Ultrafiltrate analysis confirms the specificity of the selected method for plasma ammonia determination.

The specificity of the direct enzymatic determination of plasma ammonia has hitherto not been unequivocally confirmed, because a suitable comparison method was lacking. Therefore a method variant was elaborated, which includes ultrafiltration to eliminate the high-molecular-mass components regarded as potential sources of unspecificity in the direct measurement procedure according to Rattliff, C.R. & Hall, F. F. (Select. Meth. Clin. Chem. 9, 85-90 (1982)). As the distribution of ammonia during plasma ultrafiltration is markedly influenced by pH and protein concentration, plasma pH is adjusted to 5.5 where the distribution ratio is 1 and nearly independent of actual protein concentration. Acidification significantly diminishes the spontaneous increase of ammonia in plasma at 2-4 degrees C, and the plasma ultrafiltrate is virtually stable. Taking into consideration the slow ammonia formation during sample preparation, excellent agreement was found between ammonia concentrations measured in plasma and in plasma ultrafiltrate, using samples with an apparently normal matrix (n = 30), dysproteinaemia (n = 32) or paraproteinaemia (n = 8). Our data show that the protein matrix of the sample does not cause significant unspecificity in the direct "endpoint" procedure for ammonia determination nor does it affect imprecision. In samples with added bilirubin (up to 252 mumol/l), haemolysate (haemoglobin up to 3.87 g/l) or lipid emulsion (triacylglycerol up to 3.86 mmol/l) ammonia values determined directly in plasma differed maximally by 4% from ultrafiltrate values. A simplified procedure for the ultrafiltration of plasma may be used routinely in clinical service in cases of grossly icteric, haemolytic or turbid samples.