By clicking “Accept”, you agree to the storing of cookies on your device to enhance site navigation, analyse site usage, and assist in our marketing efforts. View our Privacy Policy for more information.
Coffee & Health

The impact of paternal alcohol, tobacco, caffeine use and physical activity on offspring mental health: a systematic review and meta-analysis

K E Easey & G C Sharp, 2021.
Reproductive Health, Volume 18 (1).
October 28, 2021



There is some evidence that paternal health behaviours during and around pregnancy could be associated with offspring health outcomes. However, the impact that paternal health behaviours during pregnancy can have on offspring mental health is understudied and remains unclear.


We conducted a systematic review and meta-analysis of articles in PubMed describing studies of potentially modifiable paternal health behaviours (tobacco smoking, alcohol consumption, caffeine consumption and physical activity) in the prenatal period in relation to offspring mental health. GRADE was used to measure risk of bias.


Eight studies were included and categorized by paternal health behaviour and offspring mental health outcome investigated. The narrative synthesis provided evidence of association between paternal health behaviours around pregnancy and offspring mental health problems, with the strongest evidence shown for tobacco use. Grouped by analysis type, two separate meta-analyses showed evidence of paternal smoking during pregnancy being associated with greater odds of ADHD in offspring (OR 1.42, 95% CI 1.02-1.99; HR 1.28, 95% CI 1.19-1.39).


The small number of studies that have investigated paternal prenatal effects on offspring mental health, and the limited sample sizes of those studies, makes it challenging to draw firm conclusions. Although existing studies suggest that paternal tobacco smoking and alcohol consumption in the prenatal period are associated with poorer offspring mental health, (particularly hyperactivity/ADHD), further investigation of potential paternal effects is required, using methods that allow stronger inference to determine whether associations are causal.

More research

All research