Mid-infrared spectroscopy hyphenated with micro-dialysis is an excellent method for monitoring metabolic blood parameters as it enables the concurrent, reagent-free and precise measurement of multiple clinically relevant substances such as glucose, lactate and urea in micro-dialysates of blood or interstitial fluid. For a marketable implementation, quantum cascade lasers (QCL) seem to represent a favourable technology due to their high degree of miniaturization and potentially low production costs. In this work, an external cavity (EC) - QCL-based spectrometer and two Fourier-transform infrared (FTIR) spectrometers were benchmarked with regard to the precision, accuracy and long-term stability needed for the monitoring of critically ill patients. For the tests, ternary aqueous solutions of glucose, lactate and mannitol (the latter for dialysis recovery determination) were measured in custom-made flow-through transmission cells of different pathlengths and analyzed by Partial Least Squares calibration models. It was revealed, that the wavenumber tuning speed of the QCL had a severe impact on the EC-mirror trajectory due to matching the digital-analog-converter step frequency with the mechanical resonance frequency of the mirror actuation. By selecting an appropriate tuning speed, the mirror oscillations acted as a hardware smoothing filter for the significant intensity variations caused by mode hopping. Besides the tuning speed, the effects of averaging over multiple spectra and software smoothing parameters (Savitzky-Golay-filters and FT-smoothing) were investigated. The final settings led to a performance of the QCL-system, which was comparable with a research FTIR-spectrometer and even surpassed the performance of a small FTIR-mini-spectrometer.