Routine Cervical Length in Twins

and Perinatal Outcomes

Cynthia Gyamfi, M.D.,1 Veronica Lerner, M.D.,2 Ian Holzman, M.D.,3
and Joanne L. Stone, M.D.2

ABSTRACT

A retrospective review of twin gestations was undertaken to evaluate whether

routine cervical lengths (CLs) in such instances change pregnancy outcome. Data were
collected from the ultrasound database and chart review. Exclusion criteria included twins
reduced to singletons, twins not delivering at our institution, and incomplete information.
Twin gestations with a CL were compared with those without a CL. Outcomes of interest
included gestational age (GA) at delivery, preterm delivery (PTD), antepartum admissions,
antepartum length of stay (LOS), cerclage placement, birthweight, neonatal intensive care
unit admissions, and neonatal LOS. Two hundred sixty-two patients met inclusion criteria.
Of those, 184 had CLs and 78 did not. Comparing the CL to the no-CL group, there were
no differences with respect to GA at delivery (34.8 versus 35.3 weeks; p ¼ 0.35),
antepartum admissions (32.1 versus 23.1%; p ¼ 0.16), cerclage placement (7.1 versus
1.3%; p ¼ 0.06), or tocolysis use (28.6 versus 21.8%; p ¼ 0.26). There was no difference
between the two groups with respect to preterm labor (26 versus 19%; p ¼ 0.25), PTD < 28
weeks (8.2 versus 3.9%; p ¼ 0.21), PTD < 34 weeks (26.1 versus 25.6%; p ¼ 0.94), or PTD
< 37 weeks (76.1 versus 70.5%; p ¼ 0.34). The only significant difference was antepartum
LOS (34.5 versus 31.3 days; p < 0.001). There were no differences in neonatal outcomes.
Routine CL did not improve perinatal outcome but increased maternal antepartum LOS.

KEYWORDS: Twins, ultrasound, cervical length

T win gestations are a leading cause of perinatal
morbidity and mortality, primarily due to the high
incidence of preterm delivery (PTD). PTD occurs in
45% of twin pregnancies, with 22.4% delivering before
34 weeks.1 In the last decade, cervical length by trans-
vaginal ultrasound has become a modality for prediction
of PTD. Iams et al2 showed clearly that the risk for PTD
increased as the cervical length (CL) decreased after
reviewing CLs in more than 2500 singleton gestations.
This work has now been validated in twin gestations as
well. Guzman et al3 showed that a CL < 20 mm
between 15 and 20 weeks positively predicts delivery
prior to 34 weeks in twin gestations.
The ability to identify those patients at risk for
preterm labor before viability is important; however, an
appropriate intervention to prevent this delivery has not
yet been established. Cervical shortening leading to
PTD can manifest itself as preterm labor (defined as
preterm contractions causing cervical change), preterm
premature rupture of the membranes, or painless dilata-
tion suggestive of cervical incompetence. It is not clear
which of these forms of PTD will ensue once a shortened

1
Division of Maternal-Fetal Medicine, Department of Obstetrics and
Gynecology, Columbia University Medical Center; 2Division of
Maternal-Fetal Medicine, Department of Obstetrics, Gynecology,
and Reproductive Sciences, Mount Sinai School of Medicine; 3Divi-
sion of Neonatology, Department of Pediatrics, Mount Sinai School of
Medicine, New York, New York.
Address for correspondence and reprint requests: Cynthia Gyamfi,
M.D., Division of Maternal-Fetal Medicine, Department of Obstetrics
and Gynecology, Columbia University Medical Center, 622 West
168th Street, PH-16, New York, NY 10032.
Am J Perinatol 2007;24:65–70. Copyright # 2007 by Thieme
Medical Publishers, Inc., 333 Seventh Avenue, New York, NY 10001,
USA. Tel: +1(212) 584-4662.
Accepted: September 17, 2006. Published online: December 27,
2006.
DOI 10.1055/s-2006-958167. ISSN 0735-1631.
65
66 AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 24, NUMBER 1 2007

cervix is identified. A recent meta-analysis showed that characteristics are listed in Table 1. The patients who
when a cerclage was placed for cervical shortening in had CLs were more likely to be older and have private
twin gestations, there was an increase in PTD rates.4 insurance. They also presented significantly earlier for
This leads to the question of whether routine measure- their first ultrasound than those who did not have CLs
ment of CLs in twins change pregnancy outcome. (14.1 versus 20.8 wks; p < 0.001, respectively). Impor-
tantly, they were similar in having a history of previous
PTD and a history of two or more spontaneous abortions
STUDY DESIGN (Table 1). The average number of CLs performed was
A retrospective review of all twin gestations evaluated by 3.9; the median was 4.
ultrasound at the Mount Sinai Medical Center from When comparing the CL versus the no-CL
January 2002 to June 2004 was undertaken. Patients group, we analyzed variables that would reflect perinatal
were initially identified as having twin gestations using morbidity (Table 2). We found no differences with
International Classification of Diseases (9th revision) respect to gestational age at delivery (34.8 versus 35.3
codes. Patient data were then collected from the ultra- weeks; p ¼ 0.35); preterm labor (26 versus 19%; p ¼ 0.25);
sound database and correlated with information in or PTD at < 28 weeks (8.2 versus 3.9%; p ¼ 0.21), < 34
patient charts. Delivery data were collected from patient weeks (26.1 versus 25.6%; p ¼ 0.94), or < 37 weeks (76.1
charts, the labor and delivery log books, the labor floor versus 70.5%; p ¼ 0.34). The incidence of spontaneous
computer database, and the neonatal database. All pa- abortion at < 24 weeks in the current pregnancy was
tients with twin gestations and at least one ultrasound at similar in both groups (4.9 versus 2.6%, respectively;
our institution were included. Exclusion criteria included p ¼ 0.39). We created a composite variable, cervical risk
twins reduced spontaneously or electively to singletons, factors, which included factors related to potential cer-
twins not delivering at our institution, and those for vical manipulation. These factors included a history of a
which outcome information was incomplete. Loop Electrosurgical Excision Procedure (LEEP) or
Our main variable of interest was whether a CL cold-knife cone, history of first-trimester bleeding, pre-
had been performed in the current pregnancy. CLs were vious septoplasty, and a history of cervical ablation. The
all performed by a core group of providers in the cervical risk factors were similar in the CL and no-CL
Division of Maternal-Fetal Medicine Ultrasound Unit group (12.5 versus 11.5%, respectively; p ¼ 0.83).
at Mount Sinai Medical Center. All CLs were per- We then looked at maternal outcomes. Women
formed by the transvaginal approach as previously de- with a CL had equal amounts of antepartum admis-
scribed by Iams et al.2 The shortest CL was recorded. In sions (32.1 versus 23.1%; p ¼ 0.16), cerclage placement
2004, the division of Maternal-Fetal Medicine at Mount (7.1 versus 1.3%; p ¼ 0.06), or tocolysis use (28.6 versus
Sinai instituted a policy of routine CLs on all twin 21.8%; p ¼ 0.26). The only significant difference was
gestations. Prior to 2004, the decision to perform a CL for antepartum LOS. The CL group was admitted for
was made by the provider. During the study period, CL an average of 34.5 days compared with 31.3 days in the
was not used routinely in the paradigm for management
of preterm labor, and CLs were rarely performed after 24
weeks. Therefore, none of the CLs were performed on
Table 1 Patient Characteristics
the labor and delivery ward. Demographic information
Patient
for patients receiving CLs was compared with that of
Characteristics CL No CL
patients who did not have this measurement. Twin (n ¼ 262) (n ¼ 184) (n ¼ 78) p
gestations with a CL were compared with those without
one. Outcomes of interest included gestational age at Age (yr) 34.2 31.3 < 0.001
delivery, PTD, antepartum admissions, antepartum Private insurance (%) 87.2 71.8 0.003
length of stay (LOS), cerclage placement, birthweight, Race (%) 0.009
NICU admissions, and neonatal LOS. White 79.7 61.5
Statistical analysis was performed using x2 for the Black 5.9 16.7
dichotomous variables. Continuous variables were ana- Hispanic 9.1 17.9
lyzed with the Student’s t-test. In addition, a multi- Asian 4.3 2.6
variate logistic regression analysis was used to adjust for Other 1.1 1.3
potential confounders. The study was approved by the No. of pregnancies 2.5 2.5 0.932
Mount Sinai Institutional Review Board. History of preterm 4.3 3.8 0.873
delivery (%)
History of
2 spontaneous 11.4 7.7 0.367
RESULTS abortions (%)
We identified 262 patients who met inclusion criteria. GA at first ultrasound (wk) 14.1 20.8 < 0.001
Of those, 184 had CLs and 78 did not. The patient CL, cervical length; GA, gestational age.
 ROUTINE CERVICAL LENGTH IN TWINS AND PERINATAL OUTCOMES/GYAMFI ET AL 67

Table 2 Results were diagnosed with preterm labor requiring active

CL No CL intervention.
Variable (n ¼ 184) (n ¼ 78) p

Spontaneous abortion 4.9 2.7 0.39

DISCUSSION
< 24 wk (%)
In summary, we found that the performance of routine
PTD < 28 wk (%) 8.2 3.9 0.21
CLs in twin gestations did not change maternal or
PTD < 34 wk (%) 26.1 25.6 0.94
neonatal outcomes. We know that a shortened cervix is
PTD < 37 wk (%) 76.1 70.5 0.34
predictive of PTD. Iams et al2 were one of the first
Gestational age at 34.8 35.3 0.35
groups to show that CL correlated with PTD. Multiple
delivery (wk)
gestations, however, were excluded from this study.
Antepartum 32.1 23.1 0.16
Subsequently, Guzman et al5,6 and others have validated
admissions (%)
these findings in twin and triplet gestations.7,8 The
Antepartum LOS (d) 34.5 31.3 < 0.001
corollary was also found to be true. Imseis et al9 found
Chorionicity (%) 0.75
that a CL of > 35 mm at 24 to 26 weeks could to identify
Dichorionic/diamniotic 84.1 82.3
those patients with twin gestations who were at a low
Monochorionic/diamniotic 15.4 18.7
risk for delivery prior to 34 weeks.
Neonatal composite 31.9 26.1 0.20
However, does the detection of cervical short-
(RDS, NEC, IVH) (%)
ening change perinatal outcome in the absence of inter-
Neonatal LOS (d) 10.9 11.9 0.60
vention? Prior to 24 weeks, management of cervical
NICU admissions (%) 45.3 59.1 0.27
shortening has varied. Some practitioners place a cerc-
Birthweight (g) 2253 2288 0.61
lage, some will encourage bedrest, and others will mon-
CL, cervical length; PTD, preterm delivery; LOS, length of stay; RDS, itor the patient for contractions and proceed based on
respiratory distress syndrome; NEC, necrotizing enterocolitis; IVH,
intraventricular hemorrhage; NICU, neonatal intensive care unit. the findings. There are a few studies looking at cerclage
placement in twin gestations with shortened cervi-
no-CL group (p < 0.001). The antepartum LOS did ces.10,11 This intervention has not been shown to de-
not correlate with the number of CLs (r ¼ 0.022; crease PTD in twins. In fact, the first randomized trial
p ¼ 0.909). Neonatal outcomes, such as NICU admis- looking at cerclage as an intervention for PTD was
sions, neonatal LOS, and birthweight, were similar in performed in twin gestations by Dor et al10 in 1982.
both groups (Table 2). They randomly assigned 50 twin pregnancies conceived
Finally, we performed a logistic regression after ovulation induction to elective cerclage versus no
model to analyze which factors significantly affected cerclage. They found no difference in the rates of PTD
the rate of PTD at < 34 weeks (Table 3). We in either group. Newman et al11 also evaluated twins and
controlled for potential confounders that included cerclage placement. They performed a prospective, non-
age, history of preterm birth, the composite of cervical randomized trial offering cervical cerclage to women
risk factors, a composite for obstetrical complications, with twin gestations who had a CL of  25 mm. Their
antepartum LOS, use of tocolytics, and presence of a group found no difference in the rate of PTD in either
cerclage. We found that antepartum admission was an group.
independent risk factor for preterm birth < 34 weeks Other trials evaluating cerclage and cervical short-
(odds ratio [OR], 1.06; 95% confidence interval [CI], ening included both singletons and twins.12,13 A meta-
1.00 to 1.12; p ¼ 0.04). Tocolysis was also associated analysis looking at these trials found that placement of a
with preterm birth, likely because those patients cerclage doubles the risk of PTD in twins at < 35 weeks
(OR, 2.15; 95% CI, 1.15 to 4.01).4 Although this
finding is only based on 49 patients, a significant differ-
Table 3 Multivariate Logistic Regression Model of
ence was noted.
Factors Affecting the Rate of PTD < 34 wk
What other interventions are available prior to 24
Factor OR 95% CI p
weeks? Bedrest has not been shown to prevent PTD in
CL versus no CL 0.82 0.38–1.27 0.60 twins. Crowther et al14 randomly assigned patients with
Age 1.07 0.96–1.07 0.58 twin gestations to hospitalization for bedrest versus
History of PTD 2.43 0.57–10.4 0.23 outpatient management. There was no difference in
Cervical risks (composite) 1.19 0.43–3.42 0.75 the incidence of PTD or in neonatal morbidity. How-
Cerclage 1.78 0.43–0.78 0.42 ever, they did find a difference in birthweights. The
Tocolysis 7.91 3.65–17.1 < 0.001 hospitalized group had larger infants (mean ¼ 2430 g),
AP admissions 1.06 1.00–1.12 0.04 compared with the outpatient group (mean ¼ 2300 g;
OR, odds ratio; CI, confidence interval; CL, cervical length; PTD,
p ¼ 0.02). These findings have been confirmed in other
preterm delivery; AP, antepartum. trials. Younis et al15 also found that birthweight was
68 AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 24, NUMBER 1
increased in a group of women with twins hospitalized
for bedrest when compared with two control groups: one
on bedrest at home and the other without restrictions.
However, there were no differences in gestational dura-
tion or rate of prematurity, so they concluded that the
cost of hospitalization may not justify the small differ-
ence in birthweight.
Monitoring a patient for contractions prior to 24
weeks can lead to a slippery slope of interventions. If
someone is contracting prior to 24 weeks, is that woman
in labor? If so, should she be tocolyzed? Tocolysis, at
best, has been shown to delay PTD by 48 hours, but only
by retrospective data.16 Conceivably, the outcome for a
fetus delivered at 20 2/7 weeks is not different from that
of one delivered at 20 4/7 weeks. Tocolysis, however, in
the form of nonsteroidal anti-inflammatories, has been
used successfully for women with preterm contractions
due to large myomatous uteri.17 Broadening this indi-
cation to include all women contracting prior to 24
weeks may be futile.
An intervention for PTD after 24 weeks in twin
gestations has not been established. Progesterone, when
given either as a vaginal suppository or by weekly intra-
muscular injection, has been shown to decrease preterm
birth in a high-risk population.18,19 Both of these studies
excluded multiple gestations. There currently is a pro-
spective trial underway to evaluate this intervention in
twins and higher order multiples.
Finally, a discussion on intervention for PTD
would not be complete without the mention of toco-
lytics. A Cochrane Database Review showed that mag-
nesium sulfate, the most popular drug used in the
United States to prevent PTD, did not improve neo-
natal outcomes when compared with placebo.20 They
recommended that the drug should not be used due to
the significant side effect profile, without a documented
benefit, for preventing preterm labor. Nifedipine, a
calcium channel blocker, appeared to be a more prom-
ising drug when reviewed by the Cochrane Database.21
However, the studies showing a benefit of nifedipine
for preterm birth did not compare the drug versus
placebo.
Without an intervention to prevent preterm
birth, routine CL in twin gestations should not be
expected to change perinatal outcome. That was the
finding of this retrospective review. Without improving
perinatal outcomes, we increased maternal costs indi-
rectly by increasing the overall antepartum LOS. The
majority of the patients admitted in the no-CL group
were admitted for preterm labor. Therefore, patients in
either group received similar amounts of antenatal
corticosteroids to promote fetal lung maturity, and
the neonatal outcomes were not different. Our numbers
were small and our patients were not randomly as-
signed. This will limit the generalization of our find-
ings. In addition, prior to 2004, CLs were not routine
2007
in our institution. Therefore, there is a selection bias
prior to this year as to which patients had this particular
scan. In addition, during the study period, our institu-
tion did not use CL in the paradigm for evaluation and
management of preterm labor. This further limits
generalization because there are several institutions
for which this is not the case. However, our patients
did have similar risk factors in that the history of
preterm birth and history of more than two sponta-
neous abortions were similar. By detecting a shortened
cervix prior to 24 weeks, we did not see a difference in
spontaneous abortion at < 24 weeks or PTD after that
time. We also believe that until an effective interven-
tion is found for PTD in twins, a prospective, random-
ized, controlled study to answer the question of routine
CLs in twins would not be helpful.
REFERENCES
1. Vital and Health Statistics. Health and Demographic
Characteristics of Twin Birth: United States, 1988. Series
21: Data on Natality, Marriage, and Divorce, No. 50.
Hyattsville, MD: U.S. Department of Health and Human
Services; 1992
2. Iams JD, Goldenberg RL, Meis PJ, et al. The length of the
cervix and the risk of spontaneous premature delivery. N Engl
J Med 1996;334:567–572
3. Guzman ER, Walters C, O’Reilly-Green C, et al. Use of
cervical ultrasonography in prediction of spontaneous pre-
term birth in twin gestations. Am J Obstet Gynecol 2000;
183:1103–1107
4. Berghella V, Odibo AO, To MS, Rust OA, Althuisius SM.
Cerclage for short cervix on ultrasonography: meta-analysis
of trials using individual patient-level data. Obstet Gynecol
2005;106:181–189
5. Guzman ER, Walters C, O’Reilly-Green C, et al. Use of
cervical ultrasonography in prediction of spontaneous pre-
term birth in twin gestations. Am J Obstet Gynecol 2000;
183:1103–1107
6. Guzman ER, Walters C, O’Reilly-Green C, et al. Use of
cervical ultrasonography in prediction of spontaneous pre-
term birth in triplet gestations. Am J Obstet Gynecol
2000;183:1108–1113
7. Vayssiere C, Favre R, Audibert F, et al. Cervical length and
funneling at 22 and 27 weeks to predict spontaneous birth
before 32 weeks in twin pregnancies: a French prospective
multicenter study. Am J Obstet Gynecol 2002;187:1596–
1604
8. Shulman A, Jauniaux E, Holmes A, Maschiach S, Maymon
R. Vaginal sonography of the cervix for the prediction of
‘‘time to delivery’’ in ART twins gestations. Twin Res 2002;
5:255–259
9. Imseis HM, Albert TA, Iams JD. Identifying twin gestations
at low risk for preterm birth with a transvaginal ultrasono-
graphic cervical measurement at 24 to 26 weeks’ gestation.
Am J Obstet Gynecol 1997;177:1149–1155
10. Dor J, Shalev J, Mashiach S, Blankstein J, Serr DM.
Elective cervical suture of twin pregnancies diagnosed
ultrasonically in the first trimester following induced
ovulation. Gynecol Obstet Invest 1982;13:55–60
 ROUTINE CERVICAL LENGTH IN TWINS AND PERINATAL OUTCOMES/GYAMFI ET AL
11. Newman RB, Krombach RS, Meyers MC, McGee DL.
Effect of cerclage on obstetrical outcomes in twin gestations
with a shortened cervical length. Am J Obstet Gynecol
2002;186:634–640
12. Rust OA, Altas RO, Jones KJ, Benham BN, Balducci J. A
randomized trial of cerclage versus no cerclage among
patients with ultrasonographically detected second-trimester
preterm dilatation of the internal os. Am J Obstet Gynecol
2000;183:830–835
13. Berghella V, Odibo AO, Tolosa JE. Cerclage for prevention
of preterm birth in women with a short cervix found on
transvaginal ultrasound examination: a randomized trial.
Am J Obstet Gynecol 2004;191:1311–1317
14. Crowther CA, Verkuyl DAA, Neilson JP, Bannerman C,
Ashurst HM. The effects of hospitalization for rest on fetal
growth, neonatal morbidity and length of gestation in twin
pregnancy. Br J Obstet Gynaecol 1990;97:872– 877
15. Younis JS, Sadovsky E, Eldar-Geva T, Mildwidsky A, Zeevi
D, Zajicek G. Twin gestations and prophylactic hospital-
ization. Int J Gynaecol Obstet 1990;32:325–330
16. Elliott JP. Magnesium sulfate as a tocolytics agent. Am J
Obstet Gynecol 1983;147:277–284
69
17. Dildy GA III, Moise KJ Jr, Smith LG Jr, Kirshon B,
Carpenter RJ Jr. Indomethacin for the treatment of
symptomatic leiomyoma uteri during pregnancy. Am J
Perinatol 1992;9:185–189
18. Meis PJ, Klebanoff M, Thom E, et al. National Institute of
Child Health and Human Development Maternal-Fetal
Medicine Units Network. Prevention of recurrent preterm
delivery by 17 alpha-hydroxyprogesterone caproate. N Engl J
Med 2003;348:2379–2385
19. da Fonseca EB, Bittar RE, Carvalho MH, Zugaib M.
Prophylactic administration of progesterone by vaginal
suppository to reduce the incidence of spontaneous preterm
birth in women at increased risk: a randomized placebo-
controlled double-blind study. Am J Obstet Gynecol 2003;
188:419–424
20. Crowther CA, Hiller JE, Doyle LW. Magnesium sulphate
for preventing preterm birth in threatened preterm
labour. Cochrane Database Syst Rev 2002;(4):CD001060.
Review
21. King JF, Flenady VJ, Papatsonis DN, Dekker GA, Carbonne
B. Calcium channel blockers for inhibiting preterm labour.
Cochrane Database Syst Rev 2003;(1):CD002255. Review