Direct microfluidic antibiotic resistance testing in urine with smartphone capture: significant variation in sample matrix interference between individual human urine samples.

Rapid and portable direct tests for antibiotic resistance in human clinical samples such as urine could reduce misuse of precious antimicrobials, by allowing treatment decisions to be informed by microfluidic diagnostic tests. We demonstrate that the variable composition of human urine can significantly affect the antibiotic minimum inhibitory concentration (MIC) measured using microfluidic devices. The urine sample matrix interference was not observed in pooled normal urine, emphasising the critical importance of assessing matrix interference with a wide range of individual urine samples, rather than a few standardised or pooled controls. Both dilution into assay medium and inclusion of buffer could reduce the matrix interference, but dilution may affect analytical sensitivity by increasing the minimum bacterial cell density needed in a sample for growth to be detected, especially for miniaturised devices that test small sample volumes. We conclude it is vital to fully assess and optimise novel analytical microbiology tools using multiple individual urine samples, otherwise the high variation in matrix interference will compromise the clinical performance of these rapid diagnostics that are urgently needed to tackle the global threat of antimicrobial resistance.