ADHD Linked to Higher Mortality in Meta-Analysis
Key Takeaways
- Across age groups, people with ADHD had higher all-cause mortality than comparison groups.
- Suicide, unintentional or intentional injury, and natural-cause mortality were also higher in ADHD groups.
- Among people with ADHD, medication use was associated with lower all-cause mortality than nonuse, without establishing a treatment effect.
- Between-study heterogeneity was substantial, complicating interpretation of the pooled associations.
Researchers conducted the JAMA Network Open ADHD mortality meta-analysis, systematically reviewing peer-reviewed quantitative studies included through July 2025. The review included 19 studies and 15,920,970 participants across age groups; 12 studies provided outcomes usable in the meta-analyses, including nine in the all-cause mortality analysis. Investigators compared all-cause and cause-specific mortality in people identified as having ADHD with mortality in comparison groups. They also compared medication users with nonusers within ADHD populations. Identification was not limited to formal diagnoses, and medication exposure was not randomly assigned.
For all-cause mortality, the pooled comparison of people with ADHD against comparison groups yielded a hazard ratio (HR) of 1.71 (95% confidence interval [CI] 1.31–2.24; P<.001). This estimate describes a relative hazard across the included studies, rather than an individual's probability of death.
Cause-specific comparisons pointed in the same direction. Suicide mortality was higher in ADHD groups (HR 1.89), as was mortality from unintentional or intentional injury (HR 1.55). Natural-cause mortality was also higher (HR 1.27). These associations do not identify a single pathway accounting for the differences.
Within populations with ADHD, a three-study analysis associated medication use with lower all-cause mortality than nonuse (HR 0.75; 95% CI 0.64–0.87; P<.001); confounding may affect the estimate. Unlike the ADHD-versus-comparison-group estimates, this association reflects differences in observed medication exposure among people with ADHD.
Medication exposure was observational, so differences between users and nonusers could reflect factors other than medication. The report described substantial heterogeneity overall, but no significant heterogeneity in the natural-cause model and moderate heterogeneity with a nonsignificant Q statistic in the medication analysis; differences in study methods and potential residual confounding also complicate interpretation of the pooled associations. In particular, the medication finding cannot be read as a randomized treatment effect. These constraints also limit how closely a pooled estimate might describe a particular clinical population.
The researchers concluded that ADHD was associated with higher all-cause mortality and higher mortality from suicide, injury, and natural causes across the populations studied. Lower all-cause mortality associated with medication use remains an observational finding, not proof of a causal benefit.
Clinician Questions
Were ADHD mortality associations restricted to people with a formal ADHD diagnosis?
No. Corrigan and colleagues included studies identifying attention-deficit/hyperactivity disorder across age groups through clinical diagnosis, self-report, trait measures, or ADHD polygenic risk scores. The analyzed populations were therefore not restricted to formally diagnosed cases; the pooled findings do not provide separate estimates for each identification method.
Can ADHD mortality hazard ratios estimate an individual patient's absolute risk or medication benefit?
No. Both the ADHD-versus-comparison-group findings and the medication-use-versus-nonuse findings within ADHD are relative hazard associations.