Objectives: The concept of platform switching has been introduced to implant therapy, however long-term data are sparse. The aim of this study was to biochemically investigate the inflammatory response mediated by MMP-8 to platform switching after 3 years of loading, in order to understand the long-term effect of implant/abutment mismatching on peri-implant health. Methods: A total of 70 implants had been inserted in the posterior maxilla in 26 patients and were randomly assigned to one of the four treatment regimens (implant diameter 3.8 [control group], 4.3 [Test group 1, T-1], 4.8 [Test group 2, T-2] and 5.5mm [Test group 3, T-3]). All implants were restored using a 3.8mm abutment. In the test groups, this restoration resulted in a mismatching of 0.25-0.85mm of implant-abutment diameters. Results: Thirty-six months after prosthetic rehabilitation, peri-implant sulcular fluid samples were taken from two aspects of all implants and from periodontally healthy adjacent teeth. Samples were processed in a conventional ELISA using monoclonal antibodies recognizing the active entity of MMP-8. In the test groups, MMP-8 mean values were 2.76 ng for T-1 (SD: 2.91), 3.30 ng for T-2 (SD: 1.94) and 3.18 ng for T-3 (SD: 2.46). For the control group, MMP-8 mean value was 3.6 ng (SD: 2.23), whereas 3.38 ng (SD: 2.2) was recorded at the adjacent teeth. There were no statistically significant differences in MMP-8 values between the groups (P = 0.113, Kruskal-Wallis). Conclusions: The presence of an implant/abutment mismatching specific for this prosthetic concept is compatible with long-term peri-implant health as demonstrated by analysis of a sensitive biomarker of the peri-implant inflammatory response.