When one size does not fit all: evaluating urine culture stewardship strategies for high-risk pediatric patients

Reflex urine culture protocols reduce inappropriate antibiotic use in adults but remain understudied in high-risk pediatric populations, including hematology/oncology and neonatal intensive care unit (NICU) patients. In this study, we evaluated the predictive value of urinalysis (UA) parameters for urine culture positivity within a quaternary care pediatric medical setting. We performed a retrospective analysis of 99,348 urine samples collected between January 2016 and January 2025. Among these, 20,019 paired UA-urine culture specimens were available, including 15,410 (77.7%) clean-catch and 4,609 (22.3%) in/out catheter specimens. UA markers assessed included leukocyte esterase (LE), nitrite, white blood cells, and microscopic bacteria. Across all specimens, the absence of UA markers (LE ≥ trace or nitrite positivity) demonstrated a negative predictive value (NPV) of 98.2% and sensitivity of 89.0% for a positive urine culture (≥10⁵ CFU/mL for clean-catch, and ≥5 × 10⁴ CFU/mL for in/out catheter specimens). The addition of microscopic parameters did not improve diagnostic accuracy. In hematology/oncology patients ( n = 1,344), sensitivity was significantly reduced compared with the overall cohort (74.7% vs 89.0%, P < 0.05). Among neutropenic patients ( n = 497), NPV remained high (98.7%), but sensitivity declined to 50.0% ( P = 0.06). Overall, 1.8% of neutropenic patients with positive cultures had negative UA results, many of whom had prior antibiotic exposure. Implementation of a UA-based reflex protocol would have reduced urine culture orders by 65.3%. Negative UA reliably excluded urine culture positivity in most pediatric patients, supporting reflex culture protocols. However, reduced sensitivity in neutropenic hematology/oncology and NICU patients suggests such protocols are inappropriate for these groups. IMPORTANCE Retrospective analysis of over 20,000 pediatric urine specimens demonstrated that screening urinalysis using leukocyte esterase or nitrite reliably predicted a negative urine culture, achieving a 98.2% negative predictive value. Furthermore, the addition of microscopic white blood cell counts did not improve overall diagnostic accuracy compared with automated urinalysis parameters alone. Implementing a urinalysis-based reflex algorithm would safely reduce unnecessary urine cultures by 65.3%, optimizing diagnostic stewardship and laboratory workflows. However, diagnostic sensitivity was significantly compromised in high-risk pediatric cohorts, including hematology/oncology, neutropenic, and neonatal intensive care unit patients. These findings demonstrate that while urinalysis reflex protocols are highly effective for general pediatric populations, diagnostic stewardship algorithms must account for host immune status to avoid missing true urinary tract infections in high-risk patients.