Letter: stepwise diagnosis in covert hepatic encephalopathy - do not forget ammonia.

Abstract

Listen

Translate article

We read with great interest the article by Greinert et al.1 In this paper, the authors proposed a stepwise approach for the diagnosis of covert hepatic encephalopathy (CHE). The aim was to screen patients who should undergo the gold standard Psychometric Hepatic Encephalopathy Score (PHES), using the measurements of critical flicker frequency (CFF) and MELD score. This paper suggests that PHES testing could be avoided in patients either with a MELD score of >24 and critical flicker frequency <43 Hz or <24 and >43 Hz, respectively, with a diagnostic accuracy for CHE of 81%. While we agree with this kind of approach and the value of the critical flicker frequency in this setting, we would like to emphasise that this clinical approach is probably helpful in hospitalised patients with high MELD scores. However, the diagnosis of CHE is also useful in out-patients with low MELD scores, because CHE has a negative prognostic value and because of an impaired quality of life which can be improved with specific treatments. In that regard, we would like to suggest the potential role of blood ammonia as a first-line test in a stepwise approach. First, ammonia plays a major role in the complex pathophysiology of hepatic encephalopathy and could correlate with the severity of hepatic encephalopathy.2 High levels of ammonia predicts the occurrence of overt hepatic encephalopathy,3, 4 as does the presence of CHE and it was recently shown that ammonia levels were higher in patients with CHE than in patients without hepatic encephalopathy.5, 6 It could therefore be a useful test, more specific that psychometric testing or CFF to pre-select patients. Moreover, it is a simple blood test that could be performed simultaneously with INR, creatinine and bilirubin measurements used to calculate the MELD score. Although its positive predictive value for the diagnosis of CHE may not be optimal, partially due to logistic problems in the measurement that can be solved by taking several precautions, the negative predictive value seems to be high and especially adapted to pre-select patients (where an emphasis is put on the low false-negative rate). The 2014 EASL/AASLD guidelines indeed recommend to reconsider the diagnosis of hepatic encephalopathy in the presence of normal ammonia.7 In our experience (personal data), out of a cohort of 51 cirrhotic patients admitted to our out-patient clinics to confirm or rule out the diagnosis of CHE (made using psychometric tests in our practice), only 8% displayed a MELD score >24, making the Greinert's algorithm not really useful in this context. However, when replacing MELD score by ammonia, and considering that a CFF >43 Hz and plasma ammonia <50 μmol/L would rule out the diagnosis of CHE, 87% of our patients were correctly classified. Hence, we suggest that Greinert's algorithm should be used in hospitalised patients, whereas in patients with low MELD scores, ammonia values could help to rule out CHE. Declaration of personal and funding interests: None.