Purpose: The aim of this study was to systematically review the studies on the association between diabetes mellitus (DM) and destructive periodontal disease. Methods: The methods applied include a literature search strategy, inclusion and exclusion criteria for selecting the studies, characteristics of the studies, quality assessment and meta-analysis. Data sources included PubMed, EMBASE, SciELO and LILACS. Selected papers were articles relating to human studies investigating whether or not diabetes is a risk factor for periodontitis and if it influences the response to periodontal therapy. Those papers that were published between January 1980 and June 2007 were retrieved. Results: Of the 2440 identified studies, 49 cross-sectional and eight longitudinal studies met the inclusion criteria. Twenty-seven of the 49 cross-sectional studies that are included in this review detected more periodontal disease in diabetic subjects compared with non-diabetic subjects. The greater risk of periodontal disease progression was associated with type 2 DM, and one study associated DM with response to periodontal therapy. Methodological flaws of most of the studies included inadequate control for confounders, insufficient statistical analysis and lack of information about sampling design. Random effect model showed a significant association with clinical attachment level (mean difference = 1.00 [CI 95% = 0.15 to 1.84]) and periodontal pocket depth (mean difference = 0.46 [CI 95% = 0.01 to 0.91]) between type 2 diabetics and non-diabetics. Conclusions: Type 2 DM can be considered a risk factor for periodontitis. More studies are needed to confirm the harmful effects of type 1 DM on periodontal disease. Parole chiave: diabetes mellitus, meta-analysis, periodontal disease, periodontitis, systematic review