Porphyromonas gingivalis, Treponema denticola and Bacteroides forsythus have been implicated in periodontal disease and each possesses an enzyme capable of hydrolyzing the synthetic trypsin substrate, BANA. We have used a chairside test for BANA hydrolysis to diagnose an anaerobic periodontal infection in patients with advanced forms of clinical disease using a 15-min/55 degrees C incubation protocol. However, the BANA test performance is dependent upon the length and temperature of incubation. In the present study, we have evaluated a 5-min/35 degrees C, a 5-min/55 degrees C and a 15-min/55 degrees C incubation protocol to determine whether the performance of the BANA test could be optimized using plaque samples obtained from subjects seeking dental treatment. Logistic regression models were tested with age, smoking status, and gingivitis scores as covariates. The best fitting model obtained with the 5-min/35 degrees C protocol had a sensitivity of 71%, a specificity of 68%, a false-positive proportion of 9%, a false-negative proportion of 65%, and an overall accuracy of 80%. When maximum likelihood estimates were obtained in this model, plaques from individuals who reported that they currently smoked were 9.57x, and those who quit smoking were 4.73x more likely to have a positive BANA score than someone who never smoked. Plaques were 4.55x more likely to be BANA-positive if they were removed from sites with gingivitis. These findings indicate that the performance of the BANA test is best using the 5-min/35 degrees C incubation protocol.