Contraception 87 (2013) 655 – 660

Review article

Retention of intrauterine devices in women who acquire pelvic

inflammatory disease: a systematic review☆
Naomi K. Tepper a,⁎, Maria W. Steenland a , Mary E. Gaffield b ,
Polly A. Marchbanks a , Kathryn M. Curtis a
a
Division of Reproductive Health, Centers for Disease Control and Prevention, Atlanta, GA 30341, USA
b
Department of Reproductive Health and Research, World Health Organization, CH-1211 Geneva 27, Switzerland
Received 6 July 2012; accepted 7 August 2012

Abstract

Background: Women using intrauterine devices (IUDs) are not protected against acquiring pelvic inflammatory disease (PID). If a woman
has an IUD in place when she is diagnosed with PID, there is a theoretical concern that presence of an IUD might impact the course of
treatment. This review was conducted to evaluate the evidence regarding whether an IUD should be retained or removed if a woman
develops PID.
Study Design: The PubMed database was searched from database inception through April 2012 for all peer-reviewed articles in any language
concerning PID in women using IUDs. Articles were included if they examined women with IUDs who developed PID and compared the
clinical course of women in whom the IUD was retained versus women in whom the IUD was removed. Articles were excluded if the
infection was diagnosed before or at the time of IUD insertion. The quality of each study was assessed using the United States Preventive
Services Task Force grading system.
Results: Four fair-quality studies met inclusion criteria for this review. One randomized controlled trial showed that women with IUDs
removed had longer hospitalizations than those with IUD retention (15% versus 4%, pb.01), although there were no differences in PID
recurrences or subsequent pregnancies. Another randomized controlled trial showed no differences in laboratory parameters among women
who retained the IUD when compared with women in whom the IUD was removed. One prospective cohort study showed that there were no
differences in clinical or laboratory parameters during hospitalization; however, the IUD removal group had a higher proportion hospitalized
for more than 2 weeks compared with the IUD retention group (33% versus 19%, pb.05). One randomized controlled trial showed that
women who had the IUD removed experienced improved recovery in most clinical signs and symptoms compared with women who retained
the IUD.
Conclusions: Three fair-quality studies showed no difference in clinical or laboratory outcomes among women who retained IUDs when
compared with women who had IUDs removed, and two of these studies showed that women who had IUDs removed had longer
hospitalizations. In contrast, one fair quality study showed improved clinical signs and symptoms among women who had IUDs removed.
Overall, women who retained their IUDs had similar or better outcomes than women who had their IUDs removed.
Published by Elsevier Inc.

Keywords: Intrauterine device; Pelvic inflammatory disease; Systematic review

1. Introduction

Intrauterine devices (IUDs) are safe, long-acting and

highly effective methods of contraception. However, they do
☆
Disclaimer: The findings and conclusions in this article are those of not provide protection against pelvic inflammatory disease
the authors and do not necessarily represent the official position of the (PID). Sequelae of PID can include ectopic pregnancy, tubal
Centers for Disease Control and Prevention or the World Health
Organization.
factor infertility and chronic pelvic pain [1]. Prompt
⁎ Corresponding author. Tel.: +1 770 488 6506; fax: +1 770 488 6391. antimicrobial treatment is important to minimize the negative
E-mail address: ntepper@cdc.gov (N.K. Tepper). sequelae of pelvic infections. If a woman has an IUD in place
0010-7824/$ – see front matter. Published by Elsevier Inc.
http://dx.doi.org/10.1016/j.contraception.2012.08.011

Downloaded for Mahasiswa FK UKDW 01 (mhsfkdw01@gmail.com) at Duta Wacana Christian University from ClinicalKey.com by Elsevier on September 20, 2019.
For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.
 656
Downloaded for Mahasiswa FK UKDW 01 (mhsfkdw01@gmail.com) at Duta Wacana Christian University from ClinicalKey.com by Elsevier on September 20, 2019.
For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.

Table 1
Evidence for removal versus retention of an IUD in women who develop PID
Author, year, Study design Population IUD type(s) Outcomes Results Strengths Weaknesses Quality
location
Larsson and Randomized 928 women admitted to Cu IUD Percentage with Number and percentage with 1, 2 or ≥3 weeks Long follow-up for Randomization for IUD I, fair
Wennergren [6], controlled trial hospital with PID divided 1, 2 and ≥3 weeks hospital stay: some women removal not described

N.K. Tepper et al. / Contraception 87 (2013) 655–660

1977 Sweden into 3 groups: hospital stays 1 wk 2 wks ≥3 wks Performed intention- Distribution of covariates at
Group I: no IUD, N=632 PID recurrence to-treat analysis baseline not reported
Group I 276 (46%) 313 (50%) 43 (7%)
Group II: Cu IUD left in and subsequent Small number of women
Group II 87 (37%) 139 (59%) 10 (4%)*
place, N=236 pregnancies followed for 18 months
Group III 24 (40%) 27 (45%) 9 (15%)
Group III: Cu IUD No description of women
removed on admission, * pb.01 when compared with Group III. who were followed for
randomly selected, N=60 18 months
184 women followed for Number and percentage with recurrence P-value not stated for
18 months for recurrence and subsequent intrauterine pregnancies: comparison of recurrence
and pregnancy rates: Recurrences Pregnancies and pregnancies
Group I: no IUD, N=121 between groups
Group I 11 (9%) 15 (12%)*
Group II: Cu IUD left
Group II 4 (9%) 0
in situ, N=46
Group III 2 (12%) 3 (18%)*
Group III: Cu IUD
removed before treatment, *Percentage calculated by systematic
N=17 review authors, differs from percentage
PID diagnosis: palpable reported by study authors.
tender adnexal mass,
ESR ≥15 mm/h, lower
abdominal pain, fever
All women treated
with antibiotics
Soderberg and Randomized 53 women with IUD Cu IUD, Mean days from Mean days from admission to peak ESR: Use of objective criteria Randomization not described I, fair
Lindgren [3], controlled trial hospitalized for acute Plastic IUD, admission to peak Mean days (SD) Performed intention- Distribution of covariates at
1981 Sweden salpingitis randomized Unknown ESR to-treat analysis baseline not reported
Group A 3.0 (1.8)
to 2 groups (7 left study): Mean days from (with the exception of age)
Group B 3.2 (1.7)
Group A: IUD left in peak ESR to half Low and comparable attrition
place, N=23 of peak value rate between groups
Group B: IUD removed Treatment failures Mean days from peak ESR to half of peak value: Clinical course not reported
before treatment, N=23 or rehospitalizations Mean days (SD) Small number of women
PID diagnosis: acute within 3 months
Group A 8.2 (3.5)
pelvic pain, adnexal
Group B 8.0 (2.8)
tenderness, increased ESR
during hospitalization
≥40 mm/h No treatment failures or rehospitalizations within
All women treated with 3 months in either group
pivampicillin/doxycycline
 Teisala [4], 1989 Prospective 186 women with IUD Cu IUD, Clinical findings, Clinical and laboratory findings during Use of objective criteria No description of how II-2,
Finland cohort hospitalized for acute PID: Lippes loop, ESR, duration of hospitalization: groups determined fair
Downloaded for Mahasiswa FK UKDW 01 (mhsfkdw01@gmail.com) at Duta Wacana Christian University from ClinicalKey.com by Elsevier on September 20, 2019.

Group 1: IUD left in place, Dalkon shield, hospitalization Group 1 Group 2 No follow-up after
N=105 Saf-T-Coil hospitalization
Days of fever, 2.0 (1.4) 2.2 (1.8)
Group 2: IUD removed on Did not adjust for
mean (SD)
admission, N=81 potential confounders
Highest ESR 52.0 (24.7) 51.4 (23.9)
PID diagnosis: acute
(mm/h),
pelvic pain, uterine and
mean (SD)
adnexal tenderness,
ESR (mm/h) on 28.3 (11.7) 26.4 (10.1)
increased ESR N15 mm/h,
discharge
fever
For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.

Days of 10.8 (5.2) 12.7 (6.6)

All women treated with
hospitalization,
penicillin ± metronidazole
mean (SD)
(17 women in each group
Hospitalized 19 33
received treatment before
for N2 weeks,
admission)
percentage*
* pb.05

N.K. Tepper et al. / Contraception 87 (2013) 655–660

Altunyurt et al. [5], Randomized 138 women with mild and Coil Recovery percentage Recovery percentage at 15 days after treatment: Randomization described Small number of women I, fair
2003 Turkey blinded controlled moderate PID randomized at 15 days after IUD IUD p value Follow-up examination Distribution of covariates
trial to removal or non-removal treatment in clinical removed left in performed by clinician at baseline not reported
of IUD (12 did not symptoms, findings blinded to IUD group Diagnosis of PID not
Pelvic pain 85 48 .0001
follow-up): of gynecologic exam, Performed intention- well defined
Vaginal 66 33 .007
Group 1: IUD removed, ESR, leukocyte count to-treat analysis Attrition from each group
discharge
N=60 not stated
Dysuria/ 100 59 .0003
Group 2: IUD left in place, Short follow-up
frequency
N=66
Nausea/ 100 92 NS
Exclusion criteria: women
vomiting
with severe PID
Vaginal 80 57 NS
(peritoneal irritation,
bleeding
rebound tenderness,
Dyspareunia 96 59 .0001
abdominal rigidity or
Abdominal 93 67 .005
clinical and hematological
tenderness
instability)
Cervical 64 59 NS
All women treated with
discharge
ciprofloxacin/metronidazole/
Cervical 90 44 .0001
doxycycline.
tenderness
Follow-up
Leukocyte 5600 ± 3100 10 000 ± .001
15 days after
count (mm -3) 2300
treatment
ESR (mm/h) 14 ± 5.6 20 ± 8.3 .001

Abbreviations: ESR, erythrocyte sedimentation rate; IUD, intrauterine device; PID, pelvic inflammatory disease; SD, standard deviation.

657
658 N.K. Tepper et al. / Contraception 87 (2013) 655–660
when she is diagnosed with PID, an important consideration
is whether the IUD can be retained or should be removed.
There is a theoretical concern that presence of an IUD might
impact the course of treatment among women with PID. The
objective of this systematic review is to examine the
evidence comparing the effect of IUD removal versus
retention on clinical course among women using IUDs who
develop PID.
2. Methods
We searched the PubMed database for all peer-reviewed
articles in any language published from database inception
through April 2012 for evidence of outcomes in women
using IUDs (any type) who develop PID. The following
search strategy was used:
((((“Pelvic Inflammatory Disease”[Mesh])) OR (“pelvic
inflammatory disease” OR endometritis OR oophoritis OR
parametritis OR salpingitis))) AND (((levonorgestrel AND
(intrauterine devices[mesh] OR iud OR iucd OR ius OR iuc
OR intrauterine system OR intra-uterine system OR
intrauterine device OR intra-uterine device OR intrauterine
contraceptive OR intrauterine contraception)) OR mirena)
OR ((“Intrauterine Devices”[Mesh] OR “Intrauterine De-
vices, Medicated”[Mesh] OR “Intrauterine Devices, Cop-
per”[Mesh]) OR iud OR iucd OR ius OR iuc OR intrauterine
system OR intra-uterine system OR intrauterine device OR
intra-uterine device OR intrauterine contraceptive OR
intrauterine contraception)).
Reference lists from articles identified by the search
strategy and review articles were hand-searched in order to
identify additional articles. We did not attempt to identify
abstracts from scientific conferences.
2.1. Study selection
We included articles which examined IUD users who
developed PID. Articles were included if they compared
outcomes among women in whom the IUD was retained to
women in whom the IUD was removed. We excluded
articles describing women in whom the infection was
diagnosed before or at the time of IUD insertion. Case
reports and case series were also excluded.
2.2. Study quality assessment
The evidence was summarized and systematically
assessed by all authors. The quality of each individual
piece of evidence was assessed using the United States
Preventive Services Task Force grading system [2].
2.3. Data synthesis
Summary measures of association were not computed due
to heterogeneity among studies with respect to study design,
subject characteristics, disease definitions and outcomes.
Downloaded for Mahasiswa FK UKDW 01 (mhsfkdw01@gmail.com) at Duta Wacana Christian University from ClinicalKey.com by Elsevier on September 20, 2019.
For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.
3. Results
The PubMed search strategy identified 1534 articles, of
which 4 met inclusion criteria for this review (Table 1).
Three randomized controlled trials [3,5,6] and one prospec-
tive cohort study [4] were identified. The majority of women
in the studies used copper IUDs. Outcomes examined
included laboratory parameters and clinical course.
One randomized controlled trial from Sweden examined
women hospitalized for PID [6]. Women using copper IUDs
were randomly assigned to IUD retention or removal at
admission, although randomization was not described.
Women were divided into three groups: 632 women not
using an IUD, 236 women who retained their IUD, and 60
women who had their IUD removed at admission. Among
women who retained the IUD, the percentage of women
hospitalized for three weeks or longer was smaller (4%) than
among women who had the IUD removed (15%) (pb.01). A
group of 184 women were followed for 18 months to
examine PID recurrences and subsequent pregnancies,
although the authors do not state how these women were
selected. Among this group, there were 121 not using an
IUD, 46 in whom the IUD had been retained, and 17 in
whom the IUD had been removed. The authors state that
there were no significant differences in the number of
recurrences or subsequent pregnancies between the groups;
however, it is not clear whether statistical testing was
performed and numbers with each outcome were small.
Another randomized controlled trial from Sweden
examined the course of acute salpingitis in women using
IUDs [3]. Women using IUDs hospitalized with acute
salpingitis were randomized to IUD retention (N=27) or
removal (N=26) at admission. There were 23 women in
each group included in final analyses. The IUDs used
included plastic, copper and unknown type. There were no
statistical differences in the mean peak erythrocyte
sedimentation rate (ESR) level, time to peak ESR value
or half-time of ESR between the two groups. There were no
treatment failures or rehospitalizations within 3 months in
either group.
A third randomized controlled trial from Turkey exam-
ined 138 women using IUDs who developed mild or
moderate PID [5]. All women were using a coil IUD and
were randomized to removal or retention of the IUD. There
were 126 women available for follow-up at 15 days, 66 in
whom the IUD had been retained and 60 in whom the IUD
had been removed. At 15 days after treatment, women in the
IUD removal group had a higher rate of recovery in certain
clinical signs and symptoms than women in the retention
group. Rates of recovery were higher in the IUD removal
group for pelvic pain, vaginal discharge, dysuria, dyspar-
eunia, abdominal tenderness and cervical tenderness. The
IUD removal group also had lower mean ESR and leukocyte
counts at follow-up. There were no differences at follow-up
between the groups in nausea/vomiting, vaginal bleeding or
cervical discharge.
 N.K. Tepper et al. / Contraception 87 (2013) 655–660
One prospective cohort study from Finland examined 186
women using IUDs who were hospitalized with acute PID [4].
The IUD was retained in 105 women and removed in 81
women; no information was provided on how the decision to
retain or remove the IUD was made. The majority of women
were using copper IUDs, with the exception of four women
using Lippes loop, Dalkon shield or Saf-T-Coil. There were no
statistically significant differences in clinical and laboratory
outcomes between groups during hospitalization, including
duration of fever, highest ESR, ESR on discharge and length of
hospitalization. A greater proportion of women in the IUD
removal group had a hospital stay longer than 2 weeks
compared to the IUD retention group (33% versus 19%; pb.05).
4. Discussion
Evidence is limited regarding the clinical course of PID in
IUD users who retain versus remove the IUD. Clinical and
laboratory outcomes were not generally different between
IUD retention and IUD removal groups. Two studies showed
that women with IUDs removed had higher rates of longer
hospitalization [4,6]. In contrast, one randomized trial
showed that IUD removal was associated with higher rate
of recovery and improved laboratory values [5].
Some strengths of the evidence were that one study had
long follow-up for some subjects [6] and two studies used
objective criteria to compare outcomes [3,4].There were
several limitations to this body of evidence. Most studies had
small numbers of women [3,5,6] with limited [3,5] or no
follow-up after hospitalization [4]. Two of the randomized
trials did not describe randomization procedures [3,6]. In one
of these trials, there was a large discrepancy between the size
of the groups and no description of subject characteristics was
given [6]. The same study also did not describe how the subset
of women who were followed prospectively was selected [6].
In the cohort study, there was no description of how groups
were determined [4]. One of the studies examining ESR did
not assess any clinical outcomes of PID [3].
When assessing the impact that the IUD may have on the
treatment and clinical outcome of PID, the severity of disease
should be considered. Most cases of PID in the United States
are mild or moderate, defined as having evidence of lower
genital tract inflammation and pelvic organ tenderness but no
mass [7,8]. The location of treatment (inpatient or outpatient)
can be considered a proxy for the severity of disease. Three of
the studies in this review examined only women hospitalized
with PID [3,4,6]. The fourth study did not specify the location
of treatment or whether or not women were hospitalized;
however, women with severe PID were excluded [5].
Together, these studies provide minimal evidence regarding
women diagnosed and treated in the outpatient setting. While
historically most cases of PID were treated with intravenous
antibiotics in the inpatient setting, the majority of cases of PID
are now treated with oral outpatient therapy. It is estimated that
70-90% of women with PID in the United States receive PID
Downloaded for Mahasiswa FK UKDW 01 (mhsfkdw01@gmail.com) at Duta Wacana Christian University from ClinicalKey.com by Elsevier on September 20, 2019.
For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.
659
care through outpatient facilities [8,9]. Among women with
mild to moderate PID, clinical outcomes are similar whether
treatment occurs in the inpatient or outpatient setting [8].
Although evidence is limited, any impact of an IUD on the
course of PID may be less pronounced among women with
mild to moderate PID compared with women with severe PID
due to their milder disease state.
There is a theoretical concern that presence of an IUD
might impact the effectiveness of treatment among women
with PID. There are studies that have examined course of
treatment comparing IUD users with non-IUD users. Some
of these studies have demonstrated that women with IUDs
had increased ESR, longer hospitalizations, increased risk of
surgery and more anaerobic bacteria than women without
IUDs [10–12]. In particular, some of these clinical outcomes
were worse among women ages ≥35 years using IUDs
[11,13]. However, another study did not show differences in
clinical outcomes when comparing IUD users to users of
other contraceptive methods [14]. None of these studies
provided information on whether the IUD users retained the
IUD or had it removed. Therefore, these results should be
interpreted cautiously, as the clinical course in women who
are not using IUDs may not be equivalent to that in women
who have their IUDs removed.
An additional consideration is whether a levonorgestrel-
releasing IUD (LNG-IUD) might have a differential effect on
the course of PID as compared with copper or other types of
IUDs. Users of LNG-IUDs may have lower rates of
developing PID than users of other types [15,16]. Possible
mechanisms for this reduced risk include thickening of
cervical mucus, suppression of endometrium, and reduced
bleeding in LNG-IUD users. However, this review found no
studies of women who developed PID while using an LNG-
IUD and no studies which compared different types of IUDs.
In summary, two randomized controlled trials and one
prospective cohort study, all of fair quality, showed no
differences in clinical outcomes or laboratory parameters
among women hospitalized for PID who retained the IUD
when compared with women in whom the IUD was removed.
Two fair quality studies showed that women with IUDs
removed had longer hospitalizations. One fair quality
randomized controlled trial showed that women who had the
IUD removed experienced improved recovery in most clinical
signs and symptoms compared with women who retained the
IUD. Overall, women who retained their IUDs had similar or
better outcomes than women who had their IUDs removed.
References
[1] Simms I, Stephenson JM. Pelvic inflammatory disease epidemiology:
what do we know and what do we need to know? Sex Transm Infect
2000;76:80–7.
[2] Harris RP, Helfand M, Woolf SH, et al. Current methods of the US
Preventive Services Task Force: a review of the process. Am J Prev
Med 2001;20(3 suppl):21–35.
[3] Soderberg G, Lindgren S. Influence of an intrauterine device on the
course of an acute salpingitis. Contraception 1981;24:137–43.
660 N.K. Tepper et al. / Contraception 87 (2013) 655–660

[4] Teisala K. Removal of an intrauterine device and the treatment of acute
pelvic inflammatory disease. Ann Med 1989;21:63–5.
[5] Altunyurt S, Demir N, Posaci C. A randomized controlled trial of coil
removal prior to treatment of pelvic inflammatory disease. Eur J Obstet
Gynecol Reprod Biol 2003;107:81–4.
[6] Larsson B, Wennergren M. Investigation of a copper-intrauterine
device (Cu-IUD) for possible effect on frequency and healing of pelvic
inflammatory disease. Contraception 1977;15:143–9.
[7] Jaiyeoba O, Lazenby G, Soper DE. Recommendations and rationale for
the treatment of pelvic inflammatory disease. Expert Rev Anti Infect
Ther 2011;9:61–70.
[8] Ness RB, Soper DE, Holley RL, et al. Effectiveness of inpatient and
outpatient treatment strategies for women with pelvic inflammatory disease:
results from the Pelvic Inflammatory Disease Evaluation and Clinical Health
(PEACH) Randomized Trial. Am J Obstet Gynecol 2002;186:929–37.
[9] Bohm MK, Newman L, Satterwhite CL, Tao G, Weinstock HS. Pelvic
inflammatory disease among privately insured women, United States,
2001–2005. Sex Transm Dis 2010;37:131–6.
[10] Miettinen A, Saikku P, Jansson E, Paavonen J. Epidemiologic and
clinical characteristics of pelvic inflammatory disease associated with
Mycoplasma hominis, Chlamydia trachomatis, and Neisseria gonor-
rhoeae. Sex Transm Dis 1986;13:24–8.
[11] Viberga I, Odlind V, Berglund L. The impact of age and intrauterine
contraception on the clinical course of pelvic inflammatory disease.
Gynecol Obstet Invest 2006;61:65–71.
[12] Viberga I, Odlind V, Lazdane G, Kroica J, Berglund L, Olofsson S.
Microbiology profile in women with pelvic inflammatory disease in
relation to IUD use. Infect Dis Obstet Gynecol 2005;13:183–90.
[13] Viberga I, Odlind V, Berglund L. “Older” age is a risk factor for pelvic
inflammatory disease in intrauterine device users. Acta Obstet Gynecol
Scand 2005;84:1202–7.
[14] Svensson L, Westrom L, Mardh PA. Contraceptives and acute
salpingitis. JAMA 1984;251:2553–5.
[15] Farley TM, Rosenberg MJ, Rowe PJ, Chen JH, Meirik O. Intrauterine
devices and pelvic inflammatory disease: an international perspective.
Lancet 1992;339:785–8.
[16] Toivonen J, Luukkainen T, Allonen H. Protective effect of intrauterine
release of levonorgestrel on pelvic infection: three years' comparative
experience of levonorgestrel- and copper-releasing intrauterine de-
vices. Obstet Gynecol 1991;77:261–4.

Downloaded for Mahasiswa FK UKDW 01 (mhsfkdw01@gmail.com) at Duta Wacana Christian University from ClinicalKey.com by Elsevier on September 20, 2019.
For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.