Translating a NMR-based test for diagnosis of urinary tract infection for routine clinical use

Abstract

Urinary tract infection (UTI) is a common bacterial infection in general populations. From the economic view, it causes numerous clinician visits, hospitalizations, cost of treatment and thus financial burden in a society. Additionally, it is also significantly associated with increased rates of morbidity, mortality and antimicrobial resistance. Current medical laboratory tests, such as urine microscopy, dipstick urinalysis and urine culture, have been widely used for the diagnosis of UTI. However, urine microscopy and dipsticks usually provide low sensitivity or specificity while bacterial culture takes time and is labor-intensive.

1H Nuclear Magnetic Resonance (NMR) spectroscopy has been used to identify urinary acetic acid as a discriminatory biomarker for the diagnosis of UTI. Despite it is highly sensitivity and highly specificity, NMR spectrometer is expensive and is difficult to be installed in routine clinical laboratory. On the other hand, colorimetric assay for acetic acid through enzymatic reactions provides a new option for quantitative detection of urinary acetic acid. Translation of the NMR-based approach to the colorimetric method for the quantitation of urinary acetic acid can provide further information in developing a simple, rapid, and sensitive test for diagnosis of UTI.

A total of 191 midstream urine samples were recruited in one hospital cluster in Hong Kong. Samples showing positive urine culture with bacteria growth > 〖10〗^5 colony forming units/ml were classified as UTI cases whereas samples showing both dipstick negative and microscopy negative without previous UTI history were classified as non-UTI controls. 18 Non-UTI controls and 25 UTI cases were selected randomly to perform the quantitative detection of urinary acetic acid by an enzymatic colorimetric assay. Statistical analysis was performed using Microsoft® Excel® 2010 software.

The optimal diagnostic cutoff for urinary acetic acid was determined using receiver operating characteristic (ROC) which showed an area under ROC curve of 0.993 with a sensitivity of 100% and a specificity of 94.4% at the cutoff 367.2 μM. The two groups were compared using Mann-Whitney U test which were significantly different with a p value of <0.0001. The mean urinary acetic acid in the case was 941.4 μM and that in the control was 305.2 μM. A 3-fold difference was observed between the mean of the two groups.

We have successfully translated a NMR-based test into a enzymatic colorimetric assay which can be incorporated for routine clinical use. This method requires a spectrophotometer set up which is affordable and available in most chemical laboratory. We envisage this method can provide accurate and efficient diagnosis of bacterial UTI in the future.