Morbidity and Mortality in Neonates With Down Syndrome vs Those Without Down Syndrome by Gestational Age

Down syndrome (DS) is the most common viable trisomy and affects approximately 6000 births in the United States each year.^1,2 A significant proportion of these neonates may require admission to the neonatal intensive care unit (NICU) after delivery due to assorted comorbidities.^3–9 Previous analyses of neonates with DS have shown increased morbidity and mortality during this time period due to various factors; however, analyses have usually been limited due to single-institution or specific regional investigations.^10–13 Epidemiological trends support that the burden of a DS diagnosis varies by region and socioeconomic status.¹⁴ 

National data regarding outcomes in the neonatal period are sparse, as this group is often excluded from studies. In addition, minimal data exist comparing outcomes in neonates with vs those without DS.¹¹ The neonatal time period has the highest risk for adverse outcomes in the population with DS.¹² Comparative data in neonatal outcomes between those with and without DS would be beneficial for counseling purposes in this high-risk population, as well as for targets for improvement in therapies and care.

The goal of this study was to determine whether there were gestational age-based differences in characteristics, morbidities, and mortality in neonates with vs those without DS to the NICU using contemporary data from US children’s hospitals.

Data for this study were obtained from the Pediatric Health Information System (PHIS), an administrative database that contains inpatient, emergency department, ambulatory surgery, and observation encounter-level data from not-for-profit, tertiary care pediatric hospitals in the United States. These hospitals are affiliated with the Children’s Hospital Association (Lenexa, Kansas). Data quality and reliability are assured through a joint effort between the Children’s Hospital Association and participating hospitals. For the purposes of external benchmarking, participating hospitals provide discharge/encounter data, including demographics, diagnoses, and procedures. Nearly all of these hospitals also submit resource use data (eg, pharmaceuticals, imaging, and laboratory) into PHIS. Data are deidentified at the time of data submission, and data are subjected to several reliability and validity checks before being included in the database. The PHIS database represents 10.6% of all pediatric discharges nationwide and 15.8% of all pediatric discharges, excluding normal newborns.

Neonates discharged between January 1, 2008, and December 31, 2022, from 47 participating PHIS sites were identified. Neonates were included if they were aged 30 days or younger at admission and their index hospitalization included time in the NICU. Those missing gestational age or with a gestational age less than 22 weeks or greater than 42 weeks were excluded. Data were collapsed to the unique hospital and medical record number level to account for multiple hospitalizations from the same neonate.

Because this study used deidentified data, the investigational review board determined this study was not human subjects research under 45 CFR part 46.102(f), and evaluation was waived.

Neonates were characterized overall and by presence of DS (identified using International Classification of Diseases [ICD], Ninth Revision, Clinical Modification or International Statistical Classification of Diseases and Related Health Problems, Tenth Revision diagnosis codes; Supplemental Table 1) and stratified by gestational age. Gestational age was categorized according to the World Health Organization’s definition of prematurity: less than 28 weeks (extremely preterm), 28 to 31 weeks (very preterm), 32 to 36 weeks (moderate/late preterm), and 37 weeks or older (term).¹⁵ Characteristics recorded included gestational age, race and ethnicity, urbanicity, zip code-level median household income (2010), birth weight, small for gestational age (SGA) status (derived according to Fenton growth charts),¹⁶ Apgar scores at 1 and 5 minutes, and presence of the following: atrioventricular septal defect, atrial septal defect, ventricular septal defect, patent ductus arteriosus (PDA), duodenal atresia/stenosis, and Hirschsprung disease (Supplemental Table 1). Birth weight was set to missing if it was less than 300 g or the first percentile or greater than the 99th percentile, as assessed using Fenton growth charts.16 A variety of in-hospital outcomes were recorded based on included ICD-9/ICD-10 diagnosis codes, procedure codes, or facility charges or as defined by PHIS. These include disposition (ie, death or discharge to hospice, discharge home, or transferred), length of stay, unplanned readmission within 30 days of discharge, extracorporeal membrane oxygenation (ECMO), pulmonary hypertension (PH) or persistent pulmonary hypertension of the newborn, nitric oxide (NO) use, necrotizing enterocolitis (NEC), NEC (Bells 2+), periventricular leukomalacia, hypoxic-ischemic encephalopathy (HIE), interventricular hemorrhage or intracranial hemorrhage (ICH), retinopathy of prematurity (ROP), bronchopulmonary dysplasia (BPD), hydrops, any central line use, central line-associated bloodstream infection (CLABSI), total parental nutrition (TPN), any infection, tracheostomy use, gastrostomy use, ventilator use, oscillator use, noninvasive positive-pressure ventilation (NPPV) use, and high-flow nasal cannula (HFNC) use.

Outcomes were compared among neonates with and without DS using linear probability (ie, risk difference) models, stratified by gestational age as the absolute risk facilitates clinical interpretability. Risk differences were reported as a percentage. When examining unplanned readmission within 30 days, analysis was limited to those who were discharged home. Data were processed using SAS Enterprise Guide 8.1 (SAS Institute, Inc). Analysis was conducted using R version 4.3.1 (R Project for Statistical Computing).¹⁷ Owing to the exploratory nature of our analysis, we present the estimate and CI with no adjustment for multiple comparisons.¹⁸ 

There were 575 116 NICU stays with admission aged 30 days or younger from January 1, 2008, through December 31, 2022. A total of 1.5% (n = 8598) of these were neonates diagnosed with DS. Excluding those who were missing gestational age, 7037 neonates with DS were compared with 460 635 neonates without DS within gestational age groups (Figure 1).

Baseline characteristics were similar between groups for sex and race and ethnicity (Table 1). In each gestational age group, the percentage of SGA neonates was higher in the group with DS compared with the group without DS. Apgar scores were similar between groups across all gestational ages. The prevalence of congenital heart defects was significantly higher in the neonates with DS compared with those without DS for all gestational age groups. This was also generally true for the diagnosis of duodenal atresia/stenosis and Hirschsprung disease.

Extremely preterm neonates with DS (n = 95) had a higher risk of death and unplanned readmission within 30 days vs those without DS (n = 30 481; Figure 2, Table 2). There was a higher percentage of PH, NO usage, and gastrostomy tube (GT) placement among those with DS compared with those without DS at this gestational age. ROP was lower for extremely preterm neonates with DS. Median length of stay among survivors was 123 days for extremely preterm neonates with DS and 99 days for those without DS.

Very preterm neonates with DS (n = 304) had a higher risk of death and unplanned readmission within 30 days vs those without DS (n = 39 265; Figure 2, Table 2). Very preterm neonates with DS had a higher percentage PH, NO usage, tracheostomy placement, ventilator use, and high-frequency oscillator use but a lower percentage of nonpositive pressure use. Those with DS had a higher percentage of developing NEC (Bells ≥ 2), GT placement, central venous line (CVL) placement, and TPN use vs those without DS. Very preterm neonates with DS also had a higher percentage of infection, CLABSI, developing hydrops, and being placed on ECMO. The median length of stay among survivors was 79 days for very preterm neonates with DS and 48 days for those without DS.

Moderate/late preterm neonates with DS (n = 2168) had a higher risk of death and unplanned readmission within 30 days vs those without DS (n = 145 024; Figure 2, Table 2). Moderate/late preterm neonates with DS had a higher percentage PH, NO use, ventilator use, high-frequency oscillator use, and HFNC use. Among moderate/late preterm neonates, those with DS had a higher percentage of GT placement, CVL placement, and TPN use vs those without DS. Moderate/late preterm neonates with DS also had a higher percentage of infection, developing hydrops, and being placed on ECMO. Median length of stay among survivors was 31 days and 14 days, respectively, for moderate/late preterm neonates with and without DS.

For term neonates with DS, there was a small increased risk of death and moderately increased risk of unplanned readmission within 30 days among those with DS (n = 4470) vs those without DS (n = 245 865; Table 2). Risk of PH, NO use, ventilator use, nonpositive pressure use, and HFNC use was higher among neonates with DS relative to those without DS. There was a slightly lower percentage HIE and ICH among those with DS vs those without DS. GT use remained higher in the term neonates with DS vs those without DS. Term neonates with DS had a higher percentage of CVL placement, TPN use, infection, and developing hydrops relative to those without DS. The median length of stay among survivors was 18 and 5 days for term neonates with and without DS, respectively.

This is the largest US database study to date examining demographics, socioeconomics, morbidities, and mortality rates in neonates with DS compared with neonates without DS admitted to the NICU. Previous literature on increased risk has been limited by region and center.^10,19 In this national study, neonates with DS had increased morbidity and mortality vs neonates without DS across all gestational age groups.

Baseline characteristics were similar between those with DS and those without DS within gestational age groups. There was a higher percentage of SGA neonates and increased percentage of congenital malformations in those with DS in all gestational age groups. It is well known that neonates with DS have lower birth weights and slower growth than those without DS; however, the specifics by gestational age have not been well described.²⁰ Currently there is no standardized growth curve for premature neonates with DS, and, often, NICUs will use the Fenton growth curve until a corrected gestational age of 50 weeks is reached before switching to DS-specific growth curves.^16,21–23 In addition, the risk for congenital heart defects and abdominal malformations are also well known to be increased in the DS population. Both these findings therefore were not surprising to document.

The increased rate of cardiac anomalies in the neonates with DS compared with those without DS was prevalent in all gestational age groups. Congenital heart disease is present in up to 50% of neonates with DS and is one of the leading causes of morbidity.^4,24 The rate of PDA detection was greater than 50% for both those with and without DS in the extremely preterm group. For very preterm, moderate/late preterm, and term neonates with DS, this rate of PDA detection remained similar, at around 60%. However, the rate of PDA detection in neonates without DS in the very preterm, moderate/late preterm, and term groups was all less than 25%. This difference could be partially explained by increased surveillance in this population, as an echocardiogram is recommended for all neonates with DS, and, if admitted to the NICU, it will likely be completed there to aid in clinical management.²⁵ Neonates without DS in the NICU, particularly late preterm and term gestations, are less likely to have an echocardiogram completed and, therefore, are less likely to have a PDA discovered by imaging. However, the increased incidence of PH in neonates with DS compared with those without DS in all gestational age groups was mirrored by the increased rates in NO usage. This argues that neonates with DS not only had echo findings compatible with PH but also clinical signs and symptoms prompting treatment. Previous research has reported an increased incidence of PH in neonates with DS compared with the general neonatal population using published incidences,⁸ but none have described the differences by gestational age group as our study has. The mechanism of PH in neonates with DS is thought to be multifactorial, including genetic, congenital, and environmental causes.²⁶ Although large prospective studies focused on the etiology of PH in those with DS are minimal, multiple risk factors in this population for developing PH have been identified, including cardiac malformations, pulmonary conditions, thyroid abnormalities.^26,27 The specific etiology is likely related to challenges on the pulmonary vascular system likely related to abnormal lung development, intrinsic endothelial dysfunction, increases in pulmonary vascular resistance, and increases in pulmonary venous pressures.²⁶ These changes could be related to overexpression of chromosome 21-related genes involved in vasoactive regulation. A recent review of the Pediatric Pulmonary Hypertension Network Registry, including 158 pediatric patients with DS, reported similar survival rates to those without DS-associated PH.²⁷ Although one-half of the infants with DS were born prematurely, this study did not differentiate degrees of prematurity, and median gestational age was 37 weeks. Thus, it is possible that, with increasing prematurity, outcomes may differ.

Neonates with DS are at risk of multiple pulmonary comorbidities that can necessitate invasive support, including upper airway problems, lower airway disease, including respiratory distress syndrome, as well as vascular and lymphatic defects.²⁸ In our study, the extremely preterm neonates with DS did not have increased ventilator usage compared with neonates without DS in the NICU, which is overall not surprising given the high risk of intubation due to various factors for all neonates when born at this gestation.²⁹ However, the increased ventilator usage for neonates with DS compared with those without DS in the 3 more mature gestational age groups emphasizes that pulmonary-related morbidities persist in this group regardless of gestational age, as neonates born at these later gestational ages typically have higher rates of success with noninvasive respiratory support.³⁰ Although we are unable to determine reason for escalation to invasive ventilation based on our dataset, it is known that pneumonia and respiratory tract infections are more common and more severe in those with DS compared with healthy control individuals, providing one possible rationale for the increased level of support in older gestational age groups of neonates with DS compared with those without DS.³¹ 

Neonates with DS are also at risk of structural and functional gastrointestinal (GI) complications, which often lead to feeding problems and delay discharge home from the NICU.³² The rates of NEC in our study were modestly higher in the neonates with DS in the extremely preterm and very preterm groups compared with those without DS but similar in the moderate/late preterm and term groups. The pathophysiology of NEC, although not completely understood, is thought to be related to inflammation of the intestine leading to bacterial invasion, in turn leading to damage and necrosis of the bowel.³³ It is possible that the GI comorbidities of neonates with DS, such as Hirschsprung disease, put them at higher risk for NEC, as these conditions can promote higher levels of inflammation in the bowels and increase risk of bacterial invasion. The immature gut of premature infants adds another layer of risk but potentially follow a similar mechanism of inflammation along with lack of underdeveloped vascular system leading to ischemia and decreased immune function.^34,35 Additionally, baseline factors already known to be associated with higher risk of NEC, including CHD and low birth weight, are found more commonly in neonates with DS, which could also infer higher threats of NEC.³⁶ The explanation behind increased rates of GT placement in neonates with DS compared with those without DS in all gestational age groups is likely multifactorial given the multiple GI complications neonates with DS often face in the NICU and the fact that patients with DS are known to have feeding issues.^32,37–39 These data reinforces the need for further investigation of optimal nutritional support and to explore early feeding interventions for this high-risk population.

The diagnosis of DS has previously been shown to decrease the odds of ROP in very low birth weight survivors in smaller studies.⁴⁰ This study also found a negative association between DS and ROP in the extremely preterm group. It has been theorized that the extra copy of HSA21 leads to increased expression of genes that are potentially antiangiogenic or ROP-protectant, including endostatin, superoxide dismutase, and regulator of calcineurin 1.^40–42 However, the specifics have not been well demonstrated but do provide potential for screening and therapeutic targets in the future if this finding were to be confirmed in additional studies.

Across all gestational age groups, neonates with DS had a longer length of stay and higher risk of unplanned readmissions relative to neonates without DS. Previous studies describing hospitalization patterns for infants with DS have found risk of hospital admission 3 to 5 times higher in this population than the public, with upper and lower respiratory problems as one of the most common causes of unplanned readmissions.^43–45 These previous studies did not distinguish premature neonates or NICU graduates, who typically have more hospital readmissions.⁴⁶ Shimokaze et al reviewed death rate and cause for neonates with DS in a NICU in Japan over a 27-year period and found a 40% mortality rate in infants aged younger than 34 weeks, with the main causes attributed to BPD and PH.¹³ Our study found a significantly higher risk of mortality among neonates with DS compared with those without DS across all gestational groups, with the most severe difference being in the 2 most premature groups. The cause of death was unable to be determined, although it is likely multifactorial considering the previously mentioned morbidities associated with DS. Regardless of cause of death, these data may provide useful information for counseling purposes.

There are multiple limitations to our study. This was a retrospective study using administrative data with inherent shortcomings of such a design. Data were limited by what was available via the PHIS database, which included only children’s hospitals contributing to the database. Because these hospitals are usually not birth hospitals, patients were presumably transferred for specialized care; thus, not all neonates with DS are represented, and our findings may not be generalizable to all NICU populations. Maternal information was not available given the nature of the data collection. In general, severity of medical conditions could not be ascertained. Because clinical information is primarily from ICD-9/IDC-10 codes, nonbillable data are likely to not be recorded. The dataset used for this study was unable to distinguish between presence of a complete additional chromosome and partial translocations, limiting genotype-phenotype correlations. PHIS has a robust quality control procedure; however, we cannot verify that diagnoses were made and coded in a similar way across hospitals. We did not adjust for multiple testing and the results should be used to motivate future research rather than definitively conclude that DS is a causal effect of mortality or morbidity.

In conclusion, neonates with DS have increased morbidity and mortality compared with neonates without DS in the NICU. Prematurity is associated with a higher risk of adverse in-hospital outcomes for this vulnerable population across all gestational ages. Further research is needed to examine specific barriers to discharge and cause of death to identify areas for improvement in therapies and intervention.

This research was supported by the Ohio Perinatal Research Network (OPRN) at Nationwide Children’s Hospital. OPRN is supported by the Center for Perinatal Research at Nationwide Children’s Hospital. Funding is provided by the Abigail Wexner Research Institute at Nationwide Children’s Hospital (IRB10-00035). The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

ASD

atrial septal defect

BPD

bronchopulmonary dysplasia

CLABSI

central line-associated blood stream infection

CVL

central venous line

DS

Down syndrome

ECMO

extracorporeal membrane oxygenation

GI

gastrointestinal

GT

gastrostomy tube

HFNC

high-flow nasal cannula

HIE

hypoxic-ischemic encephalopathy

ICD

International Classification of Diseases

ICH

intracranial hemorrhage

NEC

necrotizing enterocolitis

NICU

neonatal intensive care unit

NO

nitric oxide

PDA

patent ductus arteriosus

PH

pulmonary hypertension

PHIS

Pediatric Health Information System

ROP

retinopathy of prematurity

SGA

small for gestational age

TPN

total parental nutrition