Abstract Background and Objectives We developed a gas chromatography–mass spectrometry (GC-MS) method for the analysis of organic acids (OAs) in plasma using liquid-liquid extraction with acidified methanol. This method allows for the simultaneous analysis of several metabolites associated with energy metabolism, including lactic and pyruvic acids. Here, we evaluated the effect of different preanalytical variables (anticoagulants, delayed plasma separation, fasting vs nonfasting status) on plasma organic acid quantitation. Methods Nine organic acids (lactic, pyruvic, 3-hydroxybutyric, acetoacetic, 2-ketoisovaleric, 2-keto-3-methylvaleric, 2-ketoisocaproic, succinic, and glutaric) were extracted by deproteinizing with acidified methanol, oximated to preserve ketoacids, converted to volatile trimethylsilyl derivatives, and detected by GC-MS. Compound identification was obtained by retention time and fragmentation spectra using Agilent MassHunter software. We performed several comparison studies using blood collected from self-reported healthy participants (24-60 years of age; 14 females, 13 males). The effect of fasting/nonfasting status was evaluated in 14 participants by comparing samples collected after overnight fasting and 4 hours after meal. Paired t test was used for the statistical analysis of results. Results When compared to the preferred sample type, sodium heparin (green top) plasma EDTA plasma (purple top, n = 9) and serum (red top or serum separator tube; n = 5) displayed an unacceptable variability for several analytes. The biggest differences were observed for lactic and pyruvic acids. Lactic acid was significantly lower in EDTA plasma (9%-35% decrease; P = .0001), but higher in serum (4%-63% increase; P = .014), while pyruvic acid was significantly higher in EDTA plasma (4%-145% increase; P = .0002). The measurements of lactic and pyruvic acids were also affected by delayed separation of plasma in samples kept at room temperature. We observed a marked increase in lactate and decrease in pyruvate levels (+42% and –43% mean change, respectively) in plasma separated from whole blood after 2 hours of collection compared to samples processed within 30 minutes. Plasma separated 1 hour postcollection was acceptable. Overnight fasting (12-14 hours) did not have a significant impact on ketones (3-hydroxybutyric and acetoacetic acids) or branched-chain ketoacids (2-keto-3-methylvaleric, 2-ketoisocaproic, 2-ketoisovaleric acids). In contrast, fasting samples displayed higher lactic acid concentrations compared to nonfasting samples (24% increase on average, P = .0072), although in most fasting samples (13 out of 14), lactate concentrations were still within established normal range (600-2,600 µmol/L). Conclusions We identified several preanalytical factors influencing plasma organic acid quantitation, particularly lactic and pyruvic acids. Since those are critical analytes, suggestive of primary or secondary metabolic acidosis, appropriate steps should be taken to ensure sample quality during collection and processing.
Read full abstract