The reversing association between advanced maternal age and child cognitive ability: evidence from three UK birth cohorts

Abstract

Background: Studies on advanced maternal age-defined here as age 35 or older-and children's cognitive ability report mixed evidence. Previous studies have not analysed how the time period considered in existing studies influences the association.

Methods: We analysed trends in the association between maternal age and cognitive ability using data from the 1958 National Child Development Study ( n = 10 969), the 1970 British Cohort Study ( n = 9362) and the 2000-2002 Millennium Cohort Study ( n = 11 600). The dependent variable measures cognitive ability at age 10/11 years. Cognitive scores were standardised to a mean of zero and a standard deviation of one.

Results: For the 1958-70 cohort studies, maternal ages 35 -39 were negatively associated with children's cognitive ability compared with maternal ages 25-29 (1958 cohort β = -0.06 standard deviations (SD) 95% confidence interval (CI): -0.13, -0.00; 1970 cohort β = -0.12 SD 95% CI: -0.20, -0.03). By contrast, for the 2000-2002 cohort study maternal ages 35-39 were positively associated with cognitive ability (β = 0.16 SD 95% CI: 0.09, 0.23). For maternal ages 40+, the pattern was qualitatively similar. These cross-cohort differences were explained by the fact that in the earlier cohorts advanced maternal age was associated with high parity, whereas in the 2000-2002 cohort it was associated with socioeconomically advantaged family background.

Conclusions: The association between advanced maternal age and children's cognitive ability changed from negative in the 1958 and 1970 cohorts to positive in the 2000-2002 cohort because of changing parental characteristics. The time period considered can constitute an important factor in determining the association between maternal age and cognitive ability.

Keywords: Advanced maternal age; cognitive ability; secular change.

Figure 1

Coefficients from regression models of cognitive ability on maternal age; Model 1 (Panel A), Model 2 (Panel B) and Model 3 (Panel C). 1958, 1970, and 2001 refer to the 1958, 1970 and 2001 Cohort Studies. Panel A: Model 1 adjusted for cohort member sex and multiple birth. Panel B: Model 2 adjusted for cohort member sex, multiple birth and birth order. Panel C: Model 3 adjusted for cohort member sex, multiple birth, birth order, mother’s education, mother’s marital status at the time of birth, father’s social class, mother’s smoking during pregnancy, mother had antenatal care after 12 weeks of pregnancy, mother’s height and breastfeeding. Results for the MCS are weighted to account for its complex survey design.