Quick diagnosis of hepatic encephalopathy: Fact or fiction?


  • Potential conflict of interest: Nothing to report.

To the Editor:

Limited attention has been devoted to the differential diagnosis of neurological/psychiatric disturbances in cirrhosis. Therefore, there is a risk that in these patients virtually any neurological/psychiatric abnormality will be considered, by default, to be due to hepatic encephalopathy (HE).[1]

We run a daily HE clinic, to which patients are referred for evaluation of diagnosed/suspected HE, or for pretransplant screening. Our evaluation includes laboratory tests (full blood count, liver/kidney function, electrolytes, venous ammonia, thyroid stimulating hormone [TSH], C-reactive protein [CRP], and vitamin B12), examination by a physician and a neuropsychologist, a comprehensive set of neuropsychological tests, critical flicker frequency (CFF), and electroencephalography (EEG). If needed, cerebral imaging and/or repeat evaluations after the institution of appropriate treatment are performed. A final diagnosis is reached after joint review of patient history and all tests, taking into account the known features of the HE profile (i.e., impairment of psychomotor speed, attention, executive function and inhibition, with relatively preserved memory, slowed EEG, low CFF, and high ammonia levels).

Between June 2009 and May 2013, we evaluated 276 patients (94% outpatients; 200 males, age 58 ± 11 years, Child-Pugh A 37%, B 41%, C 22%, Model for Endstage Liver Disease [MELD] 13 ± 6), 61 of whom were seen more than once. Of the 276, 99 (36%) were neuropsychiatrically unimpaired; 90 (33%) had clinically obvious neurological/psychiatric abnormalities and 87 (32%) had subclinical abnormalities (neuropsychological and/or EEG and/or CFF). Of 90 patients with overt abnormalities, 35 were deemed to have HE and 55 had other comorbidities that could contribute or fully explain their neurological/psychiatric profiles (Fig. 1). Of 87 patients with subclinical abnormalities, 47 were deemed to have minimal HE and 40 had comorbidities that could contribute or fully explain their profiles. Thus, 95/177 patients (54%) with overt or subclinical alterations had comorbidities other than HE that could contribute to these alterations. Comorbidities and confounders included recent alcohol misuse (n = 15), cerebrovascular disease (17), malnutrition (6), severe hyponatremia (2) or hypoglycemia (1), hypothyroidism (2), infection/sepsis (5), psychoactive drugs (4), and various combinations of the above in a single patient. Several of these comorbidities are unlikely to benefit from a new liver. Patients with comorbidities/confounders were significantly older than those without (59.8 ± 10.8 versus 56.8 ± 11.5 years, P = 0.04).

Figure 1.

Time-course of the neurological/psychiatric profile of a 68-year-old female patient with alcohol-related cirrhosis (abstinent > 5 years), qualified as having both overt HE and cerebrovascular disease at baseline. Dark grey plot: relative theta power, or slow activity, of the electroencephalogram (EEG);[2] light grey plot: psychometric hepatic encephalopathy score (PHES).[3] Dark grey dashed line: abnormality threshold for the relative theta power (≥35%); light grey dashed line: abnormality threshold for the PHES (≤4; Italian normative data). After the institution of ammonia-lowering treatment (lactulose and then lactulose plus rifaximin), both the EEG and psychometric performance initially improved (i.e. reduction in EEG slow activity and increase in PHES score). Over time, psychometric performance worsened in the absence of significant changes in the EEG, ammonia levels or the degree of hepatic failure, suggesting that cerebrovascular disease had started to dominate the picture. Imaging confirmed the development of significant cerebral atrophy and she is now under the joint care of a hepatologist and a geriatrician. (Critical flicker frequency data are not presented as they were obtained only on two occasions.)

Short, stand-alone tests for HE diagnosis such as CFF, brief sets of questions, and quick computerized tests have long been advocated for use in the clinic or at the bedside, and may be adequate for screening purposes. However, considerably more comprehensive evaluations seem necessary as a second step, especially for patients on a transplant list. In the same way as we do not expect an ECG, an echocardiogram, or an angiogram to provide us with comprehensive information on cardiac function as stand-alone tests, we should not expect one test to provide us with comprehensive information on cerebral function in cirrhosis. It is our impression that for the time being the HE quick diagnosis remains fiction, and the specificity of any neurological/psychiatric test a myth.

  • Sara Montagnese, Ph.D.

  • Sami Schiff, Ph.D.

  • Piero Amodio, M.D.

  • Department of Medicine University of Padova Italy