A close association between NAFLD and obesity has consistently been reported (3, 19). In this context, most subjects with NAFLD are overweight/obese and have varying components of metabolic syndrome (4, 34). Nonetheless, a significant proportion of subjects with NAFLD are lean, although the clinical implications are unclear (35). Accordingly, using nationwide, population-based, cross-sectional data, we investigated whether ASCVD risk differs according to obesity in patients with NAFLD and checked whether similar results can be reproduced in subgroups with significant liver fibrosis.
Subjects with lean NAFLD had a significantly higher ASCVD score and a greater chance of having high ASCVD risk than those with obese NAFLD. Hypertension, DM, hypo-HDL cholesterolaemia, hypertriglyceridemia, and albuminuria significantly increased the risk of ASCVD in lean compared to obese subjects with NAFLD. After appropriate adjustment, lean subjects with NAFLD had a significantly higher ASCVD risk than obese subjects with or without NAFLD. Furthermore, after adjustment, lean subjects with significant liver fibrosis had the highest ASCVD risk, followed by obese subjects with significant liver fibrosis and those without significant liver fibrosis.
This study had several strengths. First, the selected cohort was large (n > 4,000), ensuring statistical reliability and robust results. Furthermore, the proportion of lean subjects with NAFLD in the cohort (11.0%) was similar to that in previous Asian studies (9–23.5%) (13, 36). This suggests that the subjects were selected appropriately and that the results are applicable to other Asian populations, although further validation is required for other ethnic groups. In addition, the prevalence of subjects with NAFLD (37.6%) was similar to that in a recent Asian study (37). Therefore, our study population, a nationwide representative cohort, was selected appropriately based on non-invasive surrogates for statistical analyses.
Second, to the best of our knowledge, this is the first report of an independent association between ASCVD risk and NAFLD according to obesity status. NAFLD and obesity are significant risk factors for cardiovascular events (17–19). However, after controlling for important confounders, lean NAFLD was associated with a higher ASCVD score and an increased prevalence of a high ASCVD risk compared to obese NAFLD; moreover, this finding was reproduced in the subgroup with significant liver fibrosis. In addition, our results provide evidence on the prognosis of lean NAFLD. Unlike the prognosis and complications of overall NAFLD, the data on lean NAFLD are scarce. Two studies have reported a higher mortality rate for lean compared to obese NAFLD (38, 39). Moreover, cardiovascular disease accounted for approximately one quarter of the causes of death (38). In our study, subjects with lean NAFLD were at markedly increased risk for individual cardiometabolic components and ASCVD compared to those with obese NAFLD. These data indicate that lean subjects with NAFLD should be counselled about the risk for unfavourable cardiovascular outcomes and managed accordingly (40).
Third, the influence of NAFLD on ASCVD might have been biased because most subjects with NAFLD had simple hepatic steatosis, which has a very favourable prognosis. Thus, we selected subjects with NAFLD to determine whether the co-existence of significant liver fibrosis is significantly associated with ASCVD risk, because liver fibrosis can be considered a sequela of the inflammatory process of NAFLD. In addition, liver fibrosis is the single most important factor as well as a clinically relevant issue that correlates with poor outcomes. In the subgroup with NAFLD, around 20% of subjects had significant liver fibrosis, indicating that lean subjects had higher ASCVD risk than obese subjects among those with NAFLD and significant fibrosis.
We are also aware of several unresolved issues that should be addressed. First, although we used well-validated liver fibrosis and steatosis prediction models, liver imaging and histological information was not available because of the high cost of ultrasonography and ethical concerns regarding screening of a large national population-based cohort. KNHANES participants who gave informed consent underwent only serum tests. In addition, because cancer diagnosis in the KNHANSE was based on a questionnaire, not ultrasonography, subjects with HCC were excluded.
Second, because of the cross-sectional nature of the study, we could not assess the longitudinal dynamic association between changes in NAFLD, obesity status, and ASCVD risk. We were also unable to assess the effects of therapeutic interventions, such as lifestyle modification, exercise, weight loss, medications, nutritional support, and protein supplements, on NAFLD, obesity, and ASCVD risk. Nevertheless, our results reveal the need to screen patients with NAFLD, particularly lean patients, to identify those at high risk for ASCVD requiring intensive medical therapy.
Third, we used a pooled cohort risk equation to assess ASCVD risk and did not examine the risk for real clinical events during follow-up. The 10-year ASCVD risk with primary prevention is estimated in the blood cholesterol guidelines of the ACC/AHA. However, because the equation for calculating ASCVD risk might have been in the study population, our findings should be interpreted with caution.
Finally, several serum markers, such as HbA1c, were available for only a small proportion of the subjects. Thus, the incremental influence on the final results, not simply the presence of DM, could not be assessed. In addition, detailed information regarding anti-hypertensive and anti-diabetic drugs was not available, preventing analyses of their influence. Furthermore, although we excluded subjects known to have chronic liver diseases (such as viral hepatitis and alcoholic liver disease), those with other types of such diseases (such as Wilson disease, autoimmune liver disease, or primary biliary hepatopathy) might have been included, which may have biased the results. Moreover, as a result of the limitations of general medical examination, we could not adjust for dietary preference and genetic risk factors (including palatin-like phospholipase domain-containing 3 and transmembrane 6 superfamily member 2), which could affect the risk for lean NAFLD (42).
Although the characteristics and underlying pathophysiological mechanism of lean NAFLD/NASH are unclear, individuals with a low BMI are likely to have reduced lean body mass, particularly muscle mass, which could lead to unfavourable traits related to cardiovascular outcomes (43). In addition, fat tissue in a distinct depot may have protective functions, particularly in people with chronic diseases (44). Subcutaneous fat can act as a metabolic reservoir, protecting other organs from lipotoxicity and ectopic fat formation, and leg fat is associated with lower ACD risk and cardiometabolic risk factors (45).