Cannabis Use Disorder Linked to Elevated Risk of Benign Salivary Gland Tumors: New Insights from a Large Retrospective Cohort Study

Highlights

• Cannabis use disorder (CUD) is associated with over a fivefold increased risk of benign major salivary gland tumors compared to non-CUD individuals.
• The association is particularly pronounced for parotid gland tumors and persists beyond five years of follow-up.
• Study findings are robust to adjustment for age, race, tobacco use, nicotine use disorder, and prior radiation exposure.
• Further research is needed to clarify causality, mechanisms, and the impact of cannabis dose and route of administration.

Study Background and Disease Burden

Benign tumors of the major salivary glands, particularly the parotid gland, represent a clinically significant but relatively rare cause of head and neck neoplasia. While most are non-malignant, these tumors can cause facial swelling, pain, nerve involvement, and require surgical intervention with attendant risks. Epidemiologically, cigarette smoking has been established as a risk factor for benign salivary gland neoplasms, likely due to carcinogenic exposure of glandular tissue. With the growing prevalence of cannabis use and the increase in cannabis use disorder (CUD) in the US and globally, questions regarding the potential long-term health effects of cannabis—including oncologic risk—have become increasingly urgent.

Study Design

In a retrospective cohort analysis published in JAMA Otolaryngology–Head & Neck Surgery (Gallagher TJ et al., 2025), investigators queried electronic health records from 66 US-based healthcare organizations. The study population included nearly 184,000 adults with a documented history of CUD and an outpatient hospital visit between 2005 and 2025, and a comparator group of over 6 million adults without CUD. The primary outcomes were the incidence of benign tumors of major salivary glands (including parotid, submandibular, and sublingual glands) and specifically of the parotid gland. Key clinical covariates—age, race, ethnicity, tobacco use, nicotine use disorder, and prior radiation—were adjusted using propensity score matching to ensure comparability between groups.

Key Findings

The study reported a marked increase in the incidence of benign salivary gland tumors among individuals with CUD:

• For any benign major salivary gland tumor, incidence was 0.08% in the CUD group versus 0.02% in the non-CUD group.
• For parotid gland tumors, incidence was 0.05% versus 0.01%, respectively.

Relative risk (RR) estimates demonstrate the magnitude and persistence of this association:

• RR for any benign major salivary gland tumor at any time: 5.2 (CUD vs. non-CUD).
• RR for 0–5 years post-index: 4.3.
• RR for ≥5 years post-index: 5.0.
• RR for benign parotid gland tumor at any time: 5.2.
• RR for 0–5 years: 4.9; for ≥5 years: 3.9.

These findings were robust across all time points, indicating a persistent increased risk that does not diminish over time. Adjustments for confounders, including tobacco and nicotine use, strengthen the inference that the observed association is not merely a byproduct of co-occurring substance use. Notably, the study did not report effect sizes with confidence intervals in the summary, and absolute risk remains low given the rarity of these tumors.

Expert Commentary and Mechanistic Insights

The observed association between CUD and benign parotid tumors raises important clinical and biological questions. The parotid gland, being the largest salivary gland and anatomically exposed to inhaled substances, may be particularly susceptible to the carcinogenic or proliferative effects of cannabis smoke or its metabolites. Similar to the established risk profile for tobacco, chronic exposure to combustion byproducts could plausibly contribute to glandular epithelial changes or promote neoplastic transformation. However, the study design cannot establish causality, and the biological mechanisms remain speculative.

Key study limitations include reliance on diagnostic coding (potential for misclassification), lack of granular data on cannabis dose, duration, and mode of use (smoking, vaping, edibles), and absence of detailed histopathological data. There is also the possibility of residual confounding by unmeasured factors, such as alcohol use or environmental exposures. The study population’s generalizability is strengthened by its large and diverse multi-institutional sample, yet findings should be interpreted with caution until replicated in prospective cohorts.

Current clinical guidelines do not address cannabis use as a risk factor for salivary gland tumors, underscoring the novelty of these findings. The authors and other experts emphasize the need for future research with more detailed exposure assessment, prospective design, and molecular characterization of tumor subtypes.

Conclusion

This large-scale, retrospective US cohort study provides compelling epidemiological evidence that cannabis use disorder is associated with a significantly increased risk of benign major salivary gland tumors, particularly of the parotid gland, independent of tobacco and other known risk factors. While the absolute risk remains low, the persistent nature and magnitude of the association merit further investigation into causality, underlying mechanisms, and clinical implications. As cannabis use continues to rise, clinicians should remain alert to these emerging risks and counsel patients accordingly. More definitive prospective studies are essential to guide evidence-based recommendations and policy.

References

Gallagher TJ, Lin ME, Kim I, Kwon DI, Kokot NC. Cannabis Use and Benign Salivary Gland Neoplasms. JAMA Otolaryngol Head Neck Surg. 2025 Jul 17:e251955. doi: 10.1001/jamaoto.2025.1955.

Additional real-world references on cannabis and cancer risk:
– Gurney J, Shaw C, Stanley J, Signal V, Sarfati D. Cannabis exposure and risk of testicular cancer: a systematic review and meta-analysis. BMC Cancer. 2015;15:897.
– Johnson Newlson K, et al. Salivary gland tumors: epidemiology and aetiology. Otolaryngol Clin North Am. 2016;49(2):237-252.