AbstractDuring a 2–year period pocket depth, probing attachment level, gingival index, and crevicular fluid aspartate aminotransferase (AST) levels were monitored in 25 previously treated periodontitis patients. Probing attachment level change was used retrospectively to identify sites where active periodontal destruction had occurred. The ability of crevicular fluid AST activities at 600, 800, 1000, and 1200 μIU levels to recognize active disease was investigated. Probing attachment level changes observed support the concept that the pattern of periodontal disease activity is episodic and infrequent. A loss of ≥2 mm was found at 11% of all studied sites, whereas a gain of ≥2 mm was noticed for 15% of sites. 2 subjects had 3 teeth that lost ≥2 mm of attachment, whereas 15 subjects demonstrated no teeth with disease activity. The remaining 8 subjects had 1 or 2 sites that lost ≥2 mm of attachment. Bayes's theorem and ROC curves were used to exemplify the sensitivity and the specificity of AST assessments. The AST 800 μIU demonstrated a sensitivity of 0.93 and specificity 0.68 and an odds ratio of 15.4 for attachment loss ≥2 mm. Under the conditional probability of either 50%, 25% or 10% active disease prevalence, AST 800 μIU has a predictability of 73%, 50%, and 24%