Outcomes for Extremely Low-Birth-Weight Infants: Title and subTitle BreakDisappointing News

The long-term outcome for extremely low-birth-weight (ELBW) infants—weighing less than 1000 g at birth—is of great interest to parents, the public, and a broad variety of professional groups, including educators, psychologists, health care planners, as well as obstetricians, neonatologists, pediatricians, and all other medical specialists involved in the care of these children. Studies of small premature infants born before the 1990s indicated that intensive perinatal care decreased their mortality without decreasing the proportion of survivors with disabilities. As a result, the absolute number of the survivors with disabilities increased.¹

The advances in perinatal care in the 1990s, particularly the widespread use of antenatal corticosteroids and postnatal surfactant, dramatically reduced the mortality of high-risk infants. Partly because these advances help to prevent intracranial hemorrhage and improve pulmonary function, there has been hope for a reduction in the proportion, if not the absolute number, of ELBW survivors who have disabilities or other long-term problems common in these children: chronic medical problems, functional limitations, and the need for special educational or medical services. Until now, however, there have been few well-designed studies in other countries^{2 - 3} and none in the United States reporting school-age outcomes among ELBW survivors born in the 1990s.

In this issue of JAMA, Hack and colleagues⁴ report outcomes at 8 years of age for 219 ELBW survivors admitted to Rainbow Babies and Children’s Hospital in Cleveland, Ohio, in 1992-1995. These investigators assessed 92% of the eligible ELBW survivors at this age and compared them on a broad range of measures to a control group of 176 children born at term with a normal birth weight (NBW). To the extent feasible, the ELBW children were each matched with a NBW child randomly selected from those children of the same age, race, and sex and attending the same school. In relating birth-weight group to outcome, logistic regression analyses were used to adjust for socioeconomic status (assessed using maternal education and median family income for the family Census tract). Standardized parental questionnaires and objective testing were used to assess disabilities and medical diagnoses as well as cognition, academic achievement, motor skills, social functioning, chronic conditions, and use of educational and health care services. Psychometric testing was performed by assessors blinded to birth-weight group.

The results show that the ELBW group fared substantially worse than the NBW group in every type of assessment, although the difference may be less than that expected by some obstetricians and pediatricians,^{5 - 6} other professionals, or lay persons. The most severe problems occurred almost exclusively in the ELBW group; 16% of these children had major neurosensory impairments, including cerebral palsy, deafness, and blindness; 6% required a wheelchair; 1% required tube feedings; and 6% to 10% had difficulty eating, difficulty using the toilet, or were unable to play or socialize with others. Other problems observed in both groups also occurred more commonly among ELBW infants. For every 100 children in each group, 24 more children in the ELBW group had a low IQ (<85); 13 more required treatment for asthma; 31 more had poor adaptive functioning; 43 more had a functional limitation, 34 more had a chronic condition, 25 more were dependent on special equipment, diet, or medication; and 38 more were receiving special medical, educational, or other services for chronic conditions.

The increased proportion of ELBW children with these problems is superimposed on the relatively large proportion of infants in the NBW group with the same problems. For example, 42% of the NBW group was found to have a chronic condition, a rate higher than the 15% to 30% identified in other studies using the same or a similar assessment tool. Such findings are likely to be due in part to the efforts in this study to ensure that the NBW group was as disadvantaged as the ELBW group with respect to current socioeconomic status. To the extent that the responsibilities of caring for an ELBW child reduce parental income, education, and thus socioeconomic status, the problems of ELBW children relative to NBW children may be somewhat understated. To avoid this problem, controls would ideally be selected at birth (rather than at the time of assessment). However, it would be difficult and expensive to avoid loss to follow-up among such controls in long-term studies.

Differences in the design of previous long-term studies of ELBW infants make them difficult to compare with the current article by Hack et al.⁴ However, the findings, like those in studies of younger survivors evaluated by these authors⁷ and others,⁸ suggest that the proportion of ELBW survivors with adverse outcomes has not decreased among those born most recently. Indeed, this proportion may have increased, a problem likely to be due partly to the liberal postnatal use of pharmacological doses of corticosteroids to prevent or treat chronic lung disease⁹ during the 1990s. However, the change in outcome in different centers is likely to vary, given the considerable difference between centers both in their clinical practices and their baseline proportion of live-born ELBW infants who survive without impairment.¹⁰

In the absence of population-based studies, it is difficult to know the extent to which the findings are representative of all ELBW survivors in the same region or more broadly across the United States. The generalizability of the findings of Hack et al⁴ to current outcomes in other centers may be reduced by the relatively low percentage of mothers (33%) who received antenatal corticosteroids and the relatively high percentage of infants (59%) who received postnatal corticosteroids. The number and characteristics of infants transferred to this center, a factor that may also limit generalizability to other centers, is not described. Nevertheless, data collected by the National Institute of Child Health and Human Development Neonatal Research Network¹¹ indicate that this center is comparable with the other 15 network centers with respect to the patient population and the proportion of ELBW survivors with developmental impairment at 18 to 22 months. The results reported by Hack et al⁴ are likely to be representative of ELBW outcomes in urban, university-affiliated, perinatal centers in the United States. The findings, like those in well-designed studies of the smallest or most immature infants in other populations,^{2 - 3 ,12} strongly support the investigators’ conclusion that ELBW survivors in the 1990s still continue to have considerable long-term health and educational needs.

The current state-of-the-art in follow-up studies might be enhanced to increase their value and facilitate better outcomes for high-risk infants by considering the following strategies. First, follow-up assessments should be performed and reported for all survivors in the lowest gestational age categories as well as all those in the lowest birth-weight categories.² Unlike the study by Hack et al, most recent studies have been based on gestational age. However, both birth weight and gestational age influence outcome, and studies based on both would allow a broader comparison with other follow-up studies and a better assessment of the prognosis of individual infants.

Second, emphasis should be placed on performing population-based follow-up studies in the United States, as in other countries. Virtually all such studies have been performed outside the United States. These studies avoid the referral biases inherent in center-based studies, allow outcome to be related to location of birth (perinatal center or community hospital), and help to define the overall need for special medical or educational services for high-risk survivors. Well-designed^{12 - 13} population-based studies deserve full support as a public health measure.

Third, to help in addressing ethical dilemmas in the care of marginally viable infants,¹⁴ the mortality and long-term morbidity of these infants should be related to treatment decisions to forgo or withdraw intensive care. Formal methods to better assess the effect of these decisions have been proposed.^{12 ,15} In addition, family outcomes should be assessed for infants who die as well as those who survive.¹⁶ Such information is needed to understand all important consequences of these treatment decisions.

Fourth, quality of life should be assessed in long-term survivors. Despite systematically higher disability rates among ELBW vs NBW survivors, Canadian studies have reported minimal difference in their self-assessed quality of life in adolescence and early adulthood.^{17 - 18} Whether this is true in US populations is an important question.

And fifth, the use of follow-up assessments should be expanded in testing interventions used before or after^{19 - 20} nursery discharge to reduce adverse medical, neurodevelopmental, or functional outcomes.²¹ Most published follow-up data have been from purely observational studies. Yet, death or impairment at 18 months or older would arguably be the best primary outcome in many clinical trials to determine the value of major perinatal interventions.

Most of these strategies would entail considerably more effort and expense. However, the best possible follow-up studies are crucial to a clear understanding of the outcome of small premature infants, the beneficial or harmful effects of perinatal treatment decisions, and the long-term needs of these children for special medical or educational services.