Cannabis Impacts Tumor Response Rate to Nivolumab in Patients with Advanced Malignancies

There has been a significant increase in the use of immunotherapy and cannabis recently, two modalities that have immunomodulatory effects and may have possible interaction. We evaluated the influence of cannabis use during immunotherapy treatment on response rate (RR), progression-free survival (PFS), and overall survival (OS). Subjects, materials, and

In this retrospective analysis, the use of cannabis during immunotherapy treatment decreased RR, without affecting PFS or OS and without relation to cannabis composition. Considering the limitations of the study, further prospective clinical study is needed to investigate possible interaction. Implications for practice: Although the data are retrospective and a relation to cannabis composition was not detected, this information can be critical for cannabis users and indicates that caution is required when starting immunotherapy.

In this retrospective, observational study, data were collected from the files of patients treated with nivolumab in the years 2015-2016 at our hospital, and cannabis from six cannabis-supplying companies. Included were 140 patients (89 nivolumab alone, 51 nivolumab plus cannabis) with advanced melanoma, non-small cell lung cancer, and renal clear cell carcinoma. The groups were homogenous regarding demographic and disease characteristics. A comparison between the two arms was made.

In a multivariate model, cannabis was the only significant factor that reduced RR to immunotherapy (37.5% RR in nivolumab alone compared with 15.9% in the nivolumab-cannabis group (p = .016, odds ratio = 3.13, 95% confidence interval 1.24-8.1). Cannabis use was not a significant factor for PFS or OS. Factors affecting PFS and OS were smoking (adjusted hazard ratio [HR] = 2.41 and 2.41, respectively (and brain metastases (adjusted HR = 2.04 and 2.83, respectively). Low performance status (adjusted HR = 2.83) affected OS alone. Tetrahydrocannabinol and cannabidiol percentages did not affect RR in any group (p = .393 and .116, respectively).