Impact of Maternal-Fetal Environment on Outcomes Following the Hybrid Procedure in the Single Ventricle Population

Introduction: Treatment of infants with hypoplastic left heart syndrome (HLHS) remains challenging, and those affected remain with signi�cant risks for mortality and morbidity throughout their lifetimes. The maternal-fetal environment (MFE) has been shown to affect outcomes for infants with HLHS after the Norwood procedure. The hybrid procedure, comprised of both catheterization and surgical components, is a less invasive option for initial intervention compared to the Norwood procedure. It is unknown how the MFE impacts outcomes following the hybrid procedure. Methods: This is a single-center, retrospective study of infants born with HLHS who underwent hybrid palliation from January 2009-August 2021. Predictor variables analyzed included fetal, maternal, and postnatal factors. The primary outcome was mortality prior to Stage II palliation. Results: We studied a 144-subject cohort. There was a statistically signi�cant difference in mortality prior to stage II palliation in infants with prematurity, small for gestational age, and aortic atresia subtype (p<0.001, p=0.009, and p=0.008, respectively). There was no difference in mortality associated with maternal diabetes, hypertension, obesity, smoking or illicit drug use, or advanced maternal age. State and national area deprivation index scores were associated with increased risk of mortality in the entire cohort, such that infants born in areas with higher deprivation had a higher incidence of mortality. Conclusions: Several markers of an impaired MFE, including prematurity, small for gestational age, and higher deprivation index scores, are associated with mortality following hybrid palliation. Individual maternal comorbidities were not associated with higher mortality. The MFE may be a target for prenatal counseling and future interventions to improve pregnancy and neonatal outcomes in this population.


Introduction
The single ventricle population, including infants with hypoplastic left heart syndrome (HLHS) or HLHS variants, is a unique subset of patients with CHD who require surgical intervention within the rst week of life. The Norwood procedure is the most common initial surgical palliation. The hybrid stage I (HS1) palliation is an alternative intervention that includes surgical placement of bilateral pulmonary artery bands, transcatheter atrial septostomy, and placement of a stent across the ductus arteriosus. Relative to the classic Norwood procedure, bene ts to the HS1 palliation include a less invasive procedure in the neonatal period prior to maturation of the neonatal brain, shortened anesthesia exposure, avoidance of cardiopulmonary bypass and arch reconstruction, and less risk of end-organ injury including kidney injury [8]. Although often reserved at many institutions for smaller infants of younger gestational age or highrisk surgical candidates [9], the hybrid procedure is routinely the initial palliation of choice for HLHS at Nationwide Children's Hospital (NCH).
Stage I palliation, whether Norwood or hybrid, is associated with signi cant morbidity and mortality. The interstage period, the time between the rst and second stages of palliation, is a particularly vulnerable period. The Single Ventricle Reconstruction Trial analyzed a large cohort of single ventricle patients who underwent Norwood palliation, and over half of the mortality in this population occurred with the stage I palliation, while interstage mortality accounted for another one-third of the deaths [10]. Impaired MFE, de ned as maternal gestational hypertension, pre-eclampsia, gestational diabetes, and/or smoking during pregnancy, was associated with a signi cantly higher risk of death following the Norwood procedure [11]. Identi ed independent sociodemographic risk factors for interstage mortality following the Norwood procedure include teen mothers and single caregivers [12].
The impact of the MFE has not been applied the hybrid procedure. This study will investigate potential factors that may identify patients with higher risk for mortality following the hybrid procedure based on their MFE. We hypothesize that among patients with HLHS who undergo the hybrid procedure, there will be an increased risk for poor surgical outcomes in those with an impaired MFE.

Methods
We conducted a retrospective, observational cohort study at NCH for patients born between January 2009 and August 2021. Inclusion criterion was infants born with single ventricle CHD with a dominant right ventricle and aortic arch obstruction who underwent HS1 palliation at NCH with intention to treat for complete single ventricle palliation. Exclusion criteria included infants with single left ventricle, infants who underwent a subsequent two-ventricle repair or palliation without intention to treat, infants with incomplete records including those who underwent HS1 palliation at an outside institution or were lost to follow up, and infants with Trisomy 21. The primary outcome was mortality prior to stage II palliation. Data were obtained through the NCH electronic medical record in the infant's chart, which included prenatal, postnatal, and maternal data. Additional maternal records were obtained through a shared electronic medical record in the mother's chart when available. This study was approved by the Institutional Review Board at NCH.
The MFE was evaluated using fetal, maternal, and postnatal factors re ecting the fetal environment.
Collected fetal variables included gender, gestational age, and diagnosis (anatomic subtype, restriction of the atrial septum, major genetic diagnosis, and presence of a prenatal diagnosis). Gestational age was used to identify premature infants, de ned as those born prior to 37 weeks gestational age. Determination of anatomic subtype was de ned by the presence of aortic atresia. Atrial septum restriction was de ned by the need for urgent atrial septal intervention within the rst 48 hours of life.
Major genetic syndromes included in the study were Turner Syndrome, Jacobsen Syndrome, and Kabuki syndrome. There were no patients in our cohort with Trisomy 18 or CHARGE syndrome.
Maternal variables analyzed included maternal race, insurance type, nine-digit zip codes, and the presence of ve maternal risk factors, including advanced maternal age (≥ 35 years old), smoking or illicit drug use, diabetes (gestational or pregestational), hypertension (gestational or pregestational), or obesity (body mass index ≥ 30kg/m 2 ). Maternal race was categorized as Caucasian or non-Caucasian. Insurance type was categorized as Medicaid or private insurance. The nine-digit zip code was used to generate state and national area deprivation index (ADI) scores using the Neighborhood Atlas, developed by the University of Wisconsin-Madison [13,14]. Maternal risk factors were analyzed individually and as a composite, comparing the presence of any risk factors (one or more) to the absence of all risk factors.
Smoking and illicit drug use were determined based on self-reported use at any time during the pregnancy.
Postnatal data that was collected included birth weight, APGAR scores, pre-procedural data (respiratory support prior to intervention, brain-natriuretic peptide (BNP) within 24 hours prior to surgery, rst neonatal creatinine as a marker of maternal milieu, age at time of hybrid palliation), and data from the initial echocardiogram after birth (right ventricular function, tricuspid regurgitation). Birth weight was used to identify infants who were SGA, de ned as a birth weight less than the tenth percentile for their gestational age. Respiratory support prior to hybrid palliation was de ned as the need for intubation, non-invasive positive pressure, or high-ow nasal cannula pre-operatively. Echocardiograms were performed in accordance with guidelines from the American Society of Echocardiography. All echocardiograms were independently reviewed for assessment of RV function and tricuspid regurgitation.
Based on previous studies suggesting higher mortality in HLHS infants with aortic atresia [15], a separate analysis was performed on infants with this anatomic subtype. A separate analysis was also performed on a cumulative high-risk subgroup, which included infants with any of the following features: birth weight less than 2.5kg, moderate or severe tricuspid regurgitation or right ventricular dysfunction after birth, a major genetic diagnosis, or a restrictive atrial septum. Variables in these analyses were selected based on signi cance in the univariable analysis, and additional maternal risk factors were selected based on relatively high prevalence within our study.

Statistical Analyses
Continuous variables are presented as medians with interquartile range while categorical variables are presented as counts with proportion of the sample. Characteristics of infants who did and did not experience mortality were compared using Wilcoxon rank-sum tests for continuous variables or Fisher's exact tests for categorical variables. Similar comparisons were made for infants with and without aortic atresia subtype as well as infants meeting criteria for the cumulative high-risk subgroup. The relationship between the ADI and prematurity, SGA, and the presence of maternal risk factors was visualized using box plots. Logistic regression was used to estimate the adjusted effect of National ADI, aortic atresia subtype, and prematurity on mortality prior to stage II palliation. P values of < 0.05 were considered signi cant and all con dence intervals are 95% con dence intervals. All analysis was conducted in R version 4 [16, 17] and all graphs were made using ggplot2 [18].

Results
Between January 2009 and August 2021, there were 151 infants with HLHS who underwent the hybrid procedure. Seven were excluded -two secondary to a diagnosis of Trisomy 21, two due to hybrid palliation without intention to treat through the single ventricle pathway, two additional infants who underwent hybrid palliation at an outside institution, and one infant was lost to follow-up after transferring to another institution. The nal cohort consisted of 144 infants who met inclusion criteria for our study.
Fetal and maternal demographics are listed in Table 1. The majority of infants had a prenatal diagnosis of HLHS (81%) and were born full term (88%). Among the 17 infants born prematurely, the median gestational age was 35 weeks (range 31-36 weeks). Median birth weight was 3.140kg. There were 17 infants with an absolute birth weight < 2.5kg and 23 infants (16%) in the nal cohort were SGA. There were 28 infants in our study whose mothers resided from a state other than Ohio. 105 mothers (73%) had at least one of the ve maternal risk factors. Only two mothers had late prenatal care with care starting in the third trimester.  Table 2 shows the relationship of fetal, maternal, and postnatal factors with mortality. Considering fetal factors, prematurity and aortic atresia subtype (p < 0.001 and p = 0.008, respectively) were associated with increased mortality prior to stage II palliation. No mortality difference was observed in maternal race, insurance type, or maternal risk factors. Higher state and national ADI scores, re ective of more deprived neighborhoods, were also associated with mortality (p = 0.028 and p = 0.019, respectively). Across postnatal factors, increased mortality was observed among infants with lower birth weight and those meeting criteria for SGA (p = 0.039 and p = 0.009). The need for respiratory support prior to hybrid palliation was also signi cantly higher among non-survivors (p = 0.049), while measured cardiac indices (right ventricular dysfunction, tricuspid regurgitation on rst postnatal echocardiogram) were not associated with mortality following the hybrid procedure.  0 and 64, respectively). The median state ADI score for moms with one or more maternal risk factors was 5.0, compared to a score of 4.0 for moms with no maternal risk factors.
The median national ADI score for moms with one or more maternal risk factors was 69, compared to a score of 56 for moms with no maternal risk factors.
Within the two high risk subgroups identi ed, there were 79 infants with aortic atresia anatomic subtype and 45 infants in the cumulative high-risk subgroup (Table 3). Among infants with aortic atresia, mortality was 29% (23/79). Mortality was associated with prematurity, SGA, and lower birth weight. While state and national ADI scores trended higher in infants with mortality, this did not reach statistical signi cance. Among the 45 infants in the cumulative high-risk subgroup, there were 17 infants with birth weight < 2.5kg, 8 infants with a major genetic diagnosis, 8 infants with a restrictive atrial septum, 16 infants with moderate or severe tricuspid regurgitation, and 11 infants with moderate or severe right ventricular dysfunction. Within this subgroup, mortality was 29% (13/45). Mortality was again associated with prematurity and lower absolute birth weight. Similar to the aortic atresia subgroup, state and national ADI scores trended higher in infants with mortality in this subgroup but did not reach statistical signi cance.

Discussion
Despite signi cant advances in surgical and medical management in CHD, surgical palliation for infants with single ventricle CHD remains associated with signi cant mortality [10,19]. Several studies have shown an association between an impaired MFE and post-operative mortality in this population [6, 11,12]. However, these studies have primarily or exclusively included infants who underwent Norwood palliation. Our study goal was to evaluate the impact of the MFE on mortality in a large cohort of infants who underwent hybrid palliation for single ventricle CHD.
Previous studies de ned an impaired MFE by the presence of maternal gestational hypertension, preeclampsia, gestational diabetes, smoking during pregnancy, preterm birth, or SGA [6, 11]. Our study evaluated the impact of these maternal co-morbidities of impaired MFE, as well as several other fetal, maternal, and postnatal factors in order to more broadly evaluate and capture an impaired MFE, including socioeconomic factors. ADI assesses socioeconomic status as a measure that includes factors such as education, employment, housing quality, and income to identify disadvantaged neighborhoods.
Fetal factors that were associated with higher mortality in our population included prematurity and aortic atresia, consistent with previous studies [6,7,15]. There were few infants within our cohort with a restrictive atrial septum or a major genetic diagnosis, making it di cult to adequately assess the role of these factors in mortality. Lack of a prenatal diagnosis did not impact mortality. However, this does not take into consideration other important bene ts of a prenatal diagnosis including improvement in parental and care team preparation and planning, increasing the potential for prevention of signi cant infant comorbidities. Markkanen et. al demonstrated that prenatal diagnosis of HLHS is associated with improved postnatal right ventricular function, reduced metabolic acidosis, and reduced end-organ dysfunction [20]. Similarly, in our study, signi cant right ventricular dysfunction was seen exclusively in infants with a postnatal diagnosis, all of whom improved once the diagnosis was made. Within our health system as a free-standing children's hospital, linkage of maternal-fetal data can be di cult. Using the electronic medical record, including the infant and mother's charts, we were able to collect 97% of maternal data points. The presence of at least one maternal risk factor within our cohort was overwhelmingly high, present in 73% of mothers. This is clinically important and concerning for poor overall maternal state of health. Although these risk factors individually were not associated with higher mortality, mortality tended to be higher in the presence of any risk factor compared to infants of mothers with zero risk factors. These maternal risk factors were also related to ADI scores. Mothers with one or more risk factors had a higher median ADI score compared to mothers with zero risk factors. In these maternal-fetal dyads, worse pregnancy outcomes (prematurity and SGA) also had a trend toward higher median ADI scores when compared to their term and average weight for gestational age counterparts. This has been shown in other population health studies of pregnancy outcomes [21][22][23]. This suggests an important relationship between overall maternal health, the MFE, socioeconomic status, and pregnancy outcome metrics of gestational age and birthweight. The impact of these relationships is potentially magni ed when the infant is born with critical CHD such as HLHS. Future use of ADI scores may aid in the development and targeting of pre-pregnancy and prenatal intervention programs to improve the MFE for the highest risk maternal-fetal dyads.
Among infants with aortic atresia and those identi ed as cumulative high-risk, prematurity and lower birth weight remained signi cantly associated with mortality. ADI scores trended towards higher but not to statistical signi cance in the aortic atresia or cumulative high-risk groups. These analyses may be underpowered with the smaller cohort sizes. Alternatively, the increased risks conveyed by the anatomic subtype and other high-risk features may outweigh the added risk from socioeconomic status.
The model for our multivariate analysis was created as an exploratory analysis using 3 signi cant MFE factors from the univariable analysis. When adjusted for aortic atresia and prematurity, the OR for national ADI score approaches 1. Given the signi cance of this factor in the univariable analysis, this suggests that national ADI score may not be as indicative of higher mortality risk once prematurity and aortic atresia are accounted for. It is also possible that ADI provides similar information on mortality risk as prematurity, as mothers in a more deprived area with a higher ADI score may be at risk for premature delivery. However, it is important to consider that this is a continuous variable and the OR is multiplicative. A 1% increase in national ADI score increases odds of mortality by 2.4%, so a change of several percentiles can become very signi cant. This analysis also shows that in the context of our patient population, prematurity increases the odds for mortality more than aortic atresia anatomic subtype, when considering national ADI score, prematurity and aortic atresia as mortality risk indicators.

Limitations
This study is limited by its retrospective nature. As expected, there were changes in management over the course of this retrospective study that impacted mortality. There are also patients with incomplete data, particularly related to maternal risk factors and in patients with a postnatal diagnosis for whom maternal history or neonatal data were not fully available. In instances where the status of a risk factor was unknown, if the mother indicated they did not have any other risk factors, the mother was assumed to have no risk factors. Therefore, we may have under-reported the presence of maternal risk factors given limitations associated with unique mother and baby electronic medical record systems.
There is also potential referral bias, with nearly one in ve patients within our cohort being from out of state. Although we have a relatively large cohort, particularly considering the rare use of hybrid palliation on a national level, our study is still limited by patient numbers and power. There are few patients within our study with certain features, such as genetic syndromes, a restrictive atrial septum, or maternal history of diabetes, making it challenging to assess the full impact of such features on outcomes. Furthermore, our study does not include patients with fetal demise or mortality prior to hybrid palliation who may be even higher risk patients with more features of an impaired MFE.
Additionally, in the logistic regression analysis, there may exist confounding factors that were not incorporated into the model. The results of the logistic regression, namely OR and corresponding con dence intervals, should be interpreted as such.

Conclusion
This represents the rst study to explore the impact of MFE on outcomes following hybrid palliation for HLHS. We identi ed several markers of an impaired MFE, including prematurity, SGA, and lower birth weight, that are associated with mortality in HLHS following HS1 palliation. Notably, our study also identi ed the important role of socioeconomic disadvantage in mortality following hybrid palliation.
Although speci c maternal comorbidities were not individually associated with higher mortality, infants who experienced mortality were more likely to be born to mothers from a more deprived area, as re ected by higher state and national ADI scores. This should be an important consideration in prenatal counseling and can also serve as a possible target for future modi able interventions at the prepregnancy, prenatal, and postnatal stages to improve outcomes in this patient population. Moving forward, comprehensive efforts to improve maternal health and the MFE, speci cally to decrease the incidence of preterm birth and infants born SGA, will be particularly important as we look to the future of shows box plots assessing national and state area deprivation index scores by prematurity status, presence of maternal risk factors (no= none identi ed, yes= at least 1 risk factor reported), and SGA (small for gestational age). Maternal risk factors include advanced maternal age (≥35 years old), smoking or illicit drug use, gestational or pregestational diabetes, gestational or pregestational hypertension, or obesity (body mass index ≥30kg/m 2 ). This gure shows consistently higher deprivation scores in the presence of prematurity, maternal risk factors, and SGA.