Fatty liver now, diabetes and heart attack later? The liver as a barometer of metabolic health

Abstract

In this issue of the Journal, two Asian studies explore the association of nonalcoholic fatty liver disease (NAFLD) with metabolic disorders. We will first review what is already known on this subject, then examine these studies in more detail with particular consideration of their implications for patient care. NAFLD is characterized by histologic changes resembling alcohol-related liver injury occurring in persons who do not drink excessive alcohol.1 The pathologic characteristics include hepatic steatosis (fat alone), nonalcoholic steatohepatitis (NASH) (fat with hepatic inflammation, liver cell injury, often with fibrosis), and cirrhosis. There is general agreement that the pathophysiology of NAFLD/NASH is more often linked with metabolic derangements, ultimately the result of over-nutrition than with alcohol excess.1 Insulin resistance is the common thread linking these metabolic conditions, which include visceral (central) obesity, overall obesity, type 2 diabetes mellitus (T2DM), and hyperlipidemia. Thus, virtually all individuals with NAFLD have evidence of insulin resistance.2 NAFLD, and NASH in particular, is now recognized as the hepatic manifestation of the metabolic syndrome. Most (87%) fulfill criteria for the metabolic syndrome at diagnosis.3 In turn, persons with metabolic syndrome were at a four to 11–fold risk of developing NAFLD over a period of 14 months' follow up.4 Of the different metabolic syndrome components, the tightest association is with visceral (central) obesity, which is present in over 90% of cases in Asian studies when ethnic-specific criteria are used for definition. While the nexus between NAFLD and insulin resistance is invariable, what is less clear is whether the steatosis is caused by peripheral insulin resistance, by the resultant increased uptake of fatty acids from extrahepatic sites,5 or whether steatosis is the primary abnormality and hepatic insulin resistance occurs secondary to this. These potential relationships have been reviewed elsewhere.6 However, what is becoming evident experimentally is that the more severe the steatosis, the more likely is the hepatic pathology to be steatohepatitis rather than simple steatosis.7 The relationship between steatosis severity and liver injury is supported by the observations of Hsaio et al. reported in this issue of the Journal.8 Their study population involved over 16 000 adults attending a Taiwan University hospital. Using predefined hepatic sonographic criteria, subjects were defined as having no (42%), mild (53%), or severe steatosis (5%). When compared with those without steatosis, subjects with severe steatosis were more likely to have increased alanine aminotransferase (ALT) and also a greater frequency of metabolic syndrome components; the odds ratios ranged from 1.5 for hypertension and impaired fasting blood glucose to 4 for obesity and T2DM. The authors concluded that the severe fatty liver group should be screened for metabolic syndrome components. Can we use hepatic ultrasound to identify patients at risk for metabolic disease? Hepatic ultrasound is a relatively low-cost and widely available tool and detects hepatic steatosis with reasonable sensitivity. However, it is very much operator-dependent and misclassification can occur in 10–30%.9 If hepatic ultrasound is to hold its own as a metabolic screening test, then it has to be compared with simpler clinical tools. In Hsaio's study, more than half (57%) of overweight individuals (body mass index of 23–24.9 kg/m2) had NAFLD (table 3 in their study). Therefore, it is reasonable to ask whether the ultrasound is necessary or if a simple measure such as body mass index or more accurately, the waist circumference measurement (an unfortunate omission from the Hsaio study) would be sufficient grounds to order a metabolic profile. In light of these and earlier findings, are we justified in warning our patients that NAFLD is a prediabetic condition and in exhorting them to adopt lifestyle changes? Limited direct evidence is available to answer this question. In one study, three of 17 (18%) patients with NAFLD developed T2DM over a period of 3.8–8 years.10 However, population-based studies showing an association between elevated liver enzymes (a surrogate marker for NAFLD) and a future potential for development of the metabolic syndrome lend support to this contention.11,12 The Insulin Resistance Atherosclerosis Study showed that persons with levels of ALT and alkaline phosphatase (ALK) in the upper quartile were twice as likely as to develop metabolic syndrome as compared with those with ALT and ALK levels in the lower quartile.11 In the second important study in this issue of the Journal, Fan et al. identified a cohort of 358 individuals with hepatic ultrasound defined–fatty liver by reviewing the annual medical records of steel company employees.13 Persons with significant alcohol intake or with viral hepatitis B or C were excluded. A total of 788 age, sex, and occupation-matched coworkers served as controls. Both groups were followed for 4–7 years (median 6 years). Both at baseline as well as on the follow-up examination, metabolic syndrome components were present at a greater frequency among those with fatty liver than among controls. The prevalence of the metabolic syndrome was not formally assessed by predefined criteria at entry, but the authors found 26% of the fatty liver group and 12% of controls had three or more metabolic abnormalities at baseline; the corresponding figures for the follow-up study were 68% and 22%, respectively. Subjects with NAFLD were also more likely to develop new metabolic disorders than controls; a new diagnosis of T2DM was made in 20% (odds ratio [OR] 4.6), hypertriglyceridemia in 39% (OR 3.3), obesity in 48% (OR 3.4), and hypertension in 70% (OR 2.9). Further, in an attempt to establish any separate effect of fatty liver on the incidence of metabolic disorders other than those common associations with (overall) obesity, a separate analysis was performed according to the characteristics at study entry. The incidence of hypertension, hypertriglyceridemia, hypercholesterolemia, impaired fasting glucose, and T2DM were significantly higher in the NAFLD group than with controls with neither fatty liver nor obesity. The authors concluded that fatty liver is more predictive of the development of metabolic disorders due to insulin resistance than is obesity itself. While broadly accepting the findings of the present study, a few caveats apply with respect to the subsequent risk of developing T2DM. It is highly probable that there were some patients with unrecognized T2DM in the fatty liver group. Patients with isolated postprandial hyperglycemia (up to 40% in one study) would have been missed without an oral glucose tolerance test.14 Further, the study does not take into account a number of known factors which can affect insulin sensitivity such as age, ethnicity, family history diet, physical activity, smoking, and concurrent medications. Therefore, the cited risk of 20% for T2DM should be regarded as a rough estimate. Still, it is the best estimate to date and should help doctors in counseling their patients about their likelihood of future metabolic problems. Finally, what are the therapeutic implications of these two studies? The metabolic syndrome brings with it a two to three–fold risk of future cardiovascular disease.15 At the present time, it remains unclear whether the cardiovascular risk of individuals with NAFLD can be solely attributable to traditional risk factors alone or whether there is an additional contribution from the inflammatory (e.g. oxidative stress) or hormonal disturbances (hyperinsulinemia, hypoadiponectinemia) characteristic of NASH. However, there are recent data suggesting that persons with NAFLD have a higher frequency of subclinical atherosclerosis markers (carotid artery intima-media thickness),16 show evidence of impaired endothelial reactivity,17 and also exhibit features of early left ventricular diastolic dysfunction.18 Further, cardiovascular disease was the leading cause of death in at one least population-based19 and another hospital-based series of patients with NAFLD20 and also in several other metabolic follow up studies (reviewed by Targer).21 Therefore, it is reasonable to initiate discussions on lifestyle changes in patients with NAFLD, just as one would with an younger patient with hypertension or dyslipidemia, or an obese patient with a family history of T2DM. As shown in diabetes prevention studies, lifestyle intervention can help reduce the risk of diabetes in high-risk groups.22 It is time for hepatologists and patients to get serious about measures to improve metabolic health. There is already emerging evidence that just as fatty liver can harbor the onset of metabolic syndrome and its complications,23 resolution of steatosis is associated with a decrease in the prevalence of the metabolic syndrome.4 Perhaps the liver, and more particularly its fat content, will become the modern barometer of metabolic health.