Authors: Delaporte A et al.
Anesthesiology Open 1(1): e0007, January 2026
Summary
Cannabis use is increasing among older adults, particularly in regions where medical or recreational use is legal. Despite this trend, the perioperative implications of cannabis exposure remain poorly understood. Because cannabinoids interact with neural pathways involved in arousal, attention, and memory, investigators have hypothesized that cannabis exposure could influence the risk of postoperative delirium. This study examined whether self-reported cannabis use before surgery is associated with postoperative delirium in older adults undergoing major noncardiac surgery.
This single-center retrospective cohort study evaluated patients aged 60 years or older with ASA Physical Status III–IV who underwent major noncardiac surgery under general anesthesia between January 2013 and April 2024 at an academic medical center. Patients with neurosurgical procedures, ASA I–II or V status, or documented cognitive impairment were excluded. Data were obtained from an institutional perioperative database.
Preoperative cannabis exposure was identified through the “drug exposure” field in the preoperative evaluation. Only patients with documented marijuana use were included in the exposed group, and individuals with other illicit substance use were excluded. Postoperative delirium within 72 hours served as the primary outcome and was defined by a positive Confusion Assessment Method for the ICU recorded in the electronic health record.
Among 5,171 patients included in the analysis, only 79 (1.5%) had documented cannabis use. Postoperative delirium occurred in 19.0% of cannabis users compared with 11.7% of non-users (P = 0.047). In the initial multivariable logistic regression model adjusting for numerous preoperative and intraoperative confounders, cannabis use was associated with higher odds of postoperative delirium (adjusted odds ratio 3.47; 95% CI 1.03–11.65).
However, the apparent association weakened when additional analytic approaches were applied. After adjusting for the year of surgery, the odds ratio decreased to 2.67 and was no longer statistically significant. A Firth penalized logistic regression—used to address bias related to small sample sizes—yielded a similar but nonsignificant estimate (adjusted odds ratio 2.77; 95% CI 0.71–10.21). In a propensity score–matched cohort, the effect size declined further (adjusted odds ratio 1.47; 95% CI 0.68–3.17). Across all analyses, confidence intervals were wide, reflecting substantial statistical uncertainty.
The study also revealed several baseline differences between cannabis users and non-users. Cannabis users were more frequently male, had higher rates of alcohol use and depression, and experienced longer periods of intraoperative hypotension. They also received preoperative midazolam more often and had higher rates of postoperative opioid use and acute kidney injury. These factors could potentially confound the relationship between cannabis exposure and delirium.
The authors concluded that although a signal suggesting increased delirium risk appeared in the primary analysis, the association was not robust across multiple sensitivity analyses. The small number of exposed patients and potential residual confounding likely contributed to the instability of the results. Consequently, the data do not support cannabis use as an independent risk factor for postoperative delirium in this population.
The investigators emphasized several limitations. Cannabis exposure was based on self-report, which likely underestimates true use. The dataset lacked information on cannabis dose, frequency, formulation, and timing relative to surgery. Important behavioral risk factors for delirium—such as tobacco use, chronic opioid therapy, and detailed psychiatric diagnoses—were not adequately captured. The small number of cannabis users also limited statistical power and increased the possibility of model overfitting.
Overall, the study suggests that while cannabis exposure may appear associated with delirium in crude analyses, the evidence does not demonstrate an independent causal relationship. Larger prospective studies with more detailed characterization of cannabis exposure are needed to clarify whether cannabis influences postoperative neurocognitive outcomes in older adults.
Key Points
• Cannabis use among older adults undergoing surgery remains relatively uncommon but is increasing as legalization expands.
• In this cohort of more than 5,000 patients, postoperative delirium occurred more frequently among cannabis users (19%) than non-users (11.7%).
• The initial statistical association between cannabis use and delirium disappeared after more robust analytic adjustments.
• Wide confidence intervals and a small exposed subgroup limit the reliability of the findings.
• Current evidence does not support cannabis use as an independent risk factor for postoperative delirium, but larger prospective studies are needed.
Thank you to Anesthesiology Open for allowing us to summarize and discuss this article for educational purposes.