Does Cannabinoid Use Reduce Opioid Utilization Among Patients with Gastrointestinal Cancer? Evidence from Epic COSMOS.
[PURPOSE] Whether medical cannabis reduces opioid use during early chemotherapy remains unclear.
- 95% CI 1.30-1.61
APA
Worku EB, Woldesenbet SA, Pawlik TM (2026). Does Cannabinoid Use Reduce Opioid Utilization Among Patients with Gastrointestinal Cancer? Evidence from Epic COSMOS.. Cancers, 18(7). https://doi.org/10.3390/cancers18071110
MLA
Worku EB, et al.. "Does Cannabinoid Use Reduce Opioid Utilization Among Patients with Gastrointestinal Cancer? Evidence from Epic COSMOS.." Cancers, vol. 18, no. 7, 2026.
PMID
41976333
Abstract
[PURPOSE] Whether medical cannabis reduces opioid use during early chemotherapy remains unclear. We examined cannabinoid and opioid prescribing among patients with GI cancers initiating chemotherapy.
[METHODS] Patients with GI cancers initiating chemotherapy (2016-2025) were identified from the Epic COSMOS database. Cannabis exposure was defined as prescriptions for FDA-approved cannabinoids recorded in the EHR; non-prescription cannabis obtained through dispensaries were not captured. Patients with any cannabis or opioid use in the prior year or death within 90 days were excluded. Multivariable logistic regression evaluated factors associated with cannabinoid use.
[RESULTS] Among 144,981 patients, 2.4% received cannabinoids within 90 days of chemotherapy initiation. Cannabinoid recipients were more likely to receive opioids than nonrecipients (60.6% vs. 31.1%, < 0.01). In adjusted analyses, cannabinoid prescribing was more common among Black patients (aOR 1.45; 95%CI 1.30-1.61), women (aOR 1.11; 95%CI 1.03-1.19), and individuals with pancreatic cancer (aOR 3.26; 95%CI 2.97-3.58). Prescribing varied by region and declined over time, relative to 2017.
[CONCLUSIONS] Early cannabinoid prescribing was not associated with reduced opioid use. Use was more common among patients with pancreatic cancer and among Black patients, highlighting clinical and structural variation in supportive-care practices.
[METHODS] Patients with GI cancers initiating chemotherapy (2016-2025) were identified from the Epic COSMOS database. Cannabis exposure was defined as prescriptions for FDA-approved cannabinoids recorded in the EHR; non-prescription cannabis obtained through dispensaries were not captured. Patients with any cannabis or opioid use in the prior year or death within 90 days were excluded. Multivariable logistic regression evaluated factors associated with cannabinoid use.
[RESULTS] Among 144,981 patients, 2.4% received cannabinoids within 90 days of chemotherapy initiation. Cannabinoid recipients were more likely to receive opioids than nonrecipients (60.6% vs. 31.1%, < 0.01). In adjusted analyses, cannabinoid prescribing was more common among Black patients (aOR 1.45; 95%CI 1.30-1.61), women (aOR 1.11; 95%CI 1.03-1.19), and individuals with pancreatic cancer (aOR 3.26; 95%CI 2.97-3.58). Prescribing varied by region and declined over time, relative to 2017.
[CONCLUSIONS] Early cannabinoid prescribing was not associated with reduced opioid use. Use was more common among patients with pancreatic cancer and among Black patients, highlighting clinical and structural variation in supportive-care practices.