Antimicrobial activity of antibiotics in urine under different physiological conditions

Background

In vitro activity of antimicrobials is commonly assessed in well-characterized growth media like Mueller-Hinton broth (MHB); however, such media hardly reflect in vivo conditions. We set out to compare activity of selected, mostly novel, antibiotics in MHB and urine, an important compartment for bacterial infections, at different conditions.

Methods

Urine obtained from healthy volunteers was pooled and sterile filtered. Microdilution tests were performed with Escherichia coli ATCC 25922 in MHB and human urine. The pH of urine was adjusted to values ranging from 5 to 8. For simulating different glucose levels in urine of diabetic patients, urine was adjusted to glucose levels of 100 and 1000 mg/dL. Bacterial growth in different media was investigated by growth curves. Results obtained from MIC testing were confirmed by use of bacterial killing curves. Klebsiella oxytoca ATCC 700324 was used to investigate transferability of finding to other strains. Each experiment was performed 5 times.

Results

Growth of bacteria was similar for MHB and adjusted urine. Mean ratios (fold changes) of MICs for E. coli at the 6 different conditions simulated in urine compared to MIC-values in standard MHB are presented. Results were confirmed by time-killing curves and were concurrent for E. coli and K. oxytoca. Fosfomycin activity in urine was in unison higher than in MHB adjusted by glucose-6-phosphate.

Conclusions

Strong impact on activity was observed at low pH values for fluoroquinolones, trimethoprim, amikacin and tigecycline. Remarkably, all these antimicrobials act intracellulary while impact of pH on cell membrane-active antibiotics like β-lactams and colistin was much lower. Impairment of diffusion into the bacterial cell due to modification of the electric charge of the drug seems likely.