Monitoring haemodialysis using electronic nose and chemometrics

An ever-increasing number of patients have to undergo regular renal dialysis to compensate for acute or chronic renal failure. The adequacy of the treatment has a profound effect on patients' morbidity and mortality. Therefore, it is necessary to assess the delivered dialysis dose. For the quantification of the dialysis dose, two parameters are most commonly used, namely the K(t)/V value (normalised dose of dialysis) and the urea reduction rate, yet the prescribed dialysis dose often differs from the actual delivered dialysis dose. Currently, no interactive process is available to ensure optimal treatment. The aim of this study was to investigate the potential for an "electronic nose" as a novel monitoring tool for haemodialysis. Blood samples were analysed using an electronic nose, comprising an array of 14 conducting polymer sensors, and compared to traditional biochemistry. Principal component analysis and hierarchical cluster analysis were applied to evaluate the data, and demonstrated the ability to distinguish between pre-dialysis blood from post-dialysis blood independent of the method used. It is concluded that the electronic nose is capable of discriminating pre-dialysis from post-dialysis blood and hence, together with an appropriate classification model, suitable for on-line monitoring.