Cocaine doubles stroke risk, amphetamines nearly triple it in under-55s
The numbers are large enough to end a debate. In a meta-analysis encompassing more than 100 million people, Cambridge researchers found that cocaine use is associated with a 96% increase in stroke risk. Amphetamines: 122%. Cannabis: 37%. And when the analysis was restricted to people under 55, amphetamine use nearly tripled the risk.
Published March 9, 2026, in the International Journal of Stroke, the study goes beyond correlation. Using a genetic technique called Mendelian randomization, the researchers found evidence that these drugs are causing strokes, not merely co-occurring with them.
Why this study settles a long-running question
Previous research linking recreational drugs to stroke has been largely observational. Individual studies have found associations, but each one faced the same objection: maybe drug users have other risk factors -- diet, stress, income, co-occurring substance use -- that actually explain the increased stroke rates. Maybe the drugs are just a marker, not a cause.
The Cambridge team, from the Department of Clinical Neurosciences, addressed this in two ways. First, they pooled data from all available published studies into a meta-analysis, which combines results that individually might be inconclusive into a statistically more robust conclusion. Second, they used Mendelian randomization, which exploits naturally occurring genetic variants associated with substance use to test whether the relationship is likely causal.
The genetic analysis showed that cocaine use disorders were particularly associated with brain hemorrhage and cardioembolic stroke -- where a clot forms in the heart and travels to block blood flow in the brain. Cannabis use disorders were associated with stroke overall, particularly large artery stroke. Problematic alcohol use was linked to cardioembolic and large artery stroke.
"Our analysis suggests that it is these drugs themselves that increase the risk of stroke, not just other lifestyle factors among users," said Dr. Eric Harshfield, an Alzheimer's Society Research Fellow involved in the study.
The under-55 picture
Stroke is typically thought of as a disease of aging, but the age-restricted analysis tells a different story. Among people under 55, amphetamine use increased stroke risk by 174% -- nearly triple the baseline. Cocaine increased risk by 97%. Cannabis showed a more modest 14% increase in the younger group.
These numbers matter because stroke in younger adults is often treated as idiopathic -- cause unknown. The data suggest that drug use history should be part of the clinical workup when a young person presents with stroke symptoms.
The scale of potential exposure is substantial. In 2024, 8.8% of adults aged 16 to 59 in England and Wales -- roughly 2.9 million people -- reported using a recreational drug in the past year. In the United States, over half of those aged 12 and older have used drugs such as cocaine, cannabis, or opiates at least once.
How the drugs might cause strokes
The researchers outline several biological mechanisms that could explain the link. Cocaine and amphetamines cause sudden spikes in blood pressure and can trigger blood vessel spasm and constriction -- both established pathways to stroke. Heart rhythm disturbances, particularly from stimulants, can generate the cardiac blood clots that cause cardioembolic strokes.
Cannabis appears to work differently, with increased blood clotting as a potential mechanism. Amphetamines may additionally cause inflammation of blood vessel walls (vasculitis). All of these pathways are well-documented causes of both ischemic strokes (from blood clots) and hemorrhagic strokes (from bleeding).
What the study cannot tell us
The Mendelian randomization analysis could not be applied to amphetamines because there are currently no large genetic datasets with information on amphetamine usage. The genetic evidence for a causal link is therefore strongest for cocaine and cannabis, while the amphetamine association, though large in the meta-analysis, rests on observational data alone.
The study also does not quantify dose-response relationships -- how much drug use over what period produces what level of risk. The meta-analysis pooled studies with varying definitions of drug use, from single lifetime exposure to chronic use. A more granular understanding of how frequency and quantity affect stroke risk remains a gap.
Opioid use showed no statistically significant link to stroke risk, an interesting negative finding given the severity of the opioid crisis. The researchers note this does not mean opioids are safe; it means stroke specifically does not appear to be among their major cardiovascular harms.
"This is the most comprehensive analysis ever conducted on recreational drug use and stroke risk," said Dr. Megan Ritson of the Stroke Research Group at Cambridge. "These findings give us stronger evidence to guide future research and public health strategies."