This controlled, split-mouth, preclinical study was designed to evaluate outcome following surgical implantation of an allogeneic, freeze-dried demineralized bone matrix-calcium sulfate (DBM+CS) composite with a CS barrier in 3-wall intrabony periodontal defects in 4 dogs. Control conditions included surgical implantation of DBM or CS and gingival flap surgery (GFS) alone. Three-wall intrabony defects (4x4x4 mm) were surgically created at the mesial and distal aspect of the maxillary and mandibular first and third premolars, respectively. Maxillary and mandibular defects each received 1 of the 4 experimental conditions. Experimental conditions were rotated between defect sites in subsequent animals. Block sections of the defects were collected at sacrifice 8 weeks postsurgery and processed for histometric analysis. Histometric defect height (means +/- SD) for the DBM+CS, DBM, CS, and GFS groups amounted to 4.2 +/- 0.5, 4.3 +/- 0.7, 4.0 +/- 0.2, and 4.1 +/- 0.2 mm, respectively. Connective tissue adhesion (connective tissue contact to the root without apparent cementum formation) amounted to 0.4 +/- 0.3, 0.4 +/- 0.3, 0.5 +/- 0.2, and 1.6 +/- 0.5 mm for the DBM+CS, DBM, CS, and GFS groups, respectively; the DBM+CS, DBM, and CS groups being significantly different from the GFS group (P < 0.05). Cementum regeneration amounted to 3.0 +/- 0.3, 3.1 +/- 0.4, 2.5 +/- 0.4, and 1.6 +/- 0.3 mm for the DBM+CS, DBM, CS, and GFS groups, respectively; the DBM+CS, DBM, and CS groups being significantly different from the GFS group (P < 0.05). Alveolar bone regeneration amounted to 2.7 +/- 0.4, 2.7 +/- 0.3, 1.8 +/- 0.5, and 0.7 +/- 0.1 mm for the DBM+CS, DBM, CS, and GFS groups, respectively; the DBM+CS, DBM, and CS groups being different from the GFS group (P < 0.05), and the DBM+CS and DBM groups being different from the CS group (P < 0.05). None of the DBM-containing implants provided evidence of bone metabolic activity. In summary, surgical implantation of DBM and CS, alone or in combination, may result in significantly improved regeneration of alveolar bone and cementum in this preclinical model. Observed regeneration is likely unrelated to a biologic activity inherent in DBM. Rather it appears that space-providing properties of the implants supported observed regeneration.