Intelligence test at preschool-age predicts reading difficulty among school-aged very low birth weight infants in Japan.

Objective: To elucidate whether the results of an intelligence test at preschool age are predictive of reading difficulty (RD) at school age among very low birth weight infants (VLBWI).

Methods: Subjects were 48 Japanese children whose birth weight was <1500 g and who regularly visited a follow-up clinic. All subjects completed the Wechsler Intelligence Scale for Children-III (WISC-III) during the last grade of kindergarten, and four reading tasks during the second to fourth grade of elementary school. All participants had a full-scale intelligence quotient score of 85 or higher. Subjects with a standard deviation reading time score greater than 2.0 in two or more tasks were considered to have RD. We evaluated the associations between each WISC-III score and RD using logistic regression analyses. Furthermore, we performed receiver operating characteristic (ROC) analysis to determine a cutoff WISC-III score predictive of RD.

Results: In the mutually-adjusted model, the adjusted odds ratio per 1 score increase of freedom from distractibility (FD) was 0.832 (95% confidence interval: 0.720-0.962). In the ROC analysis, an FD score of <95.5 was chosen as the cutoff value for predicting RD (sensitivity, 0.77; specificity, 0.74).

Conclusion: The present study indicated that a lower FD score at preschool age, which was associated with deficits in verbal working memory and attention, is a risk factor for RD at school age among Japanese VLBWI. Further investigation is desired to clarify the cognitive deficits underlying RD in Japanese-speaking preterm children, and to establish appropriate interventions for these children.