NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Institute of Medicine (US) Committee on Understanding Premature Birth and Assuring Healthy Outcomes; Behrman RE, Butler AS, editors. Preterm Birth: Causes, Consequences, and Prevention. Washington (DC): National Academies Press (US); 2007.

Cover of Preterm Birth

Preterm Birth: Causes, Consequences, and Prevention.

Show details

11Neurodevelopmental, Health, and Family Outcomes for Infants Born Preterm

ABSTRACT

Although advances in high-risk obstetric and neonatal care have resulted in improved survival of infants born preterm, many studies have documented the prevalence of a broad range of neurodevelopmental impairments in preterm survivors. The spectrum of neurodevelopmental disabilities includes cerebral palsy, mental retardation, visual and hearing impairments, and more subtle disorders of central nervous system function. These dysfunctions include language disorders, learning disabilities, attention deficit-hyperactivity disorder, minor neuromotor dysfunction or developmental coordination disorders, behavioral problems, and social-emotional difficulties. Preterm infants are more likely to have lower intelligence quotients and academic achievement scores, experience greater difficulties at school, and require significantly more educational assistance than children who were born at term. Preterm infants have an increased risk of rehospitalization during the first few years of life and increased use of outpatient care. Among the conditions leading to poorer health are reactive airway disease or asthma, recurrent infections, and poor growth. The smallest and most immature infants have the highest risk of health problems and neurodevelopmental disabilities. Limited evidence of the impact of prematurity on families suggests that caring for a child born preterm has negative and positive effects that change over time, that these effects extend to adolescence and are influenced by different environmental factors over time, and that many areas of family well-being are affected. The prevalence of neurodevelopmental disabilities and health impairments varies. This is not surprising, in light of the multiple etiologies and complications of preterm birth and the variability of both the intrauterine and the extrauterine environments to which fetuses and children born preterm are exposed. In recognition of the increased developmental and emotional risks of children born preterm, several interventions have focused on the provision of services in the early years of life to prevent subsequent developmental and health problems. Although early interventions have a short-term impact, it has been more difficult to demonstrate more long-term benefits.

At first glance, a wealth of data seems to be available for characterization of the outcomes of infants born preterm; however, as with many other areas addressed in this report, much of this literature uses birth weight as the measure of prematurity (see Chapter 2). The use of birth weight as a selection criterion for studies of the outcomes for infants born preterm introduces a well-recognized bias by including various proportions of more mature infants who experienced intrauterine growth retardation (IUGR). Many infants with IUGR are small for gestational age when they are born full term (i.e., at 37 to 41 weeks of gestation). Most infants with birth weights of less than 1,500 grams are preterm, but those who also have IUGR are vulnerable to the complications of both IUGR and prematurity (Garite et al., 2004). A number of the more recent studies have reported on the outcomes for preterm infants by gestational age category, but as in other parts of this report, this chapter uses birth weight-specific data when information by gestational age is not available.

Finding 11-1: Most studies of the outcomes of preterm birth use birth weight criteria for the selection of study participants. Few studies report on the outcomes for preterm infants by gestational age. In addition to infants born preterm, studies with samples of infants with birth weights less than 2,500 grams include full-term infants who are small for gestational age.

When this literature is examined, it is also well to keep in mind that preterm delivery is not a disease with a fixed set of outcomes. Rather, preterm delivery increases the risk of adverse outcomes that are also seen in term infants. Nevertheless, the more preterm an infant is, the greater the risk of adverse outcomes. Thus, these outcomes are a probability for a group and not a certainty for any given infant. Although the adverse outcomes associated with preterm delivery are discussed individually, readers should be aware that individual children may experience more than one outcome. Thus, it would not be unusual for a child to have some coordination difficulty and a specific health problem like asthma. Indeed, multiple milder problems may create more functional difficulties than a single, more severe one.

Relatively few studies of the outcomes for infants born preterm provide comparison groups, and those that do have almost uniformly selected healthy full-term infants or infants with birth weights above 2,500 grams. Some have used siblings or classmates as comparisons. The study question determines the criteria used to select the comparison group. Some have proposed the use of infants or children who experienced other life-threatening conditions to provide a better sense of the disabilities and outcomes from serious neonatal health problems.

Finally, a number of biological and environmental factors may affect the risk of adverse outcomes independent of gestational age or birth weight. To the extent that such independent risk factors have been identified, including some of those that may ameliorate the risks due to prematurity, they are discussed. Nevertheless, researchers are far from understanding all these factors, and prediction of the outcome for an individual child born preterm with any degree of certainty remains impossible.

This chapter describes the outcomes of preterm birth from a life-span perspective, including the prevalence of neurodevelopmental disabilities, health-related quality of life, and functional outcomes to adolescence and early adulthood. The chapter concludes with a discussion of intervention strategies that can be used for the developmental support of children who were born preterm after discharge from the neonatal intensive care unit (NICU).

NEURODEVELOPMENTAL DISABILITIES

Among the earliest concerns about the health of premature infants was the association between preterm delivery and neurodevelopmental disabilities. Neurodevelopmental disabilities are a group of chronic interrelated disorders of central nervous system function due to malformation of or injury to the developing brain. The spectrum of neurodevelopmental disabilities includes the major disabilities: cerebral palsy (CP) and mental retardation. Sensory impairments include visual impairment and hearing impairment. The more subtle disorders of central nervous system function include language disorders, learning disabilities, attention deficit-hyperactivity disorder (ADHD), minor neuromotor dysfunction or developmental coordination disorders, behavioral problems, and social-emotional difficulties.

Early studies focused primarily on cognitive impairment, as measured by intelligence quotient (IQ) and by the detection of motor abnormalities on standardized neurological examinations. A landmark study, the Collaborative Perinatal Project of the National Institute of Neurological and Communicative Disorders and Stroke, monitored 35,000 children born before neonatal intensive care (i.e., in the late 1950s and early 1960s) for 7 years. Although only 177 children born at less than 34 weeks gestation survived, the study documented the increased risk of cognitive and motor impairment as a function of decreasing gestational age. It highlighted the need for neurodevelopmental follow-up of populations born preterm, especially as the emergence of neonatal intensive care and high-risk obstetric care dramatically reduced gestational age-specific mortality rates but not preterm birth rates (see Chapters 1, 2, and 10). The history of neonatal intensive care is not only one of miracles achieved, but also of therapeutic misadventures (Allen, 2002; Baker, 2000; Silverman, 1980; Silverman, 1998). Iatrogenic complications have contributed to adverse health and neurodevelopmental outcomes in the past, but a shift from a trial and error approach toward evidence-based medicine is establishing a more empiric basis for treating mothers and preterm infants. Although new therapies are generally evaluated in randomized clinical trials, the safety and efficacy of many currently used treatments and medications have not been adequately studied (see Chapter 10).

The resulting literature demonstrates wide variations in the prevalence of neurodevelopmental disabilities (Allen, 2002; Aylward, 2002b; Aylward, 2002a; Aylward, 2005). Much of this variation is due to methodological issues, for example, a lack of uniformity in sample selection criteria, the method and the length of follow-up, follow-up rates, and the outcome measures and the diagnostic criteria used. Variations in outcome frequencies reported also reflect differences in the population base and in clinical practice. Whenever possible, outcomes data are provided by gestational age categories for preterm infants born in the 1990s to the present. However, because the age of evaluation determines which outcomes can be assessed, recent studies of the outcomes for adolescents who were born preterm report on preterm births that occurred in the 1980s. The time lag required for follow-up makes caution necessary in generalizing reported adolescent and adult outcomes to preterm infants who survive with the technology available today. Perinatal and neonatal risk factors do not reliably predict these long-term outcomes. Therefore, research is needed to identify better neonatal predictors of neurodevelopmental disabilities, functional abilities, health and other long-term outcomes.

Motor Impairment

Cerebral Palsy

Cerebral palsy (CP) is a general term to describe a group of chronic conditions that impair control of movement and posture. CP is due to malformation of or damage to motor areas in the brain, which disrupts the brain’s ability to control movement and posture. Symptoms of CP may range from mild to severe, change over time and differ from person to person, and include difficulty with balance, walking, and fine motor tasks (such as writing or using scissors) and involuntary movements. Many people with CP also have associated cognitive, sensory, social, and emotional disabilities (NIDS, 2005).

The diagnosis of CP may not become certain until the second year of life. As many as 17 to 48 percent of preterm infants demonstrate neuromotor abnormalities during infancy (e.g., abnormal muscle tone or asymmetries) (Allen and Capute, 1989; Khadilkar et al., 1993; Pallas Alonso et al., 2000; Vohr et al., 2005). Some of these infants go on to develop significant neuromotor abnormalities and motor delays that signify CP, but most do not. Although neuromotor abnormalities tend to resolve or do not interfere with function, transient neuromotor abnormalities are associated with an increased risk of later school and behavioral problems (Drillien et al., 1980; Khadilkar et al., 1993; Sommerfelt et al., 1996; Vohr et al., 2005).

The severity of CP is determined by the type of CP, which limbs are affected, and the degree of functional limitation. Increasingly, investigators are distinguishing between mild CP and moderate to severe (i.e., disabling) CP (Doyle and Anderson, 2005; Grether et al. 2000; Vohr et al., 2005; Wood et al., 2000). Many longitudinal studies of the outcomes for preterm infants show good stability between motor assessments at 18 to 30 months of age and at school age (Hack et al., 2002; Marlow et al., 2005; Wood et al., 2000).

The smallest and most immature infants have the highest risk of CP. In their seventh report of CP in Sweden, Hagberg and associates (1996) reported an almost stepwise increase in the prevalence of CP with gestational age: 1.4 per 1,000 live births for children born at more than 36 weeks gestation, 8 per 1,000 live births for children born between 32 and 36 weeks gestation, 54 per 1,000 live births for children born between 28 and 31 weeks gestation, and 80 per 1,000 live births for children born at less than 28 weeks of gestation. Because they report prevalence as the number who have CP per 1,000 live births, infants who die are included in the denominator.

For the most immature infants, another meaningful statistic is the rate of CP among survivors. On the basis of data for preterm survivors born in the late 1980s through the 1990s, the rate of CP increases with decreasing gestational age or birth weight category (Table 11-1) (Colver et al., 2000; Cooke, 1999; Doyle et al., 1995; Doyle and Anderson, 2005; Elbourne et al., 2001; Emsley et al., 1998; Finnstrom et al., 1998; Grether et al., 2000; Hack et al., 2000 Hack et al., 2005; Hansen and Greisen, 2004; Hintz et al., 2005; Lefebvre et al., 1996; Mikkola et al., 2005; O’Shea et al., 1997; Piecuch et al., 1997a,b; Salokorpi et al., 2001; Sauve et al., 1998; Stanley et al., 2000; Tommiska et al., 2003; Vohr et al., 2000, 2005; Wilson-Costello et al., 2005; Wood et al., 2000). Only 0.1 to 0.2 percent of full-term children develop CP, whereas 11 to 12 percent born at 27 to 32 weeks of gestation and 7 to 17 percent born at less than 27 or 28 weeks of gestation develop CP (Doyle, 2001; Elbourne et al., 2001; Finnstrom et al., 1998; Lefebvre et al., 1996; Vohr et al., 2005). A comprehensive British study of preterm infants born in 1995 with gestational ages of less than 26 weeks diagnosed CP in 20 percent of the survivors at 6 years of age (Marlow et al., 2005). In the few reported survivors with birth weights of less than 500 grams, a quarter to a half developed CP (Sauve et al., 1998; Vohr et al., 2000).

TABLE 11-1. Rates of Cerebral Palsy in Preterm Children by Gestational Age Category.

TABLE 11-1

Rates of Cerebral Palsy in Preterm Children by Gestational Age Category.

Many more studies have reported on the outcomes of CP in terms of birth weight categories. In a review of 17 studies published from 1988 to 2000, Bracewell and Marlow (2002) estimated that approximately 10 percent of preterm infants with birth weights of less than 1,000 grams developed CP. An older meta-analysis of 85 studies of infants with birth weights of less than 1,500 grams estimated that 7.7 percent of survivors developed CP (Escobar et al., 1991). Studies of 18- to 20-year-olds reported that from 5 to 7 percent of those who were born with birth weights of less than 1,500 grams and up to 13 percent of those born with birth weights of less than 1,000 grams had CP (Ericson and Kallen, 1998; Hack et al., 2002; Lefebvre et al., 2005; Saigal et al., 2006a). A Swedish study of young men born as singletons from 1973 to 1975 with birth weights of less than 1,500 grams estimated an odds ratio for CP of 55 (95 percent confidence interval = 41 to 75) (Ericson and Kallen, 1998).

With continuing improvements in high-risk obstetric and neonatal intensive care over the last several decades, several studies have demonstrated small increases or decreases in the overall prevalence of CP (Colver et al., 2000; Hagberg et al., 1996; Stanley and Watson, 1992; Stanley et al., 2000). However, any improvements in gestational age- or birth weight-specific rates of CP are offset by dramatic decreases in the rates of infant mortality. The net result is that more preterm children survive, but more children have CP as well.

Many regional studies of children with CP find an overrepresentation of preterm children with CP than the number expected for their birth rates (Table 11-1) (Amiel-Tison et al., 2002; Colver et al., 2000; Cummins et al., 1993; Dolk et al., 2001; Hagberg et al., 1996; MacGillivray and Campbell, 1995; Petterson et al., 1993; Stanley and Watson, 1992). Although only 1.4 percent of infants are born at less than 32 weeks of gestation, they comprise 26 percent of children with CP. Four percent of all live births are born at 32 to 36 weeks of gestation, and they constitute 16 to 37 percent of children with CP. Although less than 10 percent of births are preterm, approximately 40 to 50 percent of children with CP are born preterm.

Although children born preterm are vulnerable to all types of CP, the most common type is spastic diplegia (Hack et al., 2000; Hagberg et al., 1996; Wood et al., 2000). Spasticity is characterized by tight muscle tone, increased reflexes, and limited movement around one or more joints. Spasticity of both lower extremities but no or very little involvement of the arms constitutes spastic diplegia. Although most children with spastic diplegia require physical therapy and medical interventions (e.g., orthopedic surgery, orthoses, or Botulinum Toxicum injections), many children with spastic diplegia are quite functional by school age. In a study of children born at less than 26 weeks gestation, 43 percent with spastic diplegia were unable to walk and 43 percent had an abnormal gait at 6 years of age (Marlow et al., 2005).

A large regional study of Swedish preterm children with CP reported that 66 percent had spastic diplegia, 22 percent had spastic hemiplegia, and 7 percent had spastic quadriplegia (Hagberg et al., 1996). Associated deficits were common: 39 percent had mental retardation, 26 percent had epilepsy, 18 percent had severe visual impairment, and 23 percent had hydrocephalus. The proportion of children with CP who had spastic diplegia decreased with increasing gestational age category: 80 percent for children born at less than 28 weeks of gestation, 66 percent for children born at between 28 and 31 weeks of gestation, 58 percent for children born at between 32 and 36 weeks of gestation, and 29 percent for children born at greater than 36 weeks of gestation. The proportion of children with hemiplegia increased with gestational age: 10 percent for children born at less than 28 weeks gestation, 16 percent for children born at between 28 and 31 weeks gestation, 34 percent for children born at between 32 and 36 weeks gestation, and 44 percent for children born at less than 36 weeks gestation.

Coordination and Motor Planning

Children with incoordination and motor planning problems are less likely to enjoy and participate in many preschool and playground activities. Minor neuromotor dysfunction is a diagnosis used to describe infants and children who have persistent neuromotor abnormalities but minimally to mildly impaired motor function. Children with minor neuromotor dysfunction may have mild motor delay but are able to walk by age 2 years and have good mobility. They have a higher risk of coordination difficulties, motor planning problems, fine motor incoordination, or sensorimotor integration problems (which at preschool and school age may be diagnosed as developmental coordination disorder) (Botting et al., 1998; Hadders-Algra, 2002; Hall et al., 1995; Khadilkar et al., 1993; Mikkola et al., 2005; Pharoah et al., 1994; Vohr and Coll, 1985). In a study of 5-year-olds born between 1996 and 1997 with birth weights of less than 1,000 grams, 51 percent had coordination problems, 18 to 20 percent had abnormal reflexes or abnormal posture, and 17 percent had exceptional involuntary movements (Mikkola et al., 2005). Sensorimotor integration problems can range from inability to tolerate certain textures of food or clothing (e.g., an inability to tolerate lumpy food or the tag on the back of a T-shirt) to difficulty following demonstrated directions (e.g., how to put on a shirt or tie shoelaces) or an inability to tolerate motion (e.g., swinging).

Preterm children, even those with normal intelligence and no CP, have more difficulties than full-term children with fine motor, visual motor, visual perceptual, and visual spatial tasks. These tasks include drawing, cutting with scissors, dressing, writing, copying figures, perceptual mapping, spatial processing, finger tapping, and pegboard performance. In a study of 5-year-old children with birth weights of less than 1,500 grams, 23 percent had impaired fine motor skills and 71 percent scored 1 standard deviation or more below average on tests of fine motor function (Goyen et al., 1998). Below-average performances in visual motor skills and visual perceptual tasks were noted for 17 and 11 percent of the children, respectively. These problems were most common in children born at less than 28 weeks of gestation. Even the more mature preterm children are at risk for these problems; a third of school-age children born at 32 to 36 weeks of gestation had poor fine motor and writing skills (Huddy et al., 2001).

Failures at gross motor, fine motor, sensorimotor, and visual perceptual activities are mild in comparison with the difficulties with mobility and adaptive skills that many children with CP face. Nonetheless, these subtle abnormalities of central nervous system function can, over time, adversely influence the child’s self-esteem and peer relationships, which in turn contribute to a cycle of frustration and despair that interferes with academic progress and social relationships. Early recognition of these subtle deficits allows modification of expectations, teaching methods, and the environment to support the development of these children and prevent adverse secondary consequences.

Cognitive Impairment

Cognitive Test Scores and Mental Retardation

Intelligence is not one skill but a composite of multiple cognitive processes, including visual and auditory memory, abstract reasoning, complex language processing, understanding of syntax, visual perception, visual motor integration, and visual spatial processing. A variety of standardized intelligence tests are available for use with children at each age level. Scores across a variety of cognitive tasks are summed to form an IQ or, for younger children, a developmental quotient (DQ) (Lichtenberger, 2005). Cognitive assessments of very young infants are limited in their predictive ability because of their reliance on assessment of visual-motor and perceptual abilities. As children mature, more verbal and abstract cognitive processes can be evaluated, and scores more accurately reflect their abilities. Cognitive tests are standardized for diverse large populations, with an IQ score of 100 considered the population mean.

The IQ score is a global score that does not include information about subtle dysfunctions. The full range of cognitive deficits seen in preterm children is not well described by the IQ score, and further cognitive analyses are necessary. Many preterm children have a wide scatter in their cognitive abilities, with excellent performance in some areas but relative weakness in other areas, and these contribute to difficulties in the classroom and at home.

Calculation of a DQ for preterm infants is complicated by whether their age should be calculated from their birth date (i.e., the chronological age) or from their due date (i.e., age corrected for the degree of prematurity). This issue is more important arithmetically the younger the infant is and the lower the gestational age at birth was. For example, a 6-month-old preterm infant born 3 months early who has skills at the normal level for a 3-month-old would have a normal DQ of 100 if it was corrected for the degree of prematurity but would be considered delayed in skill attainment, with a DQ of 50, if the chronological age was used.

For the most part, neuromaturation of the preterm infant in the NICU proceeds along the same timeline as intrauterine development (Allen, 2005a; Saint-Anne Dargassies, 1977). From biological and maturational perspectives, few environmental influences significantly accelerate neuromaturation, and most agree that one should fully correct for the degree of prematurity when preterm infants are evaluated and that this correction should be incorporated for at least the first 2 years of life (Allen, 2002; Aylward, 2002a). Whether or not one corrects for the degree of prematurity may influence IQ scores for up to 8 years (Rickards et al., 1989).

Mental retardation is a disability that originates in childhood and is characterized by significant limitations both in intellectual functioning and in adaptive behavior, as expressed in conceptual, social, and practical adaptive skills (AAMR, 2005). Intellectual functioning is considered subaverage or significantly limited when an individual’s IQ score is 2 or more standard deviations below the mean on a standardized intelligence test (generally an IQ less than 70 or 75, depending on the test). Borderline intelligence is when an individual’s IQ score is between 1 and 2 standard deviations below the mean (generally, IQs of 70 to 80 or 85).

In a study of children with mental retardation in Norway, children born at 32 to 36 weeks of gestation had a 1.4 times increased risk of mental retardation than full-term children, and this risk increased to 6.9-fold for children born at less than 32 weeks of gestation (Stromme and Hagberg, 2000). The risks of mental retardation in children born preterm compared with those in children born with normal birth weights increase from 2.3-fold for children with birth weights of 1,500 to 2,499 grams to 12-fold for children with birth weights of less than 1,500 grams, 15-fold for children with birth weights of less than 1,000 grams, and 22-fold for children with birth weights of less than 750 grams (Resnick et al., 1999; Stromme and Hagberg, 2000). Nonetheless, children born at less than 32 weeks of gestation or with birth weights of less than 1,500 grams comprised only 4 percent of children with mental retardation.

On the basis of data for preterm children born in the late 1980s and 1990s, survivors born preterm with the lowest gestational ages and birth weights have the highest risk of mental retardation and borderline intelligence (Tables 11-2 and 11-3). A recent large study of infants born at less than 26 weeks gestation in 1995 in the British Isles and evaluated at age 6 years reported that 21 percent had an IQ 2 or more standard deviations below the test mean and 25 percent had borderline intelligence (i.e., IQs 1 to 2 standard deviations below the test mean), whereas for the controls born full term the rates were 0 and 2 percent, respectively (Marlow et al., 2005).

TABLE 11-2. Mean IQ Scores in Children Born Preterm and Full Term.

TABLE 11-2

Mean IQ Scores in Children Born Preterm and Full Term.

TABLE 11-3. Proportion of Survivors of Preterm Birth With and Without Cognitive Impairment.

TABLE 11-3

Proportion of Survivors of Preterm Birth With and Without Cognitive Impairment.

Studies that compare preterm children’s performance on intelligence tests against published test norms may underestimate their cognitive disadvantage. Although cognitive tests are standardized on the basis of a mean IQ of 100 for normal populations, there is a tendency for the mean IQ score in normal or control populations to drift upward over time. Marlow et al. (2005) noted a mean cognitive score of 106 in their full-term classmate controls. With restandardization, the percentage of children born before 26 weeks of gestation who had cognitive scores 2 standard deviations or more below the full-term comparison group’s mean score rose from 21 to 41 percent.

Children born full term with normal birth weights and raised in similar environments have generally served as comparison groups in studies of the outcomes of preterm birth. In a 1989 meta-analysis, 4,000 children born with birth weights of less than 2,500 grams had a mean IQ that was 5 to 7 points lower than the mean for 1,568 controls who were born full term (Aylward et al., 1989). In more recent studies of children with birth weights of less than 1,500 or 1,000 grams, the preterm children have mean IQ scores that were 10 to 17 points, or 1 standard deviation, below those for the full-term controls (Breslau et al., 1994; Doyle and Anderson, 2005; Grunau et al., 2002; Halsey et al., 1996; Hansen and Greisen, 2004; Whitfield et al., 1997).

A 2002 meta-analysis of 16 case-control studies of children aged 5 years old or older and born from 1975 to 1988 noted significantly lower cognitive scores for 1,556 children born preterm compared with those for 1,720 controls born full term, with a weighted mean difference of 10.9 (95% CI 9.2–12.5) (Bhutta et al., 2002). When only studies that excluded severely neurologically impaired children born preterm were analyzed, the weighted mean difference was 10.2 (95% CI 9.0–11.5).

Many studies have noted a trend toward lower mean cognitive scores with decreasing gestational age and birth weight categories (Tables 11-2 and 11-3) (Bhutta et al., 2002; Doyle and Anderson, 2005; Hall et al., 1995; Halsey et al., 1996; McCarton et al., 1997; McCormick et al., 1992; Saigal, 2000c; Taylor et al., 2000; Wilson-Costello et al., 2005).

There is some controversy as to the consistency of individual DQ and IQ scores over time. Artifacts of intelligence testing contribute to this confusion. One small study of infants with birth weights of less than 1,000 grams found lower Bayley Scale of Infant Development cognitive scores at 18 to 20 months from term than at 8 months from term, and this was associated with infant behavioral characteristics and family income (Lowe et al., 2005). The Bayley test items are, by necessity, heavily weighted toward visual motor abilities during the first year, but during the second year language concepts can be evaluated. In a study with 200 children with birth weights of less than 1,000 grams, Hack and colleagues (2005a) noted a significant improvement between the Bayley scores at age 20 months and cognitive scores at age 8 years (means, 76 and 88, respectively). The proportion of children with cognitive impairment (i.e., IQ scores 2 or more standard deviations below the test mean) decreased from 39 percent at age 20 months to 16 percent at age 8 years. This difference could be an artifact of the use of different tests at different ages. Ment and colleagues (2003) reported an increase in vocabulary test scores of 10 or more points in 45 percent of children with birth weights of less than 1,500 grams when they were retested at age 96 months after initial testing at age 36 months.

Further complicating the interpretation of these differences is the upward drift of IQ scores as a function of increased time from standardization (Flynn, 1999). Improvement in IQ scores with age is most common in children born preterm who have no neurological injuries or impairment and whose mothers have high levels of educational attainment (Hack et al., 2005; Koller et al., 1997; Ment et al., 2003). Despite improvements in their IQ scores with age, these children had more academic problems than children with stable IQ scores in the average range (Hack et al., 2005).

Adolescents and young adults who were born preterm continue to demonstrate a cognitive disadvantage compared with those who were born full-term. When young adults who were born with birth weights of less than 1,000 grams were tested at a mean age of 18 years, they were found to have lower verbal, performance, and full-scale IQ scores than fullterm controls: 93 and 106, 97 and 109, and 94 and 108, respectively (p < 0.0001) (Lefebvre et al., 2005). Hack and colleagues (2002) evaluated 20-year-olds who were born with birth weights of less than 1,500 grams and found a mean IQ of 87, whereas the mean IQ was 92 for controls who were born with normal birth weights. Only half (51 percent) had an IQ score above 84, whereas 67 percent of adults who were born full term had IQ scores above 84. When Saigal and colleagues (2000c) compared 12- to 16-year-olds who were born with birth weights less than 1,000 grams with controls who had normal birth weights, preterm children had lower mean IQ scores even when children with IQ scores below 85 or neurosensory impairments were excluded (mean IQ scores 99 and 104, respectively, p < 0.001; for total sample, mean IQ scores were 89 and 102, respectively; p < 0.0001).

Preterm children with no neurological impairments demonstrate not only lower mean cognitive test scores but also more problems with specific cognitive processes than fullterm controls (Anderson and Doyle, 2003; Bhutta et al., 2002; Breslau et al., 1994; Grunau et al., 2002; Hack et al., 1993; Mikkola et al., 2005). One study of preschool children who were born at 30 to 34 weeks of gestation and who had no neurological impairments found lower scores not only on the Stanford-Binet IQ test (111 and 121, respectively; p < 0.001) but also on tests of visual perception, visual motor integration, memory for location, sustained attention, and vocabulary, as compared to a matched control group of children born at term (Caravale et al., 2005).

A number of studies have demonstrated that preterm children who were born with birth weights less than 1,000 or 1,500 grams and who had normal IQ scores have more problems with attention, executive function (i.e., organization and planning skills), memory, language, learning disabilities, spatial skills, and fine and gross motor function than controls who were born with normal birth weights (Anderson and Doyle, 2003; Aylwarda, 2002a; Goyen et al., 1998; Grunau et al., 2005; Hack and Taylor, 2000; Halsey et al., 1993; Mikkola et al., 2005; O’Callaghan et al., 1996; Ornstein et al., 1991; Rose et al., 2005; Saigal et al., 1991).

School Problems

Difficulty with cognitive processes contributes to the increased risk of school problems seen in children born preterm (Aylward, 2002a; Grunau et al., 2002). In a study of 153 children born at less than 28 weeks of gestation, only half were ready and able to enter kindergarten with their peers (Msall et al., 1992). Speech and language delays, attention deficits, and learning disabilities were common. Among 8- to 10-year-old children who were born preterm with birth weights of less than 800 or 1,000 grams, 13 to 33 percent repeated a grade, 15 to 47 percent required some special education support, and 2 to 20 percent were in special education placements (Buck et al., 2000; Gross et al., 2001; Whitfield et al., 1997).

In a longitudinal study of 813 Dutch children born at less than 32 weeks gestation or with birth weights less than 1,500 grams, at age 9 years children born preterm had more school-related problems than the general Dutch population: 32 and 14 percent, respectively, functioned below grade level; 38 and 6 percent, respectively, received special education assistance; and 19 and 1 percent, respectively, were in special education classes (Hille et al., 1994). In adolescence (age 14 years), 27 percent received special education services, whereas 7 percent of their peers received special education services (Walther et al., 2000).

By early adolescence, the children who had been born preterm with birth weights of less than 1,000 grams were 3 to 5 times more likely than the controls born fullterm to fail a grade and required 3 to 10 times more special education resources than the controls born fullterm (Klebanov et al., 1994; Saigal et al., 2000c; Taylor et al., 2000). Saigal and colleagues (2005b) found progressive increases in school problems with decreasing birth weight category: 13 percent for fullterm controls, 53 percent for children born with birth weights between 750 and 1,000 grams, and 72 percent for children born with birth weight less than 750 grams. The proportions of children who had been born with birth weights less than 1,000 grams and who were in regular classrooms without grade failures or special education resources were only 42 to 50 percent at 8 to 10 years of age and as low as 36 percent at 18 years of age (Halsey et al., 1996; Klebanov et al., 1994; Lefebvre et al., 2005; Saigal et al., 2000c).

Many difficulties in school reflect the presence of learning disabilities, which become more apparent as children who had been born preterm progress through their education. A specific learning disability is a term that refers to a heterogeneous group of disorders of one or more of the basic psychological processes involved in understanding or in using spoken or written language. These disorders may manifest as significant difficulties with the acquisition and use of listening, speaking, reading, writing, reasoning, or mathematical skills. The diagnosis of a learning disability requires comparisons of performances on tests of academic achievement, cognition, language, visual motor integration, and perceptual abilities. Although the incidence of learning disabilities varies depending on how they are defined, most estimates indicate that 10 percent or less of the general population has evidence of a specific learning disability.

Ample evidence suggests that many more children born preterm have specific learning disabilities than children of normal birth weight born fullterm. By school age, despite normal intelligence, children born with birth weights of less than 1,000 or 800 grams have a 3- to 10-fold increased risk of problems with reading, writing, spelling, or mathematics compared with the risk for their peers who had been born fullterm (Aylward, 2002a; Grunau et al., 2002; Hall et al., 1995; O’Callaghan et al., 1996; Ornstein et al., 1991; Saigal et al., 2000c). The most consistent academic difficulties associated with preterm birth are arithmetic and reading (Anderson and Doyle, 2003; Bhutta et al., 2002; Hack et al., 1994; Klebanov et al., 1994; O’Callaghan et al., 1996; Ornstein et al., 1991; Saigal et al., 2000c). The proportion of children born preterm who experience academic difficulties increases with age as the complexity of the schoolwork increases and efficiency becomes an issue in the higher grade levels (Aylward, 2002a). In a detailed analysis of the nature of the learning disabilities in 8- to 9-year-olds who had been born with birth weights of less than 1,000 grams, Grunau and colleagues (2002) suggested that the children’s problems with visual memory, visual motor integration, and verbal intelligence explained many of their difficulties with arithmetic and reading.

As with other neurodevelopmental disabilities and school problems, the prevalence of learning disabilities increases with decreasing gestational age and birth weight: 7 to 18 percent in children born full term, 30 to 38 percent in children born with birth weights 750 to 1,499 grams, 66 percent in children born at less than 28 weeks of gestation, and 50 to 63 percent in children born with birth weights less than 750 grams (Avchen et al., 2001; Aylward, 2002a; Breslau, 1995; Grunau et al., 2002; Hack et al., 1994; Halsey et al., 1996; Hille et al., 1994; Pinto-Martin et al., 2004; Taylor et al., 2000). In a study of 12- to 16-year-olds in three birth weight categories (normal, 750 to 1,000 grams, and less than 750 grams), the proportion with scores 2 or more standard deviations below the mean increased with decreasing birth weight for reading (0, 12, and 23 percent, respectively), spelling (2, 18, and 38 percent, respectively), and arithmetic (5, 32, and 50 percent, respectively) (Saigal et al., 2000c). The parents of more mature preterm children (those with gestational ages 32 to 35 weeks) reported significant problems with mathematics in 29 percent of the children, speaking in 19 percent, reading in 21 percent, and writing in 32 percent (Huddy et al., 2001).

In a study of 20-year-olds, those with birth weights less than 1,500 grams continued to have lower academic achievement scores in mathematics and reading than controls born with normal birth weights (Hack et al., 2002). Fewer of the young adults studied had graduated from high school (74 and 83 percent, respectively). The mean age at graduation was higher for those with lower birth weights (18.2 and 17.9 years, respectively). Fewer men born weighing less than 1,500 grams than men of normal birth weight attended a 4-year college (16 and 44 percent, respectively), but the rates were similar for women (33 and 38 percent, respectively). In a Canadian study of 18-year-olds born with birth weights of less than 1,000 grams, 56 percent obtained a secondary school diploma, whereas 86 percent of the controls who had been born fullterm did so (Lefebvre et al., 2005). In contrast, a recent study of 23-year-old Canadians weighing less than 1,000 grams at birth and controls born with normal birth weights found no differences in the rates of completion of high school, postsecondary education, or university education or the total number of years of education that they had completed (Saigal et al., 2006a).

Visual Impairment

As discussed in Chapter 10, retinopathy of prematurity (ROP) is a common complication of prematurity that increases with decreasing gestational age and birth weight. As a group, preterm children have a higher risk of impaired visual acuity than full-term children (Table 11-4). Myopia (i.e., nearsightedness) is one of the most common visual sequelae, and its incidence increases with the severity of ROP and with decreasing gestational age. Myopia occurs in 20 to 22 percent of children born with birth weights of less than 1,251 or 1,751 grams, and 4.6 percent have a high degree of myopia (i.e., ≥ 5 diopters) (O’Connor et al., 2002; Quinn et al., 1998).

TABLE 11-4. Sensory Impairments in Children Born Preterm.

TABLE 11-4

Sensory Impairments in Children Born Preterm.

Other visual problems include hyperopia and astigmatism (in 12 and 29 percent of children born at less than 29 weeks of gestation, respectively) (Hard et al., 2000). The need for glasses was higher in 7-year-olds born at gestational ages of less than 32 weeks than in controls born fullterm (13 and 4 percent, respectively) (Cooke et al., 2004). At 10 to 14 years of age, visual impairment was more common in children born with birth weights of less than 750 grams than in children born with birth weights of between 750 and 1,499 grams and children born with normal birth weights (31, 13, and 11 percent, respectively), as was the need for glasses (47, 24, and 27 percent, respectively) (Hack et al., 2000). In a British study of children born at less than 26 weeks of gestation, 24 percent of preterm 6-year-olds wore glasses, whereas 4 percent of controls with fullterm gestations did so (Marlow et al., 2005).

Strabismus (i.e., ocular misalignment, or crossed eyes) is also a frequent complication of prematurity. Strabismus has been reported in 3 percent of children born fullterm; 14 to 19 percent of children born with birth weights less than 1,500 grams, birth weights less than 1,750 grams, and gestational ages less than 29 and 32 weeks; and 24 percent of children born at gestational ages less than 26 weeks (Bremer et al., 1998; Hard et al., 2000; Marlow et al., 2005, O’Connor et al., 2002). The risk of strabismus increases with intraventricular hemorrhage, periventricular leukomalacia, and the severity of ROP (Bremer et al., 1998; Hard et al., 2000; Hardy et al., 1997; Marlow et al., 2005; O’Connor et al., 2002; O’Keefe et al., 2001). Treatments for strabismus include correction with glasses or surgery, or both. Repka and colleagues (1998) reported that 10 percent of children with severe ROP underwent surgery for strabismus.

The sequelae of strabismus include amblyopia (i.e., suppression of the visual input to the cortex) and a loss of binocular vision. Amblyopia has been found to occur in 1 to 4 percent of the general population, 2.5 percent of preterm infants without ROP, 12 percent of children with ROP, and 20 percent of children with severe ROP (Cats and Tan, 1989; Repka et al., 1998). The rate of the absence of stereopsis was higher in 7-year-olds born at less than 32 weeks of gestation than in controls born at fullterm (16.5 and 3.8 percent, respectively) (Cooke et al., 2004). Optic nerve atrophy and cortical visual impairment can also influence visual acuity in preterm children (Repka, 2002).

Although late or severe ophthalmic findings, including cataracts, angle closure glaucoma, and retinal detachment, are uncommon in children born preterm, they interfere with function and quality of life in children, adolescents, and adults when they occur (Kaiser et al., 2001; Machemer, 1993; Repka, 2002). Cataracts have been associated with untreated ROP and severe ROP (Kaiser et al., 2001; Repka et al., 1998). Glaucoma presents as an acute illness because of the severely increased pressure within the eye globe. The abnormal neovascular tissue of ROP can progressively and silently fold, exert traction, and tear or even detach the retina, causing a loss of vision (Kaiser et al., 2001; Machemer, 1993). Although most retinal tears or detachments occur in those with severe ROP, some children born preterm had mild or no ROP but a high degree of myopia (Kaiser et al., 2001). Surgical treatments, which include placement of a flexible band (i.e., a scleral buckle) around the eye to reduce tractional forces and vitrectomy (i.e., replacement of the vitreous fluid of the eye with sterile fluid), can restore visual function to some individuals. Severe detachment results in poor vision, with only light perception or, if the detachment is complete, no vision.

Ophthalmic morbidities are common in survivors of preterm birth, and early detection carries the best prognosis. Half of children with birth weights less than 1,751 grams had ophthalmic morbidities when they were evaluated at ages 10 to 12 years, whereas 20 percent of fullterm controls had ophthalmic morbidities (O’Connor et al., 2002). Although the risk was highest in individuals with severe ROP, ophthalmic morbidities were more common in survivors of preterm birth who had no or only very mild ROP than in controls born fullterm. Whether these ophthalmic problems contribute to the higher rate of visual perceptual deficits in the survivors of preterm birth (42 percent of 5- to 9-year-old children born at gestational ages less than 29 weeks compared with 14 percent of controls born fullterm) is not well understood (Hard et al., 2000, see section on Cognition). For some of these children, even if their visual system is intact, dysfunction of visual processing in the occipital cortex impairs their ability to perceive and understand visual patterns (drawings of figures, letters on a page).

Although the causes of ROP, ophthalmic morbidities, and visual perceptual deficits are, for the most part, unknown, identification of these morbidities allows their correction or amelioration, thereby improving functional outcomes. There is not yet adequate recognition of the need for regular ophthalmologic follow-up during the life span of survivors of preterm birth and assessment of their visual perceptual abilities when they are of preschool and school age (Hard et al., 2000; Kaiser et al., 2001; Repka, 2002).

Hearing Impairment

Children born preterm have a higher incidence of hearing loss than the general population. In the majority of studies of children born in the 1990s, severe hearing impairment occurred in 2 to 4 percent of children who had been born at less than 25 weeks gestation and 1.5 to 3 percent of children who had weighed less than 1,000 grams at birth (Doyle and Anderson, 2005; Hansen and Greisen, 2004; Hintz et al., 2005; Vohr et al., 2005). A British study of 6-year-old children born in 1995 at gestational ages of less than 26 weeks reported that 3 percent had profound sensorineural hearing impairment that could not be corrected with hearing aides, 3 percent had sensorineural hearing impairment that could be corrected with hearing aides, and 4 percent had mild hearing impairment (Marlow et al., 2005). Among the comparison group of children who had been born fullterm, 1 percent had hearing impairment that could be corrected with hearing aides and 1 percent had mild hearing impairment. A Swedish study of 18- to 19-year-olds born between 1973 and 1975 with birth weights of less than 1,500 grams found impaired hearing in 7 percent (Ericson and Kallen, 1998). An Australian study of 14-year-olds who had been born with birth weights of less than 1,000 grams reported that 5 percent required hearing aides (Doyle and Casalaz, 2001). When hearing impairment was defined as no perception of sounds at or above 40 dB in the better ear, bilateral hearing impairment occurred in 1.6/1,000 children aged 3 to 10 years, and in 7.5/1,000 of children born before 29 weeks gestation (Chapter 12, Table 12-8).

Some data suggest that preterm infants have difficulty with auditory processing and auditory discrimination, which involve multiple neural pathways to the cortex of the brain. A study of preterm infants born between 24 and 32 weeks of gestation who had normal cranial ultrasounds assessed event-related potentials (neurophysiological recordings) to evaluate learning and memory of patterns of speech sound discrimination (Therien et al., 2004). Unlike the fullterm controls, preterm infants were unable to discriminate between the speech of their mother and a stranger and demonstrated deficits in discriminating simple speech sounds and auditory recognition memory. Because speech discrimination is necessary for speech recognition, these types of deficits compromise the acquisition of language skills.

There is universal agreement regarding the importance of hearing for speech and language acquisition, and of the need to identify hearing impairment as early as possible. For this reason, most states in the United States are implementing plans to screen the hearing of all newborns (White, 2003). Because of their risk of hearing impairment and language processing problems, the hearing of all preterm infants should be screened before they are discharged from the hospital and later at the first sign of language delay or recurrent ear infections. Many infants with hearing impairment respond favorably to hearing aides, and a number of strategies can be used to teach language even when amplification is ineffective (Gabbard and Schryer, 2003; Gravel and O’Gara, 2003; Yoshinaga-Itano, 2000). Recent evidence also suggests that cochlear implants are the most successful when they are implanted early in infancy (Niparko and Blankenhorn, 2003).

Behavioral and Social-Emotional Problems

Behavior and social-emotional problems are more difficult to define clinically, and most of these data are elicited from surveys of parents and teachers. Symptoms suggestive of ADHD occur two to six times more frequently in children born preterm with birth weights of less than 1,000 grams, less than 1,500 grams, and less than 2,000 grams than in controls born fullterm (9 to 15 percent diagnosed with ADHD compared with 2 percent of controls born fullterm) (Aylward, 2002a; Bhutta et al., 2002; Breslau, 1995; Levy, 1994; Pharoah et al., 1994; Saigal et al., 2001; Stjernqvist and Svenningsen, 1995; Szatmari et al., 1990; Taylor et al., 1998). Refinement of descriptions of impairments of attention and behavior provides further insight into the problems children born preterm and their families face. In a study of 8-year-old children born at less than 28 weeks of gestation or with birth weights of less than 1,000 grams, the preterm children had significantly lower scores for processing speed, attention, and working memory and higher scores for hyperactivity than those for children born with normal birth weights (Anderson and Doyle, 2003). Middle-school children who weighed less than 750 grams at birth were found to have a much higher prevalence of symptoms of executive function disorder, including difficulties with planning, problem solving, organizing, and abstracting, which can seriously influence their function and behavior at school and at home (Taylor et al., 2000).

Parents and teachers report that preterm children with birth weights of less than 1,000 or 1,500 grams lag behind their peers in social competence and behavior and that this is unrelated to their IQ scores (Anderson and Doyle, 2003; Breslau et al., 1988). It is especially true for boys. In a large study of 8- to 10-year-old children born between 1978 and 1981, significantly more children with birth weights of less than 2,500 grams than chil dren born with normal birth weights had behavioral problems, but there were no significant differences among the low birth weight categories (i.e., 27 to 29 percent for children with birth weights of less than 1,000, 1,001 to 1,500, and 1,501 to 2,500 grams versus 21 percent for children with birth weights of greater than 2,500 grams) (McCormick et al., 1992, 1996). Specifically, the children born preterm had higher subscores for hyperactivity and whining and were perceived to be less competent in athletics, behavior, and scholastics. Both teachers and parents rated 8- to 10-year-old children born at less than 28 weeks gestation or with birth weights less than 1,000 grams as having more behavioral symptoms (especially more somatic complaints and atypical behaviors), less adaptability, and fewer social and leadership skills than controls born with normal birth weights (Anderson and Doyle, 2003).

The evidence examining the relationship between low birth weight or preterm birth and autism is mixed. While some studies suggest a positive association (Finegan and Quarrington, 1979; Hultman et al., 2002; Indredavik et al., 2004; Larsson et al., 2005; Wilkerson et al., 2002), other studies have concluded that there is not an increased risk of autism in children born low birth weight or preterm (Deykin and MacMahon, 1980; Mason-Brothers et al., 1990; Piven et al., 1993; Williams et al., 2003). One study suggests that perinatal and obstetric factors might interact to impact birth outcomes (Eaton et al., 2001).

Conduct disorders are more common in children born preterm, but so are traits such as shyness, unassertiveness, withdrawn behavior, and social skill deficits (Bhutta et al., 2002; Grunau et al., 2004; Sommerfelt et al., 1996). In a meta-analysis of 16 case-control studies of children 5 years old or older who had been born preterm, 13 (81 percent) of the studies found that children born preterm had more behavioral problems than controls born fullterm (Bhutta et al., 2002). Two-thirds of the studies found a higher prevalence of ADHD, 69 percent found a higher prevalence of externalizing symptoms (e.g., delinquency), and 75 percent found a significantly higher prevalence of internalizing symptoms (e.g., anxiety, depression, and phobias). Many of these children withdraw from challenging tasks. Many preterm children with nonverbal learning disabilities have social skill deficits that seriously influence their social interactions and peer relationships (Aylward, 2002a; Fletcher et al., 1992). At school age, children born at less than 29 weeks of gestation were more often the target of verbal bullying by their peers than fullterm controls (Nadeau et al., 2004).

Adolescents and adults born preterm are less likely to demonstrate risk-taking behaviors than controls who were born fullterm. A recent study of young adults born from 1980 to 1983 in Britain found that fewer of those born with birth weights less than 1,500 grams than controls born fullterm drank alcohol or used illicit drugs, but there were no differences in the rates of smoking or sexual activity (Cooke, 2004). In the United States, young adults with birth weights less than 1,500 grams reported lower rates of alcohol and illicit drug use than controls with normal birth weights, but the two groups showed similar rates of tobacco use (57 and 59 percent of the men, respectively, and 40 and 48 percent of the women, respectively) (Hack et al., 2002). The men in the sample were also less likely to violate the law (37 and 52 percent, respectively), and this was primarily due to the lower rates of illicit drug use and truancy. The women were less likely to have had intercourse by age 20 years (65 and 78 percent, respectively) and to have had children (13 and 24 percent, respectively). These differences in risk-taking behaviors persisted even when the data for young adults with neurosensory impairments were excluded from the analysis. Both men and women with birth weights of less than 1,500 grams reported fewer delinquent behaviors, and women reported higher rates of anxiety-depression and withdrawal behaviors, fewer friends, and poorer family relationships than controls born with normal birth weights (Hack et al., 2004). An older study of Danish teenagers found no differences in the rates of alcohol or drug use between those born weighing less than 1,500 grams and controls born with normal birth weights (Bjerager et al., 1995).

Severity of Disability

Many outcomes researchers recognize the limitations of reporting the outcomes for individuals born preterm only in terms of the diagnoses of specific neurodevelopmental disabilities and have defined and reported on the severity of disability as well. For 8-year-olds born in 1991 and 1992 in Australia, severe disability (i.e., severe CP, blindness, or IQ scores 3 or more standard deviations below the mean) occurred in 9 percent of the children with birth weights of less than 1,000 grams and 12 percent of the children with birth weights of less than 750 grams (Doyle and Anderson, 2005). Moderate disability (i.e., moderate CP, deafness requiring hearing aides, or IQ scores 2 to 3 standard deviations below the mean) occurred in 10 percent of those with birth weights less than 1,000 grams and 15 percent of those with birth weights less than 750 grams. Mild disability (i.e., mild CP or IQ scores 1 to 2 standard deviations below the mean) occurred in 25 percent of children weighing less than 1,000 grams at birth and 33 percent of children weighing less than 750 grams at birth. More than half (56 percent) of the children with birth weights of less than 1,000 grams and 40 percent with birth weights of less than 750 grams had no disabilities.

Although a relatively small proportion of children born preterm have multiple disabilities, this group of children faces significant challenges with respect to mobility, academics, and the transition toward independence. Multiple disabilities were seen in 5 percent of kindergartners born at less than 28 weeks of gestation (Msall et al., 1992). By using the definitions of handicap based on the Education for All Handicapped Children Act (P.L. 94-142), 2.5 percent of the parents of children born with normal birth weights reported that their child had multiple handicaps, whereas 5 percent of the parents of children born with birth weights of between 1,501 and 2,500 grams, 12 percent of the parents of children born with birth weights of between 1,001 and 1,500 grams, and 14 percent of the parents of children born with birth weights of less than 1,000 grams reported that their child had multiple handicaps (Klebanov et al., 1994). In a sample of infants born with birth weights less than 1,000 grams, 14 percent of the infants born at greater than 32 weeks of gestation had a severe disability (Mercier et al., 2005). In contrast, severe disability was reported in 100 percent of the infants born at less than 22 weeks of gestation, 48 percent of those born at 22 to 23 weeks of gestation, 37 percent of those born at 24 to 25 weeks of gestation, and 25 percent of those born at 26 and 27 weeks of gestation.

The presence of any neurosensory and neurodevelopmental impairment has been investigated in a number of studies of preterm infants. These impairments include: moderate to severe CP, cognitive impairment 2 or more standard deviations below the mean, blindness in both eyes, and hearing loss requiring amplification in both ears. For children born preterm in the 1990s, 28 percent born at 27 to 32 weeks gestation, 13 to 25 percent weighing less than 1,000 grams at birth, 45 percent born at less than 27 weeks of gestation, 58 percent born at 24 weeks of gestation, and 61 percent born at less than 24 weeks of gestation had neurosensory or neurodevelopmental impairment, whereas 1 percent of controls born fullterm had such impairments (Hansen and Greisen, 2004; Hintz et al., 2005; Saigal et al., 2000c; Vohr et al., 2005; Wilson-Costello et al., 2005). Hack and colleagues (2002) reported a major neurodevelopmental disability rate of only 10 percent among young adults born with birth weights of less than 1,500 grams.

The probability of survival without a neurosensory impairment or a major disability increases with increasing gestational age. As many as 56 to 77 percent of children born with birth weights less than 1,000 grams and birth weights less than 750 grams survive and are free of major disability (Doyle and Anderson, 2005; Hack and Fanaroff, 1999; Hansen and Greisen, 2004; Piecuch et al., 1997a,b; Saigal et al., 1990; Wilson-Costello et al., 2005). However, among children born at less than 26 weeks gestation, only 20 percent were free of disability at age 6 years, with 34 percent of these children having a mild disability, 24 percent having a moderate disability, and 22 percent having a severe disability (Marlow et al., 2005). This is similar to the 21 percent of survivors born at gestational ages less than 25 weeks who had no impairments reported by Hintz and colleagues (2005).

Some use the term “intact survival,” calculated as the number of survi vors who are “normal” divided by all live births. Generally, “normal” means no major cerebral palsy, mental retardation, or severe sensory impairment (i.e., no neurosensory impairment). As a concept, intact survival is most useful at the limit of viability, when the mortality rate is so high. Two recent regional studies have reported survival without major disability: 0 to 0.7 percent for those with gestational ages of less than 23 weeks, 6 percent to 35 percent for those with gestational ages of 23 weeks, 13 to 42 percent for those with gestational ages of 24 weeks, and 31 to 56 percent for those with gestational ages of 25 weeks (Doyle, 2001; Wood et al., 2000). This concept of survival without disability may be useful for discussions with prospective parents as they face the impending delivery of their infant at less than 26 weeks gestation (see Appendix C).

Disability in Late Preterm or Near-Term Infants

The rates of mortality (discussed in Chapter 10) and neurodevelopmental disability for moderately preterm infants; that is, those born between 32 and 36 weeks gestation or those weighing 1,500 to 2,499 grams at birth, are higher than those for infants born fullterm (although they are lower than those for infants born more prematurely). Although children born between 32 and 36 weeks of gestation constitute only about 8 to 9 percent of all births, they account for 16 to 20 percent of children with CP (Hagberg et al., 1996; MacGillivray and Campbell, 1995). Children born late preterm (or near term) have also been reported to have more developmental delays of infant milestones (those born at 33 to 36 weeks gestation) and more difficulty with hyperactivity, fine motor skills, mathematics, speaking, reading, and writing (those born at 32 to 35 weeks gestation) (Hediger et al., 2002; Huddy et al., 2001).

Functional Outcomes

Authors of studies of the outcomes for children born preterm have struggled with how to convey the full range of outcomes. Although the majority have focused on cognitive outcomes and clinical diagnoses of neurodevelopmental disabilities, some have taken a more practical approach and have described what survivors of preterm birth have been able to do.

The acquisition of motor and adaptive milestones is a means of conveying the functional abilities of a toddler. Wood and colleagues (2000) found that at 30 months from term 90 percent of preterm infants born at less than 26 weeks gestation could walk, 97 percent could sit, 96 percent could feed themselves with their hands, and 6 percent could speak. In a similar study of children weighing less than 1,000 grams at birth, at 18 months from term 93 percent could sit, 83 percent could walk, and 86 percent could feed themselves (Vohr et al., 2000). A study of developmental data from the National Health and Nutrition Examination Survey found higher frequencies of motor and social developmental delays in children born preterm (at less than 37 weeks gestation), including the children born moderately preterm at 33 to 36 weeks of gestation (Hediger et al., 2002). Based on their rate of attainment of milestones compared with the rate for other children of the same age, for each week of gestation below full term, the developmental scores of boys decreased by 0.1 point.

Several studies have reported on the functional abilities of 5- to 6-year-old children born preterm. In a study of 149 children born at less than 28 weeks gestation who were in kindergarten, 95 percent could walk and perform basic self-care skills and were continent during the daytime (Msall et al., 1992). Most of the children who had neurodevelopmental disabilities were still able to function well: 87 percent could walk one block, 84 percent talked in sentences, and 81 percent could perform self-care tasks. In a study of 5-year-olds weighing less than 1,500 grams at birth, the results were reported in terms of the proportion of children with severe functional limitations (e.g., mobility, self-care, and social communication) (Palta et al., 2000). Severe functional limitations were identified in more children with CP (57 percent for self-care, 89 percent for mobility, and 32 percent for social communication) than in those without CP (5 percent for self-care, 21 percent for mobility, and 8 percent for social communication). Marlow and colleagues (2005) reported on the motor functions of 6-year-olds born at less than 26 weeks gestation: 6 percent could not walk, 6 percent walked independently but with an abnormal gait, and the remainder were functional with respect to walking and hand use.

A study of 5.5-year-old children born with birth weights of less than 1,250 grams found an increase in the proportions of children with severe functional limitations with an increasing severity of ROP: 4 percent with no ROP, 11 percent with prethreshold ROP, and 26 percent with severe threshold ROP (Msall et al., 2000). Severe functional limitations were common in children with severe ROP and poor vision (77 percent for self-care, 43 percent for mobility, 50 percent for continence, and 66 percent for social communication). Children who were born preterm and who had good vision fared better: 25 percent had self-care limitations, 5 percent had mobility limitations, 4.5 percent were not continent, and 22 percent had limitations in social communication skills.

Hack and Taylor (2000) found that 14-year-old children with weights of less than 750 grams at birth had significantly higher prevalences of functional limitations and a greater need for special services than children who weighed 750 to 1,499 grams at birth and controls born fullterm. Only 3 percent of those born weighing less than 750 grams and 2 percent of those born weighing 750 to 1,499 grams (and no controls) had severe functional impairments that interfered with feeding, dressing, washing, or toileting themselves. Compared with control children born fullterm, children with birth weights less than 750 grams were more likely to have mental or emotional delays, restrictions in activity, and visual difficulties (odds ratios = 4.7, 5.1, and 3.9, respectively). They were more likely to need special education, counseling, and special arrangements at school (odds ratios = 5.0, 4.8, and 9.5, respectively).

Later functional outcomes can be expressed in terms of the highest educational level achieved and the transition to adulthood. A study that linked data from the Swedish Medical Birth Registry with the National Service Enrollment Register found that not only did 18- to 19-year-olds born from 1973 to 1975 with birth weights of less than 1,500 grams have higher rates of CP (odds ratio = 55.4), mental retardation (odds ratio = 1.7), myopia (odds ratio = 3.3), and severe hearing impairment (odds ratio = 2.5); but they also tended to leave the school system early (odds ratio = 1.6) (Ericson and Kallen, 1998). Hack and colleagues (2002) found that fewer 20-year-olds born from 1977 to 1979 with birthweights less than 1,500 grams than controls who had normal birth weights graduated from high school or earned general equivalency diplomas (74 and 83 percent, respectively) and that fewer men attended 4-year colleges (16 and 44 percent, respectively). These differences persisted even when the data for those with any neurosensory impairment were excluded from the analyses.

Other investigators have found lower rates of graduation from high school among young adults born weighing less than 1,000 or 1,500 grams compared with the rates for controls with normal birth weights (Cooke et al., 2004; Lefebvre et al., 2005). In a comparison of performance on exams at the end of secondary school in Great Britain (i.e., the General Certificate of Secondary Education), matched control graduates born fullterm scored more points per test and more total points than graduates born with birth weights less than 1,500 grams (Pharoah et al., 2003). On the other hand, among a fairly privileged group of young adults born from 1977 to 1982 with birth weights less than 1,000 grams, Saigal and colleagues found no significant differences in the levels of educational attainment or the rates of continuing with higher education compared with those for the controls born with normal birth weights (Saigal et al., 2006a).

Tideman and colleagues (2001) reported that 19-year-olds born at less than 35 weeks gestation gave similar reports of self-esteem as controls born fullterm. Sixteen- to 19-year-olds born from 1981 to 1986 with birth weights of less than 800 grams reported less self-confidence with athletics, school achievement, job confidence, and romance than controls born fullterm (Grunau et al., 2004). Despite feeling less attractive than their peers who had been born with normal birth weights, young adults born from 1980 to 1983 with birth weights of less than 1,500 grams had similar social activities and similar experiences with sex (Cooke et al., 2004). More young adults who had been born preterm were parents, but they had lower rates of participation in higher education and paid employment.

Saigal and colleagues (2006a) found no significant differences in high school graduation rates, the level of education attained, rates of employment, rates of independent living, marriage or cohabitation status, or rates of parenthood between 22- and 25-year-olds born weighing between 501 and 1,000 grams and those born with normal birth weights. Subanalyses, however, revealed that more participants born with extremely low birth weights reported that they were neither in school nor employed, although these differences disappeared when those with disabilities were excluded. These results suggest that individuals born preterm can make a successful transition into adulthood. It is noted, however, that the participants in this sample were predominantly white, were from relatively advantaged homes, and had access to universal health care.

Summary

The spectrum of neurodevelopmental disabilities and functional outcomes in children, adolescents, and young adults born preterm is wide. Likewise, the prevalences of neurodevelopmental disabilities and neurosensory impairments in people born preterm are quite variable. This is not surprising, in light of the multiple etiologies and complications of preterm birth, the variability of both the intrauterine and the extrauterine environments to which they are exposed, and the infinite genetic variations of humans.

Finding 11-2: There is tremendous variation in the outcomes reported for individuals born preterm. Much of this variation is due to a lack of uniformity in study sample selection criteria, the study methodologies used, the age of evaluation, and the measurement tools and cutoffs used.

Finding 11-3: Few long-term studies of adolescents and adults born preterm have been conducted. Good indicators of the functional development of the central nervous system of preterm infants in neonatal intensive care units are lacking, and predictors of longterm neurodevelopmental and health outcomes are inadequate.

FACTORS THAT INFLUENCE NEURODEVELOPMENTAL OUTCOMES

Many factors influence the outcomes for infants born preterm, including gestational age at birth, complications that injure the brain, variations in the clinical management of infants at various health care institutions, the family’s socioeconomic condition, and the mother’s mental health. As discussed above, children who were born more prematurely and who were sicker tend to have more health and developmental problems. The marked variations in neurodevelopmental and health outcomes for preterm infants is the result of variations in the etiologies of preterm birth, intrauterine environments, complications of prematurity, NICU management and treatments, and home environments. Although many factors influence the outcomes for individuals born preterm, the many studies that have evaluated demographic, prenatal, perinatal, and neonatal predictors have not been able to devise a method of determining the outcome for an individual infant. Predictors of outcomes can identify groups of infants with a high risk of disability who may benefit from intensive developmental support and community services. Further research to refine the predictors may also provide insights into the etiologies and the mechanisms of organ injury and recovery in preterm infants.

The processes that lead to preterm birth and the subsequent treatments necessary to support life (i.e., neonatal intensive care) can injure immature organ systems that are not ready to support life in the extrauterine environment. Maternal illness can influence fetal growth, fetal organ development, preterm delivery, and infant health and neurodevelopmental outcomes. A growing concern is the impact of infection and inflammation on the mother and the fetus: inflammatory mediators have been implicated in proposed mechanisms of preterm delivery and fetal brain injury leading to CP and cognitive impairments (Andrews et al., 2000; Dammann et al., 2002, 2005; Goldenberg et al., 2005; Gravett and Novy, 1997; Hagberg et al., 2005; Holling and Leviton, 1999; Wu and Colford, 2000) (see Chapters 6 and 10). Research is needed to understand the relationships between the genome, factors that control inflammation, and development of the brain, the neurotransmitter system within the brain, and immune system (Raju et al., 2005).

Although prenatal and intrapartum factors can adversely influence the outcomes for infants born preterm, many neonatal factors are stronger predictors (Allen, 2005b). The strongest neonatal predictors are the severity of the acute illness, measures of bronchopulmonary dysplasia/chronic lung disease (BPD/CLD), severe ROP, and signs of brain injury (Allen, 2005b; Doyle, 2001; Emsley et al., 1998; Hack et al., 2000; Hintz et al., 2005; Piecuch et al., 1997a,b; Schmidt et al., 2003; Vohr et al., 2005). The effect of BPD/CLD on neurodevelopment may be due to difficulties with nutrition and growth, management of ventilators and oxygen supplementation, or the use of postnatal steroids or other medications (Bhutta and Ohlsson, 1998; Halliday and Ehrenkranz, 2001a,b,c; Kaiser et al., 2005; Saugstad, 2005; Thomas et al., 2003; Tin, 2002; Tin and Wariyar, 2002; Wood et al., 2005). Because various management strategies influence outcomes, one factor that would be expected to influence outcomes is the institution(s) at which the infant was managed clinically (Vohr et al., 2004).

Signs of brain injury on neuroimaging studies (e.g., ultrasound or magnetic resonance imaging), neurodevelopmental examination, and analysis of infant movements are some of the strongest neonatal predictors of preterm motor and cognitive outcomes. Predictive neuroimaging findings include severe intraventricular hemorrhage, intraparenchymal hemorrhage, hydrocephalus, porencephaly, periventricular leukomalacia, and other signs of white matter injury (Doyle, 2001; Hack et al., 2000; Hintz et al., 2005; Ment et al., 2003; Piecuch et al., 1997a,b; Pinto-Martin et al., 1995; Rogers et al., 1994; Tudehope et al., 1995). Measures of infant neurological function, including neurodevelopmental examination and assessment of general movements, are independently predictive and enhance the predictive capability when they are used in combination with neuroimaging (Allen and Capute, 1989; Einspieler and Prechtl, 2005; Gosselin et al., 2005; Mercuri et al., 2005). Neuroimaging technologies have begun to demonstrate reduced brain volumes, involving the cortex (especially parietal and sensorimotor areas) and deep nuclear structures, in infants and children born preterm as compared with fullterm control children (Peterson et al., 2000; Peterson, 2003). Reduction in brain volume has been associated with white matter injury and cognitive deficits. Research into the nature of brain-behavior relationships, including relationships between brain structural and functional development, areas of the brain typically affected by brain insult and their corresponding neurodevelopmental and behavioral deficits, and how plasticity is manifested as recovery from brain injury could provide insight into better predictors of neurodevelopmental outcomes as well as neuroprotective strategies (Peterson, 2003; Raju et al., 2005).

The developing preterm infant is not immune to environmental factors, and preterm birth may interact with poverty toward further disadvantage. There is a wealth of information on how a child’s low socioeconomic status adversely affects the outcomes for the child in general (Brooks-Gunn and Duncan, 1997). It is also well documented that preterm infants are disproportionately poor (see Chapter 4). Not only does poverty increase the risk of being born preterm but it also independently increases the risk of adverse outcomes. The effect of poverty directly relates to the design of postdischarge interventions. Even controlling for birth weight or prematu rity, poverty worsens the outcomes for preterm infant, particularly cognitive outcomes (McGauhey, 1991).

There is some evidence that maternal mental health influences the outcomes for children born preterm. In general, children of depressed mothers do not fare well (independently of poverty, which increases the risk of depression) (Downey and Coyne, 1990; Zuckerman and Beardslee, 1987). As discussed above, the mothers of preterm infants experience more depression, and that affects their parenting abilities (Singer et al., 1999, 2003). Maternal mental distress early in a child’s life has long-term effects on child behavior (Gray et al., 2004).

Evidence suggests that educational and other programs may improve outcomes and have especially been shown to mitigate the effects of poverty (Barnett, 1995; Devaney et al., 1997; Yoshikawa, 1995). Potentially beneficial effects from developmental support in the NICU are discussed in Chapter 10. Early intervention programs that may affect the outcomes of premature infants after discharge from the NICU are discussed below. Research into preterm health and neurodevelopmental outcomes should address the complex and intricate relationships between biomedical, proximal, and distal environmental and behavioral influences. Consideration should be given to specific and discrete outcomes, ages at time of exposure to environmental variables, mediators and moderators of environmental influences, and use of multivariate modeling.

HEALTH AND GROWTH

Health

As with much of the literature reviewed in this volume, studies of the health of premature infants after discharge from the hospital have been characterized largely in terms of birth weight (Doyle et al., 2003b). The most frequently cited evidence of a higher risk for adverse health status among low birthweight and preterm infants is an increased risk of rehospitalization during the first few years of life. Infants born weighing 1,500 grams or less are four times more likely than normal birth weight infants to be hospitalized in the first year of life (McCormick et al., 1980) and are more likely to have a disproportionate duration of stay for these hospitalizations (Cavalier et al., 1996). Among infants born with birth weights of 2,500 grams or less, the relative risk of rehospitalization in one study was about twice that of heavier infants, and again, these infants had longer lengths of stay in the hospital (Cavalier et al., 1996).

More recently, attention has been shifted to the risk of rehospitalization associated with prematurity. Although the findings of these studies are difficult to compare with those from the older literature based on birth weight, these studies suggest that preterm infants are also more likely to be rehospitalized (Escobar et al., 1999; Martens et al., 2004). For example, among infants born preterm, those born at earlier gestations compared to those born moderately preterm were at greater risk of rehospitalization (Joffe et al., 1999). Children born with birth weights below 2,500 grams also make more use of outpatient health care (Jackson et al., 2001) and incur significantly higher medical and nonmedical costs (McCormick et al, 1991) compared to children born with normal birth weights. Furthermore, as with other measures of health care utilization, the rates of rehospitalization vary among institutions (Escobar et al., 2005; Martens et al., 2004).

The increased risk of rehospitalization for preterm and low birth weight infants is likely a reflection of their compromised health status. Children born with birth weights below 1,500 grams suffer increased morbidity (McCormick et al., 1992) compared to children with normal birth weights. The psychosocial environment is also important for children born with birth weights below 2,500 grams, as those with high psychosocial risk have worse health status than children in low and moderate risk categories (McGauhey et al., 1991). Furthermore, in comparison to children with normal birth weights, chronic health conditions have a stronger impact on the school achievement and participation, and behavior problems of children with birth weights below 2,500 grams (McGauhey et al., 1991). The impact of LBW extends into adolesence. Adolescents with birth weights below 1,500 grams have higher blood pressure than those with normal birth weights (Doyle et al., 2003a).

Growth

Besides acute and chronic conditions, infants born preterm or with birth weights below 2,500 grams also experience poorer growth. The first 3 years of life evidence a discrepancy in the growth patterns of children with birth weights below 2,500 grams, compared to those with normal birth weights (Binkin et al., 1988; Casey et al., 1991). Poor growth resulting from intrauterine, neonatal, or postnatal growth failure has been documented widely among children with birth weights below 1,500 grams (Binkin et al., 1988; Casey et al., 1991). Studies performed with adolescents who with birth weights less than 2,500 grams suggest that their anthropometric measurements are lower than those of adolescents with normal birth weights. Similarly, a study by Peralta-Carcelen and colleagues (2000) showed lower growth measures for adolescents born at birth weights less than 1,000 grams who survived without a major neurodevelopmental disability, compared to those with normal birth weights.

However, other studies suggest that there is catch-up growth. Hack and colleagues (1984) found that catch-up growth occurs during the first 2 to 3 years of life. A more recent study of children with birth weights less than 1,500 grams documented catch-up growth during the first 8 years of life, with poorer growth attainment among those who were small for gestational age (SGA) (Hack et al., 1996). Ford et al. (2000) also documented catch-up growth among children with birth weight less than 1,500 grams although they are still smaller than those born with normal birth weights. Another study also found that 18–19 year old boys with birth weights less than 1,500 grams were shorter and lighter than their counterparts with normal birth weights (Ericson and Kallen, 1998). Finally, Saigal and colleagues (2001) found that adolescents with birth weights less than 1,000 grams demonstrated patterns of catch-up growth between age 8 and adolescence.

Studies on the ultimate growth attainment or growth during the adolescent years and into early adulthood for individuals born preterm are only recently appearing in the literature. Doyle and colleagues (2004b) found compromised growth among survivors up to age 8 who were born with birth weights less than 1,000 grams, but by age 14 and up to age 20 they had reached average height and weight. Despite the persistence of lower height among survivors with birth weights less than 1,000 grams, Saigal and colleagues (2001) showed that most of their adolescents were within 2 standard deviations of the mean. Furthermore, Hack and colleagues (2003) documented gender differences. In childhood (8 years), males with birth weights less than 1,500 grams were shorter and lighter than counterparts with normal birth weight, whereas females were lighter but not significantly shorter than their counterparts with normal birth weights (Hack et al., 2003). The discrepancy for males persisted at 20 years, but females did not demonstrate subaverage height and weight at early adulthood (Hack et al., 2003). The predictors of height and weight also differed for males and females. For females, black race and chronic illness predicted weight, and maternal height and birthweight standardized score predicted height at 20 years of age (Hack et al., 2003). The predictors of height for males were the same as the females’ with duration of neonatal hospital stay and SGA birth as additional predictors of height (Hack et al., 2003).

Many of these findings suggest that by adolescence children born preterm experience catch-up growth. Since sexual maturation is an important aspect of adolescent development, studies on growth attainment for preterm infants during the adolescent years should evaluate sexual maturity. Another area for further research to elucidate the factors involved in catch-up growth among LBW survivors is nutrition. Weiler and colleagues (2002) found that despite the impact of preterm birth on attainment of normal height in young adulthood, bone mass is appropriate.

Health-Related Quality of Life

Health, as formulated by the World Health Organization (WHO, 1958) is a “state” of complete physical, mental, and social well-being and not merely the absence of disease or infirmity. Health-related quality of life (HRQL) is a narrower concept that considers the net impact or consequence of a disease or impairment and implicitly reflects the personal values of the individual (Gill and Feinstein, 1994). Measurement of HRQL can be used for comparisons with different disease conditions, as well as for cost-effectiveness and cost-utility analyses.

The challenge in pediatrics is that children are constantly developing and changing; and their personal values may also evolve over time (Rosenbaum and Saigal, 1996). Traditionally, parents or caregivers have been accepted as reliable proxy respondents on behalf of younger children or those with severe disabilities. However, there is evidence that proxy responses by parents correlate poorly with the perceptions of children. Parents have more consistent agreement with children for observable functioning (physical health), than for emotional and social functioning. Overall, parents are generally more negative, and their responses may be influenced by the burden of caregiving (Eiser and Morse, 2001). Also, health professionals have limited abilities to judge patients’ HRQL, and their values may differ from those of children and their parents (Saigal et al., 1999). Further work is required to understand how the characteristics of patient proxy responses influence agreement. However, it is acknowledged that although children’s perspectives may differ from that of parents or health professionals, they are valid, and should be accepted. There are strong arguments for obtaining HRQL perspectives from multiple respondents (Eiser and Morse, 2001). Recently a few “generic” and “disease-specific” HRQL instruments been developed to measure the physical, psychological, and social domains of health in young children from their own perspective. To date, there are limited studies on Quality of Life (QL) in children, and even fewer in children born preterm. In this section a few such studies conducted at various ages are described.

On a preschool quality of life questionnaire (TAPQOL) (Fekkes et al., 2000) administered to parents and neonatologists in the Netherlands, 1- to 4-year-old children born before 32 weeks gestation were reported to have significantly lower HRQL than the reference group (Theunissen et al., 2001). However, the study found differences between the neonatologists’ and parents’ perceptions of HRQL in terms of what conditions needed treatment. A Canadian study from British Columbia (Klassen et al., 2004) used both the Infant Toddler Quality of Life Questionnaire (ITQOL) (Landgraf et al., 1999) and the Health Status Classification System for Preschool children (Saigal et al., 2005a) and found that the 1,140 children who required NICU care at birth had poorer health status and HRQL on a range of domains compared with the findings for 393 children who had been born fullterm.

A series of studies on the HRQL of children born in Ontario Canada between 1977 and 1982 weighing less than 1,000 grams were conducted by Saigal and colleagues. In the initial study of 156 8-year-old children who weighed less than 1,000 grams at birth and 145 fullterm controls (Saigal et al., 1994b), health status was determined by health professionals using a multiattribute classification system, which also provided the levels of functional limitations (Feeny et al., 1992). This descriptive information was used to map HRQL scores using a utility function formula (HUI2), based on preferences about hypothetical health states expressed by the general public (Torrance, 1995). The children with birth weights less than 1,000 grams had functional limitations in several attributes, and mean HRQL scores were significantly lower than in controls (Saigal et al., 1994b). In a subsequent study of the same cohort (Saigal et al., 1996), directly measured preferences were elicited for the first time at adolescence using the standard gamble technique (method for eliciting patient preferences when there is uncertainty about the outcome). The teenagers born weighing less than 1,000 grams reported a higher frequency and more severe functional limitations in their health status in several domains as compared to fullterm controls. Although their overall mean HRQL scores were significantly lower, the majority of adolescents with birth weight less than 1,000 grams viewed their HRQL to be similar to normal birth weight controls (71% vs 73%). Parents of both groups rated their adolescent children’s HRQL higher than the children’s self-ratings (Saigal et al., 2000b).

In a further study to determine whether the preferences of health professionals differed systematically from those of the Ontario teenagers and their parents, hypothetical health scenarios were employed, and preferences were elicited by standard gamble technique (Saigal et al., 1999). Although there was a fair degree of agreement between the groups for milder disabilities, parents and teenagers appeared to be more “accepting” of the more severely disabling health states than health professionals. These findings have clinical implications for medical decision making around the birth of a very preterm infant.

Other investigators have measured parental perspectives of HRQL in 244 10-year-old children who weighed less than 1,250 grams at birth. These children participated in the landmark CRYO trial for retinopathy of prematurity (ROP) (CRPCG, 1988). Using the HUI2 system (Feeny et al., 1992), threshold ROP was associated with functional limitations in health and reduction in HRQL scores and, as expected, children with poorer visual outcomes had lower HRQL scores (Quinn et al., 2004). HRQL in relation to the severity of abnormality on brain ultrasounds was measured in a cohort of adolescents born preterm (Feingold et al., 2002). Paradoxically, the overall self-perceived HRQL of adolescents with a higher degree of severity of intraventricular hemorrhage was better than that of adolescents with mild (i.e., grade 0 to 2) intraventricular hemorrhage.

Recently, several studies have reported the HRQL in young adults born preterm. In a telephone survey, the quality of life of 85 Danish young adults born between 1971 and 1974 in Denmark with birth weights less than 1,500 grams was compared with that of subjects with normal birth weights by using both objective and subjective measures (Bjerager et al., 1995). Although the self-reported quality of life scores for those with physical and mental handicaps were significantly lower, the scores of nondisabled subjects were comparable to those of the group with normal birth weights. In a subsequent study in Denmark, a cohort of young adults born from 1980 to 1982 with birth weights less than 1,500 grams had similar subjective quality-of-life scores but lower objective quality-of-life scores compared to the reference group (Dinesen and Greisen, 2001). Tideman and colleagues (2001) found that on the visual analogue scale, the HRQL of 19-year-olds born at less than 35 weeks gestation was similar to the HRQL of the normal birth weight group. The quality of life of a British cohort of young adults with birth weights less than 1,500 grams assessed by SF-36 (Short Form 36 Health Survey) was found to be similar to that of normal birth weight group (Cooke 2004). Similarly, at a mean age of 23 years, Saigal and colleagues (2006b) found no differences in the self-reported mean HRQL scores in the Ontario young adults weighing less than 1,000 grams compared to normal birth weight controls.

Although measurement of quality of life appears to be popular, a well-defined theoretical framework for the accurate assessment of the child’s conceptual and developmental viewpoint is not yet available (Jenney and Campbell, 1997). Furthermore, most of the currently available quality-of-life measures focus to a large extent on how children are functioning in different domains as determined by age-appropriate roles. Although functional measures provide considerable valuable information, they are considered to be tapping health profiles and not the values of individual subjects (Gill and Feinstein, 1994). Such health profiles generally paint a picture more negative than that obtained when personal valuation is sought (Guyatt and Cook, 1994). However, in-person interviews are time consuming and expensive and impose considerable cognitive demands on respondents. Feeny et al. (2004) compared directly measured standard gamble utility scores of the Ontario cohort at adolescence with indirectly measured scores from the HUI, based on self-assessed health status but valued using community preferences. Although HUI scores matched directly measured utility scores reasonably well at the group level, at the individual level they were poor substitutes for directly measured preferences.

Finally, despite the growing literature, measurement of self-reported quality of life continues to be viewed with skepticism by the medical community and even by some parents (Hack, 1999; Harrison, 2001). Although a considerable challenges the view that the quality of life of people with disabilities is inevitably compromised, the prevailing perception appears to be largely negative.

Fetal Origins of Adult Disease

The “fetal origins” hypothesis proposes that undernutrition in utero at critical periods of development, programs or permanently alters fetal metabolism and renders the individual susceptible to future cardiovascular disease (CVD) and metabolic derangements of glucose metabolism (Barker et al., 1989a). This hypothesis originated in a retrospective study in which it was noted that men whose birth weights were below the 5th percentile had a higher risk of dying from coronary artery disease than men with a higher birth weight. Subsequent investigations with other populations have confirmed this relationship (Barker et al., 1993a; Martyn et al., 1998) and noted that it also exists in women (Osmond et al., 1993; Osmond et al., 2000; Rich-Edwards et al., 1997).

Low birth weight is reported to be associated with established cardiovascular risk factors, including dysglycemia, dyslipidemia, and hypertension (Barker et al., 1993b; Lithell et al., 1996). Upon meta-analysis, however, the association between birth weight and blood pressure or dyslipidemia is weaker than initially thought, with the larger studies showing the smallest relationship (Huxley et al., 2002). In a recent meta-analysis, the pooled relative risk for coronary artery disease among the 6,056 subjects who weighed 5.5 pounds or less at birth was 1.26 (95% CI 1.11– 1.44) compared with the risk for the 80,802 subjects who weighed more than 5.5 pounds at birth (Raju, 1995). Animal studies have provided supportive evidence in favor of this programming hypothesis, although often after the use of extreme nutritional insults during fetal life (Coates et al., 1983).

Despite the growing number of reports on an association between low birth weight, cardiovascular risk factors and surrogate markers of CVD, most studies investigating the fetal origins of adult disease have used a retrospective design for data collection (Barker et al., 1989a,b, 1993a,Barker et al., b,c; Leeson et al., 2001). Furthermore, many of these studies were conducted with highly selected subgroups of people for whom very little information on pregnancy and perinatal events was available. In addition, most investigators did not differentiate low birth weight from small for gestational age or prematurity, and did not measure postnatal growth. Thus, birth weight and early growth were used as surrogates for overall somatic growth, with out data on the presence or the absence of interim growth decelerations or subsequent catch-up growth. Furthermore, social class (paternal occupation) was based on recall by adult subjects 50 to 70 years later and the studies did not control for postnatal modifiers, such as socioeconomic, environmental or behavioral factors, or social deprivation in the early critical period of life (Joseph and Kramer, 1996; Paneth, 1994; Paneth and Susser, 1995; Paneth et al., 1996). In addition, retention rates in most study cohorts were extremely poor with only between 19 and 60 percent of the subjects available for further follow-up (Bhargava et al., 2004; Cooke, 2004; Strauss, 2000).

Only a paucity of studies have been designed to investigate specifically whether the fetal origins hypothesis is also applicable to preterm infants and not just to those who are small for gestational age. Fewtrell and colleagues (2000) examined the relationship between gestational age and size for gestational age on glucose and insulin concentrations at ages 9 to12 years in 385 children who had been born preterm with birth weights less than 1,850 grams. Low birth weight, whether it was due to being born preterm or intrauterine growth restriction, was associated with higher plasma glucose levels 30 minutes after administration of a glucose load. Recently, Hofman and colleagues (2004a) have demonstrated that 4- to 10-year- olds born preterm have metabolic abnormalities similar to those observed in infants born fullterm but small for gestational age and that these occur irrespective of whether the preterm infants are small or appropriate for gestational age. In fact, there did not seem to be an additive effect on reduced sensitivity from being born both preterm and small for gestational age. A subsequent study by the same investigators confirms the reduction in insulin sensitivity, which may be a risk factor for Type II diabetes mellitus (Hofman et al., 2004b). This reduction was similar in infants born between 24 and 32 weeks gestation, suggesting that a critical window exists in the third trimester in which insulin activity is altered. In another study by Hovi et al. (2005), young adults born with birth weights less than 1,500 grams had fasting insulin levels that were 34 percent higher than those for controls, and their mean fasting serum glucose level was also higher (but an oral glucose tolerance test was not done). Unfortunately, these studies examined only 50 percent of the cohort.

Childhood weight gain has also been shown to be an important predictor of measures of insulin secretion and resistance in some studies (Fewtrell et al., 2000). Singhal and colleagues (2003a) have shown that preterm infants with birth weight less than 1,850 grams who received nutrientenriched formula had higher fasting 32–33 split pro-insulin levels (a marker of insulin resistance) at adolescence. This effect of postnatal diet was a proxy for greater weight gain in infants in the first 2 weeks of life, independently of birth weight, gestational age, and other sociodemographic factors. The authors propose that relative undernutrition in preterm infants early in life may actually have beneficial long-term effects on insulin resistance. Similar beneficial effects on vascular structure and endothelial function were also observed (Singhal et al., 2004). These studies have raised further controversy regarding the nutritional management of very preterm infants and what should be considered “optimal” catch-up growth.

Other studies have also shown high rates of type II diabetes in individuals who were small for gestational age at birth and who later became overweight as adults (Bavdekar et al., 1999; Eriksson et al., 1999; Newsome et al., 2003). In a recent prospective longitudinal study of 1,492 Indian subjects 26 to 32 years of age, the growth of children in whom impaired glucose tolerance or diabetes later developed was characterized by a low body mass index between birth and 2 years of age, followed by an early adiposity rebound and a sustained and accelerated increase in body mass index until adulthood (Bhargava et al., 2004). In two other studies with young adults, individuals who experienced the largest increase in body mass index and those who remained overweight over time had evidence of vascular change manifest by increased common carotid intima-media thickness (CIMT) (to estimate cardiovascular risk) (Eriksson et al., 2001; Oren et al., 2003). Thus the association of low birth weight and later CVD and metabolic factors is very likely modified by postnatal factors, although this has not been adequately studied.

Body composition, specifically the distribution of the fat and lean bone mass compartments, may also be important predictors of risk of CVD, hypertension and diabetes in adult life. Fat mass and fat-free mass were lower in 8- to 12-year-old children born with birth weights less than 1,850 grams than in children born with normal birth weights (Fewtrell et al., 2004). Such findings may reflect programming of body composition by early growth and nutrition. Indeed, a higher birth weight was associated with a greater fat free mass in adolescents (Singhal et al., 2003b). The authors suggest that an association of low birth weight and lower lean mass may be the underpinnings of programming for suboptimal insulin sensitivity, lower metabolic activity, and a subsequent propensity to greater adiposity and risk of CVD (Singhal et al., 2003b).

Using whole-body magnetic resonance, Uthaya and colleagues (2005) have recently shown that by the time that infants born preterm reached their term age, they had a highly significant decrease in subcutaneous adipose tissue and significantly increased levels of intra-abdominal adipose tissue. They caution that preterm infants may be at risk of metabolic complications later in life through this increased and aberrant adiposity. In a cohort of 132 20-year-old individuals who had been born small for gestational age and average for gestational age and who were born fullterm (Levitt et al., 2005), the association between low birth weight and expres sion of adult chronic cardiometabolic disease was not dependent on birth weight alone, but was also dependent on its interaction with subsequent fat accumulation (either generally or abdominally) (Levitt et al., 2005).

Fewer studies have explored the association of preterm birth and CVD in adulthood. Irving and colleagues (2000) investigated 61 young adults who had been born with low birth weights less than 2,000 grams at a mean age of 24 years and showed that those who were small because of prematurity were also at risk of hypertension, an adverse metabolic profile (higher plasma insulin triglyceride and total cholesterol levels and lower high-density lipoprotein cholesterol levels) and hyperglycemia as adults. Among the preterm cohort, those who were small for gestational age were not measurably more disadvantaged than those who were average for gestational age. CIMT studies, however, were not performed. A study conducted in the Netherlands attempted to elucidate the effects of prenatal and infancy growth on the lipid and CIMT measures in a very preterm cohort at age 19 years (Martin et al., 2006). Their findings support an effect of current body composition rather than early growth on CVD risk. Two recent studies (Doyle et al., 2003; Hack et al., 2005a) have shown higher systolic blood pressure among very low birth weight infants in late adolescence and young adulthood. However, no relationship was found between intrauterine growth and blood pressure. Not all studies have found higher blood pressure in preterm subjects in childhood (Morley et al., 1994) or at young adulthood (Saigal et al., 2005). Further prospective, long-term studies of preterm infants monitored to adulthood are warranted to confirm whether preterm infants are at increased risk for CVD and metabolic problems as adults.

IMPACT OF PRETERM BIRTH ON FAMILIES

Families caring for a child born preterm face long-term and multilayered challenges. The limited research on this topic suggests that this impact is largely negative (Beckman and Pokorni, 1988; Cronin et al., 1995; Davis et al., 2003; Eisengart et al., 2003; Lee et al., 1991; Macey et al., 1987; McCain, 1990; McCormick et al., 1986; Singer et al., 1999; Stjernqvist and Svenningsen, 1995; Taylor et al., 2001; Veddovi et al., 2001), although some studies found positive outcomes (Macey et al., 1987; Saigal et al., 2000a; Singer et al., 1999). Furthermore, the impact varies according to sociodemographic risk factors as well as the severity of the child’s health condition (Beckman and Pokorni, 1988; Cronin et al., 1995; Davis et al., 2003; Eisengart et al., 2003; Lee et al., 1991; McCormick et al., 1986; Rivers et al., 1987; Saigal et al., 2000a; Singer et al., 1999; Taylor et al., 2001; Veddovi et al., 2001).

Most studies on the impact of caring for a preterm infant have focused on those born at less than 32 weeks gestation (Davis et al., 2003) and less than 35 weeks of gestation (Veddovi et al., 2001), although others studied infants with birth weights less than 1,500 grams or less than 1,750 grams (Eisengart et al., 2003; Macey et al., 1987; Singer et al., 1999). Others have used prematurity and low birth weight as a continuous variable (Beckman and Pokorni, 1988). The assessment of outcomes has centered on the mother’s psychological well-being in the postpartum period and suggests that the mothers of infants born preterm are at risk of experiencing depressive symptoms (Davis et al., 2003; Singer et al., 1999; Veddovi et al., 2001). Longitudinal studies of children born preterm and with low birth weights in the first 2 to 3 years of life suggest that the levels of maternal depression and psychological distress (Singer et al., 1999), as well as problems related to the child, decreased over time (Beckman and Pokorni, 1988) except among high-risk (defined as having bronchopulmonary dysplasia) infants (Singer et al., 1999). Furthermore, specific factors that may contribute to depressive symptoms include a higher medical risk for the infants, the less frequent use of informal networks to obtain information about their infants, increased use of escape-avoidance coping strategies, and less knowledge of infant development (Eisengart et al., 2003; Veddovi et al., 2001). On the other hand, factors that might buffer these mothers from depressive symptoms include a higher level of educational attainment and support from nurses (Davis et al., 2003).

Families caring for a child who was born preterm continue to manage the effects of prematurity when the children are toddlers (Lee et al., 1991; McCormick et al., 1986; Singer et al., 1999), school age (Cronin et al., 1995; Lee et al., 1991; McCain, 1990; Rivers et al., 1987; Taylor et al., 2001), and adolescents (Saigal et al., 2000a). Studies focusing on these children have mainly included children who were born weighing less than 2,500 grams (Cronin et al., 1995; Lee et al., 1991; McCormick et al., 1986; Rivers et al., 1987; Singer et al., 1999; Taylor et al., 2001); and only one focused on children born weighing less than 1,000 grams (Saigal et al., 2000a). Their findings suggest that the impact on families is long term and that the parents, siblings, finances, and family functioning are all affected (Cronin et al., 1995; Saigal et al., 2000a; Singer et al., 1999; Taylor et al., 2001). Furthermore, the families of children with more severe levels of impairment are the most affected (Cronin et al., 1995; Rivers et al., 1987; Saigal et al., 2000a; Singer et al., 1999; Taylor, 2001).

At the individual level of the impact of a preterm birth on the family, the parents of children born preterm report higher levels of emotional distress (Saigal et al., 2000a; Singer et al., 1999; Taylor et al., 2001) and strain and a compromised sense of mastery (Cronin et al., 1995). One study suggests that some of the factors that parents associate with higher stress levels might include supervision of the child, the child’s peer relationships and self-esteem, the impact of the child’s difficulties on family routines, and worrying about the child’s future (Taylor et al., 2001). The length of time that the newborn preterm infant must stay in the hospital also affects the ability of the mother to fulfill her role in the family (McCain, 1990).

Other studies suggest that there might be gender role differences in parents’ perception of problems. Mothers perceived that the preterm birth of a child had a greater impact on their sense of mastery, finances, and employment (Cronin et al., 1995). They also perceived greater satisfaction in caring for their child (Cronin et al., 1995). The mothers also perceived a greater impact when the child was born at a younger gestational age (Lee et al., 1991), experienced more physical symptoms during the pregnancy, and were more likely than the fathers to experience crisis reactions (Stjernqvist, 1992). On the other hand, fathers perceived greater uncertainty, less individual strain (Cronin et al., 1995), and greater effects at lower levels of progression of the infant’s development (Lee et al., 1991).

Beyond the impact on each of the parents individually, caring for children born preterm affects other units within the family, including the couple, the siblings, and the family as a whole (Beckman and Pokorni, 1988; Cronin et al., 1995; Macey et al., 1987; McCormick et al., 1986; Saigal et al., 2000a; Singer et al., 1999; Stjernqvist, 1992; Taylor et al., 2001). Specifically, the parent’s marital relationship is stressed (Macey et al., 1987; Stjernqvist, 1992), at times leading to divorce (Saigal et al., 2000a), and parenting difficulties emerge (Taylor et al., 2001). Siblings are affected because of the decreased attention that they receive from their parents (Saigal et al., 2000a). The family as a unit is affected by the greater likelihood of not having additional children (Cronin et al., 1995; Saigal et al., 2000a), the financial burden (Cronin et al., 1995; Macey et al., 1987; McCormick et al., 1986; Rivers et al., 1987), limits on family social life (Cronin et al., 1995; McCormick et al., 1986), high levels of adverse family outcomes (family stress and dysfunction) (Beckman and Pokorni, 1988; Singer et al., 1999; Taylor et al., 2001), and parents’ difficulty maintaining employment (Macey et al., 1987; Saigal et al., 2000a). Lower income and education place an additional burden on families caring for children born preterm (Cronin et al., 1995; McCormick et al., 1986; Taylor et al., 2001), although one study found that the higher medical risks faced by neonates had more significant impacts on socioeconomically advantaged families (Taylor et al., 2001).

Furthermore, different factors predict family stress at different ages (Beckman and Pokorni, 1988). When the neonate born preterm was 3 months of age, it was found that informal support, the number of siblings, and the family’s socioeconomic condition were the most important factors; at 6 months of age, gestational age at birth, home environment, caregiving demands, and the number of parents in the home were the most important; at 12 months of age, race, home environment, and scores on the Bayley scales of infant development were the most important; and at 24 months of age, birth weight at birth, informal social support, temperament, caregiving demands, and race were the most important (Beckman and Pokorni, 1988).

Families and parents also have positive experiences and demonstrate resilience in caring for a child with impairments related to preterm birth. A study by Saigal and colleagues (2000a) found that parents perceived positive interactions with friends and within the family stemming from their efforts to care for their child born with birth weight less than 1,000 grams. The parents also reported enhanced personal feelings and improved marital closeness (Saigal et al., 2000a). Macey and colleagues (1987) found that at 12 months (corrected for prematurity), 50 percent of the infants’ mothers perceived their marriage to be more cohesive. Other studies suggest that these parents perceive their children to be acceptable, attached, and reinforcing (Singer et al., 1999) and to have a greater appreciation for their child than was the case when the child was an infant (Rivers et al., 1987). Thus, the impact of caring for a child born preterm may also contribute to the growth of the family as well as its members.

In summary, the limited evidence presented here suggests that caring for a child born preterm has negative and positive impacts on the family that change over time, that these impacts extend to adolescence and are influenced by different environmental factors across time, and that many areas of family well-being are affected. However, because of the limitations of these studies, further research is needed. First, these findings are limited in their generalizability because of a lack of ethnic and socioeconomic diversity in the samples and because a higher proportion of mothers than fathers were surveyed. Research should strive to balance these sociodemographic factors in the samples used. Second, the measures used to determine the effects of a child born preterm on the family and the child’s functional health were not uniform across studies. For example, the effects on the family were measured as the economic burden, parental symptomatology, and parenting stress, among others. Similarly, the child’s health and functional health status were assessed on the basis of the presence of serious health conditions in one study, whereas other studies formally assessed functional health status by the use of validated measures.

Future studies could advance knowledge in this area by developing a measure that would capture the particular health and functional health challenges that these families and the children born preterm face. In a recent review of functional health outcomes of preterm children, Donohue (2002) suggested that measures that are sensitive to the child’s developmental stage should be developed for children and that the measures for parents should focus on the peculiarities of their role as caregivers for these children.

Third, the few longitudinal studies reviewed in this section suggest that future research is needed to study changes in the impact of a child born preterm on the family over time. The fourth limitation noted in the studies reviewed were the variations in the gestational ages and the birth weights of the infants. Researchers should be encouraged to focus on prematurity by gestational age in addition to birth weight, so that the variations in the impacts on families can be ascertained by gestational age. Finally, studies of the impacts of an infant born preterm on families during the child’s infancy should assess outcomes beyond maternal depressive symptoms in the post-partum period.

POST-NICU DISCHARGE INTERVENTIONS

In recognition of the increased developmental and emotional risks for children born preterm, several interventions have focused on the provision of services in the early years of life to prevent subsequent developmental and health problems. Coordinated, community-based, multidisciplinary programs for early intervention, based on the findings of some seminal studies, have been established for children and their families. The types and severities of the conditions affecting children with disabilities are varied, and so are the intensity and the extent of the services provided. Research suggests that these programs may be effective in improving some cognitive outcomes in individual children and can also lead to important improvements in family function (Berlin et al., 1998; Majnemer, 1998; McCormick et al., 1998; Ramey et al., 1992; Ramey and Ramey, 1999). However, longterm follow-up of the children in some of these studies has shown mixed results, with some evidence that differences apparent within 3 years of an intervention all but disappear after time.

Early Infant and Childhood Interventions

Several longitudinal studies have attempted to ascertain the effects of early intervention on the emotional, physical, and developmental outcomes in children born preterm or with disabilities. The Infant Health and Development Program (IHDP) is a multicentered, randomized, controlled, U.S. nationwide study of preterm infants born in 1985 at gestational ages of less than 37 weeks and with birth weights of less than 2,500 grams and their families. Infants and their families were randomly assigned to either the intervention group (n = 377) or the follow-up-only (FUO) group (n = 608) within two birth weight strata: less than 2,000 grams and 2,000 to 2,499 grams. For their first 3 years, both groups received medical, developmental, and social assessments, as well as referrals for services such as health care. An educational intervention for infants and families in the intervention group consisted of home visits (weekly during the first year and every other week thereafter), enrollment in a child development center at 12 months from the due date, and parent group meetings (Ramey et al., 1992). The educational sessions at home and the center encouraged parents to use games and activities to promote their child’s cognitive, language, and social skills development; and parents were provided with information on health, safety, and child-rearing topics.

At the outcome evaluation at 36 months from the due date, the children in the intervention group had higher cognitive scores (14 points higher for those with birth weights of 2,000 to 2,499 grams and 7 points higher for those with birth weights of less than 2,000 grams) and fewer behavioral problems than the children in the FUO group (Brooks-Gunn et al., 1992b; McCormick et al., 1993). Receptive language, visual motor, and spatial skills were also improved for those in the intervention group. Even among the smallest infants; that is, those with birth weights of less than 1,500 grams and less than 1,000 grams, IQ scores were higher and behavior was better with the early intervention. The effects were the greatest for the highest-risk children, whose parents had no more than a high school education or were of an ethnic-racial minority status (a,Brooks-Gunn, 1992b). The effects were long lasting for children with birth weights of 2,000 to 2,499 grams. The children of well-educated mothers did not benefit from the intervention (McCormick et al., 1998). Mothers who had less than a high school education reported less emotional distress as a result of the intervention (Klebanov et al., 2001). These findings suggest that early intervention programs should especially target children and families at risk for poor outcomes.

The cohort in the IHDP study was again evaluated at 5, 8, and 18 years of age (a,Brooks-Gunn, 1992b; McCarton et al., 1997; McCormick et al., 2006). Among children with birth weights of 2,000 to 2,499 grams, the differences in IQ scores, behavior, and math and reading achievement persisted, although for the IQ scores the difference decreased to 4 points. In adolescence, the intervention group reported lower rates of engagement in risky behavior (e.g., substance use or delinquency). These findings are consistent with those of other long-term studies of single-site educational interventions for poor healthy children (Campbell et al., 2002; Reynolds et al., 2001; Belfield et al., 2006). The lack of a persistent difference in those with birth weights of less than 2,000 grams raises questions about their subsequent experiences and the need for more sustained support for neurologically vulnerable children.

Avon Premature Infant Project

The United Kingdom Avon Premature Infant Project was a randomized controlled trial in which the parents of 284 infants born at less than 33 weeks gestation received a home-based developmental education program, a social support intervention, or standard care (Johnson et al., 2005). A fullterm reference population served as a control group. Although there were some differences in cognitive, motor, and behavioral outcomes at 2 years of age, there were no differences at 5 years of age (mean age, 58 months and 15 days) among the intervention groups. The children born preterm had poorer cognitive performance than their peers born fullterm. Further analyses, in which the outcomes data were adjusted for social factors, did not reveal any differences between the intervention groups or between subgroups classified by a range of perinatal variables. The authors concluded that the small advantage shown at 2 years of age is no longer detectable at 5 years of age and questioned the effectiveness of early intervention in sustained cognitive, behavioral, and motor functions.

National Early Intervention Longitudinal Study

The National Early Intervention Longitudinal Study (NEILS), sponsored by the Office of Special Education Programs of the U.S. Department of Education, is monitoring more than 3,338 children who have disabilities or who are at risk for disabilities and their families through their experiences in early intervention and into early elementary school. Information about the characteristics of the children and their families, the services that they receive, and the outcomes that they experience is being collected. A nationally representative sample of children between birth and 31 months of age and their families who began early intervention services for the first time between September 1997 and November 1998 has been recruited for the study. NEILS is focusing on, among other issues, the early intervention services that participating children and families receive and the outcomes that participating children and families experience. Because this study will also assess how outcomes relate to variations in child and family characteristics and the services that they received, it has particular relevance for infants born preterm and their families.

A three-stage stratified sampling procedure was used to identify the original sample for the study. Twenty states were selected on the basis of the number of children served in early intervention and the region of the country. These states represented considerable variation with regard to the lead agency and whether or not the agencies served children at risk. The second stage involved the selection of counties on the basis of the estimated number of children served in Part C programs.1 Three to seven counties were selected within each state, for a total of 93 counties. The children ranged in age from birth to 30 months when they began receiving early intervention services (between 1997 and 1998).

The initial results from this study have been favorable. In a 2004 report by Bailey et al., it was found that most parents considered early intervention to have had a significant impact on their families, reporting that their families were much better off (59 percent) or somewhat better off (23 percent) as a result of the help and information that the early intervention program provided. Most parents (96 percent) also believed that they were able to help their children learn and develop, although in comparison, when they were asked about their perceived competence in caring for their child’s basic needs, fewer (64 percent) reported strong agreement and more (32 percent) reported simple agreement.

A separate assessment of functional status over the time of the intervention showed that the proportion of children with vision, hearing, or motor skills problems stayed constant over the period of the intervention. However, some children who had problems with communication when the services began showed improvement over time (Markowitz, 2004). The assessment also found that 96 percent of the families reported that the intervention helped them become more proficient in working with professionals and advocating for their child’s needs.

Summary

Although the short-term impact of early interventions has been well demonstrated, the findings of evaluations of the long-term impact of early interventions for preterm infants have been ambiguous. Long-term prenatal and perinatal cohort studies conducted before the introduction of neonatal intensive care concluded that social factors and the quality of the home environment can compensate for the disadvantages encountered perinatally and neonatally (Wolke, 1998). Recent evidence shows that intervention providing social and environmental enhancement through home visits and child development programs, is associated with catch-up in cognitive and behavioral development in large preterm infants, especially those from socieconomically disadvantaged backgrounds (Brooks-Gunn et al., 1994; Olds and Kitzman, 1993; Ramey and Ramey, 1999). This suggests that these larger preterm infants may not have persistent central nervous system insults. In contrast, although early interventions may have an impact on the outcomes for smaller preterm infants, biological factors may be the best predictors of cognitive and behavioral outcomes at school age.

However, McCormick (1997) and others have argued that the lack of comparability across studies that use such a broad categorization of morbidity is but one methodological flaw recurring in the follow-up literature. Other methodological problems include the failure to characterize the study samples by the eligibility for the study and the number of losses in the cohort, the failure to provide sufficient information with which the representativeness of the sample can be assessed, and the failure to use appropriate controls (McCormick, 1997). In addition, the outcomes being assessed may be too limited. Finally, even for the outcomes selected, many studies fail to incorporate a specific underlying pathogenic or conceptual model to identify potential factors influencing the relationship between the initial state (i.e., prematurity or low birth weight) and the outcomes observed (McCormick, 1997).

Finding 11-4: Early childhood educational and other therapeutic research interventions have been demonstrated to improve outcomes for some infants born preterm; however, it is critical to determine the appropriate intensity, type of service, personnel, and curricula to achieve improvement in interventions.

CONCLUSION

There is a wide range of health and neurodevelopmental outcomes for infants born preterm, and many resources are required to provide the necessary medical, neurodevelopmental and educational support for the children and support for their families. More outcomes data are reported by birth weight categories than by gestational age categories, but until better measures of organ maturation are available, information regarding gestational age is necessary for medical decision making and parent counseling when a preterm delivery is anticipated. Because of their long-term impact, health care providers should focus not on preterm birth but on degree of organ maturity at birth and on short and long-term neurodevelopmental, functional, and health outcomes. Just as the etiologies of preterm birth are multifactorial, the neurodevelopmental, functional, and health outcomes of infants born preterm are determined by interactions among the genome, intrauterine environment, high-risk obstetric and neonatal intensive care provided, the home environment, and available community resources. Future research to develop better predictors of outcomes should focus on the relationships between brain structural and functional development, areas of the brain typically affected by brain insult and corresponding neurodevelopmental and behavioral deficits, and how organ recovery and plasticity occur. Better predictors of outcomes will allow for improved parent counseling, enhance safety of trials of maternal and infant interventions by providing more immediate feedback, and facilitate planning for use of comprehensive follow-up and early intervention resources. Until preterm birth can be prevented, much work also needs to be done to develop treatment strategies that prevent injury to the brain and other organs and support the infant’s ongoing development.

Part C of the Individuals with Disabilities Education Act elevated the family component of early intervention to a new level. This legislation replaced the Individualized Education Program for children ages 3 to 21 years with the Individualized Family Service Plan for infants and toddlers with disabilities.

Footnotes

1

Part C of the Individuals with Disabilities Education Act elevated the family component of early intervention to a new level. This legislation replaced the Individualized Education Program for children ages 3 to 21 years with the Individualized Family Service Plan for infants and toddlers with disabilities.

Copyright © 2007, National Academy of Sciences.
Bookshelf ID: NBK11356
PubReader format: click here to try

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (3.1M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...