LAUSR

PURPOSE Different implant-abutment connections are available and it has been claimed they could have an effect on marginal bone loss. The aim of this review is to establish if implant connection configuration influences peri-implant bone loss (PBL) after functional loading. METHODS A specific question was formulated according to the Population, Intervention, Control, and Outcome (PICO): Does the type of implant-abutment connection (external, internal, or conical) have an influence on peri-implant bone loss? A PubMed/MEDLINE electronic search was conducted to identify English language publications published in international journals during the last decade (from 2006 to 2016). The search was conducted by using the Medical Subject Headings (MeSH) keywords "dental implants OR dental abutment AND external connection OR internal connection OR conical connection OR Morse Taper." Selected studies were randomized clinical trials and prospective studies; in vitro studies, case reports and retrospective studies were excluded. Titles and abstracts and, in the second phase, full texts, were evaluated autonomously and in duplicate by two reviewers. RESULTS A total of 1649 articles were found, but only 14 studies met the pre-established inclusion criteria and were considered suitable for meta-analytic analysis. The network meta-analysis (NMA) suggested a significant difference between the external and the conical connections; this was less evident for the internal and conical ones. Platform-switching (PS) seemed to positively affect bone levels, non-regarding the implant-connection it was applied to. CONCLUSIONS Within the limitations of this systematic review, it can be concluded that crestal bone levels are better maintained in the short-medium term when internal kinds of interface are adopted. In particular, conical connections seem to be more advantageous, showing lower peri-implant bone loss, but further studies are necessary to investigate the efficacy of implant-abutment connection on stability of crestal bone levels.