To the Editor.
—We read the article in the July Archives (138:1077-81, 1978) with great interest. The authors claimed that on the basis of their own experience and that of others, the method would seem to offer the greatest promise in identifying unsuspected upper urinary tract infections among patients with asymptoTest Results Positive Negative Totals Disease 16 10 26 present Disease 39 223 262 absent Totals 55 233 288 matic bacteriuria. Based on the criteria from Galen and Gambino,1 we determined the value of the determination of antibody-coated bacteria in upper urinary tract infection. Our analysis showing the sensitivity, specificity, accuracy, and predictive value is shown below.There were 288 specimens tested. Of these, 26 had a discharge diagnosis of acute pyelonephritis; of these, 16 had a positive test for antibody-coated bacteria. The other 262 samples were from cases not diagnosed as acute pyelonephritis; 39 of these had