SEARCH

SEARCH BY CITATION

Keywords:

  • amnion;
  • chorion;
  • fetal membrane thickness;
  • preterm delivery;
  • ultrasound

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Objective

To evaluate whether measurement of the thickness of the fetal membranes by high-resolution ultrasound is a useful marker to predict preterm delivery.

Methods

One hundred and fifty-eight women with singleton pregnancies at 18–35 gestational weeks were enrolled consecutively at our referral center for obstetric care and the thickness of their fetal membranes was measured using high-resolution ultrasound equipment. Data were analyzed to determine whether there were significant differences between those delivering at term and those delivering preterm. Receiver–operating characteristics (ROC) curves were used to determine the best cut-off point of membrane thickness for predicting preterm birth.

Results

Women who delivered preterm had greater fetal membrane thickness than did those who delivered at term (1.67 ± 0.27 mm vs. 1.14 ± 0.30 mm, P < 0.0001). For the best cut-off indicated by ROC curve analysis (1.2 mm), the sensitivity and specificity for predicting preterm birth were 100% (95% CI, 80.3–100) and 69.5% (95% CI, 61.2–77.0), respectively, and positive and negative likelihood ratios were 3.3 and 0.0, respectively.

Conclusion

Sonographic measurement of fetal membrane thickness could be helpful in the prediction of preterm delivery. Copyright © 2008 ISUOG. Published by John Wiley & Sons, Ltd.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Preterm labor is considered a syndrome with multiple etiologies resulting in the activation of the common terminal pathway of parturition (membrane activation, myometrial contractility and cervical ripening)1. Although a number of different pathogenic mechanisms (infection, hemorrhage, stress) may initiate preterm labor, all appear to involve disruption of the chorionic decidual interface2–5. Several biochemical markers that are released by the disruption of this interface (fetal fibronectin, plasma matrix metalloproteinase-9) and biophysical markers (e.g. cervical length) have been proposed as predictors of spontaneous preterm birth6–16.

Inflammation of chorionic and amniotic membranes has been documented at histological examination after delivery in about one third of preterm deliveries; the extent and severity of the villous edema have been shown to have a strong positive correlation with neonatal morbidity and mortality17. Prenatal identification of membrane inflammation is thus desirable, and biophysical and biochemical markers of inflammation are presently under investigation to identify patients at risk of preterm delivery. Because the sensitivity of each single marker in predicting preterm birth is only moderate, combinations of different markers and multiple-marker tests are more likely to provide better diagnostic ability.

The aim of this study was to determine if measurement of the thickness of the amniochorionic membranes by high-resolution ultrasound could be a useful marker to predict preterm delivery, with a view to ultimately using it in combination with biochemical markers.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

We evaluated 158 singleton pregnant women admitted consecutively to Siena University Hospital, at the Department of Pediatrics, Obstetrics and Reproductive Medicine, a referral center for high-risk pregnancies, between February and November 2005. We excluded women whose pregnancy was complicated by cerclage, placenta previa or major fetal anomalies, those with a gestational age at the time of ultrasound examination of < 18 weeks or > 35 weeks, and those who delivered by Cesarean section. The local ethics committee approved the study, and informed consent was obtained from all patients prior to enrolment.

In this study, each woman underwent one ultrasound evaluation between 18 and 35 weeks of gestation, during which the thickness of the fetal membranes was measured and fetal biometric and biophysical parameters were assessed. Ultrasound examinations were performed by two different operators using a MyLab50 Esaote (MyLab Family group Esaote SpA, Genova, Italy) ultrasound machine, equipped with a 2.5–6.6-MHz transabdominal convex multifrequency probe. For measurement of the thickness of the fetal membranes, the transducer was positioned perpendicular to the maternal abdomen. The region of interest was magnified (to occupy > 75% of the screen) and the measurement was taken at about 3 cm from the umbilical cord insertion (Figure 1), positioning the lower border of the horizontal line of the upper caliper on the line that defines the external margin of the chorion and the upper border of the horizontal line of the lower caliper on the line that defines the external margin of the amnion (Figure 2).

thumbnail image

Figure 1. Ultrasound image showing where the measurement of the fetal membrane thickness was taken. To better identify the umbilical cord insertion it is useful to superimpose the color Doppler box.

Download figure to PowerPoint

thumbnail image

Figure 2. After magnification of the region of interest to better identify amnion and chorion, the measurement was taken by positioning the calipers on their external margins.

Download figure to PowerPoint

Gestational age was assessed by last menstrual period and confirmed by first-trimester ultrasound. Preterm delivery was defined as occurring between 25 + 0 and 36 + 6 weeks of gestation. Labor was diagnosed when women experienced regular and painful uterine contractions at least twice every 10 min for at least 1 h, and in the presence of cervical effacement or dilatation.

Statistical analysis

Data were expressed as mean ± SD. To determine whether there were statistically significant differences between those delivering at term and those delivering preterm, we used the Mann–Whitney U-test, two-tailed Fisher's exact test and Spearman's correlation coefficient test. Statistical significance was assumed whenever P < 0.05. Cut-off points for identifying the best value of membrane thickness for prediction of preterm birth were chosen by receiver–operating characteristics (ROC) curve analysis. For the best cut-off indicated by ROC curve analysis, we calculated specificity, sensitivity, positive and negative predictive values with their respective 95% CIs, likelihood ratios and the area under the curve.

Intra- and interobserver reproducibility of ultrasound measurement of fetal membrane thickness was tested in 15 consecutive normal pregnancies. Three consecutive measurements of the thickness of the fetal membranes were taken by each operator for each woman. Intraobserver variability was expressed as the difference between the highest and the lowest measurements obtained by one operator. Interobserver variability was expressed as the difference between the means of the three measurements of each observer.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Clinical findings

Of the 158 enrolled patients, 59.5% (n = 94) were nulliparous. The mean ± SD maternal age was 32.9 ± 4.9 years, gestational age at the time of sonographic measurement of the thickness of the membranes was 27.7 ± 5.7 weeks and gestational age at delivery was 38.1 ± 3.1 weeks; 17/158 (10.8%) delivered preterm (< 37 weeks), including 10 (6.3%) before 33 weeks. The interval between measurement of fetal membrane thickness and delivery did not exceed 10 weeks (70 days). Table 1 summarizes the clinical characteristics of the study population in relation to time of delivery.

Table 1. Clinical and obstetric characteristics of the study population according to term (n = 141) or preterm (PTD, n = 17) delivery
CharacteristicTerm delivery (≥ 37 weeks)Preterm delivery (25–36 weeks)P
  • Values are mean ± SD or n (%).

  • *

    When fetal membrane thickness was measured.

  • t-test. GA, gestational age. NS, not significant.

Maternal age (years)32.9 ± 4.832.6 ± 6.1NS
Nulliparous83 (58.9)11 (64.7)NS
History of spontaneous PTD4 (4.28)1 (5.9)NS
GA at ultrasound (weeks)*27.5 ± 3.829.1 ± 3.7NS
GA at delivery (weeks)38.9 ± 1.430.9 ± 4.0< 0.0001
Birth weight (g)3238 ± 4731691 ± 810< 0.0001
Male fetal gender71 (50.4)10 (58.8)NS

Fetal membrane thickness

There was no correlation between fetal membrane thickness and gestational age at ultrasound examination in the whole group of 158 (P = 0.107, r = 0.128), in the group who delivered at term (P = 0.213, r = 0.106) or in those delivering preterm (P = 0.345, r = − 0.243).

Women who delivered preterm had a greater membrane thickness than did those who delivered at term (1.67 ± 0.27 vs. 1.14 ± 0.30 mm, P < 0.0001) (Figure 3). There was a significant and inverse correlation between membrane thickness and gestational age at delivery (P = 0.0001; r = − 0.302) (Figure 4a). Moreover, there was a statistically significant and inverse correlation between fetal membrane thickness and time elapsed between this measurement and delivery (P < 0.0001; r = − 0.306).

thumbnail image

Figure 3. Scatterplot showing mean (horizontal line) membrane thickness in term (▪) and preterm (▴) deliveries. The difference was statistically significant (P < 0.0001).

Download figure to PowerPoint

thumbnail image

Figure 4. Scatterplots showing correlations between fetal membrane thickness and gestational age at delivery (a) (P = 0.0001, r = − 0.302) and ultrasound examination–delivery time interval (b) (P < 0.0001, r = − 0.306). Some data points include two or more cases.

Download figure to PowerPoint

For the best cut-off indicated by ROC curve analysis (fetal membrane thickness of 1.2 mm), the sensitivity and specificity were 100% (95% CI, 80.3–100%) and 69.5% (95% CI, 61.2–77%), respectively, with positive and negative likelihood ratios of 3.3 and 0.0, respectively (Figure 5). The area under the ROC curve was 0.897.

thumbnail image

Figure 5. Receiver–operating characteristics curve for fetal membrane thickness and risk of preterm delivery in the whole study population (n = 158). The area under the curve was 0.897.

Download figure to PowerPoint

Inter- and intraobserver reproducibility were good: interobserver variability was 6.7% and intraobserver variability was 6.5%.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

One of the main goals of modern obstetric care is the identification of new strategies to improve the outcome of preterm infants. The identification of biochemical and biophysical markers associated with preterm labor is considered important in order to identify women at high risk for preterm delivery, allowing targeted intervention. In this study we have shown for the first time that women who later undergo preterm delivery have thicker fetal membranes than do those who go on to deliver at term, perhaps suggesting that local events operate in determining the length of gestation. It is possible that premature activation of mechanisms involved in parturition occur in women who deliver preterm and that this may be associated with abnormal thickness of fetal membranes. In fact, biochemical and biomolecular studies support the hypothesis that labor is an inflammation-like condition1, 18, which might lead us to expect signs of inflammation. Regarding the use of ultrasound to detect such signs, it is well known that in in-vivo models, the thickness of certain internal membranes has been measured sonographically to assess and/or monitor the presence of human pathologies related directly to inflammation; for example, measurement of leptomeninges (pia-glial plate) thickness in the case of meningitis19 and measurement of vascular endothelial (intima–media) thickness in vasculitis20. Such evidence together with our findings support the hypothesis that thickening of fetal membranes is a sign of inflammation related to the biochemical events of parturition.

With respect to our methodology, it must be borne in mind that measurement of membrane thickness in a location near the cervix rather than at other locations could be more representative of an ascending infectious process and that such a measurement would ideally be done by transvaginal sonography, allowing positioning of the transducer directly on the cervix, with higher ultrasound frequencies and better definition. However, there are several points that must be taken into consideration: first, in our experience, the transvaginal measurement of fetal membranes has high interobserver variability due to the difference in compression of the endocavitary transducer on the cervix, or due to the presence of uterine contractions or funneling. In addition, it is not always possible by transvaginal sonography to identify fetal membranes near the uterine cervix or the internal cervical os and their measurement can be affected by the position of the placenta in cases of placenta previa or low-lying (both anterior or posterior) placenta. Another consideration is that the measurement of fetal membrane thickness in locations outside the chorionic plate (i.e. where membranes are in strict contact with the maternal decidua) becomes very difficult due to the lower contrast existing between the fetal membranes and the maternal decidua, thus further increasing the interobserver variability.

Our study has also shown that prematurely thickened membranes precede preterm delivery by up to 70 days, allowing early identification of women at risk. The sensitivity of this approach was high (100%) and the specificity was good (69.5%): women with thickened membranes have a high risk (likelihood ratio, 3.3) of delivering preterm and the risk increases as membrane thickness increases. When the sonographic thickness was below the cut-off (1.2 mm) identified by ROC curve analysis, premature delivery did not occur, even when women were complaining of contractions (LR 0.0). In the group of patients presenting other risk factors for preterm delivery this approach could be useful to better identify those who need to be monitored intensively and, eventually, hospitalized.

Future prospective studies are needed to evaluate the performance of the cut-off identified by our ROC curve analysis for the prediction of preterm delivery, and to establish whether this non-invasive approach to studying amniochorionic membranes will be useful, alone or in combination with other (biochemical and/or biophysical) markers, in the early prediction of preterm delivery.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References