Neurodevelopment and Predictors of Outcomes of Children With Birth Weights of Less Than 1000 g:  1992-1995 FREE

PERINATAL TREATMENT of extremely low-birth-weight (ELBW; <1000 g) children in the 1990s has, in addition to surfactant therapy, been associated with new ventilator strategies and an increased use of pharmacological agents, including antenatal and postnatal steroids and indomethacin.^1- 3 Previous reports from our perinatal center studied the worsening early childhood outcomes of children with birth weights of less than 750 g born during 1993 through 1995, compared with previous years.^4- 5 With the exception of the National Institute of Child Health and Human Development (NICHD) Neonatal Network multicenter study of children born during 1993 through 1994,⁶ the most recent reports of the neurologic and developmental status of children with birth weights of less than 1000 g pertain to children born during 1990 through 1992.^7- 12

Because of the paucity of information on the outcomes of surviving children with birth weights of less than 1000 g born since 1992, and to have sufficient numbers of children to identify predictors of outcome, we sought to examine the early childhood cognitive and neurosensory status of all the children with birth weights of less than 1000 g treated at our perinatal center from January 1, 1992, through December 31, 1995. We furthermore sought to identify the significant correlates of outcome in this cohort.

Three hundred thirty-three ELBW infants without major congenital malformations were admitted to the neonatal intensive care unit at Rainbow Babies and Childrens Hospital, University Hospitals of Cleveland, Cleveland, Ohio, during the 4-year study period.

Two hundred forty-one children (72%) survived to 20 months' corrected age, of whom 221 (92%) underwent complete neurodevelopmental assessments at that time. Of the 20 children not included in the study population, 13 were unavailable for follow-up and 7 were not able to undergo testing (3 had neurosensory abnormalities, and 4 displayed severe behavioral difficulties and/or lack of cooperation).

A description of the maternal demographic risk factors and infant birth data is presented in Table 1. Demographic data included maternal age, race, marital status, and a composite Social Risk Score that includes marital status (married, 0; single, 1), race (white, 0; black, 1), and education (high school, 0; less than high school, 1).¹⁴ Major neonatal complications and therapies during the initial neonatal hospitalization are presented in Table 2. Chronic lung disease was defined as an oxygen dependence at 36 weeks' corrected age (postmenstrual plus postnatal age)¹⁵; episodes of sepsis (blood culture positive for sepsis plus clinical signs of sepsis); meningitis (spinal fluid culture positive for meningitis); necrotizing enterocolitis (according to Bell et al¹⁶); jaundice (maximum bilirubin level >171 µmol/L [>10 mg/dL]); and the most severe abnormal finding on cranial ultrasounds performed serially during the hospital stay (grade III or IV hemorrhage), periventricular leukomalacia (PVL), or persistent ventricular dilation at the time of discharge home.¹⁷ Apnea was defined as the need for theophylline or assisted ventilation for this condition.

At 20 months' corrected age, 24 children (11%) were in foster care, 2 (1%) remained in the chronic care facility, and the remaining 195 (88%) lived at their parent's home. A physical and neurologic examination and the revised Bayley Scales of Infant Development (BSID II) were administered at this time (mean±SD corrected age, 19.7 ± 2 months^18- 19). Major neurologic abnormalities included cerebral palsy (spastic diplegia, hemiplegia, hemidiplegia [ie, triplegia] or quadriplegia), hypertonia, hypotonia, and shunt-dependent hydrocephalus without other neurologic abnormality. Hypotonia and hypertonia were included in the category of major neurologic abnormality, as these conditions are considered by some to represent a variant of cerebral palsy. Shunt-dependent hydrocephalus without neurologic abnormality was also considered an impairment.²⁰ Sensory abnormality included unilateral or bilateral blindness or deafness.

There were no significant differences in birth data or neonatal risk factors between the 221 ELBW children followed up to 20 months' corrected age and the 20 children not undergoing testing; however, the latter group was at higher social risk (P=.04). Informed consent was obtained from the parents of the study children.

At 20 months' corrected age, 24 children (11%) were in foster care, 2 (1%) remained in the chronic care facility, and the remaining 195 (88%) lived at their parent's home. A physical and neurologic examination and the revised Bayley Scales of Infant Development (BSID II) were administered at this time (mean±SD corrected age, 19.7 ± 2 months^18- 19). Major neurologic abnormalities included cerebral palsy (spastic diplegia, hemiplegia, hemidiplegia [ie, triplegia] or quadriplegia), hypertonia, hypotonia, and shunt-dependent hydrocephalus without other neurologic abnormality. Hypotonia and hypertonia were included in the category of major neurologic abnormality, as these conditions are considered by some to represent a variant of cerebral palsy. Shunt-dependent hydrocephalus without neurologic abnormality was also considered an impairment.²⁰ Sensory abnormality included unilateral or bilateral blindness or deafness.

There were no significant differences in birth data or neonatal risk factors between the 221 ELBW children followed up to 20 months' corrected age and the 20 children not undergoing testing; however, the latter group was at higher social risk (P=.04). Informed consent was obtained from the parents of the study children.

The t test was used to compare continuous measures and the χ² test or Fisher exact test to compare categorical data. Outcomes considered included rates of neurologic and sensory abnormality, Mental (MDI) and Psychomotor Developmental Index scores on BSID II, and rates of neurodevelopmental impairment defined as neurosensory impairment and/or an MDI score of less than 70. Logistic regression was used to examine the association between sociodemographic and individual neonatal risk factors on outcomes. In these analyses, sex, social risk, and birth weight were included as covariates. In separate logistic analyses, all neonatal risk factors, as well as sex, social risk, and birth weight, were considered as predictors of outcome and were selected for a final model using forward stepwise regression.

Information concerning the neurosensory outcomes is presented in Table 3, and the results of the BSID II are presented in Table 4. Twenty percent of the children had a major neurologic abnormality and 42%, a subnormal (<70) MDI score. Overall, 105 children (48%) had neurodevelopmental impairment (defined as a subnormal MDI score, neurologic abnormality, blindness, or deafness). Sixty-one children (28%) had 1 impairment; 28 (13%), 2 impairments; 13 (6%), 3 impairments; and 3 (1%), 4 impairments.

There were no significant differences in outcomes between the 68 children (31%) who weighed from 500 to 749 g at birth and the 153 (69%) who weighed from 750 to 999 g. Fifty-eight (38%) of the 750-999-g birth-weight group had an MDI score of less than 70 vs 34 (50%) of the group with birth weights of less than 750; 22 (14%) vs 11 (16%), respectively, had cerebral palsy; and 26 (17%) vs 18 (26%), respectively, had overall neurologic abnormality. Blindness occurred in 1 child in each group (1% each) and hearing loss in 13 (9%) vs 7 (10%) children, respectively.

There were also no differences in outcomes of the 171 children of appropriate size for gestational age and 50 small for gestational age (birth weight, <−2 SD¹³). Sixty-nine (40%) of the children of appropriate size for gestational age had an MDI score of less than 70 vs 23 (46%) of the children small for gestational age; 33 (19%) vs 11 (22%), respectively, had overall neurologic abnormality; and 17 (10%) vs 3 (6%), respectively, had deafness.

Figure 1 and Figure 2 illustrate the 100-g birth weight and gestational age–specific rates of neurodevelopmental impairment. The gestational age-specific outcomes pertain only to children with birth weights of less than 1000 g. Outcomes of the 28 children of at least 29 weeks' gestation are not presented, as the number of children at each of these ages was too small to provide meaningful results.

Multiple birth, Caesarean section, antenatal steroid therapy, and a history of amnionitis had no measurable effect on outcomes. The association of the various individual neonatal risk factors to outcomes, controlling for sex, social risk, and birth weight, are presented in Table 5. In addition to postnatal steroid therapy, controlling for chronic lung disease was predictive of a subnormal MDI score (odds ratio [OR], 2.47; 95% confidence interval [CI], 1.25-4.89) and neurologic abnormality (OR, 4.49; 95% CI, 1.66-12.14).

Results of the multiple step-wise logistic regression analysis are presented in Table 6. Significant predictors of a subnormal MDI score were male sex, social risk, and chronic lung disease; of neurologic abnormality, abnormal ultrasound findings and chronic lung disease; and of deafness, male sex, sepsis, and jaundice.

Our study indicates that ELBW children born since 1992 have high rates of neurosensory and developmental abnormality in early childhood. Neurosensory impairments prevalent in this population include cerebral palsy in 15% and deafness in 9%. As can be expected, most children with major neurosensory impairments have subnormal cognitive development. However, developmental outcomes are poor even for children free of major neurosensory impairments, of whom 62% have subnormal or borderline MDI scores on the BSID II. Although gestational age–specific outcomes improved at 28 weeks' gestation, there were very few differences in outcomes when the children were examined by 100-g birth-weight groups. Social and biological (sex and neonatal) risk factors predicted a subnormal MDI score, whereas neurologic abnormality and deafness were predicted by biological risk factors only.

Our results are in general agreement with those from the NICHD multicenter study in which 17% of the children had cerebral palsy, 11% had deafness, and 37% had subnormal cognitive scores.⁶ Of the 221 children described in our study, 102 also participated in the NICHD study. Other studies from the United States,¹⁰ Canada,¹² Sweden,⁷ Australia⁹ and New Zealand¹¹ that pertain to ELBW children born during 1990 through 1992 report better outcomes similar to those of ELBW children born during the late 1970s and 1980s, including cerebral palsy in 7% to 10% and subnormal cognitive function in 10% of children.^{10,12,21- 23}

The reason for the poorer outcomes for children born since 1992 is unclear. Lorenz et al²⁴ recently reported on the association of neurologic impairments with increased survival rates. Our survival rate of 72% at 20 months' corrected age for children born during 1992 through 1995 exceeds those of 60%, 57%, and 56% for children born during 1990 through 1992 in Canada,¹² Sweden,⁷ and Australia,⁹ respectively. Increased survival is also associated with an increase in chronic lung disease,^2,25 which together with the use of postnatal steroid therapy also might have contributed to the poorer neurodevelopmental outcomes.^26- 27 Children unavailable for follow-up are at greatest risk for neurodevelopmental problems.^28- 29 Our excellent return rate of 92% might have ensured the inclusion of such children, contributing to more negative overall outcomes.

Additional factors that might explain our outcomes are the relatively high rate of social risks for our population^30- 32 and the inclusion of outborn infants, ie, infants transported to our neonatal intensive care unit after birth.³³ Compared with the 185 children born at our perinatal center, the 36 outborn children had a higher rate of subnormal MDI score (58% vs 38%; P=.04) but similar rates of neurosensory abnormalities and overall impairment (58% vs 45%). Our tertiary urban perinatal center also includes an inner-city population with many social risk factors. The use of the BSID II since 1992 may partly explain our lower MDI scores but would not account for the relatively high rates of cerebral palsy and deafness.^18,34 Longer follow-up to school age will be necessary to confirm these findings.

Factors that had a significant effect on the 20-month outcomes in our study include male sex, social risk, and the neonatal complications of severely abnormal cerebral findings on ultrasound, chronic lung disease, sepsis, and jaundice. We did not examine the independent effects of postnatal steroid therapy on outcome, as steroid use is closely associated with chronic lung disease and its severity. Additional factors that were not examined, but which have been reported to influence outcomes, include hypoglycemia, breast-milk feeding, and neonatal hypothyroxinemia.^35- 37

The effect of sex is considered to be associated partly with the greater severity of illness in male infants.^38- 42 Severely abnormal cerebral findings on ultrasound are well-recognized predictors of neurologic abnormality.^6,43- 46 We presented the bivariate effects of the individual abnormal ultrasound findings, as they are used to counsel parents during the neonatal hospital stay (Table 5). The rates of neurodevelopmental impairment for children with a grade III or IV bleed, PVL, and persistent ventricular dilation were similar (69%, 75%, and 71%, respectively). These findings are in agreement with published reports.^{44- 45,47- 49} However, such predictions are confounded by the co-occurrence of the various abnormal ultrasound findings.⁵⁰ Nearly half of our children with PVL or persistent ventricular dilation also had a grade III or IV bleed. Our relatively small sample size did not allow for an examination of the effects of the severity and localization of the PVL or the degree of persistent ventricular dilation that may influence outcomes.⁴⁴

Chronic lung disease is a recognized predictor of poor outcomes, even when controlling for other risk factors.^6,43,51- 54 Factors contributing to these poor outcomes include periods of hypoxia, hypotension, and the detrimental effects of poor growth and prolonged hospitalization.

Our findings associated with deafness need to be interpreted with caution, since we did not routinely perform formal hearing tests; however, since deafness usually is diagnosed before 2 years of age, we doubt that any cases were missed. Sepsis is considered a marker for severe illness, which explains its high association with neurodevelopmental impairment, including deafness.³⁸ The use of aminoglycosides to treat infants with sepsis may also have contributed to deafness.^55- 57 The association between jaundice and deafness is considered to be due to bilirubin deposition in the brainstem.^58- 61 Sepsis, a risk factor for the development of toxic effects of bilirubin, may have contributed to these effects.^62- 63

The prevention of prematurity and/or the improvement of the socioenvironmental milieu in which children develop, which could improve outcomes, seems at present to be unattainable in the United States.⁶⁴ Furthermore, early enrichment programs, although beneficial, have not resulted in major improvements in the outcomes of the smallest preterm children.^65- 66 There is thus an urgent need for research into the etiology and prevention of neonatal morbidity. Ongoing surveillance is also mandatory to identify iatrogenic factors that may have deleterious effects on the outcome of ELBW children.

Accepted for publication December 20, 1999.

This study was supported in part by General Clinical Research Center grant MO1 RR00080 from the National Institutes of Health, National Center for Research Resources, Bethesda, Md.

We appreciate the help of Eloise Scott, Dee Kay Grant, and Bonnie Siner, RN, for research assistance; Nori Mercuri Minich for data analysis; and Cecily Lewis for secretarial assistance.

Corresponding author: Maureen Hack, MB, ChB, Division of Neonatology, Rainbow Babies and Childrens Hospital, 11100 Euclid Ave, Cleveland, OH 44106 (e-mail: mxh7@po.cwru.edu).