Maternal smoking in pregnancy moderates the effect of the brain-derived neurotrophic factor (BDNF) gene polymorphism on childhood IQ across multiple ages

Maternal smoking during pregnancy and intelligence quotient of offspring aged 18 and 30 years: Evidence from two birth cohorts in southern Brazil

Maternal smoking during pregnancy causes several harmful effects, including deficits in the intelligence quotient (IQ), a measure associated with academic achievements and higher socioeconomic position. We aimed to measure the association between maternal smoking during pregnancy and offspring's IQ in two birth cohorts from Pelotas, Brazil. Data from the 1982 and 1993 birth cohorts were analyzed. Smoking during pregnancy was ascertained at the perinatal follow-up, and IQ was measured at 30 years (1982) and 18 years (1993). Simple and multiple linear regressions were used for crude and adjusted analysis. Mediation analysis was performed using birth weight and breastfeeding as possible mediators. Data analysis was conducted in 2020. There were 3611 and 4050 participants from the 1982 and 1993 cohorts, respectively. After adjustment for potential confounders, the inverse association between smoking during pregnancy and children's IQ remained significant in both cohorts. Offspring of mothers who smoked during pregnancy had an average of 1.32 less IQ points (95%CI: -2.23, -0.242) in the 1982 cohort, and an average of 1.66 less IQ points (95%CI: -2.42, -0.90) in the 1993 cohort. Neither cohort showed association with paternal smoking (negative controls) after adjustment. Breastfeeding's mediated effects accounted for 26.2% and 23.9% of the association in the 1982 and 1993 cohorts, respectively, while birth weight's accounted for 6.8% and 30.1%, respectively; indirect effects were not significant. The inverse association between maternal smoking and IQ and the lack of association with paternal smoking reinforces our findings of a negative association between exposure and outcome.

Maternal smoking during pregnancy and intelligence quotient in offspring: A systematic review and meta-analysis

BACKGROUND

Exposure to tobacco during pregnancy may disrupt fetal brain development and impact offspring cognitive development.

AIMS

To perform a systematic review and meta-analysis on maternal smoking during pregnancy and intelligence quotient (IQ) in childhood, adolescence, and adulthood.

METHODS

We searched PubMed, Lilacs, PsycINFO, and Web of Science. Original articles evaluating tobacco use/exposure during pregnancy and the offspring's IQ as the outcome. The review protocol is registered in PROSPERO (number CRD 42,019,116,257). For the meta-analysis, we included studies with information on the regression coefficient and its confidence interval (CI) or standard error. Random effects model was used for pooling the estimates.

RESULTS

25 studies were included in the review, and of these 14 met the inclusion criteria for the meta-analysis. The overall pooled estimate showed that subjects who were exposed to maternal smoking during pregnancy presented lower IQ scores, compared to those not exposed to maternal smoking (β -1.30; 95 % CI -1.74, -0.86; I² = 87.8 %); IQ scores were also lower in crude (β -5.46; 95 % CI -7.31, -3.60; I²: 79.0 %) and adjusted pooled estimates (β =-0.45; 95 % CI -0.76, -0.13; I² = 80.4 %), for the group exposed to maternal smoking. In the stratified analysis, an inverse association was also observed in studies with large sample size (n≥1000 participants) (β=-0.49; 95 % CI -0.96, -0.02), among those performed with adolescents (β=-1.16; 95 % CI -2.18, -0.14), and among those adjusted for maternal education (β=-0.57; 95 % CI -1.05, -0.08).

CONCLUSIONS

Our findings suggest that exposure to tobacco during pregnancy may have negative effects on IQ. However, the findings of this meta-analysis should be interpreted with caution.

Genotype-environment interaction on human cognitive function conditioned on the status of breastfeeding and maternal smoking around birth

We estimated genotype by environment interaction (G × E) on later cognitive performance and educational attainment across four unique environments, i.e. 1) breastfed without maternal smoking, 2) breastfed with maternal smoking, 3) non-breastfed without maternal smoking and 4) non-breastfed with maternal smoking, using a novel design and statistical approach that was facilitated by the availability of datasets with the genome-wide single nucleotide polymorphisms (SNPs). There was significant G × E for both fluid intelligence (p-value = 1.0E-03) and educational attainment (p-value = 8.3E-05) when comparing genetic effects in the group of individuals who were breastfed without maternal smoking with those not breastfed without maternal smoking. There was also significant G × E for fluid intelligence (p-value = 3.9E-05) when comparing the group of individuals who were breastfed with maternal smoking with those not breastfed without maternal smoking. Genome-wide significant SNPs were different between different environmental groups. Genomic prediction accuracies were significantly higher when using the target and discovery sample from the same environmental group than when using those from the different environmental groups. This finding demonstrates G × E has important implications for future studies on the genetic architecture, genome-wide association studies and genomic predictions.