BACKGROUND AND AIMS: Endoscopic ultrasound-guided biliary drainage (EUS-BD) has emerged as an alternative in cases of endoscopic retrograde cholangiopancreatography (ERCP) failure. Two types of EUS-BD methods for achieving biliary drainage when ERCP fails are choledochoduodenostomy (CDS) or hepaticogastrostomy (HGS). However, there is no consensus if one approach is better than the other. Therefore, we conducted a systematic review and meta-analysis to evaluate these 2 main EUS-BD methods. METHODS: We searched MEDLINE, Embase, Scopus, Cochrane database, LILACS from inception through April 8, 2017, using the following search terms in various combinations: biliary drainage, biliary stent, transluminal biliary drainage, choledochoduodenostomy, hepaticogastrostomy, endoscopic ultrasound-guided biliary drainage. We selected studies comparing CDS and HGS in patients with malignant biliary obstruction with ERCP failure. Pooled odds ratio (OR) were calculated for technical success, clinical success, and adverse events and difference of means calculated for duration of procedure and survival after procedure. RESULTS: A total of 10 studies with 434 patients were included in the meta-analysis: 208 underwent biliary drainage via HGS and the remaining 226 via CDS. The technical success for CDS and HGS was 94.1% and 93.7%, respectively, pooled OR=0.96 [95% confidence interval (CI)=0.39-2.33, I=0%]. Clinical success was 88.5% in CDS and 84.5% in HGS, pooled OR=0.76 (95% CI=0.42-1.35, I=17%). There was no difference for adverse events OR=0.97 (95% CI=0.60-1.56), I=37%. CDS was about 2 minutes faster with a pooled difference in means of was -2.69 (95% CI=-4.44 to -0.95). CONCLUSION: EUS-CDS and EUS-HGS have equal efficacy and safety, and are both associated with a very high technical and clinical success. The choice of approach may be selected based on patient anatomy.