The aim of the present trial was to study the effect of meticulous supragingival plaque control on (i) the subgingival microbiota, and (ii) the rate of progression of attachment loss in subjects with advanced periodontal disease. An intra-individual group of sites exposed to non-surgical periodontal therapy served as controls. 12 patients with advanced periodontal disease were subjected to a baseline examination (BL) including assessments of oral hygiene status, gingival condition (BoP), probing depth, clinical attachment level and subgingival microbiota from pooled samples from each quadrant. The assessments were repeated after 12, 24 and 36 months. Following BL, a split mouth study was initiated. The patients received oral hygiene instruction, supragingival scaling and case presentation. 2 quadrants in each patient were identified as "test" and the remaining 2 as "control" quadrants. Subgingival therapy was performed in all bleeding sites in the control quadrants. Oral hygiene instructions and plaque control exercises were repeated once every 2 weeks during the initial 3 months of the study. Thereafter the plaque control program was repeated once every 3 months for the duration of the 3 years. Sites demonstrating loss of clinical attachment > or =2 mm in the test quadrants were treated subgingivally. The results showed that in both test and control quadrants repeated oral hygiene instructions and supragingival plaque removal procedures resulted in low plaque scores throughout the study. The gingival bleeding scores and the frequency of periodontal pockets > or =4 mm was, however, significantly higher in the test quadrants than in the control quadrants. At the end of the 3 year study, the control quadrants showed significantly more reduced (> or =2 mm) pockets than the test quadrants, 265 versus 96. The number of sites in the test quadrants showing probing attachment loss > or =2 mm was more than 4x greater than in the control quadrants (59 versus 13). The microbiological findings indicate a more pronounced reduction only for P. gingivalis in the control quadrants. None of the other 4 marker bacteria consistently reflected or predicted the clinical parameters. The present study shows that only supragingival plaque control fails to prevent further periodontal tissue destruction in subjects with advanced periodontal disease.