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).