Clues from the carotids: An appraisal of cardiovascular disease risk in non‐alcoholic fatty liver disease

In this issue of the Journal of Gastroenterology and Hepatology, Wang et al. report that patients with non-alcoholic fatty liver disease (NAFLD) and raised serum alanine aminotransferase (ALT) have an increased risk of carotid atherosclerosis.1 Their study involved 170 subjects recruited from the Buddhist Tzu Chi General hospital in Taiwan. The diagnosis of NAFLD was made by liver ultrasound and exclusion of other causes of hepatic steatosis. The NAFLD group was further subdivided based on the levels of ALT into those with normal (n = 72) and those with raised ALT levels (n = 29). All their subjects were screened for the presence of metabolic disorders (obesity, hypertension, dyslipidemia, type 2 diabetes). In addition, they underwent ultrasound evaluation of both common carotid arteries for evidence of carotid atherosclerosis. Measurements of carotid intima-media thickness (IMT) were obtained and screening for carotid plaques was carried out; participants with carotid IMT readings of greater than 1 mm were designated as having carotid atherosclerosis. Persons with normal liver ultrasound findings served as controls (n = 69). As expected, their subjects with NAFLD were more likely to fulfil criteria for metabolic syndrome (MetS) than controls (55% with raised ALT, 22% with normal ALT and 4% of controls). The authors go on to show that subjects with NAFLD, particularly those with raised ALT, have a significantly higher prevalence of carotid atherosclerosis than controls (41% vs 18%, respectively). Further, within the NAFLD group, an incremental increase in ALT was an independent predictor of carotid atherosclerosis; for every 10 IU increase in ALT, the odds ratio for developing these carotid changes was 1.44. The authors conclude that raised ALT can serve as a surrogate marker for cardiovascular disease (CVD) in persons with fatty liver. Are these findings novel and, if confirmed, what are the implications? Before addressing these questions, it is necessary to review our current understanding of the place of NAFLD among metabolic disorders and its association with CVD. Although association with MetS has been alluded to earlier, definitive studies establishing NAFLD as the hepatic manifestation of MetS have been more recent.2 In addition, patients with NAFLD are at risk of developing other metabolic disorders with follow up. For instance, there is a three- to fourfold likelihood of developing type 2 diabetes and hypertension within a decade of diagnosis.3, 4 While these studies have firmly placed NAFLD amidst the constellation of disorders that constitute MetS, the real test of inclusion is whether NAFLD is also associated with a greater risk of CVD and type 2 diabetes. After all, it has been known for many years that persons with MetS have a two- to threefold risk of developing CVD and a fivefold risk of type 2 diabetes.5 Of these two disorders, the association with type 2 diabetes (more specifically, insulin resistance) is well recognized; type 2 diabetes or impaired glucose tolerance is present in 30% to 40% of patients with fatty liver.2 With respect to CVD risk, the most compelling data come from natural history studies of NAFLD.6, 7 While these studies have been instructive in highlighting the potential for liver-related complications among certain subgroups, they also uniformly document CVD as one of the leading causes of death. In recent years, the NAFLD-cardiovascular link has undergone further scrutiny. Because it is difficult to conduct large, prospective and longitudinal studies on cardiovascular outcomes, several research groups have focused instead on surrogate markers for CVD. These include studies of intramyocardial resistance,8 endothelial flow-mediated vasodilation,9 basilar artery pulsatility index10 and carotid IMT measurement.11-19 Of these, the latter isperhaps the most widely studied surrogate marker for CVD in fatty liver cohorts. In addition to measuring IMT, carotid plaques are also visualized. Carotid IMT is not synonymous with carotid atherosclerosis and should be regarded as a biomarker of ‘vascular age’. However, it is often used interchangeably (as in this study by Wang et al.) to connote cardiac atherosclerosis. Its value as a biomarker of CVD is supported by a recent meta-analysis; for every 0.1 mm common carotid IMT difference, the CVD risk was increased by 1.15.20 In terms of absolute risk, carotid IMT readings greater than 1.06 mm were associated with a 2% risk of myocardial infarction and stroke.21 Like many biological variables (serum cholesterol, blood glucose, blood pressure), there is likely to be a continuum of risk rather than an abrupt ‘cut-off’. How do we interpret the carotid IMT findings in the current study? In keeping with other recent publications (Table 1), there were significant differences in carotid IMT between patients with NAFLD and controls. The authors have wisely chosen a relatively high cut-off of 1 mm to represent significant carotid disease. Many, but not all, studies listed in the table have identified NAFLD as an independent predictor of increased carotid IMT and, by inference, CVD risk. To reiterate this important point, this finding implies that the CVD risk for patients with NAFLD is greater than that which could be attributed to the individual MetS components alone. Unfortunately, a limitation of the study by Wang et al. is that the cohort was probably too small to conduct a multivariate analysis of the entire cohort to answer this question. However, as mentioned above, they confined their analysis to the NAFLD group; age, male gender and ALT emerged as significant predictors of carotid IMT. The relevance of age and gender to the process of atherosclerosis is well documented and need not be considered any further. However, incremental ALT changes and the risk of CVD bear comment. Subanalyses of large population studies have shown that elevated liver enzymes (gamma-glutamyl transpeptidase and also ALT) correlate with future risk of type 2 diabetes, MetS and also CVD.22 More recently, even persons with ALT levels within the ‘normal’ range were found to be at risk of these complications.23 Thus, the findings of Wang et al. would be consistent with these observations. However, their proposal that ALT could serve as a cardiovascular disease biomarker poses several practical problems. Besides definition of the ‘normal’ range and the need to exclude other liver diseases, this would also exclude patients with NAFLD and a normal ALT. The latter group are also at risk of developing CVD; in their own study, 82% of their subjects with normal ALT had carotid plaques, 22% had features of MetS, and 17% had carotid IMT values of greater than 1 mm. If the authors mean to imply that raised ALT is indicative of increased liver histological severity (in the absence of liver biopsy in their study), they find some support from one study that found a correlation between severity of liver disease and carotid IMT.13 The problem is that the entire spectrum of NAFLD has been documented in persons with normal ALT.2 One cannot ‘best guess’ the underlying liver histology based on ALT readings. Moreover, other studies have not found a relationship of liver histology to NAFLD.24 The jury is still out on this! While we have alluded to the NAFLD-CVD nexus right through this editorial, do we have conclusive data to prove or disprove whether there is an ‘additional’ CVD risk attributable to the fatty liver state itself? Potentially, the presence of fatty liver could enhance the atherogenic risk by creating and perpetuating the inflammatory milieu through the effects of reactive oxygen species, cytokine alterations, adhesion and procoagulant molecule expression and abnormal lipoprotein metabolism.11 Unfortunately, at the present time, the jury is still out with respect to this important question. One of the problems relates to the number of confounding factors that must be controlled for in multivariate analysis; sophisticated measures of visceral adiposity, insulin sensitivity and also non-traditional risk factors, such as coronary calcium, are but a few of these variables which need to be taken into consideration. These considerations do not detract from the need for careful metabolic assessment of persons with fatty liver. Prospective studies can only determine the magnitude of CVD risk in patients with NAFLD. However, the link with well-described disease constellations (metabolic syndrome), biomarkers (carotid IMT, inflammatory molecules etc.) and retrospective data showing CVD among the principal causes of death in natural history studies, suggest that it is time to regard NAFLD as an additional coronary artery disease equivalent. Therefore, it is becoming increasingly clear that in clinical practice, some form of cardiovascular risk assessment will become mandatory, using one of the validated methods such as the Framingham risk score. Individuals at high risk could be targeted for intensive intervention through lifestyle modification as well as with medications (e.g. lipid-lowering, anti-inflammatory drugs). Carotid IMT is not yet part of routine CVD risk testing, but studies such as those by Wang et al. draw attention to latent atherosclerotic disease in fatty liver-afflicted individuals and thereby provide opportunities for early intervention.