We have performed a two-part evaluation of assays for beta-hydroxybutyrate (beta HB). In Part I, we evaluated the analytical acceptability of a dry film, self-contained analyzer for beta HB measurement (KetoSite; GDS Diagnostics, Elkhart, Ind). The assay response is linear and between-day imprecision is acceptable. Acetoacetate (AcAc) interferes severely with this assay. This interference could be minimized by dilution of serum before the assay. With this predilution protocol, KetoSite beta HB results compare favorably with an automated Sigma (St. Louis, Mo) beta HB assay. In Part II, we compared the clinical utility of beta HB measurements (Sigma) with the qualitative measurement of ketones using the nitroprusside test (Acetest; Bayer, Elkhart, Ind) for the diagnosis and management of diabetic ketoacidosis (DKA). Test results for both assays were abnormal for all patients with DKA on admission. However, frequent monitoring with either test during the management of uncomplicated DKA added little if any additional clinical information over routine measurements of serum glucose and total carbon dioxide.