Early Risk Stratification in Hospitalized Community-Acquired UTI: An 8-Item Bedside Score for Bacteremia and 30-Day Mortality

Background Community-acquired urinary tract infection (CA-UTI) is a common cause of sepsis in hospitalized adults, yet early identification of concomitant bacteremia and short-term mortality risk at presentation remains challenging. We aimed to define independent predictors of bacteremia and 30-day mortality and to derive a pragmatic bedside score for early risk stratification. Methods We conducted a retrospective cohort study of consecutive adults hospitalized with CA-UTI at a tertiary university hospital in Türkiye (January 2023–June 2025), including only the first eligible episode per patient. Bacteremia was defined as a pathogenic blood culture consistent with a urinary source obtained within 48 hours of arrival. Baseline clinical variables, comorbidities, severity indices, and admission biomarkers were extracted from electronic health records. Group comparisons used Shapiro–Wilk testing for normality, Student’s t-test or Mann–Whitney U test for continuous variables, and χ² or Fisher’s exact test for categorical variables. Predictors of bacteremia were identified using multivariable logistic regression with backward likelihood-ratio selection (age and sex forced). For 30-day mortality, a parsimonious multivariable logistic regression model was specified a priori to mitigate overfitting. Discrimination was assessed by receiver-operating characteristic (ROC) analysis with AUROC (95% CI), and AUROCs were compared by DeLong’s test. An unweighted 8-item risk-factor count score was constructed from admission variables. Results Among 358 patients (mean age 64.9 years), 117 (32.7%) had bacteremia. Independent predictors of bacteremia included male sex (female sex protective; aOR 0.31), shorter symptom duration (aOR per day 0.65), prior ESBL/carbapenem-resistant organism colonization/infection (aOR 2.88), Charlson Comorbidity Index ≥ 2 (aOR 9.27), diabetes mellitus (aOR 3.43), qSOFA ≥ 2 (aOR 6.61), CRP > 100 mg/L (aOR 2.09), and procalcitonin ≥ 0.5 ng/mL (aOR 2.01). The full model showed good discrimination (AUROC 0.83), comparable to the simplified score (AUROC 0.80; DeLong p = 0.32). Bacteremia rates were 3.8% (0–1 factors), 24.7% (2–3), and 75.0% (≥ 4). Overall 30-day mortality was 6.1%; after adjustment, bacteremia (aOR 9.74) and qSOFA ≥ 2 (aOR 21.82) independently predicted death. Conclusions In hospitalized CA-UTI, bacteremia is frequent and is best predicted by host vulnerability, early organ dysfunction, and inflammatory biomarkers rather than chronological age alone. An 8-item bedside score using readily available admission variables provides clinically actionable bacteremia risk stratification and may support early triage decisions; external validation is warranted.