Selective inhibition of recolonization of Porphyromonas gingivalis was investigated by topical application of monoclonal antibody (MAb). To select a MAb to P. gingivalis with the potential for recognizing most strains of P. gingivalis, we examined seven MAbs, one of which (MAb 61BG 1.3) recognized all 22 laboratory strains and serotypes of P. gingivalis tested as well as 105 human clinical isolates. A comparative study of the number of P. gingivalis bacteria identified by conventional culture and immunofluorescence with MAb 61BG 1.3 showed a very significant correlation between the two methods (Spearman r = 0.85, P < 0.001). Fourteen patients with periodontitis, who harbored P. gingivalis in their subgingival plaque, were treated by root planing and with metronidazole to suppress any detectable P. gingivalis. In this double-blind study, the patients were then divided randomly into two groups; one was treated with MAb to P. gingivalis, and the other was treated with saline. Each patient had four subgingival applications of 3 micrograms of MAb (or saline) per tooth at 1, 3, 7, and 10 days after P. gingivalis was suppressed. The number of P. gingivalis bacteria was then monitored, and significantly less recolonization of the sites with the most severe periodontitis was found in the MAb-treated patients than in the control patients (P < 0.01). This was evident at 6 and 9 months after the application of MAb, but by 12 months, P. gingivalis, was also found to recolonize these sites in two of the MAb-treated patients. The effect of MAb was specific to P. gingivalis, since the numbers of spirochetes were not significantly different between the two groups. However, no significant difference in any clinical periodontal indices between the immunized and control patients at 6 and 12 months was observed. This is the first demonstration that a putative periodontal pathogen can be selectively prevented from recolonization for up to 9 months in sites with the most severe periodontitis. This strategy could be used to establish directly in humans whether a microorganism is involved in the pathogenesis of periodontitis, by repeated application of the corresponding MAb at about 6-month intervals and by comparing the clinical indices between the MAb-treated and control patients.