Articles

Salpingotomy versus salpingectomy in women with tubal

pregnancy (ESEP study): an open-label, multicentre,
randomised controlled trial
Femke Mol, Norah M van Mello, Annika Strandell, Karin Strandell, Davor Jurkovic, Jackie Ross, Kurt T Barnhart, Tamer M Yalcinkaya,
Harold R Verhoeve, Giuseppe C M Graziosi, Carolien A M Koks, Ingmar Klinte, Lars Hogstrm, Ineke C A H Janssen, Harry Kragt, Annemieke Hoek,
Trudy C M Trimbos-Kemper, Frank J M Broekmans, Wim N P Willemsen, Willem M Ankum, Ben W Mol, Madelon van Wely, Fulco van der Veen,
Petra J Hajenius, for the European Surgery in Ectopic Pregnancy (ESEP) study group*

Summary
Background Tubal ectopic pregnancy can be surgically treated by salpingectomy, in which the aected Fallopian tube
is removed, or salpingotomy, in which the tube is preserved. Despite potentially increased risks of persistent
trophoblast and repeat ectopic pregnancy, salpingotomy is often preferred over salpingectomy because the preservation
of both tubes is assumed to oer favourable fertility prospects, although little evidence exists to support this
assumption. We aimed to assess whether salpingotomy would improve rates of ongoing pregnancy by natural
conception compared with salpingectomy.

Published Online
February 3, 2014
http://dx.doi.org/10.1016/
S0140-6736(14)60123-9
See Online/Comment
http://dx.doi.org/10.1016/
S0140-6736(14)60129-X
*Members listed at end of paper

Methods In this open-label, multicentre, international, randomised controlled trial, women aged 18 years and older
with a laparoscopically conrmed tubal pregnancy and a healthy contralateral tube were randomly assigned via a
central internet-based randomisation program to receive salpingotomy or salpingectomy. The primary outcome was
ongoing pregnancy by natural conception. Dierences in cumulative ongoing pregnancy rates were expressed as a
fecundity rate ratio with 95% CI, calculated by Cox proportional-hazards analysis with a time horizon of 36 months.
Secondary outcomes were persistent trophoblast and repeat ectopic pregnancy (expressed as relative risks [RRs] with
95% CIs) and ongoing pregnancy after ovulation induction, intrauterine insemination, or IVF. The researchers who
collected data for fertility outcomes were masked to the assigned intervention, but patients and the investigators who
analysed the data were not. All endpoints were analysed by intention to treat. We also did a (non-prespecied) metaanalysis that included the ndings from the present trial. This trial is registered, number ISRCTN37002267.
Findings 446 women were randomly assigned between Sept 24, 2004, and Nov 29, 2011, with 215 allocated to
salpingotomy and 231 to salpingectomy. Follow-up was discontinued on Feb 1, 2013. The cumulative ongoing
pregnancy rate was 607% after salpingotomy and 562% after salpingectomy (fecundity rate ratio 106, 95% CI
081138; log-rank p=0678). Persistent trophoblast occurred more frequently in the salpingotomy group than in
the salpingectomy group (14 [7%] vs 1 [<1%]; RR 150, 201134). Repeat ectopic pregnancy occurred in 18 women
(8%) in the salpingotomy group and 12 (5%) women in the salpingectomy group (RR 16, 0833). The number of
ongoing pregnancies after ovulation induction, intrauterine insemination, or IVF did not dier signicantly between
the groups. 43 (20%) women in the salpingotomy group were converted to salpingectomy during the initial surgery
because of persistent tubal bleeding. Our meta-analysis, which included our own results and those of one other
study, substantiated the results of the trial.
Interpretation In women with a tubal pregnancy and a healthy contralateral tube, salpingotomy does not
signicantly improve fertility prospects compared with salpingectomy.
Funding Netherlands Organisation for Health Research and Development (ZonMW), Region Vstra Gtaland Health
& Medical Care Committee.

Introduction
About 12% of all pregnancies are ectopic.1 Most ectopic
pregnancies are located in the Fallopian tube, and
surgery is often used as treatment.2 Excision of the
aected tubeie, salpingectomyis regarded as the
standard surgical procedure.35
As early as 1914, Beckwith Whitehouse questioned
whether this sacrice of the tube was justied in all
cases. After studying the histopathology of tubal
pregnancies, he showed that salpingotomyie, removal

of the pregnancy while preserving the aected tube

was also a feasible intervention.6 From 1957 onwards,
this concept of preservation of the aected organ was
promoted in favour of ablative surgery.7
Despite the fact that randomised studies had not been
done, salpingotomy was widely adopted on the basis of a
presumed favourable outcome with respect to future
reproductive capacity, although physicians recognised the
potential drawback of a repeat ectopic pregnancy in the
same tube.8 In 1984, another disadvantage of salpingotomy

www.thelancet.com Published online February 3, 2014 http://dx.doi.org/10.1016/S0140-6736(14)60123-9

Centre for Reproductive

Medicine (F Mol PhD,
M van Wely PhD,
F van der Veen PhD) and
Department of Obstetrics and
Gynaecology
(N M van Mello PhD,
W M Ankum PhD, B W Mol PhD,
P J Hajenius PhD), Academic
Medical Centre, University of
Amsterdam, Amsterdam,
Netherlands; University
Medical Centre Groningen,
University of Groningen,
Groningen, Netherlands
(A Hoek PhD); Radboud
University Medical Centre,
Nijmegen, Netherlands
(W N P Willemsen PhD); Leiden
University Medical Centre,
Leiden, Netherlands
(T C M Trimbos-Kemper PhD);
Utrecht University Medical
Centre, Utrecht, Netherlands
(F J M Broekmans PhD); Onze
Lieve Vrouwe Gasthuis,
Amsterdam, Netherlands
(H R Verhoeve PhD); Mxima
Medical Centre, Veldhoven,
Netherlands (C A M Koks PhD);
Reinier de Graaf Hospital, Delft,
Netherlands (H Kragt PhD);
Antonius Hospital,
Nieuwegein, Netherlands
(G C M Graziosi PhD); Groene
Hart Hospital, Gouda,
Netherlands
(I C A H Janssen PhD);
Sahlgrenska University
Hospital, Gteborg, Sweden
(A Strandell PhD,
K Strandell MD); Skaraborg
Hospital, Skvde, Sweden
(L Hogstrm MD); NU Hospital
Group, Trollhttan, Sweden
(I Klinte MD); Kings Early
Pregnancy Unit, Kings College

Articles

Hospital, London, UK
(D Jurkovic MD, J Ross MD);
Wake Forest University School
of Medicine, Winston-Salem,
NC, USA (T M Yalcinkaya MD);
Penn Fertility Care, Perelman
School of Medicine, University
of Pennsylvania, Philadelphia,
PA, USA (K T Barnhart MD);
Department of Obstetrics and
Gynaecology, University
College London Hospital,
London, UK (D Jurkovic); and
School of Paediatrics and
Reproductive Health,
University of Adelaide,
Adelaide, SA,
Australia (B W Mol)
Correspondence to:
Dr Femke Mol, Centre for
Reproductive Medicine,
Academic Medical Centre,
University of Amsterdam,
1105 AZ Amsterdam,
Netherlands
f.mol@amc.nl

was reportedthe incomplete removal of the ectopic

pregnancy (ie, persistent trophoblast), necessitating
additional treatment.9
Whether salpingotomy provides better fertility
prospects than salpingectomy and whether the presumed
potential benet outweighs the disadvantages remains
unclear. In view of this uncertainty, we aimed to compare
the procedures with respect to future fertility prospects
in a randomised controlled trial.

Methods
Study design and participants
The European Surgery in Ectopic Pregnancy (ESEP)
study was an open-label, multicentre, randomised
controlled trial in university hospitals and other teaching
and non-teaching hospitals in the Netherlands, Sweden,
the UK, and the USA. ESEP started as a Dutch-SwedishBritish collaboration, and two centres in the USA joined
during the study period.
Women were eligible for the trial if they had a
presumptive diagnosis of tubal pregnancy and were
scheduled for surgery. To reach this presumptive diagnosis,
an algorithm based on transvaginal ultrasonography
with serum human chorionic gonadotropin (hCG)
measurements was followed in all clinics, as recommended
by national and international guidelines.35,10 Women who
were younger than 18 years, were haemodynamically
unstable, had no desire for future pregnancy, or were
pregnant after in-vitro fertilisation (IVF) were excluded, as
were women with only one tube, or with contralateral tubal
occlusion or a hydrosalpinx documented at a previous
hysterosalpingography or laparoscopy.
At surgery, the presence of a tubal pregnancy had to be
conrmed. If tubal rupture was present, women were
still eligible for the trial as long as the tubal rupture did

not aect the possibility of doing a salpingotomy. Women

were not eligible if, in the opinion of the surgeon, the
condition of the contralateral tube rendered future
pregnancy unlikely in case the assigned treatment would
be salpingectomy (eg, because of hydrosalpinx, severe
peritubal adhesions, or malformations). Thus, only
women with a tubal pregnancy amenable to either
treatment intervention and with a healthy contralateral
tube were enrolled.
One institutional review board in each country approved
the study protocol,11 after which the boards of directors of
all other participating centres provided local approval.
Participants provided written informed consent.

Randomisation and masking

During surgery, after conrmation of the diagnosis of
tubal pregnancy and conrmation that the contralateral
tube was healthy, consenting women were randomly
assigned to salpingotomy or salpingectomy via a central
internet-based program running a computer-generated
randomisation sequence. Randomisation was stratied
by hospital, the womans age (<35 or 35 years), and
history of tubal disease (ie, previous ectopic pregnancy,
tubal surgery, or pelvic inammatory disease). The
randomisation sequence was not accessible by the
recruiters. The allocation code was disclosed after the
patients initials were entered and inclusion criteria were
conrmed on the website; the unique number generated
could not be deleted afterwards. The study was openlabel because the nature of the intervention meant that
masking patients to the assigned intervention was not
possible. The researchers who collected data for fertility
outcomes were masked to the assigned intervention, but
those who analysed the data were not.

Procedures
450 women with tubal ectopic
pregnancies enrolled
4 excluded because data
were unavailable
446 randomly assigned

215 assigned to salpingotomy

164 received assigned intervention
51 received salpingectomy
43 had persistent bleeding
3 had a reintervention
5 had persistent trophoblast

11 lost to fertility follow-up

215 included in the intention-to-treat analysis

164 included in the per-protocol analysis

Figure 1: Trial prole

231 assigned to salpingectomy

231 received assigned intervention

13 lost to fertility follow-up

231 included in the intention-to-treat analysis

231 included in the per-protocol analysis

Both interventions were done in accordance with local

procedural standards used in the participating hospitals.
In women assigned to salpingotomy, a linear
salpingotomy was done. To prevent bleeding, an injection
of epinephrine, terlipressin, ornipressin, or any analogue,
or the application of brin glue was allowed. In women
assigned to salpingectomy, all techniques for complete
salpingectomy were allowed, including clamping, cutting
and suturing, and use of bipolar forceps and scissors,
bipolar cutting forceps, vessel sealing instruments,
ultrasonic devices, and endoloop snares. Whenever
necessary, salpingotomy was converted to salpingectomy
and laparoscopy was converted to open surgery.
After surgery, all women were informed about their
study group assignment and the intervention that they
received. To identify persistent trophoblast, serum hCG
(expressed in IU/L)was measured postoperatively until
undetectable concentrations were reached in both study
groups.12 To assess fertility after surgery, researchers
contacted the participants by telephone, email, or postal
mail every 6 months for 36 months. The participants

www.thelancet.com Published online February 3, 2014 http://dx.doi.org/10.1016/S0140-6736(14)60123-9

Articles

completed a questionnaire about the occurrence and

outcome of subsequent pregnancies until an ongoing
pregnancy occurred. Repeated attempts by telephone,
email, or postal mail were made when participants did
not respond to the initial follow-up contact.
The primary outcome was ongoing pregnancy by
natural conception. An ongoing pregnancy was dened
as an intrauterine pregnancy visible on ultrasound at a
gestational age of 12 weeks or more with fetal cardiac
activity, or a pregnancy that resulted in a livebirth. We
calculated the time to the rst ongoing pregnancy in
months, from the date of surgery to the rst day of the
last menstrual period before the conception that led to
the ongoing pregnancy. If an ongoing pregnancy did not
occur, follow-up ended at the last date of contact, or at the
moment when either IVF or reconstructive tubal surgery
was done. Natural conceptions that occurred after failed
IVF treatment were registered, but these pregnancies
were not included as primary endpoint events in the
initial analysis.
Secondary outcomes were persistent trophoblast, rst
repeat ectopic pregnancy, and rst ongoing pregnancy
after ovulation induction, intrauterine insemination, or
IVF. Persistent trophoblast was dened as rising or
plateauing serum hCG concentrations postoperatively
that necessitated systemic methotrexate treatment or
surgical intervention.12 Repeat ectopic pregnancy was
dened as any ectopic pregnancy or a persisting pregnancy
of unknown location for which surgery or medical
treatment with methotrexate was necessary.
We also did a non-prespecied meta-analysis that
included the results of our trial. We searched PubMed
and the Cochrane Central Register of Controlled Trials
for reports of randomised trials that compared
salpingotomy with salpingectomy for women with tubal
pregnancy and that included an outcome that assessed
future fertility, persistent trophoblast, or repeat ectopic
pregnancy. We used the search terms ectopic
pregnancy and randomized controlled trial to identify
reports published from Jan 1, 1966, to May 1, 2013. No
language restrictions were used. Two authors (FM and
PJH) reviewed identied articles for relevance and
quality, and abstracted the data. We assessed studies for
quality on the basis of the Cochrane Collaborations
method for assessing risk of bias.13 We calculated hazard
ratios (HRs) for time to subsequent ongoing pregnancy
by natural conception and relative risks (RRs) with
95% CIs for persistent trophoblast and repeat ectopic
pregnancy.

Salpingotomy
(n=215)
Mean age (years)
Age 31 years

Salpingectomy
(n=231)

309 (55)

309 (55)

110 (51%)

118 (51%)

Risk factors for tubal disease

Known tubal disease*

6 (3%)

4 (2%)

History of chlamydia

26 (12%)

22 (10%)

History of pelvic inammatory disease

5 (2%)

9 (4%)

History of ectopic pregnancy

9 (4%)

5 (2%)

41 (19%)

54 (23%)

3 (1%)

2 (1%)

History of termination of pregnancy

Intrauterine device in situ
Symptoms
None

13 (6%)

10 (4%)

Pelvic pain only

38 (18%)

34 (15%)

Vaginal bleeding only

Pelvic pain and vaginal bleeding

24 (11%)

25 (11%)

138 (64%)

155 (67%)

139 (65%)

154 (67%)

Ultrasound ndings
Ectopic mass
Mean size of ectopic mass (cm)
Fetal heart beat present
Median preoperative serum hCG (IU/L)

26 (13)

24 (14)

18 (8%)

26 (11%)

2181 (8604298)

2409 (9206036)

Location of tubal pregnancy at surgery

Ampulla

163 (76%)

177 (77%)

Fimbriae

14 (7%)

26 (11%)

Isthmus

31 (14%)

22 (10%)

Tubal rupture present during surgery

1 (<1%)

Data are n (%), mean (SD), or median (IQR). We noted no signicant imbalances between groups. *Tubal disease seen
on previous hysterosalpingography or laparoscopy. Mean size of ectopic mass was calculated from data for 125 (90%)
of 139 women with an ectopic mass in the salpingotomy group and 129 (84%) of 154 women with an ectopic mass in
the salpingectomy group. Preoperative serum human chorionic gonadotropin (hCG) was recorded in 212 (99%) of
215 women in the salpingotomy group and 226 (98%) of 231 women in the salpingotomy group. Did not aect
salpingotomy.

Table 1: Baseline characteristics of the women*

Salpingotomy
(n=215)
Conversion to open surgery

3 (1%)

Salpingectomy
(n=231)
3 (1%)

Conversion to salpingectomy

43 (20%)

NA

Blood transfusion

14 (7%)

7 (3%)

Initial admission
Repeat laparoscopy with salpingectomy for suspected bleeding*

2 (1%)

Readmission*
Repeat laparoscopy with salpingectomy for suspected bleeding

1 (<1%)

Repeat laparoscopy with salpingectomy for persistent trophoblast

5 (2%)

Other surgical reintervention

4 (2%)

2 (1%)

10 (5%)

3 (1%)

Readmission only

Statistical analysis

Data are n (%). *Repeat laparoscopy and readmissions were regarded as serious adverse events.

We assumed the cumulative proportion of women with

an ongoing pregnancy after salpingectomy would be 40%
after 36 months, with a median time to ongoing
pregnancy of 14 years.14 We regarded an increase of
15 percentage points (ie, from 40% to 55%) in women
treated with salpingotomy to be clinically relevant

Table 2: Adverse events

enough to overcome the potential disadvantages of

persistent trophoblast and repeat ectopic pregnancy. This
increase corresponded to a reduction in the median time
to ongoing pregnancy from 14 to 10 years. On this

www.thelancet.com Published online February 3, 2014 http://dx.doi.org/10.1016/S0140-6736(14)60123-9

Articles

basis, we needed to enrol 404 women for a power of 80%

with a two-sided signicance of 005. Since we
anticipated a 10% loss to follow-up, we aimed to enrol
450 women.15
The analysis was done according to the intention-totreat principle. We derived the cumulative ongoing
pregnancy rates from life-table analysis and expressed
dierences as a fecundity rate ratio with 95% CI,
calculated through Cox proportional-hazards analysis
with a time horizon of 36 months. We constructed
Kaplan-Meier curves to estimate the cumulative
probability of an ongoing pregnancy over time. The logrank test was used to test dierences between the
Kaplan-Meier curves for signicance. Persistent
trophoblast, repeat ectopic pregnancy, and ongoing

Cumulative ongoing pregnancy rate (%)

100

Salpingotomy
Salpingectomy

75

50

Role of the funding source

25

0
0
Number at risk
Salpingotomy 215
Salpingectomy 231

6
158
172

12
18
24
Time after random assignment (months)
128
126

92
103

65
95

30

36

48
80

40
66

Figure 2: Kaplan-Meier curves for time to ongoing pregnancy by natural conception

Shaded areas show 95% CIs. Median time to ongoing pregnancy by natural conception after salpingotomy was
20 months (95% CI 1723) and after salpingectomy was 26 months (95% CI 1537); cumulative rate of ongoing
pregnancy by natural conception after 36 months was 607% after salpingotomy and 562% after salpingectomy;
log-rank p=0678; =0172 (one degree of freedom).

Salpingotomy Salpingectomy Relative risk

(n=215)
(n=231)
(95% CI)
1 (<1%)

p value

Persistent trophoblast

14 (7%)

150 (201134)

001

Repeat ectopic pregnancy

18 (8%)

12 (5%)

16 (0833)

019

Ipsilateral tube

7 (3%)

3 (1%)

25 (0796)

018

Contralateral tube

8 (4%)

7 (3%)

12 (0534)

069

Persisting pregnancy of unknown location

3 (1%)

2 (1%)

16 (0395)

060

Ovulation induction

3 (1%)

Intrauterine insemination

1 (<1%)

In-vitro fertilisation

7 (3%)

2 (1%)

Ongoing pregnancy by:

Data are n (%), unless otherwise indicated.

Table 3: Secondary outcomes

pregnancy after ovulation induction, intrauterine

insemination, or IVF were expressed as RRs with
95% CIs. We also did a per-protocol analysis of the
primary outcome in which we compared women who
underwent the assigned salpingotomy (ie, those who did
not have a conversion to salpingectomy) with those who
had been randomly assigned to and underwent
salpingectomy.
We did prespecied subgroup analyses for maternal
age, history of a previous ectopic pregnancy, preoperative
serum hCG concentration, and size of the ectopic
pregnancy on ultrasound. Continuous variables were
dichotomised on the basis of median values. For each
subgroup, we calculated fecundity rate ratios with
95% CIs using Cox proportional-hazards analysis. We
investigated interactions between these variables and
assigned treatment with respect to time to ongoing
pregnancy, with a p value (also calculated by the Cox
proportional-hazards analysis) of less than 005 regarded
as showing an interaction.
For the meta-analysis, we extracted the HRs for
ongoing pregnancy and the RRs for persistent trophoblast
and repeat ectopic pregnancy, and pooled the data using
the random-eect method.
Statistical analyses were done in Stata (version 11.2),
and the meta-analysis was done with Comprehensive
Meta-Analysis software (version 9).

38 (08179)

010

The sponsors of the study had no role in study design,

data collection, data analysis, data interpretation, or
writing of the report. The corresponding author had
full access to all the data in the study and had nal
responsibility for the decision to submit for publication.

Results
446 women were randomly assigned between Sept 24,
2004, and Nov 29, 2011, with 215 allocated to salpingotomy
and 231 to salpingectomy. Of the 215 women in the
salpingotomy group, 164 (76%) underwent the assigned
intervention as planned, and the remaining 51 women in
the group received salpingectomy. 43 women were
converted to salpingectomy during the initial surgery
because of persistent tubal bleeding, three had a
salpingectomy at reintervention because of suspected
tubal bleeding, and ve had a salpingectomy because of
persistent trophoblast. Baseline characteristics were
similar between the study groups (table 1). Table 2 shows
the frequencies of adverse events, such as conversion to
laparotomy or salpingectomy, blood transfusions, readmittances, and reinterventions.
Of the 446 women who underwent random assignment
and were analysed, 222 had an ongoing pregnancy by
natural conception; 108 after salpingotomy and 114 after
salpingectomy. The cumulative ongoing pregnancy rate
by natural conception within a time horizon of 36 months
was 607% after salpingotomy and 562% after

www.thelancet.com Published online February 3, 2014 http://dx.doi.org/10.1016/S0140-6736(14)60123-9

Articles

salpingectomy (fecundity rate ratio 106, 95% CI

081138; log-rank p=0678; gure 2). Persistent
trophoblast occurred signicantly more frequently in the
salpingotomy group than in the salpingectomy group,
but the numbers of repeat ectopic pregnancies did not
dier signicantly (table 3).
The per-protocol analysis incorporated the 164 women
who underwent a completed salpingotomy and the
231 women who were assigned to and received
salpingectomy. 201 women in the per-protocol analysis
had an ongoing pregnancy by natural conception; 87 after
salpingotomy and 114 after salpingectomy. The
cumulative ongoing pregnancy rate by natural conception
was 623% after salpingotomy and 562% after
salpingectomy (fecundity rate ratio 110, 95% CI
083146; log-rank p=0492).
The prespecied subgroup analyses showed no
signicant benecial eect of salpingotomy on the
cumulative rates of ongoing pregnancy by natural
conception in any of the subgroups (table 4).
The results of all the analyses were similar when the
four women from the hospital that was unable to provide
data were included (data not shown).
Our meta-analysis, which included our own results and
data from another study,16 substantiated our nding that
cumulative rates of ongoing pregnancy by natural
conception were similar between salpingotomy and
salpingectomy (gure 3). Meta-analysis for the outcome
of persistent trophoblast was not possible since only data
from our study were available. The risk of repeat ectopic
pregnancy was not signicantly increased after
salpingotomy (gure 3).

Salpingotomy

Salpingectomy Fecundity rate ratio

(95% CI)

Interaction
p value

Age group
<31 years

57/105 (54%)

55/113 (49%)

116 (080168)

056

31 years

51/110 (46%)

59/118 (50%)

095 (065139)

History of previous ectopic pregnancy

Yes*

0/9

No

1/5 (20%)

108/206 (52%)

113/226 (50%)

095

110 (084143)

Preoperative serum hCG

<2335 IU/L

64/109 (59%)

63/110 (57%)

114 (080161)

057

2335 IU/L

43/103 (42%)

48/116 (41%)

098 (065148)

Size of ectopic mass on ultrasound

<21 cm

29/61 (48%)

30/62 (48%)

099 (059165)

042

21 cm

34/64 (53%)

27/67 (40%)

136 (082227)

Data are n/N, unless otherwise indicated. Continuous variables were dichotomised by their medians. *A fecundity rate
ratio could not be calculated because no pregnancies occurred among the nine women with a history of previous
ectopic pregnancy. Preoperative serum human chorionic gonadotropin (hCG) was recorded in 212 (99%) of
215 women in the salpingotomy group and 226 (98%) of 231 women in the salpingotomy group Mean size of
ectopic mass was calculated from data for 125 (90%) of 139 women with an ectopic mass in the salpingotomy group
and 129 (84%) of 154 women with an ectopic mass in the salpingectomy group.

Table 4: Ongoing pregnancy by natural conception in prespecied subgroups

Hazard ratio (95% CI)

Z value

p value

Future fertility rate

Mol et al (2013)

1057 (08121375)

0414

0680

Fernandez et al16 (2013)

1130 (07321745)

0552

0581

Total

1079 (08561361)

0647

0518

Relative risk (95% CI)

Z value

p value

I2=0%

001

01

10

100

Repeat ectopic pregnancy

Discussion
In this randomised controlled trial, salpingotomy did not
improve cumulative rates of ongoing pregnancy by
natural conception in women with a tubal pregnancy and
a healthy contralateral tube, but was associated with an
increased risk of persistent trophoblast. The hypothesis
that women with tubal pregnancy and a healthy
contralateral tube might benet from salpingotomy was
based on the assumption that two tubes provide a better
chance of future pregnancy than a solitary tube. Our
ndings reject this hypothesis, since one properly
functioning tube seems to be equally sucient for timely
conception.
The ESEP study is the largest trial so far to compare
salpingotomy with salpingectomy in women with a
healthy contralateral tube. Results from another recent
randomised controlled trial (DEMETER)16 had a similar
result with respect to cumulative ongoing pregnancy
rates (HR 113, 95% CI 073174). Our meta-analysis of
both trials, which included a total of 649 women, showed
no signicant dierence in cumulative ongoing
pregnancy
rates
between
salpingotomy
and
salpingectomy (panel). Although we did not note
statistical heterogeneity between the two trials, we expect

Mol et al (2013)

1612 (07953266)

1324

0185

Fernandez et al16 (2013)

0647 (01882222)

0692

0489

Total

1287 (06972375)

0806

0420

I2=0%
001

01

10

100

Figure 3: Meta-analysis of studies comparing salpingotomy with salpingectomy in women with tubal pregnancy

substantial clinical heterogeneity due to dierences

between the populations studied, the interventions
assessed, and the outcomes measured. In the ESEP
study, women with contralateral tubal disease were
excluded because ndings from previous cohort studies14
suggested that these women might have a better fertility
outlook from conservative surgery, whereas such women
were not excluded from the DEMETER study. The
DEMETER study combined salpingotomy with a single
dose of systemic methotrexate, whereas ESEP used
salpingotomy alone.2,3,5 In the ESEP study, the primary
outcome was ongoing pregnancy by natural conception,
whereas in the DEMETER study a composite outcome
that included miscarriages and pregnancy terminations
was used. In the DEMETER study, time to pregnancy was
calculated from desire for pregnancy, whereas we used
time from random assignment. Finally, we regarded a

www.thelancet.com Published online February 3, 2014 http://dx.doi.org/10.1016/S0140-6736(14)60123-9

Articles

Panel: Research in context

Systematic review
We searched PubMed and the Cochrane Central Register of
Controlled Trials for reports of randomised trials that
compared salpingotomy with salpingectomy for women with
tubal pregnancy and that included an outcome that assessed
future fertility. Our search covered reports published from Jan
1, 1966, to May 1, 2013, and used the search terms ectopic
pregnancy and randomized controlled trial. No language
restrictions were used. Two authors (FM and PJH) reviewed
identied articles for relevance and quality, and abstracted the
data. We assessed studies for quality on the basis of the
Cochrane Collaborations method for assessing risk of bias.13
We identied one study,16 the results of which suggested that
salpingotomy does not dier signicantly from salpingectomy
with respect to future fertility. We included this study and our
own trial results in our meta-analysis.
Interpretation
Findings from our trial and meta-analysis show that
salpingotomy does not signicantly improve fertility
prospects compared with salpingectomy in women with
tubal pregnancy and a healthy contralateral tube.

dierence of 15 percentage points in ongoing pregnancy

rate as clinically relevant for our sample size calculation,
whereas the DEMETER study used a dierence of
20 percentage points. Results of a cohort study17 reported
in 2012 suggest that the cumulative rate of intrauterine
pregnancy is nine percentage points higher after
salpingotomy than after salpingectomy, but this
dierence was undetectable in our meta-analysis of
ndings from the two trials. In our opinion, new
randomised studies will be unlikely to have sucient
power to draw denitive conclusions on the basis of this
small dierence.
Our initial sample size calculation assumed that
276 women would have an ongoing pregnancy by natural
conception. We discontinued the study follow-up in
February, 2013, because we judged that further
continuation was futile; at the time of discontinuation, 29
(7%) women were still in follow-up. A sensitivity analysis
showed no dierence on the primary outcome when we
assumed that these 29 women all achieved an ongoing
pregnancy by natural conception (data not shown).
The numbers of repeat ectopic pregnancies were low in
both the ESEP study and the DEMETER study.16 The
point estimates were in opposite directionswe noted
more repeat ectopic pregnancies after salpingotomy,
whereas the DEMETER investigators noted more after
salpingectomybut neither result was signicant. In
our meta-analysis, the non-signicant summary point
estimate for repeat ectopic pregnancy was in the direction
of more repeat ectopic pregnancies after salpingotomy.
Our study had some limitations. Women were randomly
assigned during surgery, and learned afterwards whether
6

or not they were included in the trial and which

intervention they received. Patients, therefore, were not
masked to the assigned intervention. The researchers who
collected data for fertility outcomes were unaware of the
treatment allocation. We believe that this approach
provides a reasonable balance between the need to inform
participants and the safeguarding of unbiased assessment
of the primary outcome.
20% of women in the salpingotomy group were
converted to salpingectomy during the initial surgery, a
similar proportion to that reported in the DEMETER
study.16 Few data are available with which to compare this
conversion frequency, which is otherwise prone to
selection biaseg, by the size of the tubal pregnancy or
the serum hCG concentration.18 This conversion
frequency increases the statistical uncertainty about the
absence of an eect on the primary outcome, but our perprotocol analysis, which included only women who
underwent the assigned intervention, showed similar
point estimates and CIs for the primary outcome. This
result substantiates the ndings from our intention-totreat analysis and suggests that even without conversions
to salpingectomy (ie, a conversion rate of zero) a positive
eect of salpingotomy on fertility is unlikely. Our study
was a pragmatic trial, which is representative of clinical
practice. In future studies, the experience of each
surgeonie, the number of previous interventions
undertakenshould be recorded and taken into account,
but it remains dicult to determine the degree of
experience that equates to a high-quality standard for
surgical skill.
Our results suggest that salpingectomy should
generally be preferred to salpingotomy in women with
tubal pregnancy and a healthy contralateral tube. This
conclusion is lent support by the results of a patient
preference study that showed a strong preference
towards salpingectomy.19 Since our results cannot exclude
the possibility of a very small benet from salpingotomy,
women with a strong preference for maximising their
future pregnancy prospects might still opt for
salpingotomy. However, taking the data of from the ESEP
study and our meta-analysis into account, we believe that
salpingectomy should be the preferred surgical treatment
in women with a tubal pregnancy and a healthy
contralateral tube.
Contributors
FM, WMA, PJH, FvdV, BWM, and AS designed the trial. FM, NMvM,
AS, DJ, KTB, and TMY coordinated the trial. KS, JR, HRV, GCMG,
CAMK, IK, LH, ICAHJ, HK, AH, TCMT-K, FJMB, and WNPW collected
the data. FM and MvW analysed the data. FM drafted the report. All
authors interpreted the data, revised the report, and approved the nal
submitted version.
European Surgery in Ectopic Pregnancy (ESEP) study group
In addition to the authors, the members of the ESEP study group were:
A B Dijkman (Boven IJ Hospital, Amsterdam, Netherlands),
A L Thurkow (St Lucas Andreas Hospital, Amsterdam, Netherlands),
H J H M van Dessel (Twee Steden Hospital, Tilburg, Netherlands),
P J Q van der Linden (Deventer Hospital, Deventer, Netherlands),
F W Bouwmeester (Waterland Hospital, Purmerend, Netherlands),

www.thelancet.com Published online February 3, 2014 http://dx.doi.org/10.1016/S0140-6736(14)60123-9

Articles

G J E Oosterhuis (Medical Spectrum Twente, Enschede, Netherlands),

J J van Beek (VieCuri Medical Centre, Venlo, Netherlands),
M H Emanuel (Spaarne Hospital, Hoofddorp, Netherlands), H Visser
(Ter Gooi Hospital, Blaricum, Netherlands), J P R Doornbos (Zaans
Medical Centre, Zaandam, Netherlands), P J M Pernet (Kennemer
Gasthuis, Haarlem, Netherlands), J Friederich (Gemini Hospital,
Den Helder, Netherlands), F Pettersson (Halland Hospital, Halmstad,
Sweden), Z Sabetirad (Karlstad Central Hospital, Karlstad, Sweden),
K Nilsson (Kvinnokliniken, rebro University Hospital, rebro,
Sweden), G Tegerstedt (South General Hospital, Stockholm, Sweden),
and J J Platz-Christensen (NU Hospital Group, Trollhttan, Sweden).

7
8
9
10

11

Conicts of interest
We declare that we have no conicts of interest.
Acknowledgments
This study was supported by grants from the Netherlands Organisation
for Health Research and Development (ZonMw grants 92003328 and
90700154) and the Region Vstra Gtaland Health & Medical Care
Committee (Sweden). We thank the gynaecologists and research nurses
from the participating hospitals for their dedication and assistance, and
we are grateful to all the women who participated in the trial.

12

References
1
Barnhart KT. Clinical practice. Ectopic pregnancy. N Engl J Med
2009; 361: 37987.
2
Hajenius PJ, Mol F, Mol BW, Bossuyt PM, Ankum WM,
van der Veen F. Interventions for tubal ectopic pregnancy.
Cochrane Database Syst Rev 2007; 1: CD000324.
3
Newbatt E, Beckles Z, Ullman R, Lumsden MA, on behalf of the
Guideline Development Group. Ectopic pregnancy and miscarriage:
summary of NICE guidance. BMJ 2012; 345: e8136.
4
Guidelines and Audit Committee of the Royal College of
Obstetricians and Gynaecologists. Green-top guideline number 21:
the management of tubal pregnancy. London: Royal College of
Obstetricians and Gynaecologists, 2004.
5
Nederlandse Vereniging voor Obstetrie en Gynaecologie. Richtlijn
tubaire EUG, diagnostiek en behandeling. Utrecht: Nederlandse
Vereniging voor Obstetrie en Gynaecologie, 2001.
6
Whitehouse B. Salpingotomy versus salpingectomy in the treatment
of tubal gestation. Proc R Soc Med 1922; 15: 1722.

14

13

15
16

17

18

19

Miller NF. Gynecology; whither bound. Am J Obstet Gynecol 1957;

74: 699704.
Rosenblum JM, Dowling RW, Barnes AC. Treatment of tubal
pregnancy. Am J Obstet Gynecol 1960; 80: 27477.
Richards BC. Persistent trophoblast following conservative
operation for ectopic pregnancy. Am J Obstet Gynecol 1984; 150: 100.
American College of Obstetricians and Gynecologists. ACOG
Practice Bulletin no 94: medical management of ectopic pregnancy.
Obstet Gynecol 2008; 111: 147684.
Mol F, Strandell A, Jurkovic D, et al, for the European Surgery in
Ectopic Pregnancy (ESEP) study group. The ESEP study:
salpingostomy versus salpingectomy for tubal ectopic pregnancy;
the impact on future fertility: a randomised controlled trial.
BMC Womens Health 2008, 8: 11.
Hajenius PJ, Mol BW, Ankum WM, van der Veen F, Bossuyt PM,
Lammes FB. Clearance curves of serum human chorionic
gonadotrophin for the diagnosis of persistent trophoblast.
Hum Reprod 1995; 10: 68387.
Higgins JPT, Altman DG. Chapter 8: assessing risk of bias in
included studies. In: Higgins JPT, Green S, eds. Cochrane
handbook for systematic reviews of interventions, version 5.0.1
(updated September, 2008). The Cochrane Collaboration, 2008.
Mol BW, Matthijsse HC, Tinga DJ, et al. Fertility after conservative
and radical surgery for tubal pregnancy. Hum Reprod 1998;
13: 180409.
Schoenfeld DA. Sample-size formula for the proportional-hazards
regression model. Biometrics 1983; 39: 499503.
Fernandez H, Capmas P, Lucot JP, Resch B, Panel P, Bouyer J, for
the GROG. Fertility after ectopic pregnancy: the DEMETER
randomized trial. Hum Reprod 2013; 28: 124753.
de Bennetot M, Rabischong B, Aublet-Cuvelier B, et al. Fertility
after tubal ectopic pregnancy: results of a population-based study.
Fertil Steril 2012; 98: 127176, e13.
Hajenius PJ, Engelsbel S, Mol BW, et al. Randomised trial of
systemic methotrexate versus laparoscopic salpingostomy in tubal
pregnancy. Lancet 1997; 350: 77479.
van Mello NM, Mol F, Opmeer BC, et al. Salpingotomy or
salpingectomy in tubal ectopic pregnancy: what do women prefer?
Reprod Biomed Online 2010; 21: 68793.

www.thelancet.com Published online February 3, 2014 http://dx.doi.org/10.1016/S0140-6736(14)60123-9