BACKGROUND:Positive effects on the clinical outcome of moderately rough implant surfaces are described. Intercomparison of clinical data, however, is rarely found. PURPOSE:The aim of this study was to compare the clinical results of two macroscopically identical implants, the one with a turned, machined and the other with an etched surface. MATERIALS AND METHODS:In a retrospective cohort study, the included implants followed the criteria: standard surgical protocol, >12 months in situ; minimally rough self-threading implants with a turned, machined surface (Mk II Nobel Biocare AB, Göteborg, Sweden], n=210); etched implants of the same macrodesign (3i Implant Innovations Inc., Palm Beach Gardens, FL, USA], n=151), length > or = 10 mm. Clinical data and implant success were rated. Resonance frequency analysis (RFA) and Periotest (Siemens AG, Bensheim, Germany) were measured and related to the corresponding implant survival rate in the respective group. RESULTS:The total number of implants was 361, of which 264 (73%) were subject to clinical reexamination. RFA and Periotest could be recorded in 25% of the implants. Neither clinically relevant nor statistically significant differences between the surface designs were found in the RFA (64 +/- 8.6 vs 63 +/- 9.7), in Periotest (-2 +/- 3.3 vs -1 +/- 5.1), and in mean survival periods (49 months, 95% confidence interval CI]: 46-51 months, for the turned vs 46 months, 95% CI: 43-49 months, for the double-etched implant). After osteoplastic procedures, a significantly higher rate of implant losses in the turned, machined implant group was observed (17 vs 1) with a mean survival period of 43 (40-46) months for the turned and 46 (45-48) months for the double-etched implants. CONCLUSION:No difference between implants with two different minimally rough surfaces was found. A positive effect of surface roughness is observed in poor quality bone, but the pivotal proof of this effect is still lacking.