Abstract
BACKGROUND: We evaluated the harm reduction-based cannabis substitution hypothesis, that increased cannabis use may lead to decreased harmful alcohol or tobacco use, among people with HIV (PWH). SETTING/METHODS: Data are from the Centers for AIDS Research Network of Integrated Clinical Systems (CNICS) cohort between 2009 and 2023. PWH completed longitudinal assessments of cannabis, alcohol, and tobacco use. Two trajectories of change in cannabis use were evaluated: increasing use (initiation and increased use frequency vs. non-increasing use) and decreasing use (abstinence and reduced use frequency vs. non-decreasing use). Associations between changes in frequency of cannabis use and subsequent frequency of alcohol and tobacco use were estimated using time-updated, multivariable linear mixed models. Joint longitudinal and survival models were used to examine associations with hazardous alcohol use. RESULTS: Among 12,143 PWH, 31.9 % reported cannabis use, 17.5 % hazardous alcohol use, and 35.6 % cigarette use at baseline. Compared to those who did not increase use, increasing cannabis use was associated with 0.67 additional drinking days/month (95 %CI:0.42-0.92), 0.25 additional heavy episodic drinking days/month (95 %CI:0.04-0.46), and a 2.2 times higher odds of hazardous alcohol consumption (95 %CI:1.67-2.44). Increasing cannabis use was also associated with 0.60 additional cigarettes/day (95 %CI:0.27-0.93). Initiation of cannabis resulted in similar estimates. Conversely, decreasing cannabis use, with and without abstinence, was associated with less alcohol and cigarette consumption. CONCLUSIONS: Increasing cannabis use was associated with more drinking, more hazardous drinking, and higher cigarette intake. Contrary to the cannabis substitution hypothesis, increased cannabis use did not result in a reduction of alcohol or tobacco use among PWH.