Diagnosis of inborn errors of metabolism using 1H NMR spectroscopic analysis of urine.

Abstract

Nuclear magnetic resonance (NMR) spectroscopy is most widely known for its applications in the area of chemical or biomolecular structure elucidation. Improvements in field strength and spectrometer sensitivity now permit analysis of small molecular species in body fluids, including organic acids, amino acids, toxins and drugs (Iles and Burns 1995). NMR measures the radiofrequency electromagnetic radiation that is absorbed when it interacts with atomic nuclei. Because hydrogen ( 1 H) is the most abundant element in biologically important compounds, most NMR is specific to this element. The most widely used form of NMR spectroscopy is liquid-state NMR, which yields high spectral resolution and detailed structural information, and has the capability to provide a full biochemical analysis of biofluids (Leibfritz 1996); Nicholson et al 1995; Smith and Blandford 1995; Videen and Ross 1994). It is predicted that high-resolution NMR spectroscopy, assisted by pattern recognition methods, will become an important diagnostic tool in medicine and will permit rapid identification of biological compounds at a concentration as low as 10 μmol/L. Urine is particularly useful because it is free of macromolecules such as proteins or lipids that may complicate the analysis (Holmes et al 1994). We have examined the application of 1 H NMR spectroscopic analysis of urine to the diagnosis of inherited metabolic defects of amino acid and organic acid metabolism. Since no pretreatment of the urine is necessary, the analysis is nonselective, and both water-soluble and water-insoluble compounds can be examined simultaneously. A novel refinement in the technique is the development of a neural network for automated identification of urinary metabolites.