This paper examines the effects of smoking on the treatment outcomes of two nonsurgical therapies: (1) scaling and root planing alone (SRP) or (2) controlled-release of subgingivally delivered doxycycline hyclate in a polylactic acid based polymer gel. Subjects from 2 9-month multicenter studies were classified as nonsmokers (never smoked: 100 subjects), former smokers (137 subjects), and current smokers (> or = 10 cigarettes/day: 121 subjects). Clinical parameters were analyzed for treated sites with baseline probing depths > or = 5 mm and for a subset of treated sites with baseline probing depths of > or = 7 mm. Clinical parameters (plaque levels, clinical attachment levels, pocket depths, and bleeding on probing) were analyzed at baseline, 4, 6, and 9 months. In the doxycycline treated group in general, there were neither marked significant differences in clinical attachment gain nor differences in probing depth reduction among the 3 smoking groups. On the other hand, in the scaling and root planing treated group in general, there were significant differences in clinical attachment gain and pocket depth reduction, with non-smokers responding better than former smokers and current smokers at 6 and 9 months. These differences in clinical response between scaling and root planing alone versus controlled-release of locally-delivered doxycycline hyclate among these 3 smoking groups are discussed in relation to treatment implications for smokers.