Risk of adverse pregnancy outcome in isolated single umbilical artery diagnosed at the mid-trimester anomaly scan: a large Danish retrospective cohort study

Abstract Objective To examine the association of isolated single umbilical artery (iSUA) confirmed at the mid-trimester anomaly scan and adverse pregnancy outcome and congenital malformations with up to 10 years postnatal follow up. Methods This retrospective cohort study included 116,501 singleton pregnancies consecutively enrolled in first trimester screening for aneuploidies and mid-trimester anomaly scan at three University Hospitals in the Capital Region of Copenhagen, Denmark. Data from the Danish Fetal Medicine Database (2008–2017) were verified by manually scrutinizing pre- and postnatal records. The main outcomes of interest were intrauterine fetal demise (IUFD), small for gestational age (SGA), preterm delivery, cesarean section and unrecognized pre- and postnatal congenital malformations. Results In total, 775 pregnancies with iSUA were identified. Isolated SUA were associated with a significantly increased risk of IUFD (OR 4.16, 95% CI 2.06–8.44), SGA < 3rd centile (aOR 2.41, 95% 1.85–3.14) and SGA < 10th centile (aOR 1.84, 95% CI 1.53–2.21), but not with preterm delivery or cesarean section. The laterality of the missing artery was not associated with SGA. In total, 4.3% of pregnancies with iSUA had unrecognized congenital malformations. 1.5% with iSUA had congenital cardiovascular malformations, which were considered minor. Conclusion Isolated SUA is associated with IUFD and SGA, supporting surveillance during third trimester. If, during the mid-trimester scan, the sonographer achieves thorough, extended cardiac views and finds no additional malformation other than SUA, fetal echocardiography seems not to be needed.


Introduction
Single umbilical artery (SUA) is a condition where the umbilical cord contains only one artery instead of two.The main theories regarding the pathogenesis of SUA include primary agenesis, persistence of the allantoic artery or secondary atrophy; the last-mentioned mechanism being the one thought of as the most common [1].It is well documented that SUA is associated with chromosomal abnormalities and congenital malformations mainly in the heart and urogenital system [2,3].The incidence of SUA is approximately 0.5% at the time of the mid-trimester anomaly scan in singleton pregnancies with completed combined first trimester screening for aneuploidies [2].Around 5% of these cases with SUA have additional congenital malformations [2].The term isolated single umbilical artery (iSUA) generally applies when no further malformations than SUA or chromosomal abnormalities are detected [2].
Several studies have reported an increased risk of small for gestational age (SGA) in pregnancies with iSUA [3][4][5][6][7].Furthermore, an increased risk of preterm delivery and intrauterine death has been reported in a recent systematic review [8].It has also been reported that up to 3% of cases with iSUA have unrecognized congenital malformations detected after the midtrimester anomaly scan [4].However, the significance of iSUA is not well established.Systematic reviews have concluded, that large cohort studies are needed to reach definitive results and conclusions about the significance of iSUA [8][9][10].These results can revise clinical guidelines and improve the quality of prenatal counseling and management in pregnancies with iSUA.
The aim of this study was to examine the significance of iSUA diagnosed during the routine midtrimester anomaly scan in singleton pregnancies with a completed first trimester combined screening for aneuploidies.The risks of adverse pregnancy outcome including intrauterine fetal demise (IUFD), SGA, preterm delivery, cesarean section and congenital malformations detected later in the pregnancy or postnatally were explored.

Methods
A retrospective cohort study was conducted based on a cohort of singleton pregnancies consecutively enrolled into the first trimester screening program for aneuploidies in a 10-year period (1 st January 2008 -31 st December 2017) at three University Hospitals (Hvidovre Hospital, Rigshospitalet and Herlev Hospital) in the Capital Region of Copenhagen, Denmark.The cohort was restricted to pregnancies attending both the first trimester combined screening for aneuploidies (nuchal translucency measurement, free beta-hCG and PAPP-A MoM) and the routine mid-trimester anomaly scan (performed between GA 18 þ 0 -22 þ 6) with a viable fetus and delivery in one of the three hospitals.
Data was obtained from the Danish Fetal Medicine Database (DFMD).Since 1 st of January 2008, the DFMD has automatically collected data through Astraia databases (Astraia software, Munich, Germany) containing all prenatal screening results (maternal characteristics, biochemical variables, ultrasonic biometrics, fetal malformations etc.) obtained from all hospital departments of obstetrics and gynecology in Denmark [11].Any malformation detected in the fetus is categorized using the International Classification of Diseases, 10 th revision code system (ICD-10).
Data on pregnancy outcomes and results of preand postnatal chromosome analysis are automatically collected in the DFMD from national registers [12].In the DFMD all data are linked to a unique personal identification number (CPR number), which has been given to national citizens, newborns or upon immigration to Denmark since 1968 [13].
Sonographers, who have obtained The Fetal Medicine Foundation Certificate of Competence in the 18-23 weeks scan, carried out the scans [14].According to protocol, the routine mid-trimester anomaly scan included examination of the two umbilical arteries by colour-flow.In cases with SUA, we registered the laterality of the present artery if recorded during the prenatal scans.According to national guidelines, the mid-trimester scan included a systematic evaluation of the fetal heart.The sonographers assessed the situs, rhythm, location and axis of the heart as well as the four-chamber view, ventricular outflow tract views and the three-vessel trachea view.Until 2015, all hospitals subsequently followed the diagnosis of SUA with fetal echocardiography at 21 weeks by a specialist (at Hvidovre Hospital during the full period of inclusion).All three hospitals offered two additional growth scans in third trimester.The midwife routinely assessed the placenta and cord at every birth, registering the number of vessels (SUA or a three-vessel cord), and performed a standard examination of the child.According to national guidelines, children born with iSUA and no additional clinical postnatal findings were not routinely examined by a pediatrician.In cases with iSUA and clinical suspicion of a cardiovascular malformation, a postnatal echocardiography was performed by a specialist.
Isolated SUA (iSUA) was defined as pregnancies with SUA, but without additional malformations or chromosomal abnormalities discovered before or at the routine mid-trimester anomaly scan.If any malformation was suspected at the mid-trimester anomaly scan, but later rejected at an expert scan, cases were classified as iSUA.Cases classified as iSUA with a malformation diagnosed later during pregnancy or postnatally, were classified as iSUA with an unrecognized malformation.The reference group consisted of the complete, consecutively enrolled cohort except the cases with iSUA.Accordingly, cases with other malformations and chromosomal abnormalities were included in the reference group.
Cases with SUA were identified using an ICD-10 code linked with SUA (ICD-10 codes: DQ753, DQ270, DQ270A, DQ270B, DQ273, DQ278, DQ279).Additionally, a local query was established to identify pregnancies in local fetal medicine databases (Astraia), where an ICD-10 code was missing.The query searched the records in Astraia for the terms '"SUA Ã '" or '"Single Umb Ã .'"These pregnancies potentially with SUA were all manually validated, including every visit during prenatal scans at the three hospitals.If SUA was not confirmed by manual validation in the medical records (prenatal files, autopsy or postnatal physical examination) we classified the case as having a three-vessel cord.
The outcomes of interest in cases with iSUA were IUFD beyond GA 22 þ 0 and SGA defined as a birth weight deviation z-score < À1.88 (<3 rd centile) or zscore < À1.29 (<10 th centile) based on the mean birth weight in normal pregnancies according to GA and corrected for gender [15,16].Other outcomes were preterm delivery (GA < 34 þ 0 and < 37 þ 0), cesarean section (acute and elective) and congenital malformations recognized prenatally after the routine mid-trimester anomaly scan or within a minimum of one-year follow-up after birth.
Cases with iSUA, where a malformation had been detected after the mid-trimester anomaly scan, were manually validated pre-and postnatally to confirm or correct the findings.Any imprecision regarding manual validation passed a second opinion.In cases of spontaneous or induced abortion or infant death, where an autopsy had been made, differences from the prenatal diagnosis were noted.Minor malformations considered irrelevant were excluded from the analyses (Table 2).

Statistical analysis
Groups were compared using Chi-Square test, t-test or Mann-Whitney U-test as appropriate.We calculated the unadjusted odds ratios (OR) and 95% confidence intervals (CI) to express the relationship between iSUA and each outcome.For the outcomes IUFD and preterm delivery <34 þ 0 only the univariate analysis applied because of few cases.Multivariate logistic regression analysis was used to calculate the adjusted OR (aOR) and 95% CI using a backward regression model approach.Adjusted confounders were maternal age, BMI, smoking status (yes/no or stopped in first trimester), conception method (Natural/artificial reproductive technology (ART)).
Stepwise, we controlled each potential confounder and interaction parameters in relation to iSUA and the respective outcome, only including covariates having a p-value <.05 in the final model.We adjusted for SGA (<10 th centile) in cases with preterm delivery and cesarean section.All statistical analyses were conducted with IBM SPSS Statistics 25, Chicago IL, USA.

Approval
As stated in the Danish law of ethical treatment of medical research, medical research studies including data extracted entirely from databases do not require approval by the Danish National Committee on Health Research Ethics (NVK).The study aligns with these criteria regarding extracted data §14, pc. 2. Access and reuse of data from databases was approved by the Danish Patient Safety Authority (31-1521-345).The Data Protection Agency in Denmark approved the management and storing of data (P-2020-572).

Results
The study population consisted of 116,501 singleton pregnancies.The initial broad search on ICD-10 codes and local query resulted in 2,006 cases being selected for manual validation of the SUA diagnosis.A total number of 842 pregnancies with SUA (0.72%) were identified.Of these, 12 cases (1.4%) was diagnosed with SUA later in pregnancy (n ¼ 10) or postnatally (n ¼ 2) despite a normal mid-trimester scan.In the remaining 830 (98.6%) pregnancies with SUA, 775 pregnancies had iSUA (93.4%) confirmed by ultrasound at the mid-trimester anomaly scan and were available for analysis.The laterality of the present artery was available in 593 cases with iSUA (76.5%).In 367 cases with iSUA (61.9%) we found the right umbilical artery present, while in 226 cases with iSUA (38,1%) we found the left umbilical artery present (p < .001).
The maternal-and pregnancy related characteristics for this cohort is outlined in Table 1.In pregnancies with iSUA, 100 (12.9%) were conceived by ART which was significantly higher compared to 9,381 (8.1%) pregnancies in the reference group (p < .001),while significantly more fetuses with iSUA were females (53.8%) compared to the distribution of females (48.5%) in the reference group (p < .003).

Discussion
The main findings in this study of 775 pregnancies with iSUA were a significantly increased risk of IUFD (OR ¼ 4.16) and SGA.Regarding SGA, we found a significantly increased risk of SGA < 3 rd percentile (aOR ¼ 2.41) and SGA < 10 th percentile (aOR ¼ 1.84).We found 4.3% with iSUA in the second trimester had unrecognized malformations.1.5% had unrecognized congenital cardiovascular malformations, all considered minor.
Strengths of this study include the 116,501 singleton pregnancies, all of which were consecutively enrolled into the first trimester screening program for aneuploidies and the mid-trimester anomaly scan.This enhanced the generalizability of our findings to other populations with a similar screening program.We performed an initial broad search of ICD-10 codes and a sensitive search with the local query validating more than 2000 cases in order to detect as many with SUA.Verifying all prenatal journals on pregnancies with SUA and postnatal journals in cases with IUFD or malformations up to 10 years postnatally, ensured a clear timeline and consistent high quality.Limitations to our study include the incomplete postnatal verification of the diagnosis SUA.The sensitivity and specificity when using prenatal ultrasound to diagnose SUA has shown to be 94,3% and 99,9%, respectively, and a potential misclassification would tend to underestimate our results [4].We did not calculate ORs in Table 2 as the frequency of unrecognized malformations in the reference group was not manually validated.The DFMD was less complete in regards of data on fetal malformations the first two years of inclusion from 2008 to 2010 [12].
To our knowledge, this is the first study about iSUA and adverse outcomes in pregnancies systematically screened in both first-and second trimester.Previous studies are based on cohorts without or with only partial 1. trimester screening [2,3,17].In our study, we found that iSUA increased the risk of IUFD by four times.A recent meta-analysis by Dagklis et al. [8] included 914 cases with iSUA and found a statistical significant association regarding iSUA and intrauterine death (OR 2.62, 95% CI 1.43-4.79).Additional large cohort studies besides the two included in the metaanalysis supports our findings [3,18,19].Kaulbech et al. [3], Ebbing et al. [18] and Gutvirtz et al. [19] found a significant higher risk of perinatal mortality in iSUA with ORs of 3.11, 5.1 and 7.3, respectively.All studies found iSUA associated with placental abnormalities, while Ebbing et al. [18] and Gutvirtz et al. [19] additionally found associations with cord abnormalities, which we also observed in cases with IUFD, see Supplementary Table S1.
We found increased risk of SGA in iSUA, in agreement with the majority of previous studies [3][4][5][6][7][8].Our findings regarding the laterality of the missing artery closely resembled the distribution found by Friebe-Hoffmann et al. [1], finding the right umbilical artery present in 64.2% of cases.We found no association between laterality and SGA in cases with iSUA.In contrast, a smaller study found a lower mean birthweight and height only in iSUA with the right umbilical artery present [20].
In contrast to other large cohort studies by Kaulbech et al. [3], Ebbing et al. [18] and Gutvirtz et al. [19] we did not find increased risk of cesarean section and preterm delivery in cases with iSUA, supported by a few smaller studies [9,21,22].The meta-analysis by Dagklis et al. [8] reported significant associations regarding cesarean section (OR 1.64) and preterm birth (OR 2.48) but substantial heterogeneity was noted.The different conclusions may be attributed to differences in study design, as we included only pregnancies with first trimester combined screening.Furthermore, we did not exclude pregnancies with malformations and chromosomal abnormalities from our reference group since the diagnosis was not manually validated.This could cause a possible underestimation (conservative estimate) of our risk estimates for adverse outcomes.Khalil et al. [4] found that 2.6% (4/156) of iSUA cases had anomalies at birth.In contrast, we listed each of the 33 (4.3%) cases with unrecognized malformations including the time of detection to improve the management in pregnancies with iSUA.It is well documented that around 19-32% of all congenital abnormalities are detected for the first time in the third trimester or postnatally despite a second trimester malformation scan [23,24].In alignment with our findings, the most common findings were ventricular septal defects, duplex kidney, gastrointestinal atresi, hypospadi etc.
The need for fetal echocardiography in iSUA has been debated.Smaller studies on mixed populations found no indication of fetal echocardiography, when no chromosomal abnormalities or malformations were found at the mid-trimester anomaly scan [2,[25][26][27][28].In our study, 12 cases (1.5%) with iSUA had unrecognized cardiovascular malformations and all were considered minor malformations.All were detected postnatally except one ventricular septal defect found by prenatal, fetal echocardiography despite a normal mid-trimester anomaly scan.Our findings aligned with a recent study by Tai et al. finding a congenital heart disease birth prevalence of 1.2-1.8% in 326 patients with iSUA [29].
Isolated SUA diagnosed during the routine mid-trimester anomaly scan in singleton pregnancies with completed first trimester combined screening for aneuploidies was associated with increased risk of IUFD and SGA, justifying surveillance on fetal growth during the third trimester.After the mid-trimester anomaly scan, we did not find a large number of major, unrecognized malformations in any affected organ system including the cardiovascular system.Finding SUA at the mid-trimester scan should prompt the sonographer to search thoroughly for other fetal malformations.The amount and severity of unrecognized malformations highly depends on the expertise of the sonographers and the population's enrollment in prenatal screening programs.In a similar setting to ours, with standardized educated and experienced sonographers, when the sonographer achieves sufficient standard cardiac views and no malformation is present, additional fetal echocardiography seems not to be needed.
There are few studies on postnatal outcomes in infants born with iSUA.Recent cohort studies found children born with iSUA had increased risk of hospitalassociated genitourinary infections and long-term respiratory morbidity [30,31].One study found no evidence of SUA affecting growth and neurodevelopment, yet another found SUA associated with impaired school achievements [32,33].Additional longterm follow up studies in children and adolescents born with iSUA are warranted.

Table 1 .
Baseline maternal-and pregnancy related characteristics.

Table 2 .
Congenital malformations recognized after the mid-trimester malformation scan in 33 cases with unrecognized malformations.Each case can have multiple malformations affecting one or several organ systems.Ã Irrelevant minor malformations excluded from total cases and analyses.