Attention: Minimal hepatic encephalopathy and road accidents


  • Potential conflict of interest: Nothing to report.

  • See Article on Page 1164

In this issue of HEPATOLOGY, Bajaj et al.1 provide a comparison between the estimated costs of driving accidents in patients with cirrhosis and minimal hepatic encephalopathy (MHE), and the costs of searching for/managing MHE with different screening and treatment strategies. The analysis suggests that screening for MHE with a test of attention and inhibition, i.e., the Inhibitory Control Test (ICT), and treating diagnosed patients with lactulose may be the most cost-effective strategy in this clinical setting. The approach utilized in the article, which goes beyond patients and their immediate well-being, to include an outlook on disease and management consequences on society, has merit, especially in an era when health systems are under considerable pressure to remain or become cost-effective.


HE, hepatic encephalopathy; ICT, Inhibitory Control Test; MHE, minimal hepatic encephalopathy.

Liver cirrhosis, far from being an isolated disorder of the liver, has well-known consequences on brain/mental functioning. In their overt expression of delirium or coma, liver-related mental alterations have been known since the ancient Greek and Roman times. Hippocrates recognized a “symptomatic delirium” related to “jaundice and suppression of natural periodical evacuation.”2 In 1765, Morgagni described coma in cirrhosis,3 which was subsequently termed portal-systemic encephalopathy,4 and later hepatic encephalopathy (HE).5 In the 1950s, Parsons-Smith et al.6 demonstrated that approximately 40% of in-patients with cirrhosis exhibit electroencephalographic abnormalities despite not showing obvious mental alterations on clinical examination. Along the same lines, it was subsequently shown that these patients also have impaired performance on neuropsychological tests,7 the prevalence of which depends on the explored cognitive domains,8, 9 and the reduction in functional hepatic mass and in liver perfusion.7, 10 These forms of cognitive impairment due to liver failure and portal-systemic shunting, in the absence of clinically apparent neurological/psychiatric dysfunction, are referred to as MHE.

Brain dysfunction adversely influences the well-being of patients with cirrhosis, and their performance. However, HE, and even more so MHE, are often neglected by hepatologists in their routine practice.11 Fortunately, the interest in these syndromes and their effect on activities of daily living, especially driving, has grown over recent years.12, 13 The attention devoted to the relationship between MHE and driving is more than justified, because motor vehicle accidents are associated with considerable morbidity and mortality, as well as direct and indirect economic and social costs (Table 1).

Table 1. Number of Road Accidents, Killed and Injured Persons in Europe and North America in 2008
CountryPopulation (x106)Number of Accidents, Killed and Injured Persons
  1. From Statistics of Road Traffic Accidents in Europe and North America, Vol. 52, UN, New York and Geneva, 2011.


Driving errors account for 71%-98% of motor vehicle accidents,14, 15 thus the assessment of driving ability is crucial. In most countries, restraints are applied to alcohol consumption and speed, as these are recognized risk factors for driving errors. In contrast, legal systems have devoted limited attention to the cognitive and behavioral elements related to driving, with the exception of full-blown mental dysfunction. MHE, which is fluctuating and not easily or homogeneously diagnosed, hardly falls under this category, and is not formally regulated in most countries, at least to our knowledge.

Patients with cirrhosis and HE are generally optimistic about their driving abilities.16, 17 In a recent study by Kircheis et al.,17 100% of patients with mild overt HE and 96% of those with MHE were convinced they were good or very good drivers, compared with 92% of control subjects. In contrast to their convictions, the actual driving ability of patients with MHE is reduced based on any of the assessment criteria adopted so far, which include: (1) neuropsychological testing of cognitive domains that are thought to be implicated in driving skills,18 (2) simulated driving on virtual navigators,12 and (3) on-the-road driving.17, 19 However, whereas patients with MHE may have reduced driving ability taken as a group, the predictive value of the various techniques on actual driving ability seems limited on a single-patient basis.17 This clearly impinges on the appropriateness of providing direct advice, or applying restrictions to a daily life activity that has such profound impact on a patient's job, their earning capacity, their social life, etc. The interesting case of a taxi driver who worked apparently without problems while affected by minimal HE was reported by Srivastava et al.20 in one of the first published studies on HE and driving. This is not surprising, because the behavioral effects of brain damage are due to both the severity of brain damage and the so-called cognitive reserve. The latter describes the resilience of the mind to objective, anatomical/functional brain damage. This phenomenon, which was recently proven to occur also in patients with cirrhosis and HE,21 is probably related to the life-long changes in brain connectivity triggered by chronic training in different activities of daily life.

The identification of subjects with MHE is based on tools measuring cognitive/brain dysfunction. This is not a simple procedure in clinical practice, for several reasons. The specificity of impaired cognitive performance for the diagnosis of MHE is rather low, because a number of medical, social, educational, and cultural issues interfere with cognition. Patients may be impaired in relation to their own premorbid standard or potential, even if their performance falls within the range of the pertinent reference population. In these individuals, the response to ammonia-lowering treatment may disclose the existence of MHE. Finally, the complexity of predicting driving ability on a single-patient basis suggests that: (1) ad-hoc neuropsychological tests designed to assess driving skills may be more useful than tests designed to diagnose MHE, and (2) where the same tests are applied, the cutoffs that are useful to predict driving ability may be different from those which diagnose MHE.

In the present study, Bajaj et al. go beyond these issues, demonstrating that MHE, regardless of a number of details that require further definition, is worth treating in order to prevent driving accidents, and that the costs of screening and treating MHE are reasonable in relation to the savings derived from the reduction in accident rates. Obviously, the cost-effectiveness analysis of a set of diagnostic and therapeutic interventions in patients with suspected MHE is based on a number of assumptions/simplifications, which represent the foundations of the pharmacoeconomic model.22 Several of these assumptions are reasonable, or even proven; a few of them, however, may by less solid. If these were modified, the results might change quite considerably, and beyond the limits tested in the study by the sensitivity analysis. For example, the diagnosis of MHE depends on the techniques adopted, and the procedures which should be used to exclude concomitant or alternative causes of neuropsychological dysfunction are debated. Secondly, it is uncertain whether patients can be kept on the proposed therapeutic regimes long term, without significant side effects. Similarly, there is a possibility that patients who recover from MHE may withdraw from treatment, or that treatment may positively affect driving skills in less than 78% of cases, as postulated in the study. Indeed, even if MHE and reduced driving skills are related, they cannot be considered one and the same thing, since the assumption that ammonia-lowering strategies may affect driving to the same extent that they affect psychometric performance is not sufficiently proven. The combined effect of variations in some of these base-case parameters, or in the structure of the decision tree, might lead to partially different conclusions to the study.22 These limitations aside, which pertain to most of the pharmacoeconomic literature, the information provided by Bajaj et al. is welcome, as it might: (1) stimulate further, formal studies on the real-life effect of MHE screening/treatment on accident rates, and (2) attract the attention of the pertinent regulatory bodies on the relationship between MHE and driving, which has such profound implications for single patients, and for society at large.