High Dynamic Range Bacterial Cell Detection in a Novel "Lab-In-A-Comb" for High-Throughput Antibiotic Susceptibility Testing of Clinical Urine Samples

Abstract

Antibiotic resistance in urinary tract infection is a major global challenge, and improved cost-effective and rapid antibiotic susceptibility tests (AST) are urgently needed to inform correct antibiotic selection. Although microfluidic technology can miniaturise AST, the high dynamic range of pathogen density found in clinical urine samples makes direct testing of clinical samples - rather than testing colonies from overnight agar plates - extremely challenging. We evaluated for the first time how pathogen concentration in urine affects microfluidic AST using a novel microplate-compatible high-throughput microfluidic AST system termed \"Lab-on-a-Comb\". When tested with clinical E. coli isolates at standardised density, these devices gave identical antibiotic susceptibility profiles to standard disc diffusion and microtitre plate tests. Bacterial detection directly in synthetic urine spiked with clinical E. coli UTI isolates was possible over a very large dynamic range of starting cell densities, from 103 - 108 CFU/mL which covers the range of pathogen cell densities found in patient urine. The lowest cell density where cell growth was reproducibly detected optically was 9.6x102 CFU/mL, corresponding to one single CFU detected in a 1 L microcapillary-an unprecedented level of sensitivity. Cell growth kinetics followed a simple Monod model with fast growth limited by the substrate availability and an estimated doubling time of 24.5 min, indicating optimal E. coli growth conditions within these microfluidic devices. There was a trade-off between sensitivity and speed of detection, with 105 CFU/mL detection possible within 2h, but 6h incubation required at 103 CFU/mL.