Objective
The objective of the study was to test whether ultrasound-measured fetal adrenal gland volume (AGV) and fetal zone enlargement (FZE) predicts preterm birth (PTB) better than cervical length (CL).
Study Design
Three-dimensional and 2-dimensional ultrasound were used prospectively to measure fetal AGV, FZE, and CL in women with preterm labor symptoms. We corrected AGV for fetal weight (cAGV). The ratio between whole gland depth (D) and central fetal zone depth (d) (d/D) was used to measure FZE. Ability of cAGV, d/D, and CL to predict PTB 7 days or less was compared.
Results
Twenty-seven of 74 women (36.5%) presenting between 21 and 34 weeks had PTB of 7 days or less. FZE greater than 49.5% was the single best predictor for PTB (sensitivity/specificity 100%/89%) compared with cAGV (81%/87%) and CL (56%/60%; P < .05). Prediction was independent of obstetrics history and tocolytic use.
Conclusion
The 2-dimensional measurement of the adrenal gland FZE is highly effective performing superior to CL in identifying women at risk for PTB within 7 days.
Preterm birth (PTB) remains a major cause of perinatal morbidity and mortality worldwide. Compared with term deliveries, early PTB (<34 weeks’ gestation) carries a 7-fold increased risk of neonatal death. Following PTB, survivors can experience significant long-term cognitive, behavioral, emotional, sensory, and motor deficits. Thus, identification of women at risk for PTB is one of the critical prerequisites for effective intervention and improvement in outcome. For the last several decades, significant effort has been focused toward discovery of an accurate method to predict PTB. Traditional predictors such as obstetric risk factors and clinical presentation are helpful but seldom completely define the population that will truly deliver preterm.
For Editors’ Commentary, see Table of Contents
Convincing data have shown that 2-dimensional (2D) ultrasound measurement of cervical length (CL) can identify women at risk for PTB. Accordingly, CL is now widely used in clinical practice for risk estimation. However, as understanding of the mechanisms of preterm labor (PTL) have evolved, obstetricians have learned that, in some women, cervical shortening is a phenomenon that carries no increased risk for prematurity. Therefore, the search for early and accurate markers that distinguish between physiologic processes and abnormal activation of the labor cascade has been ongoing.
There is evidence to support the view that activation of the fetal hypothalamic-pituitary-adrenal axis, and the cross talk between a variety of placental and fetal adrenal gland endocrine signaling pathways play an important role in initiation of the normal parturition process. Biochemical activation causes increased dehydroepinadrosterone-sulfate production in the central zone of the fetal adrenal gland (fetal zone). Accordingly the whole fetal adrenal gland increases in size and this increase is predominantly due to significant enlargement of the central fetal zone.
Arguments in support of this process have been provided by an autopsy study, which demonstrated that neonates that delivered in the setting of idiopathic PTB had significantly higher adrenal gland weight than those that delivered secondary to fetal/maternal hemorrhage. In a prior study, we were able to show that the 3-dimensional (3D) ultrasound measurement of fetal adrenal gland volume (AGV) allows prenatal identification of this process. In that study the increased volume of the whole fetal adrenal gland significantly correlated with the risk for PTB.
However, whereas 3D ultrasound studies of the AGV may provide valuable insight in studying the pathophysiology of fetal adaptation to intrauterine stressors and fetal contribution to the preterm labor process, they are technically challenging and have not achieved generalized clinical application. In addition, this method evaluates the whole gland rather than the physiologically relevant fetal zone. In addition, a comparative analysis of the predictive value of AGC vs CL has not yet been performed. Herein we sought to test the hypothesis that ultrasound evaluation of fetal adrenal gland volume and the central fetal zone size are predictive of PTB better than cervical length measurement.
Materials and Methods
Study design
This was a prospective observational study performed at the University of Maryland School of Medicine and Yale University School of Medicine from 2005 to 2009. Consecutive patients with singleton pregnancies between 21 and 34 weeks that presented with signs of preterm labor or preterm premature rupture of membranes (PPROM) were recruited for the study. Exclusion criteria included suspected fetal growth restriction (sonographically estimated fetal weight <10th percentile), maternal medical conditions (eg, hypertension, preeclampsia, diabetes, or thyroid or adrenal diseases), and presence of fetal heart rate abnormalities at enrollment (ie, bradycardia or prolonged variable decelerations). This study protocol was approved by the University of Maryland Institutional Review Board and the Yale University Human Investigation Committee. Written informed consent was obtained from all participants prior to enrolment. The 3D data and outcomes of 32 patients were previously reported.
Gestational age (GA) was established based on the last menstrual period and/or an ultrasound evaluation prior to 20 weeks. Preterm labor was defined as the presence of regular uterine contractions with cervical effacement and/or advancing cervical dilatation. We confirmed PPROM by visualization of amniotic fluid vaginal pooling at the time of the sterile speculum examination. Positive nitrazine and ferning tests were also considered diagnostic.
Ultrasound fetal biometry, the CL measurement of 3D AGV acquisitions were performed for each patient at the time of admission. Clinical management was at the discretion of the obstetric providers. Patients with PPROM were managed expectantly in the absence of signs or symptoms of clinical chorioamnionitis (fever >38.0°C, abdominal tenderness, fetal tachycardia) and/or abnormalities of fetal heart rate (variable or late decelerations) and/or placental abruption.
Following admission, corticosteroid and antibiotic administrations were recommended in accordance with American Congress of Obstetrics and Gynecology recommendations. For most cases (80%), acquisition of the AGV was performed prior to the administration of antenatal steroid. Per institutional protocol, patients were monitored twice daily for fetal heart abnormalities and/or uterine contractions by cardiotocography. Following evaluation the clinical course of each patient was recorded prospectively until birth. The interval (days) was computed from 2D/3D ultrasound examination to delivery of the fetus, and delivery outcomes were recorded.
Two-dimensional ultrasound and CL measurement
Ultrasonographic estimated fetal weight was calculated using biparietal diameter, head circumference, abdominal circumference, and femur length. The length of the closed portion of cervix was measured transvaginally in patients with intact membranes or translabially if PPROM was present.
Three-dimensional fetal adrenal gland data acquisition and volume calculation
Fetal AGV acquisition was performed with the Voluson 730 and E8 systems (Voluson Expert; General Electric Medical Systems, Milwaukee, WI), equipped with a 4-8 MHz curved array transducer as previously described. Three-dimensional blocks were analyzed by a single investigator (O.M.T.) blinded to pregnancy outcome. Calculation of AGV was performed using VOCAL (Virtual Organ Computer-aided AnaLysis, 4D view; General Electric Medical Systems) software package as previously described. The corrected AGV (cAGV) was calculated by using the following formula: cAGV = AGV/estimated fetal weight. As previously described, the cAGV is a GA-independent parameter.
Calculation of fetal zone enlargement
To provide proof of concept regarding our proposed methodology, the whole gland and fetal zone dimensions were determined by using the volumetric blocks used to calculate the AGV. The fetal zone was identified in transverse, sagittal, and coronal planes as an echogenic demarcated area ( Figure 1 ). These 3 planes were used to measure the length (L), width (W), and depth (D) of the whole fetal adrenal gland. Similarly, the fetal zone’s length (l), width (w), and depth (d) were also measured. L and l were measured in transverse or sagittal planes. W and w were measured in transverse or coronal planes. D and d were measured in sagittal or coronal planes. Measurements were obtained in those planes that gave the best definition of tissue interfaces.
Three consecutive measurements of fetal adrenal gland and fetal zone were obtained, and the mean of 3 measurements were used for final analysis. The relative size of the fetal zone was calculated for each orthogonal plane by dividing the fetal zone dimension by the gland dimension at the widest point (l/L, w/W, and d/D).
The intraobserver coefficients of variation for calculation of the cAGV and fetal zone dimensions were 1.5% and 3.5%, respectively.
Data analysis
Three ultrasound parameters, the cAGV, relative 2D size of the fetal zone, and the length of the closed portion of the cervix, were related to delivery within 7 days (primary outcome). History of prior PTB, PPROM, tocolytic use, and delivery outcome findings were related to each test variable.
Normality of data was evaluated with the Kolmogorov-Smirnov test. Categorical variables were analyzed using χ 2 or Fisher’s exact tests as appropriate. Continuous variables were analyzed using Mann-Whitney or student t test according to their distributions. Pearson correlation was used to measure colinearity between the selected independent variables as well as other relevant relationships between dependent and independent variables. Predictive cutoff levels for each test were calculated by receiver operator characteristics (ROC) curve analysis and curves were compared to assess the predictive value of each. Regression analysis was used for covariate analysis. SPSS 11 (SPSS Inc, Chicago, IL), MedCalc (Broekstraat, Belgium), and Excel 2007 (Microsoft, Richmond, CA) were used for analysis.
Results
Seventy-four patients at high risk for PTB were recruited for the study. Demographics and clinical and outcome characteristics are presented in Table 1 . Of the women evaluated in this study, 27 (36.5%) delivered within 7 days from the time of our evaluation. As expected, women who gave birth longer than 7 days from the time of our evaluation more frequently had intact membranes, were delivered at a higher GA, and had babies of greater birthweight. The majority of women who delivered less than 34 weeks GA had a delivery interval less than 7 days.
| Characteristics | All cases (n = 74) | Delivery ≤7 d (n = 27) | Delivery >7 d (n = 47) | P value a |
|---|---|---|---|---|
| Demographic | ||||
| Age, y, mean (SD) | 26.0 (6.38) | 25.7 (5.63) | 26.2 (6.83) | NS |
| Parity, median (range) | 1 (0–4) | 0 (0–3) | 1 (0–4) | NS |
| Race (n, %) | NS | |||
| Black | 39 (53) | 15 (56) | 24 (51) | |
| White | 25 (34) | 8 (31) | 17 (36) | |
| Hispanic | 7 (9) | 3 (11) | 4 (9) | |
| Other | 3 (4) | 1 (2) | 2 (4) | |
| Gestational age at enrollment, wks, mean (SD) | 27.6 (3.49) | 27.6 (3.71) | 27.4 (3.38) | NS |
| History of prior preterm birth, n (%) | 26 (35) | 8 (31) | 18 (39) | NS |
| Presence of PPROM, n (%) | 21 (28) | 12 (46) | 9 (19) | .03 |
| Outcome | ||||
| Gestational age at delivery, wks, mean (SD) | 32.1 (4.87) | 28.1 (3.40) | 34.5 (3.96) | < .0001 |
| Delivery weight, g, mean (SD) | 1779 (864.59) | 1098.7 (504.27) | 2225 (786.22) | < .0001 |
| Cesarean section, n (%) | 22 (30) | 8 (30) | 14 (30) | NS |
| Delivery <37 wks, n (%) | 58 (78) | 27 (100) | 33 (70) | NS |
| Delivery <34 wks, n (%) | 44 (59) | 26 (96) | 18 (38) | < .0001 |
a Comparisons performed between delivery of groups 7 days or less and delivery longer than 7 days.
There was a significant correlation between GA and all 3 dimensions of the adrenal gland fetal zone (l, w, and d) (Pearson correlation R = 0.42, 0.50, and 0.54, respectively; P < .001 for all). Similarly, we observed a significant correlation between GA and the total gland measurements (L, W, and D) (R= 0.56, 0.52, and 0.41, respectively; P < .0001 for all). However, there was no relationship between GA and the relative dimension of the adrenal gland fetal zone (R= −0.04 for the l/L ratio, 0.07 for the w/W ratio, and 0.12 for the d/D ratio, respectively; P > .05 for all).
To determine the best indicator of fetal zone enlargement, we first correlated the cAGV individually with l/L, w/W, and d/D ratios. All 3 ratios correlated significantly with the cAGV (R = 0.53, 053, and 0.60, respectively; P < .0001 for all) ( Figure 2 ). Next, using ROC curve analysis, we calculated the predictive performance of the different ratios for PTB within 7 days of delivery ( Table 2 ). By using ROC curve comparative analysis, we determined that the d/D ratio showed superior predictive value compared with the l/L and w/W ratios (z = 2.49, P = .013 and z = 2.08, P = .037, respectively). Accordingly, the d/D ratio was used as a measure of fetal zone enlargement (FZE) and for the remainder of the analysis.
