An extensive body of literature has investigated the efficacy of antimicrobial regimens used to treat pelvic inflammatory disease (PID), leaving many clinicians confused about how to choose among them. This study provides a formal appraisal of these reports. Thirty-four treatment studies published between 1966 and 1992 were identified, using Medline and bibliographies, and evaluated qualitatively and quantitatively in a metaanalysis. Twenty-one studies met the criteria for inclusion in this evaluation: appropriate system for making the diagnosis of PID, standardized assessment of clinical outcome, and entry and follow-up evaluation for lower genital tract infection with Neisseria gonorrhoeae and Chlamydia trachomatis. This metaanalysis identifies a considerable range of quality in study methods and research design and underscores the limitations inherent in comparing such data. Despite this, a number of antimicrobial regimens appear to have very good short-term clinical and microbiologic efficacy. Pooled clinical cure rates range from 75% to 94% and pooled microbiologic cure rates range from 71% to 100%. A cost comparison is provided, and future research priorities are suggested.