as appropriate. Statistical significance was set at P ≤ 0.05.
Patient demographics grouped by the type of device is shown in
Table I. Of the 75 patients, 54 (72%) used an LNG-IUS and 21
(28%) a Cu-IUD. No statistically significant differences emerged
between the two groups in the various demographic characteristics.
History of vaginal delivery (VD)
Yes 48 (89)
VD/woman, median 2 (0 – 6)
No 6 (11)
History of Caesarean section (CS)
Yes 13 (24)
CS/woman, median 0 (0 – 3)
No 41 (76)
History of miscarriage (M)
Yes 12 (22)
M/woman, median 0 (0 – 3)
No 42 (78)
History of termination of pregnancy (TOP)
Yes 13 (24)
TOP/woman, median 0 (0 – 2)
No 41 (76)
Previous IUD use
Yes 10 (19)
No 10 (19)
Data not available 34
History of abdominal surgery (incl. curettage)
Yes 29 (54)
No 25 (46)
Endometriosis 3 (6)
Salpingectomy (ectopic 0
Curettage 13 (24)
Multiple curettages 5/13 (38)
Leep treatment 5 (9)
Gastrointestinal surgery 4 (7)
The data are presented as n (%) unless stated otherwise.
aFincoid 1, FlexiT 1, NovaT380Ag 13, unspecified type of Cu-IUD 6.
Table I Demographics of patients with surgically
removed perforating IUDs/IUSs.
Time from insertion to diagnosis and
The median time from insertion to diagnosis was 5 months (range 0
days to 69 months). Although 26 (49%) patients experienced
immediate symptoms (,5 days) after insertion, only 12 of them (46%) sought
treatment within a month (median 53 days, range 0 days to 15
months). The manner of presentation against time from insertion to
diagnosis is shown in Table II. After diagnosis, most patients were
treated without delay, the median time being 21 days (range 0 days
to 16 months).
The majority of women (n ¼ 53; 71%) were symptomatic and sought
treatment (Fig. 1). Symptoms were mainly mild, either abnormal
bleeding or mild lower abdominal pain, or both. Asymptomatic
patients (n ¼ 22; 29%) were diagnosed during routine follow-up
visits (n ¼ 16) or because of unintended pregnancy (n ¼ 6). One
intra-abdominal IUD was found during a hysterectomy performed
for other reasons; the patient had forgotten the presence of the IUD.
The onset of symptoms was precisely defined for only 28 (53%) of
the symptomatic patients, representing 37% of all women in the study.
Symptoms were immediate in the majority of these cases: within 24 h
in 21 patients (75%) and within 1 – 5 days for 5 patients (18%). All 26
patients with onset of symptoms within 5 days presented with pain; in
6 cases combined with bleeding. The only patient with symptoms of
infection had had the IUD inserted 4 years prior to symptoms and
presented with an unintended pregnancy.
Altogether, there were 11 pregnancies (15%), 10 intrauterine and 1
tubal. Pregnancies occurred more often following perforation with a
Cu-IUD (n ¼ 7/21; 33%) than with the LNG-IUS (4/54; 7%, P ¼
0.009; Fig. 1). The patient with a tubal pregnancy used a LNG-IUS
and she underwent laparoscopic surgery as a result of acute
intra-abdominal bleeding. Five pregnant women (45%) experienced
symptoms of pain and/or bleeding. While two women with
Cu-IUDs continued their pregnancies successfully, seven (Cu-IUD ¼
4, LNG-IUS ¼ 3) chose termination of pregnancy and one patient
with a Cu-IUD had a miscarriage. Ten IUDs in the pregnant women
were located intra-abdominally and one Cu-IUD intramurally.
The main reason for referral was missing threads (n ¼ 50; 67%); for 24
patients (32%) this was the only finding. Other reasons were abnormal
bleeding and/or abdominal pain (n ¼ 32; 43%) and pregnancy (n ¼
11; 15%). Seven (9%) intramural or partially perforated devices
could not be removed by pulling visible threads. All these women
requested the removal of the device; three were asymptomatic, two
experienced disturbing bleeding and two bleeding and pain. Diagnosis
was made by a specialist in all cases, either at the treating hospital (n ¼
64; 85%) or by a private gynaecologist with access to vaginal US (n ¼
11; 15%). Personnel at primary health care or family planning clinics
Table II Manner of presentation for women with known time of insertion (n 5 59).
Figure 1 Clinical symptoms of the patients [n ¼ 75 (LNG-IUS 54,
had suspected uterine perforation at the time of referral in 12 patients
The primary diagnostic examination was commonly vaginal US (n ¼
70; 93%) and if not sufficient for diagnosis, the secondary examination
was most often abdominal X-ray (n ¼ 57; 76%). If needed,
hysteroscopy (n ¼ 20; 27%) or curettage (n ¼ 8; 11%) was used for both
diagnosis and treatment. Curettage was performed in six pregnant
patients for termination of pregnancy and in two patients with
abnormal bleeding. Only after at least one of the above mentioned
diagnostic examinations was laparoscopy the treatment method.
Computerized tomography (CT) was used once, as US and X-ray
failed to give a clear location of an IUD.
Location of IUD, treatment and adhesions
The majority of patients (n ¼ 70; 93%) underwent abdominal surgery;
69 patients underwent laparoscopy and 1 pregnant patient had the
device removed during Caesarean section. In addition to the 68
intra-abdominal IUDs, two partially perforating devices were removed
by laparoscopy. All intramurally embedded IUDs and one partially
perforating IUD were removed by means of hysteroscopy (n ¼ 5; 7%). The
locations of the removed devices are shown in Table III.
Surgical findings were mainly minimal. No symptomatic IUD-related
infections were diagnosed, but in four Cu-IUD cases the device
located intra-abdominally was embedded in pus. One pregnant
patient with a LNG-IUS showed symptoms of infection, which were
found to be unrelated to the IUS in laparoscopy. In addition, in two
women small haematomas and in two women areas of uterine
erythema related to acute perforation were found. These four
women with LNG-IUS perforation suffered acute pain at insertion.
The majority of patients had no intra-abdominal adhesions
(Table III). Adhesions were observed in 7 (35%) of the asymptomatic
women and in 14 (28%) of the symptomatic women. Adhesions
occurred more often following perforation with Cu-IUDs (11/19,
58%) versus LNG-IUSs (10/51; 20%, P ¼ 0.002). In addition,
adhesions in LNG-IUS patients were more filmy and small and in 4 out
of 10 cases (40%) unrelated to the location of the device. In contrast,
in all patients with Cu-IUD perforation and adhesions the device was
embedded in adhesions and the adhesions were denser. The location
of the device (omental versus pelvic) did not statistically correlate with
adhesion formation (P ¼ 0.54). Adhesions found at surgery were most
common in women treated 1 – 6 months after insertion. Table IV
shows adhesions in correlation with time from insertion to diagnosis.
The majority (n ¼ 53; 71%) of the women were treated as
daycases. However, 11 women (15%) stayed 2 days because of hospital
policy and 10 (13%) stayed 3 – 7 days because of suspected infection
(n ¼ 3) or acute abdominal pain shortly after insertion (n ¼ 7). A
total of 23 patients (31%) underwent multiple procedures;
hysteroscopy or curettage prior to laparoscopy. The procedure was
performed simultaneously in 11 patients (48%), while 12 patients (52%)
spent two separate days at the hospital.
We found that uterine perforation with modern devices is rarely
dangerous, as reported in recent publications (van Grootheest et al.,
2011; Gill et al., 2012). The majority of women experienced
symptoms resulting in a consultation. The most common symptom was
abdominal pain with or without abnormal bleeding, but one-third of the
women were asymptomatic. Although half of the symptomatic
patients experienced immediate symptoms, many waited before
seeking treatment. This reflects the fact that women are usually
counselled concerning post-insertion abdominal pain lasting a few days.
After the diagnosis, one-third waited for a procedure for more than
1 month, reflecting the fact that symptoms often were neither acute
The patients were identified in the National Care Register, using the
operative codes for removal of an intra-abdominal or intrauterine
foreign body. We assume that the majority of patients undergoing
abdominal surgery in Finland were identified during the study period of
15 years, as these types of procedures have their own surgical
codes. In this study, however, we found mainly women treated for
suspected perforation and attending follow-ups, suspected pregnancy
Table III Findings at laparoscopic surgery.
Table IV Intra-abdominal adhesions at laparoscopy
in women with known time of insertion.
or unsuccessful removal of an IUD/IUS. Women not treated, but only
followed, were not identified, excluding the possibility to analyse
conservative follow-up as a treatment option.
The diagnosis of IUD/IUS perforation is usually straightforward
after suspicion of a misplaced device arises. A combination of
vaginal US and abdominal X-ray is usually sufficient to diagnose a
perforation. In the present study, vaginal US was used in nearly all cases,
and abdominal X-ray was needed for diagnosis in 75% of the cases. If
vaginal US is unavailable, X-ray makes it possible to differentiate
between a perforation and an expulsion when threads are not
found and uterine perforation is suspected. In hospital settings, CT
or magnetic resonance imaging is good alternative for further
preoperative examination if suspicion of visceral complication arises
(Boortz et al., 2012).
The traditional form of treatment has been laparotomy, but with
developing surgical techniques, laparoscopy, being less invasive and
more safe, is nowadays the treatment of choice. In our study, all patients
with an intra-abdominal device could successfully be treated by
laparoscopy, including those with adhesions. All adhesions, either
IUD-associated or clearly non-related, were local and occurred more often
with a Cu-IUD. Although adhesions were found in 30% of the
laparoscopies, none led to laparotomy, unlike in the large case series reviewed
by Gill et al. (2012), where adhesions were a common cause of
laparotomy. Similarly, we found no visceral complications caused by the
perforating IUDs/IUSs and no correlation between symptoms and
intra-abdominal adhesions in surgery. Although the groups were small,
adhesions were surprisingly most common between 1 and 6 months
from insertion. It is, however, impossible to predict when these had
formed. It is widely accepted that adhesion formation starts immediately
after surgery (Brokelman et al., 2011). Partially perforating or embedded
devices could be removed by minimally invasive hysteroscopy, a
procedure significantly easier for both the patient and the physician.
Both a misplaced Cu-IUD and a LNG-IUS increase the risk of
contraceptive failure. However, we found a significant difference in pregnancy
rates between women using a Cu-IUD versus a LNG-IUS. This may
reflect the different mechanisms of action of the devices. The
contraceptive effect of Cu-IUDs relies on local sterile endometrial inflammation as
well as on the toxic effect of Cu ions on spermatozoa (Ortiz and
Croxatto, 2007). These mechanisms may also explain why adhesions
occurred more often in connection with Cu-IUDs. The absence of these
uterine reactions in cases of perforating Cu-IUDs increases the risk of
unplanned pregnancy. In contrast, the mechanism of action of the
LNG-IUS is suppression of the endometrium and thickening of the
cervical mucus (Luukkainen et al., 2001; Lewis et al., 2010). With the
LNG-IUS in utero, plasma levels of LNG are low (Nilsson et al., 1986)
and changes in ovarian function are minimal (Luukkainen et al., 2001).
An intraperitoneally displaced LNG-IUS, however, results in circulating
levels of LNG similar to those seen during the use of combined oral
contraceptives (Haimov-Kohman et al., 2003a). It has been suggested that
these levels are high enough to prevent ovulation and thus pregnancy,
as well as adhesion formation, owing to the anti-inflammatory properties
of LNG (Haimov-Kohman et al., 2003b). Nevertheless, 7% of the
patients with misplaced LNG-IUSs were pregnant at the time when
the perforation was diagnosed.
The majority of women with IUD/IUS perforation experience mild
symptoms of abnormal bleeding or mild lower abdominal pain, or
both. These symptoms or missing threads in asymptomatic or pregnant
women using an IUD/IUS should lead to suspicion of perforation. The
diagnosis is easily made by way of vaginal US and/or abdominal X-ray.
The method of treatment is laparoscopy. Surgical findings are mainly
minimal, with possible adhesions mostly being filmy or local and
treatable by way of modern laparoscopic techniques. Adhesions as well as
pregnancies are more common in connection with Cu-IUDs, which
might be explained by the different mechanisms of action of Cu-IUDs
versus the LNG-IUS. Asymptomatic women may need no treatment
at all if the risk of pregnancy is low or contraception is not needed. In
fertile-aged, sexually active women an additional form of contraception
is important, as unplanned pregnancies occur, especially in association
with misplaced Cu-IUDs. However, further studies based on long-term
data are needed to assure the safety of this treatment option. In the case
of expectant management detailed patient information is essential and
the decision-making must be shared with the woman.
Stiftelsen Dorothea Olivia, Karl Walter och Jarl Walter Perkle´ns Minne
and Nylands Nation are gratefully acknowledged.
All authors participated equally in the study design, analysis of the
data, manuscript drafting and critical discussion.
Helsinki University Central Hospital research funds.
Conflict of interest
O.H. has lectured and designed educational events with Bayer AG and
MSD, and serves occasionally on scientific advisory boards for these
companies. S.S. has lectured in educational events organized by
MSD Finland (part of Merck & Co. Inc.) and served on the advisory
board for contraception of this company. The other authors have
no conflicts of interest to declare.
Adoni A , Ben Chetrit A. The management of intrauterine devices following uterine perforation . Contraception 1991 ; 43 : 77 - 81 .
Andersson K , Mattson LK , Rybo G , Stadberg E. Intrauterine release of levonorgestrel-a new way of adding progestogen in hormone replacement therapy . Obstet Gynecol 1992 ; 79 : 963 - 967 .
Andersson K , Ryde-Blomqvist E , Lindell K , Odlind V , Milsom I. Perforations with intrauterine devices . Contraception 1998 ; 57 : 251 - 255 .
Boortz HE , Margolis DJA , Ragavendra N , Patel MK , Kadell BM . Migration of intrauterine devices: radiologic findings and implications for patients care . Radiographics 2012 ; 32 : 335 - 352 .
Brokelman W , Lensvelt M , Rinkel I , Klinkenbijl J , Rejnen M. Peritoneal changes due to laparoscopic surgery . Surg Endosc 2011 ; 25 : 1 - 9 .
Caliskan E , O¨ ztu ¨rk N , Dilbaz O¨ , Dilbaz S. Analysis of risk factors associated with uterine perforation by intrauterine devices . Eur J Contracept Reprod Health Care 2003 ; 8 : 150 - 155 .
Gill R , Mok D , Hudson M , Shi X , Birch D , Karmal S. Laparoscopic removal of an intra-abdominal intrauterine device: case and systematic review . Contraception 2012 ; 85 : 15 - 18 .
Haimov-Kohman R , Amsalem H , Adoni A , Lavy Y , Spitz I. Management of a perforated levonorgestrel-medicated intrauterine device-a pharmacokinetic study: case report . Hum Reprod 2003a ; 18 : 1231 - 1233 .
Haimov-Kohman R , Doviner V , Amsalem H , Prus D , Adoni A , Lavy Y. Intraperitoneal levonorgestrel-releasing intrauterine device following uterine perforation: the role of progestins in adhesion formation . Hum Reprod 2003b ; 18 : 990 - 993 .
Heikinheimo O , Gemzell-Daniselsson K. Emerging indications for the levonorgestrel-releasing intrauterine system (LNG-IUS) . Acta Obstet Gynecol Scand 2012 ; 91 : 3 - 9 .
Heinemann K , Barnett C , Dinger J. Risk Factors for Uterine Perforations during IUD Insertion: Interim Results from the EURAS-IUD Study 2011 . http://www.pharmacoepi.org/meetings/27thconf/presentations/ Risk%20Factors% 20for%20Uterine%20Perforations%20during%20IUD% 20Insertion-%20Interim%20Results%20from%20the%20EURAS-IUD% 20Study . pdf (29 January 2013 , date last accessed) .
International Classification of Diseases (ICD) . World Health Organization (WHO) 2010 : http://apps.who.int/classifications/icd10/browse/2010/ en (29 January 2013 , date last accessed) .
Kaislasuo J , Suhonen S , Gissler M , La¨hteenma¨ki P, Heikinheimo O. Intrauterine contraception: incidence and factors associated with uterine perforation-a population-based study . Hum Reprod 2012 ; 27 : 2658 - 2663 .
Lewis R , Taylor D , Natavio M , Melamed A , Felix J , Mishell D Jr . Effects of the levonorgestrel-releasing intrauterine system on cervical mucus quality and sperm penetrability . Contraception 2010 ; 82 : 491 - 496 .
Luukkainen T , Pakarinen P , Toivonen J. Progestin-releasing intrauterine systems . Semin Reprod Med 2001 ; 19 : 355 - 363 .
Markovitch O , Klein Z , Gidoni Y , Holzinger M , Beyth Y. Extrauterine mislocated IUD: is surgical removal mandatory? Contraception 2002 ; 66 : 105 - 108 .
Milsom I. The levonorgestrel-releasing intrauterine system as an alternative to hysterectomy in peri-menopausal women . Contraception 2007 ; 75 : 152 - 154 .
Nilsson CG , La¨hteenma¨ki P , Luukkainen T , Robertson D. Sustained intrauterine release of levonorgestrel over five years . Fertil Steril 1986 ; 33 : 139 - 148 .
Nordic Medico-Statistical Committee (NOMESCO). Classification of Surgical Procedures. Version 1.15 2011 no. 93:2010 Copenhagen, 2010 .
Olivieira da Silva M , Jahn A , Olsen J , Karro H , Temmerman M , Gissler M , Bloemenkamp K , Fronteira I. The Reproductive Health Report-the state of sexual and reproductive health within the European Union . Eur J Contracept Reprod Health Care 2011 ; 16 (S1): S1 - S70 .
Ortiz M , Croxatto H. Copper - T intrauterine device and levonorgestrel intrauterine system: biological bases of their mechanism of action . Contraception 2007 ; 75 : 16 - 30 .
van Grootheest K , Sachs B , Harrison-Woolrych M , Caduff-Janosa P , van Puijenbroek E. Uterine perforation with the levonorgestrel-releasing intrauterine device: analysis of reports from four national pharmacovigilance centres . Drug Saf 2011 ; 34 : 83 - 88 .
van Haudenhoven K , van Kaam KJAF , van Grootheest AC , Salemans THB , Dunselman GAJ . Uterine perforation in women using a levonorgestrelreleasing intrauterine system . Contraception 2006 ; 73 : 257 - 260 .
World Population Bureau . Family Planning Worldwide , 2008 Data Sheet. http://www.prb. org/pdf08/fpds08.pdf (29 January 2013 , date last accessed) .
Zakin D , Stern WZ , Rosenblatt R. Complete and partial uterine perforation and embedding following insertion of intrauterine devices I. Classification, complications, mechanism, incidence and missing strings . Obstet Gynecol Surv 1981 ; 36 : 335 - 353 .