Plasma amino acids as predictors of the severity and outcome of sepsis

Sepsis is a major catabolic insult resulting in a peripheral energy deficit which is made up in part by increased breakdown of lean body mass and oxidation of amino acids, principally the branched chain amino acids. The prognosis in any given case of sepsis is difficult to predict, but should theoretically be related to the degree of disturbance in peripheral energy deficit, which may in turn, be related to plasma amino acid pattern. In order to study whether this hypothesis was correct, plasma amino acids and some of their metabolic byproducts, the beta-hydroxyphenylethanolamines, were studied in 25 septic patients, and were used as discriminant variables in a series of computer performed discriminant analyses and multiple regressions. The two functions tested were the degree of metabolic septic encephalopathy as a determinant of the severity of sepsis and the final outcome in the septic patient. Plasma amino acid patterns exhibited elevated levels of the aromatic and sulfur containing amino acids, phenylalanine, tryosine, tryptophan, methionine, cysteine, and taurine, normal concentrations of alanine, and low normal concentrations of the branched chain amino acids, valine, leucine and isoleucine. Arginine levels, as previously noted, were very low. Patients not surviving the septic episode exhibited higher concentrations of aromatic and sulfur containing amino acids, while patients surviving sepsis had higher concentrations of the branched chain amino acids and arginine. When the degree of encephalopathy as a determinant of the severity of sepsis and step wise discriminant analysis with multiple crescent techniques were used, the best discriminant function between patients with and without encephalopathy was found to result from the interaction of cysteine, methionine, phenylalanine, isoleucine, leucine, and valine. These amino acids gave a correct classification in 82% of patients with no encephalopathy, and 80% of patients with septic encephalopathy. When the same amino acids were used for the discriminant analysis for patients dying of sepsis and patients surviving, the best discriminant function was achieved by using plasma concentrations of alanine, cysteine, methionine, isoleucine, arginine, tyrosine and phenylalanine resulting in 91% of the nonsurvivors, and 79% of the survivors correctly classified. The results suggest a close and significant relationship between the deranged energy metabolism and muscle protein breakdown in sepsis, and the outcome. This further suggests a central role for certain amino acids in perhaps predicting the severity of sepsis and its outcome.