Glucosuria: accuracy and precision of laboratory diagnosis by dip stick analysis.

Two urine dip-stick assays for glucose were studied and compared to a quantitative hexokinase procedure. We determined sensitivity, specificity, predictive value of "normal" and "abnormal" dip stick results, assay efficiency, technologist precision, and method association. Precision was poor. Attempts to estimate urine glucose concentration with the dip sticks frequently produced gross inaccuracies. We believe the dip stick glucose assay should be considered qualitative only; as a qualitative test, the dip sticks are fairly efficient at distinguishing glucose concentrations of less than 0.3 g/L from concentrations greater than 0.3 g/L.