Mortality, In-Hospital Morbidity, Care Practices, and 2-Year Outcomes for Extremely Preterm Infants in the US, 2013-2018

Key Points

Question  Among extremely preterm infants born at US academic medical centers between 2013 and 2018, what were mortality, in-hospital morbidity, and 2-year neurodevelopmental outcomes?

Findings  In this observational study based on a prospective registry of 10 877 infants born at 22-28 weeks’ gestational age in 2013-2018 in 19 US academic medical centers, survival to discharge occurred in 78.3% and was significantly improved compared with a historical rate of 76.0% among infants born in 2008-2012. Among infants born at less than 27 weeks’ gestational age who survived to follow-up assessment at 2 years, 49.9% had been rehospitalized and severe neurodevelopmental impairment occurred in 21.2%.

Meaning  Among extremely preterm infants born at US academic medical centers from 2013 to 2018, survival to discharge significantly improved compared with infants born in 2008-2012, but among those born at less than 27 weeks’ gestational age, rehospitalization and neurodevelopmental impairment at 2 years were common.

Importance  Despite improvement during recent decades, extremely preterm infants continue to contribute disproportionately to neonatal mortality and childhood morbidity.

Objective  To review survival, in-hospital morbidities, care practices, and neurodevelopmental and functional outcomes at 22-26 months’ corrected age for extremely preterm infants.

Design, Setting, and Participants  Prospective registry for extremely preterm infants born at 19 US academic centers that are part of the Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network. The study included 10 877 infants born at 22-28 weeks’ gestational age between January 1, 2013, and December 31, 2018, including 2566 infants born before 27 weeks between January 1, 2013, and December 31, 2016, who completed follow-up assessments at 22-26 months’ corrected age. The last assessment was completed on August 13, 2019. Outcomes were compared with a similar cohort of infants born in 2008-2012 adjusting for gestational age.

Exposures  Extremely preterm birth.

Main Outcomes and Measures  Survival and 12 in-hospital morbidities were assessed, including necrotizing enterocolitis, infection, intracranial hemorrhage, retinopathy of prematurity, and bronchopulmonary dysplasia. Infants were assessed at 22-26 months’ corrected age for 12 health and functional outcomes, including neurodevelopment, cerebral palsy, vision, hearing, rehospitalizations, and need for assistive devices.

Results  The 10 877 infants were 49.0% female and 51.0% male; 78.3% (8495/10848) survived to discharge, an increase from 76.0% in 2008-2012 (adjusted difference, 2.0%; 95% CI, 1.0%-2.9%). Survival to discharge was 10.9% (60/549) for live-born infants at 22 weeks and 94.0% (2267/2412) at 28 weeks. Survival among actively treated infants was 30.0% (60/200) at 22 weeks and 55.8% (535/958) at 23 weeks. All in-hospital morbidities were more likely among infants born at earlier gestational ages. Overall, 8.9% (890/9956) of infants had necrotizing enterocolitis, 2.4% (238/9957) had early-onset infection, 19.9% (1911/9610) had late-onset infection, 14.3% (1386/9705) had severe intracranial hemorrhage, 12.8% (1099/8585) had severe retinopathy of prematurity, and 8.0% (666/8305) had severe bronchopulmonary dysplasia. Among 2930 surviving infants with gestational ages of 22-26 weeks eligible for follow-up, 2566 (87.6%) were examined. By 2-year follow-up, 8.4% (214/2555) of children had moderate to severe cerebral palsy, 1.5% (38/2555) had bilateral blindness, 2.5% (64/2527) required hearing aids or cochlear implants, 49.9% (1277/2561) had been rehospitalized, and 15.4% (393/2560) required mobility aids or other supportive devices. Among 2458 fully evaluated infants, 48.7% (1198/2458) had no or mild neurodevelopmental impairment at follow-up, 29.3% (709/2419) had moderate neurodevelopmental impairment, and 21.2% (512/2419) had severe neurodevelopmental impairment.

Conclusions and Relevance  Among extremely preterm infants born in 2013-2018 and treated at 19 US academic medical centers, 78.3% survived to discharge, a significantly higher rate than for infants born in 2008-2012. Among infants born at less than 27 weeks’ gestational age, rehospitalization and neurodevelopmental impairment were common at 2 years of age.

Mortality and most short-term morbidities among extremely preterm infants have declined in recent decades,^1-5 but ongoing review of medical and neurodevelopmental outcomes is needed as care strategies and technologies continue to evolve. Published reports of trends in neurodevelopmental outcomes among extremely preterm infants who survive differ, highlighting challenges to interpretation due to variation in age at assessment, assessment instruments, and other factors.^6-10 Moreover, additional postdischarge milestones including functional and health-related outcomes are increasingly recognized as important for children born extremely preterm and their families.^10-14

Quiz Ref IDThe Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network tracks in-hospital outcomes for infants born before 29 weeks’ gestational age at network centers. At 22-26 months’ corrected age,¹⁵ the neurological and developmental status of those born before 27 weeks are assessed by certified examiners. Since 1991, the network has periodically reported extremely preterm infant in-hospital mortality and morbidity outcomes and longer-term developmental outcomes. Network investigators have not previously published a comprehensive update on both short-term and postdischarge follow-up outcomes. This study was conducted to assess mortality, in-hospital morbidities, and care practices for infants born in 2013-2018 and neurodevelopmental, functional, and health-related outcomes at 22-26 months’ corrected age for infants born in 2013-2016.

The Neonatal Research Network’s extremely preterm infant registry enrolls infants born weighing 401 g to 1000 g and/or born between 22 weeks 0 days’ and 28 weeks 6 days’ gestation at network hospitals. Trained research coordinators prospectively collect maternal and neonatal data from birth until discharge home, transfer, death, or 120 days. For infants transferred or still hospitalized at 120 days, vital status is collected until 1 year of age. Surviving infants born before 27 weeks are eligible for comprehensive follow-up assessment at 22-26 months’ corrected age. The institutional review board at each participating hospital approved participation in the registry. Waiver of consent for enrollment in the extremely preterm infant registry was granted at most affiliated hospitals, but parental consent was required at 5 hospitals (4 written, 1 oral). Most hospitals required written parental consent for participation in the follow-up study, but 5 hospitals allowed participation under waiver of consent.

Infants born at each of 19 network centers between January 1, 2013, and December 31, 2018, at gestational ages of 22-28 weeks and enrolled in the registry were included in the present study. Postdischarge outcomes assessed at the 2-year follow-up visit were available for infants born between January 1, 2013, and December 31, 2016. The final in-hospital outcome was determined on December 20, 2019; the final 2-year assessment was completed on August 13, 2019.

Maternal demographic, pregnancy, and delivery information and infant characteristics were recorded. Gestational age was determined by best obstetric estimate or, if unavailable, by neonatal estimate. Maternal race and ethnicity were self-identified from a list of options; race and ethnicity were included because they are known to be associated with differences in care and outcomes in preterm infants.¹⁶ Infants with birth weights below the 10th percentile for gestational age¹⁷ were classified as small for gestational age. Clinical interventions and morbidities during the birth hospitalization were recorded for infants surviving longer than 12 hours.

Quiz Ref IDInfant survival was recorded to 12 hours and to discharge or 1 year if still hospitalized. Morbidities recorded included necrotizing enterocolitis (modified Bell stage ≥IIA^18,19), early-onset sepsis and/or meningitis (≤72 hours), late-onset sepsis and/or meningitis (>72 hours), intracranial hemorrhage (grade III or IV²⁰), cystic periventricular leukomalacia, retinopathy of prematurity, and bronchopulmonary dysplasia. Bronchopulmonary dysplasia was defined using the optimal definition (definition 15) from Jensen et al.²¹ This definition classifies bronchopulmonary dysplasia based on the mode of support received at 36 weeks’ postmenstrual age or at discharge home if earlier: no bronchopulmonary dysplasia is classified as no support or breathing room air; grade 1, nasal cannula at a flow rate of 2 L/min or lower; grade 2, nasal cannula at a flow rate higher than 2 L/min or noninvasive positive airway pressure; and grade 3, invasive mechanical ventilation. Active treatment was defined as intubation, surfactant therapy, respiratory support, chest compressions, epinephrine, volume resuscitation, blood pressure support, and/or parenteral nutrition.²²

The 2-year study visit included structured interviews with children’s primary caretakers to review medical history, rehospitalizations, and medical equipment use. A standardized neurologic examination and the Bayley Scales of Infant and Toddler Development, Third Edition (Bayley-III)²³ were administered by certified examiners who completed annual training to ensure interrater reliability.²⁴ Bayley-III cognitive, language, and motor composite scores are normalized to a mean of 100 (SD, 15), with lower scores indicating a higher degree of impairment. Severity of motor impairment was determined by the Gross Motor Function Classification System (GMFCS) of Palisano et al.²⁵ Mild cerebral palsy was defined as GMFCS level 1, moderate as GMFCS level 2 or 3, and severe as GMFCS level 4 or 5.

Mild or no neurodevelopmental impairment (NDI) was defined as a Bayley-III cognitive composite score of 85 or higher, a Bayley-III motor composite score of 85 or higher, and GMFCS level 0 or 1. Moderate NDI was defined as a Bayley-III cognitive composite score or motor composite score of 70 to 84 or GMFCS level 2 or 3. Severe NDI was defined as a Bayley-III cognitive composite score or motor composite score lower than 70, GMFCS level 4 or 5, bilateral blindness, or bilateral severe functional hearing impairment.^6,8

The following outcomes were determined at the 2-year assessment by history from caregivers and/or medical records: rehospitalization after initial discharge and reason for hospitalization; feeding assistance required; need for supportive medical equipment, such as apnea monitor, oxygen, or other respiratory support; and need for mobility aid, such as braces, walker, or adaptive stroller or wheelchair.

Descriptive statistics were used to summarize maternal characteristics, neonatal characteristics at birth, and survival for all infants; in-hospital treatments and morbidities for infants surviving more than 12 hours; and outcomes at follow-up for infants evaluated. Survival was analyzed separately for actively treated infants. Follow-up outcomes were included if infants underwent neurologic examination and/or Bayley-III assessment at 18-30 months’ corrected age, allowing for 4 months outside the targeted age.

Regression models were used to assess whether outcomes differed across gestational ages; reported P values were adjusted for the covariates listed below. Logistic, generalized logit, and linear models were used to assess whether binary proportions, categorical outcomes with more than 2 levels, or means differed by gestational age, with statistical significance determined by Wald χ² or F tests. Quantile regression was used to examine differences in medians across gestational ages for outcomes with highly skewed distributions, with statistical significance determined by a likelihood ratio χ² test. Models assessing survival, morbidities, and treatments included gestational age, study center, small for gestational age, sex, multiple gestation, and maternal race and ethnicity. Covariates known to affect outcomes were chosen a priori and entered in models simultaneously. Study center was entered as a fixed effect in all models. Because of the role of socioeconomic factors in postdischarge child development, maternal education was added to models examining outcomes at follow-up and was entered with a category for missing/unknown data to minimize loss of observations from the models. Each other covariate was missing for less than 1% of the cohort. In total, less than 2% of observations were excluded from most models (maximum of 6% from 2 models) due to missing covariates or outcomes. Unless otherwise indicated, reported P values indicate whether there were differences in outcomes by gestational age overall with all gestational age group comparisons considered together. Tests of overall differences by gestational age are reported in the text and in table footnotes. Comparisons between 2 proportions at different gestational ages were reported as risk (prevalence) differences with 95% Miettinen-Nurminen confidence intervals and P values by χ² test indicating statistical significance for the pairwise comparison. P values, all 2-sided, were adjusted as described above. A P < .05 was considered significant. Analyses were performed using SAS version 9.4 (SAS Institute Inc). Because of the potential for type I error due to multiple comparisons, findings for the analyses should be interpreted as exploratory.

Percentages in the current cohort were compared with percentages reported by Stoll et al¹ for births at 22-28 weeks’ gestational age in Neonatal Research Network centers during 2008-2012, when outcomes were reported for the earlier cohort. Risk differences adjusted for gestational age and 95% confidence intervals were estimated for overall proportions using binomial models.²⁶ Differences in proportions for infants in 1 gestational age group were unadjusted. The outcome definitions were the same in the 2 studies with 2 exceptions. First, in the current study, no assumption was made when an infant’s final vital status was unknown, while in the study by Stoll et al,¹ these infants were assumed to have survived to discharge (44 infants [0.5%]). Second, early- and late-onset sepsis and/or meningitis included sepsis and/or meningitis in the current cohort, but the reported proportions in the cohort of Stoll et al¹ did not include meningitis; this difference had minimal impact because few infants had meningitis alone.

We studied 10 877 infants born between January 1, 2013, and December 31, 2018, at Neonatal Research Network study centers with gestational ages of 22-28 weeks. The median maternal age was 29 (IQR, 24-33) years; 26.5% (n = 2886) of infants were multiples, 49.0% (n = 5327) were female and 51.0% (n = 5550) were male, 9.3% (n = 1016) were small for gestational age, and 5.5% (603/10 868) had major birth defects (Table 1). Overall, 88.1% (9571/10 867) of mothers received corticosteroids, 75.0% (8135/10 847) received antibiotics, and 79.4% (8622/10 858) received magnesium sulfate. Although 64.8% (7043/10 867) of infants were born by cesarean delivery overall, only 3.1% (17/549) of infants at 22 weeks and 39.7% (430/1083) at 23 weeks were born by cesarean delivery.

Overall, 9966 (91.6%) infants survived more than 12 hours and 8495 of 10 848 (78.3%) survived to discharge or to 1 year if still hospitalized (Table 2; eFigure 1 in Supplement 1). Survival to discharge increased significantly with gestational age (P < .001) (Table 2; eFigure 2 in Supplement 1). Of live-born infants at 22 weeks, 10.9% (60/549) survived compared with 6.6% (22/334) in 2008-2012 (difference, 4.3%; 95% CI, 0.6%-8.1%; P = .03); at 23 weeks, survival increased significantly to 49.4% (535/1083) from 32.3% (252/779) in 2008-2012 (difference, 17.1%; 95% CI, 12.6%-21.5%; P < .001).¹ Survival at 28 weeks was 94.0% (2267/2412). Only 36.5% of infants born at 22 weeks’ gestational age were actively treated, but 88.5% born at 23 weeks, 97.9% born at 24 weeks, and more than 99% born at 25-28 weeks received active treatment. Although only 10.9% of infants live-born at 22 weeks survived to discharge or 1 year, 30.0% (60/200) of those actively treated survived. Among those born at 23 weeks, 49.4% of all live-born infants and 55.8% (535/958) of actively treated infants survived.

Delayed umbilical cord clamping or cord milking was documented for 40.6% (1982/4881) of births (Table 3). Treatments including intubation/ventilation, nutrition/hydration, and/or medication were withheld, limited, or withdrawn for 10.1% (1005/9956) of infants at some time after 12 hours of age.

A total of 84.3% (8388/9951) of infants who survived more than 12 hours underwent mechanical ventilation during the hospitalization, and 79.8% received surfactant; these proportions varied significantly with gestational age (P < .001 for each). Overall, 18.6% (1791/9624) of infants received postnatal corticosteroids for bronchopulmonary dysplasia, 10.1% (1009/9959) of infants received inhaled nitric oxide, and 18.4% (1799/9796) received vitamin A; proportions varied significantly by gestational age (P < .001 for each). Medication was given to 23.4% (2331/9954) of infants for closure of a patent ductus arteriosus, and 7.4% (741/9953) had surgery or cardiac catheterization for patent ductus arteriosus closure. Overall, 78.4% (7807/9960) of infants underwent transfusion with red blood cells; more than 95% of infants born at 22-25 weeks’ gestational age had transfusion, as did 49.3% (7807/9960) of those born at 28 weeks. Probiotics were received by 9.6% (443/4630) of infants, but use varied by center; no infants received probiotics at 12 centers, less than 6% received probiotics at 4 centers, and more than 30% received probiotics at 3 centers. Overall, 91.7% (9019/9830) of infants received human milk in the first 28 days of life. Parenteral nutrition was received for a median of 17 (IQR, 11-30) days. Infants remained in the hospital for a median of 88 (IQR, 63-117) days; those born at 23-25 weeks’ gestation had median hospital stays of 104 to 121 days. The median length of hospitalization for all surviving infants was 93 (IQR, 72-122) days.

The proportion of infants diagnosed as having each morbidity during the birth hospitalization varied significantly by gestational age (P < .001 for each) (Table 3). Necrotizing enterocolitis was diagnosed in 8.9% (890/9956) of infants, and 3.9% (386/9956) received surgery. Early-onset sepsis and/or meningitis was diagnosed in 2.4% (238/9957) of infants. Late-onset sepsis and/or meningitis was diagnosed in 19.9% (1911/9610) of infants surviving more than 72 hours. The overall incidence of severe intracranial hemorrhage was 14.3% (1386/9705); severe intracranial hemorrhage was diagnosed in 38.2% (52/136) of infants born at 22 weeks compared with 5.3% (123/2342) born at 28 weeks (difference, 33.0%; 95% CI, 25.2%-41.4%; P < .001). Cystic periventricular leukomalacia was diagnosed in 4.6% (444/9731) of infants. The overall incidence of retinopathy of prematurity was 54.9% (4713/8585); 12.8% (1099/8585) had severe retinopathy of prematurity (stage ≥3). Retinopathy of prematurity was diagnosed in 92.9% (65/70) of infants born at 22 weeks compared with 29.8% (660/2216) of those born at 28 weeks (difference, 63.1%; 95% CI, 54.4%-67.7%; P < .001). Of the infants who survived to 36 weeks’ postmenstrual age, 49.8% (4307/8641) had bronchopulmonary dysplasia based on continued oxygen use, including more than 75% of infants of 22-24 weeks’ gestational age. Based on the mode of support received at 36 weeks’ postmenstrual age,²¹ 60.7% (5045/8305) of infants had bronchopulmonary dysplasia of grade 1 or higher and 8.0% (666/8305) had grade 3. The distribution of bronchopulmonary dysplasia grade differed by gestational age (P < .001). Infants received respiratory support for a median of 41 days, decreasing significantly from 93 days at 22 weeks to 17 days at 28 weeks.

Compared with the 2008-2012 cohort reported by Stoll et al,¹ the incidence of necrotizing enterocolitis in the present cohort was significantly lower (8.9% vs 10.3%; adjusted difference, −1.6%; 95% CI, −2.5% to −0.8%), as was the incidence of late-onset sepsis and/or meningitis (19.9% vs 24.4%; adjusted difference, −4.9%; 95% CI, −6.0% to −3.7%), severe intracranial hemorrhage (14.3% vs 14.6%; adjusted difference, −1.0%; 95% CI, −1.9% to −0.1%), and retinopathy of prematurity (54.9% vs 56.3%; adjusted difference, −2.1; 95% CI, −3.5% to −0.7%). The incidence of bronchopulmonary dysplasia (supplemental oxygen use at 36 weeks’ postmenstrual age) was significantly higher (49.8% vs 44.7%; adjusted difference, 4.3%; 95% CI, 2.8%-5.8%) in the present cohort (Table 3). There was no significant difference between the current cohort and the 2008-2012 cohort in the incidence of early-onset sepsis and/or meningitis, cystic periventricular leukomalacia, or severe retinopathy of prematurity.

Among 2930 surviving infants born in 2013-2016 at gestational ages of 22-26 weeks who were eligible for follow-up, 2566 (87.6%) were seen at visits between January 22, 2015, and August 13, 2019, and were included in analysis of outcomes (eFigure 1 in Supplement 1); 2458 infants (83.9%) had all information needed to complete the NDI evaluation. Mothers of children seen at follow-up had an older median age and were more likely to have a college degree and to be of Black or White race compared with those lost to follow-up, and more children seen at follow-up were born at lower gestational ages and with lower birth weights (eTable 1 in Supplement 1).

A total of 48.7% (1198/2458) of the children evaluated had no or mild NDI, 29.3% (709/2419) had moderate NDI, and 21.2% (512/2419) had severe NDI (Figure, A; eTable 2 in Supplement 1). For infants born at 23-26 weeks’ gestation, the proportion with no or mild NDI increased significantly from 31.0% (88/284) at 23 weeks to 58.1% (518/892) at 26 weeks (difference, 27.1%; 95% CI, 20.6%-33.2%; P < .001). Of 29 infants born at 22 weeks’ gestation and evaluated for NDI, 13 (44.8%) had no or mild NDI, 7 (24.1%) had moderate NDI, and 9 (31.0%) had severe NDI. eFigure 3 in Supplement 1 shows a complete accounting of all infants who were eligible for follow-up assessment, including those who died before assessment or were lost to follow-up. The proportions of children who died or who survived with varying degrees of NDI are shown for all children who died or were evaluated (eFigure 4 in Supplement 1) and for the subset of children who were actively treated at birth (Figure, B).

Mean Bayley-III cognitive, language, and motor composite scores differed significantly overall by gestational age (P < .001 for each) (Table 4; eFigures 5-7 in Supplement 1). The proportion of children with cognitive composite scores of less than 70 decreased from 25.8% (8/31) for those born at 22 weeks to 8.3% (75/905) for those born at 26 weeks (difference, −17.5%; 95% CI, −35.1% to −5.2%; P < .001). Twenty-five percent (607/2431) of children had Bayley-III language composite scores of less than 70, with higher proportions among those born at 23 and 24 weeks. Seven of 30 children (23.3%) born at 22 weeks had a motor composite score of less than 70, while the proportions were 29.1% (81/278) for those born at 23 weeks and 9.6% (85/882) for those born at 26 weeks. Among children born at 22, 23, or 24 weeks’ gestational age, mean cognitive, language, and motor scores were significantly lower for those born at 23 weeks compared with those born at 24 weeks (P < .01 for each), but statistically significant differences were not found between scores for children born at 22 weeks’ gestational age compared with those born at 23 or 24 weeks (Table 4; eFigures 5-7 in Supplement 1). The proportion of children with language and motor scores of less than 70 was significantly higher for children born at 23 weeks’ gestational age compared with those born at 24 weeks (P = .006 and P = .03, respectively), but the proportion with cognitive scores of less than 70 did not differ among children born at 22, 23, or 24 weeks (P = .20).

The proportion of children with cerebral palsy of any severity was 17.9% (457/2555) and decreased significantly from 48.4% (15/31) among those born at 22 weeks to 11.9% (112/938) among those born at 26 weeks (difference, −36.5%; 95% CI, −53.4% to −19.9%; P < .001) (Table 4). Most of the affected children had mild cerebral palsy, but 2.1% to 9.7% had severe cerebral palsy, varying inversely with gestational age. Overall, 1.5% (38/2555) of children had blindness in both eyes; 2.5% (64/2527) used hearing aids or cochlear implants.

A total of 49.9% (1277/2561) of children were rehospitalized after initial discharge. The 1277 children who were rehospitalized had 2980 admissions. Respiratory disease was the most common reason for readmission (54.0%), followed by surgery (15.4%) and infection (9.3%) (eTable 3 in Supplement 1).

By 22-26 months’ corrected age, 81.6% (2090/2561) of children were feeding themselves independently; this proportion varied significantly with gestational age (P < .001) (Table 4). Oral feeding requiring more than occasional assistance was reported for 7.3% (187/2561) of children. Limited oral feeding requiring some food by an alternate route or no oral feeding was reported for 11.1% (284/2561).

At 22-26 months’ corrected age, 14.9% (382/2561) of children were using an apnea monitor, oxygen, ventilator or continuous positive airway pressure, feeding tube, tracheostomy, and/or pulse oximeter, with the proportion significantly decreasing from 22.6% (7/31) of those born at 22 weeks’ gestational age to 10.5% (99/941) of those born at 26 weeks (difference, −12.1%; 95% CI, −29.4% to −0.6%; P < .001) (Table 4). The most common assistive device used at 2 years’ corrected age (11.5% [294/2561]) was a feeding tube (including gastrostomy). Oxygen was used by 5.3% (135/2561) of infants, and 1.7% (43/2561) required a ventilator or continuous positive airway pressure. In addition, 15.4% (393/2560) of children required use of a mobility aid or other supportive equipment.

Quiz Ref IDAmong extremely preterm infants born in 2013-2018 and treated at 19 US academic medical centers, 78.3% survived to discharge, a significantly higher rate than for infants born in 2008-2012.¹ Rates of mortality and in-hospital morbidities were higher among those born at earlier gestational ages. Among infants born at less than 27 weeks’ gestational age, rehospitalization and severe NDI were common at 2 years.

With improving survival of children born extremely preterm, there is increasing focus on broader outcomes including neurodevelopment. Although it is important to evaluate these findings with respect to other reports, comparing neurodevelopmental outcome studies is fraught with significant challenges. Cohort inclusion criteria including gestational age, birth years, and other factors may differ, as may assessment instruments or versions of instruments. Similarly, clinical and sociodemographic factors and definitions of NDI vary by study, further confounding comparisons.²⁷

Such factors may result in substantial differences in reported rates of NDI and severe NDI. Nevertheless, these findings are similar to those reported by Synnes et al.⁷ Their report from a more mature cohort (up to 29 weeks’ gestational age) reported slightly lower overall NDI and severe NDI rates; however, rates of severe NDI at 24 weeks (about 30%), 25 weeks (about 20%), and 26 weeks (about 18%) were consistent with our findings. The EPICure 2 study followed up children born before 26 weeks’ gestational age in the United Kingdom in 2006 to age 2.5 to 3 years and reported a 19% severe NDI rate but used a definition restricted to developmental quotients more than 3 SDs below the mean.¹⁰ The Swedish EXPRESS study followed up children born before 27 weeks’ gestational age in 2004-2007, reporting that 58% had some level of disability at 30 months’ corrected age. Only 11% were reported to have severe disability, but the definition used included Bayley-III composite cognitive, language, or motor scores more than 3 SDs below the mean.²⁸ Kono et al⁴ presented 3-year outcomes from the Neonatal Research Network of Japan for children born before 25 weeks’ gestational age in 2008-2012. They reported an NDI rate of 38.1% among those with neurodevelopmental follow-up using a developmental test standardized for Japanese children, but the follow-up rate was only 60.6%.

Quiz Ref IDComposite end points such as NDI and disability at toddler age are commonly used outcomes for clinical studies in neonatology. Contemporary early neurodevelopmental outcome data are also used in parent counseling at the time of threatened extremely preterm delivery and in the neonatal intensive care unit (NICU) during discussions regarding long-term transitions. However, these composite outcomes may be confusing to parents, conflate multiple end points of varying value to parents, and frame discussions in a negative way.^12,29,30 Parents may be more interested in understanding “real-life” outcomes to develop a personalized and functional picture for the future, yet these outcomes are inconsistently reported. Information regarding post-NICU rehospitalizations, feeding, equipment, and medication needs may inform both complex home care preparations and health-related quality-of-life considerations and thus may be more meaningful for families.^14,31 Hospitalization after NICU discharge was common, with nearly half of all children being rehospitalized at least once by 22 to 26 months, consistent with previous reports.^32-34Quiz Ref ID Profound functional neurosensory impairments were rare for the cohort overall, including blindness and severe functional hearing impairment. The rate of severe cerebral palsy was also low.

The overall survival for infants of 22-28 weeks’ gestational age has continued to increase since 2008-2012,¹ even though more infants born at 22 weeks were being actively treated in this more recent cohort. The survival of infants live-born at 22 weeks is similar to more recently reported survival among infants born at 22 weeks from the Vermont Oxford Network.³⁵ Higher survival for infants born at 22 weeks has been reported from Japan⁴ and Sweden.⁵

Strengths of this study include the large number of infants with high follow-up rates and rigorous prospective data collection including neurodevelopmental assessments by certified evaluators. This study reports survival both for all live-born infants and for only those actively treated. In addition, it reports often overlooked 2-year neurodevelopmental and functional outcomes of importance to families.^29,30

This study has several limitations. First, the Neonatal Research Network cohort is hospital-based rather than population-based. Although large and diverse, both in terms of geography and sociodemographics, this cohort does not represent the entire US preterm population but, rather, a selected population from academic medical centers. Second, the 2-year end points presented provide only a limited window into the future. Individual prediction of later childhood outcomes at 2 years has limitations.³⁶ The complexity of developmental and cognitive milestones increases throughout childhood and adolescence, and engagement of both parents and children in health-related quality-of-life and other outcome assessments should continue through later childhood and into adulthood.^14,31,37

Among extremely preterm infants born in 2013-2018 and treated at 19 US academic medical centers, 78.3% survived to discharge, a significantly higher rate than for infants born in 2008-2012. Among infants born at less than 27 weeks’ gestational age, rehospitalization and severe neurodevelopmental impairment were common at 2 years.

Corresponding Author: Edward F. Bell, MD, Department of Pediatrics, University of Iowa, 200 Hawkins Dr, Iowa City, IA 52242 (edward-bell@uiowa.edu).

Correction: This article was corrected on June 7, 2022, for a technical problem in Supplement 2.

Accepted for Publication: December 10, 2021.

Author Contributions: Ms Hansen and Dr Das had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Bell, Hintz, Vohr, Stoll, Rysavy.

Acquisition, analysis, or interpretation of data: Bell, Hintz, Hansen, Bann, Wyckoff, DeMauro, Walsh, Stoll, Carlo, Van Meurs, Rysavy, Patel, Merhar, Sánchez, Laptook, Hibbs, Cotten, D’Angio, Winter, Fuller, Das.

Drafting of the manuscript: Bell, Hintz, Vohr.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Hansen, Bann, Das.

Obtained funding: Bell, Wyckoff, Walsh, Stoll, Van Meurs, Patel, Merhar, Sánchez, Hibbs, D’Angio, Das.

Administrative, technical, or material support: Hintz, Wyckoff, Stoll, Carlo, Van Meurs, Sánchez, Hibbs.

Supervision: Bell, Hintz, Wyckoff, Van Meurs, Sánchez, Hibbs, Cotten, D’Angio, Das.

Conflict of Interest Disclosures: Drs Bell, Hintz, Walsh, Vohr, Rysavy, Merhar, Laptook, Hibbs, and Fuller reported receiving grant(s) from the National Institutes of Health (NIH) outside the submitted work. Dr DeMauro reported receiving grants from the NIH outside the submitted work and a grant from the Thrasher Research Fund. Dr Carlo reported receiving grants from the NIH outside the submitted work and grants from the Thrasher Research Fund, the University of Virginia, the Hudson Alpha Institute for Biotechnology, the Foundation for the National Institutes of Health, and the Gates Foundation. Dr Van Meurs reported receiving grants from the NIH outside the submitted work and grants from the Thrasher Foundation and serving on a scientific advisory board for Greenwich Biosciences. Dr Patel reported serving on the data monitoring committee for a trial by Infant Bacterial Therapeutics/Premier Research. Dr Sánchez reported receiving grants from Merck and AstraZeneca (MedImmune). Dr Cotten reported receiving grants from the NIH outside the submitted work and a grant from the Robertson Foundation. No other disclosures were reported.

Funding/Support: The NIH, the Eunice Kennedy Shriver National Institute of Child Health and Human Development, the National Center for Research Resources, and the National Center for Advancing Translational Sciences provided grant support for the Neonatal Research Network Generic Database and Follow-up studies to maintain high-risk preterm infant registries through the following cooperative agreements: U10 HD021373, UG1 HD021364, UG1 HD021385, UG1 HD027851, UG1 HD027853, UG1 HD027856, UG1 HD027880, UG1 HD027904, UG1 HD034216, UG1 HD036790, UG1 HD040492, UG1 HD040689, UG1 HD053089, UG1 HD053109, UG1 HD068244, UG1 HD068263, UG1 HD068270, UG1 HD068278, UG1 HD068284, UG1 HD087226, UG1 HD087229, UL1 TR000006, UL1 TR000041, UL1 TR000042, UL1 TR000077, UL1 TR000093, UL1 TR000105, UL1 TR000442, UL1 TR000454, and UL1 TR001117.

Role of the Funder/Sponsor: Eunice Kennedy Shriver National Institute of Child Health staff had input into the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review or approval of the manuscript; and decision to submit the manuscript for publication.

Group Information: The members of the Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network appear in Supplement 2.

Disclaimer: The content of the article is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Additional Contributions: Participating Neonatal Research Network sites collected data and transmitted them to RTI International, the data coordinating center for the network, which stored, managed, and analyzed the data for this study. We are indebted to our medical and nursing colleagues and the infants and their parents who agreed to take part in this study.