The study team, led by Anh P. Nguyen of the Institute for Health Research at Kaiser Permanente Colorado, assembled a matched cohort of 100,140 people aged 12 years or older. Of these, 16,693 had received a diagnosis of cannabis use disorder, a clinical condition marked by impaired control over cannabis use, continued use despite harmful consequences, and physiological dependence in some patients. Each person with the disorder was compared against individuals without the diagnosis drawn from the same health system, allowing the researchers to isolate the association between the condition and mortality while accounting for shared characteristics of the patient population.

Follow-up extended through June 2024, with a median observation period of 7.7 years. Death records came from state vital statistics, enabling the team to classify not only whether people died but why. Over the follow-up window, 1,197 individuals with cannabis use disorder died, representing 7.2 percent of that group. The raw numbers were stark: people with the disorder died at a crude rate of 899.5 per 100,000 person-years, compared with 374.6 per 100,000 person-years among matched individuals without the diagnosis, a difference of more than a factor of two before any statistical adjustment.

Because people with substance use disorders often carry other health and social burdens, the investigators used Cox proportional hazards regression to adjust for confounding factors and estimate adjusted hazard ratios. Even after this statistical correction, the signal remained strong. Cannabis use disorder was associated with a 52 percent higher risk of death from any cause, with an adjusted hazard ratio of 1.52 and a 95 percent confidence interval of 1.30 to 1.77. The confidence interval excludes the null value of 1.0, indicating that the association is unlikely to be explained by random variation in the data.

The researchers then split deaths into two broad categories: injury-related and illness-related. Injury deaths, which include overdoses, accidents, and self-harm, occurred at a crude rate of 295.3 per 100,000 person-years among people with cannabis use disorder versus 62.5 among matched comparators, nearly a fivefold difference in unadjusted rates. After adjustment, the disorder was associated with a 91 percent higher risk of injury-related death, with a hazard ratio of 1.91 and a confidence interval of 1.37 to 2.67. Illness-related mortality, covering deaths from disease, was also elevated, at 604.2 versus 312.1 per 100,000 person-years in crude terms, with an adjusted hazard ratio of 1.40.

Drilling down into specific causes of death revealed a differentiated pattern. The strongest association was with drug overdose: people with cannabis use disorder had nearly 2.6 times the risk of dying from an overdose, with a hazard ratio of 2.59 and a confidence interval of 1.30 to 5.19. This finding aligns with a growing body of evidence that cannabis use disorder frequently co-occurs with other substance use disorders, and that polysubstance use multiplies overdose danger. Deaths from accidents were also significantly elevated, with a hazard ratio of 1.96, consistent with concerns about impairment-related injuries in an era of expanding cannabis availability.

Cancer mortality emerged as another significant contributor, with an adjusted hazard ratio of 1.71 and a confidence interval of 1.20 to 2.46. The authors and related literature note that cannabis is most commonly consumed by smoking, and recent observational studies have begun to examine links between cannabis smoking and lung cancer as well as broader cancer risk in the context of legalized access. By contrast, the study found no statistically significant elevation in deaths from self-harm, cardiovascular disease, or respiratory disease. The hazard ratios for those causes were 1.18, 1.25, and 1.29 respectively, but all of their confidence intervals crossed 1.0, meaning the data could not distinguish those estimates from no increased risk.

The findings arrive at a moment of rapid change in the American cannabis landscape. Self-reported cannabis use in the United States has risen markedly over recent decades, and prior research has documented increases in cannabis use disorder following recreational legalization in several states. National surveys have also shown that the prevalence of marijuana use disorders grew substantially between the early 2000s and the early 2010s. Yet cannabis use disorder remains underrecognized in routine care, and evidence on mortality among affected individuals has historically been limited, often relying on treatment-seeking populations or emergency department cohorts rather than broad health-system samples.

The new study adds to a small but growing literature. A 2025 analysis published in JAMA Network Open linked cannabis use disorder-related emergency department visits and hospitalizations to elevated five-year mortality, and a 2025 systematic review and meta-analysis of cohort studies reported increased all-cause mortality associated with cannabis use. The Colorado cohort strengthens this evidence base by using matched comparison individuals from the same integrated health system, by capturing deaths through official vital records rather than electronic health data alone, and by distinguishing injury-related from illness-related causes of death within a single analytic framework.

The authors caution that the study is observational: it identifies associations, not proof of causation, and cannabis use disorder may act partly as a marker for a constellation of risks, including co-occurring substance use, mental health conditions, and social determinants of health that themselves predict premature death. Still, the researchers conclude that a diagnosis of cannabis use disorder may serve as a clinically meaningful signal of elevated all-cause, injury-related, and illness-related mortality risk, and they call for comprehensive prevention strategies to address the heightened risk in this population. As cannabis potency and availability continue to climb, the study suggests that identifying and treating cannabis use disorder in primary care settings could become an increasingly important piece of the mortality-prevention puzzle.

Subject of Research: Mortality risk associated with cannabis use disorder in a matched Colorado health system cohort

Article Title: Cannabis Use Disorder and Mortality Risk: A Matched Cohort Study from a Colorado Health System

Article References: Cannabis Use Disorder and Mortality Risk: A Matched Cohort Study from a Colorado Health System. (n.d.). https://doi.org/10.1007/s11606-026-10845-1

Image Credits: AI Generated

DOI: 10.1007/s11606-026-10845-1

Keywords: cannabis use disorder, mortality, matched cohort study, drug overdose, cancer, injury-related death, Cox proportional hazards, Colorado, public health, substance use, epidemiology, Journal of General Internal Medicine