Advertisement

Independent Association Between Nonalcoholic Fatty Liver Disease and Cardiovascular Disease in the US Population

  • Maria Stepanova
    Affiliations
    Center for Liver Diseases, Department of Medicine, Inova Fairfax Hospital, Falls Church, Virginia

    Betty and Guy Beatty Center for Integrated Research, Inova Health System, Falls Church, Virginia
    Search for articles by this author
  • Zobair M. Younossi
    Correspondence
    Reprint requests Address requests for reprints to: Zobair M. Younossi, MD, MPH, Betty and Guy Beatty Center for Integrated Research and Claude Moore Health Education and Research Building, 3300 Gallows Road, Falls Church, Virginia 22042. fax: (703) 776-4386
    Affiliations
    Center for Liver Diseases, Department of Medicine, Inova Fairfax Hospital, Falls Church, Virginia

    Betty and Guy Beatty Center for Integrated Research, Inova Health System, Falls Church, Virginia
    Search for articles by this author
Published:January 16, 2012DOI:https://doi.org/10.1016/j.cgh.2011.12.039

      Background & Aims

      Nonalcoholic fatty liver disease (NAFLD) and cardiovascular diseases (CVDs) have common metabolic risk factors. Despite reports from clinical studies, the association between NAFLD, cardiovascular disease, and cardiovascular mortality are not clear at the population level.

      Methods

      We analyzed data from the National Health and Nutrition Examination Survey III, conducted from 1988 to 1994, and compared hepatic ultrasound and mortality data. Participants were classified into those with NAFLD (moderate or severe hepatic steatosis, based on ultrasound analysis, without any evidence of other liver disease; n = 2492) and those without (absence of NAFLD or any other chronic liver diseases: controls). The prevalence of CVD was compared between subjects with and without NAFLD. Additional comparisons were made between NAFLD patients who had increased levels of liver enzymes and those who had normal levels. Independent predictors of CVD and cardiovascular mortality also were studied.

      Results

      During the follow-up period (median, 171 mo), 12.21% of the National Health and Nutrition Examination Survey III participants died; cardiovascular mortality was 3.76%. Regardless of whether levels of liver enzymes were increased or not, individuals with NAFLD were older, predominantly male, more likely to be Hispanic, and less likely to be African American than controls. They also had a higher prevalence of all components of metabolic syndrome and CVD. Regardless of levels of liver enzymes, NAFLD was associated independently with CVD, after adjusting for major demographic, clinical, and metabolic confounders (odds ratio, 1.23; 95% confidence interval, 1.04–1.44). The independent association of NAFLD with cardiovascular mortality was not statistically significant.

      Conclusions

      NAFLD is associated independently with an increased risk of CVD. However, NAFLD did not increase cardiovascular mortality over a 14-year period.

      Keywords

      Abbreviations used in this paper:

      CI (confidence interval), CVD (cardiovascular disease), CVM (cardiovascular mortality), HCV (hepatitis C virus), MI (myocardial infarction), NAFLD (nonalcoholic fatty liver disease), NHANES (National Health and Nutrition Examination Survey), NCHS (US National Center for Health Statistics), OR (odds ratio)
      See editorial on page 568.
      The relationship between nonalcoholic fatty liver disease (NAFLD) and cardiovascular diseases (CVDs) has long been suspected.
      • Targher G.
      • Bertolini L.
      • Poli F.
      • et al.
      Nonalcoholic fatty liver disease and risk of future cardiovascular events among type 2 diabetic patients.
      • Lin Y.C.
      • Lo H.M.
      • Chen J.D.
      Sonographic fatty liver, overweight and ischemic heart disease.
      • Targher G.
      • Bertolini L.
      • Padovani R.
      • et al.
      Relations between carotid artery wall thickness and liver histology in subjects with nonalcoholic fatty liver disease.
      In fact, NAFLD and CVD often are considered as 2 different manifestations of metabolic syndrome and share the same risk factors such as visceral obesity, type II diabetes, dyslipidemia, and insulin resistance.
      • Targher G.
      Non-alcoholic fatty liver disease, the metabolic syndrome and the risk of cardiovascular disease: the plot thickens.
      Indeed, a number of studies have shown a higher prevalence of atherosclerosis in individuals with NAFLD.
      • Lizardi-Cervera J.
      • Aguilar-Zapata D.
      Nonalcoholic fatty liver disease and its association with cardiovascular disease.
      Possible molecular mechanisms include the increased production of inflammatory cytokines, often observed in the setting of NAFLD, possibly accompanied by abnormal levels of lipoproteins, endothelial dysfunction, and oxidative stress.
      • Villanova N.
      • Moscatiello S.
      • Ramilli S.
      • et al.
      Endothelial dysfunction and cardiovascular risk profile in nonalcoholic fatty liver disease.
      • Van Gaal L.F.
      • Mertens I.L.
      • De Block C.E.
      Mechanisms linking obesity with cardiovascular disease.
      • Targher G.
      • Marra F.
      • Marchesini G.
      Increased risk of cardiovascular disease in non-alcoholic fatty liver disease: causal effect or epiphenomenon?.
      Furthermore, some studies confirmed that NAFLD is associated with an increased risk of CVD, even after adjustment for major cardiovascular and metabolic risk factors.
      • Hamaguchi M.
      • Kojima T.
      • Takeda N.
      • et al.
      Nonalcoholic fatty liver disease is a novel predictor of cardiovascular disease.
      Both tertiary-care
      • Rafiq N.
      • Bai C.
      • Fang Y.
      • et al.
      Long-term follow-up of patients with nonalcoholic fatty liver.
      and population-based studies
      • Stepanova M.
      • Rafiq N.
      • Younossi Z.M.
      Components of metabolic syndrome are independent predictors of mortality in patients with chronic liver disease: a population-based study.
      have shown that CVD is either the most common, or is among the 2 most common causes of death for individuals with NAFLD. However, the suggested independent association of NAFLD with cardiovascular mortality (CVM)
      • Dunn W.
      • Xu R.
      • Wingard D.L.
      • et al.
      Suspected nonalcoholic fatty liver disease and mortality risk in a population-based cohort study.
      has not been proven in other studies using population-based data.
      • Stepanova M.
      • Younossi Z.M.
      Association of chronic liver diseases and cardiovascular mortality—the authors' reply.
      • Musso G.
      • Gambino R.
      • Cassader M.
      • et al.
      Meta-analysis: natural history of non-alcoholic fatty liver disease (NAFLD) and diagnostic accuracy of non-invasive tests for liver disease severity.
      • Calori G.
      • Lattuada G.
      • Ragogna F.
      • et al.
      Fatty liver index and mortality: the Cremona study in the 15th year of follow-up.
      It is important to note, however, that population-based findings reporting an association of NAFLD and CVD/CVM published to date used the clinical definition of NAFLD using liver enzyme levels and laboratory tests only. Although valuable, this approach can result in a smaller and probably biased sample of patients presumed to have NAFLD. Thus, the suspected association of NAFLD with CVM at the population level requires validation using a more established approach for the diagnosis of NAFLD.
      The aim of our study was to examine the association of radiologically proven NAFLD with cardiovascular events, including cardiovascular disease and cardiovascular mortality, using recent population-based data.

      Methods

       Study Population

      For this study, we used the National Health and Nutrition Examination Survey III (NHANES III) conducted between 1988 and 1994. The survey data were collected by the US National Center for Health Statistics (NCHS) of the Centers for Disease Control and Prevention via household interviews, physical examination, and laboratory tests, and contains information describing the health and nutritional status of the US population as of the time of the data collection. In addition, the data collections with hepatic ultrasound data for NHANES III participants were released recently. Therefore, the NHANES III participants included in this study were required to have the hepatic ultrasound data together with the relevant demographic, clinical, and examination data available. Participants with data insufficient for ruling in/out NAFLD and CVD were excluded from the study.

       NAFLD Diagnosis

      For eligible NHANES III participants, which included adults between 20 and 74 years of age at the time of examination, the archived hepatobiliary ultrasound video images recently were re-reviewed by NCHS and the results of their assessment were published.

      NHANES 1988–1994: Hepatic/Gallbladder Ultrasound and Hepatic Steatosis Data Documentation, Codebook, and Frequencies. Available at: http://www.cdc.gov/nchs/nhanes/nhanes3/HGUHS.htm. Accessed August 18, 2011.

      From the data collection, to establish the diagnosis of NAFLD, we used the parameter representing the presence of fat within the hepatic parenchyma, which was graded as normal-mild or moderate-severe. Quality control and quality assurance procedures were used by NCHS to standardize the readings from 3 ultrasound readers who had no access to any other clinical or laboratory data of the participants.

      NHANES 1988–1994: Hepatic/Gallbladder Ultrasound and Hepatic Steatosis Data Documentation, Codebook, and Frequencies. Available at: http://www.cdc.gov/nchs/nhanes/nhanes3/HGUHS.htm. Accessed August 18, 2011.

      In our study, subjects were considered to have NAFLD if moderate to severe hepatic steatosis was noted on the ultrasound in the absence of any other possible causes of chronic liver disease such as excessive alcohol use (defined as self-reported consumption of ≥20 g/d in men and ≥10 g/d in women during the year preceding the examination for NHANES III), iron overload (defined as transferrin saturation of ≥50%), or a positive hepatitis B surface antigen or anti-hepatitis C virus (HCV) test (anti-HCV by enzyme-linked immunosorbent assay and HCV RNA by polymerase chain reaction).
      Furthermore, because abnormal level of aminotransferases could be indicative of a progressed form of NAFLD or nonalcoholic steatohepatitis, the individuals with a diagnosis of NAFLD were classified further into those with increased liver enzyme levels (defined as alanine aminotransferase level ≥40 U/L or aspartate aminotransferase level ≥37 U/L in men, alanine aminotransferase or aspartate aminotransferase ≥31 U/L in women) and those with normal liver enzyme levels.
      Individuals without hepatic steatosis on the ultrasound and any other cause of chronic liver disease (as described earlier) were presumed to have no chronic liver disease and were used as the control cohort.

       Other Definitions

      Four major race or ethnic groups included non-Hispanic whites, non-Hispanic blacks, Hispanics, and “other,” which included Aleut, Eskimo, American Indian, Asian, or Pacific Islander. We also defined obesity as a body mass index greater than 30 and insulin resistance as a homeostasis of model assessment score of 3.0 or higher. In addition, diabetes mellitus and history of smoking were consistent, as we have reported previously.
      • Stepanova M.
      • Rafiq N.
      • Younossi Z.M.
      Components of metabolic syndrome are independent predictors of mortality in patients with chronic liver disease: a population-based study.
      For the purpose of this study, we defined CVD as self-reported history of congestive heart failure, stroke, angina, or myocardial infarction (MI). Furthermore, family history of CVD was defined as a history of MI before the age of 50 in a parent or a sibling.

       Mortality Follow-up Data

      For all eligible NHANES III participants, the mortality follow-up data were collected from a publicly available NHANES III–Linked Mortality File provided by NCHS. In addition to mortality status, the linked mortality file contains the length of the follow-up period between the time of examination for NHANES III and the reported death or the end of the follow-up period (whichever is earlier), as well as causes of death. In the most recent data collection, the NHANES III participants aged 17 years and older had mortality follow-up evaluation until December 31, 2006. Individuals without available mortality follow-up data were ineligible for the study.
      The NHANES III Linked Mortality file uses the Underlying Cause of Death 113 code (UCOD113) to recode all deaths according to International Classification of Diseases 9th and 10th revision criteria. For the purpose of this study, we evaluated cardiovascular mortality (UCOD113 58–63, 67, 70–74), which covered causes of death such as ischemic heart diseases, heart failure, atherosclerosis, cerebrovascular diseases, aortic aneurysm, and other diseases of arteries, arterioles, and capillaries.

       Statistical Analyses

      As recommended by the NHANES III Analytic Guidelines,
      the sampling weights were used to account for nonresponse and unequal selection probabilities for certain categories of the population. Thus, after weighting on the basis of age, sex, level of education, and race or ethnic group, the distribution of participants was representative of that of the US population. In addition, stratum and sampling units accounted for the survey design effects using Taylor series linearization.
      As a first step, individuals with NAFLD as well as its subgroups with increased and normal liver enzyme levels were compared separately with non-NAFLD controls. For the purpose of the study, continuous variables were compared between the cohorts using a t test for a contrasted mean, and the prevalence of various parameters, including demographic parameters, metabolic syndrome components, and CVD, was compared by the stratum-specific chi-square test for independence. P values of .05 or less were considered potentially significant.
      At the next step, the presence of NAFLD or its subgroups was tested for independent association with CVD in the US population and certain subpopulations. Logistic regression was used to identify independent predictors of CVD, and all studied demographic, clinical, and laboratory parameters were used as potential confounders.
      Finally, to identify risk factors independently associated with cardiovascular mortality, a Cox proportional hazard model was used. Proportional hazard assumption was controlled for all potential mortality predictors.
      All analyses were run with SAS 9.1 and SUDAAN 10.0 (SAS Institute, Inc, Cary, NC). The protocol of the study was approved by the Inova Institutional Review Board.

      Results

      Of the initial study population (20,050 adult participants from NHANES III), 2492 individuals (18.77% ± 0.76%) fulfilled our clinical and radiologic definition of NAFLD, 426 (15.86% ± 1.15%) of them had NAFLD with increased liver enzyme levels, and 9121 were used as non-NAFLD controls.
      The results of the comparison of the NAFLD cohort with the non-NAFLD controls are summarized in Table 1. As expected, individuals with NAFLD were older, more likely to be male or Hispanic, and less likely to be African American. They also had a higher prevalence of obesity, diabetes mellitus, and insulin resistance.
      Table 1Comparison of Subpopulations With and Without NAFLD
      All NAFLDNAFLD with increased liver enzyme levels
      For patients with NAFLD with normal liver enzyme levels, the comparison is made with non-NAFLD controls, whereas patients with NAFLD with increased liver enzyme levels are compared with controls and those with NAFLD with normal liver enzyme levels.
      NAFLD with normal liver enzyme levels
      For patients with NAFLD with normal liver enzyme levels, the comparison is made with non-NAFLD controls, whereas patients with NAFLD with increased liver enzyme levels are compared with controls and those with NAFLD with normal liver enzyme levels.
      Non-NAFLD controls
      N249242620669121
      Age
       35–44, %22.04 ± 1.6027.24 ± 3.9721.06 ± 1.8323.91 ± 0.79
       45–54, %20.33 ± 1.23
      Different from controls (P ≤ .05).
      21.68 ± 3.8120.08 ± 1.40
      Different from controls (P ≤ .05).
      14.53 ± 0.61
       55–64, %18.78 ± 1.29
      Different from controls (P ≤ .05).
      11.78 ± 2.4620.10 ± 1.36
      Different from controls (P ≤ .05).
      11.81 ± 0.56
       65–74, %15.53 ± 1.25
      Different from controls (P ≤ .05).
      8.34 ± 2.0416.89 ± 1.35
      Different from controls (P ≤ .05).
      10.26 ± 0.65
      Male, %52.83 ± 1.62
      Different from controls (P ≤ .05).
      53.60 ± 4.2752.68 ± 1.59
      Different from controls (P ≤ .05).
      46.13 ± 0.72
      White, %75.66 ± 2.6171.33 ± 3.60
      Different from controls (P ≤ .05).
      76.48 ± 2.6477.03 ± 2.17
      Black, %8.48 ± 1.32
      Different from controls (P ≤ .05).
      6.26 ± 1.34
      Different from controls (P ≤ .05).
      8.90 ± 1.4010.71 ± 1.17
      Hispanic, %7.61 ± 1.38
      Different from controls (P ≤ .05).
      13.76 ± 2.45
      Different from controls (P ≤ .05).
      6.45 ± 1.21
      Different from controls (P ≤ .05).
      4.89 ± 0.81
      Obesity, %47.56 ± 1.85
      Different from controls (P ≤ .05).
      60.25 ± 4.53
      Different from controls (P ≤ .05).
      45.17 ± 1.89
      Different from controls (P ≤ .05).
      17.10 ± 0.75
      Diabetes, %14.49 ± 0.93
      Different from controls (P ≤ .05).
      20.82 ± 2.85
      Different from controls (P ≤ .05).
      13.30 ± 1.07
      Different from controls (P ≤ .05).
      3.03 ± 0.32
      Insulin resistance, %54.80 ± 1.70
      Different from controls (P ≤ .05).
      74.01 ± 3.44
      Different from controls (P ≤ .05).
      51.18 ± 1.82
      Different from controls (P ≤ .05).
      16.46 ± 0.83
      Smoking, %22.68 ± 1.29
      Different from controls (P ≤ .05).
      16.43 ± 3.82
      Different from controls (P ≤ .05).
      23.85 ± 1.42
      Different from controls (P ≤ .05).
      31.49 ± 1.30
      Family history of CVD, %10.77 ± 0.9411.19 ± 2.4110.69 ± 0.999.30 ± 0.61
      Any cardiovascular disease, %38.18 ± 1.68
      Different from controls (P ≤ .05).
      38.62 ± 4.3238.09 ± 1.81
      Different from controls (P ≤ .05).
      29.26 ± 0.88
      History of congestive heart failure, %3.24 ± 0.57
      Different from controls (P ≤ .05).
      1.21 ± 0.583.62 ± 0.70
      Different from controls (P ≤ .05).
      1.26 ± 0.15
      History of stroke, %2.48 ± 0.40
      Different from controls (P ≤ .05).
      1.99 ± 1.012.57 ± 0.42
      Different from controls (P ≤ .05).
      1.15 ± 0.17
      History of angina, %35.74 ± 1.60
      Different from controls (P ≤ .05).
      36.82 ± 4.3735.53 ± 1.74
      Different from controls (P ≤ .05).
      28.22 ± 0.85
      History of MI, %5.12 ± 0.71
      Different from controls (P ≤ .05).
      3.32 ± 1.135.46 ± 0.77
      Different from controls (P ≤ .05).
      2.07 ± 0.22
      a For patients with NAFLD with normal liver enzyme levels, the comparison is made with non-NAFLD controls, whereas patients with NAFLD with increased liver enzyme levels are compared with controls and those with NAFLD with normal liver enzyme levels.
      b Different from controls (P ≤ .05).
      Assessing the history of CVD revealed a higher prevalence of CVD among those with NAFLD (Table 1). In addition, all the subgroups of CVD (history of congestive heart failure, stroke, angina, and MI) were all significantly more prevalent among individuals with NAFLD when compared with non-NAFLD controls.
      Among those with NAFLD and increased liver enzyme levels, metabolic syndrome components were even more prevalent (Table 1). However, all other demographic and clinical parameters did not show any significant difference when NAFLD with increased liver enzyme levels was compared with NAFLD with normal liver enzyme levels. On the other hand, when individuals with NAFLD with normal liver enzyme levels (Table 1) were compared with those without NAFLD, a pattern of differences similar to that observed for the overall NAFLD diagnosis emerged.

       Independent Association of NAFLD and CVD

      A summary of independent predictors of CVD is provided in Table 2. As expected, older age, male sex, obesity, presence of type 2 diabetes, history of smoking, and family history of early MI all were associated independently with an increased risk of CVD. In addition, NAFLD also was found to be associated independently with a higher risk of CVD (odds ratio [OR], 1.23; 95% confidence interval [CI], 1.04–1.44). Contrary to the recently published finding,
      • Calori G.
      • Lattuada G.
      • Ragogna F.
      • et al.
      Fatty liver index and mortality: the Cremona study in the 15th year of follow-up.
      an association of NAFLD with CVD after adjustment for insulin resistance also was detected (OR, 1.19; 95% CI, 1.02–1.41).
      Table 2Independent Predictors of Cardiovascular Diseases in the US Population
      All NAFLD
      In each column, the respective form of NAFLD is used as a potential predictor of CVD.
      NAFLD with increased liver enzyme levels
      In each column, the respective form of NAFLD is used as a potential predictor of CVD.
      NAFLD with normal liver enzyme levels
      In each column, the respective form of NAFLD is used as a potential predictor of CVD.
      Age1.01 (1.01–1.02)1.02 (1.01–1.02)1.01 (1.01–1.02)
      Male sex1.16 (1.01–1.33)1.17 (1.02–1.35)1.18 (1.03–1.35)
      Caucasian1.34 (1.01–1.79)1.33 (1.00–1.78)1.38 (1.02–1.87)
      African American1.23 (0.92–1.65)1.21 (0.91–1.63)1.28 (0.94–1.75)
      Hispanic1.09 (0.81–1.47)1.10 (0.81–1.48)1.13 (0.82–1.55)
      Obesity1.32 (1.14–1.54)1.37 (1.18–1.60)1.32 (1.14–1.54)
      Diabetes mellitus1.29 (1.02–1.63)1.33 (1.05–1.69)1.36 (1.08–1.72)
      Smoking1.18 (1.01–1.40)1.18 (1.00–1.39)1.18 (1.00–1.40)
      Family history of CVD2.02 (1.72–2.36)2.02 (1.73–2.36)2.06 (1.75–2.43)
      Diagnosis of NAFLD
      In each column, the respective form of NAFLD is used as a potential predictor of CVD.
      1.23 (1.04–1.44)1.23 (0.81–1.85)1.20 (1.02–1.42)
      NOTE. OR (95% CI) for CVD is shown.
      a In each column, the respective form of NAFLD is used as a potential predictor of CVD.
      In one of the prior studies, the association of cardiovascular events and NAFLD for the middle-aged population was suggested.
      • Dunn W.
      • Xu R.
      • Wingard D.L.
      • et al.
      Suspected nonalcoholic fatty liver disease and mortality risk in a population-based cohort study.
      To evaluate this finding using a more extensive population-based data collection, we chose to repeat the analysis for 4 subpopulations separately based on the baseline age, specifically, for those groups between ages 35 and 44, 45 and 54, 55 and 64, and 65 and 74 years. Our results show that although NAFLD was found to be associated independently with a higher risk of CVD in the entire US population, there was only a trend for such an association detected for the age group of 35 to 44 years (OR, 1.40; 95% CI, 0.96–2.02; P = .0767). All other major risk factors for CVD remained independently associated with the outcome in every separate age group.

       NAFLD and Cardiovascular Mortality

      Of eligible participants, 1844 have died during the follow-up period (mortality rate, 12.21% ± 0.60%; mean length of follow-up evaluation, 171 mo). Of those, 591 events were considered cardiovascular mortality (rate, 3.76% ± 0.29%).
      The rates of both overall and cardiovascular mortality were higher among individuals with NAFLD regardless of their liver enzyme levels than in the non-NAFLD controls: the overall mortality rate was 16.52% ± 1.00% in patients with NAFLD, 17.56% ± 1.05% in patients with NAFLD with normal liver enzyme levels, and 11.21% ± 0.64% in controls (all P < .0001); cardiovascular mortality was 5.62% ± 0.59% in patients with NAFLD, 6.18% ± 0.67% in patients with NAFLD with normal liver enzyme levels, and 3.33% ± 0.29% in controls (all P < .0001). For the NAFLD cohort, CVD was the most prevalent cause of death (5.62% ± 0.59%), followed by solid organ malignancies as the next most common cause of death (3.38% ± 0.44%). In fact, as could be seen, approximately half of the difference in overall mortality for the NAFLD cohort could be explained by the higher rate of CVM in that cohort.
      For the non-NAFLD controls, CVD and solid organ malignancies were equally common causes of death (3.33% ± 0.29% and 3.30% ± 0.28%; respectively).
      Despite these associations, after adjustment for major mortality confounders (age, male sex, ethnicity, obesity, diabetes, smoking, and family history of MI), the association of NAFLD with the risk of CVM was no longer significant (Table 3). After similar analyses, neither NAFLD with increased liver enzyme levels nor NAFLD with normal liver enzyme levels were found to be associated with CVM, and no such association was found when different age groups were studied separately as well (data not shown).
      Table 3Independent Predictors of Cardiovascular Mortality in the US Population
      All NAFLD
      In each column, the respective form of NAFLD is used as a potential risk factor for CVM.
      NAFLD with increased liver enzyme levels
      In each column, the respective form of NAFLD is used as a potential risk factor for CVM.
      NAFLD with normal liver enzyme levels
      In each column, the respective form of NAFLD is used as a potential risk factor for CVM.
      Age1.12 (1.10–1.13)1.12 (1.10–1.13)1.12 (1.10–1.13)
      Male sex1.71 (1.34–2.19)1.71 (1.34–2.18)1.73 (1.34–2.22)
      Caucasian1.54 (1.03–2.30)1.54 (1.03–2.30)1.54 (1.03–2.30)
      African American1.97 (1.31–2.94)1.97 (1.32–2.93)1.95 (1.31–2.90)
      Obesity1.37 (1.00–1.86)1.36 (1.01–1.82)1.33 (0.96–1.84)
      Diabetes mellitus1.90 (1.35–2.69)1.91 (1.37–2.67)1.93 (1.34–2.76)
      Smoking2.06 (1.72–2.48)2.05 (1.70–2.47)2.02 (1.69–2.41)
      Family history of CVD1.64 (1.18–2.30)1.67 (1.19–2.35)1.62 (1.14–2.29)
      Diagnosis of NAFLD
      In each column, the respective form of NAFLD is used as a potential risk factor for CVM.
      0.92 (0.70–1.19)0.66 (0.31–1.39)0.95 (0.73–1.23)
      NOTE. The adjusted hazard ratio (95% CI) is shown.
      a In each column, the respective form of NAFLD is used as a potential risk factor for CVM.

      Discussion

      Increasing data suggest an association between NAFLD and cardiovascular diseases that could not be attributable solely to the risk factors common for both of these chronic diseases. A number of cross-sectional studies have confirmed this association and suggested that individuals with diagnosed NAFLD should be assessed for the risk of CVD.
      • Targher G.
      Non-alcoholic fatty liver disease, the metabolic syndrome and the risk of cardiovascular disease: the plot thickens.
      Other studies have proven that the risk of cardiovascular diseases increases even more when hepatic damage progresses, and described possible molecular mechanisms for the pathogenesis of CVD affected by progressing NAFLD.
      • Targher G.
      • Marra F.
      • Marchesini G.
      Increased risk of cardiovascular disease in non-alcoholic fatty liver disease: causal effect or epiphenomenon?.
      • Perseghin G.
      The role of non-alcoholic fatty liver disease in cardiovascular disease.
      • Sung K.C.
      • Ryan M.C.
      • Wilson A.M.
      The severity of nonalcoholic fatty liver disease is associated with increased cardiovascular risk in a large cohort of non-obese Asian subjects.
      It also recently was proposed that treating NAFLD potentially could help in reducing the risk of CVD.
      • Maurantonio M.
      • Ballestri S.
      • Odoardi M.R.
      • et al.
      Treatment of atherogenic liver based on the pathogenesis of nonalcoholic fatty liver disease: a novel approach to reduce cardiovascular risk?.
      In this study, we used an extensive data collection representative of the US population to validate the association of CVD and NAFLD. The data include a large number of individuals with available clinical, laboratory, and radiologic data as well as about 14 years of mortality follow-up evaluation. Our data show that even after adjustment for demographic disparity and underlying metabolic conditions including type II diabetes and obesity together with other known cardiovascular risk factors such as unfavorable family history and smoking, NAFLD was found to be associated independently with an increased risk of cardiovascular diseases (OR, 1.23; 95% CI, 1.04–1.44) regardless of the presence or absence of increased liver enzyme levels. An association close to statistical significance also was found for the group of individuals in the age group of 35 to 44 years (OR, 1.40; 95% CI, 0.96–2.02), but this finding requires validation.
      Although the association of cardiovascular mortality with NAFLD also has been suspected, the nature and strength of this association remains controversial.
      • Rafiq N.
      • Bai C.
      • Fang Y.
      • et al.
      Long-term follow-up of patients with nonalcoholic fatty liver.
      • Stepanova M.
      • Rafiq N.
      • Younossi Z.M.
      Components of metabolic syndrome are independent predictors of mortality in patients with chronic liver disease: a population-based study.
      • Dunn W.
      • Xu R.
      • Wingard D.L.
      • et al.
      Suspected nonalcoholic fatty liver disease and mortality risk in a population-based cohort study.
      In this study, we confirmed that CVD remains the number one cause of death among individuals with NAFLD. Still, most likely owing to a relatively short follow-up evaluation, we could not confirm an independent association of NAFLD with CVM at the population level. We believe, however, that longer follow-up data may provide additional information that could shed light on the possible association of NAFLD with CVD in the US population.
      The implication of these findings could be substantial. In fact, NAFLD has been associated with lipotoxicity, insulin resistance, and a low-grade inflammatory state (both in the liver and in the adipose tissue). This environment (lipotoxicity, insulin resistance, and low-grade inflammation) not only could lead to adverse liver outcomes (liver-related mortality) but also to adverse cardiac outcomes affecting the coronary vasculature and potentially the myocardial metabolism.
      • Bugianesi E.
      Nonalcoholic fatty liver disease (NAFLD) and cardiac lipotoxicity: another piece of the puzzle.
      These later processes could lead to CVD and potentially increased cardiovascular mortality. In this context, we propose that individuals with a confirmed diagnosis of NAFLD could benefit from an evaluation of their cardiovascular risks and, potentially, from regular CVD risk surveillance. In addition, treatment strategies targeting pathogenic pathways of NAFLD may improve not only the long-term liver outcomes but also long-term cardiac outcomes for patients with NAFLD. Nevertheless, these treatment strategies must be assessed for long-term efficacy and safety in well-designed randomized clinical trials.
      This was a population-based study that assessed the association of CVD/CVM with radiologically proven NAFLD. One strength of the study was that it was based on a large sample representative of the US population with complete demographic, clinical, and mortality data. Furthermore, the diagnosis of NAFLD in our study was based on hepatic ultrasound rather than increased liver enzyme levels only, which potentially could result in a less-biased and more representative sample of individuals with hepatic steatosis. The study's limitations, however, included the absence of clinical follow-up evaluation of individuals with CVD. Although there was substantial follow-up mortality data, a longer follow-up period also may be required for establishing an independent association with CVM.
      The ongoing epidemic of obesity leads to consistently increasing rates of all components of metabolic syndrome, as well as NAFLD and CVD, especially in the younger population. Specifically, since the time of the data collection in the early 1990s, the rate of NAFLD and other metabolic syndrome components in the United States have at least doubled.
      • Younossi Z.M.
      • Stepanova M.
      • Afendy M.
      • et al.
      Changes in the prevalence of the most common causes of chronic liver diseases in the United States from 1988 to 2008.
      Given the confirmed association of radiologically proven NAFLD with the higher risk of cardiovascular events, it is possible that the increasing prevalence of NAFLD, in addition to its hepatic consequences, also may lead to poorer prognosis because of higher cardiovascular risks and, therefore, will result in substantial mortality, morbidity, and resource use.

      References

        • Targher G.
        • Bertolini L.
        • Poli F.
        • et al.
        Nonalcoholic fatty liver disease and risk of future cardiovascular events among type 2 diabetic patients.
        Diabetes. 2005; 54: 3541-3546
        • Lin Y.C.
        • Lo H.M.
        • Chen J.D.
        Sonographic fatty liver, overweight and ischemic heart disease.
        World J Gastroenterol. 2005; 11: 4838-4842
        • Targher G.
        • Bertolini L.
        • Padovani R.
        • et al.
        Relations between carotid artery wall thickness and liver histology in subjects with nonalcoholic fatty liver disease.
        Diabetes Care. 2006; 29: 1325-1330
        • Targher G.
        Non-alcoholic fatty liver disease, the metabolic syndrome and the risk of cardiovascular disease: the plot thickens.
        Diabet Med. 2007; 24: 1-6
        • Lizardi-Cervera J.
        • Aguilar-Zapata D.
        Nonalcoholic fatty liver disease and its association with cardiovascular disease.
        Ann Hepatol. 2009; 8: S40-S43
        • Villanova N.
        • Moscatiello S.
        • Ramilli S.
        • et al.
        Endothelial dysfunction and cardiovascular risk profile in nonalcoholic fatty liver disease.
        Hepatology. 2005; 42: 473-480
        • Van Gaal L.F.
        • Mertens I.L.
        • De Block C.E.
        Mechanisms linking obesity with cardiovascular disease.
        Nature. 2006; 444: 875-880
        • Targher G.
        • Marra F.
        • Marchesini G.
        Increased risk of cardiovascular disease in non-alcoholic fatty liver disease: causal effect or epiphenomenon?.
        Diabetologia. 2008; 51: 1947-1953
        • Hamaguchi M.
        • Kojima T.
        • Takeda N.
        • et al.
        Nonalcoholic fatty liver disease is a novel predictor of cardiovascular disease.
        World J Gastroenterol. 2007; 13: 1579-1584
        • Rafiq N.
        • Bai C.
        • Fang Y.
        • et al.
        Long-term follow-up of patients with nonalcoholic fatty liver.
        Clin Gastroenterol Hepatol. 2009; 7: 234-238
        • Stepanova M.
        • Rafiq N.
        • Younossi Z.M.
        Components of metabolic syndrome are independent predictors of mortality in patients with chronic liver disease: a population-based study.
        Gut. 2010; 59: 1410-1415
        • Dunn W.
        • Xu R.
        • Wingard D.L.
        • et al.
        Suspected nonalcoholic fatty liver disease and mortality risk in a population-based cohort study.
        Am J Gastroenterol. 2008; 103: 2263-2271
        • Stepanova M.
        • Younossi Z.M.
        Association of chronic liver diseases and cardiovascular mortality—the authors' reply.
        Gut. 2011 Feb 15; ([Epub ahead of print])
        • Musso G.
        • Gambino R.
        • Cassader M.
        • et al.
        Meta-analysis: natural history of non-alcoholic fatty liver disease (NAFLD) and diagnostic accuracy of non-invasive tests for liver disease severity.
        Ann Med. 2010 Nov 2; ([Epub ahead of print])
        • Calori G.
        • Lattuada G.
        • Ragogna F.
        • et al.
        Fatty liver index and mortality: the Cremona study in the 15th year of follow-up.
        Hepatology. 2011; 54: 145-152
      1. NHANES 1988–1994: Hepatic/Gallbladder Ultrasound and Hepatic Steatosis Data Documentation, Codebook, and Frequencies. Available at: http://www.cdc.gov/nchs/nhanes/nhanes3/HGUHS.htm. Accessed August 18, 2011.

      2. Analytic and reporting guidelines: the Third National Health and Nutrition Examination Survey, NHANES III (1988–94). CDC/NCHS, Hyattsville, MD1996
        • Perseghin G.
        The role of non-alcoholic fatty liver disease in cardiovascular disease.
        Dig Dis. 2010; 28: 210-213
        • Sung K.C.
        • Ryan M.C.
        • Wilson A.M.
        The severity of nonalcoholic fatty liver disease is associated with increased cardiovascular risk in a large cohort of non-obese Asian subjects.
        Atherosclerosis. 2009; 203: 581-586
        • Maurantonio M.
        • Ballestri S.
        • Odoardi M.R.
        • et al.
        Treatment of atherogenic liver based on the pathogenesis of nonalcoholic fatty liver disease: a novel approach to reduce cardiovascular risk?.
        Arch Med Res. 2011 Aug 12; ([Epub ahead of print])
        • Bugianesi E.
        Nonalcoholic fatty liver disease (NAFLD) and cardiac lipotoxicity: another piece of the puzzle.
        Hepatology. 2008; 47: 2-4
        • Younossi Z.M.
        • Stepanova M.
        • Afendy M.
        • et al.
        Changes in the prevalence of the most common causes of chronic liver diseases in the United States from 1988 to 2008.
        Clin Gastroenterol Hepatol. 2011; 9: 524-530.e1