OBJECTIVE: To investigate the accuracy of history taking to diagnose asymptomatic bacteriuria (ASB) in diabetic women, and the added value of leukocyturia. METHODS: Data were obtained from a multicenter study including 465 women with diabetes. Many patient characteristics were considered as potential diagnostic determinants. A urinary leukocyte count and a urine culture (the criterion standard) were performed. Logistic regression analyses were performed and areas under the receiver operating characteristic curves (AUC) were calculated. RESULTS: For women with type 1 diabetes (n = 236; ASB 11%), duration of diabetes and glycosylated hemoglobin (GHb) were powerful predictors of ASB. The AUC of the model including these two variables was 0.66 (95% confidence interval (CI) 0.53-0.78). After addition of leukocyturia, the AUC increased considerably to 0.78 (95% CI 0.68-0.88; p = 0.018). For women with type 2 diabetes (n = 229; ASB 19%), age and the number of symptomatic urinary tract infections (UTIs) in the previous year were the strongest predictors of ASB. The AUC of the model including these variables was 0.70 (95% CI 0.61-0.80). After addition of leukocyturia, the AUC increased to 0.79 (95% CI 0.71-0.86; p = 0.023). CONCLUSION: In diabetic women, ASB can be diagnosed using two easily obtainable variables (duration of diabetes and GHb for women with type 1 diabetes, and age and the number of UTIs in the previous year for women with type 2 diabetes) in combination with a urinary leukocyte count. This results in a model with sufficient accuracy (AUC > 0.75).
OBJECTIVE: To investigate the accuracy of history taking to diagnose asymptomatic bacteriuria (ASB) in diabeticwomen, and the added value of leukocyturia. METHODS: Data were obtained from a multicenter study including 465 women with diabetes. Many patient characteristics were considered as potential diagnostic determinants. A urinary leukocyte count and a urine culture (the criterion standard) were performed. Logistic regression analyses were performed and areas under the receiver operating characteristic curves (AUC) were calculated. RESULTS: For women with type 1 diabetes (n = 236; ASB 11%), duration of diabetes and glycosylated hemoglobin (GHb) were powerful predictors of ASB. The AUC of the model including these two variables was 0.66 (95% confidence interval (CI) 0.53-0.78). After addition of leukocyturia, the AUC increased considerably to 0.78 (95% CI 0.68-0.88; p = 0.018). For women with type 2 diabetes (n = 229; ASB 19%), age and the number of symptomatic urinary tract infections (UTIs) in the previous year were the strongest predictors of ASB. The AUC of the model including these variables was 0.70 (95% CI 0.61-0.80). After addition of leukocyturia, the AUC increased to 0.79 (95% CI 0.71-0.86; p = 0.023). CONCLUSION: In diabeticwomen, ASB can be diagnosed using two easily obtainable variables (duration of diabetes and GHb for women with type 1 diabetes, and age and the number of UTIs in the previous year for women with type 2 diabetes) in combination with a urinary leukocyte count. This results in a model with sufficient accuracy (AUC > 0.75).