Screening for ketonemia in patients with diabetes

To determine the sensitivity and specificity of the urine ketone dip test as a screening test for ketonemia in hyperglycemic patients and to compare the performance of the urine ketone dip test with the anion gap and serum bicarbonate level.

This was a prospective study conducted in an urban, university-affiliated public hospital emergency department. Inclusion criteria consisted of (1) patients with known diabetes and hyperglycemia (glucose level>200 mg/dL) and any complaint of illness, or (2) patients with hyperglycemia and symptoms of undiagnosed diabetes mellitus. Urine ketone dip test, serum ketone, and electrolyte levels were determined on all subjects. Sensitivity, specificity, and predictive values along with 95% confidence intervals (CIs) were calculated.

The study group comprised 697 patients, including 98 patients with diabetic ketoacidosis (DKA) and 88 with diabetic ketosis (DK). The sensitivity, specificity, positive, and negative predictive values of the urine ketone dip test for the detection of DKA were 99% (95% CI 97% to 100%), 69% (95% CI 66% to 73%), 35% (95% CI 29% to 41%), and 100% (95% CI 99% to 100%), respectively. For DKA and DK, the sensitivity, specificity, positive, and negative predictive values of the urine ketone dip test were 95% (95% CI 90% to 97%), 80% (95% CI 76% to 83%), 63% (95% CI 57% to 69%) and 98% (95% CI 96% to 99%). The anion gap and serum bicarbonate level were less sensitive but more specific than the urine ketone dip test for the detection of DKA and DK.

The urine ketone dip test has high sensitivity for detecting DKA and a high negative predictive value for excluding DKA in hyperglycemic patients with diabetes with any symptoms of illness. The urine ketone dip test is a better screening test for DKA and DK than the anion gap or serum bicarbonate.