Background: Periodontal disease is an inflammatory reaction to the bacteria in dental plaque. The present study compared the prevalence of periodontal disease in patients using as a diagnostic either probing depth measurements, an inflammatory marker such as numbers of white blood cells in plaque samples, or microbiological markers such as the microscopic count and the benzoyl-DL-arginine naphthylamide (BANA) test.
Methods: Teeth with the most inflammation and/or deepest pockets in each quadrant were probed and subgingival plaque was sampled from 1,043 consecutive new patients enrolled in a private practice. Multivariate "diagnostic" models were developed based upon the probing depth (general linear models), percentage of white blood cell-positive and percentage of BANA-positive plaques (logistic regression models) to determine the prevalence of patients with periodontal disease.
Results: Plaque samples were removed from 3,694 sites. Fifty-two percent of sampled pockets were >4 mm; 49% of sites were inflamed, using the presence of white blood cells, and 28% were infected using the BANA test. Diagnostic models were highly significant at P<0.0001. The white blood cell model was the most parsimonious as demonstrated by the lowest Akaike information criteria statistic and had the highest receiver operator characteristic (ROC) curve relative to the probing depth and BANA models.
Conclusions: Periodontal disease can be diagnosed chairside by the presence of white blood cells in plaque samples, a finding that reflects the inflammatory nature of the disease process. This approach would reduce the misclassification of subjects as having periodontal disease (130 patients in the present study who had pockets) but minimal evidence of an inflammatory response.