It has been shown that the mineral trioxide aggregate (MTA) used to seal lateral/furcal perforations stimulates the deposition of newly formed cementum. Nevertheless, when the site of the perforation is contaminated, the healing process might occur under less favorable conditions. This study evaluated the repair healing process of noncontaminated and contaminated lateral perforations filled with MTA and the effect of previously filling the contaminated perforations with a bactericidal agent. Thirty lateral root perforations were prepared in endodontically treated dog's teeth, thus forming 3 groups with 10 specimens each. In group 1 the perforations were immediately sealed with MTA. In group 2 the perforations were left open for 7 days and thereafter sealed with MTA. In group 3 the perforations were left open for 7 days, filled temporarily with a calcium hydroxide-based paste for 14 days, and then sealed with MTA. The animals were killed after 90 days, and the pieces were prepared for histomorphologic and histomicrobiologic evaluations. The statistical analysis showed that group 1 had significantly better repair than groups 2 (P < .05) and 3 (P < .05), which validates the superior results obtained when MTA was immediately used to seal root perforations. Groups 2 and 3 had statistically similar repair to each other (P > .05). There were a larger number of cases of complete or partial biologic seal in group 1 compared with the contaminated groups. It might be concluded that the lateral root perforations sealed with MTA after contamination presented worse repair than the noncontaminated, immediately sealed perforations. The temporary filling with a bactericidal agent (calcium hydroxide-based paste) did not improve the repair of perforations exposed to contamination, and the contaminated groups presented similar results to each other.