| |
Fatty Liver Predicted Heart Disease
|
| |
| |
"Endothelial dysfunction and cardiovascular risk profile in nonalcoholic fatty liver disease"
Hepatology
August 2005
".... Our study provides evidence that an endothelial dysfunction is present in NAFLD (nonalcoholic fatty liver disease), when compared with controls matched for age and sex and with a similar prevalence of features of metabolic syndrome...
..... The pathogenesis of NAFLD is probably multifactorial, but insulin resistance, either genetically determined or obesity related, is a cornerstone...
..... hyperlipidemia might be a crucial link between hepatic steatosis, insulin resistance, the insulin resistance syndrome, and endothelial dysfunction....
..... triglycerides....were higher and more frequently in excess of 150 mg/dL. Fasting glucose was not different, but 120-minute glucose during oral glucose tolerance test was moderately increased, and a higher number of cases fitted the criteria of impaired glucose tolerance (IGT) or diabetes.....
..... Obese subjects and NASH patients frequently have circulating markers of inflammation, and atherosclerosis is systematically associated with inflammation. Oxidative stress also has been implicated....
.... A critical issue in this type of study is the selection of an appropriate control group. We enrolled consecutive NAFLD cases, after exclusion of patients with overt diabetes or BMI 35 kg/m2. In the selection of control subjects, we also excluded cases with steatosis at ultrasounds....
.... Epidemiological data suggest that cardiovascular mortality is increased in patients with a diagnosis of fatty liver of nonalcoholic or unspecified cause.....recent study showed that NAFLD cases have more carotid atherosclerosis than controls, with enlarged mean intima-media thickness and higher plaque prevalence.
......By multivariate analysis, the presence of NAFLD was an independent predictor of an increased intima-media thickness. A modest increase in glucose levels appear to be associated with carotid atherosclerosis, and increased free fatty acid levels have been associated with enlarged intima-media thickness.
.....Both hyperglycemia and fatty acids are on average higher than normal in NAFLD, possibly leading to accelerated atherosclerosis....."
AUTHORS: Nicola Villanova 1, Simona Moscatiello 1, Stefano Ramilli 2, Elisabetta Bugianesi 3, Donatella Magalotti 2, Ester Vanni 3, Marco Zoli 2, Giulio Marchesini 1 *
1Unit of Metabolic Diseases, Alma Mater Studiorum University of Bologna, Bologna, Italy
2Division of Internal Medicine, Alma Mater Studiorum University of Bologna, Bologna, Italy
3Department of Gastroenterology, University of Turin, Italy
INTRODUCTION
Nonalcoholic fatty liver disease (NAFLD) is increasingly recognized as a public health problem in the medical community.[1] The number of patients at risk of advanced liver disease through nonalcoholic steatohepatitis (NASH) and cryptogenic cirrhosis is increasing because of the epidemics of obesity and diabetes related to western lifestyle,[2] associated with fatty liver[3][4] and raised alanine aminotransferase levels.[5-7]
The pathogenesis of NAFLD is probably multifactorial, but insulin resistance, either genetically determined or obesity related, is a cornerstone.[8] A lot of experimental and clinical data point to NAFLD as the hepatic expression of the metabolic syndrome (MS).[9] Because of the underlying metabolic disorder, NAFLD patients are expected to have a higher risk of vascular and coronary heart disease as well. Hyperinsulinemia and insulin resistance are risk factors for atherosclerosis,[10] and the cluster of conditions forming MS has cardiovascular disease as the common final event.[11-13] Epidemiological studies in fatty liver have suggested a higher-than-expected cardiovascular mortality,[14] but very few data are available on the cardiovascular risk profile of these patients.
The arterial endothelium is also a target for the atherosclerotic process. Atherosclerosis is associated with endothelial dysfunction in the very early stages of the disease process.[15-17] Aging and exposure to risk factors such as hypercholesterolemia, hypertension, and smoking are also associated with endothelial dysfunction.[18] Several studies have shown that MS is significantly associated with endothelial dysfunction,[18][19] impairing the vascular response to physiologic and pharmacologic stimuli.
The percent changes in the caliber of large arteries induced by the ischemia inflation and deflation of a sphygmomanometer cuff around the limb distal to the scanned part of the artery may be easily and noninvasively quantified.[20] The response is mediated by nitric oxide released by endothelial epithelium in response to the shear stress generated by artery occlusion, and it is considered a measure of endothelial function.[21] It correlates with more invasive methods of endothelial function.[22]
The aim of this study was to evaluate the flow-mediated vasodilation (FMV) of the brachial artery in NAFLD. The cardiovascular risk profile of the cohort of NAFLD cases was also measured and compared with an age- and sex-matched control group.
Abstract
Nonalcoholic fatty liver disease (NAFLD) is consistently associated with features of the metabolic syndrome, a condition carrying a high risk of cardiovascular events.
We measured the vasodilatory response of the brachial artery in response to ischemia (a test of endothelial function) (FMV) as well as cardiovascular risk profile in 52 NAFLD cases and 28 age- and sex-matched controls. The 10-year risk of coronary events was calculated according to the Framingham equation and the scores derived from the PROCAM study and NCEP-ATPIII proposals.
FMV was 6.33% ± 5.93% in NAFLD versus 12.22% ± 5.05% in controls (P < .0001), and higher in pure fatty liver (9.93%) compared with nonalcoholic steatohepatitis (4.94%) (P = .010). No differences were observed in flow-independent vasodilation (response to sublingual nitroglycerin). Percent FMV was negatively associated with insulin resistance (homeostasis model assessment) in the whole population (r = -0.243; P = .030).
In logistic regression analysis, NAFLD was associated with a percent FMV in the lower tertile (OR, 6.7; 95% CI, 1.26-36.1), after adjustment for age, sex, body mass index, and insulin resistance.
Among NAFLD patients, low FMV was associated with nonalcoholic steatohepatitis (adjusted OR, 6.8; 95% CI, 1.2-40.2). The 10-year probability of cardiovascular events was moderately increased in NAFLD, and particularly in nonalcoholic steatohepatitis.
In conclusion, our study provides evidence of endothelial dysfunction and increased risk of cardiovascular events in NAFLD. The risk of advanced liver disease is well recognized in NAFLD patients, but the large majority of cases might experience cardiovascular disease in the long term, indirectly limiting the burden of liver failure.
AUTHOR DISCUSSION
Our study provides evidence that an endothelial dysfunction is present in NAFLD, when compared with controls matched for age and sex and with a similar prevalence of features of MS. Only obesity was systematically more prevalent in our patients, associated with higher insulin and insulin resistance. Data suggest that insulin resistance might be involved in decreased percent FMV, as well as in the increased cardiovascular risk profile.
A critical issue in this type of study is the selection of an appropriate control group. We enrolled consecutive NAFLD cases, after exclusion of patients with overt diabetes or BMI 35 kg/m2. In the selection of control subjects, we also excluded cases with steatosis at ultrasounds. Accordingly, we failed to have a complete match between patients and controls for obesity; however, differences in percent FMV were maintained when obese subjects were removed from analysis (controls, 12.2% ± 4.7%; NAFLD, 6.8% ± 4.7%; P = .0003). Obesity might be responsible for hyperinsulinemia and insulin resistance, but genetically determined defects in insulin action might be present in NAFLD,[31] as supported by data in selected cohorts of non-obese, non-diabetic subjects.[32]
Endothelial dysfunction and increased cardiovascular risk were expected in NAFLD, a condition strictly associated with MS. Epidemiological data suggest that cardiovascular mortality is increased in patients with a diagnosis of fatty liver of nonalcoholic or unspecified cause[14]; similarly, a recent study showed that NAFLD cases have more carotid atherosclerosis than controls, with enlarged mean intima-media thickness and higher plaque prevalence. By multivariate analysis, the presence of NAFLD was an independent predictor of an increased intima-media thickness.[33] A modest increase in glucose levels appear to be associated with carotid atherosclerosis,[34] and increased free fatty acid levels have been associated with enlarged intima-media thickness.[35] Both hyperglycemia and fatty acids are on average higher than normal in NAFLD,[36] possibly leading to accelerated atherosclerosis.
Percent FMV has been used extensively to assess endothelial function in several disease states and is well validated.[37] Methodological issues are critical, and guidelines for a correct measurement have been issued.[38][39] The variability of measurements carried out at intervals by the same operator is satisfactory, whereas the inter-observer variability is exceedingly high.[40] Accordingly, in the current study all tests were performed by the same operators in two different units, after careful training. Measured mean values were on average slightly higher than previously reported in several methodological studies,[40][41] but the young age of our population and the low prevalence of complicating diseases are a likely explanation.
We ascribed the defect in percent FMV to the presence of early atherosclerosis. Ideally, a control group of patients with non-cirrhosis liver disease of different causes is needed to rule out the possibility that liver disease itself may cause endothelial dysfunction. Atherosclerosis is rare in patients with advanced liver disease,[42] and the low risk was ascribed to the biochemical and circulatory changes induced by the failing liver. Also in the presence of diabetes, the risk of death of diabetes-related events is low in cirrhosis.[43] The hypothesis of a primary liver disease-related defect is therefore very unlikely but remains to be tested. We also calculated the risk of cardiovascular events at 10 years, using the scores derived from 3 population studies.[25][28][29] All these scores are extensively used to categorize patients with metabolic disease, and the Framingham score was previously shown to correlate with biochemical markers of endothelial dysfunction.[44] Their use in NAFLD patients may be biased by the young age of the population and by the low number of events observed in these classes, and should be considered with caution.
In our analysis, the risk of NAFLD patients was moderately increased, and percent FMV was associated with risk. The natural history of NAFLD is so far largely unknown. Although the risk of advanced liver disease is likely in a proportion of patients with diabetes and obesity,[45][46] most of these cases probably will experience coronary and vascular disease, indirectly limiting the burden of liver failure. In a prospective cohort study, no cardiac events were registered in a limited number of NASH-associated cirrhosis,[47] but these patients do not represent the whole spectrum of NAFLD. Our data in subjects with less advanced disease and an epidemiological study[14] are in keeping with the hypothesis that NAFLD not progressed to cirrhosis may predispose to cardiac events, but only long-term follow-up studies will solve the issue.
Reduced percent FMV was associated with the number of features of MS, as well as with NAFLD and NASH after adjustment for age, sex, BMI, and the degree of insulin resistance. Although FMV figures largely overlap between control subjects and NAFLD, the presence and the severity of liver disease were systematically associated with more altered endothelial function. Also, high triglycerides entered the regression, suggesting that hyperlipidemia might be a crucial link between hepatic steatosis, insulin resistance, the insulin resistance syndrome, and endothelial dysfunction. Obese subjects and NASH patients frequently have circulating markers of inflammation,[48][49] and atherosclerosis is systematically associated with inflammation. Oxidative stress also has been implicated. Further studies will assess the role of circulating molecules causing endothelial dysfunction on both progressive liver disease and in insulin resistance-dependent systemic atherosclerosis.
RESULTS
Clinical and Laboratory Data.
Most NAFLD cases were males of a relatively young age. In spite of the careful matching, they differed from controls in the degree of obesity, particularly in the presence of visceral adiposity. Biochemical parameters were not systematically different, with the notable exception of triglycerides, which were higher and more frequently in excess of 150 mg/dL. Fasting glucose was not different, but 120-minute glucose during oral glucose tolerance test was moderately increased, and a higher number of cases fitted the criteria of impaired glucose tolerance (IGT) or diabetes in response to load (NAFLD: IGT, 30%; Diabetes, 4%; Controls: IGT, 18%). A positive family history of diabetes was more prevalent in NAFLD patients. Fasting insulin levels were markedly increased, as was the HOMA-R index of insulin resistance. A larger percentage of patients had HOMA-R values exceeding 3.0, a cutoff corresponding to the upper quartile of an unselected control population.[30]
The prevalence of MS was more than doubled in this NAFLD cohort with a low prevalence of glucose alterations and obesity.
At histology, the hepatic fat content averaged 44%, with two thirds of cases having more than 33% fat accumulation, and more than having fibrosis stage 1 or higher. Thirty-nine cases were classified as NASH, on the basis of the presence of fibrosis (grade 1 or more) or necroinflammation (grade 2 or more), whereas 13 were classified as pure fatty liver.
Brachial Artery Measurements.
At baseline, no differences were observed in the diameter of the brachial artery (NAFLD, 3.87 ± 0.51 mm vs. 3.62 ± 0.69 in controls; P = .069). In response to ischemia, the mean diameter increased to 4.31 mm in NAFLD and to 4.25 in controls (P = 0.662), corresponding to an average percent FMV of 6.33% ± SD 5.93% in NAFLD and 12.22% ± 5.05% in controls (P < .0001, Fig. 1). The percent FMV was also remarkably higher in fatty liver (9.93%) compared with NASH cases (4.94%) (P = .010).
In response to nitroglycerin, the diameter of the brachial artery increased both in NAFLD (4.49 ± 0.64 mm) and in controls (4.26 ± 0.75; P vs. NAFLD = .155), and the percent FMV was not different.
In the whole group, percent FMV was remarkably larger in subjects without MS (9.43% ± 6.36%), compared with subjects with MS (5.66 ± 5.17; P = .016). Similar differences were found when percent changes in brachial artery diameter were compared in subjects with and without enlarged waist (P = .028), high glucose (P = .012), and high triglycerides (P = .0015), not in relation to high blood pressure (P = .350) and low HDL cholesterol (P = .063). Percent FMV was negatively associated with HOMA-R in the whole population (r = -0.248; P = .027).
Association of Percent FMV With Clinical, Anthropometric, and Histological Parameters.
Individual values of percent FMV were grouped in tertiles, and values in the lower tertile (<5% vasodilation) were tested for association with clinical, anthropometric, and biochemical variables. Low percent FMV was significantly associated with BMI, with several components of MS, and with insulin resistance.
In multivariate analysis, the risk of a low percent FMV was systematically associated with the number of features of MS, and the presence of a single feature nearly doubled the risk. Moreover, a percent FMV in the lower tertile carried a 7-fold higher risk of NAFLD in the whole population, and a 6-fold higher risk of NASH within NAFLD cases, after adjustment for age, sex, BMI, and the degree of insulin resistance. In the last two models, high triglycerides (>150 mg/dL), but no other parameter of MS, entered the regression.
Cardiovascular Risk Profile.
Because of a skewed distribution of data and a risk score below 1% in several cases, the scores were categorized using predefined cutoffs as very low (<1%), low (1%-3%), intermediate (4%-6%), high (7%-10%), and very high (>10%). The 10-year probability of a cardiovascular event was moderately increased in NAFLD compared with controls, and the differences were significant according to both the Framingham and the PROCAM equations (Table 3). No differences were observed when the risks were compared between subjects with pure fatty liver and NASH cases (Framingham, P = 0.933; PROCAM, P = 0.623; ATPIII, P = 0.483; chi-squared test).
The risk score category was not systematically associated with percent FMV in our total population, and separately in controls and NAFLD cases (not reported in details, Kruskal-Wallis test). On the contrary, the HOMA-R index of insulin resistance was positively correlated with cardiovascular risk (Framingham, r = 0.305, P = .007; PROCAM, r = 0.261, P = .020; ATPIII, r = 0.347, P = .002, Spearman rank correlation).
Patients and Methods
Patients.
The current study was carried out in 52 consecutive patients (42 males and 10 females) with NAFLD attending the metabolic unit of the University Hospital of Bologna and in 28 control subjects without metabolic diseases (20 males and 8 females).
NAFLD cases were identified on the basis of chronically raised alanine aminotransferase levels (>1.5× upper normal values for 6 months or more) and a bright liver at ultrasound scan, in the absence of any cause of liver disease. To exclude other causes of liver disease, all subjects underwent a complete laboratory investigation for viral hepatitis (hepatitis B and C viral markers), autoimmune hepatitis, or primary biliary cirrhosis (non-organ-specific antibodies), celiac disease (anti-gliadin antibodies), genetic disease (1 antitripsin, ceruloplasmin). In addition, they had no history of current or past excessive alcohol drinking as defined by an average daily consumption of more than 20 g alcohol. Subjects with overt diabetes or body mass index (BMI) above 35 kg/m2 also were excluded. At the time of study, NAFLD patients had no clinical, biochemical, or histological evidence of cirrhosis.
Both controls and NAFLD subjects had complete clinical examination, anthropometric measurements, laboratory tests (Table 1), and a routine liver ultrasonographic scan. Laboratory investigations included an oral glucose tolerance test. Family history of diabetes (first- and second-degree relatives) and of cardiovascular disease (first-degree relatives: males younger than 55 years, females younger than 60 years) was also registered, as were current pharmacological treatments (lipid-lowering agents, antihypertensive drugs, oral hypoglycemic agents) and cigarette smoking.
At ultrasounds, 4 parameters were tested: (1) diffuse hyperechoic echotexture (bright liver); (2) increased liver echotexture compared with the kidneys; (3) vascular blurring; (4) deep attenuation.[23] All NAFLD cases had a percutaneous liver biopsy, performed on day-hospital admission. Liver biopsy was scored according to the criteria proposed by Brunt et al.[24]
The control subjects had no evidence of fatty liver at ultrasounds. They were matched with NAFLD cases for age and sex. They were selected in a high BMI range, after exclusion of subjects with fatty liver at ultrasounds.
The 5 components of MS were available in all subjects; the presence of 3 or more of the following criteria characterized subjects with MS[25]: (a) fasting glucose 110 mg/dL or treated for diabetes; (b) central obesity (waist circumference > 102 cm [males] and > 88 cm [females]); (c) arterial pressure 130/85 mm Hg or pharmacologically treated; (d) triglyceride levels > 150 mg/dL or current use of fibrates; (e) HDL-cholesterol < 40 mg/dL (males) and < 50 mg/dL (females).
The study was part of a complete assessment of metabolic abnormalities in NAFLD patients entering a treatment protocol. All subjects gave their informed consent to metabolic assessment, liver biopsy, and brachial artery study. The study was approved by the Senior Staff Committee of the University hospital. This board regulates noninterventional studies and is comparable to an Institutional Review Board.
Methods.
Body weight was measured in light clothing and without shoes to the nearest 0.5 kg. Height was measured at the nearest 0.5 cm. BMI was calculated as weight (kg) divided by height (m2). Waist circumference was measured at the nearest 0.5 cm at the shortest point between the lower rib margin and the iliac crest, whereas hip circumference was similarly obtained at the widest point between hip and buttocks.
Blood pressure measurements were obtained according to Guidelines of the International Society of Hypertension.[26] Three blood pressure readings were obtained at 1-minute intervals, and the second and third systolic and diastolic pressure readings were averaged and used in the analyses.
Venous plasma glucose, both fasting state and in the course of an oral glucose tolerance test, was measured in duplicate with an automated analyzer. Basal and glucose stimulated insulin was measured by an immuno-enzymometric assay (AIA-PACK IRI, AIA-1200 system, Tosoh Co., Tokyo, Japan) with intra- and inter-assay coefficients of variation for quality control < 7%. Insulin resistance was calculated by means of the homeostasis model assessment method (HOMA-R).[27] Fasting total cholesterol, high-density lipoprotein (HDL)-cholesterol, and triglycerides were measured by common standard laboratory techniques (CHOL, HDL-C [plus 2nd generation], and TG assays (Roche Diagnostics, Indianapolis, IN).
The risk of cardiac events at 10 years was estimated by means of the Framingham equation[28] and by the score derived from the Prospective Cardiovascular Munster (PROCAM) study[29] and the ATPIII score.[25] They are based on history or biochemistry and clinical data, but differ slightly in the relative weight and type of predictive variables.
Ultrasound Examination of the Brachial Artery.
High-resolution external ultrasound was used to measure brachial artery diameter in the basal state, in response to an increase in blood flow (causing shear stress) during reactive hyperemia (leading to flow-mediated endothelium-dependent dilation, FMV) and in response to sublingual nitroglycerin, an endothelium-independent, direct smooth-muscle dilator.[20]
The artery was scanned and the diameter measured during 3 conditions: at baseline, during reactive hyperemia (induced by inflation and deflation of a sphygmomanometer cuff around the limb, distal to the scanned part of the artery), and finally after administration of sublingual nitroglycerin using a normal antianginal dose of 400 g (causing endothelium-independent smooth muscle-mediated vasodilation).
The time required to obtain a high-quality baseline scan was at least 10 minutes after supine rest; after 10 minutes of rest, ischemia was induced by inflation to 250 mmHg and the deflation of a sphygmomanometer cuff around forearm. Finally, after a new period of 10 minutes of rest, 400 g nitroglycerin was administered sublingually, giving a vasodilation of the artery by a mechanism that is independent of the endothelium. The scanning period used in our laboratory is 15 seconds before and 60 seconds after cuff deflation.
Data were expressed as absolute diameter values (in millimeters) and as percent increase above basal in response ischemia (percent FMV) and sublingual nitroglycerin.
Statistical Analysis.
Data were processed on a personal computer and analyzed using StatView 5.0 (SAS Institute, Inc., Cary, NC). Patients were grouped according to categorical variables (sex, presence/absence of MS and its individual components). Differences between controls and NAFLD cases were tested for significance using unpaired t test (two-tailed) or nonparametric analysis (Mann-Whitney U test or Kruskal-Wallis test). Contingency test and Fisher's exact test also were used, whenever appropriate, to compare prevalence. Linear regression analysis and Spearman rank correlation were used to examine the relationship between variables. Finally, logistic regression analysis was used to test the association of a reduced percent FMV (lower tertile, FMV < 5%) with anthropometric, clinical, and biochemical parameters. In multivariate analysis, 3 models were tested, to identify the association of reduced percent FMV with the number of features of metabolic syndrome, with NAFLD, and finally with NASH. The first 2 models were applied to the whole data set; the third was applied only to NAFLD patients.
All data in the text and in the tables are given as means ± SD, when not otherwise indicated. Values of P < .05 were considered statistically significant.
|
|
| |
| |
|
|
|
