In this study the effectiveness of a quantitative test of beta-hydroxybutyrate (beta-HBA) against a commercial test for urine ketone bodies (UKB) in monitoring diabetic ketoacidosis (DKA) was evaluated to verify whether this ketone-testing method was able to reduce the monitoring costs and professional burden of nurses and physicians. Thirty-three children with severe (arterial pH < or = 7.2) or moderate (7.2 < pH < or = 7.3) DKA were studied. Sixteen patients were randomly monitored with blood beta-HBA (group 1) and 17 by UKB (group 2). Contrary to UKB, beta-HBA levels appeared correlated with: HbA1c values on admission (r = 0.99; p = 0.0001); latent period before diagnosis of diabetes (r = 0.95; p = 0.0001); changes in arterial pH (r = -0.82; p = 0.0001) and blood bicarbonate values (r = -0.63; p = 0.001) during the treatment for DKA. Required time to achieve the resolution of ketosis in group 1 patients was related to the values of beta-HBA on admission (r = 0.84; p < 0.001). Determination of beta-HBA showed that ketosis in group 1 patients cleared 4 to 9.5 hours earlier than in group 2. Due to this early normalization, the patients of group 1 left the Intensive Care Unit 6.5 +/- 1.5 hr earlier than those of group 2. This led to 22 hr saved for clinical assessment and 375 laboratory investigations for a total saving of 2940 euros including costs for laboratory tests (29.8%) and clinical assessment (70.2%). Quantitative determination of beta-HBA levels seems to offer useful information for monitoring DKA in newly-diagnosed diabetic children and for reducing time and costs in an Intensive Care Unit.