Stroke prevention has long focused on classical vascular risk factors — hypertension, atrial fibrillation, diabetes — but a growing body of evidence implicates recreational substance use as an underappreciated independent driver. A rigorous new analysis combining meta-analytic and causal inference methods now quantifies those risks with greater precision than prior work, with implications for clinicians, public health messaging, and younger adults who may perceive illicit drug use as low-risk for cardiovascular outcomes.
Drawing on 32 observational studies spanning more than 100 million participants, along with two-sample Mendelian randomization (MR) leveraging genome-wide association data, the analysis examined seven illicit substances across ischemic and hemorrhagic stroke subtypes. Cocaine use was associated with nearly double the stroke odds (OR = 1.96), while amphetamine use carried an even larger signal (OR = 2.22). Cannabis showed a more modest but still statistically significant elevation (OR = 1.37), though with notable heterogeneity and potential small-study bias flagging caution. Opioids showed no significant association in the meta-analytic arm. Critically, the MR component — designed to filter confounding by using genetic proxies for substance dependence — provided additional causal evidence for cannabis, lending biological plausibility beyond the observational layer.
Mendelian randomization is not infallible: pleiotropy (genetic variants influencing multiple traits) and weak-instrument bias can distort findings. The observational leg of this analysis is also subject to detection bias, since drug users may interact more frequently with healthcare systems where strokes get recorded. Nevertheless, the convergence of two methodologically distinct approaches — traditional meta-analysis and genetic causal inference — substantially strengthens the inference here. Amphetamines are particularly notable given rising prescription stimulant use and methamphetamine prevalence; the 2.22-fold signal warrants prospective investigation into mechanisms, likely involving vasospasm, hypertensive surges, and cardioembolism. This analysis is best characterized as confirmatory and consolidating rather than paradigm-shifting, but its scale and methodological rigor make it the most comprehensive evidence synthesis on this question to date.