Lucy Riglin, Beate Leppert, Christina Dardani, Ajay K Thapar, Frances Rice, Michael C O'Donovan, George Davey Smith, Evie Stergiakouli, Kate Tilling, Anita Thapar
5 min
ADHD and depression frequently co-occur, but it remains unclear whether this relationship is causal—meaning ADHD directly increases the risk of later depression—or if it is primarily driven by shared genetic factors or other unmeasured confounders. This study aimed to disentangle these possibilities by examining the link between childhood ADHD and later depression using both longitudinal data and genetic analysis.
The researchers employed two complementary methods. First, they used the Avon Longitudinal Study of Parents and Children (ALSPAC), a large birth cohort, to track the relationship between ADHD at age 7 and recurrent depression between ages 18 and 25. Second, they performed two-sample Mendelian Randomization (MR) to test whether genetic liability for ADHD causes depression, using data from large-scale Genome-Wide Association Studies (GWAS).
The longitudinal analysis revealed that children with ADHD were significantly more likely to experience recurrent depression in young adulthood. This association remained robust even after controlling for factors like sex, early social adversity, and maternal depression. Notably, the risk for depression was highest among individuals whose ADHD symptoms persisted into adulthood. The MR analysis provided further support for a causal effect of ADHD on major depressive disorder. However, when the researchers used a broader, less specific definition of depression, the results were less consistent, suggesting that the strength of the causal link may depend on how depression is defined and measured.
These findings suggest that ADHD is not just a childhood neurodevelopmental condition but a potential precursor to later mental health struggles. The results imply that effective, ongoing management of ADHD—particularly as patients transition into adulthood—might serve as a preventative measure against the development of depression. The study highlights the importance of clinical vigilance for depression in adults who have a history of ADHD.
BACKGROUND: Attention-deficit hyperactivity disorder (ADHD) is associated with later depression and there is considerable genetic overlap between them. This study investigated if ADHD and ADHD genetic liability are causally related to depression using two different methods. METHODS: First, a longitudinal population cohort design was used to assess the association between childhood ADHD (age 7 years) and recurrent depression in young-adulthood (age 18-25 years) in N = 8310 individuals in the Avon Longitudinal Study of Parents and Children (ALSPAC). Second, two-sample Mendelian randomization (MR) analyses examined relationships between genetic liability for ADHD and depression utilising published Genome-Wide Association Study (GWAS) data. RESULTS: Childhood ADHD was associated with an increased risk of recurrent depression in young-adulthood (OR 1.35, 95% CI 1.05-1.73). MR analyses suggested a causal effect of ADHD genetic liability on major depression (OR 1.21, 95% CI 1.12-1.31). MR findings using a broader definition of depression differed, showing a weak influence on depression (OR 1.07, 95% CI 1.02-1.13). CONCLUSIONS: Our findings suggest that ADHD increases the risk of depression later in life and are consistent with a causal effect of ADHD genetic liability on subsequent major depression. However, findings were different for more broadly defined depression.
Sam: That's the key methodological move. Mendelian randomization lets you use instruments specific to ADHD liability and ask what they predict about depression, independently of the shared variance. If the instruments were just proxying the genetic overlap, you'd expect the effect to wash out once you account for that. The fact that it doesn't — that the ADHD-specific genetic signal still predicts major depression — is what supports a directional interpretation rather than a pure shared-etiology story.
Alex: What are the main limitations?
Sam: Two that matter most. On the MR side, the standard pleiotropy concern: the instruments are assumed to affect depression only through ADHD, not through independent pathways. If those variants have pleiotropic effects — influencing general neurodevelopmental risk, for instance — the causal estimate is biased upward. The authors run sensitivity analyses, but you can't fully rule that out.
Alex: And the longitudinal arm?
Sam: Attrition, and it's not trivial. When they use listwise deletion — complete cases only — the association disappears. That's a meaningful finding in itself. It almost certainly reflects non-random dropout: the participants most likely to have ADHD and develop depression are also the most likely to leave the study. So the positive result in the longitudinal analysis depends on how you handle missing data, which should temper confidence in that arm.
Alex: So the MR result is doing more of the causal heavy lifting.
Sam: That's a fair characterization. The longitudinal data provides temporal precedence, but the causal inference is more credibly grounded in the Mendelian randomization. And even there, the finding is specific: ADHD liability predicting major, recurrent depression — not depression broadly construed. That specificity is actually reassuring from a validity standpoint, but it also limits how far you can generalize.
Alex: What's the practical implication for someone working in this space?
Sam: The main takeaway is that the ADHD-to-depression pathway looks like more than a shared-etiology artifact. There's evidence for a directional effect, which has real implications for intervention timing. If ADHD is genuinely upstream of depression risk — not just correlated with it — then effective ADHD treatment in childhood might have downstream protective effects on depression in adulthood. That's a testable hypothesis this paper motivates, even if it doesn't answer it directly.
Alex: And the phenotype sensitivity finding is worth sitting with. If your outcome definition changes your causal conclusion, that's a methodological signal for the field more broadly.
Sam: It's a reminder that in psychiatric genetics, the phenotype is never just an administrative detail. How you operationalize the outcome shapes what you find — and in this case, it shapes whether you find a causal signal at all. That's something any researcher using MR in this space needs to take seriously.
Alex: A well-designed triangulation study with an honest account of where each method's assumptions strain — that's worth the read. Thanks for listening to ResearchPod.