Leveraging a large, population-based cohort with genotyped parent-offspring trios and pairs, this study provides the first comprehensive investigation of genetic nurture effects across multiple substances and across adulthood, revealing important insights into the persistence and mechanisms underlying intergenerational transmission of substance use. We highlight four key findings. First, genetic nurture significantly influenced smoking quantity (cigarettes per day and pack-years), but not smoking initiation, cannabis initiation, or daily alcohol intake, demonstrating substance specificity. In contrast, direct genetic transmission played a substantial role across all measured outcomes. Second, while direct genetic effects on smoking quantity diminished over time, genetic nurture effects remained consistent throughout adulthood. Third, parent-of-origin analyses revealed that the magnitudes of genetic nurture effects on smoking quantity were nearly identical for mothers and fathers. Finally, parental smoking quantity mediated genetic nurture effects, with maternal smoking having a stronger impact on offspring smoking than paternal smoking, particularly among daughters.

Our results demonstrate the specificity of genetic nurture in substance use: significant effects were observed for smoking quantity (cigarettes per day, pack-years) but not for initiation of smoking or cannabis, and daily alcohol intake. This pattern aligns with evidence that genetic and environmental contributions vary across different dimensions of substance use^6,7,23–25, such as initiation, quantity, and dependence. The observed genetic nurture for smoking quantity reflects passive gene-environment correlation (rGE)²⁶, whereby parental genotypes shape the rearing environment in which smoking is frequently modeled. Lifetime CPD and pack-years capture sustained smoking patterns involving long-term regulation and reinforcement^27,28. These cumulative behaviors are more likely influenced by the enduring family environment shaped by parental genetics.

In contrast, the absence of genetic nurture for tobacco or cannabis initiation, and daily alcohol intake may reflect the environmental factors operating independently of parental genetics, the transient nature of these behaviors, and measurement characteristics. Tobacco and cannabis initiation represent discrete behavioral events occurring within peer networks and immediate social contexts^29–31, while daily alcohol consumption fluctuates with social situations and life events (e.g., being a college student, celebrations or holidays)^32,33. Rather than passive rGE (i.e., genetic nurture), these behaviors may be driven by other types of rGE, such as active rGE, whereby individuals select environments (e.g., peer groups, social settings) that align with their own genetic predispositions, facilitating initial substance use^34,35. Methodologically, dichotomous measures of initiation may lack sensitivity to detect subtle genetic nurture effects, despite large sample sizes (N = 15,853 for smoking initiation). Similarly, our short-term assessment of alcohol use (past 30 days) may not adequately capture the stable, long-term drinking patterns³⁶ that are more likely to reflect early family influences. This is consistent with evidence that genetic nurture is more salient for problematic alcohol use¹⁴ than for normative consumption¹⁵, highlighting its primary role in sustained, heavy or problematic use.

Another novel contribution of our study is the longitudinal analysis demonstrating that genetic nurture effects on smoking quantity persist across adulthood, while the effects of direct genetic transmission attenuate with age. The persistence of genetic nurture across the adult lifespan aligns with developmental models³⁷ emphasizing long-term parental influences on offspring outcomes. The age-related attenuation of direct genetic effects might reflect the growing influence of non-familial environments (e.g., reduced smoking due to health concerns or lifestyle changes) that suppress genetic predispositions as individuals age. These divergent temporal patterns are consistent with previous research suggesting that genetic and environmental effects on substance use are dynamic and may shift over time^38,39.

Our findings advance understanding of how genetic nurture operates across the lifespan. A prior study found genetic nurture effects on smoking at age 24 but not at 29 using separate cross-sectional analyses¹⁵, whereas our within-individual repeated measures design provides greater power to detect subtle, persistent effects that might be obscured in single-time-point comparisons⁴⁰. This underscores the critical importance of longitudinal designs with repeated measures to elucidate how genetic transmission and genetic nurture unfold over time. Future studies replicating these temporal dynamics in independent samples are warranted.

Although maternal and paternal genetic nurture effects were equal in magnitude overall, our mediation analysis revealed that the specific pathway through observable smoking behavior was stronger for mothers. This stronger maternal mediation likely reflects mothers’ primary role in shaping the day-to-day home environment and child-rearing practices during critical developmental periods⁴¹. Mothers typically spend more time with children than fathers^42,43, which may amplify the impact of maternal smoking through increased opportunities for behavioral modeling^44,45 consistent with social learning theory⁴⁶. Additionally, mothers may exert unique influences through prenatal smoking exposure⁴⁷, which has been linked to neurobiological changes in offspring, including alterations in early brain development⁴⁸ and epigenetic patterns⁴⁹, thereby increasing the risk of later substance use^50,51. Fathers’ genetic nurture effects were also partly mediated by their smoking quantity, though these effects were weaker than maternal pathways. This suggests that paternal genetic nurture likely operates through additional pathways not included in our model, such as psychiatric disorders⁵² and father-child relationship⁵³. Future research should explore alternative mediation pathways to gain a more comprehensive understanding of the environmental mechanisms underlying genetic nurture.

Additional mediation analyses by offspring sex revealed that maternal mediation significantly exceeded paternal mediation among daughters but not among sons. This finding may reflect same-sex parental modeling processes⁵⁴, whereby daughters more strongly identify with and emulate maternal behaviors compared to paternal behaviors. However, the non-significant result in sons should be interpreted cautiously, as it likely reflects reduced statistical power in the smaller male sample rather than a true absence of effect, given the overlapping confidence intervals and comparable point estimates between sexes. Therefore, these exploratory findings warrant replication in larger samples.

The strengths of our study include the large population-based cohort, a novel approach integrating genotypic data from parent-offspring trios and pairs, and the use of diverse substance use measures. Moreover, we employed longitudinal modeling with repeated measures at three time points and structural equation modeling to investigate parent-of-origin effects in genetic nurture and mediation pathways. However, several limitations remain. First, our sample included only Dutch participants of European descent, potentially limiting generalizability. Second, PGSs explain only a small proportion of genetic liability to substance use, and the modest genetic nurture effect sizes observed reflect this limitation. Third, as with most observational studies, statistical power to detect small effects in interaction (e.g., PGS_NT×Age) and subgroup analyses (e.g., sex differences in mediation) was limited. Additionally, fewer fathers than mothers had both genotypic and phenotypic data available, which may have further constrained power for paternal analyses. Finally, retrospective self-reports on substance use, such as pack-years, may be subject to recall bias or underreporting⁵⁵. However, measures like cigarettes per day and pack-years have demonstrated strong reliability and validity in assessing lifetime smoking exposure⁵⁶.