Research ajog.

org

OBSTETRICS
Mode of delivery of twin gestation with very
low birthweight: is vaginal delivery safe?
Eran Barzilay, MD, PhD; Shali Mazaki-Tovi, MD; Uri Amikam, BSc; Hila de Castro, MD;
Jigal Haas, MD; Ram Mazkereth, MD; Eyal Sivan, MD; Eyal Schiff, MD; Yoav Yinon, MD

OBJECTIVE: The purpose of this study was to determine whether
planned vaginal delivery is associated with increased risk of perinatal
death and morbidity in twin pregnancies that are complicated by a very
low birthweight of the second twin.
STUDY DESIGN: We conducted a retrospective cohort study of twin
pregnancies in which the second twins birthweight was
1500 g.
One hundred ninety-three twin gestations met the study criteria; pa-
tients were classified into 2 groups according to the planned mode of
delivery: (1) cesarean delivery (n 142) and (2) vaginal delivery (n
51). In the vaginal delivery group, 21 pairs were in cephalic-cephalic
presentation at the time of delivery; 28 pairs were cephalic-
noncephalic, and 2 pairs were noncephalic-noncephalic. Composite
adverse neonatal outcome was defined as the presence of neonatal
death, respiratory distress syndrome, sepsis, necrotizing enterocolitis,
or intraventricular hemorrhage grade 3-4.
RESULTS: Trial of vaginal delivery was successful for both twins in
90.5% of cephalic-cephalic twins and 96.4% in cephalic-noncephalic
twins. The rate of intraventricular hemorrhage was significantly higher
in the vaginal delivery group (29.4% vs 8.5%, respectively; P .013;
adjusted odds ratio [OR], 3.65; 95% confidence interval [CI],
1.32e10.1). The increased risk of intraventricular hemorrhage in the
vaginal delivery groups was evident in both twin A (17.6% vs 7.0%;
P .029) and twin B (15.7% vs 4.9%; P .014); however, these
differences were not significant after adjustment for possible con-
founders (twin A: adjusted OR, 1.79; 95% CI, 0.58e5.55; twin B:
adjusted OR, 2.13; 95% CI, 0.63e7.25). In addition, subgroup
analysis revealed that both cephalic-cephalic and cephalic-
noncephalic twins who were delivered vaginally had increased risk
for intraventricular hemorrhage. There were no significant differences
between the cesarean and vaginal delivery groups in the rates of Apgar
score <7 at 5 minutes, arterial cord pH <7.1, composite adverse
neonatal outcome, and neonatal mortality rate. However, the rate of
respiratory distress syndrome was significantly lower in the vaginal
delivery group (66.7% vs 69%; P .042; OR, 0.34; 95% CI,
0.12e0.96).
CONCLUSION: Vaginal delivery of very low birthweight twins is asso-
ciated with an increased risk of intraventricular hemorrhage, regard-
less of presentation. Because of the small sample size and the
retrospective cohort design, large prospective randomized studies are
needed.
Key words: intraventricular hemorrhage, mode of delivery, twin de-
livery, very low birthweight

Cite this article as: Barzilay E, Mazaki-Tovi S, Amikam U, et al. Mode of delivery of twin gestation with very low birthweight: is vaginal delivery safe? Am J Obstet Gynecol
2015;212:.

T he rate of twin gestations has

increased over the last 30 years,
mainly because of the growing rate of
have led to a signicant increase in the
rate of delivery of very low birthweight
(VLBW; <1500 g) twins.5,6
deliveries are associated with increas-
ed fetal risk, mainly for the second
twin,7-13 other cohort studies,14-19 2
pregnancies that have been achieved via The optimal method of delivery of metaanalyses20,21 and a small prospec-
assisted reproductive techniques.1-4 The twins is controversial; several studies tive randomized study22 have shown no
rise in twin pregnancies accompanied by have addressed this question. Although difference in perinatal outcome between
the higher risk of premature delivery, some retrospective population-based planned vaginal and planned cesarean
which characterizes such pregnancies, studies have reported that vaginal deliveries. Barrett et al23 recently have
reported on a large prospective ran-
domized controlled trial, in which a
From Sackler School of Medicine, Tel-Aviv University, Tel Aviv (all authors), and the Department of
Obstetrics and Gynecology (Drs Barzilay, Mazaki-Tovi, de Castro, Haas, Sivan, Schiff, and Yinon) and
total of 2804 twin pregnancies between
the Department of Neonatology, Edmond and Lili Safra Childrens Hospital (Dr Mazkereth), Sheba 32 and 39 weeks gestation were assigned
Medical Center, Tel-Hashomer, Israel. randomly to planned vaginal or planned
Received Dec. 4, 2014; revised Feb. 9, 2015; accepted March 17, 2015. cesarean delivery. In this study, planned
The authors report no conict of interest. vaginal delivery did not increase risk for
Presented at the 33rd annual meeting of the Society for Maternal-Fetal Medicine, San Francisco, CA, neonatal morbidity or death.
Feb. 11-16, 2013. However, the data in the literature on
Corresponding author: Eran Barzilay, MD, PhD. eran.barzilay@gmail.com the optimal mode of delivery for severe
0002-9378/$36.00  2015 Elsevier Inc. All rights reserved.  http://dx.doi.org/10.1016/j.ajog.2015.03.030 premature or VLBW twins are limited.
Despite the limited information available,

MONTH 2015 American Journal of Obstetrics & Gynecology 1.e1

Research Obstetrics
some authors have advocated for cesarean
deliveries in cases of a second twin in a
nonvertex presentation and with an esti-
mated weight of <1500-2000 g.24,25 This
view largely relies on extrapolation of
data from breech deliveries of singleton
gestation.26,27
Other investigators have shown
adverse outcome in VLBW twins who
undergo vaginal deliveries and proposed
that cesarean delivery is the optimal
route of delivery for all twins who are
expected to have a birthweight <1500 g,
regardless of presentation.28
Decisions regarding mode of delivery
in VLBW twins are based on very limited
data. This is especially true for preg-
nancies with a VLBW second twin in a
nonvertex presentation. Therefore, we
aimed to evaluate the association be-
tween the mode of delivery and perinatal
morbidity and death in twin pregnancies
that are complicated by VLBW of the
second twin.
M ETHODS
All twin deliveries in a single tertiary care
medical center from August 2004 to
April 2011 were reviewed. Inclusion
criteria included (1) twin gestation and
(2) second twin birthweight of
1500 g.
Exclusion criteria included (1) gesta-
tional age at delivery of <24 weeks, (2)
fetal death of 1 or both twins before la-
bor, and (3) major malformation diag-
nosed in 1 or both twins.
The standard of care in our medical
center regarding twin delivery is to allow
vaginal delivery of cephalic-cephalic and
cephalic-noncephalic twins, regardless
of their estimated weight or the gesta-
tional age, as long as the estimated
weight of twin B is not signicantly
higher (20%) than twin A. The de-
livery is supervised by an experienced
obstetrician under continuous fetal
monitoring, and the preferred method
for delivering the noncephalic second
twin is total breech extraction with or
without internal podalic version.
A total of 193 twin gestations met the
study criteria. Patients were classied
into 2 groups according to the planned
mode of delivery: cesarean delivery (n
142) and vaginal delivery (n 51). In
the vaginal delivery group, 21 pairs
1.e2 American Journal of Obstetrics & Gynecology MONTH 2015
were cephalic-cephalic; 28 pairs were
cephalic-noncephalic, and 2 pairs were
noncephalic-noncephalic. The charts of
all women and their infants were
reviewed for the variables of interest.
Maternal characteristics and their preg-
nancy outcomes were abstracted from
the obstetric electronic charts. The
following neonatal outcomes were
examined: Apgar scores at 5 minutes,
cord blood PH, death, birth trauma
(spinal cord injury, skull fracture, frac-
ture of a long bone, peripheral nerve
injury, and subdural or intracerebral
hemorrhage), respiratory distress syn-
drome (RDS), sepsis, necrotizing en-
terocolitis (NEC), and intraventricular
hemorrhages (IVH). Composite adverse
neonatal outcome was dened as the
presence of neonatal death, RDS, sepsis,
NEC, or IVH grade 3-4.
Because decisions regarding the
planned mode of delivery are made per
pregnancy for both twins, we chose to
analyze our data as outcomes per preg-
nancy. Outcome measures were dened as
outcomes for twin A, twin B, or any twin.
Comparison of continuous variables
between the 2 groups was conducted
using Mann-Whitney U test or the Stu-
dent t test, as appropriate. Chi-square
test was used for comparison of cate-
goric variables. Logistic regression anal-
ysis was used to examine the relationship
between mode of delivery and neonatal
outcome measures. Adjustment was
conducted for gestational age at delivery,
chorionicity, and antenatal steroid
treatment because these factors have
been shown to be associated signicantly
with perinatal death and neonatal
morbidity and thus may have a con-
founding potential.29,30 The regression
model was limited to 4 variables to ac-
count for the small sample size.31 How-
ever, we also used an alternate regression
model that adjusted for maternal age,
parity, and birthweight in addition to
gestational age at delivery, chorionicity,
and antenatal steroid treatment. This
model was not used in subgroup analysis
according to presentation because the
subgroups were signicantly smaller.
Signicance was accepted at a probabil-
ity value of < .05. Statistical analyses
were conducted with the IBM Statistical
ajog.org
Package for the Social Sciences (IBM
SPSS version 19; IBM Corporation Inc,
Armonk, NY).
The study was approved by the local
institutional review board.
R ESULTS
A total of 193 twin deliveries that were
complicated with VLBW of the second
twin were identied during the study
period. Of them, 142 delivered by cesarean
section without trial of vaginal delivery,
and 51 underwent trial of delivery. The
vaginal delivery group included 2 cases of
noncephalic-noncephalic twin pregnan-
cies that attempted vaginal delivery de-
spite our consultation to have a cesarean
delivery. The main indications for cesar-
ean delivery were a noncephalic rst twin
(35.3%), intrauterine growth restriction
(17.6%), maternal request (12.5%), sus-
pected fetal distress (11%), and previous
cesarean delivery (5.9%). Demographic
and clinical characteristics of the patients
are presented in Table 1. The groups did
not differ with regards to maternal age,
parity, body mass index, rate of gestational
diabetes mellitus, male/female ratio, and
rate of exposure to antenatal steroids.
However, the median gestational age at
delivery was 31.4 weeks in the cesarean
delivery group compared with 30.4 weeks
of gestation in the vaginal delivery group
(P .025). Despite the signicant differ-
ence in gestational age at delivery, median
birthweights of both twins were not
signicantly different between the 2
groups (twin A: cesarean delivery 1417.5 g
vs vaginal delivery: 1335 g; P .183; twin
B: 1258 vs 1195 g; P .654). The rate of
monochorionic twins was signicantly
higher in the cesarean delivery group
compared with the vaginal delivery group
(33.8% vs 11.8%; P .034). The mono-
chorionic pregnancies in the cesarean
group included 4 cases of monoamniotic
pregnancies. The higher rate of mono-
chorionic pregnancies was adjusted for in
our regression model.
In 8 cases in the vaginal delivery
group, labor was induced because of
intrauterine growth restriction. Out-
come in induced twin pregnancies was
generally favorable (no cases of RDS,
IVH, or death; 1 neonate with NEC),
most likely because of the relatively
ajog.org Obstetrics Research
advanced gestational age at delivery of
these cases (median gestational age, 34.1 TABLE 1
wk; interquartile range, 33.5e34.5). Demographic and clinical parameters
Mode of delivery in the trial of vaginal Cesarean delivery Vaginal delivery
delivery group is displayed in Table 2. The Variable (n [ 142) (n [ 51) P valuea
b
rates of vaginal delivery of both twins Maternal age, y 31 (29e35) 31 (27e34.5) .281
were high for both the cephalic-cephalic Body mass index, kg/m 2
26.2 (23.6e29.1) 24.4 (21.8e27.8) .121
twins and the cephalic-noncephalic
Parity  1, n (%) 47 (33.1) 13 (25.5) .633
(90.5% and 96.4%, respectively). Only 1
case (from the cephalic-cephalic sub- Monochorionic, n (%) 48 (33.8) 6 (11.8) .034
group) resulted in cesarean delivery of Gestational diabetes 17 (12) 8 (15.7) .498
twin B after a vaginal delivery of twin A, mellitus, n (%)
and 2 cases (1 in each subgroup) ended in Exposure to steroids, n (%) 142 (100) 49 (96.1) .722
cesarean delivery of both twins. There
Gestational age at 31.4 (29.4e33) 30.4 (28e31.9) .025
were no cases of documented birth delivery, wkb
trauma in either group, and there were no
Birthweight, gb
cases of intrapartum death. Excluding the
cases that culminated in a cesarean de- Twin A 1417.5 (1135e1700) 1335 (1042e1562) .183
livery, all second twins in cephalic- Twin B 1258 (997e1395) 1195 (1007e1372) .654
noncephalic twin pregnancies were
Male/female ratio, n/N
delivered by total breech extraction.
Table 3 compares the neonatal Twin A 71/71 26/25 .904
outcome between the 2 groups. The Twin B 69/73 26/25 .77
rate of low 5-minute Apgar score and a b
Calculated with the Mann-Whitney U test or the chi-squared test; Data are expressed as median (interquartile range).
low arterial cord pH (pH <7.1) were Barzilay. Mode of delivery in VLBW twins. Am J Obstet Gynecol 2015.
similar between the 2 groups. Similarly,
there were no signicant differences in
the rates of mortality, composite adver- 0.28; 95% CI, 0.09e0.88) and in twin A Table 4). A signicantly increased risk for
se outcome, sepsis, NEC, or neonatal (OR, 0.34; 95% CI, 0.13e0.9); that the IVH was demonstrated in the vaginal de-
intensive-care unit admission. risk for RDS in twin A was signicantly livery group compared with the cesarean
IVH was signicantly more common lower in the vaginal delivery group (OR, delivery group, both for the cephalic-
in the vaginal delivery group (29.4% vs 0.34; 95% CI, 0.13e0.88). cephalic twins (33.3% vs 8.5%; P .022;
8.5%; P < .001) with an adjusted odds Because the presentation of the fetuses OR, 4.35; 95% CI, 1.24e15.38) and the
ratio (OR) of 3.65 (95% condence in- in twin pregnancies is pivotal in cephalic-noncephalic twins (28.6% vs
terval [CI], 1.32e10.1). The rate of decision-making regarding mode of 8.5%; P .025; OR, 4.88; 95% CI,
grade 3-4 IVH for any twin was also delivery and because the main clinical 1.22e19.23). However, IVH grade 3-4
higher in the vaginal delivery group controversy is regarding delivery of seemed to be more common only
(11.8% vs 2.1%), however, statistical cephalic-noncephalic twin pregnancies, among the cephalic-noncephalic sub-
signicance was lost after adjustment we conducted a subanalysis according group (17.9% vs 2.1%, respectively; P <
for confounders (adjusted OR, 2.56; to the presentation of the twins in .001), although this difference did not
95% CI, 0.46e14.08). The rate of RDS, the vaginal delivery group (either reach statistical signicance after adjust-
although not considerably different be- cephalic-cephalic or cephalic-noncephalic; ment (OR, 6.58; 95% CI, 0.96e45.45). A
tween the groups, appeared to be
signicantly lower in the vaginal de-
livery group compared with the cesar-
TABLE 2
ean delivery group after adjustment for
Mode of delivery in the vaginal delivery-intention group
confounders (66.7% vs 69%; P .042;
Cephalic-cephalic Cephalic-noncephalic
OR, 0.34; 95% CI, 0.12e0.96). When
Variable (n [ 21), n (%) (n [ 28), n (%)
the regression model was altered to
include more confounders, which Vaginal delivery for both twins 19 (90.5) 27 (96.4)
included birthweight, maternal age, and Cesarean delivery for both twins 1 (4.8) 1 (3.6)
parity (regression model 2), the only Cesarean delivery for twin B 1 (4.8) 0
differences that were noted in our re- after vaginal delivery for twin A
sults were that the risk for composite Data are expressed as number (percentage).
outcome was signicantly lower in the Barzilay. Mode of delivery in VLBW twins. Am J Obstet Gynecol 2015.
vaginal delivery group in any twin (OR,

MONTH 2015 American Journal of Obstetrics & Gynecology 1.e3

Research Obstetrics ajog.org

TABLE 3
Neonatal outcome
Cesarean delivery Vaginal delivery Adjusted Adjusted odds ratio (95% CI)
Variable (n [ 142), n (%) (n [ 51), n (%) P value P value Model 1 Model 2
Any twin
5-Min Apgar score <7 9 (6.3) 2 (3.9) .523 .09 0.2 (0.03e1.29) 0.26 (0.04e1.83)
Umbilical artery pH <7.1 a
2/74 (2.7) 0/31 (0) .362 .995 NA b
NAb
Composite outcome 107 (75.4) 37 (72.5) .693 .076 0.4 (0.14e6.91) 0.28 (0.09e0.88)
Death 8 (5.6) 4 (7.8) .575 .679 0.74 (0.18e3.08) 1.29 (0.25e6.55)
Respiratory distress syndrome 98 (69) 34 (66.7) .757 .042 0.34 (0.12e0.96) 0.25 (0.08e0.76)
Necrotizing enterocolitis 7 (4.9) 2 (3.9) .77 .809 0.82 (0.16e4.17) 1.37 (0.23e8.28)
Sepsis 18 (35.3) 48 (33.8) .847 .385 0.71 (0.33e1.53) 0.61 (0.26e1.47)
Intraventricular hemorrhage 12 (8.5) 15 (29.4) < .001 .013 3.65 (1.32e10.1) 4.86 (1.54e15.34)
Intraventricular hemorrhage 3 (2.1) 6 (11.8) .005 .281 2.56 (0.46e14.08) 4.88 (0.51e47.1)
grade 3-4
Neonatal intensive care 129 (90.8) 46 (90.2) .891 .751 0.78 (0.17e3.55) 0.69 (0.14e3.46)
unit admission
Twin A
5-Min Apgar score <7 4 (2.8) 0 .226 .997 NAb NAb
Umbilical artery pH <7.1 a
2/91 (2.2) 0/32 (0) .398 .997 NA b
NAb
Composite outcome 87 (61.3) 31 (60.8) .952 .116 0.49 (0.2e1.19) 0.34 (0.13e0.9)
Death 3 (2.1) 3 (5.9) .183 .716 0.66 (0.07e5.99) 0.86 (0.08e9.75)
Respiratory distress syndrome 83 (58.5) 29 (56.9) .844 .08 0.46 (0.19e1.1) 0.34 (0.13e0.88)
Necrotizing enterocolitis 2 (1.4) 1 (2.0) .784 .819 1.33 (0.11e15.87) 4.14 (0.17e102.4)
Sepsis 21 (14.8) 11 (21.6) .264 .871 0.92 (0.35e2.43) 0.68 (0.23e2)
Intraventricular hemorrhage 10 (7.0) 9 (17.6) .029 .311 1.79 (0.58e5.55) 2.16 (0.6e7.82)
Intraventricular grade 3-4 2 (1.4) 4 (7.8) .023 .321 2.66 (0.38e18.52) 3.47 (0.34e35.7)
Neonatal intensive care 111 (78.2) 42 (82.4) .527 .882 0.9 (0.22e3.56) 0.59 (0.13e2.74)
unit admission
Twin B
5-Min Apgar score <7 6 (4.3) 2 (3.9) .919 .152 1.06 (0.46e2.44) 0.22 (0.02e2.43)
Umbilical artery pH <7.1 a
1/88 (1.1) 0/37 (0) .515 .996 NA b
NAb
Composite outcome 94 (73.4) 34 (66.7) .951 .115 0.45 (0.17e1.21) 0.38 (0.13e1.1)
Death 6 (4.2) 4 (7.8) .317 .972 1.03 (0.23e4.57) 5.24 (0.26e105.3)
Respiratory distress syndrome 79 (55.6) 32 (62.7) .378 .499 0.73 (0.29e1.84) 0.41 (0.14e1.19)
Necrotizing enterocolitis 7 (4.9) 1 (2.0) .362 .389 0.39 (0.05e3.32) 0.67 (0.07e6.5)
Sepsis 37 (26.1) 12 (23.5) .722 .241 0.61 (0.27e1.39) 0.76 (0.3e1.93)
Intraventricular hemorrhage 7 (4.9) 8 (15.7) .014 .224 2.13 (0.63e7.25) 4.47 (0.99e20.1)
Intraventricular hemorrhage 2 (1.4) 3 (5.9) .084 .948 0.92 (0.08e10.53) 1.06 (0.04e27.1)
grade 3-4
Neonatal intensive care 124 (87.3) 45 (88.2) .866 .809 0.85 (0.23e3.07) 0.82 (0.22e3.1)
unit admission
Logistic regression was performed to correct for possible confounders that included gestational age, antenatal steroids, and chorionicity (model 1) or gestational age, antenatal steroids, chorionicity,
maternal age, birthweight, and parity (model 2).
NA, not available.
a
Umbilical artery pH was not available for all cases; the effective denominator is noted for each group; b Odds ratios were not calculated for cases with extremely high probability values (>.99).
Barzilay. Mode of delivery in VLBW twins. Am J Obstet Gynecol 2015.

1.e4 American Journal of Obstetrics & Gynecology MONTH 2015

ajog.org Obstetrics Research

TABLE 4
Neonatal outcome: subanalysis according to presentation
Vaginal delivery Adjusted odds ratio
Cesarean delivery cephalic-cephalic Adjusted (95% confidence
Variable (n [ 142), n (%) only (n [ 21), n (%) P value P value interval)
Composite outcome 107 (75.4) 16 (76.2) .934 .343 0.46 (0.09e2.3)
Death 8 (5.6) 1 (4.8) .87 .625 0.58 (0.06e5.18)
Respiratory 98 (69) 15 (71.4) .823 .251 0.4 (0.09e1.9)
distress syndrome
Intraventricular hemorrhage 12 (8.5) 7 (33.3) .001 .022 4.35 (1.24e15.38)
Intraventricular 3 (2.1) 1 (4.8) .464 .79 1.5 (0.07e30.3)
hemorrhage grade 3-4
Vaginal delivery
cephalic-noncephalic
only (n 28)
Composite outcome 107 (75.4) 19 (67.9) .408 .072 0.32 (0.09e1.11)
Death 8 (5.6) 2 (7.1) .756 .744 0.74 (0.12e4.63)
Respiratory 98 (69) 17 (60.7) .391 .041 0.27 (0.08e0.95)
distress syndrome
Intraventricular hemorrhage 12 (8.5) 8 (28.6) .003 .025 4.88 (1.22e19.23)
Intraventricular 3 (2.1) 5 (17.9) < .001 .055 6.58 (0.96e45.45)
hemorrhage grade 3-4
Logistic regression was performed to correct for possible confounders that included gestational age, antenatal steroids, and chorionicity.
Barzilay. Mode of delivery in VLBW twins. Am J Obstet Gynecol 2015.

decreased risk for RDS in the vaginal de-
livery group was signicant after adjust-
ment only in the cephalic-noncephalic
subgroup (60.7% vs 69%; P .041; OR,
0.27; 95% CI, 0.08e0.95). There were no
signicant differences compared with the
cesarean delivery group in all other
outcome measures (NEC, sepsis, neonatal
intensive care unit admission, low 5-
minute Apgar score, and low arterial pH)
for either cephalic-cephalic twins or
cephalic-noncephalic twins (data not
shown).
Both cases of noncephalic-noncephalic
pregnancies that attempted vaginal de-
livery achieved vaginal delivery without
birth trauma or entrapment of the after-
coming head. One pair was born at 24
weeks of gestation (both twins died during
the rst 2 days of life). The other patient
delivered at 31 weeks of gestation (both
infants had RDS; twin A had sepsis, but
neither twin had IVH).
Next, we sought to assess whether our
results will change if we divided the
groups by mode of delivery instead of
intended mode of delivery. There were
144 twin pregnancies that delivered both
twins by cesarean delivery and 48 twin
pairs that delivered both twins vaginally.
The median gestational age was lower in
the vaginal delivery group compared
with the cesarean delivery group (30.4 vs
31.3 weeks); albeit, this difference was
not statistically signicant (P .073).
Similar to analysis by intended mode of
delivery, vaginal delivery was associated
with a signicant increase in risk for IVH
compared with cesarean delivery, a trend
for increased risk for IVH grade 3-4 and
a signicant reduction in risk for RDS
(Table 5).
Finally, we examined whether ex-
cluding the cases with a presenting
noncephalic presentation will change
our results. Similar to our primary
analysis, vaginal delivery was associated
with a statistically signicant increase in
risk for IVH, a trend for increased risk
for IVH grade 3-4 and a signicant
reduction in risk for RDS (Table 6). The
data depicted in Tables 5 and 6 was also
analyzed with the use of the second
regression model. In both cases, when
the regression model was altered to
include more confounders, the only
difference that was noted in our results
was that the risk for composite outcome
was signicantly lower in the vaginal
delivery group.
C OMMENTS
In this study, we examined the safety of
vaginal delivery of twins that was
complicated by VLBW of the second
twin. There were no signicant differ-
ences in the rates of low Apgar score, low
arterial cord pH, composite adverse
neonatal outcomes, or death between the
vaginal delivery group and the cesarean
delivery group. However, we have
demonstrated that vaginal delivery of
VLBW twins is associated with an
increased risk of IVH. This increase in
risk was similar in cephalic-cephalic and
cephalic-noncephalic twins. The differ-
ence in rate of IVH grade 3-4 did not
reach statistical signicance, possibly
because of the small sample size. On the
other hand, the risk for RDS appeared to
be decreased in the vaginal delivery

MONTH 2015 American Journal of Obstetrics & Gynecology 1.e5

Research Obstetrics ajog.org

TABLE 5
Neonatal outcome: subanalysis according to actual mode of delivery
Adjusted odds ratio
Cesarean delivery Vaginal delivery (95% confidence interval)
Variable (n [ 144), n (%) (n [ 48), n (%) P value Adjusted P value Model 1 Model 2
Composite outcome 109 (75.7) 34 (70.8) .503 .07 0.39 (0.14e1.08) 0.28 (0.09e0.87)
Death 8 (5.6) 4 (8.3) .491 .809 0.84 (0.2e3.5) 1.45 (0.29e7.32)
Respiratory distress syndrome 100 (69.4) 31 (64.6) .531 .038 0.36 (0.12e0.94) 0.24 (0.08e0.74)
Intraventricular hemorrhage 13 (9) 14 (29.2) .001 .017 3.53 (1.26e9.9) 4.58 (1.47e14.26)
Intraventricular hemorrhage 3 (2.1) 6 (12.5) .003 .217 2.96 (0.53e16.39) 5.73 (0.61e53.5)
grade 3-4
Logistic regression was performed to correct for possible confounders that included gestational age, antenatal steroids, and chorionicity (model 1) or gestational age, antenatal steroids, chorionicity,
maternal age, birthweight, and parity (model 2).
Barzilay. Mode of delivery in VLBW twins. Am J Obstet Gynecol 2015.

group compared with the cesarean de-
livery group. Thus, vaginal delivery
may have a protective effect against RDS,
as has been suggested in previous
studies.32-34
Because low-grade IVH is considered
to have less clinical impact compared
with high-grade IVH, the clinical sig-
nicance of an increase in IVH rate
without a signicant increase in the rate
of high-grade IVH is yet to be deter-
mined. Nevertheless, studies have shown
that low-grade IVH can impact the
neonatal brain35 and that low-grade
IVH may be associated with neuro-
developmental impairment of preterm
infants.36
Data on mode of delivery in twin
pregnancies complicated by VLBW
twins is very limited. Recently, a large
prospective randomized trial compared
delivery strategies for twins between 32
and 38 weeks of gestation and reported
on similar neonatal outcomes between
planned cesarean delivery and planned
vaginal delivery. However, the inclusion
criteria for this trial included estimated
fetal weight of 1500-4000 g; therefore,
this study was not designed to address
the issue of mode of delivery of VLBW
twins.23 Our inclusion criteria were
based on birthweight and not estimated
fetal weight. Taking into account the
accuracy of fetal-weight estimations,
some authors suggested to use a 2000-g
cutoff when considering mode of de-
livery.25 However, the twin birth trial
demonstrated the safety of breech de-
livery of the second twin when estimated
fetal weight was 1500 g.23
Morales et al34 examined neonatal
outcome of 156 VLBW twin pairs who
were born from 1981-1986 and did
not nd an advantage for cesarean de-
livery. In accordance with our ndings,
the aforementioned study demonstrated
an increase in IVH grade 3-4 after
vaginal delivery when at least 1 twin
was noncephalic. However, the increase
in risk was not statistically signicant
after correction for confounders. A
population-based study that included all
twin births in the United States from
1995-1997 has shown increased mo-
rbidity and death risk in vaginal de-
liveries of twins compared with cesarean
deliveries13 with a more pronounced
death risk in VLBW infants. However,
the data in this study were derived from
many centers with great variability in the

TABLE 6
Neonatal outcome: a subanalysis restricted to cases with a cephalic presentation of the first twin
Adjusted odds ratio
Cesarean delivery Vaginal delivery Adjusted (95% confidence interval)
Variable (n [ 142), n (%) (n [ 49), n (%) P value P value Model 1 Model 2
Composite outcome 107 (75.4) 35 (71.4) .588 .068 0.387 (0.14e1.07) 0.28 (0.09e0.87)
Death 8 (5.6) 3 (6.1) .899 .534 0.62 (0.13e2.85) 1.14 (0.2e6.47)
Respiratory distress syndrome 98 (69) 32 (65.3) .631 .036 0.33 (0.12e0.93) 0.24 (0.08e0.74)
Intraventricular hemorrhage 12 (8.5) 15 (30.6) < .001 .006 4.42 (1.55e12.63) 6.03 (1.81e20.03)
Intraventricular hemorrhage 3 (2.1) 6 (12.2) .004 .191 3.14 (0.57e17.4) 6.15 (0.57e65.8)
grade 3-4
Logistic regression was performed to correct for possible confounders that included gestational age, antenatal steroids, and chorionicity (model 1) or gestational age, antenatal steroids, chorionicity,
maternal age, birthweight, and parity (model 2).
Barzilay. Mode of delivery in VLBW twins. Am J Obstet Gynecol 2015.

1.e6 American Journal of Obstetrics & Gynecology MONTH 2015

ajog.org
level of experience of twin delivery and
the level of neonatal care.
Another noteworthy outcome of our
study is the high success rate of
vaginal delivery (90.5% for cephalic-
cephalic twins and 96.4% for cephalic-
noncephalic twins). This success rate is
markedly higher than what was previ-
ously reported in the literature15,19 and
is higher than the global success rate for
trial of vaginal delivery in twin preg-
nancies in our center during the study
period (84%; unpublished results). Ho-
wever, these publications referred to
term or near-term twin deliveries, and
we believe that the low birthweight may
contribute to the high success rate that
was observed.
The concerns that this study raises
regarding the risk of IVH in VLBW twins
who undergo vaginal delivery may also
foster concerns regarding vaginal de-
livery in VLBW singleton pregnancies.
The association between mode of de-
livery and the risk of IVH in VLBW
singleton pregnancies has been studied,
with controversial ndings. Although
several small studies suggested that ce-
sarean delivery may decrease the risk
for IVH in VLBW singleton pregnan-
cies,37-40 a study of 4658 VLBW newborn
infants41 and other smaller studies42-44
did not nd an association between
IVH and mode of delivery. Because an
association between IVH and mode of
delivery in singleton pregnancies has not
been established and considering our
small sample size, we believe that, at this
stage, the ndings of our study should be
interpreted with caution.
The inherent disadvantages of a
retrospective cohort study design limited
our ability to control for potential con-
founding factors. The patients in the
vaginal delivery group delivered earlier
compared with patient in the cesarean
delivery group, which could potentially
explain the increased rate of IVH among
vaginally delivered twins.
However, multiple regression analysis
adjustment for gestational age at delivery
and other confounding factors revealed
that vaginal delivery was associated
independently with a higher rate of IVH.
Furthermore, the neonatal outcome in
our study was not inuenced by the
presentation of the second twin because
both cephalic-cephalic and cephalic-
noncephalic twins who were delivered
vaginally had increased rates of IVH.
These ndings are consistent with the
aforementioned randomized trial of
larger twins that showed similar
neonatal outcomes for vaginal delivery
vs cesarean delivery, regardless of the
presentation of the second twin.23
Because 1 of the disadvantages of the
nonrandomized design, the cesarean
group included cases that may have been
more compromised. Cases with sus-
pected fetal distress will be less likely to
undergo a trial of labor. However,
because this may cause a bias towards the
null hypothesis, it does not explain the
increased risk for IVH that was found in
the vaginal delivery group. Moreover,
residual confounding could also occur
as the result of a lack of data regarding
fetal heart-rate monitoring, intra-
amniotic infection, and placental path-
ologic condition.
In view of the obvious lack of suf-
cient evidence relevant to mode of de-
livery in VLBW twin pregnancies, we
believe that this study is of great value
and provides important data that may
assist the medical staff when they are
counseling patients regarding the
preferred mode of delivery in these cases.
In conclusion, in this study, we
demonstrated that vaginal delivery of
VLBW twins may increase risk of IVH,
regardless of presentation. Because of
our ndings regarding the safety of
vaginal delivery of VLBW twins and
because of the limits of the small sample
size and the retrospective cohort design,
large prospective randomized studies
that will address the question of mode of
delivery in VLBW twins are needed
urgently. - 14. Fox NS, Silverstein M, Bender S,
REFERENCES
1. Bergh T, Ericson A, Hillensjo T, Nygren KG,
Wennerholm UB. Deliveries and children born
after in-vitro fertilisation in Sweden 1982-95: a
retrospective cohort study. Lancet 1999;354:
1579-85.
2. Ferraretti AP, Goossens V, Kupka M, et al.
Assisted reproductive technology in Europe,
2009: results generated from European regis-
ters by ESHRE. Hum Reprod 2013;28:2318-31.
MONTH 2015 American Journal of Obstetrics & Gynecology
Obstetrics Research
3. Sunderam S, Kissin DM, Flowers L, et al.
Assisted reproductive technology surveillance:
United States, 2009. MMWR Surveill Summ
2012;61:1-23.
4. Wilcox LS, Kiely JL, Melvin CL, Martin MC.
Assisted reproductive technologies: estimates
of their contribution to multiple births and
newborn hospital days in the United States. Fertil
Steril 1996;65:361-6.
5. Janvier A, Spelke B, Barrington KJ. The
epidemic of multiple gestations and neonatal
intensive care unit use: the cost of irresponsi-
bility. J Pediatr 2011;159:409-13.
6. Bassil KL, Shah PS, Barrington KJ,
Harrison A, da Silva OP, Lee SK. The changing
epidemiology of preterm twins and triplets
admitted to neonatal intensive care units in
Canada, 2003 to 2008. Am J Perinatol 2012;29:
237-44.
7. Wen SW, Fung Kee Fung K, Oppenheimer L,
Demissie K, Yang Q, Walker M. Neonatal
morbidity in second twin according to gesta-
tional age at birth and mode of delivery. Am J
Obstet Gynecol 2004;191:773-7.
8. Wen SW, Fung Kee Fung K, Oppenheimer L,
Demissie K, Yang Q, Walker M. Neonatal mor-
tality in second twin according to cause of death,
gestational age, and mode of delivery. Am J
Obstet Gynecol 2004;191:778-83.
9. Armson BA, OConnell C, Persad V,
Joseph KS, Young DC, Baskett TF. De-
terminants of perinatal mortality and serious
neonatal morbidity in the second twin. Obstet
Gynecol 2006;108:556-64.
10. Smith GC, Shah I, White IR, Pell JP,
Dobbie R. Mode of delivery and the risk of
delivery-related perinatal death among twins at
term: a retrospective cohort study of 8073
births. BJOG 2005;112:1139-44.
11. Smith GC, Pell JP, Dobbie R. Birth order,
gestational age, and risk of delivery related
perinatal death in twins: retrospective cohort
study. BMJ 2002;325:1004.
12. Smith GC, Fleming KM, White IR. Birth order
of twins and risk of perinatal death related to
delivery in England, Northern Ireland, and Wales,
1994-2003: retrospective cohort study. BMJ
2007;334:576.
13. Yang Q, Wen SW, Chen Y, Krewski D, Fung
Kee Fung K, Walker M. Neonatal death and
morbidity in vertex-nonvertex second twins ac-
cording to mode of delivery and birth weight. Am
J Obstet Gynecol 2005;192:840-7.
Klauser CK, Saltzman DH, Rebarber A. Active
second-stage management in twin pregnan-
cies undergoing planned vaginal delivery in a
U.S. population. Obstet Gynecol 2010;115:
229-33.
15. Schmitz T, Carnavalet Cde C, Azria E,
Lopez E, Cabrol D, Gofnet F. Neonatal out-
comes of twin pregnancy according to the
planned mode of delivery. Obstet Gynecol
2008;111:695-703.
16. Grisaru D, Fuchs S, Kupferminc MJ, Har-
Toov J, Niv J, Lessing JB. Outcome of 306 twin
deliveries according to rst twin presentation
1.e7
Research Obstetrics
and method of delivery. Am J Perinatol 2000;17:
303-7.
17. Adam C, Allen AC, Baskett TF. Twin de-
livery: inuence of the presentation and method
of delivery on the second twin. Am J Obstet
Gynecol 1991;165:23-7.
18. Gocke SE, Nageotte MP, Garite T,
Towers CV, Dorcester W. Management of the
nonvertex second twin: primary cesarean sec-
tion, external version, or primary breech extrac-
tion. Am J Obstet Gynecol 1989;161:111-4.
19. Breathnach FM, McAuliffe FM, Geary M,
et al. Prediction of safe and successful vaginal
twin birth. Am J Obstet Gynceol 2011;205:237.
e1-7.
20. Crowther CA. Caesarean delivery for the
second twin. Cochrane Database Syst Rev
2000:CD000047.
21. Hogle KL, Hutton EK, McBrien KA,
Barrett JF, Hannah ME. Cesarean delivery for
twins: a systematic review and meta-analysis.
Am J Obstet Gynecol 2003;188:220-7.
22. Rabinovici J, Barkai G, Reichman B,
Serr DM, Mashiach S. Randomized manage-
ment of the second nonvertex twin: vaginal de-
livery or cesarean section. Am J Obstet Gynecol
1987;156:52-6.
23. Barrett JF, Hannah ME, Hutton EK, et al.
A randomized trial of planned cesarean or
vaginal delivery for twin pregnancy. N Engl J Med
2013;369:1295-305.
24. Udom-Rice I, Skupski DW, Chervenak FA.
Intrapartum management of multiple gestation.
Semin Perinatol 1995;19:424-34.
25. Chervenak FA, Johnson RE, Youcha S,
Hobbins JC, Berkowitz RL. Intrapartum man-
agement of twin gestation. Obstet Gynecol
1985;65:119-24.
26. Duenhoelter JH, Wells CE, Reisch JS,
Santos-Ramos R, Jimenez JM. A paired
1.e8 American Journal of Obstetrics & Gynecology MONTH 2015
controlled study of vaginal and abdominal de-
livery of the low birth weight breech fetus. Obstet
Gynecol 1979;54:310-3.
27. Goldenberg RL, Nelson KG. The premature
breech. Am J Obstet Gynecol 1977;127:240-4.
28. Barrett JM, Staggs SM, Van Hooydonk JE,
Growdon JH, Killam AP, Boehm FH. The effect
of type of delivery upon neonatal outcome in
premature twins. Am J Obstet Gynecol
1982;143:360-7.
29. Mwansa-Kambafwile J, Cousens S,
Hansen T, Lawn JE. Antenatal steroids in pre-
term labour for the prevention of neonatal deaths
due to complications of preterm birth. Int J
Epidemiology 2010;39(suppl 1):i122-33.
30. Vergani P, Russo FM, Follesa I, et al. Peri-
natal complications in twin pregnancies after 34
weeks: effects of gestational age at delivery and
chorionicity. Am J Perinatol 2013;30:545-50.
31. Sainani K. The limitations of statistical
adjustment. PM R 2011;3:868-72.
32. Jain NJ, Kruse LK, Demissie K,
Khandelwal M. Impact of mode of delivery on
neonatal complications: trends between 1997
and 2005. J Matern Fetal Neonatal Med
2009;22:491-500.
33. Levine EM, Ghai V, Barton JJ, Strom CM.
Mode of delivery and risk of respiratory dis-
eases in newborns. Obstet Gynecol 2001;97:
439-42.
34. Morales WJ, OBrien WF, Knuppel RA,
Gaylord S, Hayes P. The effect of mode of de-
livery on the risk of intraventricular hemorrhage in
nondiscordant twin gestations under 1500 g.
Obstet Gynecol 1989;73:107-10.
35. Morita T, Morimoto M, Yamada K, et al.
Low-grade intraventricular hemorrhage disrupts
cerebellar white matter in preterm infants: evi-
dence from diffusion tensor imaging. Neurora-
diology 2015 [Epub ahead of print].
ajog.org
36. Klebermass-Schrehof K, Czaba C,
Olischar M, et al. Impact of low-grade intraven-
tricular hemorrhage on long-term neuro-
developmental outcome in preterm infants.
Childs Nerv Syst 2012;28:2085-92.
37. Minguez-Milio JA, Alcazar JL, Auba M, Ruiz-
Zambrana A, Minguez J. Perinatal outcome and
long-term follow-up of extremely low birth
weight infants depending on the mode of de-
livery. J Matern Fetal Neonatal Med 2011;24:
1235-8.
38. Ment LR, Oh W, Philip AG, et al. Risk factors
for early intraventricular hemorrhage in low birth
weight infants. J Pediatr 1992;121:776-83.
39. Ment LR, Oh W, Ehrenkranz RA, Philip AG,
Duncan CC, Makuch RW. Antenatal steroids,
delivery mode, and intraventricular hemorrhage
in preterm infants. Am J Obstet Gynecol
1995;172:795-800.
40. Philip AG, Allan WC. Does cesarean section
protect against intraventricular hemorrhage in
preterm infants? J Perinatol 1991;11:3-9.
41. Riskin A, Riskin-Mashiah S, Bader D, et al.
Delivery mode and severe intraventricular hem-
orrhage in single, very low birth weight, vertex
infants. Obstet Gynecol 2008;112:21-8.
42. Paul DA, Sciscione A, Leef KH, Stefano JL.
Caesarean delivery and outcome in very low
birthweight infants. Aust N Z J Obstet Gynaecol
2002;42:41-5.
43. Malloy MH, Onstad L, Wright E. The effect of
cesarean delivery on birth outcome in very low
birth weight infants: National Institute of Child
Health and Human Development Neonatal
Research Network. Obstet Gynecol 1991;77:
498-503.
44. Low JA, Galbraith RS, Sauerbrei EE, et al.
Maternal, fetal, and newborn complications
associated with newborn intracranial hemor-
rhage. Am J Obstet Gynecol 1986;154:345-51.