Because the correlation between ammonia levels and the severity of hepatic encephalopathy remains controversial, we prospectively evaluated the correlation in 121 consecutive patients with cirrhosis.

The diagnosis of hepatic encephalopathy was based on clinical criteria, and the severity of hepatic encephalopathy was based on the West Haven Criteria for grading of mental status. Arterial and venous blood samples were obtained from each patient. Four types of ammonia measurements were analyzed: arterial and venous total ammonia, and arterial and venous partial pressure of ammonia. Spearman rank correlations (r_s) were calculated.

Of the 121 patients, 30 (25%) had grade 0 encephalopathy (no signs or symptoms), 27 (22%) had grade 1, 23 (19%) had grade 2, 28 (23%) had grade 3, and 13 (11%) had grade 4 (the most severe signs and symptoms). Each of the four measures of ammonia increased with the severity of hepatic encephalopathy: arterial total ammonia (r_s = 0.61, P ≤0.001), venous total ammonia (r_s = 0.56, P ≤0.001), arterial partial pressure of ammonia (r_s = 0.55, P ≤0.001), and venous partial pressure of ammonia (r_s = 0.52, P ≤0.001).

Ammonia levels correlate with the severity of hepatic encephalopathy. Venous sampling is adequate for ammonia measurement. There appears to be no additional advantage of measuring the partial pressure of ammonia compared with total ammonia levels.