The Development of Simple Scoring System to Predict Urinary Tract Infection (UTI) in Patients with Stroke

Abstract

Urinary tract infection is a frequent problem after stroke. Although prior scoring systems for UTI after stroke have been developed, we developed a simple scoring system for all types of stroke in our own. The study was designed on retrospective data. The population includes 1496 patients with stroke who had been admitted at the Neurology Department of Pyongyang Medical College Hospital between January 2010 and August 2019. The patients were diagnosed with confirmed CT and MRI. Urinary tract infection (UTI) was diagnosed through urine culture: more than 100,100 colony-forming units per millimeter in patients with signs and symptoms. The UTI prediction scoring system was developed by means of the variables available on admission. The variables with significant difference between the non-UTI group and the UTI group were age (non-UTI versus UTI, 56.4 ± 7.2 vs. 59.0 ± 12.8; p<0.001), female (244 (24.2) vs. 176 (36.1), p<0.001), 300 ≦ SI (smoking index) (16 (2.4) vs. 48 (12.0), p<0.001), alcohol > 25 g/d (292 (29.0) vs. 184 (37.7), p<0.001), poststroke hyperglycemia (120 (10.3) vs. 163 (33.4), p<0.001), indwelling of urinary catheter (157 (15.6) vs. 351 (72.0), p<0.001), GCS (Glasgow Coma Scale) on admission (11.2 ± 3.9 vs. 8.5 ± 4.0, p=0.038), and WFNS (World Federation of Neurosurgeons) (in subarachnoid hemorrhage) on admission (2.9 ± 1.7 vs. 3.5 ± 1.5, p<0.001). The UTI prediction score ranged from 0 to 8 and produced an AUC (area under curve) of 0.800. The optimal cutoff point was 2.5 (sensitivity 64.3% and specificity 79.9%). So, the score ≧ 3 was the optimal score for the prediction of UTI after stroke.