Purpose: The effect of different implant impression techniques on the accuracy of casts has been investigated mostly in vitro, and clinically relevant evidence is scarce. The purpose of this study was to investigate the effect of implant impression techniques-specifically, splinted versus nonsplinted-on the accuracy of fit of fixed implant prostheses in edentulous patients. Materials and Methods: This clinical study included 12 edentulous patients (13 edentulous arches). All patients had undergone computer-guided, prosthetically driven implant surgery. Splinted (with acrylic resin) and nonsplinted pickup implant impression techniques were used to generate two different casts. Intraoral verification jigs were made to fabricate a third index cast (prosthesis fabrication cast); these made up a control group. All patients were definitively rehabilitated with one-piece zirconia prostheses. The accuracy of fit of each prosthesis was evaluated indirectly by examining them clinically and radiographically while they were fit on the generated casts. Results: Of the 13 splinted casts, 12 presented with accurate clinical fit when the zirconia prosthesis was seated on its respective cast. Only 6 of the 13 nonsplinted casts showed accurate clinical fit. The zirconia prostheses fit accurately on all respective casts of the control group (prosthesis fabrication cast) as well as intraorally. The differences between the test groups and between the nonsplinted and control groups were statistically significant. No statistically significant differences were found between the splinted and control groups. Conclusion: There is clinical evidence that the splinted impression technique generates more accurate implant impressions and master casts than the nonsplinted technique for complete-arch, one-piece fixed prostheses. Keywords: accuracy of fit, dental implants, implant impressions, nonsplinted techniques, splinted techniques