The purpose of this investigation was to evaluate the effect of non-surgical periodontal treatment and adjunctive systemic minocycline therapy on the level of neutral protease activity in whole saliva of adults with periodontitis. A test group of 21 adult patients with moderate to severe periodontitis was compared to a control group of 5 adults with healthy periodontium. Four test groups were examined: 1) scaling and root planing (SRP), probing depth = 4 to 5 mm; 2) SRP, PD > or = 6 mm; 3) SRP and adjunctive systemic minocycline therapy, PD = 4 to 5 mm; 4) SRP and adjunctive systemic minocycline therapy, PD > or = 6 mm. Clinical parameters and levels of neutral protease in whole saliva were assessed at baseline and on the sixth week after the non-surgical periodontal treatment. Neutral protease activity was measured by spectrofluorimetric method. Statistical tests of Mann-Whitney and Spearman Rank correlation coefficient were used in the evaluation of the mean values of measurements. The mean values of protease activity were significantly higher in the test groups than in the control group at baseline. Six weeks after non-surgical therapy, patients with 4 to 5 mm probing depth had approximate values of protease activity comparable to the control group. Hence it can be argued that these patients did not need minocycline HCL as an adjunctive therapy. However, non-surgical therapy had limited effects on both clinical parameters and enzyme activities for subjects with > or = 6 mm probing depth; on the other hand, gingival inflammation and enzyme activities were reduced significantly by the usage of minocycline as adjunctive therapy in these patients. According to our results, neutral protease activity in saliva is related to probing depth and gingival bleeding index, and not related to age and epithelial cell number. For these reasons, systemic minocycline therapy might be useful as an adjunct to non-surgical therapy in the presence of deep pockets, especially for reinfected cases. Further investigations are needed to confirm this suggestion.