Cannabis is the third most used psychoactive substance worldwide. The legal status of cannabis is changing in many Western countries, while we have very limited knowledge of the public health impact of cannabis-related harms. There is a need for a summary of the evidence of harms and risks attributed to cannabis use, in order to inform the definition of cannabis risky use. We have conducted a systematic review of systematic reviews, aiming to define cannabis-related harms. We included systematic reviews published until July 2018 from six different databases and following the PRISMA guidelines. To assess study quality we applied the AMSTAR 2 tool. A total of 44 systematic reviews, including 1,053 different studies, were eligible for inclusion. Harm was categorized in three dimensions: mental health, somatic harm and physical injury (including mortality). Evidence shows a clear association between cannabis use and psychosis, affective disorders, anxiety, sleep disorders, cognitive failures, respiratory adverse events, cancer, cardiovascular outcomes, and gastrointestinal disorders. Moreover, cannabis use is a risk factor for motor vehicle collision, suicidal behavior and partner and child violence. Cannabis use is a risk factor for several medical conditions and negative social consequences. There is still little data on the dose-dependency of these effects; evidence that is essential in order to define, from a public health perspective, what can be considered risky use of cannabis. This definition should be based on quantitative and qualitative criteria that informs and permits the evaluation of current approaches to a regulated cannabis market. (c) 2020 Elsevier B.V. and ECNP. All rights reserved.