OBJECTIVE To provide a context for the interpretation of lactic acidosis risk among patients using metformin, we measured rates of lactic acidosis in patients with type 2 diabetes before metformin was approved for use in the U.S.

RESEARCH DESIGN AND METHODS Using electronic databases of hospital discharge diagnoses and laboratory results maintained by a large, nonprofit health maintenance organization (HMO), we identified possible lactic acidosis events in three geographically and racially diverse populations with type 2 diabetes. We then reviewed hard-copy clinical records to confirm and describe each event and determine its likely cause(s).

RESULTS From <41,000 person-years of experience, we found four confirmed, three possible, and three borderline cases of lactic acidosis. In each case, we identified at least one severe medical condition that could have caused the acidosis. The annual confirmed event rate is similar to published rates of metformin-associated lactic acidosis.

CONCLUSIONS Lactic acidosis occurs regularly, although infrequently, among persons with type 2 diabetes, at rates similar to its occurrence among metformin users. The medical conditions with which both metformin-associated and naturally occurring lactic acidosis cooccur are also its potential causes. The observed association between metformin and lactic acidosis may be coincidental rather than causal. This possibility merits further study.