Population

This is a substudy from the CBHS, which is a prospective, multicenter cohort study including participants from the second trimester of pregnancy (gestational week 18–22) from April 2016 until October 2018. All parents at the second-trimester anatomy scan were offered the opportunity to participate in the study. All pregnant women in Denmark are offered an anatomy scan in the second trimester (gestational week 20), and >95% of pregnant women attend this screening.

The participants were included at the following 3 hospitals in Copenhagen: Rigshospitalet, Herlev Hospital, and Hvidovre Hospital. All parents gave written informed consent to participate in the CBHS, including the substudies. The neonates underwent an echocardiography within the first 4 weeks of life, and the design of the study has previously been described. The study was approved by the Regional Ethics Committee of the Capital City Region of Denmark (H-16001518) and the Danish Data Protection Agency (I-Suite no.: 04546, ID-no. HGH-2016-53). The CBHS is registered with Clinical Trial registration: clinicaltrials.gov identifier NCT02753348.

The exclusion criteria for this study were as follows: (1) twins, (2) abnormal findings on the postnatal echocardiography resulting in a referral to a pediatric cardiologist, (3) chromosomal abnormalities, (4) gestational age (GA) at scan before 18+0 or after 22+6 weeks as the second trimester anatomy scan is usually performed between 18 and 22 weeks , and (5) missing birthweight.

Ultrasonography and cardiovascular images

Obstetrical ultrasound scans in Denmark are performed by sonographers certified by the Fetal Medicine Foundation, London. For this study, we used Voluson E8 scanners, GE Medical Systems, Zipf, Austria. The GA was determined by the crown-rump length measured in the first trimester GA weeks 11 to 14. At the second trimester anatomy scan, all the standardized fetal biometry were stored as still images in the obstetrical database, Astraia (Astraia GmBH, Munich). The normal ultrasonographic screening of the fetal heart follows the national guideline that, since 2010, includes the 4-chamber view, the outflow tracts, and the 3-vessel view. In 2017, the second-trimester image documentation of the heart was extended at the 3 hospitals recruiting patients for the CBHS, thus including the placement of the heart in the thoracic cavity, the end-diastolic 4-chamber view with closed atrioventricular (AV) valves, the left ventricular outflow tract with a closed aortic valve (AoV), the right ventricular outflow tract with a closed pulmonary valve (PV), the 3-vessel view (3VV), the 4-chamber view with color Doppler flow across the AV-valves (had to be identified but were not stored until 2017), and the sagittal view of the aortic arch (new). After the inclusion of all the participants neonatally in the CBHS, the prenatal cardiovascular measurements were performed retrospectively on still images with measuring tools in Astraia by 2 investigators blinded to diagnoses and the outcome of the pregnancy. The following heart biometry were measured in Astraia: the AoV diameter at the end-systole, the PV diameter at the end-systole, the lumen diameter of the ascending aorta (AAo) in the 3VV, the cardiac axis in the thorax, the left ventricular (LV) length, the LV diameter, the right ventricular (RV) length, and the RV diameter. Furthermore, the width and the length of the heart were measured from epicardium to epicardium in the 4-chamber view ( Supplemental Figure ). We only measured on images with well-defined structures. The measurements were taken from inner-wall to inner-wall. In the 4-chamber view, the RV length was measured from the AV-valve to the beginning of the moderator band. We had few available measurements for the aortic isthmus in the sagittal view (missing 84.9%) and the axis (missing 80.9%); why analyses of those parameters were not performed in this study.

Definition of growth restriction

SGA and IUGR were classified according to the 2020 International Society of Ultrasound in Obstetrics and Gynecology (ISUOG) guidelines on growth restriction. They were defined retrospectively on the basis of the birthweight and the third trimester Doppler flow pulsatility index (PI). SGA was defined as a birthweight between the 3rd and 10th percentiles with normal Doppler flow PIs in the umbilical artery and the uterine artery (UtA-PI) and also the cerebroplacental ratio. IUGR was defined as a birthweight below the 3rd percentile or between the 3rd and 10th percentiles with abnormal Doppler flow PIs. SGA and IUGR were calculated using the Marsal’s formula , with the z-score=−1.28 for the 10th percentile and z-score=−2 for the 3rd percentile, respectively. The prenatal Doppler flow measurements were retrospectively assessed from patient files in Astraia if available in all cases with a birthweight <10th percentile. Abnormal Doppler flow was defined as UA-PI or UtA-PI >95th percentile and cerebroplacental ratio <5th percentile. Of the cases between the 3rd and 10th percentiles, 40.6% had available flow assessments in the 3rd trimester. In Denmark, 3rd trimester screening is offered only by indication. All cases with a birthweight <10th percentile had all their prenatal scans reviewed retrospectively to classify them correctly as either SGA or IUGR.

Statistical analysis

All calculations were performed using the statistical program R V.3.6.1 (R Studio Team) and R Studio V.1.2.5001 (R Studio Team). All the data are expressed as means with standard deviations (SD), medians with interquartile range (IQR), absolute values, or percentages. Comparisons were performed using the Student’s t test, χ ² test, or the Mann-Whitney U test where appropriate. After Bonferroni correction, a significance level of 0.005 was chosen (0.05/11=0.005). The mean differences in cardiovascular biometry were calculated using linear regression analysis with 95% confidence intervals (CI) and were expressed as z-scores. The analyses were adjusted for the GA at the time of the second-trimester scan, as the measured sizes are closely associated with the GA. The mean differences were also analyzed adjusting for both the GA and the abdominal circumference (AC) at the time of the second-trimester scan. Instead of adjusting for all 3 biometry (biparietal diameter, femur length, and AC), we only adjusted for AC, as possible errors for each measurement would be multiplied if using all 3 biometry. Apart from the listed biometry, analyses on the ratio between the length and the width of the heart and the LV length and width were performed as proxies for the shape of the heart. We did not have a UA-PI measurement from all the pregnancies. Hence, we also pooled the SGA and IUGR data and repeated the same analyses for all cases with a birthweight <10th percentile. To examine whether the SGA and IUGR fetuses had a smaller AC than the controls in the second trimester, we calculated the z-score, from the predicted AC adjusted for GA. A z-score =-1.28 was equal to the 10th percentile.

Furthermore, we computed the receiver operating characteristic (ROC) curves, and the area under the curve (AUC) was also calculated. The ROC curves were computed after backward regression including both biomarkers and maternal characteristics. The sensitivities were reported for set false positive rates of 5%, 10%, and 20%, respectively.