Outcomes in early term neonates requiring extracorporeal membrane oxygenation

Introduction

Gestational age is a key determinant of survival and developmental outcomes in newborns. Accurately classifying gestational age has important implications for obstetric and neonatal management, as well as for public understanding of the optimal time for delivery [1]. Historically, term gestation has been defined as 37^0/7 to 41^6/7 weeks [2], and it is well established that infants born prematurely, at less than 37 weeks of gestation, have worse overall health outcomes [3]. There is increasing evidence from population health research that infants born at the earlier end within term period, referred to as early term (ET), are also at significant risk for lower survival, impaired growth, and worse neurodevelopmental outcomes than counterparts born at the later end within term period [4], [5], [6], [7], [8], [9]. Recognizing this, in 2013 (and reaffirmed in 2022) the American College of Obstetricians and Gynecologists (ACOG) suggested subclassifying term gestation into ET, defined as 37^0/7 to 38^6/7 weeks; full term (FT), defined as 39^0/7 to 40^6/7 weeks; and late term, defined as 41^0/7 to 41^6/7 weeks of gestation [10]. This subclassification is useful when counseling families, managing patients, and planning future research. This is increasingly relevant considering that the number of ET deliveries has risen significantly in recent years. Data from the National Vital Statistics System published in January 2024 shows there was a 20 % increase in ET deliveries in the United States, increasing from 24 % in 2014 to 29 % in 2022. There was additionally a 12 % increase in premature deliveries and a 6 % decrease in FT deliveries during this time [11]. After a decade of decline [12], this increase in premature and ET births and poor outcomes in ET infants compared to FT counterparts is an alarming and expensive public health issue. While the adverse outcomes associated with ET birth are well documented in general neonatal populations [4], [5], [6], [7], [8], [9], the implications for high-risk interventions such as extracorporeal membrane oxygenation (ECMO) in this group are unclear particularly when compared to FT counterparts.

There are favorable ECMO outcomes in neonates with respiratory conditions like meconium aspiration syndrome (MAS) and/or persistent pulmonary hypertension of the newborn (PPHN), with reported survival around 83 % for neonates undergoing ECMO for respiratory indications as per the latest summary report from the Extracorporeal Life Support Organization (ELSO) Registry [13]. Though it is widely accepted that infants with MAS and/or PPHN have very favorable outcomes among neonates requiring ECMO, it is unknown how ET infants fare compared to FT infants. In 2011 Ramachandrappa et al. published a review of ELSO Registry data from 1986 to 2006 comparing mortality in premature, ET, and ‘term’ infants on ECMO for respiratory indications. They found ECMO mortality to be the highest in premature infants and higher in ET infants than the ‘term’ group [14]. However, the authors defined ‘term’ as 39^0/7 to 42^6/7 weeks gestation which differs from the currently recommended ACOG subclassification of term gestation which was put out in 2013 and reaffirmed in 2022 [10]. Additionally, since the Ramachandrappa et al. publication, medical management of MAS and PPHN has improved with more common use of surfactant, inhaled nitric oxide, and alternative ventilator strategies such as high frequency oscillator use in eligible infants [15]. Given this, we carried out a retrospective comparative study of 10 years of ELSO Registry data with the aim of characterizing ECMO survival and outcomes between ET and FT neonates requiring ECMO for MAS and/or PPHN using the modern gestational age subclassification from ACOG in a post-surfactant era ECMO cohort. We hypothesized that ET infants would have lower ECMO survival and higher complication rates compared to FT infants, despite improvements in neonatal care over the past decade.

Methods

The ELSO Registry was queried for all neonates who received ECMO for MAS and/or PPHN between 2007 and 2017. We analyzed respiratory ECMO runs for all newborns less than 28 days old who required ECMO with a primary diagnosis of MAS and/or PPHN based on select ICD-9 (770.1, 763.84, 777.1, 770.11, 770.12, 747.89, 747.83, 416.0) and ICD-10 (P24.9, P03.82, P03.82, P24.00, P24.01, Q28.8, P29.3, I27.0) codes in this retrospective cohort. Gestational age was the primary grouping variable: neonates born at 37^0/7 to 38^6/7 weeks were classified as ET and those born at 39^0/7 to 40^6/7 weeks were classified as FT in accordance with the current ACOG recommendation for the subclassification of term gestational age [10]. In order to minimize heterogeneity in the sample, patients were excluded if they were born at gestational ages outside of these parameters, received ECMO for cardiac causes or a principal respiratory diagnosis aside from MAS and/or PPHN, received extracorporeal cardiopulmonary resuscitation, or required more than one ECMO run or a change in mode of ECMO. These exclusion criteria were applied to create a homogenous cohort focused solely on term-equivalent neonates undergoing ECMO for MAS and/or PPHN. The primary outcomes were ECMO survival and survival to hospital discharge. ECMO survival was defined as discontinuation of ECMO due to patient improvement with expected recovery per the ELSO Registry definitions. Survival to discharge was defined as discharged from the hospital alive per the ESLO Registry definitions. Secondary outcomes were complications on ECMO including mechanical, hemorrhagic, cardiovascular, pulmonary, metabolic, infectious, renal, or limb complications as coded by the ELSO Registry definitions (Appendix) [16]. Data were analyzed using Mann-Whitney U testing for continuous variables and Fisher’s Exact test for categorical variables. Multivariate logistic regression was performed to assess odds of ECMO survival in the MAS/PPHN cohort after adjusting for covariates that differed significantly between ET and FT groups. Odds ratios with 95 % confidence intervals were calculated. Model performance was evaluated using Nagelkerke R² to estimate explained variance. Statistical significance was assumed at p<0.05. Data analyses were performed using the IBM Statistical Package for the Social Sciences (IBM SPSS 25.0, IBM, Armonk, NY). This study did not require review by the NYU Langone Health Institutional Review Board per institutional policy regarding analysis of deidentified data in conjunction with the ELSO Data Use Agreement which authorizes the sending and receiving of deidentified data sets between institutions [17]. This study was carried out in accordance with the Declaration of Helsinki.

Results

A total of 2,551 infants met inclusion criteria. Of these, 805 (32 %) were ET and 1,746 (68 %) were FT. The ET group had significantly lower birth weight and were more likely to be of white race. The ET group had a slightly lower 1-min Apgar score and a slightly higher 5-min Apgar score than the FT group (Table 1). The ET group received less neuromuscular blockade and surfactant administration, had higher age at cannulation, and lower mean airway pressure (Table 2). No other significant differences were noted between groups in baseline characteristics or pre-ECMO clinical variables which are summarized in Tables 1 and 2. Regarding the overall time course, the ET infants had longer time between intubation and ECMO start, and between discontinuation of ECMO and extubation (Table 3).

For the primary outcomes, ET infants demonstrated significantly lower ECMO survival (90 vs. 94 %, p<0.01) and survival to discharge (80 vs. 88 %, p<0.01) compared to FT infants. Among secondary outcomes, ET infants had significantly higher rates of neurologic complications on ECMO (15 vs. 12 %, p=0.024), and increased need for hemofiltration (33 vs. 29 %, p=0.033). There were no significant differences observed in rates of mechanical, hemorrhagic, cardiovascular, pulmonary, metabolic, infectious, renal, or limb complications on ECMO between groups (Table 4).

We constructed a multivariate regression model to identify factors that are independently associated with odds of ECMO survival. Variables noted to be both clinically relevant and statistically different between ET and FT groups in univariate analysis were selected as covariates for the multivariate regression model. Covariates included in the model are gestational age classification (ET and FT), 5-min Apgar score, age at ECMO cannulation, use of neuromuscular blockade, mean airway pressures, neurologic complications, and need for hemofiltration. Our multivariate logistic regression model showed that ET gestational age was an independent negative predictor of ECMO survival (OR=0.65, 95 % CI: 0.47–0.90, p<0.01). Similarly, the need for hemofiltration or development of neurologic complications on ECMO, regardless of gestational age, were independently associated with decreased odds of ECMO survival (OR=0.42, 95 % CI: 0.31–0.58, p<0.001 and (OR=0.19, 95 % CI: 0.13–0.26, p<0.001) respectively (Table 5). All other covariates in the model were not found to independently predict odds of ECMO survival. The model explained 12 % of the variance in ECMO survival (Nagelkerke R²=0.12).

Discussion

This is the first study presenting a novel, focused analysis of the ELSO Registry comparing ECMO morbidity and survival outcomes between ET and FT newborns with MAS and/or PPHN in a post-surfactant era cohort using the most current ACOG gestational age definitions. We found ET infants had significantly lower ECMO survival and survival to discharge as well as higher rates of neurologic complications and the need for hemofiltration compared to FT infants. These findings highlight a previously underrecognized vulnerability among ET neonates undergoing ECMO. Further, our results are aligned with the growing body of literature showing poorer general outcomes in ET infants [4], [5], [6], [7], [8], [9] and extends the evidence to critically ill neonates receiving ECMO.

Importantly, our multivariate logistic regression model shows that ET gestational age is an independent driver of worse ECMO survival compared to FT gestational age. The model also shows that the need for hemofiltration and the development of neurologic complications on ECMO are both independently associated with worse ECMO survival regardless of gestational age. Although the model shows that requiring hemofiltration or developing neurologic complications on ECMO are each more significant predictors of poor ECMO survival than ET gestational age alone, ET gestational age is still an important negative predictor of ECMO survival. This reflects that ET gestational age is a biologically less resilient state that is particularly vulnerable when there is neonatal respiratory failure. As the number of ET births continue to rise in the United States [11], our study underscores the need to carefully assess the timing of delivery as well as consider gestational maturity as a risk factor in ECMO risk stratification and counseling.

The noted higher incidence of neurologic complications on ECMO in ET infants may be explained by their known central nervous system immaturity and the increased risk of intracranial hemorrhage with decreasing gestational age [18], 19]. Compared to the FT group, the ET group did have a significantly lower mean birth weight (by approximately 300 g) which is expected as birth weight is generally directly correlated with gestational age, though the clinical significance of this small difference in birth weight is unclear. Additionally, the mean birth weights for both the ET and FT groups are considered appropriate for gestational age per the Fenton 2025 Third-generation Preterm Growth Charts for male and female infants [20]. This smaller size may potentially contribute to increased susceptibility to ECMO complications in ET infants, including neurologic injury and renal dysfunction requiring hemofiltration. The increased need for hemofiltration in ET infants could reflect lower baseline renal function associated with lower birth weight [21], however there was no significant difference in renal complications between groups making this less likely as the only explanation. Smaller infants are at higher risk for hemolysis on ECMO [22], however there was no significant difference between groups in metabolic complications which includes hemolysis as per ELSO definitions [16]. In general, while the difference in birth weight with varying gestational age is expected, it still remains a key contributor to risk stratification. It is unclear if the noted worse outcomes in ET infants are due to inherent risk factors associated with lower gestational age and birth weight alone or related to prenatal risk factors that led to earlier delivery. Further studies are needed to evaluate if subtle differences in maturity, body composition, or anthropometric measurements such as birth weight in early term newborns contribute meaningfully to ECMO-related morbidity and mortality.

Apart from discrepancy in birth weight, the ET group had slightly more white subjects, though there was a white race majority in both groups. The slight statistical differences in Apgar scores between ET and FT infants is likely not clinically significant. There were no other significant differences in baseline characteristics between groups, however it is important to note that the ELSO Registry contains limited maternal and prenatal details which limits our ability to report and analyze further demographic information. In reviewing pre-ECMO clinical variables, the ET group received less neuromuscular blockade and surfactant administration, had higher age at cannulation, and lower mean airway pressure compared to the FT group. This suggests that the infants in the ET group may have appeared less critically ill than those in the FT group prior to ECMO cannulation, yet the ET group suffered worse ECMO outcomes and neurologic complications. This raises questions about clinical perceptions and thresholds to administer escalating treatment modalities. The finding that more FT infants received surfactant may reflect concern for more aspiration-related surfactant inactivation in the setting of MAS in the FT group compared to the ET group. Additionally, the ET group may have had less severe parenchymal disease or different underlying etiology of respiratory failure explaining the lower rate of surfactant administration in the ET group. These findings highlight important differences in clinical management patterns. However, since the ELSO Registry database does not contain information on exclusions or decision making processes, further studies are needed to assess how clinicians’ perception and management pathways influence the decision to place an ET infant on ECMO. Compared to the FT group, the ET group was also noted to have a longer time between discontinuation of ECMO and extubation (Table 3) which suggests the infants in this group likely required further management and stabilization once off ECMO compared to FT counterparts. This is consistent with existing evidence that ET infants suffer more respiratory distress than FT counterparts [5] and with the findings of this study showing worse survival to discharge and complications on ECMO.

Notably, this study excluded neonates with other respiratory diagnoses, multiple ECMO runs, and cardiac indications, allowing us to isolate the effects of gestational maturity in a homogeneous clinical population. Previous studies often combined broader gestational categories or less specific diagnoses, limiting the clinical applicability of their findings [15]. In contrast, our study offers a more targeted and generalizable understanding of ECMO outcomes in the specific, high-survival MAS/PPHN group, using contemporary classification and management standards. Limitations of this study include that this is a retrospective review of data from the ELSO Registry database. Although the ELSO Registry represents the largest repository of ECMO data worldwide giving our study a large sample size and generalizability, our analysis is potentially limited by reporting bias and variation in data quality. We are unable to account for variability in management practices that may have occurred over the 10-year study period and across the many institutions that report to the ELSO Registry. We also are limited to analyzing ECMO complications as coded by the ELSO Registry and subclassification of respiratory diagnoses such as MAS vs. PPHN. Therefore we are unable to provide further details on the clinical significance of these complications or perform etiology specific subgroup analyses. Finally, characterizing differences in the long-term outcomes of ET and FT infants who survive ECMO would be of tremendous interest, but these data are not presently available from the ELSO Registry.

Conclusions

This ten-year retrospective analysis of the ELSO Registry revealed that ET infants with MAS and/or PPHN have worse ECMO survival and survival to discharge, and increased neurologic complications and need for hemofiltration than FT infants on ECMO. These novel findings showing ET gestational age as a clear risk factor for poor ECMO survival and outcomes adds to the emerging literature on overall worse health outcomes in ET infants compared to FT counterparts [4], [5], [6], [7], [8], [9] and extends the evidence to critically ill neonates requiring ECMO. Steps should be taken to assess the reasoning behind the increasing trend in ET delivery in the United States [11] and to minimize unnecessary ET deliveries in order to promote best neonatal health outcomes.