Elsevier

The Lancet

Volume 380, Supplement 3, 23 November 2012, Page S23
The Lancet

Abstracts
Early-life determinants of cognitive ability in childhood and old age

https://doi.org/10.1016/S0140-6736(13)60379-7Get rights and content

Abstract

Background

Factors early in life, even prenatally, could affect subsequent health and development of disease. Studies have consistently shown a small but significant correlation between birthweight in the normal range and childhood cognitive ability. However, only two studies have investigated whether this correlation persists into older age, showing no significant effect. The aim of this investigation was to assess whether factors early in life (birthweight, gestational age, parity, antenatal complications, neonatal complications, and social group), affect cognition in childhood and whether these effects persist.

Methods

The Aberdeen 1936 Birth Cohort (ABC1936) is a longitudinal study of cognitive ageing in a well-characterised cohort of individuals born in 1936 (n=506 of 658 alive and resident in Aberdeen, were recruited in 2000). The study is ongoing with wave 4 of testing completed in 2011 (n=216). Only wave 1 (baseline) data are presented here. In wave 1, participants underwent a battery of cognitive and physical tests at age 64–66 years. Results for Raven's Progressive Matrices (RPM), a test of non-verbal reasoning, are reported here. The test was administered by a trained research nurse and individuals had 20 min to complete as many of the multiple choice problems as possible. All 658 individuals invited to participate in ABC1936 had IQ measured with the Moray House Test when aged 11 years as part of the nationwide Scottish Mental Survey 1947. Some were born in hospital, and therefore had prospectively obtained data available for perinatal effects. Birth data were obtained for all hospital births in Aberdeen in 1936 (n=1250). Results are reported for individuals who had birth data, IQ at age 11 years, and IQ at age 64–66 years available. Associations between cognitive ability scores (at both ages) and potential predictors (sex, social group at birth, birthweight, gestational age, parity, antenatal or neonatal complications) were tested. Factors significantly associated with cognitive ability (p<0·1) were included in multivariate regression models for the prediction of childhood IQ and RPM scores.

Findings

Of participants with childhood IQ available, 140 (21%) had birth data available. Of those, 108 (77%) had IQ scores at age 64–66 years. Mean IQ scores did not differ between those with birth data and those without at either age 11 years (with birth data, mean 38·3 [SD 13·8]; without birth data, mean 40·3 [14·0]; p=0·06) or age 64–66 years (with birth data, mean 34·7 [8·2]; without birth data, mean 35·8 [8·8]; p=0·21). Birthweight was not significantly correlated with childhood cognitive ability (r=0·07; p=0·39). Multivariate regression analysis showed that low perinatal social group (unskilled manual work compared with others; β=–0·23; p=0·007) and increased parity (β=–0·21; p=0·015) were associated with decreased test scores in childhood (adjusted R2=0·09). Neither birthweight (r=–0·003; p=0·77) nor perinatal social group (p=0·56) were significantly associated with RPM scores. In older age, increased childhood cognitive test scores (β=0·47; p<0·0001) and male sex (β=0·31; p<0·0001) were associated with improved performance (adjusted R2=0·35).

Interpretation

In this fairly small cohort, birthweight and childhood or older age IQ were not associated. Childhood IQ was predicted by social group and parity, and was the main predictor of older age IQ. Longitudinal data from more than six decades, including prospectively recorded birth data, are rare. The reasons for differences between this cohort and others in Scotland and worldwide should be explored, and could include cohort effects, residual confounding factors, and selection bias. Early-life programming of cognitive function is not a significant determinant in this cohort.

Funding

The Alzheimer's Research Trust funded data collection during wave 1 of ABC1936; RLS is funded by an NRS Career Researcher Fellowship.

References (0)

Cited by (0)

View full text