Although the association between attention-deficit/hyperactivity disorder (ADHD) and drug use disorder (DUD) is well documented, it is unclear whether it is causal or results from familial confounding.

In this study we included all 551 164 individuals born in Sweden between 1991 and 1995 and used linked data from multiple nationwide registries to identify those with ADHD prior to age 15 years (1.71%). We used Cox proportional hazards models to investigate the future risk for DUD as a function of an ADHD registration and then compared the results from the entire population with the results from a co-relative design. Using the Swedish Multi-Generation Register, we identified all full-sibling, half-sibling and first-cousin pairs discordant for ADHD.

In the population sample, ADHD had a substantially increased risk for future DUD with a hazard ratio (HR) of 3.34 after accounting for gender and parental education. Examining discordant cousin pairs, discordant half-siblings and discordant siblings, those with ADHD had HRs for DUD of 3.09, 2.10 and 2.38 respectively. Controlling for the number of ADHD registrations, ADHD patients with and without stimulant treatment were similarly associated with later DUD risk.

ADHD diagnosed before 15 years of age was strongly related to future risk for DUD. The magnitude of this association was modestly reduced in relative pairs discordant for ADHD, suggesting that the ADHD–DUD association is partly causal and partly a result of familial confounding. We found no evidence to suggest that this association resulted from stimulant treatment.