The measurement of ammonia in whole blood.

The recent widespread interest in blood ammonia determinations in normal and abnormal clinical states (McDermott and Adams, 1954; Schwartz, Phillips, Gabuzda, and Davidson, 1953; White, Phear, Summerskill, and Sherlock, 1955), together with discrepant reports from various laboratories, have prompted a critical examination of the Conway microdiffusion technique in blood analysis. Method The method adopted was essentially that described by Conway (1950) with the following modifications. No. 2 units were used with N/4000 Ba(OH)2 in the burette, 0.1 ml. N/2000 HCI in the centre compartment of the unit, and 0.2 ml. saturated K2C03 and 0.2 ml. standard solution or blood in the outer compartment. The measuring error was estimated at 5%. In order to avoid difficulties with the end-point, the blood/carbonate mixture was removed by water pump aspiration immediately before titrating the residual acid. In this series of analyses blood was withdrawn from the patient with a 5 ml. syringe and No. 1 needle without venous stasis, collected into a bottle containing 50 units of dried heparin, rapidly mixed, and samples withdrawn for analysis two minutes after the beginning of the venipuncture. With practice this could be timed very accurately. No significant difference was found between heparinized blood and fresh unclotted samples. In some experiments ammonium sulphate was added to the blood using the stock ammonium sulphate solution (100 jug. NH3-N per ml.) in microlitre quantities, thereby producing negligible dilution of the blood samples. Results Serial estimations of the ammonia in standard ammonium sulphate solutions, in samples of freshly shed heparinized normal blood, and in heparinized normal blood, to which measured quantities of ammonium sulphate had been added, were made after given intervals of incubation at room temperature (230 C.). These results are shown in Figs. 1-4. It is interesting to note that, when the difference between the curves for normal blood and that for blood plus added ammonia is plotted, a very accurate estimate of the added ammonia can be obtained. Estimations of the ammonia content of blood from cirrhotic patients were made in the same manner. Two typical curves are shown (Fig. 5).