BACKGROUND: This study assesses the microbiologic effects of a two-phase antimicrobial periodontal therapy and tested microbiologic, clinical, and biologic markers as prognostic indicators for clinical success. METHODS: Eighty patients with chronic or aggressive periodontitis received periodontal treatment supplemented with 375 mg amoxicillin plus 500 mg metronidazole, three times daily for 7 days. In group A, antibiotics were given during the first non-surgical phase (T1); in group B, antibiotics were given during the second surgical phase (T2). Six microorganisms, group assignment, demographic and clinical variables, peak values of 15 cytokines, and nine acute-phase proteins in serum were evaluated as potential predictors of at least one site with probing depth (PD) >4 mm and bleeding on probing (BOP) at 12 months post-therapy. RESULTS: T1 decreased the counts of Porphyromonas gingivalis, Tannerella forsythia, Prevotella intermedia (Pi), and Treponema denticola significantly more in group A than group B. Aggregatibacter actinomycetemcomitans and Parvimonas micra (Pm) showed a significant decrease only if the treatment was supplemented with antibiotics, i.e., T1 in group A, or T2 in group B. After T2, differences between groups were no longer significant. A multivariable model including four parameters revealed a predictive value of Pm (odds ratio [OR] = 4.38, P = 0.02) and Pi (OR = 3.44, P = 0.049) and yielded moderate accuracy for predicting the treatment outcome (area under the curve = 0.72). Host-derived factors and treatment sequence were not significantly associated with the outcome. CONCLUSIONS: Long-term microbiologic outcomes of periodontal therapy with adjunctive antibiotics either in T1 or T2 were similar. Detection of Pm before therapy was a predictor for persistence of sites with PD >4 mm and BOP at 12 months post-treatment.