Cannabis use and the risk of tuberculosis: a systematic review

Cannabis has been identified as a possible risk factor in some tuberculosis (TB) outbreaks. As the most widely used (largely) illegal substance in Western countries this may be an important public health concern. We aim to systematically review the evidence on the association between cannabis use and TB (latent infection and active disease) to inform ongoing and future TB prevention and control strategies.

We conducted a systematic review. We searched Ovid Medline, Embase and PsycInfo, together with the World Health Organization website and Google Scholar, for all years to January 2018. Reference lists and conference abstracts were hand-searched, a forward citation search was conducted on the Web of Science, and experts were contacted. Two authors independently screened studies for inclusion, extracted data and assessed risk of bias using an adapted version of ROBINS-I ("Risk of Bias in Non-randomised Studies - of Interventions"). Data were narratively synthesised.

Of 377 records identified, 11 studies were eligible. Study designs were heterogeneous. Six studies utilised a relevant comparator group. Four of these investigated the association between cannabis use and latent TB infection; all provided some evidence of an association, although only two of these had adjusted for confounders. The remaining two comparator studies investigated the association between cannabis use and active TB disease; neither found evidence of an association after adjusting for confounding. All six studies were at "Serious" risk of bias. The five studies which did not utilise a relevant comparator group were all indicative of TB outbreaks occurring among cannabis users, but the quality of the evidence was very weak.

Evidence for an association between cannabis use and TB acquisition is weak. The topic warrants further robust primary research including the collection of consistent and accurate exposure information, including cannabis use practices, dose and frequency, and adjustment for confounders.