Birthweight and blood pressure in five European birth cohort studies: an investigation of confounding factors.

BACKGROUND: It has been suggested that the association between birthweight and blood pressure has been overstated as a result of publication bias and, within studies, a lack of adjustment for potentially important maternal and socioeconomic confounding factors and 'overadjustment' for current body size. This study investigates the impact of potential confounding variables on the birthweight-blood pressure association in birth cohort studies from different time periods and geographical locations in Europe. METHODS: Data from five European birth cohort studies (from Finland, the UK, and the Faroe Islands) taking part in the European Birth-Lifecourse-Studies (EURO-BLCS) project were analysed. Birthweight was measured at birth in all cohorts and confounding variable information was collected prospectively at subsequent follow-ups in all cohorts. Regression models were used to assess the unadjusted association between birthweight and blood pressure and then to assess the impact of potential maternal and socioeconomic confounding variables and adjustment for later body size. Analyses were carried out in the same way across all five cohorts. RESULTS: Birthweight was consistently negatively associated with systolic blood pressure (SBP) across all cohorts. Gestational age and possibly maternal pre-pregnancy weight, but not socioeconomic status, may be important confounding factors of the relationship between birthweight and SBP. The size of the birthweight-SBP association in adulthood may be larger than in childhood before adjustment for current body size, although a cohort effect cannot be ruled out. CONCLUSION: This study highlights the value of future cross-cohort comparisons in the investigation of the foetal origins of adult disease.

Variation at the insulin gene VNTR (variable number tandem repeat) polymorphism and early growth: studies in a large Finnish birth cohort.

Variation at the insulin gene (INS-)VNTR (variable number of tandem repeats) minisatellite polymorphism has been reported to be associated with both early growth and adult metabolic phenotypes. However, the samples studied have been small and the relationship between INS-VNTR variation and parameters of early growth inconsistent, with four previous studies producing conflicting results. We have studied the relationship between INS-VNTR class (measured by genotyping the nearby -23HphI variant with which it is in tight linkage disequilibrium) and early growth in 5,646 members of the Northern Finnish Birth Cohort of 1966. Comparing class III homozygotes with other genotypes using multivariate linear regression analysis, we found no significant associations with any early growth measure (birth weight, birth length, ponderal index, and head circumference at 1 year), even after stratifying subjects by growth trajectory during infancy and/or birth order. For example, among infants with limited postnatal growth realignment (n = 2,470), class III/III infants were no heavier at birth (difference [+/-SE] in the means [fully adjusted], 58 +/- 51 g; P = 0.26) than class I/- infants. No significant associations were detected following reanalysis with an additive model (for example, for birth weight, beta = 20 g [95% CI -3 to 44], P = 0.09). Studies of this large population-based cohort have failed to generate convincing evidence that INS-VNTR variation influences early growth.