The aim of this study is to investigate peri-implant and intraconnection microflora of healthy implants restored with cemented and screwed superstructures.Patients with two to three implants restored with cemented or screwed restorations and 5 years of follow-up were recruited. Samples were taken from peri-implant sulci, adjacent teeth, and the inner portion of connections. Prevalence of positive sites and bacterial loads for 10 microorganisms were obtained with quantitative real-time polymerase chain reaction. Implant connection permeability to the studied microorganisms was estimated using a standardized bacterial contamination index. Statistical analysis was performed using a generalized estimating equations model, Wald χ(2) test, and the least significant difference test.The final sample consisted of 18 patients (55 implants) in the cemented group and 22 patients (46 implants) in the screw-retained group. Regarding prevalence of positive sites, significant differences between groups were only found for Tannerella forsythia, which was 8.7 times more frequent at peri-implant sulci of cemented than screw-retained prostheses. Bacterial loads of Porphyromonas gingivalis, T. forsythia, Parvimonas micra, and total bacterial load were significantly higher at peri-implant sulci for the cemented group; at the inner portion of connections, values were significantly higher for P. micra and Fusobacterium nucleatum for the screw-retained group. Contamination index values demonstrated higher permeability to most microbes in the cemented group.Internal implant surfaces were microbiologically contaminated for both cemented and screw-retained superstructures. Differences were found between the two methods of prosthetic retention: the cemented group presented significantly higher bacterial loads in the peri-implant sulcus but significantly lower bacterial loads at the inner portion of the implant connection.