Purpose: To compare the subgingival microbiota around two differently designed implant systems that were in function for more than 12 years in a randomised split-mouth study design, and to compare the outcome with natural dentition. Materials and methods: A total of 18 partially edentulous patients received at least two TiOblast™ (Astra Tech) and two Brånemark (Nobel Biocare) implants following a split-mouth design. At the last follow-up visit, periodontal parameters (probing depth, bleeding on probing and plaque) were recorded and intraoral radiographs were taken to calculate bone loss. Subgingival plaque samples were collected for culture, qPCR and checkerboard DNA-DNA hybridisation analysis. These data were related to implant design and bone loss. This study setup allowed a comparison of 34 Astra Tech (Impl A) with 32 Brånemark (Impl B) implants. Results: During the 12-year follow up, five patients dropped out. One Brånemark implant was lost before abutment connection in a dropout patient. Mean bone loss between loading and year 12 was 0.7 mm (range: -0.8-5.8) (Impl A), and 0.4 mm (range: -1.1-4.1) (Impl B). No significant microbiological differences (qualitative and quantitative) could be observed between both implant types. Compared to teeth, subgingival plaque samples from implants did not reach the concentration of pathogens, even after 12 years of function. Conclusions: These data show that both implant systems (with differences in macro-design and surface characteristics), in patients with good oral hygiene and a stable periodontal condition, can maintain a successful treatment outcome without significant subgingival microbiological differences after 12 years of loading. The presence of periodontopathogens did not necessarily result in bone loss. Keywords: implant systems, microbiology, peri-implant, split-mouth design, subgingival