Several meta-analyses of observational studies have addressed the association between risk and protective factors and cannabis/cocaine/opioid use disorders, but results are conflicting. No umbrella review has ever graded the credibility of this evidence (not significant/weak/suggestive/highly suggestive/convincing). We searched Pubmed-MEDLINE/PsycInfo, last search September 21, 2020. We assessed the quality of meta-analyses with the AMSTAR-2 tool. Out of 3,072 initial references, five were included, providing 19 associations between 12 putative risk/protective factors and cannabis/cocaine/opioid use disorders (cases: 4539; N = 1,118,872,721). While 84 % of the associations were statistically significant, none was convincing. One risk factor (smoking) had highly suggestive evidence for association with nonmedical use of prescription opioid medicines (OR = 3.07, 95 %CI:2.27 to 4.14). Convincing evidence emerged in sensitivity analyses on antisocial behavior and cannabis use disoder (OR 3.34, 95 %CI 2.53–4.41). Remaining associations had weak evidence. The quality of meta-analyses was rated as moderate in two (40 %), low in one (20 %), and critically low in two (40 %). Future research is needed to better profile risk/protective factors for cannabis/cocaine/opioid use disorders disorders informing preventive approaches.