Technical pitfalls in measurement of venous plasma NH3 concentration.

Measurement of venous plasma NH3 in normal subjects by the ion-exchange method of Forman [Clin. Chem. 10, 497 (1964)] in a hospital clinical laboratory gave a mean value of 640 mug/liter (range, 300-1320 mug/liter; intraassay, intra-individual, and inter-individual coefficients of variation, 8, 47, and 47%, respectively). The following conditions adversely affect the reproducibility of the test: pollution of laboratory atmosphere and glassware by NH3-containing detergents; smoking by patient or analyst; delay, turbulence, or use of heparin lock in venipuncture; delay or warming of plasma above degrees C before mixing it with resin; and delay in colorimetric analysis of resin eluate. When these sources of error were eliminated, the mean value for normals was reduced to 330 mug/liter, the range was narrowed to 220-470 mug/liter, and the above-mentioned CV's were 5, 16, and 17%, respectively. With the precautions cited, furthermore, the intra-assay and intra-individual CV's for fasting NH3 concentration in cirrhotic patients were similarly reduced. An NH3 tolerance test was done by administering a standard dose of NH4Cl to patients and measuring venous plasma NH3 at 0, 15, 30, 60, and 90 min; the NH3 tolerance was quantified from the area under the curve relating concentration to time (mug - min/liter). As measured in the clinical laboratory, NH3 tolerance of cirrhotic patients showed intra-assay and intra-individual CV's of 50 to 90%. When the tolerance tests were repeated in the same subjects with the laboratory precautions listed above, these CV's were reduced to 8-15%.