RESUMO
OBJECTIVES: We compared the semiquantitative measurement of acetoacetate using urinary dipsticks with the bedside quantitative fingerprick measurement of the principal ketone bodies 3-beta-hydroxybutyrate, for the diagnosis of ketoacidosis. METHODS: This is a one year retrospective study of patients who presented with hyperglycemia levels of 250 mg/l or greater in the Emergency Department setting. We compared the sensitivity, specificity, and predictive value of ketonuria and ketonemia for the diagnosis of ketoacidosis (urine or blood ketone bodies, blood bicarbonates <20 mmol/l, anion gap >16 meq/l) in a sample of patients for whom the levels of ketone bodies in the blood and urine as well as serum electrolytes were available. RESULTS: We studied 355 hyperglycemic patients. The median time between arrival and dipstick testing was 21 min, and was greater than 2 h in more than 10% of cases. Comparison between ketonuria and ketonemia was performed in 173 patients (6% with diabetic ketoacidosis). Ketonuria equal to or less than one cross or a 3-beta-hydroxybutyrate value lower than 3 mmol/l enabled ketoacidosis to be excluded (negative predictive value 100%). At two-cross cutoff points for ketonuria and at the 3 mmol/l cutoff point for ketonemia, the two tests had the same sensitivity (100%), but the specificity of 3-beta-hydroxybutyrate (94%) was significantly higher (P<0.0001) than that of ketonuria (77%). The best positive predictive value for ketonemia was obtained at the 5 mmol/l cutoff point (100%) and for ketonuria at the three-cross cutoff point (26%). At the three-cross cutoff point for ketonuria and at the 5 mmol/l cutoff point for ketonemia, the two tests had the same negative likelihood ratio (0.1), but the positive likelihood ratio of 3-beta-hydroxybutyrate (infinite) was higher than that of ketonuria. CONCLUSION: The measurement of 3-beta-hydroxybutyrate in capillary blood is faster and more effective than the use of dipsticks in the urine to detect ketoacidosis in the Emergency Department setting.
