It is not generally possible to measure most organic acids in the serum of critically ill patients, due to rapid metabolism and methodological problems. Only the regular measurement of lactic acid and the arterial ketone body ratio (acetoacetate/beta-hydroxybutyrate, AKBR) have been introduced in clinical practice, but these parameters can represent only a part of the disturbed metabolism. In pediatric patients, a chromatographical urine analysis has been established for detection of inborn errors of metabolism, which allows the determination of more than 50 organic acids simultaneously (gas chromatographic (GC) analysis in combination with mass spectrometry (MS)]. In continuous treatment of acute renal failure, hemofiltrate is always available, but it contains only low protein concentrations and after the filtration process, metabolism is rapidly stopped. The sieving coefficient of lactic acid is nearly one in hemofiltration. The aim of our study was to compare results of the regular and CG/MS methods in blood and hemofiltrate for lactic acid, and to find other organic acids of possible clinical importance. We investigated serum (lactic acid) and hemofiltrate of 40 critically ill patitens, similar to the urine analysis method for infants. All patients suffered from acute renal failure and were treated by continuous veno-venous hemofiltration (CVVH). The conditions of treatment were standardized (spontaneous ultrafiltration in the first hour), and the material (blood/hemofiltrate) was taken one hour after the beginning of extracorporeal circulation. Statistical methods included correlation analysis, nonparametric ANOVA with Wilcoxon scores (ranks of data), and stepwise discriminant analysis. Regular and GC/MS methods in hemofiltrate showed a good correlation for lactic acid. The best correlation with lactic acid was found for 4-hydroxy-phenyllactic acid (n=20, r=0.866), 2-hydroxy-valeric acid (n=22, r=0.7491) and 2-hydroxybutyric acid (n=32, r=0.5148). Age, sex, diagnosis, and APACHE II score play a subordinate role, but the presence of glyceric and citric acid possibly have prognostic importance [nonparimetric ANOVA with Wilcoxon scores (ranks of data)], as does the combination of 3-hydroxypropionic acid, glyceric acid, and threonic-acid-4-lacton (stepwise discriminant analysis). It can be concluded that in acute renal failure, the measurement of lactic acid and AKBR can reflect only a small part of disturbed metabolism. Hemofiltrate can be a useful medium in describing metabolic processes in critically ill patients with acute renal failure. Some inherited metabolic diseases in infants (phenylketonuria, maple syrup disease) and ketoacidosis show similar metabolic modifications.