Nutrition in early life interacts with genetic risk to influence preadult behaviour in the Raine Study

Lars Meinertz Byg; Carol A Wang; John Attia; Andrew J O Whitehouse; Wendy H Oddy; Jonathan J Hirst; Craig E Pennell

doi:10.1038/s43856-025-01339-y

Nutrition in early life interacts with genetic risk to influence preadult behaviour in the Raine Study

Commun Med (Lond). 2026 Feb 3;6(1):76. doi: 10.1038/s43856-025-01339-y.

Authors

Lars Meinertz Byg^{1

2}, Carol A Wang^{1

2}, John Attia³, Andrew J O Whitehouse⁴, Wendy H Oddy⁵, Jonathan J Hirst¹, Craig E Pennell^{6

7

8}

Affiliations

¹ Mothers and Babies Research Center, Hunter Medical Research Institute, Newcastle, NSW, Australia.
² School of Medicine and Public Health, University of Newcastle, Newcastle, NSW, Australia.
³ Department of General Medicine, John Hunter Hospital, Newcastle, NSW, Australia.
⁴ The Kids Research Institute Australia, University of Western Australia, Perth, WA, Australia.
⁵ Menzies Institute for Medical Research, University of Tasmania, Hobart, TAS, Australia.
⁶ Mothers and Babies Research Center, Hunter Medical Research Institute, Newcastle, NSW, Australia. Craig.Pennell@newcastle.edu.au.
⁷ School of Medicine and Public Health, University of Newcastle, Newcastle, NSW, Australia. Craig.Pennell@newcastle.edu.au.
⁸ Department of Maternity and Gynaecology, John Hunter Hospital, Newcastle, NSW, Australia. Craig.Pennell@newcastle.edu.au.

Abstract

Background: Early life nutrition is associated with child behaviour; however, the interplay with genetic vulnerability is understudied. We hypothesised that psychiatric genetic risk interacted with early nutrition to predict behavioural problems in childhood and adolescence.

Methods: The Raine Study participants with genetic information aged 2-17 were repeatedly evaluated with the child behaviour checklist total problems score (CBCL_TOT). Breastfeeding duration was recalled at age 1, 2 and 3 follow-up, and toddler diet derived by an age-1 24-h maternal recall (EAT1, scale 0-70, SD 10, higher scores proxying healthy diet). We derived polygenic scores (PGS) impacting general psychopathology: attention-deficit hyperactivity disorder (ADHD), depression, chronic multisite pain (CMSP), total behaviour problems and birthweight. In confounder-adjusted mixed-effects models of CBCL_TOT throughout follow-up we examined nutrition-by-PGS interactions.

Results: In 1393 participants, a borderline signal suggests that 1 month longer breastfeeding reduces CBCL_TOT by -0.108 (95% CI [-0.184, -0.0289]) exclusively in individuals with a higher CMSP PGS (Interaction p = 0.03). In 1310 participants, a strong signal suggests that 1 EAT1 point increase results in a reduced CBCL_TOT by 0.121 points (95% CI [-0.171, -0.0704]) exclusively in individuals with a lower ADHD PGS (Interaction p = 0.0005). Post hoc analysis suggests that plant-based food consumption drives the favourable EAT1-CBCL_TOT association.

Conclusions: Nutrition in early life and psychiatric genetic risk may interact to determine lasting child behaviour. Contrary to our hypothesis, we find dietary benefits in individuals with lower ADHD PGS, necessitating replication. We also highlight the possibility of including genetics in early nutrition intervention trials for causal inference.

Plain language summary

It has been proposed that both early nutrition and genetic risk influence behaviour in childhood and adolescence. However, the interplay between the two remains poorly understood. We examined the interaction between genetic and early nutritional data with repeated parent ratings of behaviour from birth until age 17 years in Australian children. We saw fewer behavioural problems in those with a healthy diet at one year of age if the children had a low risk of developing ADHD based on their genetics. In contrast, longer breastfeeding duration primarily reduced behavioural problems in those with a higher risk of chronic pain based on their genetics. Our results suggest that tailoring early nutrition based on genetic risk could reduce pre-adult behaviour problems. However this association should be further tested in clinical studies.