A metabonomic approach to early prognostic evaluation of experimental sepsis by (1)H NMR and pattern recognition.

ABSTRACT

This study proposes an NMR-based metabonomic approach to early prognostic evaluation of sepsis. Forty septic rats receiving cecal ligation and puncture (CLP) were divided into the surviving group and nonsurviving group on day 6, while 20 sham-operated rats served as the control group. Serum samples were collected from septic and sham-operated rats at 12 h after surgery and analyzed using (1)H NMR spectroscopy. Orthogonal partial least squares (OPLS) were applied and showed clustering according to predefined groups, indicating that NMR-based metabolic profiling could reveal pathologic characteristics in the serum of sham-operated, surviving, and nonsurviving septic rats. In addition, six characteristic metabolites including lactate, alanine, acetate, acetoacetate, hydroxybutyrate, and formate, which are mainly involved in energy metabolism, changed markedly in septic rats, especially in the nonsurvivors. Using these metabolites, a predictive model for prognostic evaluation of sepsis was constructed using a radial basis function neural network (RBFNN) with a prediction accuracy of about 87% by test samples. The results indicated that the NMR-based metabonomic approach is a potential technique for the early prognostic evaluation of sepsis.