The goal of this proposal is twofold: (1) To study the neurobehavioural development of children with a known pre-/neonatal and postnatal history of environmental PCB-exposure in order to improve risk-assessment for PCBs; (2) To extend the period of validation for early indicators of neurodevelopmental deficit beyond 18 months of age (done in a previous project) until six years of age.
Measurable objectives: Associations between early and later PCB-exposure and neuromotor or cognitive functions will be described and tested for statistical significance by means of multiple linear regression analysis. The strength of observed associations will be evaluated relative to the influence of other important factors (as e.g. social background, home environment etc.). In addition, the quality of analytical and outcome measures will be assessed by means of ringtests as well as by interrater reliability exercises.

Design of Study: In this study existing cohorts will be followed in two parts covering the upper age limits of
42 months (Part I: D|sseldorf, Faroe Islands) and 72 months (Part II: Dutch groups), respectively. Within each
of these segments between 300 - 400 children will be studied.
PART I (D|sseldorf, Faroe Islands): In this section of the follow-up study cohorts from D|sseldorf (N = 150
- 180) and the Faroe Islands (N = 160 - 190) will be retested at ages 30 and 42 months. The overall sample
size is expected at about 312 at age 30 months and 295 at 42 months of age. The basic independent variables
are PCB-levels in plasma and in maternal milk, as available from the ongoing prospective studies. Additional
PCB-levels will be measured at age 42 months. Main functions tested include neurology, language and cognitive
development (Bayley, Kaufman and Stanford-Binet).
PART II (Groningen,Rotterdam): In this part of the follow-up study the cohorts from Groningen (N=212) and
Rotterdam (N=207), each of them subdivided into 50% formulafed and breastfed babies, will be retested at
schoolage (6 years), after having already been retested at age 3 1/2 years (42 months) previously. Again, the
basic independent variables are prenatal PCB-concentrations (cordblood, maternal blood) and early postnatal
levels in maternal milk as measured during the early stages of recruitment between 1989 and 1991; PCB-levels
in serum are also available from the 42 months examination, and will also be measured at 6 years of age. Main
dependent variables are neurology, cognitive development (Kaufman Scales) and behaviour ratings (Child
Behavior Check List).
Statistical data analysis for PARTS I and II: Apart from descriptive summary data the main part of the overall
analysis will be based on multiple logistic regression-models for binary data, and on multiple linear
regression-models for continuous data. Adjustment for relevant confounders, as e.g. social status, obstetrical
optimality, maternal IQ and quality of the home environment, will be an important aspect of final statistical
modelling. Particular care is taken to also consider co-exposure to other neurotoxicant chemicals, as e.g. lead
(German and Dutch studies) and methylmercury (Danish/Faroe Island study element).