Prediction of complicated urinary tract infections in patients with type 2 diabetes: a questionnaire study in primary care

Background: Relatively few data are available to predict a complicated course of community-acquired complicated urinary tract infections (UTIs) in patients with diabetes type 2 (DM2). The aim of this study was to assess predictors for a complicated course of UTIs in DM2 patients in primary care. Method: We conducted a cross-sectional questionnaire study among DM2 patients aged over 45 years as part of an educational trial. The combined outcome measure was a complicated course of UTI, defined as a self-reported episode of acute pyelonephritis, prostatitis or recurrent cystitis in the 12 months before the trial. Patients with an outcome were all verified by review of medical records. A prediction model was derived with multivariable logistic regression analysis. Results: Of the 1151 trial participants, 94 (8%) had a self-reported community-acquired complicated course of UTIs and 62 (66%) of these were medically-attended. Independent predictors for a complicated course were age above 60 years (adjusted odds ratio (OR): 1.74; 95% confidence interval (CI): 0.99-3.03), chronic use of antibiotics (adjusted OR: 5.50; 95% CI: 2.31-13.08), more than 6 physician contacts in previous year (adjusted OR: 3.60; 95% CI: 2.00-6.49), hospitalization in previous year (adjusted OR: 1.36; 95% CI: 1.00-1.85), renal disease (adjusted OR: 4.92; 95% CI: 1.59-15.18) and incontinence of urine (adjusted OR: 3.78; 95% CI: 1.93-7.38). Area under the receiver-operating curve was 0.72 (95% CI: 0.66-0.78). Analysis according to medically attended complicated UTIs did not change our findings. Conclusion: Easily obtainable predictors from medical history can be used to accurately predict a complicated course of UTIs in DM2 patients.