Clinical significance of early (< 20 weeks) vs. late (20–24 weeks) detection of sonographic short cervix in asymptomatic women in the mid-trimester

Authors

  • E. Vaisbuch,

    Corresponding author
    1. Perinatology Research Branch, Intramural Division, NICHD/NIH/DHHS, Hutzel Women's Hospital, Bethesda, MD and Detroit, MI, USA
    2. Department of Obstetrics and Gynecology, Wayne State University/Hutzel Women's Hospital, Detroit, MI, USA
    • Perinatology Research Branch, NICHD, NIH, DHHS, Wayne State University/Hutzel Women's Hospital, 3990 John R, Box 4, Detroit, MI 48201, USA
    Search for more papers by this author
  • R. Romero,

    Corresponding author
    1. Perinatology Research Branch, Intramural Division, NICHD/NIH/DHHS, Hutzel Women's Hospital, Bethesda, MD and Detroit, MI, USA
    2. Center for Molecular Medicine and Genetics, Wayne State University, Detroit, MI, USA
    • Perinatology Research Branch, NICHD, NIH, DHHS, Wayne State University/Hutzel Women's Hospital, 3990 John R, Box 4, Detroit, MI 48201, USA
    Search for more papers by this author
    • R.R. and F.G. have contributed to this work as part of their official duties as employees of the United States Federal Government.

  • O. Erez,

    1. Perinatology Research Branch, Intramural Division, NICHD/NIH/DHHS, Hutzel Women's Hospital, Bethesda, MD and Detroit, MI, USA
    2. Department of Obstetrics and Gynecology, Wayne State University/Hutzel Women's Hospital, Detroit, MI, USA
    Search for more papers by this author
  • J. P. Kusanovic,

    1. Perinatology Research Branch, Intramural Division, NICHD/NIH/DHHS, Hutzel Women's Hospital, Bethesda, MD and Detroit, MI, USA
    2. Department of Obstetrics and Gynecology, Wayne State University/Hutzel Women's Hospital, Detroit, MI, USA
    Search for more papers by this author
  • S. Mazaki-Tovi,

    1. Perinatology Research Branch, Intramural Division, NICHD/NIH/DHHS, Hutzel Women's Hospital, Bethesda, MD and Detroit, MI, USA
    2. Department of Obstetrics and Gynecology, Wayne State University/Hutzel Women's Hospital, Detroit, MI, USA
    Search for more papers by this author
  • F. Gotsch,

    1. Perinatology Research Branch, Intramural Division, NICHD/NIH/DHHS, Hutzel Women's Hospital, Bethesda, MD and Detroit, MI, USA
    Search for more papers by this author
    • R.R. and F.G. have contributed to this work as part of their official duties as employees of the United States Federal Government.

  • V. Romero,

    1. Department of Obstetrics and Gynecology, Wayne State University/Hutzel Women's Hospital, Detroit, MI, USA
    Search for more papers by this author
  • C. Ward,

    1. Department of Obstetrics and Gynecology, Wayne State University/Hutzel Women's Hospital, Detroit, MI, USA
    Search for more papers by this author
  • T. Chaiworapongsa,

    1. Perinatology Research Branch, Intramural Division, NICHD/NIH/DHHS, Hutzel Women's Hospital, Bethesda, MD and Detroit, MI, USA
    2. Department of Obstetrics and Gynecology, Wayne State University/Hutzel Women's Hospital, Detroit, MI, USA
    Search for more papers by this author
  • P. Mittal,

    1. Perinatology Research Branch, Intramural Division, NICHD/NIH/DHHS, Hutzel Women's Hospital, Bethesda, MD and Detroit, MI, USA
    2. Department of Obstetrics and Gynecology, Wayne State University/Hutzel Women's Hospital, Detroit, MI, USA
    Search for more papers by this author
  • Y. Sorokin,

    1. Department of Obstetrics and Gynecology, Wayne State University/Hutzel Women's Hospital, Detroit, MI, USA
    Search for more papers by this author
  • S. S. Hassan

    1. Perinatology Research Branch, Intramural Division, NICHD/NIH/DHHS, Hutzel Women's Hospital, Bethesda, MD and Detroit, MI, USA
    2. Department of Obstetrics and Gynecology, Wayne State University/Hutzel Women's Hospital, Detroit, MI, USA
    Search for more papers by this author

Abstract

Objective

The aim of this study was to determine whether the risk of early spontaneous preterm delivery (PTD) in asymptomatic women with a sonographic cervical length of ≤ 15 mm in the mid-trimester changes as a function of gestational age at diagnosis.

Methods

This cohort study included 109 asymptomatic patients with a sonographic cervical length of ≤ 15 mm diagnosed at 14–24 weeks of gestation. Women with a multifetal gestation, cerclage and a cervical dilatation of > 2 cm were excluded. The study population was stratified by gestational age at diagnosis (< 20 weeks vs. 20–24 weeks) and by cervical length (≤ 10 mm vs. 11–15 mm). The primary outcome variables were PTD at < 28 and < 32 weeks of gestation and the diagnosis-to-delivery interval.

Results

The median gestational age at diagnosis of a short cervix before 20 weeks and at 20–24 weeks was 18.9 and 22.7 weeks, respectively. Women diagnosed before 20 weeks had a higher rate of PTD at < 28 weeks (76.9% vs. 30.9%; P < 0.001) and at < 32 weeks (80.8% vs. 48.1%; P = 0.004), and a shorter median diagnosis-to-delivery interval (21 vs. 61.5 days, P = 0.003) than those diagnosed at 20–24 weeks. The rate of amniotic fluid sludge was higher among patients diagnosed with a short cervix at < 20 weeks of gestation than in those in whom it was diagnosed between 20 and 24 weeks (92.3% vs. 48.2%; P < 0.001).

Conclusions

Asymptomatic women with a sonographic cervical length of ≤ 15 mm diagnosed before 20 weeks of gestation have a dramatic and significantly higher risk of early preterm delivery than women diagnosed at 20–24 weeks. These findings can be helpful to physicians in counseling these patients, and may suggest different mechanisms of disease leading to a sonographic short cervix before or after 20 weeks of gestation. Copyright © 2010 ISUOG. Published by John Wiley & Sons, Ltd.

Introduction

Cervical sonography is the most objective and reliable method to assess cervical length1–6. The shorter the cervical length is in the mid-trimester, the higher the risk of spontaneous preterm delivery (PTD)1–4, 7–23. However, no agreement exists as to what is the precise definition of a sonographic short cervix2–4, 24. Using a cut-off of 15 mm at 22–24 weeks of gestation, Heath et al.3 reported a 4% risk and a likelihood ratio of 2.7 for delivery at ≤ 32 weeks, while Hassan et al.4, among asymptomatic women diagnosed with a cervical length of ≤ 15 mm between 14 and 24 weeks of gestation, reported a sensitivity and positive predictive value for a preterm delivery at ≤ 32 weeks of 8.2% and 47.6%, respectively. The prevalence of a sonographic cervical length of ≤ 15 mm ranges from 0.6%4 at 14–24 weeks to 1–1.7%3, 25–28 at 20–24 weeks.

The importance of identifying women with a sonographic short cervix in the mid-trimester derives from evidence that those patients, regardless of whether or not they have a history of preterm birth, may be candidates for therapeutic interventions such as vaginal progesterone29–31, cervical cerclage32, 33, antibiotics34 or indomethacin35. However, a short cervix is syndromic in nature and can be caused by multiple etiologies36, 37 such as subclinical intrauterine infection34, 38, 39 and/or inflammation40, 41 and cervical insufficiency. Thus, although results of randomized clinical trials evaluating the use of progesterone or cerclage have been promising29, 33, 42, these therapeutic interventions may be beneficial only in a subset of patients.

Recently, we reported that the presence of sonographic dense aggregates of particulate matter in the amniotic fluid close to the internal cervical os (‘sludge’) is associated with microbial invasion of the amniotic cavity (MIAC), impending preterm delivery and histologic chorioamnionitis43. Furthermore, amniotic fluid sludge has been identified in the mid-trimester in asymptomatic women at risk for PTD and is an independent risk factor for preterm prelabor rupture of membranes (PPROM) and PTD44.

The aims of this study were to determine whether the risk of early spontaneous PTD in asymptomatic patients with a short cervix (≤ 15 mm) in the mid-trimester varies according to gestational age and cervical length at diagnosis, and whether this risk is altered by the presence of amniotic fluid sludge.

Methods

Study population

This retrospective cohort study included pregnant women followed in our Cervix Clinic between January 2002 and June 2008 as part of a longitudinal protocol whose aim was to identify biochemical factors for the prediction of adverse pregnancy outcome. Patient characteristics were obtained by research nurses at the first antenatal visit and were updated prospectively. Transvaginal ultrasound (TVUS) examination, as well as digital assessment of the cervix, were performed every 2–4 weeks beginning as early as 14 weeks of gestation. The ultrasound findings were recorded and stored in a dedicated database. A subset of patients included in this study were diagnosed with a short cervix during a routine TVUS examination in a community clinic and were sent to our hospital for further evaluation. Patients who agreed to participate were enrolled into our protocol and a confirmatory TVUS examination was performed. After the diagnosis of a short cervix, patients were referred to the ‘Labor and Delivery’ ward for further evaluation and management. Both the sonographic cervical length and the result of the digital vaginal examination were available to the physicians, and the management of these patients was at the discretion of the attending physician.

Using a computer-based search of our clinical and sonographic databases, asymptomatic patients with a viable fetus and a documented TVUS cervical length of ≤ 15 mm diagnosed between 14 and 24 weeks of gestation, were identified. Patients with one or more of the following conditions were excluded: multifetal pregnancy; preterm labor or PPROM at the time of diagnosis; a cervical cerclage (placed before or after the diagnosis of a short cervix); cervical dilatation > 2 cm by digital examination; placenta previa; and fetuses with chromosomal and/or congenital anomalies.

All women provided written informed consent before participating in the study protocol. The use of clinical and ultrasound data for research purposes was approved by the Institutional Review Boards of Wayne State University and the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD/NIH/DHHS).

Definitions and study procedures

Gestational age was determined by the last menstrual period and confirmed by ultrasound examination, or by ultrasound examination alone if the sonographic determination of gestational age was not consistent with menstrual dating by more than 1 week. Gestational age at diagnosis was defined as the earliest gestational age at which a documented sonographic cervical length of ≤ 15 mm was recorded. Patients with an a priori risk for PTD (high-risk patients) were defined as those having at least one of the following: one or more previous PTDs (≤ 35 weeks); one or more late mid-trimester spontaneous miscarriages (≥ 16 weeks); and prior cervical surgery (loop electrosurgical excision procedure or cone biopsy)45. Spontaneous preterm labor was defined by the presence of regular uterine contractions occurring at least twice every 10 min, associated with cervical change before 37 completed weeks of gestation. Data regarding pregnancy outcome were obtained from the hospital records, and an attempt to retrieve the delivery data was made for patients who did not deliver in our hospital. Patients who were lost to follow-up and those for whom no delivery data were available were censored from the last documented follow-up visit. Early PTD was defined as delivery before 32 completed weeks of gestation. Patients defined as having spontaneous PTD included those with spontaneous preterm labor, PPROM and those in whom labor was induced or augmented as a result of chorioamnionitis (defined according to the diagnostic criteria proposed by Gibbs et al.46). Histologic chorioamnionitis was diagnosed based on the presence of inflammatory cells in the chorionic plate and/or chorioamniotic membranes47, and acute funisitis was diagnosed by the presence of neutrophils in the wall of the umbilical vessels and/or Wharton's jelly, using the criteria previously described48.

Sonographic assessment of the cervix

TVUS examination was conducted using commercially available two-dimensional (2D) and three-dimensional (3D) ultrasound systems (Acuson Sequoia, Siemens Medical Systems, Mountain View, CA, USA; Voluson 730 Expert or Voluson E8, GE Healthcare, Milwaukee, WI, USA) equipped with endovaginal transducers with frequency ranges of 5–7.5 MHz and 5–9 MHz, respectively. All sonographic examinations of the cervix were performed by registered diagnostic medical sonographers using a previously described technique1, 2, and reviewed by an experienced physician. Amniotic fluid sludge was identified by the presence of dense aggregates of particulate matter in proximity to the internal cervical os49. An experienced sonographer, blinded to clinical outcome, reviewed the 2D images and 3D volume datasets of the cervix to determine the presence or absence of amniotic fluid sludge.

Statistical analysis

The primary outcome variables included PTD at < 28 weeks and < 32 weeks of gestation, as well as the diagnosis-to-delivery interval. Secondary outcome variables included PTD at < 24, < 35 and < 37 weeks of gestation. Patients were stratified by gestational age at diagnosis (< 20 weeks vs. 20–24 weeks) and by cervical length (≤ 10 mm vs. 11–15 mm). Patients who had an indicated PTD for a diagnosis (i.e. severe pre-eclampsia, fetal growth restriction or fetal death) that was not directly related to the initial diagnosis of a short cervix (i.e. proven or clinical chorioamnionitis, or PPROM) were removed from the statistical analyses at the time of onset of induction of labor.

Between-group comparisons were performed using Fisher's exact test for categorical variables and the Mann–Whitney U-test for continuous variables. Multivariate stepwise logistic regression analyses were performed to determine the relationship between the primary outcome variables (i.e. PTD at < 28 and < 32 weeks of gestation) and possible explanatory variables, including maternal age (years), prepregnancy body mass index (BMI), parity, smoking status, an a priori risk for PTD, gestational age at diagnosis (weeks), cervical length (mm) and the presence of amniotic fluid sludge. Kaplan–Meier survival analyses were performed to assess the diagnosis-to-delivery interval, according to gestational age, upon diagnosis of a short cervix, as well as by the presence or absence of an a priori risk for PTD. A P < 0.05 was considered statistically significant. The statistical package used was SPSS v.12.0 (SPSS Inc., Chicago, IL, USA).

Results

Demographic and clinical characteristics of the study population

During the study period, 200 asymptomatic women were identified with a sonographic cervical length of ≤ 15 mm between 14 and 24 weeks of gestation. Of those, 115 met the inclusion criteria. The reasons for exclusion were: multifetal gestation (n = 24), cervical cerclage (n = 46), cervical dilatation > 2 cm (n = 8), anhydramnios with suspected PPROM at the time of diagnosis (n = 4) and fetal anomalies (n = 3). Four patients were lost to follow-up after the initial diagnosis, one patient withdrew consent to participate in the study and another had two pregnancies during the study period that met the inclusion criteria and only the first was included in the study. Thus, 109 patients were included in the final analyses.

Demographic and clinical characteristics of the study population are shown in Table 1. Fifty-five women (50.5%) were nulliparous, while 41 women (37.6%) had a history of at least one PTD and/or a late mid-trimester miscarriage. The median gestational age at diagnosis of a sonographic cervical length of ≤ 15 mm was 21.9 (range, 16.4–24) weeks. Twenty-six women (23.9%) were diagnosed with a sonographic cervical length of ≤ 15 mm before 20 weeks, and 83 (76.1%) between 20 and 24 weeks of gestation. There were no significant differences in the demographic and clinical characteristics of women diagnosed with a cervical length of ≤ 15 mm before 20 weeks and those diagnosed between 20 and 24 weeks of gestation (Table 1).

Table 1. Demographic and clinical characteristics of the study population
 GA at diagnosis of short cervix 
 14–24 weeks (n = 109)< 20 weeks (n = 26)20–24 weeks (n = 83)P*
  • Data are given as median (interquartile range) or % (n).

  • *

    Diagnosis at < 20 vs. 20–24 weeks.

  • BMI, body mass index; CL, cervical length; D&C, dilatation and curettage; GA, gestational age; NS, not significant.

Age (years)26 (21.2–31.5)25.5 (22–34.1)26 (21–31)NS
Ethnic origin
 African–American88.8 (96)92.3 (24)86.7 (72)NS
 Other11.2 (13)7.7 (2)13.3 (11)NS
Smoker21.6 (22/102)28 (7/25)19.5 (15/77)NS
Prepregnancy BMI (kg/m2)27.5 (24.1–31.2)28.8 (25.9–34.8)26.4 (23.3–30.6)NS
Nulliparous50.5 (55)46.2 (12)51.8 (43)NS
History of one or more D&C61.8 (63/102)66.7 (16/24)60.3 (47/78)NS
History of spontaneous preterm delivery37.6 (41)38.5 (10)37.3 (31)NS
Prophylactic 17α-hydroxyprogesterone caproate13.8 (15)20 (5)12 (10)NS
GA at diagnosis (weeks)21.9 (20.1–23.3)18.9 (17.7–19.1)22.7 (21.4–23.4)< 0.001
CL at diagnosis (mm)11 (5–13.5)8.5 (1.5–12)11 (6–14)NS

Association between gestational age at diagnosis of short cervix, cervical length and PTD

Women diagnosed sonographically with a short cervix at < 20 weeks had a shorter median diagnosis-to-delivery interval (21 days, interquartile range (IQR) 9–63.5 vs. 61.5 days, IQR 27.2–106.7; P = 0.003) and a lower median gestational age at delivery (21 weeks, IQR 20–26.6 vs. 32 weeks, IQR 26.8–37.9; P < 0.001) than those diagnosed between 20 and 24 weeks of gestation (Table 2). The rates of PTD delivery at < 28 weeks (76.9% vs. 30.9%; P < 0.001) and < 32 weeks (80.8% vs. 48.1%; P = 0.004), as well as at < 24 (65.4% vs. 6.0%; P < 0.001), < 35 (80.8% vs. 57.1%; P = 0.031) and < 37 (92.3% vs. 69.7%; P = 0.021) weeks, were all higher in patients diagnosed at < 20 weeks than those diagnosed between 20 and 24 weeks (Table 2). Figure 1 displays the Kaplan–Meier survival curves of diagnosis-to-delivery interval, according to gestational age at diagnosis, among asymptomatic patients diagnosed with a cervical length of ≤ 15 mm. The two survival curves differed significantly (log rank, P = 0.004).

Figure 1.

Kaplan–Meier survival curves of diagnosis-to-delivery interval among asymptomatic patients diagnosed with a cervical length of ≤ 15 mm, before 20 gestational weeks (equation image) (n = 26) or between 20 and 24 weeks (n = 83) (equation image). Five patients were censored because they were lost to follow-up, two because of preterm induction of labor and 35 because they reached 37 completed weeks of gestation. The two survival curves differed significantly (log rank, P = 0.004).

Table 2. Gestational age (GA) at delivery, diagnosis-to-delivery interval and rate of spontaneous preterm delivery at < 24, < 28, < 32, < 35 and < 37 weeks of gestation among patients diagnosed with a sonographic cervical length of ≤ 15 mm between 14 and 24 weeks of gestation, stratified according to GA at diagnosis
 GA at diagnosis of short cervix 
 14–24 weeks (n = 109)< 20 weeks (n = 26)20–24 weeks (n = 83)P*
  • Data are given as median (interquartile range) or % (n).

  • *

    Diagnosis at < 20 weeks vs. 20–24 weeks.

GA at delivery (weeks)29.5 (24.3–36.3)21 (20–26.6)32 (26.8–37.9)< 0.001
Diagnosis-to-delivery interval (days)54 (20–105)21 (9–63.5)61.5 (27.2–106.7)0.003
Spontaneous preterm delivery
 < 24 weeks20.2 (22)65.4 (17)6.0 (5)< 0.001
 < 28 weeks42.1 (45/107)76.9 (20)30.9 (25/81)< 0.001
 < 32 weeks56.2 (59/105)80.8 (21)48.1 (38/79)0.004
 < 35 weeks63.1 (65/103)80.8 (21)57.1 (44/77)0.031
 < 37 weeks75.5 (77/102)92.3 (24)69.7 (53/76)0.021

Stratification by cervical length at diagnosis revealed that patients with a cervical length of ≤ 10 mm had a lower median gestational age at delivery (26.9 weeks, IQR 22.2–33.7 vs. 33.9 weeks, IQR 26.4–37.7; P = 0.011) and a shorter median diagnosis-to-delivery interval (33 days, IQR 13–77 vs. 77 days, IQR 33.5–110.5; P = 0.002) than those with a cervical length of 11–15 mm (Table 3). Similarly, women with a cervical length of ≤ 10 mm had a higher rate of PTD at < 28 and < 32 weeks, as well as at < 24 and < 35 weeks of gestation, than those with a cervical length of 11–15 mm (Table 3).

Table 3. Gestational age (GA) at delivery, diagnosis-to-delivery interval and rate of preterm delivery at < 24, < 28, < 32, < 35 and < 37 weeks of gestation among patients diagnosed with a sonographic cervical length of ≤ 15 mm between 14 and 24 weeks of gestation, stratified according to cervical length
 Cervical length 
 ≤ 10 mm (n = 55)11–15 mm (n = 54)P
  1. Data are given as median (interquartile range) or % (n). NS, not significant.

GA at delivery (weeks)26.9 (22.2–33.7)33.9 (26.4–37.7)0.011
Diagnosis-to-delivery interval (days)33 (13–77)77 (33.5–110.5)0.002
Spontaneous preterm delivery
 < 24 weeks29.1 (16)11.1 (6)0.019
 < 28 weeks54.5 (30)28.8 (15/52)0.007
 < 32 weeks72.7 (40)38 (19/50)< 0.001
 < 35 weeks75.9 (41/54)49 (24/49)0.005
 < 37 weeks79.6 (43/54)70.8 (34/48)NS

When patients were stratified by both gestational age and cervical length at diagnosis, the rate of PTD at < 28 and < 32 weeks was higher in women with a cervical length of ≤ 10 mm than in those with cervical length of 11–15 mm only among patients diagnosed between 20 and 24 weeks of gestation (Table 4).

Table 4. Rate of preterm delivery at < 24, < 28, < 32, < 35 and < 37 weeks of gestation among patients diagnosed with a sonographic cervical length (CL) of ≤ 15 mm between 14 and 24 weeks of gestation, stratified by gestational age (GA) at diagnosis and CL
 GA at diagnosis of short cervix
 < 20 weeks (n = 26) 20–24 weeks (n = 83) 
 CL ≤ 10 mm (n = 16)CL 11–15 mm (n = 10)PCL ≤ 10 mm (n = 39)CL 11–15 mm (n = 44)P
  1. Data are given as median (interquartile range) or % (n). NS, not significant.

GA at delivery (weeks)20.823.6NS28.335.20.02
 (20.1–25.9)(19.7–35) (25.5–37.5)(28.7–38.1) 
Diagnosis-to-delivery interval (days)1728.5NS42830.01
 (6.7–50.7)(14–114.5) (18–101.5)(48–110) 
Spontaneous preterm delivery
 < 24 weeks75 (12)50 (5)NS10.3 (4)2.3 (1)NS
 < 28 weeks81.3 (13)70 (7)NS43.6 (17)19 (8/42)0.02
 < 32 weeks87.5 (14)70 (7)NS66.7 (26)30 (12/40)0.001
 < 35 weeks86.7 (13)70 (7)NS71.1 (27/38)43.6 (17/39)0.02
 < 37 weeks93.8 (15)90 (9)NS73.7 (28/38)65.8 (25/38)NS

Association between gestational age and cervical length at diagnosis of short cervix and amniotic fluid ‘sludge’

TVUS demonstrated the presence of amniotic fluid sludge in 58.7% (64/109) of the study population at the time of diagnosis. The rate of amniotic fluid sludge was higher among patients diagnosed at < 20 weeks of gestation than among those diagnosed between 20 and 24 weeks (92.3% (24/26) vs. 48.2% (40/83); P < 0.001). Similarly, the rate of amniotic fluid sludge was higher among patients with a cervical length of ≤ 10 mm than in those with a cervical length of 11–15 mm (70.9% (39/55) vs. 46.3% (25/54); P = 0.009). Patients with amniotic fluid sludge had a lower median gestational age at delivery (26.1 weeks, IQR 21.7–32.4 vs. 35.1 weeks, IQR 28.3–38.3; P < 0.001) and a shorter median diagnosis-to-delivery interval (28 days, IQR 14–73 vs. 84 days, IQR 39–111; P < 0.001) than those without this finding (Table 5). The rate of PTD at < 28 and < 32 weeks, as well as at < 24, < 35 and < 37 weeks of gestation were all significantly higher in patients with amniotic fluid sludge than in those without it (Table 5). These differences remained significant or near-significant after further stratifying the study populations by cervical length at diagnosis (≤ 10 mm and 11–15 mm) (Table 5).

Table 5. Rate of spontaneous preterm delivery at < 28, < 32, < 35 and < 37 weeks of gestation among asymptomatic patients with a sonographic cervical length (CL) of ≤ 15 mm diagnosed between 14 and 24 weeks, stratified according to the presence or absence of amniotic fluid sludge and cervical length (CL)
 Amniotic fluid sludge 
 PresentAbsentP
  1. Data expressed as n, median (interquartile range) or % (n). AF, amniotic fluid; GA, gestational age.

Cervical length ≤ 15 mm
 n6445 
 GA at delivery (weeks)26.1 (21.7–32.4)35.1 (28.3–38.3)< 0.001
 Diagnosis-to-delivery interval (days)28 (14–73)84 (39–111)< 0.001
 Spontaneous preterm delivery
  < 24 weeks31.3 (20)4.4 (2)0.001
  < 28 weeks58.1 (36/62)20 (9)< 0.001
  < 32 weeks71 (44/62)34.9 (15/43)< 0.001
  < 35 weeks76.7 (46/60)44.2 (19/43)0.001
  < 37 weeks86.4 (51/59)60.5 (26/43)0.003
Cervical length 11–15 mm
 n2529 
 Spontaneous preterm delivery
  < 24 weeks20 (5)3.4 (1)0.054
  < 28 weeks52.2 (12/23)10.3 (3)0.001
  < 32 weeks56.5 (13/23)22.2 (6/27)0.013
  < 35 weeks63.6 (14/22)37 (10/27)0.064
  < 37 weeks85.7 (18/21)59.3 (16/27)0.045
Cervical length ≤ 10 mm
 n3916 
 Spontaneous preterm delivery
  < 24 weeks38.5 (15)6.3 (1)0.017
  < 28 weeks61.5 (24)37.5 (6)0.1
  < 32 weeks79.5 (31)56.3 (9)0.08
  < 35 weeks84.2 (32/38)56.3 (9)0.028
  < 37 weeks86.8 (33/38)62.5 (10)0.043

Outcome according to a priori risk for preterm delivery in asymptomatic patients between 14 and 24 weeks of gestation with a short cervix

Among the study population, 42 patients (38.5%) had a history consistent with an a priori higher risk of PTD. Besides maternal age and the percentage of patients who received prophylactic 17α-hydroxyprogesterone caproate, there were no significant differences in the demographic and clinical characteristics of low-risk and high-risk patients (Table 6). The Kaplan–Meier survival curves of patients with an a priori risk of PTD and low-risk patients differed significantly (log rank, P = 0.003) (Figure 2). Patients at low risk had a better ‘survival’ than those with an a priori risk for PTD.

Figure 2.

Kaplan–Meier survival curves of patients with an a priori risk of spontaneous preterm delivery (high-risk patients, n = 42) (equation image) and those without such a risk (low-risk patients, n = 67) (equation image). Five patients were censored because they were lost to follow-up, two because of preterm induction of labor and 35 because they reached 37 completed weeks of gestation. The two survival curves differed significantly (log rank, P = 0.003).

Table 6. Demographic and clinical characteristics of patients stratified according to their a priori risk for spontaneous preterm birth
 Low risk (n = 67)High risk (n = 42)P
  1. Data are given as median (interquartile range) or % (n). BMI, body mass index; CL, cervical length; GA, gestational age; NS, not significant.

Age (years)23.5 (21–29)30 (23–35)0.006
Prepregnancy BMI (kg/m2)28.4 (24.1–31.4)26.3 (24–31.1)NS
Smoker17.7 (11/62)27.7 (11/40)NS
Prophylactic 17α-hydroxyprogesterone caproate6 (4)27.5 (11/40)0.003
CL at diagnosis (mm)10 (5–13)11.5 (5–14)NS
GA at diagnosis of short cervix (weeks)21.7 (20.3–23.1)22.5 (19.8–23.3)NS
Amniotic fluid sludge52.2 (35)69 (29)NS
Gestational age at delivery (weeks)30.3 (25.1–38.4)28.6 (23.4–34.4)0.046
Diagnosis-to-delivery interval (days)60 (22.5–112.5)48 (16–78)0.039
Spontaneous preterm delivery
 < 24 weeks16.4 (16)26.2 (11)NS
 < 28 weeks39.4 (26/66)45.2 (19)NS
 < 32 weeks48.4 (31/64)68.3 (28/41)0.035
 < 35 weeks54.8 (34/63)75.6 (31/41)0.026
 < 37 weeks63.9 (39/41)92.7 (38/41)< 0.001

Of note, overall, the median diagnosis-to-delivery interval did not differ significantly between women who received prophylactic 17α-hydroxyprogesterone caproate and those who did not (56 days, IQR 42–68 vs. 48 days, IQR 20–108, respectively; P = 0.8).

Frequency of histologic chorioamnionitis and association with gestational age at diagnosis and diagnosis-to-delivery interval

Placental histopathologic examination was available for 97 (89%) patients. Patients who delivered within 14 days from diagnosis of an asymptomatic short cervix had a higher rate of histologic chorioamnionitis (maternal and/or fetal side) than those who delivered after 14 days (84.2% (16/19) vs. 55.1% (43/78), P = 0.016). Similarly, among women diagnosed between 20 to 24 weeks of gestation, the rate of histologic chorioamnionitis was higher in patients who delivered within 14 days than in those who delivered after more than 14 days from diagnosis (100% (8/8) vs. 51.6% (33/64), P = 0.008). In contrast, among patients diagnosed before 20 weeks, the rate of histologic chorioamnionitis was similar in patients who delivered within 14 days and those who delivered after 14 days (72.7% (8/11) vs. 71.4% (10/14), respectively; P = 0.6).

Factors independently associated with the outcome variables

Multivariate stepwise logistic regression analyses were used to explore the relationship between possible explanatory variables, including maternal age, prepregnancy BMI, parity, smoking status, a priori risk for PTD, gestational age at diagnosis, cervical length and the presence of amniotic fluid sludge as well as the primary outcome variables (PTD at < 28 and < 32 weeks of gestation). In these models, only gestational age (P = 0.007 and P = 0.016, respectively) and cervical length at diagnosis (P = 0.001 and P < 0.001, respectively) were independent predictors for both outcome variables. In addition, the presence of amniotic fluid sludge was an independent predictor for PTD at < 28 weeks (P = 0.048), while an a priori risk for PTD was an independent predictor for PTD at < 32 weeks (P = 0.016) (Table 7). All other possible explanatory variables had a non-significant contribution to the model.

Table 7. Multivariate logistic regression (stepwise) of gestational age (GA) at diagnosis, cervical length, amniotic fluid sludge and a priori risk for preterm delivery as explanatory variables for the outcome variables of spontaneous preterm delivery at < 28 and < 32 weeks of gestation among asymptomatic patients diagnosed with a short cervix (≤ 15 mm) between 14 and 24 weeks of gestation
 Outcome variable
 Delivery < 28 weeksDelivery < 32 weeks
Explanatory variableOdds ratio (95% CI)POdds ratio (95% CI)P
GA at diagnosis of short cervix (weeks)0.7 (0.54–0.91)0.0070.69 (0.51–0.93)0.016
Cervical length (mm)0.86 (0.78–0.94)0.0010.80 (0.71–0.9)< 0.001
Presence of amniotic fluid sludge2.8 (1.01–7.77)0.0482.09 (0.77–5.69)0.15
A priori risk for preterm delivery1.32 (0.5–3.5)0.573.67 (1.28–10.54)0.016

Discussion

Principal findings of this study

There were four principal findings of this study: asymptomatic women with a sonographic short cervix (≤ 15 mm) before 20 weeks of gestation had a shorter diagnosis-to-delivery interval and a higher rate of spontaneous early preterm delivery than those diagnosed at 20–24 weeks of gestation; almost two-thirds of the women diagnosed at < 20 weeks of gestation delivered before the threshold of viability (< 24 weeks), and less than one-quarter were still pregnant at 28 weeks; gestational age at diagnosis and cervical length were independently associated with a PTD at < 28 and < 32 weeks; and amniotic fluid sludge was an independent predictor for delivery at < 28 weeks, and an a priori risk for PTD was an independent predictor for delivery at < 32 weeks of gestation.

The risk for preterm delivery varies as a function of the gestational age at diagnosis17, 50, 51. Indeed, Berghella et al.51 calculated (using logistic regression) the prevalence of PTD at < 35, < 32 and < 28 weeks according to gestational age at diagnosis (15–28 weeks) and cervical length measurement among 705 women with prior risk factors for PTD (2601 measurements). The authors concluded that the risk of PTD increases as a function of the cervical length as well as with earlier gestational age at diagnosis. Of note, there is a paucity of data51, 52 concerning an accurate estimate of the prevalence of PTD in asymptomatic women with a cervical length of ≤ 15 mm diagnosed before 24 weeks of gestation.

The present study was explicitly designed to help fill this gap in the literature and extends the important findings reported by Berghella et al.51. Our study included a relatively large number of asymptomatic women with a short cervix that was diagnosed at ≤ 24 weeks of gestation, thus allowing us to provide a comprehensive assessment and a more precise estimation of the risk of PTD in this subset of patients. Furthermore, the study population included not only pregnant women with an a priori risk for PTD, but also those without such a risk, as well as nulliparous women.

Similarly to the report by Berghella at al.51, we found that the gestational age at diagnosis of an asymptomatic short cervix and the cervical length were both important and independent predictors for the risk of PTD at < 28 and < 32 weeks of gestation (Table 7). Women with a short cervix diagnosed before 20 weeks of gestation had a significantly higher rate of PTD at < 28, < 32, < 35 and < 37 weeks, as well as a shorter diagnosis-to-delivery interval than women diagnosed at 20–24 weeks. Of note, the rate of PTD at < 28 and < 32 weeks among asymptomatic patients with a sonographic short cervix diagnosed before 20 weeks (77% and 81%, respectively) or at 20–24 weeks (31% and 48%, respectively) of gestation was higher in our study than that predicted by Berghella et al.51 (< 20 weeks: 34–48% and 27–46%, respectively; 20–24 weeks: 20–31% and 11–23%, respectively).

Importantly, we found that almost two-thirds of women diagnosed before 20 weeks delivered before 24 weeks, and that less than one-quarter were still pregnant at 28 weeks. In contrast, almost 70% of patients diagnosed with a short cervix between 20 and 24 weeks delivered after 28 weeks of gestation. This information is of clinical importance when counseling these patients regarding the prognosis of the pregnancy and management options.

Significance of an a priori risk of PTD

The current study included patients both at high risk and low risk for PTD. A history of preterm birth19 or of cervical surgery45 has been previously demonstrated to have predictive value in assessment of the risk for PTD in a patient with a short cervix. Consistent with these observations, the findings of the present study indicate that an a priori risk for PTD was associated with a shorter diagnosis-to-delivery interval and a lower gestational age at delivery than in low-risk patients. Yet, an a priori risk for PTD had a more prominent effect on the prediction of PTD for patients who remained undelivered 6–8 weeks after diagnosis (Figure 2). Moreover, logistic regression model analysis showed that an a priori risk for preterm delivery was an independent predictor for PTD at < 32 weeks, but not for an earlier preterm birth. The reason for this differential effect is not clear, but it is possible that the lack of association for an earlier PTD reflects different mechanisms of disease for patients with a short diagnosis-to-delivery interval (i.e. intra-amniotic infection) and those with a longer interval (i.e. cervical insufficiency).

Consistent with previous studies2, the diagnosis-to-delivery interval (Figure 3), gestational age at delivery, as well as the rate of PTD before 28 and 32 weeks of gestation, did not differ significantly between nulliparous and parous women (data not shown).

Figure 3.

Kaplan–Meier survival curves of nulliparous (equation image) and parous (equation image) women with a sonographic short cervix diagnosed at ≤ 24 weeks of gestation. Five patients were censored because they were lost to follow-up, two because of preterm induction of labor and 35 because they reached 37 completed weeks of gestation. The two survival curves did not differ significantly (log rank, P = 0.3).

Significance of amniotic fluid sludge

Amniotic fluid sludge43, 44, 53–55 occurs in about 1% of uncomplicated pregnancies43. Recently, our group reported a prevalence of amniotic fluid sludge of 22.6% among patients with preterm labor, an occurrence associated with MIAC, impending PTD and histologic chorioamnionitis43. Furthermore, the presence of amniotic fluid sludge in the mid-trimester in asymptomatic patients at risk for PTD is an independent risk factor for PTD at < 28, < 32 and < 35 weeks of gestation, PPROM, MIAC and histologic chorioamnionitis44.

In the present study, the prevalence of amniotic fluid sludge was almost twice as high among patients diagnosed at < 20 weeks (92%) than among those diagnosed later in pregnancy (48%). In agreement with the abovementioned observations, women with an asymptomatic short cervix and amniotic fluid sludge had a shorter diagnosis-to-delivery interval and a higher rate of PTD than those without such findings. Moreover, amniotic fluid sludge was independently associated with the diagnosis-to-delivery interval and the rate of PTD at < 28 weeks, but not at < 32 weeks.

Why do women with a short cervix diagnosed before 20 weeks have a shorter diagnosis-to-delivery interval than women diagnosed later in pregnancy?

The observational nature of the present study does not allow inferring a causal relationship between the timing of diagnosis of short cervix and the interval time to delivery. Nevertheless, several explanations can be suggested. First, multiple etiologies can account for the presence of an asymptomatic short cervix56, including intra-amniotic infection/inflammation34, 39, 41, 57, congenital cervical anomalies58–61, cervical surgery27, 45, 62–69, cervical insufficiency10, 70–72 and others56. It is conceivable that the etiologic factors leading to a sonographic short cervix early in the mid-trimester (i.e. intra-amniotic infection/inflammation) differ from those leading to the same finding later in pregnancy (i.e. cervical insufficiency). This view is supported by the study of Guzman et al.14, in which 89 women at risk for pregnancy loss and PTD were followed longitudinally by TVUS. The authors reported that women with a cervical length of ≤ 20 mm diagnosed between 15 and 19 weeks of gestation appeared to quickly enter a linear phase of shortening of the cervix, whereas those diagnosed between 20 and 24 weeks showed a period of cervical length stability before entering the shortening phase14. Furthermore, among patients with preterm labor or PPROM, the lower the gestational age at presentation, the higher the risk found of intra-amniotic infection/inflammation73–75. Second, the rate of amniotic fluid sludge was higher in women diagnosed with a short cervix before 20 weeks of gestation than those diagnosed later (20–24 weeks). Amniotic fluid sludge has been associated with intra-amniotic infection/inflammation in the form of biofilms55, which may prevent detection of infection using routine culture techniques. It is possible that by using more sensitive molecular, non-culture-based techniques76, 77 identification of MIAC will be possible even in cases of amniotic fluid sludge in the form of biofilms. Third, a short cervix may be a risk factor for ascending infection78, 79. It is possible that the inherent defense mechanisms (i.e. antimicrobial peptides80–82) are not yet fully developed in early pregnancy; thus, a similar infectious insult can lead to one consequence early and to another later in pregnancy.

Strengths and limitations

The main strength of the study was its large sample size, as the prevalence of a cervical length of ≤ 15 mm before 24 weeks of gestation is relatively low. Moreover, we included only asymptomatic patients without advanced cervical dilatation or cervical cerclage. The limitations of the study stem from its observational nature. Management of patients was at the discretion of the attending physicians. Another limitation was that the exact gestational age at which the cervix became shorter than ≤ 15 mm can only be determined by performing very frequent ultrasound examinations. The consequence of this limitation is that patients who already had a short cervix before 20 weeks of gestation may have been diagnosed at 20–24 weeks and be allocated to the latter group. However, such potential bias should, in fact, mitigate the significant differences between the groups (type 2 error).

In conclusion, our findings indicate that asymptomatic women with a short cervix, diagnosed before 20 weeks of gestation, have a shorter diagnosis-to-delivery interval and a higher rate of PTD than those diagnosed between 20 and 24 weeks. Furthermore, almost two-thirds of patients diagnosed before 20 weeks will deliver before the threshold of viability. We propose that different etiological factors may account for the disparity between the two subsets of patients. Further studies are warranted in order to gain a better understanding of the underlying mechanism responsible for the observations presented herein.

Acknowledgements

This research was supported by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development, NIH, DHHS. The authors wish to acknowledge the contributions of the Registered Diagnostic Medical Sonographers staff and the Research Nurses of the Perinatology Research Branch.

Ancillary