Background and aims: Several studies have indicated that there is an association between maternal prenatal substance use and offspring externalising disorders. However, it is uncertain whether this relationship is causal. Therefore, we updated a previously conducted systematic review to determine if the literature supports 1) a causal role of maternal prenatal substance use on offspring externalising disorders and 2) whether these associations differ across externalising disorders. Methods: We searched Web of Science, Embase, PsycINFO and Medline databases. We included studies that examined smoking, alcohol or caffeine use during pregnancy as an exposure, and diagnosis of attention-deficit hyperactivity disorder (ADHD), conduct disorder (CD) and oppositional-defiant disorder (ODD) in offspring as an outcome. Studies on non-English language, fetal alcohol syndrome and comorbid autism spectrum disorders were excluded. Risk of bias assessment was conducted using Newcastle Ottawa Scale (NOS) and where possible meta-analysis was conducted for studies classed as low risk of bias. Results: We included 63 studies. All studies were narratively synthesised, and 7 studies were meta-analysed on smoking and ADHD. The majority of studies (46 studies) investigated the association between smoking and ADHD. Studies which accounted for genetic effects indicate that the association between smoking and ADHD is unlikely to be causal. Studies on alcohol exposure in all the outcomes reported inconsistent findings and no strong conclusions on causality can be made. Studies on caffeine exposure were mostly limited to ADHD and these studies do not support a causal effect. Conclusions: There is no causal relationship between maternal smoking during pregnancy and attention-deficit hyperactivity (ADHD) in offspring. However, given that the majority of identified studies investigated the association between ADHD and smoking exposure, findings with alcohol and caffeine exposures and conduct disorder (CD) and oppositional-defiant disorder (ODD) need more research, especially using more genetically sensitive designs.