BackgroundAlthough urinary tract infections (UTIs) are extremely common, isolation of causative uropathogens is not always routinely performed, with antibiotics frequently prescribed empirically. This study determined the susceptibility of urinary isolates from two Health and Social Care Trusts (HSCTs) in Northern Ireland to a range of antibiotics commonly used in the treatment of UTIs. Furthermore, we determined if detection of trimethoprim resistance genes (dfrA) could be used as a potential biomarker for rapid detection of phenotypic trimethoprim resistance in urinary pathogens and from urine without culture.MethodsSusceptibility of E. coli and Klebsiella spp. isolates (n = 124) to trimethoprim, amoxicillin, ceftazidime, ciprofloxacin, co-amoxiclav and nitrofurantoin in addition to susceptibility of Proteus mirabilis (n = 61) and Staphylococcus saprophyticus (n = 17) to trimethoprim was determined by ETEST® and interpreted according to EUCAST breakpoints. PCR was used to detect dfrA genes in bacterial isolates (n = 202) and urine samples(n = 94).ResultsResistance to trimethoprim was observed in 37/124 (29.8%) E. coli and Klebsiella spp. isolates with an MIC90 > 32 mg/L. DfrA genes were detected in 29/37 (78.4%) trimethoprim-resistant isolates. Detection of dfrA was highly sensitive (93.6%) and specific (91.4%) in predicting phenotypic trimethoprim resistance among E. coli and Klebsiella spp. isolates. The dfrA genes analysed were detected using a culture-independent PCR method in 16/94 (17%) urine samples. Phenotypic trimethoprim resistance was apparent in isolates cultured from 15/16 (94%) dfrA-positive urine samples. There was a significant association (P < 0.0001) between the presence of dfrA and trimethoprim resistance in urine samples containing Gram-negative bacteria (Sensitivity = 75%; Specificity = 96.9%; PPV = 93.8%; NPV = 86.1%).ConclusionsThis study demonstrates that molecular detection of dfrA genes is a good indicator of trimethoprim resistance without the need for culture and susceptibility testing.
Read full abstract