Background
Nonalcoholic fatty liver disease (NAFLD) is the most common liver disease worldwide, with a prevalence exceeding 15% in China [
1]. It is an acquired metabolic stress-related liver disorder due to the comprehensive effects of multiple factors such as abnormalities in glucose and lipid metabolism, insulin resistance, and inflammation. In the majority of patients, NAFLD is commonly associated with metabolic comorbidities such as obesity, diabetes mellitus, and dyslipidemia [
2]. Studies have shown that NAFLD patients, especially those with nonalcoholic steatohepatitis (NASH), are at increased risk of mortality from liver diseases (13%), and more commonly from cardiovascular diseases (25%) and malignancies (28%) [
3]. Early identification of clinical risk factors (such as biochemical indicators, imaging changes, etc.) by monitoring disease progression will help early identification of high-risk patients (NASH, cardiovascular diseases) and timely intervention.
18F-fluorodeoxyglucose (FDG) positron emission tomography (PET)/computed tomography (CT) can reflect fasting glucose metabolism in a variety of tissues and organs and is an ideal tool theoretically for screening high-risk patients for NAFLD. Thus, it is widely used in tumor, cardiovascular, and neurological fields. However, for complex metabolic disorders, it is inconclusive whether abnormalities in a variety of tissues and organs, especially in extra-hepatic glucose metabolism, are related to the progression or complications of NAFLD. Therefore, the present study aimed to investigate the association between hepatic steatosis and glucose metabolic abnormality in a variety of organs/tissues assessed by using
18F-FDG PET/CT in a population-based cohort.
Discussion
Our study indicated that NAFLD patients are usually accompanied by obesity, diabetes, insulin resistance, dyslipidemia and hypothyroidism, etc., which is now regarded as a comprehensive hepatic manifestation of metabolic syndrome [
2]. By definition, hepatic steatosis is the component and hallmark of NAFLD. We found that in NAFLD patients, hepatic steatosis is independently associated with elevated hepatic enzymes, increased VAT volume, and decreased myocardial FDG uptake, but not with hepatic FDG uptake.
18F-FDG is an emerging widely used radiopharmaceutical agent for imaging inflammation because activated inflammatory cells display increased FDG accumulation. Some studies [
18,
19] have found that hepatic
18F-FDG uptake is significantly higher in NAFLD patients than the controls, suggesting that the development mechanism of NAFLD may be related to non-specific uptake of
18F-FDG by NASH inflammatory cells. However, other studies [
20,
21] found that there was no significant correlation between NAFLD and
18F-FDG uptake, which is consistent with our findings that there was no significant difference in hepatic
18F-FDG uptake between NAFLD patients with abnormal liver function (ALT ≥50 u/L) and NAFLD patients with normal liver function (
P > 0.05). The theoretical basis behind this finding may be 1) a potential ‘dilutional’ effect of hepatic fat on the FDG signal [
18]; 2) most NAFLD patients are accompanied with insulin resistance, which could lead to decreased hepatic glycogen synthesis and enhanced glycogen decomposition; and 3) NASH not only expresses as inflammatory cell infiltration but also is accompanied by hepatocyte injury and collagen fiber deposition [
22], resulting in a decrease in
18F-FDG uptake, which compensate the effect of non-specific uptake of
18F-FDG by hepatic inflammatory cells.
Over the past 10 years, it has also become increasingly clear that NAFLD is a multisystem disease that affects a variety of extra-hepatic organs [
23]. In this study, we for the first time utilized PET/CT to observe the differences in glucose metabolism of different tissues/organs between NAFLD patients and non-NAFLD subjects at fasting state and found that NAFLD patients had decreased myocardial glucose metabolism and slightly increased mediastinal blood pool and spleen SUVmean. In addition, we also found that decreased myocardial glucose metabolism was a risk factor for NAFLD (OR = 0.497,
P < 0.05) independent of BMI, diabetes, dyslipidemia, and insulin resistance, consistent with previous research results [
11,
21‐
23]. But the mechanism remains unknown. Some studies showed that individuals with NAFLD and decreased myocardial glucose uptake on FDG PET had higher risk of left ventricular diastolic dysfunction [
21,
24,
25]. Tang et al. reported that the low myocardial SUV was independently associated with NAFLD, non-calcified plaque, and significant coronary stenosis [
26].
We thought that the existence of systemic insulin resistance in NAFLD patients was the potential mechanism of decreased myocardial glucose metabolism. In this study, a slight but significant increase in blood pool SUVmean in NAFLD patients is indirect evidence of peripheral tissue insulin resistance and glucose metabolism utilization disorders. At the insulin resistance state, the serum concentration of free fatty acids (FFAs) elevated [
3], which was another myocardial energy substrate. The elevated FFAs can inhibit pyruvate dehydrogenase (PDH) and glucokinase, which then led to the inhibition of myocardial glucose oxygenation [
27]. Abnormal myocardial energy substrate utilization and altered metabolic pathways will lead to “metabolic remodeling” and subsequent cardiac structure and dysfunction [
21]. Perseghin et al. [
28] found in their study using cardiac
31P-MR spectroscopy that NAFLD patients had reduced myocardial phosphocreatine/adenosine triphosphate ratio, suggesting that NAFLD patients have myocardial energy metabolism disorder. Our study further verified this view from the perspective of myocardial energy substrate utilization and found that there is a mild linear correlation between hepatic steatosis and FDG uptake in the myocardium. We thought that this might be an early manifestation of cardiac metabolic remodeling in patients with NAFLD, and cardiac metabolic remodeling may take place precede the development of functional and structural remodeling of the heart [
29]. The current evidence from the published prospective studies supports that NAFLD, irrespective of its diagnostic methodology, was significantly associated with an increased risk of fatal and nonfatal cardiovascular events [
30]. But direct evidence of the correlation between myocardial glucose reduction and cardiovascular risk remains to be confirmed.
Our study once again confirmed that VAT volume and serum ALT are independently associated with NAFLD. Although the median ALT value of NAFLD patients was within the normal range, the degree of hepatic fat infiltration was significantly higher in patients with ALT ≥50 u/L than in patients with ALT < 50 u/L (
P < 0.05). As we know, ALT is a serum marker for liver inflammation or injury and frequently elevated in NAFLD patients. But persistently elevated ALT levels may be the risk factor for the progression of NAFLD [
3]. VAT surrounds the abdominal organs in the abdominal cavity. Its volume increase is a risk factor for abdominal obesity, insulin resistance and other important cardiovascular diseases and also plays an important role in hepatic steatosis, inflammation and fibrosis. Excessive visceral fat accumulation could release various bioactive adipocytokines [
31], which are prone to induce chronic low-grade inflammation and macrophage accumulation in VAT. In addition, infiltrated macrophages could produce pro-inflammatory cytokines and nitric oxide, leading to adipocytokine dysregulation [
32]. Decreased adiponectin level is associated with NAFLD [
32]. Although we found that BMI, abdominal fat volume and SAT volume were significantly higher in NAFLD patients than in non-NAFLD subjects (
P < 0.05), they were not independent correlation factors for NAFLD, suggesting that increased VAT volume is more clinically significant for NAFLD.
Limitations
The current study also has several limitations. First, this study is a retrospective study. NAFLD was diagnosed using CT imaging and liver degeneration was not confirmed by liver biopsy. Meanwhile, although CT is far more convenient than MR, it is not sensitive to low fat concentration and less sensitive than MR-based techniques for diagnosis [
33]. Second, there were only 33 NAFLD patients, which accounted for 17.3% of the study population. This ratio is close to the NAFLD prevalence (15%) in the community of China [
1]. Third, PET myocardial dynamic imaging after euglycemic-hyperinsulinemic clamp is the “gold standard” for determination of myocardial insulin resistance, but the imaging process is complicated and time-consuming. Previous studies have indicated that fasting myocardial glucose uptake is correlated with insulin resistance [
34], but the pathophysiological mechanism underlying the correlation between myocardial glucose uptake and NAFLD and the prognosis of NAFLD still need to be explored prospectively and confirmed by long-term follow-up.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.