Accuracy of copper-reduction and glucose-oxidase tests for various glucose concentrations

Three glucose-oxidase (Diastix) and two copper-reduction (Clinitest) urine tests for glucose were compared to determine the differences in accuracy for various concentrations of glucose. Fifty volunteers were assigned on a random basis to one of five urine-testing methods and to a testing order for eight glucose samples. The samples were assayed spectrophotometrically to verify the actual concentrations of glucose, ranging form 0.1 to 20.0%. Scale points were assigned to correspond to the different glucose concentrations. The deviation of the volunteers' determinations above or below the actual concentrations were recorded as error points. A one-way analysis was conducted on the error points to determine it there were significant differences among the five testing methods. Significant differences were found among the five testing methods (p less than 0.01). The five-drop Clinitest method was significantly less accurate than the 1:5 Diastix and the two-drop Clinitest methods. Significant differences were found among the five testing methods at low and medium glucose concentrations but not at the high concentrations. In the glucose concentration range of 0.1-0.25%, the Diastix method with no dilution was significantly less accurate (p less than 0.01) than the two-drop Clinitest method. In the 0.5-1.25% range, the two-drop Clinitest was more accurate than Diastix 1:5, 1:10, and five-drop Clinitest. Overall, the two-drop Clinitest procedure was most accurate, particularly at the lower concentrations of glucose.