Diabetic ketoacidosis and, to a lesser degree, diabetic ketosis are potentially dangerous complications of diabetes. Seriously ill patients are easy to detect clinically, but well-appearing diabetic patients who present with vague symptoms such as malaise, nausea and hyperglycemia with ketonemia are more difficult to identify. Detection of diabetic ketoacidosis or diabetic ketosis is accomplished on the basis of the urine ketone dip test, the anion gap or the serum bicarbonate level.

Schwab and associates compared the sensitivity, specificity and predictive values of these three screening tests in the detection of ketonemia, with or without acidosis, in diabetic patients. The study was conducted in an emergency department of a large hospital. Included in the study were patients who had been diagnosed with diabetes and had a serum glucose level greater than 200 mg per dL (11.1 mmol per L), with any symptoms of illness; patients who had symptoms suggestive of undiagnosed diabetes and a serum glucose level greater than 200 mg per dL; and patients who had diabetes and presented without any medical complaints (e.g., for medication refills). Patients were tested for the presence of ketonemia using all three screening techniques. If serum ketones were detected, an arterial blood gas analysis was performed. The criteria for the diagnosis of diabetic ketoacidosis were a serum glucose level greater than 200 mg per dL, metabolic acidosis with a corrected pH of less than 7.3, a serum bicarbonate concentration less than 15 mEq per L and ketonemia, positive at 1:2 dilution or greater. The criteria for the diagnosis of diabetic ketosis were a serum glucose concentration greater than 200 mg per dL, a positive serum ketone test and lack of acidosis as already defined.

Of the 697 patients enrolled in the study, 98 were diagnosed with diabetic ketoacidosis and 88 with diabetic ketosis. The urine ketone dip test was highly sensitive for diabetic ketoacidosis and diabetic ketosis with or without acidosis. A negative urine dip test eliminated the need for any other screening, although it should be noted that the hyperglycemic, hyperosmolar nonketotic state can occur with extreme hyperglycemia and absent serum ketones. If the urine dip test is positive for ketones, a serum ketone test should be performed to confirm or exclude the diagnosis of diabetic ketosis or diabetic ketoacidosis.

Limitations of the urine ketone dip stick test include false-positive results caused by captopril and other free sulfhydryl drugs. Other ketotic states such as alcoholic and starvation ketosis can also produce a positive urine ketone test without corresponding serum ketones, affecting the specificity of the urine dip test for detecting ketonemia.

The authors conclude that the urine ketone dip test is highly sensitive, has a high negative-predictive value and is cost effective, making it the most accurate and useful screening test for the diagnosis of diabetic ketoacidosis. Although it is less reliable in diagnosing diabetic ketosis, as an indicator it is superior to the anion gap or the serum bicarbonate level. A negative urine dip test reliably excludes the diagnosis of diabetic ketoacidosis or diabetic ketosis, while a positive test should prompt further testing, treatment, or both.