New insights into the link between NAFLD and Atherosclerotic Coronary Heart Disease?

Nonalcoholic fatty liver disease (NAFLD) is a systemic disorder that has been associated with a variety of chronic conditions including diabetes, renal disease and most significantly, cardiovascular diseases[1]. Mounting data now support a strong link between NAFLD and in particular, atherosclerotic cardiovascular disease (ASCVD) independent of known traditional risk factors[1–3]. The importance of this association is underscored by the observation that cardiovascular disease, in general, is the leading cause of death in individuals with NAFLD[4].

In the current issue of Journal of Hepatology, Lee and colleagues report on the cross-sectional association between prevalent NAFLD and prevalent subclinical coronary atherosclerosis from a large single center study of 5,121 consecutive asymptomatic participants who self-referred for coronary CT angiography (CCTA) evaluation between 2007 and 2011 as part of a government-issued general health examination in Korea. The main finding is that NAFLD was independently associated specifically with non-calcified (vs. calcified) atherosclerotic plaque (adjusted odds ratio 1.27, 95% confidence interval 1.08–1.48), which in clinical populations of patients without liver disease is highly associated with future ASCVD events[5].

There is a large existing body of literature on the relationship between NAFLD and subclinical atherosclerosis. NAFLD has been associated with impaired endothelial function, a higher prevalence of vulnerable coronary artery plaques, and atherosclerosis as measured by carotid intima media thickness (CIMT) and coronary artery calcification (CAC) independent of multiple cardio-metabolic risk factors[6]. However, the study by Lee and colleagues is the largest to date to report an association between NAFLD and severity of atherosclerosis.

A variety of invasive and non-invasive techniques are available to measure subclinical ASCVD[5]. These techniques can ascertain parameters such as luminal diameter or stenosis, vessel wall thickness, plaque volume, and the specific distribution and localization of atherosclerotic disease. One method of detection for subclinical coronary atherosclerosis is to perform chest CT, which allows the detection and quantification of CAC and can help define the burden of atherosclerosis in individuals before they develop clinical events. CAC presence has consistently been shown to predict clinical ASCVD events (e.g., heart attack, stroke) in both men and women across a variety of racial-ethnic groups and is notably a stronger predictor of incident ASCVD events than carotid ultrasound measures[5]. Another approach in measuring subclinical coronary atherosclerosis is to perform coronary CT angiography (CCTA), as done in the current study, which allows not only for detection of CAC in general, but also has the ability to identify multiple high-risk features of coronary plaque (such as non-calcified plaque and large plaque burden, positive remodeling, and low attenuation). These “vulnerable plaques” are thought to be at highest increased risk for a subsequent acute ASCVD event[7].

The presence of any CAC, which indicates that at least some atherosclerotic plaque is present, is defined by an Agatston score >0. However, an Agatston score < 10 is consonant with the absence of a significant coronary obstructive lesion (likelihood < 10%). Clinically significant plaque, frequently an indication for more aggressive risk factor management, is often defined by an Agatston score ≥100. An Agatston score ≥400 has been noted to be an indication for further diagnostic evaluation (e.g., exercise testing) for coronary artery disease. According to both the American and European guidelines on cardiovascular disease prevention, when treatment decisions are uncertain after 10-year risk is estimated, then the patient and clinician should take into consideration additional factors that modify the risk estimate, including an elevated CAC score[8, 9].

In the current study, the prevalence of CAC (defined as CAC>10) was 33.9%, although the prevalence of clinically significant CAC (e.g., ≥ 100) was low (9.3%). NAFLD was associated with prevalent CAC and also with any atherosclerotic, calcified and mixed plaques on univariate analysis. However, these associations were no longer significant in multivariable analyses controlled for demographics, ASCVD risk factors and high sensitivity C-reactive protein. There have been conflicting results on the relationship between NAFLD and CAC depending on the population studied and the covariates adjusted for[10]. As in the current study, most of the published studies have been conducted in Asian populations and all are cross-sectional, thus causality cannot be inferred. In addition, the adjustment for established cardiovascular risk factors, metabolic risk factors and other potential confounders have often been incomplete. Lee and colleagues chose to adjust for prevalent comorbidities (e.g., obesity, hypertension, diabetes, hyperlipidemia) rather than for continuous markers of these conditions (e.g., body mass index, systolic blood pressure, total cholesterol, fasting glucose, etc.). Importantly, there may be differential associations between NAFLD and CAC when risk factors are controlled versus uncontrolled, particularly among participants who may have been receiving statin medications which might lead to atherosclerotic plaque remodeling. Unfortunately, the proportion of participants who were on statin medications was not available in this study.

In contrast to the null findings between NAFLD and CAC, NAFLD was consistently associated with non-calcified plaque, which as mentioned previously has the highest risk of rupture and future ASCVD events in clinical populations. Puchner et al. recently reported an independent association between CT-diagnosed NAFLD and high-risk coronary plaque in a post-hoc analysis of subjects with chest pain who were randomized to the CCTA arm of the Rule Out Myocardial Infarction using Computer Assisted Tomography (ROMICAT) II trial[7]. Unfortunately, a main limitation of the study by Lee et al. is a lack of clinical outcome data to assess whether or not high risk coronary plaque detected in asymptomatic individuals with NAFLD actually associates with hard ASCVD clinical events. It’s also important to highlight that compared with the established value of CAC scanning for risk reclassification in asymptomatic patients, there are limited data regarding the utility of CCTA in asymptomatic people and current guidelines do not recommend its use as a screening tool for assessment of cardiovascular risk in asymptomatic people[11].

Lee et al. used both ultrasound and the fatty liver index to diagnose NAFLD and associations were consistent regardless of the method used, adding additional validity to their findings. However, ultrasound has overall low sensitivity for diagnosis of NAFLD and can only detect fat when at least 30% of the liver parenchyma is involved[12]. A helpful addition to this manuscript would have been to demonstrate a dose-response relationship between severity of hepatic steatosis and ASCVD. Very few large studies have assessed the differential effect of degree of NAFLD severity on risk of ASCVD. Lee et al. attempted to assess NAFLD severity through use of the NAFLD fibrosis score (NFS)[13]. A low NFS (<−1.455) strongly suggests the absence of liver fibrosis. In multivariable analyses, elevated NFS (≥ −1.455) was independently associated with non-calcified plaque, but not with calcified plaque. Notably, there was no significant difference in NFS between NAFLD and non-NAFLD participants. This may be due to misclassification of a participant as non-NAFLD due to the low sensitivity of ultrasound and/or due to the low overall population prevalence of advanced fibrosis in this study. Thus, we cannot determine from the current study whether or not NAFLD severity is associated with differential risk for high-risk coronary lesions.

Patients with NAFLD have numerous established risk factors for ASCVD including insulin resistance, hypertension, atherogenic dyslipidemia, obesity and chronic kidney disease[1, 6]. Furthermore, NAFLD is associated with many nontraditional and emerging ASCVD risk factors including increased serum levels of uric acid[14], pro-inflammatory markers such as C-reactive protein and IL-6[15], and pro-coagulant factors including fibrinogen, von Willebrand factor and plasminogen activator inhibitor-1[16]. NAFLD is also associated with lower plasma levels of adiponectin, a protein with anti-atherogenic and anti-diabetic properties[17]. Attenuation of all observed associations after adjustment for comorbidities and C-reactive protein in the current study supports the hypothesis that at least some of the association between NAFLD and ASCVD is mediated by a variety of diverse putative mechanisms that are highly related to the metabolic syndrome.

In summary, although there is an established association of NAFLD with ASCVD, whether or not NAFLD actually causes ASCVD remains to be determined. In fact, a recent Mendelian randomization and meta-analysis of 279,013 individuals of Danish descent did not find an association between lifelong genetically high liver fat content (e.g., PNPLA3) and ischemic heart disease suggesting that the epidemiologic association between NAFLD and at least ischemic heart disease, may be due to confounding or reverse causation[18]. Regardless, the findings by Lee et al. provide important insight into the potential link between NAFLD and ASCVD by demonstrating that NAFLD is associated not just with any atherosclerosis, but with more advanced coronary atherosclerotic plaque. These findings must be interpreted with caution given the cross-sectional study design, the single racial-ethnic population studied, and the lack of a spectrum of NAFLD disease severity present in this asymptomatic population. Future high quality prospective studies in diverse racial-ethnic groups with NAFLD are needed to further elucidate whether or not NAFLD is in fact causally related to ASCVD, and whether or not addition of CCTA findings (or CAC) to traditional ASCVD risk scores actually meaningfully reclassifies risk for ASCVD events in the already high-risk NAFLD population.