Mandibular advancement device (MAD) therapy is the most commonly used second-line treatment for obstructive sleep apnea (OSA), but MAD may be ineffective in a subgroup of patients. The aim of this systematic review is to identify predictors of the efficacy of oral appliance (OA) therapy for OSA in adult patients. This review focuses on performing the quantitative analysis by subgroups based on the response criteria used in the literature and based on the type of device. PubMed, EMBASE, Scopus, Web of Science and Cochrane databases was conducted to identify potentially relevant studies published until Dec 2021. The search identified 1343 preliminary references. A total of 99 studies met the eligibility criteria and were included in the review, and 60 in the meta-analysis. The quality of studies was assessed using the Newcastle–Ottawa scale and the Cochrane scale. Based on meta-analysis, and considering a low to moderate evidence profile according to the GRADE scale, responders are younger patients, with smaller neck circumference, lower body mass index. Responders have shorter maxillary length, lower anterior and posterior facial height, a shorter distance from the hyoid bone to the third cervical vertebra, a shorter airway length, a smaller minimum airway cross-sectional area and a higher minimum oxygen saturation during sleep. Responders needed a lower optimal continuous positive airway pressure than non-responders. The type of device has not affected the results of the meta-analysis. The criterion “AHI <10 and reduction AHI >50%” is the one that provides the “weight” of significance for several variables. This criterion should be taken into consideration for future studies to predict OSA treatment by OA.