Antibiotic use and antimicrobial activity in urine of febrile patients

Abstract

OBJECTIVES: Prior antibiotic use may decrease sensitivity of bacterial cultures leading to diagnostic difficulties. Our objectives were to detect antimicrobial activity in urine of patients presenting with fever and to correlate urine antimicrobial activity to blood culture results and antibiotic history. DESIGN, SETTING AND METHODS: This is carried out as part of a fever study during a one-year period. A sample of urine is collected within 24 hours of admission before antibiotics from consenting, febrile (>38°C) patients. Antimicrobial activity in urine is detected by applying urine-soaked discs onto agar plates inoculated with standard bacterial cultures. All patients have extensive microbiological investigations including high-quality blood cultures (with leptospiral cultures), dengue and other viral serology. RESULTS: From 117 patients recruited in 4 months, 76(65%) reported taking medication for their fever, of which 40(53%) said this was an antibiotic. Standard E. coli and S. pyogenes cultures detected antimicrobial activity in 32(27%) and 25(21%) of urines respectively. Using both strains together increased detection to 39%. Only 19(47.5%) reporting prior antibiotic use had urine antimicrobial activity. From 41 denying prior medication, 8(20%) had urine antimicrobial activity. From 43 suspected of having bacterial infection, 46% had urine antimicrobial activity. No patients with positive bacterial cultures had detectable urine antimicrobial activity, whereas 53% of negative bacterial cultures had positive urine antimicrobial activity. CONCLUSIONS: Clinical history is unreliable in assessing prior antibiotic use. Detectable urine antimicrobial activity may lead to negative blood culture results. Therefore in suspected bacterial infections, an assessment of urine antimicrobial activity along with microbiological cultures will reduce the diagnostic dilemma.