A capillary blood ammonia bedside test following glutamine load to improve the diagnosis of hepatic encephalopathy in cirrhosis

Abstract

Background: Hepatic encephalopathy (HE) is a frequent and severe complication of cirrhosis. A single determination of ammonia in venous blood correlates poorly with neurological symptoms. Thus, a better biological marker is needed.

Aim: To make a diagnosis of HE, we explored the value of ammonia in capillary blood, an equivalent to arterial blood, measured at bedside following an oral glutamine challenge.

Methods: We included 57 patients (age 56 yrs; M/F: 37/20) with cirrhosis (alcoholic = 42; MELD score 13.8 [7-29], esophageal varices = 38) and previous episodes of HE (n = 19), but without neurological deficits at time of examination, and 13 healthy controls (age 54 yrs). After psychometric tests and capillary (ear lobe) blood ammonia measurements, 20 gr of glutamine was administered orally. Tests were repeated at 60 minutes (+ blood ammonia at 30'). Minimal HE was diagnosed if values were > 1.5 SD in at least 2 psychometric tests. Follow-up lasted 12 months.

Results: The test was well tolerated (nausea = 1; dizziness = 1). Patients showed higher values of capillary blood ammonia over time as compared to controls (0'-30'-60 minutes: 75, 117, 169 versus 52, 59, 78 umol/L, p < 0.05). At baseline, 25 patients (44%) had minimal HE, while 38 patients (67%) met the criteria for HE at 60 minutes (chi2: p < 0.01). For the diagnosis of minimal HE, using the ROC curve analysis, baseline capillary blood ammonia showed an AUC of 0.541 (CI: 0.38-0.7, p = 0.6), while at 60 minutes the AUC was 0.727 (CI: 0.58-0.87, p < 0.006). During follow-up, 18 patients (31%) developed clinical episodes of HE. At multivariate analysis, the MELD score (1.12 [1.018-1.236]), previous episodes of HE (3.2[1.069-9.58]), but not capillary blood ammonia, were independent predictors of event.

Conclusions: In patients with cirrhosis and normal neurological examination, bedside determination of ammonia in capillary blood following oral glutamine load is well tolerated and achieves a better diagnostic performance for minimal HE than basal capillary ammonia levels. However, capillary blood ammonia is a poor predictor of development of clinically overt HE.

Figure 1

Design of the study.

Figure 2

Changes in capillary blood ammonia during the glutamine challenge in patients with cirrhosis (n = 57) and in healthy subjects (n = 13). # p < 0.05 versus values in healthy subjects; * p < 0.05 versus baseline values.

Figure 3

Panel A: Evolution of ammonia in samples obtained simultaneously in capillary and venous blood in a subgroup of 8 patients with cirrhosis. # p < 0.05 versus baseline. Panel B: Capillary blood ammonia changes over the test period in patients with cirrhosis according to the MELD score using a cut-off of 12.

Figure 4

Blood ammonia levels according to the presence (MEH+) or absence (MEH-) of minimal HE at baseline and 60 minutes aftert the oral glutamine challenge in patients with cirrhosis.

Figure 5

ROC curves with respect to sensitivity and specificity of baseline (Panel A) and post glutamine load (Panel B) capillary blood ammonia levels for the diagnosis of minimal HE. At 60 minutes, using a cut-off level of 260 umol/l, the sensitivity was 79% and the specificity was 50%.