Journal of Perinatology (2007) 27, 749–753

r 2007 Nature Publishing Group All rights reserved. 0743-8346/07 $30

www.nature.com/jp

ORIGINAL ARTICLE
Cervical length and the risk of spontaneous labor at term
LL Tolaymat1, VH Gonzalez-Quintero2, L Sanchez-Ramos1, A Kaunitz1, P Wludyka3, MJ O’Sullivan2 and
D Martin2
1
Department of Obstetrics and Gynecology, University of Florida College of Medicine-Jacksonville, Jacksonville, FL, USA; 2Department
of Obstetrics and Gynecology, University of Miami, Miami, FL, USA; and 3Department of Mathematics and Statistics, University
of North Florida, Jacksonville, FL, USA

measurement to the outcome of pregnancy has been assessed in

Objective: To investigate the relationship between cervical length (CL) at
37 to 40 weeks and delivery within 7 days and delivery by 41 weeks.
Study design: We performed transvaginal ultrasound to measure CL in
women with singleton gestations at 37 to 40 weeks. We then used a
receiver operating characteristic curve (ROC) to assess the relationship
between CL and delivery within 7 days and delivery by 41 weeks.
Result: For the 120 women included in the analysis, the mean CL
(±s.d.) was 25.3±9.8 mm. The logistic regression model to predict each
of the outcomes includes gestational age at ultrasound (GA-US) and CL.
Neither birthweight, nor parity seems to affect the probability of delivery
within 7 days. The ROC curve was used to assess the probability of
spontaneous labor within 7 days at each CL measurement. The likelihood
ratio of delivery within 7 days when CL is p10 mm is 12.
Conclusion: CL measurement at 37–40 weeks is an independent
predictor of delivery within 7 days and delivery by 41 weeks regardless of
GA-US. This information can be utilized when counseling patients
regarding the management of term pregnancies.
Journal of Perinatology (2007) 27, 749–753; doi:10.1038/sj.jp.7211819;
published online 30 August 2007
Keywords: cesarean; vaginal delivery; cervical length; ultrasound
Introduction
Over the last few decades, cervical assessment has moved from
digital examination to sonographic evaluation. Although
transabdominal ultrasound was initially used for evaluation of the
cervix in the second and third trimester, appropriately performed
transvaginal ultrasound produces better and clearer images of the
internal and the external os. The relevance of cervical length (CL)
Correspondence: Dr L Tolaymat, Department of Obstetrics and Gynecology, University of
Florida College of Medicine at Jacksonville, 653-1 West 8th Street, Jacksonville, FL 32209,
USA.
E-mail: lama.Tolaymat@jax.ufl.edu
Received 29 September 2006; revised 3 July 2007; accepted 24 July 2007; published online 30
August 2007
women with suspected cervical incompetence,1 symptoms of
preterm labor,2 in asymptomatic women to evaluate the risk of
preterm delivery,3 and in women before induction of labor to
predict induction outcome.4–7 Recently, studies have examined CL
at term and attempted to correlate it with spontaneous labor.
To date, four English language reports (all European) have assessed
the relationship between CL at term (X37 weeks’ gestation) and
pregnancy outcomes. Three of these studies enrolled women at one
specific week of gestation: two at 37 weeks8,9 and one at 39–40
weeks.10 Only one study enrolled women (N ¼ 93) between 37 and
40 weeks’ gestation.11
To assess the value of CL measurement after 37 weeks as a
predictor of interval to spontaneous labor, we measured CL in
low-risk pregnant women at 37 to 40 weeks and examined the
relationship between CL with two outcomes: spontaneous labor
within 7 days and delivery before 41 weeks.
Methods
This study was approved by the University of Miami Institutional
Review Board and was funded by the American College of
Obstetricians and Gynecologists, District IV 2001 Special Award.
Enrollment was conducted from 1st May 2001 to 31st January 2002.
Candidates received prenatal care and planned to deliver
at Jackson Memorial Hospital. The primary outcome was delivery
within 7 days after spontaneous labor and the secondary outcome was
delivery by 41 weeks’ gestation. Spontaneous labor was defined as
regular uterine contractions with increasing intensity and frequency
associated with cervical change leading to admission for delivery.
Inclusion criteria included: a singleton pregnancy at 37 to 40
weeks’ gestation, a cephalic presentation, intact membranes and no
evidence of contractions at enrollment. Low-risk women who met
these inclusion criteria were recruited after providing informed
consent. Participants underwent a personal interview to obtain
demographic variables and a transvaginal ultrasound within one
week of enrollment to evaluate the cervix. Gestational age was
 CL and the risk of spontaneous labor at term
LL Tolaymat et al
750
based on the date of the onset of the last menses when the age was
within one week of a first-trimester ultrasound, or within two weeks
of a second-trimester ultrasound. Otherwise, gestational age was
based on ultrasound alone.
Exclusion criteria included: placenta previa, multiple gestation,
cerclage, preterm labor requiring tocolysis, ruptured membranes,
planned cesarean, labor or a major fetal anomaly. Transvaginal
ultrasound results were not available to the managing obstetrician
unless a fetal death, oligohydramnios or an abnormal presentation
was noted. The obstetricians managed the rest of the pregnancy
without the involvement of the study team.
The length of the cervix was measured using transvaginal
real-time ultrasonography with an empty bladder. Sonograms
were obtained with an HDI 5000 system (Philips Medical
Systems, Bothell, WA, USA) curved array transvaginal probe
(6 MHz). The transducer was placed in the anterior fornix of the
vagina. Three anatomic landmarks defined the appropriate
sagittal view: the internal os, the external os and the endocervical
canal. The image was enlarged while visualizing the three
landmarks simultaneously. Gentle pressure exerted on the
cervix by the transducer was reduced followed by minimal pressure
to allow visualization of the three landmarks. This procedure
was repeated three times and the shortest measurement was
recorded.
Registered Diagnostic Medical Sonographers (RDMS ) who
performed the study examinations first underwent specific didactic
training regarding cervical assessment to facilitate standardization;
then, they submitted five transvaginal cervical sonographic
examinations performed on patients who were not
studyparticipants. Once the investigators determined that each
sonographer was assessing Cl appropriately, sonographers
performed study examinations. A single investigator, blinded to
pregnancy outcome, reviewed all sonographic images to ensure
proper assessment of CL.
We used the Student’s t-test to compare the means of two groups
(parous and nulliparous) and the w2-test to analyze categorical
variables. Two-sided P-values are reported throughout. To obtain
the optimum cutoff point for CL for predicting spontaneous labor
within 7 days and delivery by 41 weeks, we constructed receiver
operating characteristic curves (ROC). The probability of
spontaneous labor within 7 days and the probability of delivery by
41 weeks were then assessed by markers of diagnostic accuracy and
95% confidence intervals (CI). The 95% CI for sensitivity and
specificity were constructed using the asymptotic standard error
(normal approximation). We then used logistic regression to create
a model including variables such as CL measurement, parity,
maternal age, birthweight and gestational age at ultrasound.
SAS version 9.1 (SAS Institute Inc. Cary, NC, USA) was used for the
statistical analysis. In addition, Analyze-It (Microsoft Excel for
Windows), Analyse-it Software Ltd. Leeds, England was used to
construct the ROC curves.
Journal of Perinatology
Results
During the study period, 215 subjects were recruited. Eleven were at
a gestational age outside the originally specified range (8 less than
37 and 3 at 41 weeks), 19 did not return for their transvaginal
ultrasound, 7 were lost to follow-up, 9 had malpresentation
diagnosed by the transvaginal ultrasound after recruitment and
5 subsequently declined VBAC. Forty-four subjects were induced: 5 for
postdates, 17 for oligohydramnios, 6 for pregnancy-induced
hypertension and 16 for spontaneous rupture of membranes before
the onset of labor, leaving 120 evaluable participants for analysis.
The racial distribution of these women was 75% white and 25%
black. The ethnic distribution was 75% Hispanic, 10% African-
American, 9% Haitian, 4% Jamaican and 2% other. Nulliparous
women comprised 55% of the study participants.
The overall prevalence of spontaneous labor within 7 days in
this cohort was 32.8 and 73.1% delivered by 41 weeks. CL was
normally distributed with a mean (±s.d.) of 25.3±9.8 mm. The
CL was slightly shorter in nulliparous compared to parous women
(23.8±10.0 vs 27.1±9.4 mm; P ¼ 0.06).
The logistic regression model to predict delivery within 7 days
included gestational age at ultrasound and shortest CL. Neither
birthweight, nor parity appeared to affect the probability of
spontaneous labor within 7 days (Table 1). We then applied the
logistic model subcategorizing our population by parity, gestational
age and CL by increments of 5 mm. Table 2 details the sensitivity,
specificity, positive predictive value (PPV) and negative predictive
value (NPV) for each subcategory. We constructed the ROC curve to
estimate the probability of spontaneous labor within 7 days at each
CL measurement (Figure 1). The Wilcoxon estimation of the area
under the curve was 0.767 (95% CI 0.641–0.893, P-value
<0.0001). Using different cutoff points for CL, sensitivity,
specificity, PPV and NPV are outlined in Table 2. When CL is
10 mm (or less) at 37–40 weeks, the likelihood of delivery within
7 days is 12. The likelihood ratio of delivery by 7 days for each
CL cutoff is outlined in Table 2.
The overall prevalence of delivery by 41 weeks in our cohort was
73.1%. The logistic regression model to predict delivery by 41 weeks
included gestational age at ultrasound and shortest CL. Neither
birthweight nor parity seemed to affect the probability of delivery
by 41 weeks (Table 3). Applying the logistic model then
Table 1 Predictors of spontaneous labor within 7 days using logistic regression
model
Estimate s.e. w2 P-value
GA at US 1.4709 0.3092 20.9351 <0.0001
Birthweight 0.00121 0.000659 3.4029 0.0651
Cervical length 0.1087 0.0287 14.3813 <0.0001
Nulliparity 0.0478 0.2625 0.0332 0.8555
Abbreviations: GA, gestational age; US, ultrasound.
 CL and the risk of spontaneous labor at term
LL Tolaymat et al
751

Table 2 Predictive values for cervical length intervals for spontaneous labor within 7 days for each subcategory of the cohort

Strata CL (mm) LR Sensitivity (%) Specificity (%) PPV (%) NPV (%)

All cohort (n ¼ 120) 30 1.4 87.5 38.8 41.7 86.1

25 2.0 77.5 62.5 50.8 84.7
22 2.8 65.0 77.5 59.1 81.6
20 3.0 60.0 80.0 60.0 80.0
15 5.0 37.5 92.5 71.4 74.7
10 12.0 15.0 98.8 85.7 69.9

Multiparous (n ¼ 54) 30 1.5 82.4 45.9 41.2 85.0

25 2.6 76.5 70.3 54.2 86.7
22 3.1 58.8 81.1 58.8 81.1
20 3.3 52.9 83.8 60.0 79.5
15 3.6 29.4 91.9 62.5 73.9

Nulliparous (n ¼ 66) 30 1.4 91.3 32.6 42.0 87.5

25 1.8 78.3 55.8 48.6 82.8
22 2.7 69.6 74.4 59.3 82.1
20 2.8 65.2 76.7 60.0 80.5
15 6.2 43.5 93.0 76.9 75.5
10 11.2 26.1 97.7 85.7 71.2
Abbreviations: CL, cervical length; GA, gestational age; US, ultrasound.

1 Wilcoxon estimation of the area under the curve was 0.632

0.9 (95%; CI 0.518–0.746, P-value <0.0118).
0.8
Sensitivity (true positives)

0.7 Discussion
0.6 This study found that in singleton pregnancies, CL at 37–40 weeks
is an independent predictor of the outcomes spontaneous labor
0.5
within 7 days and delivery by 41 weeks regardless of the specific
0.4 gestational age at the time of ultrasound or parity. The overall
0.3 prevalence of spontaneous labor within 7 days in our cohort was
low (32.8%), likely reflecting the exclusion of women who
0.2
No discrimination complained of contractions at the time of enrollment.
0.1 prob Four previously published studies assessed CL at term
0 (X37 weeks’ gestation), attempting to evaluate the relationship
0 0.2 0.4 0.6 0.8 1 between sonographic measurement of CL and time to delivery.8–11
1-Specificity (false positives) Vimercati et al.8 in 2001 measured CL in 120 nulliparous women
Figure 1 Likelihood of spontaneous labor within 7 days. Receiver operating starting at 37 weeks and then weekly. They found that CL at 39 and
characteristic curve for cervical length (CL) to predict spontaneous labor within at 40 weeks is shorter in women who delivered prior to 41 weeks.
7 days.
They also found that before 39 weeks’ gestation, CL did not predict
time to delivery. One explanation for the difference between their
subcategorizing our population by parity, gestational age, and findings and ours may reflect the reproductive history and ethnic
CL measurements of 5 mm, we found that CL of 25 mm has a differences in the study populations; their population consisted of
sensitivity of 56% and a specificity of 67% in predicting delivery by nulliparous Italian women while our population included
41 weeks. Using different cutoff points for CL, sensitivity, specificity, multiparous women as well with the majority being (75%)
PPV and NPV are outlined in Table 4. Hispanic. Rozenberg et al.10 in 2000 enrolled 128 women with
We constructed the ROC curve to estimate the probability of singleton pregnancies at 39–40 weeks and found an association
delivery by 41 weeks at each CL measurement (Figure 2). The between CL and the likelihood of spontaneous onset of labor within

Journal of Perinatology
 CL and the risk of spontaneous labor at term
LL Tolaymat et al
752

Table 3 Predictors of delivery by 41 weeks using logistic regression model 1

Estimate s.e. w2 P-value 0.9

0.8
GA at US 0.6737 0.2381 8.0085 0.0047

Sensitivity (true positives)

Birthweight 0.00046 0.000569 0.6474 0.4211 0.7
Cervical length 0.0619 0.0264 5.5110 0.0189
0.6
Nulliparity 0.0372 0.2438 0.0232 0.8788
0.5
Abbreviations: GA, gestational age; US, ultrasound.
0.4
Table 4 Predictive values of cervical length for delivery by 41 weeks for each 0.3
subcategory
0.2 No discrimination
Strata CL Sensitivity Specificity PPV NPV
0.1 Shortest_CL
(mm) (%) (%) (%) (%)
0
All (n ¼ 120) 30 73.1 40.7 81.0 30.6 0 0.2 0.4 0.6 0.8 1
25 55.9 66.7 85.2 30.5 1-Specificity (false positives)
22 39.8 74.1 84.1 26.3
Figure 2 Likelihood of delivery by 41 weeks. Receiver operating characteristic
20 36.6 77.8 85.0 26.3
curve (ROC) using cervical length (CL) to predict delivery by 41 weeks.
15 20.4 92.6 90.5 25.3
10 7.5 100 100 23.9
differences can be due to differences in patient population,
Multiparous (n ¼ 54) 30 62.5 35.7 73.5 25.0 ultrasound operators or simply chance.
25 47.5 64.3 79.2 30.0 In 2003, Ramanathan et al.9 published their findings of CL at
22 32.5 71.4 76.5 27.0 37 weeks’ gestation as a predictor of delivery after 40 weeks and 10
20 30.0 78.6 80.0 28.2 days. They found that CL of <20 mm at 37 weeks is associated with
15 17.5 92.9 87.5 28.3 0% incidence of delivery after 40 weeks and 10 days. They did not
assess CL at any other gestational age. Finally, Bayramoglu et al.11
Nulliparous (n ¼ 66) 30 81.1 46.2 86.0 37.5 in 2005 reported on a cohort of 93 women at 37–40 weeks and
25 62.3 69.2 89.2 31.0 assessed the relationship between CL and delivery within 7 days.
22 45.3 76.9 88.9 25.6
Using the ROC curve, they identified optimal cutoff measurements
20 41.5 76.9 88.0 24.4
(29.5, 27.5, 25.5 and 24.5 mm) at 37, 38, 39 and 40 weeks
15 22.6 92.3 92.3 22.6
10 13.2 100 100 22.0
respectively.
To date, our study represents the largest report that assesses the
GA at USX38 weeks 30 76.6 23.8 69.2 31.3 relationship between CL and delivery within 7 days and delivery by
(n ¼ 68) 41 weeks that includes pregnant women with a wider range of
25 61.7 57.1 76.3 40.0 gestational age (37–40 weeks). Our findings demonstrate the
22 51.1 66.7 77.4 37.8 utility of CL in predicting either outcome, independent of specific
20 46.8 71.4 78.6 37.5 gestational age at ultrasound. One possible use of these findings is
15 23.4 90.5 84.6 34.5 to predict the likelihood of delivery by 41 weeks in a pregnant
10 6.4 100 100 32.3 woman with a previous cesarean delivery. A CL of 10 mm or less
has a PPV of 90% of delivering by 41 weeks. However, a limitation
GA at US<38 weeks 30 69.6 100 100 30.0
of using our findings for this application is that our population
(n ¼ 52)
included only had 5/120 patients with previous cesarean who
25 50.0 100 100 20.7
20 26.1 100 100 15.0
underwent a trial of labor. Women with previous cesarean delivery
15 17.4 100 100 13.6 may have different characteristics from nulliparous women and
multiparous women with a previous vaginal delivery. Another
Abbreviations: CL, cervical length; GA, gestational age; US, ultrasound.
possible use for these data is in women whose fetuses require
immediate neonatal resuscitation that is not available at a hospital
7 days. They chose a CL of 26 mm based on an ROC curve, but the in their hometown. Such women may optimally be referred for
reported sensitivity, specificity, PPV and NPV were 56.5, 71.1, 76.1 delivery at a hospital away from where they reside. A CL of at least
and 50% respectively. We found a cutoff of 25 mm has the best 30 mm in a nulliparous pregnant woman has an NPV of 91%, which
sensitivity and NPV of 77.5 and 84.7% respectively (Table 3). These means there is a 91% chance of not delivering within the next 7 days

Journal of Perinatology

(Table 2), a prediction which could prove reassuring for clinicians
and patients in making decisions regarding maternal transport.
Despite the size of our study, we recognize that our cohort was
underpowered to demonstrate differences in CL screening between
various subgroups (for example, parity, gestational age and
race/ethnicity) of low-risk women with singleton pregnancies.
Nonetheless, our findings may prove useful as a pilot study to direct
future studies in investigating the significance of a specific cutoff in
this estimation, rather than observing a cohort and trying to find an
‘optimal’ cutoff. One CL cutoff value that seems to recur in our study
and that of Ramanathan et al. is 10 mm. Future studies could
determine the value of this cutoff in assessing the likelihood of
delivery within 7 days and delivery by 41 weeks in women with a
previous cesarean delivery who are interested in a trial of labor.
The additional cost to prenatal care may not justify the routine
use of CL measurement in low-risk patients with no specific
indication for this assessment. In contrast, for patients who are
planning to relocate for the delivery; the cost of the CL
measurement may be offset by the need for fewer days in a hotel.
In addition, many practitioners have an ultrasound in their office
and can assess CL without incurring major costs. Future studies of
this intervention should consider such costs as well as particular
subgroups of women who might benefit from this information.
Acknowledgments
This study is funded by the American College of Obstetricians and Gynecologists,
District IV 2001 Special Award and presented in part at the Annual Clinical
Meeting of American College of Obstetricians and Gynecologists in 2002.
CL and the risk of spontaneous labor at term
LL Tolaymat et al
753
References
1 Bernstine RL, Sung HL, Crawford WI, Shimek MP. Sonographic evaluation of the
incompetent cervix. J Clin Ultrasound 1981; 9: 417–420.
2 Iams JD, Paraskos J, Landon MB, Teteris JN, Johnson FF. Clinical sonography in
preterm labor. Obstet Gynecol 1994; 84: 40–46.
3 Iams JD, Goldenberg RL, Meis PJ, Mercer BM, Moawad A, Das A et al. The length
of the cervix and the risk of spontaneous delivery. N Engl J Med 1996; 334:
567–572.
4 Ware V, Raynor BD. Transvaginal ultrasonographic cervical measurement as a
predictor of successful labor induction. Am J Obstet Gynecol 2000; 1882: 1030–1032.
5 Boozarjomhri F, Timor-Tritsch I, Chao CR, Fox HE. Transvaginal ultrasonographic
evaluation of the cervix before labor: presence of cervical wedging is associated with
shorter duration of induced labor. Am J Obstet Gynecol 1994; 171: 1081–1087.
6 Reis FM, Gervasi MT, Florio P, Bracalente G, Fadalti M, Servi FM et al. Prediction of
successful induction of labor at term: role of clinical history, digital examination,
ultrasound assessment of the cervix and fetal fibronectin assay. Am J Obstet Gynecol
2003; 189: 1361–1367.
7 Rane SM, Guirgis RR, Higgins B, Nicolaides KH. Models for the prediction of successful
induction of labor based on pre-induction sonographic measurement of cervical
length. J Matern Fetal Neonatal Med 2005; 17: 315–322.
8 Vimercati A, Greco P, Lopalco P, Loizzi V, Scioscia M, Mei L et al. The value of
ultrasonographic examination of the uterine cervix in predicting post-term pregnancy.
J Perinat Med 2001; 29: 317–321.
9 Ramanathan G, Yu C, Osei E, Nicolaides KH. Ultrasound examination at 37 weeks’
gestation in the prediction of pregnancy outcome: the value of cervical assessment.
Ultrasound Obstet Gynecol 2003; 22: 598–603.
10 Rozenberg P, Goffinet F, Hessabi M. Comparison of the Bishop score,
ultrasonographically measured cervical length, and fetal fibronectin assay in predicting
time until delivery and type of delivery at term. Am J Obstet Gynecol 2000; 182:
108–113.
11 Bayramoglu O, Arslan M, Yazici G, Erdem A, Erdem M, Bayramoglu K et al. Prediction
of spontaneous onset of labor at term: the role of cervical length measurement and
funneling of internal os detected by transvaginal ultrasonography. Am J Perinatol
2005; 22: 35–39.
Journal of Perinatology