The clinical and pathophysiological connections between NAFLD, insulin resistance, obesity and T2DM have been widely explored (30). However, few studies, if any, have applied three different measures of insulin sensitivity in combination with MRI and the ELF-test for estimation of liver fat fraction and liver fibrosis, respectively, to assess the relationship between levels of hepatic steatosis and fibrosis with insulin sensitivity in patients with severe obesity and T2DM. Our data show that most (> 75%) of the patients with severe obesity and T2DM had moderate-to-high grade of liver steatosis, and there was a high proportion (80%) of patients with moderate fibrosis. With increasing LFF, we observed reduced insulin sensitivity measured by HOMA2S and Matsuda ISI. On the contrary, there was no association between LFF and intravenous glucose tolerance test-derived insulin sensitivity. Notably, the grade of fibrosis was independent of LFF, and we observed no associations between grade of fibrosis and the different measures of insulin sensitivity.
Similar to our study, a vast number of papers have reported a high prevalence of NAFLD in patients with both T2DM and obesity (1, 31). The degree of hepatic steatosis is tightly linked to obesity and patients with obesity are simultaneously at increased risk of associated medical problems such as T2DM, dyslipidemia and hypertension, a cluster of metabolic conditions influencing the susceptibility for development of NAFLD. In a Finnish study of 140 patients, of which half of the patients had T2DM, patients with T2DM had 80% more liver fat compared with those without T2DM, when matched for age, sex and weight (32). Several meta-analyses have also shown that NAFLD increases the risk the of incident T2DM (9, 33), and there is some evidence suggesting that patients with T2DM are at higher risk of developing liver fibrosis (31, 34, 35). Thus, the high proportion of patients with steatosis and moderate to severe fibrosis in our cohort of patients with severe obesity and T2DM was a predictable finding.
Unfavourable fat distribution, adipose tissue dysfunctionality and insulin resistance constitute the basis of metabolic disturbances such as NAFLD (30). Current literature suggests that the pathogenic drivers are not identical among all patients and includes multiple metabolic, genetic and microbiome related factors (22), thus the pathophysiological mechanisms behind the development of NAFLD seems to be best explained by the “multi-hit hypothesis” (36). Insulin resistance within the liver and in the extra-hepatic tissue is implicated in the pathogenesis, and recent data indicate that hyperinsulinemia could be seen as both a consequence and a cause of NAFLD (37). Hyperinsulinemia may cause increased de novo lipogenesis in the liver, and results in accumulation of ectopic fat in peripheral tissues, generating macrophage infiltration and a pro-inflammatory state that promotes insulin resistance (5, 22, 30, 38). In line with this notion, HOMA2S which reflects the balance between hepatic glucose output and insulin secretion in the basal state and is a measure of mostly hepatic insulin sensitivity (39), was moderately associated with LFF. Similarly, Matsuda ISI which also includes fasting measurements, correlated well with LFF. In contrast, MinMod SI, which includes measurements 20 minutes after rise in glucose levels and reflects mostly glucose disposal in the skeletal muscles, did not correlate with LFF. Due to the cross-sectional design of this study, no casuality between liver fat content and insulin sensitivity can be established. However, our findings indicate that LFF is primarily associated with hepatic, and not peripheral insulin sensitivity. The clear differences in insulin sensitivity, observed in our study, between patients with medium-to-high grade steatosis and patients with no or low-grade steatosis, are comparable with a study by Lomonaco and colleagues (11). Among 154 patients with obesity with or without T2DM, the suppression of free fatty acids during a euglycemic-hyperinsulinemic clamp, a measure of hepatic insulin sensitivity, was negatively correlated with intrahepatic TG content.
An interesting finding in our study was the significant difference in duration of T2DM between the groups of steatosis, with a shorter duration in the medium-to-high grade steatosis group compared with the no or low-grade group. In line with this observation, a shorter duration of T2DM was found to be an independent predictor of NAFLD in a large cohort study from Scotland (40). The association of shorter duration of T2DM and NAFLD is assumed to be caused by a greater degree of hyperinsulinemia in early TD2M as the hyperinsulinemia may drive the uptake of free fatty acids in the hepatocytes (41). In addition, long duration of T2DM may also reflect earlier diabetes debut or high age, which both indicate impaired beta cell function. It is therefore possible that patients with no or low-grade steatosis have a higher degree of hepatic insulin sensitivity, but a greater degree of beta cell dysfunction.
In patients with both severe obesity and T2DM, the co-existence of these conditions increases the risk of liver fibrosis due to the imbalance in the lipid metabolism and the formation of lipotoxic lipids, cellular stress, inflammation and cell death. Published data among patients with T2DM suggest a prevalence of biopsy proven advanced fibrosis globally of 17% (31). With non-invasive methods the prevalence is estimated to be up to 37% (34). Several studies have described association between T2DM and liver fibrosis, and to some extent, predictive factors associated with the development of fibrosis (34). A study by Petta et al. reported that the risk of severe fibrosis, discriminated by age, was driven by low HDL, impaired fasting glycemia/diabetes and obesity at lower age, while impaired fasting glycemia/diabetes and low HDL were predictors of severe fibrosis at older age (42). The major part of patients in our study were reported to have moderate fibrosis and there were no difference between the groups of fibrosis regarding HDL levels. We observed that patients with severe fibrosis were older, had longer duration of T2DM and had higher levels of AST. Both AST and ALT showed linkage to liver fat content, presenting higher values in the medium to high-grade steatosis group, but only AST demonstrated a significant difference between the different groups of fibrosis. Patients with advanced fibrosis may have normal range of liver enzymes (43). Our findings supports the study by Mansour et al. (34) showing that AST, but not ALT and the AST/ALT ratio, was positively associated with fibrosis. However, it is important to note that normal values of liver function enzymes do not rule out the possibility of a significant stage of fibrosis (43, 44).
In addition to the ELF test, which is a proprietary fibrosis panel based on extracellular matrix proteins, we also applied the non-invasive NAFLD fibrosis score for assessment of fibrosis, a score developed specifically for NAFLD considering the parameters age, hyperglycemia, BMI, platelet count, albumin, and AST/ALT ratio. The ELF test has shown to be a better tool for assessment of fibrosis stage in NAFLD compared with the NAFLD fibrosis score (45), but the NAFLD fibrosis score largely matched the results of the ELF test in our study.
Our study had some limitations. Most patients were of White ethnicity (97%), thus the results may not be generalizable to other ethnicities. There is a moderate number of participants in the study. MRI with the modified Dixon method for estimation of LFF have demonstrated reduced accuracy when there is a low degree of steatosis in the liver. LFF values may even be negative as the signal intensity in liver is normalized to spleen. Hence, our study would have benefited from liver biopsy proven validation of the LFF values, especially for those in the no or low-grade steatosis group. Nevertheless, compared to a liver biopsy, MRI is non-invasive, without risk of pain, bleeding or risk of infection. The strength in our findings regarding hepatic steatosis and insulin sensitivity is the use of several insulin sensitivity indices, recognizing that HOMA2S primarily is an insulin sensitivity-surrogate evaluating NAFLD for patients without T2D (46). The European Liver Fibrosis study group has validated the ELF score. By now it is an expensive test, but it has shown good performance in differentiating fibrosis stages in NAFLD, with a better accuracy when combined with the NAFLD fibrosis score (over 90% in distinguishing severe fibrosis) (47). However, the performance can be influenced by age and gender (48).