One of the most debated issues in the field of hepatic encephalopathy (HE) is how to establish in a simple and practical manner its mildest manifestations. The article by Bajaj et al. in this issue of Hepatology is interesting, since it proposes the use of an App that can be downloaded to a smartphone and measures the time required to correctly identify the color of a series of symbols and printed words indicating a different color (e.g., the word “red” printed in blue requires that the individual press the blue button), thus overcoming a semantic-perceptual conflict (i.e., the Stroop task). The study compares several tests that are currently used for the diagnosis of minimal HE with the results of the App. The study opens new perspectives and sensitizes physicians and patients to the existence of mild brain dysfunction that is difficult to diagnose by physical examination.
Traditionally, the term minimal HE is used to identify patients with abnormalities in neuropsychological or neurophysiological tests with a normal neurological exam.[2-4] However, there are marked difficulties in assessing what is a normal neurological exam. Therefore, a new term has been proposed to overcome this limitation by combining minimal HE and grade I HE: covert HE. The term covert HE refers to brain dysfunction caused by liver insufficiency and/or portal-systemic shunting that does not cause temporal/spatial disorientation or asterixis. Interestingly, MHE (the mildest form of covert HE) has been shown to be relevant in patients with cirrhosis. MHE heralds an increased risk of overt HE and even death[6-8] and is associated with a reduced ability to perform complex and potentially harmful tasks, such as driving, an increased risk for falls, and a reduced quality of life. In addition, therapies for HE may improve driving and quality of life in patients with MHE.[12, 13]
Therefore, screening patients for the presence of MHE could be justified. However, this is not simple in clinical practice. A formal neuropsychological assessment requires both a neuropsychologist and time. The Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) has been used in the U.S. for the diagnosis of MHE, but it may not be sensitive enough because it is a battery implemented to detect cognitive decline in general. The PSE syndrome test, a paper-and-pencil test battery, that provides the Psychometric Hepatic Encephalopathy Score (PHES) has been validated in several countries, but has not been standardized in the U.S. Objective approaches based on the change of oscillatory brain properties evaluated by quantitative electroencephalogram (EEG) or functional neuroimaging are not yet widely used or available, although they are promising and have the great advantage of being independent of both patient cooperation and education level.[15-17]
The first cognitive manifestations of HE consist of impairments in the speed/accuracy of complex attention tasks, suggesting involvement of the circuitry between the anterior cerebral cortex and basal ganglia.[6, 18] In fact, the delay in reaction time in patients with HE does not depend initially on motor dysfunction, but rather on an impairment in response selection, which is revealed via psychometrical tasks requiring a great deal of sustained attention, inhibition, switching, and working memory, such as the Stroop task.[6, 20]
Bajaj et al. in this issue of Hepatology suggest a simple and insightful approach by downloading and using the Stroop task on a smartphone. This elegant study shows that cirrhosis patients performed this downloaded version of the Stroop task slower than controls, a finding that had already been demonstrated in patients with cirrhosis.
Unfortunately, the observation of delayed time of performance of the Stroop test, which Bajaj et al. proved to be extremely valid on a population basis, does not allow immediate conclusions in a single individual, because other factors such as age and education have an impact on its performance. In single individuals, deviation from expected age- and education-adjusted values is a preferable way to assess cognitive ability. An example in general medicine is bone density that is expressed in units of deviation from the expected values adjusted for age and gender, since absolute values may be less informative.
A limitation of the study is the high educational level and the rather limited age range in controls and patients. It is predictable that “normal” subjects will not be able to perform the task within the proposed cutoff when the test is applied to the general population, where the prevalence of less educated or older individuals is higher.
Additionally, differences in color sensitivity, expertise in smartphone use, linguistic/ethnic origin might have confounding effects and should be considered when the use of the test will be extended. Importantly, the nonspecific nature of any psychometric test should be emphasized, in as much as it establishes cognitive dysfunction but not its etiology. Mild cognitive impairment, usually an aging-related dysfunction that may progress to Alzheimer's disease or cerebrovascular impairment, has a prevalence of 30% in the general population over 65 years. Other causes of cognitive dysfunction are micronutrient deficiencies, obstructive apnea syndrome, drug/alcohol addiction, hyponatremia, hypercalcemia, hypercapnia, medications, etc.
Therefore, further work is required to optimize the test and properly validate its use, also controlling for the possible confounding effects of different screen sizes, illumination, and interference by other software running on the equipment, noise in the testing set, expertise in smartphone utilization, etc.
At any rate, subjects who perform the test within the indicated “normal” values, by virtue of having been evaluated in highly educated, motivated, and relatively young people, are very likely free of MHE; i.e., the test has an excellent negative predictive value. In contrast, a positive result may indicate MHE, another disease, or even be present in some otherwise normal individuals, particularly in the aged or low-educated ones.
In this technological era, the work of Bajaj et al. is likely to open new options to rule out neurocognitive impairment in patients with cirrhosis and maybe even to rule in the presence of such impairment via a simple test that can be applied in the office. Further studies will be necessary to confirm the value of an algorithm that includes this App in the process of diagnosing minimal or covert HE.
Piero Amodio, M.D.1
Juan Cordoba, M.D., Ph.D.2,3
1Department of Medicine-DIMED, University of Padova, Padova, Italy
2Liver Unit, Hospital Vall Hebron, Universitat Autònoma de Barcelona, Barcelona, Spain
3Professor of Medicine at Hospital Vall d'Hebron, Liver Unit, Universitat Autònoma de Barcelona, Barcelona, Spain