Novel microfluidic systems for raman spectroscopy based antibiotic susceptibility tests

In the present work powerful microfluidic-based approaches for clinical-spectroscopic diagnostics of bacteria were developed. With special attention paid to urinary tract infections microfluidic systems were designed and realized that enable Raman spectroscopic identification of bacteria and their antibiotic resistance directly from patients’ material. The background was to provide innovative faster and yet sensitive and specific alternatives to the standard microbiological practice. A combined dielectrophoresis Raman setup was established. It took only 35 minutes to successfully distinguish the two typical pathogens of urinary tract infections from patients’ urine samples, Escherichia coli and Enterococcus faecalis. A centrifugal microfluidic platform was manufactured. Within 1.5 hours the important characteristics of Escherichia coli and Enterococcus faecalis recovered from the urinary tract of patients were acquired. The issue of polymicrobial infections was addressed. A non-linear hybrid statistical approach based on Partial Least Squares regression in combination with Maximum-Likelihood Parameter estimation methods led to accurate predictions of various mixing ratios of bacteria. The applicability of the dielectrophoresis-Raman setup as a measurement platform for antibiotic susceptibility tests was verified. With the development of an advanced statistical model based on Partial Least Squares Regression and Linear Discriminant Analysis, vancomycin resistant enterococci could be highly sensitively and specifically diagnosed within 1.5 hours. The technology was integrated into a closed microfluidic system. The functionality of the microfluidic system was successfully demonstrated by the determination of fluoroquinolone-resistances in Escherichia coli. It could also be shown that the Raman data of the quadrupole chips were compatible with those of the microfluidic system and hence, the method was shown to be highly robust.