Evaluation of impression accuracy for a four-implant mandibular model--a digital approach

Implant-supported prosthodontics requires precise impressions to achieve a passive fit. Since the early 1990s, in vitro studies comparing different implant impression techniques were performed, capturing the data mostly mechanically. The purpose of this study was to evaluate the accuracy of three different impression techniques digitally. Dental implants were inserted bilaterally in ten polymer lower-arch models at the positions of the first molars and canines. From each original model, three different impressions (A, transfer; B, pick-up; and C, splinted pick-up) were taken. Scan-bodies were mounted on the implants of the polymer and on the lab analogues of the stone models and digitized. The scan-body in position 36 (FDI) of the digitized original and master casts were each superimposed, and the deviations of the remaining three scan-bodies were measured three-dimensionally. The systematic error of digitizing the models was 13 μm for the polymer and 5 μm for the stone model. The mean discrepancies of the original model to the stone casts were 124 μm (±34) μm for the transfer technique, 116 (±46) μm for the pick-up technique, and 80 (±25) μm for the splinted pick-up technique. There were statistically significant discrepancies between the evaluated impression techniques (p ≤ 0.025; ANOVA test). The splinted pick-up impression showed the least deviation between original and stone model; transfer and pick-up techniques showed similar results. For better accuracy of implant-supported prosthodontics, the splinted pick-up technique should be used for impressions of four implants evenly spread in edentulous jaws.