RESUMEN
Aim: To evaluate the marginal fit of protocol bars milled from digital models obtained by conventional molding followed by bench scanning or digital molding with an intraoral scanner. Methods: Four morse-cone implants and the mini-pillars were installed in a 3D printed mandible model (master model). Digital models of the master model were obtained by (n=10): (Group A - Conventional) conventional (analog) molding of the master model followed by bench scanning or (Group B - Digital) molding of the master model with an intraoral scanner. All-on-four protocol bars were designed and milled from the digital models for both groups and screwed into the master model. Scanning electron microscopy (SEM) images from the distal, central, and mesial regions of each implant were obtained and the implant-protocol bar marginal fit was measured in an image software (Image J). The mean misfit of each region was analyzed by two-factor ANOVA, Tukey test, and Student's t-test (0,05 = 0.05). Results: The digital approach (B) showed higher misadaptation than the conventional approach (A, p < 0.05), regardless of the region evaluated. In group A, the central region showed higher maladjustment than the mesial region (p<0.05), however, there were no differences among regions of group B (p>0.05). Conclusion: The conventional method of acquiring digital models using the bench scanner produced bars for the All-On-Four protocol with better marginal fit than the digital models obtained with an intraoral scanner