Intra-amniotic infection, diagnosed by microbial invasion of the amniotic cavity (MIAC) and/or the presence of intra-amniotic inflammation (IAI), is related to adverse perinatal outcome in women with preterm labor. Due to the subclinical nature of IAI, a correct diagnosis depends on amniocentesis, which is an invasive method not performed as a clinical routine. The aim of this study was to evaluate if cervical length measured by transvaginal sonography could assist in the identification of women at high risk for IAI.
Cervical length was assessed by transvaginal sonography in 87 women with singleton pregnancies in preterm labor (< 34 weeks of gestation). Cervical (n = 87) and amniotic (n = 55) fluids were collected. Polymerase chain reactions for Ureaplasma urealyticum and Mycoplasma hominis, and culture for aerobic and anaerobic bacteria, were performed. Interleukin (IL)-6 and IL-8 were analyzed by enzyme-linked immunosorbent assay.
IAI was present in 25/55 (45%) of the patients presenting with preterm labor who underwent amniocentesis. Women with IAI had a significantly shorter cervical length (median, 10 (range, 0–34) mm) than had those without IAI (median, 21 (range, 11–43) mm) (P < 0.0001). Receiver–operating characteristics curve analysis showed that a cervical length (cut-off of 15 mm) predicted IAI (relative risk, 3.6; CI, 1.9–10.0) with a sensitivity of 72%, specificity of 83%, positive predictive value of 78% and negative predictive value of 78%. Cervical length was also significantly associated with preterm birth up to 7 days from sampling and at ≤ 34 weeks.
Preterm labor (PTL) is a frequent cause of admission to hospital during pregnancy, but only few of the patients 8–38%1–3 admitted to hospital with suspicion of PTL do deliver preterm. There are so far no reliable clinical tools for discrimination between harmless early contractions and true preterm labor.
It has been shown that intra-amniotic infection, defined as microbial invasion of the amniotic cavity (MIAC), is present in 12.8%4–6 of all women with PTL. Microbes are, however, difficult to detect, and as many as 45% of cases in PTL present with intra-amniotic inflammation (IAI) (defined as high interleukin (IL)-6 and/or IL-8 levels)7. Furthermore, IAI is strongly correlated with perinatal morbidity, periventricular brain white matter lesions and cerebral palsy8. Therefore, it seems important in the management of patients in PTL to identify not only patients at risk of preterm delivery but also those at risk of IAI and MIAC.
Transvaginal sonography (TVS) has been used for many years to observe the cervix in women in PTL and it is considered to be a method that offers a reliable and simple way of assessing objectively the cervix, with high acceptance by women9–13. Other studies have shown an increased number of positive high vaginal cultures as the cervices shortened14, which suggests an association between a short cervix and intrauterine infection. The suggested mechanism is that microbial invasion into the choriodecidual interface predisposes the woman to the release of inflammatory stimuli and cervical shortening, but it could also be that an already short cervix predisposes the woman to an ascending intrauterine infection. Several studies have shown that the risk of spontaneous preterm birth (SPTB) is inversely correlated with cervical length1–3, 15–21, and a recently published study showed a relationship between MIAC and a short cervix22. There are, however, no previously published studies on the relationship between sonographic cervical length and IAI, although we and others have shown a cervical cytokine response associated with MIAC, IAI and PTB23–27.
The aim of this study was to determine the relationship between sonographic cervical length and IAI, MIAC and SPTB up to 7 days from sampling or at ≤ 34 weeks of gestation.
This was a prospective cohort study. Caucasian women with singleton pregnancies in PTL who presented at Sahlgrenska University Hospital during 1996–2001 at a gestational age of 22 + 0 weeks to 33 + 6 weeks were invited to enrol. PTL was defined as regular uterine contractions (at least two uterine contractions every 10 min for ≥ 30 min, confirmed by external tocometry) in combination with one of three cervical changes documented by digital examination (1: ≤ 2 cm length + ≥ 1 cm dilatation; 2: ≤ 2 cm length + cervical softening; 3: ≥ 1 cm dilatation + cervical softening) and/or cervical length < 30 mm measured by TVS. All women had intact membranes at enrolment. The study was approved by the local ethics committee at the University in Göteborg and all patients gave informed consent before enrolment.
Cervical ripening was assessed by digital examination. All scans were performed in a standardized way using a Siemens Sonoline SI 450 (Siemens Medical Solutions, Erlangen, Germany) ultrasound machine, equipped with a 5.0-MHz, 115/150 vaginal probe. The cervical length was measured with the woman in the dorsal lithotomy position after she had emptied her bladder. When the cervical canal was visualized, the probe was withdrawn to avoid pressure and distortion and elongation of the cervix. A sagittal view showing the entire cervix (endocervix and vaginal cervix), with the echogenic endocervical mucosa along the length of the cervical canal, was obtained. The calipers were used to measure the distance between the notches made by the junction of the anterior and posterior cervical walls at the internal and external os. Three measurements were performed (in millimeters) and the shortest distance was noted.
Women with preterm prelabor rupture of membranes, known uterine malformations, fetal malformations, significant vaginal bleeding, imminent delivery, cervical cerclage or fetal distress were excluded.
Gestational age was determined by routine ultrasound in the second trimester (at 16–19 weeks of gestation) in all women with the exception of three who had the gestational age determined by the date of their last menstrual period. Tocolytic therapy (intravenous terbutaline and/or indomethacin, the latter if the gestational age was < 28 weeks) was administered according to the local department protocol. Corticosteroids were administered to encourage fetal lung maturity after assessment in 74% (64/87) of the women.
Cervical mucus was obtained within 12 h of admission with a Cytobrush (Cytobrush Plus GT, Medscan Medical AB, Malmö, Sweden) from the external cervical os in all patients. The cervical fluid was weighed and kept in a refrigerator (+ 4 °C) until it was processed, within 5 h of sampling. The Cytobrush with the mucus was submerged in 1.0 mL sodium chloride (NaCl) and shaken for 30 min at + 4 °C, followed by centrifugation at 855 g at + 4 °C for 10 min and storage at − 80 °C until analysis. These samples were analyzed for cytokines and Ureaplasma urealyticum and Mycoplasma hominis (by polymerase chain reaction (PCR)). An additional sample of cervical mucus collected by a cotton-tipped swab was placed immediately in Ames-modified Stuart medium and sent to the microbiological laboratory for culturing. Culturing was performed using various agar plates for anaerobes, gonococci, streptococci, staphylococci, Gram-negative bacteria and yeast. Incubation was performed under aerobic, anaerobic and carbon dioxide (CO2) conditions. Sabouraud plates were cultured at 30 °C and all other plates at 37 °C. In addition, bacterial isolates were gram-stained. Non-Lactobacillus-dominated microbiota was recorded as well as the presence of specific bacteria such as Group B Streptococcus, streptococci, Escherichia coli, enterococci, coagulase-negative staphylococci, Staphylococcus aureus, Gardnerella vaginalis, Prevotella species and Candida albicans.
Ultrasound-guided transabdominal amniocentesis was performed in 55 women under antiseptic conditions within 12 h after admittance. A 0.7-mm diameter needle was used and 30–50 mL amniotic fluid was aspirated. The aspirate was placed immediately in a refrigerator (+ 4 °C) and centrifuged within 5 h of sampling at 855 g at + 4 °C for 10 min. The supernatant was stored at − 80 °C until analysis.
A sample of uncentrifuged amniotic fluid was transported immediately to the microbiological laboratory for PCR analysis of U. urealyticum and M. hominis and for aerobic and anaerobic culture. MIAC was defined as positive PCR and/or growth of any bacteria in the amniotic fluid, except for coagulase-negative Staphylococcus, which was considered to be a skin contamination. However, coagulase-negative Staphylococcus in amniotic fluid from patients with IAI (high levels of IL-6 and/or IL-8) was considered to be MIAC, as described in our previous study7. The bacterial isolates obtained by culturing of the amniotic fluid were also characterized and described in the same study.
IAI for women in PTL was defined in our previous study as ≥ 1.5 ng/mL IL-6 and/or ≥ 1.3 ng/mL IL-8 in the amniotic fluid7. The levels of IL-6 and IL–8 in cervical and amniotic fluid were analyzed by enzyme-linked immunosorbent assay (ELISA). Commercially available paired antibodies and standards (from R&D Systems, Minneapolis, MN, USA) were used in a sandwich ELISA set-up. The samples of cervical and amniotic fluid were diluted 1 : 5, 1 : 20 and 1 : 100 and run in duplicate. The inter-assay and intra-assay variation was calculated at < 10% based on analysis of several samples on three separate occasions. The coefficient of variation for amniotic fluid (29–76%) was higher for low values (≤ 700 pg/mL). The coefficient of variation for the cervical fluid was 34% for IL-6 and 28% for IL-8. The ELISA detection limit was 30 pg/mL for both IL-6 and IL-8, but because the samples were run at a 1 : 5 dilution, the lower limit of detection was 150 pg/mL for both IL-6 and IL-8.
Two investigators (R.M.H., B.J.) scrutinized the medical records and entered maternal and perinatal data into a database. We extracted from the medical records epidemiological data including maternal demographic characteristics (age, profession, marital status, height, weight), reproductive history (parity, abortions, previous preterm gestations, previous genital and sexually transmitted diseases) and medical complications (urinary tract infections, cigarette smoking, alcohol use, and treatment with antibiotics).
Clinical chorioamnionitis was defined, according to the criteria proposed by Gibbs et al.28, as fever ≥ 37.8 °C on two occasions at least 4 h apart and if two or more of the following criteria were present: uterine tenderness, malodorous vaginal discharge, fetal tachycardia (> 160 bpm), maternal tachycardia (> 100 bpm) and maternal leukocytosis (> 15 000 cells/mm3).
Sensitivity, specificity and positive and negative predictive values were calculated for different cervical lengths in relation to SPTB (up to 7 days from sampling and at ≤ 34 weeks), IAI, MIAC, and cervical and amniotic cytokines. Receiver–operating characteristics (ROC) curves were produced to establish the best cut-off levels for cervical length in relation to delivery within 7 days, IAI and MIAC. ROC curves were also constructed for cervical IL-6 and IL-8 in relation to IAI.
Calculations were made using the computer programs StatView 5.01 (SAS Institute Inc, Cary, NC, USA), InStat 2.01 (Graph Pad Software, San Diego, CA, USA) and SPSS 13.0 (SPSS Inc, Chicago, IL, USA). Continuous variables were analyzed with Mann–Whitney U-test and proportions with Fisher's exact test. Spearman's rank correlation test was used for analysis of correlation between continuous variables. Kaplan–Meier survival curves were produced. A P-value of < 0.05 or a CI not including 1 was considered statistically significant. Logistic regression was used to investigate the independent association between cervical length and IAI. Variables entered in the regression analysis included gestational age at admission, maternal age, parity, and levels of cervical IL-6 and IL-8.
A total of 87 Swedish women who presented with threatening PTL and intact membranes met the inclusion criteria. All patients were offered amniocentesis, and this was performed successfully in 55 of the 87 patients, with complete cultures/PCR and analysis of cytokines. In five patients the amount of amniotic fluid was insufficient for analysis of cytokines and in one patient the procedure was interrupted because the woman complained of pain. The remaining 26 patients declined to undergo amniocentesis. The clinical characteristics of the women are summarized in Table 1.
Table 1. Clinical characteristics of the study population
Median (range) or % (n)
n = 83 due to missing values from four women.
Latency is the time elapsed from amniocentesis, measurement of cervical length and culture of cervical and amniotic fluid, to delivery. IAI, intra-amniotic inflammation; MIAC, microbial invasion of the amniotic cavity.
The median cervical length in women delivering within 7 days of sampling (10; range, 0–29.0) mm was significantly shorter than was that in those delivering after 7 days (21.5; range, 0–43) mm (P < 0.0001) (Figure 1a).
45% (25/55) of the women who underwent amniocentesis had IAI (defined as IL-6 ≥ 1.5 ng/mL and/or IL-8 ≥ 1.3 ng/mL7). Women with IAI had a significantly shorter cervical length (median, 10 (range, 0–34) mm) than had those without IAI (median, 21 (range, 11–43) mm) (P < 0.0001) (Figure 1b).
Kaplan–Meier survival analysis was performed to assess the examination-to-delivery interval in relation to cervical length as measured by TVS. Women with a cervical length of ≤ 15 mm had a significantly shorter interval from admission to delivery than did patients with cervical length > 15 mm (P = 0.02 log rank) (Figure 2).
ROC curve analysis showed that the curve constructed for cervical length in relation to IAI was above the 45° angle, indicating a significant relationship (area under the curve = 0.83, P < 0.001) (Figure 3). Cervical length ≤ 15 mm was still significantly associated with IAI after adjustment for gestational age at admission, maternal age, parity, and levels of cervical IL-6 and IL-8. The optimal cut-off value for cervical length was 15 mm for prediction of IAI with a sensitivity and specificity of 72% and 83%, respectively, and positive and negative predictive values of 78% and 78%, respectively (Table 2), meaning that 78% (18/23) of the women with a cervical length of ≤ 15 mm had IAI compared with 22% (7/32) of the women with a cervical length of > 15 mm (P < 0.0001). 83% (15/18) of the women who had both IAI and cervical length of ≤ 15 mm delivered within 7 days of assessment. For comparison, ROC-curve analyses were also performed for cervical IL-6 and IL-8 in relation to IAI: a cervical level of IL-6 of 1.7 ng/mL and a cervical level of IL-8 of 6.7 ng/mL were the best cut-off levels for predicting IAI.
Table 2. Diagnostic indices for cervical length (≤ 15 mm) and levels of cervical IL-6 (≥ 1.7 ng/mL) and cervical IL-8 (≥ 6.7 ng/mL) and their ability to predict intra-amniotic inflammation
Cervical length ≤ 15 mm
Cervical IL-6 ≥ 1.7 ng/mL
Cervical IL-8 ≥ 6.7 ng/mL
IL, interleukin; NPV, negative predictive value; PPV, positive predictive value.
Sensitivity (% (n))
Specificity (% (n))
PPV (% (n))
NPV (% (n))
The differences between the groups on the basis of cervical length (cut-off of 15 mm) concerning the presence of MIAC or microbes in the cervical fluid were not statistically significant (Table 3).
Table 3. Comparison of results with the study population divided according to cervical length
Cervical length ≤ 15 mm (n = 33)
Cervical length > 15 mm (n = 54)
Four values missing.
One value missing.
Latency is the time elapsed from amniocentesis, measurement of cervical length and culture of cervical and amniotic fluid, to delivery. IAI, intra-amniotic inflammation; IL, interleukin; MIAC, microbial invasion of the amniotic cavity.
The most important conclusion of this study is that examination of the length of the uterine cervix with sonography provides prognostic information about IAI and SPTB. To our knowledge, this is the first study to show a relationship between cervical length and IAI, defined as high levels of IL-6 and/or IL-8.
Our choice to present the rate of SPTB within 7 days was related to the fact that prolongation of the pregnancy by 7 days made it possible to administer glucocorticoids for fetal lung maturation, which in turn has been shown to be associated with reduced neonatal mortality and morbidity29.
Studies investigating the use of TVS examination of the uterine cervix to measure cervical length have focused mostly on its ability to predict SPTB. It has been shown to be a simple method that is safe, reproducible, predictive of SPTB9–11 and well accepted by women12, 13. Our study supports the concept that cervical length measured by TVS can be used as a tool to predict SPTB in women with PTL and intact membranes. Like other studies2, 3, 15, 18, our ROC curve established a cervical length of 15 mm to be the optimal cut-off for prediction of SPTB. The rate of SPTB within 7 days of assessment in the women with a cervical length of ≤ 15 mm was 58% (19/33) compared with the overall rate of 31%; this is in accordance with data reported by Gomez et al.22. Thus, using cervical length as a marker of SPTB improved the diagnosis of true preterm labor; however, 42% (14/33) of women with a short cervix in our study remained undelivered at 7 days and were subjected to fetal monitoring, treatment and hospitalization without obvious benefit and at considerable psychological and economic cost for the families involved.
IAI, and to a certain degree MIAC, have been shown to predict perinatal morbidity, including periventricular brain white matter lesions and cerebral palsy8, 30, 31. There is increasing awareness that subclinical infection plays a role in the etiology of PTL, although it is certainly not the only pathway that leads to preterm delivery and it is probably important to separate this group from other etiological groups. The group with IAI represents a risk group that is more likely to benefit from antibiotic/anti-inflammatory treatment, and in which tocolytic therapy presumably should be avoided, although these questions need to be addressed in future randomized trials. So far, the optimal method with which to determine whether there is infection/inflammation in the amniotic cavity is by amniocentesis and analysis of the amniotic fluid. We found that cervical length was a good predictor of IAI, and cases presenting with a cervical length of ≤ 15 mm had a 78% risk of IAI. This result is in agreement with the study of Gomez et al.22 that included 401 women with intact membranes in spontaneous PTL. It was found that the shorter the cervical length and the lower the gestational age, the greater the risk of MIAC. We found merely a trend of an association between cervical length and MIAC, which is most likely explained by the limited number (n = 55) of patients in our study.
In our previous study23, we found a significant relationship with the cervical cytokines IL-6/IL-8 and MIAC, and in this study we found a significant relationship between the level of cervical IL-6 and a short cervix (≤ 15 mm). The levels of the chemokine IL-8 did not differ between the group of patients with cervical length ≤ 15 mm and the group of women who had a longer cervix (Table 3). Spearman's rank test showed, however, that cervical length was inversely associated with the level of cervical IL-8. These relationships between the cytokines and cervical length could be an indirect way of showing that there also existed an association between cervical length and MIAC in our population of women in PTL and with intact membranes, confirming the results of Gomez et al.
The means of identifying women in PTL in whom a sub-clinical infection/inflammation is the cause of their contractions is still imprecise. This study has shown that measurement of cervical length by TVS can assist in the identification of women with a high risk of IAI.
We acknowledge the assistance of colleagues at the clinic in helping in the recruitment of the study patients and the technical assistance of the midwife Ellen Samuelsson. The study was supported by the Swedish Medical Research Council (09455), The Göteborg Medical Society, The Frimurare Barnhus Foundation, and by Swedish government grants to researchers in the public health service (ALFGBG 2863).