BJOG: an International Journal of Obstetrics and Gynaecology DOI: 10.1111 /j .1471-0528 .2005.0 0746.

x
October 2005, Vol. 112, pp. 1344– 1348

Selective feticide in complicated monochorionic twin pregnancies

using ultrasound-guided bipolar cord coagulation
Romaine Robyr, Masami Yamamoto, Yves Ville
Objective To review our experience with selective feticide in complicated monochorionic (MC) twin preg-
nancies, using ultrasound-guided cord coagulation with a bipolar forceps.
Design Retrospective analysis.
Setting All consecutive umbilical cord coagulations performed at our institution in the second trimester of
pregnancy between November 1999 and 2003.
Population Consecutive cases of complicated MC pregnancies with an indication for selective termination.
Methods Ultrasound-guided coagulation of the umbilical cord with a 2.5-mm bipolar forceps under local
anaesthesia.
Main outcome measures Indications, gestational age at the procedure, perinatal outcome and neonatal
follow up.
Results Forty-six patients with MC twin pregnancies underwent this procedure. Indications included twin
reverse arterial perfusion sequence (n ¼ 17), severe malformation in one twin (n ¼ 7) and agonal
presentation or cerebral anomalies of one twin in twin-to-twin transfusion syndrome (TTTS) after laser
treatment or serial amniodrainage (n ¼ 22). The procedure resulted in six intrauterine fetal demise (IUFD,
13%), with a rate of 41% and 3% when performed at 16 –17 weeks or later, respectively (Fisher P ¼ 0.002).
Preterm rupture of the membranes (PROM) before 28 weeks and between 28 and 34 weeks occurred in 9%
and 14% of the cases, respectively. All neonatal deaths (four) occurred in cases with PROM at 28 weeks or
earlier. Paediatric follow up showed that all infants discharged alive but one were neurologically normal at
3 – 42 months, which corresponds to 70% of the 46 cases.
Conclusions This technique is effective when the natural history is likely to severely affect the development
of the normal co-twin. The overall intact survival rate was 70% and our results support justification of later
surgery. Prematurity remains a significant complication of the procedure.

INTRODUCTION including embolisation, cord ligation,3 external cord com-

Monochorionic (MC) twining occurs in two-thirds of
monozygotic twin pregnancies, accounting for approxi-
mately one-fifth of all twin pregnancies. Specific compli-
cations of MC pregnancies include twin-to-twin transfusion
syndrome (TTTS), twin reversed arterial perfusion (TRAP)
sequence, conjoined twins and congenital anomalies such
as cerebral destructive lesions.1
Monochorionicity makes cord occlusive techniques
the only reasonable options when selective feticide is
contemplated.2 In cases with a risk of in utero fetal demise
of one twin, selective cord occlusion of the affected twin
may benefit the healthy twin by preventing the conse-
quences of exsanguination of this fetus into its dead co-twin.
Several cord occlusion techniques have been described2
pression, radiofrequency ablation and bipolar diathermy
of the cord. Preterm rupture of the membranes is a com-
mon complication in all of them. Cord coagulation in the
affected twin using a bipolar forceps may be the most re-
producible technique from 20 weeks onwards; but large
numbers managed in a consistent way need to be reported
to assess fairly the risks and benefits of these procedures.
Earlier reports of cord coagulation series include series of
TRAP sequence, TTTS and malformations of one twin.
There are 47 cases in four series.
We report our experience with 46 consecutive cases of
bipolar cord coagulation, their complications and neonatal
follow up of the survivors. The results were evaluated chro-
nologically in order to assess any potential learning curve
effect.

Department of Obstetrics and Gynecology, Paris-Ouest METHODS

University, Poissy, France
Correspondence: Professor Y. Ville, Department of Obstetrics and
Gynecology, Paris-Ouest University, CHI Poissy-St-Germain, 10 Rue de
Champ Gaillard, Poissy 78300, France.
D RCOG 2005 BJOG: an International Journal of Obstetrics and Gynaecology
All consecutive cases with cord coagulations from our
first case in November 1999 until December 2003, per-
formed by a single operator, were included. Monochorionicity
www.blackwellpublishing.com/bjog
 SELECTIVE FETICIDE IN MONOCHORIONIC PREGNANCIES REQUIRES CORD OCCLUSION
was ascertained by first trimester ultrasound examination
and confirmed by placental examination after delivery. Both
twins were alive at the time of the procedure.
The indications for the procedure were TRAP sequence,
severe malformation of one twin and agonal condition or
cerebral anomalies in one twin in TTTS treated by laser
coagulation of chorionic plate anastomoses or serial amnio-
reduction. In TTTS cases, agonal was defined as where one
twin showed hydropic features or had episodes of bra-
dycardia with absent or reversed end diastolic flow in the
umbilical artery. Rescue laser could not be performed be-
cause of placental position or heavily bloodstained amni-
otic fluid. Cord coagulation was chosen when there was
either suspicion or evidence of persistence of anastomoses.
For TRAP sequence cases, growing acardiac mass similar
in size to the normal twin or larger was the main indication.
All procedures were performed by the same operator
(YV) using a constant technique. Procedures were per-
formed in the operating room using aseptic technique. Skin
preparation was carried out with 10% iodated polyvidone.
Prophylactic cefazolin (1 g iv) or clindamycin (600 mg iv)
was given 1 hour before surgery. Maternal sedation was
achieved with 1 mg flunitrazepam orally and 100 mg in-
domethacin rectally was given to prevent preterm labour.
Twenty millilitres of 1% xylocaine without adrenaline
were administered locally down to the myometrium and
the procedure was ultrasound guided. A small skin inci-
sion was determined prior to the percutaneous introduc-
tion of a nine French sharp trocar (Karl Storz, Tuttlingen,
Germany) loaded in a 10 French catheter. This was a
10 French disposable catheter originally designed for vas-
cular access for central vein catheterization with a very thin-
walled plastic sheath and a port house with a silicone valve
system for a leakproof seal (Angiocath, Terumo, Leuven,
Belgium). Amnioinfusion of 400 – 800 mL was routinely
used prior to the introduction of the trocar when there was
oligohydramnios within the sac of the affected twin.
Bipolar umbilical cord coagulation was performed using
a 2.5-mm diameter bipolar forceps (Everest Medical, Min-
nesota) similar to previous descriptions.4 – 7 The procedure
aimed at coagulating the cord at the placental or abdominal
insertion. The easier access point was chosen. The insertion
of the trocar was at least 10 cm distant from the expected
point of coagulation to let the forceps open and grasp the
cord. A longitudinal portion of the cord was exposed to the
bipolar forceps, which was introduced with its jaws per-
pendicular to the cord. The correct grasp of the cord was
checked by pulling the bipolar forceps and checking the
passive movements of the cord and of the fetus when the
abdominal insertion was grasped. After grasping the cord an
initial coagulation was performed using 20 W for 10 sec-
onds. The effectiveness of the procedure was judged by the
appearance of echogenic steam bubbles and increased
echogenicity of the coagulated portion of the cord. In the
absence of bubbles, the power was increased in 5 or 10 W
increments to a maximum of 50 W. When the appropriate
D RCOG 2005 BJOG: an International Journal of Obstetrics and Gynaecology 112, pp. 1344 – 1348
1345
power was obtained, the coagulation was determined using
1 or 2 minutes, without releasing the cord. The cord was
coagulated for an average of 10 minutes, even when the
flow stopped earlier. The cardiac activity of the co-twin was
checked by ultrasound during the procedure, immediately
after the procedure, 2 and 12 hours after the procedure. From
case 15 onwards the middle cerebral artery peak systolic
velocity was measured in order to detect fetal anaemia in
the surviving twin 2 and 12 hours after the procedure.
The women were discharged 24– 48 hours after the pro-
cedure. Ultrasound was performed every two weeks. Preg-
nancy outcomes were collected. Most patients delivered in
other hospitals in France.
Paediatric follow up was performed on a clinical basis.
No standardised imaging investigations were performed in
clinically normal infants. Neonatal information and paedi-
atric follow up were obtained following complete clinical
evaluation by the referring paediatrician as well as by direct
contact with the parents.
RESULTS
Forty-six cases were treated during this period, including
42 MC diamniotic and 4 monoamniotic pairs of twins at a
median gestation of 20 (range 16– 35) weeks. The main
indications were life-threatening malformation of one twin
(n ¼ 7), TRAP sequence (n ¼ 17) and severe TTTS with
either agonal features or cerebral anomalies in either the
donor or the recipient twin after laser or serial amniore-
duction treatment (n ¼ 22). None of the co-twins showed
any anatomical defect at the time of the procedure. The
cases are described in Table 1.
Surgical failure was defined as a procedure that could
not stop the flow in the cord. This occurred in two cases
(4%). The first case (N°4) presented at 27 weeks with TRAP
sequence in which the larger diameter of the acardiac mass
exceeded the abdominal circumference of the pump twin.
Blood flow in the acardiac mass could not be stopped and
the patient presented one week later with rupture of the
membranes in spontaneous labour. She delivered sponta-
neously of a 1200-g baby which developed severe respira-
tory distress syndrome and died by day three. The second
case (N°35) was also a case of TRAP in which the pro-
cedure was performed at 16 weeks. Despite amnioinfusion
of 400 mL in this sac, the position of the acardiac mass and
shortness of its cord did not permit grasping of the cord.
The normal fetus died on the following day.
Fetal demise of the healthy twin following cord coag-
ulation occurred in six cases (13%). Fetal demise rate in
cord coagulation cases at 16 and 17 weeks was 41% (5 of
12), compared with 3% (1 of 34) in all cases performed
at 18 weeks or later (Fisher exact test P ¼ 0.002). The
sixth case occurred at 22 weeks in a case of recurrence of
TTTS after laser coagulation of the chorionic plate anas-
tomoses that was treated definitively by cord coagulation.
1346 R. ROBYR ET AL.

Table 1. Description of cases in chronological order of presentation.

Case GA Indication Condition at birth GA at delivery Birthweight Pediatric follow up

1 16 TRAP Alive 36 2300 Normal 5 years

2 24 TTTS with hydrocephalus (r) TOP 32 1870
3 18 TRAP TOP 29 NA
4 27 TRAP Alive 28 1200 Neonatal death, RDS
5 19 TTTS with hydrocephalus Alive 37 2920 Normal 3 months
6 19 Recurrence of TTTS (d) Alive 33 2020 Normal 3 years 6 months
7 18 TRAP Alive 41 3450 Normal 2 years 9 months
8 25 Recurrence TTTS (d) Alive 28 1200 Normal 2 years 9 months
9 21 TTTS with Dandy Walker (d) Alive 33 2310 Normal 2 years 6 months
10 16 Anencephaly IUFD 18 NA
11 22 TTTS with hydrops (r) Alive 26 530 Death at 10 months, BDP
12 16 TRAP IUFD 16 61
13 18 TRAP Alive 39 3310 Normal 2 years
14 16 TRAP Alive 40 3440 Normal 7 months
15 16 TRAP IUFD 16 105
16 24 TTTS with hydrocephalus (d) Alive 31 1515 Development delay 1 year 4 months
17 17 TRAP Alive 36 2330 Normal 13 months
18 18 TTTS with severe IUGR (d) Alive 34 2200 Normal 13 months
19 26 TTTS with heart failure (r) Alive 29 1410 Normal 1 year 3 months
20 16 Acrania Alive 33 2035 Normal 10 months
21 34 Dandy Walker Alive 37 3560 Normal 12 months
22 19 Anencephaly Alive 39 2940 Normal 7 months
23 30 Hydrocephalus Alive 34 2000 Normal 11 months
24 22 TTTS with porencephaly (d) Alive 34 1940 Normal 8 months
25 24 TRAP Alive 34 2260 Normal 8 months
26 16 TRAP Alive 38 3700 Normal 5 months
27 21 TTTS with hydrocephalus (d) Alive 29 1155 Normal 6 months
28 18 TTTS with bradycardia (d) Alive 35 2200 Normal 4 months
29 22 TRAP Alive 38 2590 Normal 1 month
30 20 Spina bifida Alive 38 2685 Normal 2 months
31 18 Multiple malformations Alive 27 650 Neonatal death/prematurity
32 25 Recurrence of TTTS Alive 31 1660 Normal 3 months
33 20 TRAP Alive 29 1170 Normal 3 months
34 16 TRAP Alive 38 2300 Normal 1 month
35 16 TRAP IUFD 17 70
36 20 TTTS with heart failure (r) Alive 27 980 Neonatal death, ICH
37 19 TRAP Alive 39 2800 Normal 6 months
38 22 Recurrence of TTTS (d) IUFD 22 800
39 23 Recurrence TTTS (d) Alive 26 840 Normal 2 months
40 17 TTTS ventriculomegaly (d) IUFD 17 160
41 32 TTTS with cerebral atrophy (r) Alive 34 1915 Normal 3 months
42 25 TTTS with hydrops (r) Alive 29 1130 Normal 2 months
43 25 TTTS with bradycardia (d) Alive 34 2000 Normal 2 months
44 17 TRAP Alive 26 700 Neonatal death, ICH
45 19 TTTS with severe IUGR (d) Alive 29 1160 Normal 1 month
46 26 TTTS with severe IUGR (d) Alive 36 2400 Normal 1 month

TRAP ¼ transfusion reversal arterial perfusion; IUGR ¼ intrauterine growth restriction; NA ¼ not available; ICH ¼ intracranial haemorrhage; BPD ¼
bronchopulmonary dysplasia; d ¼ donor; r ¼ recipient; TOP ¼ termination of pregnancy; RDS ¼ respiratory distress syndrome.

Intrauterine fetal demise (IUFD) occurred within 24 hours
in four cases, and within two weeks in all cases. Post-
mortem examination was available in four cases, with no
evidence of placental abruption, feto-fetal haemorrhage or
cord entanglement.
Termination of pregnancy was requested and performed
in two cases. In one case (N°2) of TTTS, subsequent exam-
ination of the surviving co-twin revealed evidence of hy-
perechogenic kidneys and poor renal function. In another
case (N°3), cerebral atrophy, confirmed by MR scanning,
developed in the survivor.
D RCOG 2005 BJOG: an International Journal of Obstetrics and Gynaecology 112, pp. 1344 – 1348
Rupture of the membranes before 34 weeks was the main
complication recorded, occurring in eleven cases (23.4%).
This occurred within one month of the procedure in five
cases (10.4%) and later in six cases (13%). Four cases (8.7%)
occurred before 28 weeks of gestation, and two of these in-
fants subsequently died of respiratory complications.
Three surviving twins developed signs of anaemia and
were given intrauterine blood transfusions within 24 hours
of the procedure.8 All were diagnosed by middle cerebral
artery PSV measurements, considered as severe. Follow up
and neonatal outcome were normal in all three cases.
 SELECTIVE FETICIDE IN MONOCHORIONIC PREGNANCIES REQUIRES CORD OCCLUSION
Neonatal death occurred in four cases (8.6%), all born
before 29 weeks and related to prematurity. There were two
long term complications in infants surviving the neonatal
period. One child died at aged 10 months from chronic lung
disease (N°11) and another (N°16) delivered at 31 weeks of
gestation and six weeks after cord coagulation had devel-
opmental delay at aged 18 months, although without any
abnormality of brain imaging. This was a complex preg-
nancy with systemic lupus erythematosus.
DISCUSSION
In MC pregnancies, life-threatening condition of one
twin in utero can cause severe haemodynamic imbalance
and intrauterine death of the healthy co-twin in at least 25%
of the cases and around 25% of survivors have neurolog-
ical sequelae.1 This is thought to be due to exsanguination
of the normal twin into its co-twin and placenta.9 Selective
feticide in MC pregnancies can only be performed by using
cord occlusion techniques.10 Although cases of cord coag-
ulation using bipolar diathermy have been reported, larger
numbers are required in order to provide accurate counsel-
ing to couples considering this procedure. Of nine reported
cases of TRAP, seven co-twins survived.4 – 6 All six infants
with severe malformations in the co-twin survived.5 Where
TTTS was the indication, although eligibility criteria dif-
fered, 26 of 32 co-twins survived.4 – 7 The overall survival
rate was 82%. Our data are consistent with these smaller
series with 74% survival, including long term follow up.
The main complication was fetal death and early gesta-
tion appears to be a risk factor. This suggests that cord
coagulation should be postponed until 18 weeks. However,
early prenatal diagnosis may have selected more severe
cases, which may not have continued to a later gestation.
Surgical failure as defined by us was uncommon. The
main factor contributing to this was the difficulty in grasp-
ing the cord, in part explained by the large size and position
of the acardiac mass.
Preterm prelabour rupture of the membranes was com-
mon and contributed to our perinatal mortality. This may
suggest that the procedure should be considered later in
gestation when the condition of the affected twin is not
immediately life threatening of the healthy fetus. However,
this can raise both ethical and legal considerations in some
circumstances. It may run the risk of preterm labour with
possible delivery while awaiting optimal procedure timing.
Although our results are comparable to other small se-
ries, we analysed the occurrence of complications in groups
of 10 cases in chronological order. Fetal demise rate re-
mained constant over time.
Only local anaesthesia and sedation was required. There
were no serious maternal complications.
At least three survivors were found to be anaemic fol-
lowing the procedure presumably in relation with an incom-
plete coagulation of the cord in the affected twin. Indeed
D RCOG 2005 BJOG: an International Journal of Obstetrics and Gynaecology 112, pp. 1344 – 1348
1347
color Doppler stops being useful for monitoring the qual-
ity of coagulation as soon as the target fetus dies and this
may happen before occlusion can be completed. This sug-
gests closer follow up of cases using middle cerebral artery
Doppler.8 This small series cannot assess the benefit of
early recognition and treatment of fetal anaemia.
One important lesson from this series is that the proce-
dure is not finished when flow has stopped in the cord,
because this may also be produced by the grasping or by a
spasm. Cord coagulation should last for as long as neces-
sary to completely obliterate the cord, and this is likely to
occur after the flow stop and probably also after the cardiac
arrest. The minimum time needed to achieve coagulation
could not be defined in this study, but changes in echoge-
nicity in the cord could guide the operators that the proce-
dure is completed. It is likely that direct visualisation of
the cord and of the forceps during coagulation may de-
crease the duration of coagulation procedure and the need
for subsequent transfusion because complete occlusion by
the forceps could be controlled by direct vision.
Overall paediatric follow up from birth to 42 months was
normal in all survivors but one, which may suffer from an
unrelated condition.
Umbilical cord coagulation of the affected MC twin led
to the birth of a liveborn child in 83% of the cases, with
deaths related to prematurity in 11%. The overall long term
survival without handicap was 70%. Given the indications
for the procedure this represents good overall outcome. The
two main contributing factors for poor outcome appear to
be performance of the procedure before 18 weeks of ges-
tation and the occurrence of preterm PROM leading to
preterm delivery and neonatal death. We recommend that,
where appropriate, the procedure should be delayed until as
late in gestation as possible, and only exceptionally per-
formed at gestations less than 18 weeks.
References
1. Bejar R, Vigliocco G, Gramajo H, et al. Antenatal origin of neu-
rologic damage in newborn infants. Multiple gestations. Am J Obstet
Gynecol 1990;162(5):1230 – 1236.
2. Ville Y. Opinion. Selective feticide in monochorionic pregnancies:
toys for the boys or standard of care? Ultrasound Obstet Gynecol
2003;22(5):448 – 450.
3. Quintero RA, Romero R, Reich H, et al. In utero percutaneous umbil-
ical cord ligation in the management of complicated monochorionic
multiple gestations. Ultrasound Obstet Gynecol 1996;8(1):16 – 22.
4. Gallot D, Laurichesse H, Lemery D. Selective feticide in monocho-
rionic twin pregnancies by ultrasound-guided umbilical cord occlu-
sion. Ultrasound Obstet Gynecol 2003;22(5):484 – 488.
5. Nicolini U, Poblete A, Boschetto C, et al. Complicated monochorionic
twin pregnancies: experience with bipolar cord coagulation. Am J
Obstet Gynecol 2001;185(3):703 – 707.
6. Deprest J, Audibert F, Van Schoubroeck D, et al. Bipolar coagulation
of the umbilical cord in complicated monochorionic twin pregnancy.
Am J Obstet Gynecol 2000;182(2):340 – 345.
7. Taylor MJ, Shalev E, Tanawattanacharoen S, et al. Ultrasound-guided
umbilical cord occlusion using bipolar diathermy for Stage III/IV
1348 R. ROBYR ET AL.

twin – twin transfusion syndrome. Prenat Diagn 2002;22(1):70 –
76.
8. Senat MV, Loizeau Couderc S, Bernard JP, Ville Y. The value of
middle cerebral artery peak systolic velocity in the diagnosis of fetal
anemia after intrauterine death of one monochorionic twin. Am J
Obstet Gynecol 2003;189(5):1320 – 1324.
9. Fusi L, McParland P, Fisk N, et al. Acute twin – twin transfusion: a
possible mechanism for brain-damaged survivors after intrauterine
death of a monochorionic twin. Obstet Gynecol 1991;78(3):517 – 520.
10. Challis D, Gratacos E, Deprest J. Cord occlusion techniques for se-
lective termination in monochorionic twins. J Perinat Med 1999;
27(5):327 – 338.
Accepted 20 June 2005

D RCOG 2005 BJOG: an International Journal of Obstetrics and Gynaecology 112, pp. 1344 – 1348