Assessment of an automated chemiluminescence nitrogen analyzer for routine use in clinical nutrition.

An automated method of chemiluminescence analysis of nitrogen used routinely for 4 yr. Liquid samples (urine, enteral, and parenteral feeds) required simple dilution, whereas feces required a modified acid-digestion procedure, before analysis. For urine samples, the coefficient of variation was within batch from 0.9-3.6%, and between batch 4.3-7.6%. At a sample injection rate of 2 microliter/sec, the useful dynamic range, for urine diluted 1:200, was 0-14 g N/liter. Precision for fecal nitrogen analysis was 3.8-6.7% for samples of low to high nitrogen content. The correlation between this technique and an established Kjeldahl method for fecal analysis was studied (r = 0.96, slope = 1.30). The discrepancy between the methods was due to inefficient conversion of nitrogen to NH4+ during Kjeldahl digestion of feces, rather than systematic errors in chemiluminescence analysis. Reliability was as good as for other automated clinical analyzers and sample cost was ca. 0.22 pounds. It has proved possible to analyze approximately 80 samples in the working day. The efficiency of measuring 24-hr urine urea-nitrogen (UUN) and total urine nitrogen (TUN) in patients on general wards was measured. Results were obtained on 87% of TPN days, but large variations were noted in UUN/TUN from less than 30% to greater than 90% (average 75.7%) in patients receiving TPN, and from less than 55% to 100% (average 83.8%) in patients receiving enteral nutrition. In contrast, UUN/TUN was 87.0% and 84.0% in healthy subjects, fasted or receiving iv nutrition, respectively. We therefore expect that clinical nutritionists will find increasing applications for this method of nitrogen analysis.