BackgroundFetal alcohol spectrum disorder (FASD) is a significant public health concern, yet there is no internationally agreed set of diagnostic criteria or summary of underlying evidence to inform diagnostic decision-making. This systematic review assesses associations of prenatal alcohol exposure (PAE) and outcomes of diagnostic assessments, providing an evidence base for the improvement of FASD diagnostic criteria.MethodsSix databases were searched (inception–February 2023). Case-controls or cohort studies examining associations between participants with/without PAE or a FASD diagnosis and the domains of physical size, dysmorphology, functional neurodevelopment and/or brain structure/neurology were included. Excluded studies were non-empirical, sample size < 10, PAE determined via biological markers only, or no suitable comparison group. Summary data were extracted and associations between outcomes and standardised levels of PAE or FASD diagnosis determined using random-effects meta-analyses. Certainty of the evidence was assessed using GRADE.ResultsOf the 306 included studies, 106 reported physical size, 43 dysmorphology, 195 functional neurodevelopment and 110 structural/neurological outcomes, with 292 different outcomes examined. There was a dose–response relationship between PAE and head circumference, as well as measures of physical size, particularly at birth. There was also an association between higher PAE levels and characteristic sentinel facial dysmorphology, as well as many of the current functional neurodevelopmental outcomes considered during diagnosis. However, data were often lacking across the full range of exposures. There was a lack of evidence from studies examining PAE to support inclusion of non-sentinel dysmorphic features, social cognition, speech-sound impairments, neurological conditions, seizures, sensory processing or structural brain abnormalities (via clinical MRI) in diagnostic criteria. GRADE ratings ranged from very low to moderate certainty of evidence.ConclusionsThis comprehensive review provides guidance on which components are most useful to consider in the diagnostic criteria for FASD. It also highlights numerous gaps in the available evidence. Future well-designed pregnancy cohort studies should specifically focus on dose–response relationships between PAE and dysmorphology, neurodevelopment and brain structure/neurological outcomes.Systematic review registrationPROSPERO: CRD42021230522.