Introduction Anastomotic creation of autogenous arteriovenous fistulas can be performed in different ways, side-to-end or side-to-side. However, there is a paucity of evidence to recommend them. The aim of this study is to compare both anastomosis types in elbow arteriovenous fistulas. Material and methods A prospective observational national multicenter study (ISRCTN62033470) was designed, including patients receiving a native arteriovenous fistula in the elbow using side-to-end or side-to-side anastomosis, between September, 2016 and September, 2017, with six-month postoperative follow-up period. Patient characteristics, surgical details, and follow-up data (primary, assisted primary and secondary patency, maturation, functionality, complications) were recorded and compared between both anastomosis groups using Kaplan–Meier curves estimations, at one and six-month follow-up, and finally a multivariate analysis with Cox regression was performed. Results Three centers participated in the study, including 133 cases (96 side-to-end, 37 side-to-side). The cephalic vein was more often used for side-to-end (58.3%) and basilic for side-to-side (78.4%; P < 0.001). Side-to-end anastomoses were faster to create (65.1–75.1 min; P = 0.009). During follow-up, 23 cases were lost (transplanted, dead, ligated, or lost), with no differences at one month. At six months, primary patency was better for the side-to-end group (78.5 − 55.9%; P = 0.038), but it was not confirmed as an independent predictor in the multivariate analysis. Furthermore, no significant differences in assisted primary or secondary patency, maturation or functionality were seen. Patients with side-to-side anastomosis more often required vein superficialization (2.1–16.2%; P = 0.002) and presented more frequent puncture hematomas (4.9–30.0%; P = 0.015). Conclusions Anastomosis type was not significantly related to different outcomes in the follow-up. Even though side-to-end anastomosis showed better primary patency at six months with lower need of vein superficialization and fewer puncture hematomas during follow-up, it was not confirmed as an independent predictor in the multivariate analysis, and similar assisted primary and secondary patency, maturation, and functionality rates have been seen after arteriovenous fistula creation.