Hepatic steatosis is closely associated with the metabolic syndrome. We assessed for an independent association between hepatic steatosis and atherogenic dyslipidemia after adjustment for obesity, physical activity, hyperglycemia, and systemic inflammation. We studied 6,333 asymptomatic subjects without clinical cardiovascular disease undergoing a health screen in Brazil from November 2008 to July 2010. Hepatic steatosis was diagnosed by ultrasound. Atherogenic dyslipidemia was defined using 2 definitions: criteria for (1) metabolic syndrome or (2) insulin resistance (triglyceride/high-density–lipoprotein cholesterol ratio of ≥2.5 in women and ≥3.5 in men). In hierarchical multivariate regression models, we evaluated for an independent association of hepatic steatosis with atherogenic dyslipidemia. Hepatic steatosis was detected in 36% of participants (average age 43.5 years, 79% men, average body mass index 26.3 kg/m 2 ). Subjects with hepatic steatosis had similar levels of low-density–lipoprotein cholesterol, with significantly lower level of high-density–lipoprotein cholesterol and higher level of triglyceride compared with those without steatosis. Hepatic steatosis remained significantly independently associated with atherogenic dyslipidemia of both definitions (metabolic syndrome [odds ratio 2.47, 95% confidence interval 2.03 to 3.02] and insulin resistance [odds ratio 2.50, 95% confidence interval 2.13 to 2.91]) after multivariate adjustment. Stratified analyses showed a persistent independent association in nonobese subjects, those without metabolic syndrome, those with normal high-sensitivity C-reactive protein, nonalcohol abusers, and those with normal liver enzymes. Hepatic steatosis was significantly associated with atherogenic dyslipidemia independent of obesity, physical activity, hyperglycemia, and systemic inflammation after multivariate adjustment. In conclusion, this adds to the growing body of evidence that hepatic steatosis may play a direct metabolic role in conferring increased cardiovascular risk.
An emerging body of evidence suggests that hepatic steatosis may play a central pathophysiologic role in atherogenic dyslipidemia, chiefly as a substrate for increased production of large, triglyceride (TG)-rich, very-low-density lipoprotein (VLDL) particles. This ultimately leads to an increased number of smaller denser low-density–lipoprotein cholesterol (LDL-C) particles, which are more efficiently incorporated into arterial walls, thereby promoting progression of atherosclerotic plaques. Indeed, those patients with elevated levels of VLDL have been shown to have roughly 3× the risk of coronary heart disease compared with those with normal VLDL levels. Clinically, atherogenic dyslipidemia is typically characterized by elevated plasma TGs, decreased high-density–lipoprotein cholesterol (HDL-C), increased small LDL-C, and increased apolipoprotein B levels. Whether hepatic steatosis is directly associated with this atherogenic phenotype independent of other existing metabolic risk factors is not well known. We hypothesized that in a large unselected clinical sample, the presence of hepatic steatosis would be independently associated with atherogenic dyslipidemia. To address this question, we assessed the association between ultrasound-diagnosed hepatic steatosis and atherogenic dyslipidemia after adjusting for obesity, physical activity, and other components of the metabolic syndrome such as diabetes, glucose levels, and systemic inflammation.
Methods
Our study population consisted of adult men and women who received a mandatory routine employment health screen at the Preventative Medicine Center of the Hospital Israelita Albert Einstein in Sao Paulo, Brazil from November 2008 to July 2010. All study participants were asymptomatic and free of known cardiovascular disease (CVD). The health screen protocol consisted of a clinical evaluation, laboratory evaluation, and abdominal ultrasound. All participating subjects provided demographics, medical history, smoking status, quantitative alcohol consumption, physical activity, and medication use at the time of clinical evaluation. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki as reflected by approval by the Institutional Review Board of Hospital Israelita Albert Einstein and was granted a waiver for informed consent.
A standard questionnaire was used to gather medical history. Diabetes and hypertension were ascertained by previous physician diagnosis of these conditions or use of glucose or blood pressure–lowering medications. Those subjects with a systolic blood pressure >140 mm Hg or a diastolic blood pressure >90 mm Hg were included as subjects with hypertension. Blood pressure was measured during the clinical evaluation with a calibrated sphygmomanometer using the standard method recommended by the American Heart Association. Waist circumference was recorded with a plastic anthropometric tape held parallel to the ground, measuring the narrowest diameter between the iliac crest and the costal margin. Smoking status was defined as current smoker versus current nonsmoker. Alcohol use was quantified by using the Alcohol Use Disorders Identification Test and was stratified into 2 categories based on score (normal alcohol use [0 to 7] and alcohol abuse [>7]). Physical activity was determined by the International Physical Activity Questionnaire and was stratified into 4 categories (sedentary, little activity, active, and very active).
Blood specimens were collected after overnight fasting. Plasma glucose, liver aminotransferases, total cholesterol (TC), TGs, and γ-glutamyltranspeptidase levels were obtained using standardized laboratory testing on a Vitros automated platform (Johnson & Johnson Clinical Diagnostics, Rochester, New York). HDL-C level was determined using a precipitation method and LDL-C was calculated ([TC − HDL-C − TG]/5) for TG levels up to 400 mg/dl. For those with a TG level of >400 mg/dl, LDL-C level was measured directly by ultracentrifugation. Abnormal liver function testing was defined as alanine aminotransferase level of ≥40 IU/L, aspartate aminotransferase level of ≥40 IU/L, or γ-glutamyltranspeptidase level of ≥50 IU/L. High-sensitivity C-reactive protein (hs-CRP) levels were measured by immunonephelometry (Dade-Behring), and hs-CRP levels of ≥3 mg/L were defined as elevated. All laboratory tests were performed at the Central Laboratory of the Hospital Israelita Albert Einstein.
Abdominal ultrasounds were performed on all subjects after a minimum of 6-hour fasting. All images were obtained using a Siemens ACUSON XP-10 device (Mountain View, California). Hepatic steatosis was diagnosed by the presence of an ultrasonographic pattern of a hyperechoic liver with evidence of contrast between hepatic and renal parenchyma using a method that has been previously described. All acquired images were reviewed and read separately by 2 board-certified radiologists who were blinded to laboratory test data. Using multiple ultrasonographic views, a qualitative diagnosis (presence or absence) of hepatic steatosis was made if the subject met the defined criteria of altered echotexture. Degrees of severity of steatosis were not measured. Previous studies have shown that ultrasonography can identify and quantitate hepatic steatosis with accuracy similar to that of computed tomography, magnetic resonance, and liver biopsy.
Obesity was defined as a body mass index (BMI) of ≥30 kg/m 2 or waist circumference of >88 cm in women and >102 cm in overweight men with a BMI of >25 kg/m 2 . The Harmonizing Definition was used for the definition of metabolic syndrome including subjects with ≥3 of the following factors: subjects with elevated waist circumference (>88 cm in women or >102 cm in men or BMI of >30 kg/m 2 ), elevated blood pressure (≥130 mm Hg systolic and/or ≥85 mmHg diastolic or antihypertensive medications), fasting blood glucose level of ≥100 mg/dl or use of drug treatment for hyperglycemia, elevated TGs level (≥150 mg/dl), or low HDL-C level (<40 mg/dl in men or <50 mg/dl in women) or the use of lipid-lowering medications.
Atherogenic dyslipidemia was defined using 2 separate previously defined definitions: (1) dyslipidemia of the metabolic syndrome was defined as a TG level of ≥150 mg/d and HDL-C level of <40 mg/dl in men or <50 mg/dl in women or the use of lipid-lowering medications and (2) dyslipidemia of insulin resistance was defined as a TG/HDL-C ratio of ≥2.5 in women and ≥3.5 in men. Both definitions were used separately for the analysis to determine the association of hepatic steatosis with atherogenic dyslipidemia.
All statistical analyses were performed using STATA version 12 (StataCorp, College Station, Texas). Baseline characteristics of subjects with and without hepatic steatosis were compared using Wilcoxon t test for continuous variables, Pearson chi-square test for categorical variables, and the nonparametric Kruskal-Wallis test for skewed variables. We performed multivariate linear regression analyses to assess the association of hepatic steatosis with absolute change in levels of TC, LDL-C, HDL-C, TG, and non-HDL-C and ratios of TG/HDL and TC/HDL. Multivariate logistic regression analyses were used to evaluate the association of hepatic steatosis with dyslipidemia of the metabolic syndrome and dyslipidemia of insulin resistance. Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated.
For all regression analyses, a hierarchical model approach was used, first adjusting for age, gender, alcohol use, and physical activity (model 2). To further adjust for obesity, components of the metabolic syndrome, and systemic inflammation, simultaneous adjustments were made for BMI, waist circumference, hypertension, antihypertensive medications, diabetes, diabetes medications, hyperglycemia, log hs-CRP, smoking status, and lipid-lowering medications (model 3). Sensitivity analyses were performed that excluded those subjects on lipid-lowering medications to reduce confounding. Finally, the associations between hepatic steatosis and dyslipidemia were stratified by obesity, log hs-CRP, metabolic syndrome, liver function tests, and alcohol use.
Results
The study population consisted of 6,333 asymptomatic mostly white subjects of European descent. Baseline characteristics of the subjects were stratified by the presence or absence of hepatic steatosis as diagnosed by liver ultrasound ( Table 1 ). Data for all variables were available for all subjects with the exception of hs-CRP level (n = 5,418) and physical activity (n = 5,697). All subjects were included in the analysis.
| Characteristics | Hepatic Steatosis | p | |
|---|---|---|---|
| No (n = 4,045) | Yes (n = 2,288) | ||
| Age (yrs) | 42.0 (9.5) | 46.1 (9.1) | <0.001 |
| Men (%) | 70.9 | 93.9 | <0.001 |
| Systolic blood pressure (mm Hg) | 115.8 (11.6) | 124.1 (12.3) | <0.001 |
| Diastolic blood pressure (mm Hg) | 75.0 (7.6) | 80.2 (7.5) | <0.001 |
| Waist circumference (cm) | 86.7 (10.1) | 100.4 (10.0) | <0.001 |
| BMI (kg/m 2 ) | 24.8 (3.1) | 29.0 (3.8) | <0.001 |
| Fasting glucose (mg/dl) | 87.0 (8.5) | 93.3 (11.9) | <0.001 |
| hs-CRP (mg/L) ∗ | 1.0 (0.5 to 2.0) | 1.7 (1.0 to 3.0) | <0.001 |
| AST (IU/L) | 27 (24 to 32) | 32 (28 to 38) | <0.001 |
| ALT (IU/L) | 29 (23 to 37) | 42 (33 to 56) | <0.001 |
| GGT (IU/L) | 26 (20 to 37) | 40 (30 to 56) | <0.001 |
| Hypertension (%) | 7.8 | 21.3 | <0.001 |
| Antihypertensive medications (%) | 7.6 | 20.7 | <0.001 |
| Diabetes mellitus (%) | 0.4 | 2.1 | <0.001 |
| Diabetes medications (%) | 0.4 | 1.4 | <0.001 |
| Lipid-lowering medications (%) | 7.0 | 11.7 | <0.001 |
| Current smoking (%) | 8.0 | 10.4 | 0.001 |
| Alcohol use (%) | |||
| Normal use (AUDIT score of 0–7) | 85.8 | 80.6 | <0.001 |
| Alcohol abuse (AUDIT score of >7) | 14.2 | 19.4 | <0.001 |
| Physical activity (%) † | |||
| Sedentary | 19.5 | 28.9 | <0.001 |
| Little activity | 33.8 | 36.2 | <0.001 |
| Active | 35.4 | 29.8 | <0.001 |
| Very active | 11.3 | 5.1 | <0.001 |
| Obesity (%) ‡ | 9.7 | 45.9 | <0.001 |
∗ For hs-CRP, n = 3,452 for no hepatic steatosis and 1,966 for hepatic steatosis.
† For physical activity, n = 3,643 for no hepatic steatosis and 2,054 for hepatic steatosis.
‡ Obesity is defined as a BMI of ≥30 kg/m 2 or waist circumference >88 cm in women or >102 cm in men if BMI is ≥25 kg/m 2 .
Hepatic steatosis was significantly associated with higher TC, lower HDL-C, and higher TG levels and other markers of atherogenic dyslipidemia ( Figure 1 ), which persisted after multivariate adjustment ( Table 2 ). A difference in LDL-C level persisted after adjustment in model 2 but was lost after full multivariate regression in model 3.
| Lipid Parameter | Model 1 | Model 2 | Model 3 |
|---|---|---|---|
| β coefficient (95% CI) | β coefficient (95% CI) | β coefficient (95% CI) | |
| TC (mg/dl) | 9.43 (7.50 to 11.37) | 6.33 (4.15 to 8.51) | 4.49 (1.87 to 7.12) |
| LDL-C (mg/dl) | 7.39 (5.67 to 9.10) | 3.28 (1.35 to 5.22) | 0.73 (−1.58 to 3.03) |
| HDL-C (mg/dl) | −8.74 (−9.33 to −8.16) | −5.55 (−6.16 to −4.95) | −2.96 (−3.69 to −2.23) |
| TGs (mg/dl) ∗ | 0.40 (0.37 to 0.42) | 0.31 (0.28 to 0.33) | 0.23 (0.20 to 0.26) |
| Non-HDL-C (mg/dl) | 18.2 (16.2 to 20.1) | 11.9 (9.70 to 14.08) | 7.47 (4.83 to 10.11) |
| TG/HDL ratio ∗ | 0.57 (0.54 to 0.61) | 0.43 (0.39 to 0.46) | 0.29 (0.25 to 0.34) |
| TC/HDL ratio ∗ | 0.22 (0.21 to 0.24) | 0.15 (0.13 to 0.16) | 0.09 (0.07 to 0.10) |
Subjects with hepatic steatosis had significantly higher odds of dyslipidemia (metabolic syndrome definition: OR 3.41, 95% CI 2.90 to 4.02 and insulin resistance definition: OR 3.66, 95% CI 3.22 to 4.15) in model 2 ( Figure 2 ). After additionally adjusting for other predictors of dyslipidemia (model 3), this association retained significance (metabolic syndrome definition: OR 2.47, 95% CI 2.03 to 3.02 and insulin resistance definition: OR 2.50, 95% CI 2.13 to 2.91). A sensitivity analysis excluding subjects on lipid-lowering medications had similar results (metabolic syndrome definition: OR 2.60, 95% CI 2.10 to 3.20 and insulin resistance definition: OR 2.54, 95% CI 2.15 to 3.00).
