A total of 1005 subgingival and extracrevicular samples from 201 male recruits, 18-25 yr old, were selectively cultivated for Actinobacillus actinomycetemcomitans. The organism was isolated in 55 subjects (27%); 9.5% of pooled subgingival plaque samples from first molars, 14% cheek mucosa, 20% dorsum of tongue and 20% saliva samples were culture-positive. In order to divide the study population into distinct clinical categories, cluster analysis was performed, based on previous caries experience, probing pocket depth categories, bleeding scores, visible plaque and calculus. Two clusters (n = 86 and n = 92, respectively) were identified with no or minimal periodontal disease (mean +/- standard deviation % of periodontal probing depth 1-2 mm 78.7 +/- 10.4% and 57.4 +/- 12.6%, respectively; virtually no periodontal probing/depth in excess of 4 mm) and a relatively low DMF-S (22 +/- 13). A third cluster (n = 22) had, in contrast, a high DMF-S (47.7 +/- 17.2) and a relatively high % of periodontal pockets of > or = 5 mm (5.9 +/- 5.2%). Prevalence of A. actinomycetemcomitans in this cluster was 41%, while the organism was found in 23% and 27% in the minimally diseased populations (p < 0.15). Whereas no heterogeneity of associations between subgingival and extracrevicular occurrence of the organism could be ascertained in different clusters, the organism was significantly more often identified in extracrevicular material, especially dorsum of tongue samples, compared with subgingival plaque (McNemar's chi2 = 12.45, p < 0.001). Multiple linear regression analysis revealed the number of A. actinomycetemcomitans positive samples as well as the % of sites bleeding on probing being positively associated with the % of sites with a probing pocket depth of > or = 5 mm (R2 = 0.345, p < 0.0001). The present large-scale investigation points to the wide distribution of this putative periodontopathogen in young individuals with minimal periodontal disease.