A Hospital-Data Study Linked Documented Cannabis Use to Lower COVID-19 Mortality

A 2023 conference study reported that hospitalized COVID-19 patients with documented cannabis use experienced lower rates of several serious complications than patients without a recorded history of use. The findings, presented at the American College of Chest Physicians’ CHEST meeting in Honolulu, came from an observational analysis—not a clinical trial—and cannot establish that cannabis caused better outcomes.

In the CHEST conference abstract, researchers analyzed records for 322,214 patients in the National Inpatient Sample, a large U.S. hospital-discharge database. The analysis identified 2,603 patients—about 0.8 percent of the sample—with documented marijuana use.

Compared with non-users, the cannabis-use group was younger and more likely to have a record of tobacco use. Non-users, meanwhile, had higher rates of several conditions associated with severe illness, including obesity, hypertension, diabetes, and obstructive sleep apnea. These differences are important because they can influence COVID-19 outcomes independently of cannabis use.

Before adjustment, patients with documented cannabis use had lower rates of intubation, acute respiratory distress syndrome, acute respiratory failure, severe sepsis with multiorgan failure, in-hospital cardiac arrest, and death. Mortality was 2.9 percent among cannabis users compared with 13.5 percent among non-users in the unadjusted analysis.

The researchers then used one-to-one propensity matching to compare cannabis users with otherwise similar non-users based on age, race, sex, and 17 comorbidities, including chronic lung disease. After matching, cannabis use remained associated with lower odds of intubation, acute respiratory distress syndrome, acute respiratory failure, severe sepsis with multiorgan failure, and death. The adjusted odds ratio for mortality was 0.48, with a 95 percent confidence interval of 0.33 to 0.69.

However, the study has substantial limitations. The National Inpatient Sample is built primarily from hospital billing and discharge data. It does not provide the detailed information needed to determine how often patients used cannabis, how recently they used it, what products they consumed, or whether they smoked, vaped, or used edibles. A diagnosis code for cannabis use also may not capture all users consistently.

Because the study was retrospective and observational, unmeasured differences between the groups could explain some or all of the association. The cannabis-use group’s younger age and lower burden of several chronic illnesses may have contributed to its outcomes, even after statistical adjustment. The results therefore should not be interpreted as evidence that smoking cannabis prevents COVID-19 or improves its treatment.

Other research has not produced a consistent picture. A 2024 JAMA Network Open cohort study of 72,501 patients found that cannabis use was associated with higher risks of hospitalization and intensive-care admission after accounting for tobacco use, vaccination, comorbidities, and other factors. It found no significant association between cannabis use and all-cause mortality. Differences in study populations, definitions of cannabis use, and methods may help explain why the findings do not align.

Laboratory research has also been cited in discussions of cannabis and COVID-19, but it addresses a different question. In a 2022 study, Oregon State University researchers found that the cannabinoid acids CBGA and CBDA interfered with SARS-CoV-2 entry into cells in laboratory experiments. The work did not test marijuana consumption in people and did not show that smoking cannabis or taking a commercial cannabis product prevents infection.

There are also separate respiratory-health concerns associated with inhaling cannabis smoke. The Centers for Disease Control and Prevention says smoked cannabis can damage lung tissue and contains many of the same toxins and irritants found in tobacco smoke. During the pandemic, the FDA and Federal Trade Commission warned sellers against making unapproved claims that CBD or other products could prevent or treat COVID-19.

The CHEST analysis is best viewed as a signal for further research rather than a medical recommendation. More rigorous studies would be needed to determine whether any cannabinoid has a clinically useful role in COVID-19 care—and whether any potential benefit differs by compound, dose, route of administration, or timing of use.