Guidelines for clinical trials demonstrating equivalence or superiority for treatments for periodontitis are badly needed because of the great variety of drugs, agents, and devices now being developed. This paper focuses on three design issues. These are primary outcome variables and their measurement, disease-active vs disease-inactive sites and patients, and study duration. Determinants for selection of outcome variables include the biologic events to be observed, changes that are specific for periodontitis, and methods chosen to detect those changes. The primary outcome variables specific for periodontitis and appropriate for use in clinical trials are periodontal attachment level and alveolar bone status. Improved methods for measurement of both with excellent accuracy and reproducibility are now becoming available. Studies performed on untreated patients over the past decade demonstrate clearly that disease-active and disease-inactive pockets exist, at any given point in time most are inactive, disease progression is episodic and in most patients infrequent, and a rather small portion of the population--possibly around 5%--are unusually susceptible to rapid disease progression. These observations need to be taken into account in enrolling subjects into periodontitis clinical trials. Conducting a prestudy to identify actively diseased sites and susceptible subjects, or screening to enrich the proportion of active sites, is recommended. Determination of study duration is a very complex issue. It is related to the length of time required for maximal change and stabilization to occur in the biological events to be observed, the outcome variable(s) used to detect change, and the nature of the therapeutic interventions to be studied. No single duration is applicable to all periodontitis clinical trials. Large gaps in our knowledge about the design of periodontitis trials still exist, and additional research is needed.