Study Need and Importance

Inappropriate treatment for a presumed UTI in children with neurogenic bladder is a problem. This patient population will be continually exposed to antibiotics and hospitalizations across their life course, putting them at risk for developing multidrug antibiotic resistance. In turn, this may lead to increased unplanned health care utilization and more severe presentations, while also contributing to the overall worsening antimicrobial resistance patterns seen across health care settings. But choosing when to treat our patients with neurogenic bladder is vexing. Our current diagnostic options have limitations in this population because of the higher rate of asymptomatic bacteriuria and chronic cystitis. Abnormal urinalysis (UA) or culture results simply do not correlate with symptomatic UTIs.¹

In an effort to make some sense of when to treat, Madden-Fuentes et al proposed consensus-based criteria for UTI among these patients.² These criteria require the presence of 2 or more urological symptoms (including fever ≥38 °C, abdominal pain, back/flank pain, malodorous/cloudy urine, new/worsened urinary incontinence, and pain with urination/catheterization) plus pyuria (>10 white blood cells/high power field) and a positive urine culture of >100 k colony-forming units growth.²

Though this definition has gained traction in recent published works,³ its influence on practice has not been studied. We therefore sought to assess the rate of antibiotic receipt for presumed UTI and antibiotic overtreatment among children with spina bifida presenting to our own institution’s emergency department (ED).

What We Found

Across 236 ED encounters, the consensus-based criteria for UTI were met in only 14% of encounters, while antibiotics were initiated in 58%. Of the encounters in which antibiotics were initiated, 80% did not meet full criteria for UTI once cultures were resulted. Receipt of antibiotics was associated with fever, patient-reported cloudy urine, and all UA abnormalities analyzed, but pyuria was the strongest clinical variable associated with receipt of antibiotics—conferring an 8.3-fold higher likelihood of antibiotic receipt (see Table). In contrast, meeting full spina bifida criteria for UTI per se was not associated with receipt of antibiotics. These findings show that overtreatment is a common problem in this population and that nonspecific UA abnormalities may be overly relied upon on for decision-making compared to patient symptoms.

Limitations

This study was limited by its retrospective nature, particularly in relation to chart accuracy related to the assessment and documentation of symptoms. Furthermore, we did not analyze all vital signs or serum laboratory values (which were inconsistently collected), which may also influence antibiotic receipt.

Interpretation for Patient Care

The overtreatment rate of 80% identified in this study is substantially higher than what has been found among neurologically intact children similarly evaluated for UTI in the ED, which has ranged between 45%-56% in prior studies.^4,5 Pyuria was the most important factor associated with antibiotic receipt, which is consistent with prior studies identifying pyuria as a driver of overtreatment among other populations with asymptomatic bacteriuria.⁶ It should be noted that pyuria at this current threshold (>10 white blood cells/high power field) is largely historical and was not made with the neuropathic bladder population in mind.⁷ These results suggest that the threshold for what is considered “significant” pyuria in this population should be rethought.

Diagnosing a UTI in our patients with neurogenic bladder is a diagnostic challenge even for the most seasoned clinicians; however, several initial steps can be taken to help improve overtreatment. In the absence of fevers, sepsis, or other concerning patient risk factors (such as renal insufficiency or vesicoureteral reflux), attempts should be made to monitor patients with unclear symptoms.⁸ Aggressive management of constipation, hydration, and bladder irrigation are useful first steps that can be undertaken while culture results and patient symptoms are monitored at home.⁹ Callback systems for patient monitoring, while challenging, have been shown to reduce total antibiotic receipt.¹⁰ When patients are treated empirically, a reliable workflow for cessation of antibiotics after a negative culture is essential. Institutions should work to create collaborative and criteria-driven protocols for diagnosis and treatment of this population.