PLoS ONE (Jan 2020)

A diagnostic algorithm for detection of urinary tract infections in hospitalized patients with bacteriuria: The "Triple F" approach supported by Procalcitonin and paired blood and urine cultures.

  • Kathrin Rothe,
  • Christoph D Spinner,
  • Birgit Waschulzik,
  • Christian Janke,
  • Jochen Schneider,
  • Heike Schneider,
  • Krischan Braitsch,
  • Christopher Smith,
  • Roland M Schmid,
  • Dirk H Busch,
  • Juri Katchanov

DOI
https://doi.org/10.1371/journal.pone.0240981
Journal volume & issue
Vol. 15, no. 10
p. e0240981

Abstract

Read online

For acute medicine physicians, distinguishing between asymptomatic bacteriuria (ABU) and clinically relevant urinary tract infections (UTI) is challenging, resulting in overtreatment of ABU and under-recognition of urinary-source bacteraemia without genitourinary symptoms (USB). We conducted a retrospective analysis of ED encounters in a university hospital between October 2013 and September 2018 who met the following inclusion criteria: Suspected UTI with simultaneous collection of paired urinary cultures and blood cultures (PUB) and determination of Procalcitonin (PCT). We sought to develop a simple algorithm based on clinical signs and PCT for the management of suspected UTI. Individual patient presentations were retrospectively evaluated by a clinical "triple F" algorithm (F1 ="fever", F2 ="failure", F3 ="focus") supported by PCT and PUB. We identified 183 ED patients meeting the inclusion criteria. We introduced the term UTI with systemic involvement (SUTI) with three degrees of diagnostic certainty: bacteremic UTI (24.0%; 44/183), probable SUTI (14.2%; 26/183) and possible SUTI (27.9%; 51/183). In bacteremic UTI, half of patients (54.5%; 24/44) presented without genitourinary symptoms. Discordant bacteraemia was diagnosed in 16 patients (24.6% of all bacteremic patients). An alternative focus was identified in 67 patients, five patients presented with S. aureus bacteremia. 62 patients were diagnosed with possible UTI (n = 20) or ABU (n = 42). Using the proposed "triple F" algorithm, dichotomised PCT of < 0.25 pg/ml had a negative predictive value of 88.7% and 96.2% for bacteraemia und accordant bacteraemia respectively. The application of the algorithm to our cohort could have resulted in 33.3% reduction of BCs. Using the diagnostic categories "possible" or "probable" SUTI as a trigger for initiation of antimicrobial treatment would have reduced or streamlined antimicrobial use in 30.6% and 58.5% of cases, respectively. In conclusion, the "3F" algorithm supported by PCT and PUB is a promising diagnostic and antimicrobial stewardship tool.