Evaluation of NAFLD fibrosis, FIB-4 and APRI score in diabetic patients receiving exenatide treatment for non-alcoholic fatty liver disease

There is a closely relationship between the development and progression of nonalcoholic fatty liver disease (NAFLD) or metabolic associated fatty liver disease (MAFLD) and obesity and diabetes. NAFLD fibrosis scores should be routinely used to rule out patients with advanced fibrosis. High scores may help identify patients at higher risk of all causes andliverrelated morbidity and mortality. The aim of this study was to investigate the association between exenatide and fibrosis scores. The effect of exenatide treatment on fibrosis scores was evaluated in type 2 diabetes mellitus (DM) patients with MAFLD. Evaluation was made of 50 patients with type 2 DM and MAFLD. The NFS, FIB4 and APRI scores were calculated before and after 6 months of treatment. After 6 months of exenatide treatment, the NFS and APRI scores were determined to have decreased significantly. Exenatide was observed to control blood glucose, reduce body weight and improve fibrosis scores in MAFLD patients with type 2 diabetes.


Materials and methods
Patient selection. This retrospective study included 70 patients who were diagnosed with type 2 DM and treated with exenatide plus metformin at our center between January 2017 and October 2019. All patients included in the study had body mass index (BMI) > 35 kg/m 2 , and were receiving metformin treatment agents because of the national healthcare system reimbursement conditions for exenatide in Turkey. NAFLD diagnosis was made from ultrasonographic findings. A total of 20 patients were excluded from the study; 10 with a change of treatment in the last 6 months and no ultrasonographic imaging before exenatide treatment, and 10 with drug intolerance due to side-effects ( Fig. 1). All the subjects included in the study were prescribed a standard 1400 kcal diabetic diet in addition to 150 min/week mild-moderate exercise, and follow-up examinations were made at 3-month intervals. The initial exenatide treatment was at a dose of 5 μg twice a day, and this was increased to 10 μg twice daily after 4 weeks.
Clinical and biochemical measurements. A record was made for all the study subjects of baseline demographic data, medications and clinical characteristics. Baseline and final follow-up values of weight, height, and BMI were recorded. The biochemical examination applied to all patients before and after the treatment included glycated haemoglobin (HbA1c), fasting plasma glucose (FPG), post-prandial glucose (PPG), creatinine, lipid profile, aspartate aminotransferase (AST), alanine aminotransferase (ALT), gamma-glutamyltransferase (GGT),and alkaline phosphatase (ALP). Blood samples were taken between 8:00 and 11:00 a.m. from each patient after a 10 h fasting. HbA1c was measured using the high-performance liquid chromatography (HPLC) method. Reference ranges were defined as FPG:74-100 mg/dl, AST: 0-40 U/l, ALT: 0-41 U/l, ALP: 40-

Results
Evaluation was made of a total of 50 type 2 DM patients with NASH who had received exenatide treatment for six months. The patients comprised 39 (78%) females and 11 (22%) males with a mean age of 52.9 ± 9.8 years.
The mean duration of diabetes was 10 ± 7.4 years. All the patients received metformin therapy, and 24 (48%) patients were using insulin therapy in addition to metformin. Medication for hypertension was being taken by 31 patients and their blood pressure was controlled with this medication. Abdominal ultrasonography (USG) was applied to all patients before starting exenatide treatment. The USG imaging revealed grade 1 hepatosteatosis (HS) in 5 (10%) patients, grade 2 HS in 34 (68%) patients, and grade 3 HS in 11 (%22) patients. The full demographic and baseline clinical data of the patients are reported in Table 1. There was a mean reduction of 8 kg in weight, and 3 kg/m 2 in BMI in the 6th month. A significant decrease from the baseline values was determined for HbA1c (from 9.3 to 8%, p < 0.001), FPG (from 202 to 162 mg/dl, p < 0.001) and PPG (from 284 to 217 mg/dl, p = 0.001). After 6 months of exenatide treatment, the mean serum levels of triglyceride decreased significantly, but there was no significant change in HDL-C, LDL-C and total cholesterol levels. A significant decrease was determined in mean AST (from 25.6 to 21 U/l, p = 0.02) and ALT (from 29.4 to 21.7 U/l, p = 0.001) levels in the 6th month. GGT levels decreased significantly after exenatide treatment, and there was no significant change in the ALP level. NFS, FIB4 and APRI scores were calculated before and after 6 months of treatment. A significant decrease was determined in NFS (from − 0.007 to − 0.22, p = 0.002) and APRI score (from 0.23 to 0.20, p = 0.02) after 6 months of treatment and there was no significant change in FIB-4 score ( Table 2). According to the NFS score, 10% (n:5) of patients were low-risk, 66% (n:33) were intermediate-risk, and 24% (n:12) were at high-risk  (Table 3). In the analysis after dividing the patients according to gender, there was a significant difference in serum ALT level, BMI, NFS and APRI scores in female after six months of treatment (Table 4). When patients were divided into 2 groups as HbA1c < 9% and > 9%, the difference was significant in terms of reduction in HbA1c, BMI, serum ALT level and NFS score in both groups after 6 months of treatment (Table 5). Multivariate linear regression analysis was performed for further analysis of whether weight loss or the changes in glycemic parameters were factors that might predict changes in fibrosis scores. There was no correlation between the in ΔNFS and ΔBMI (r:0.225, p:0.208) or ΔHbA1c (r:-0.108 p:0.535) values. No correlation was determined between the in ΔAPRI and ΔBMI (r:0.185, p:0.218) or ΔHbA1c (r:-0.004 p:0.978) values. Weight loss and the changes in glycemic parameters observed with exenatide treatment did not predict change in fibrosis scores (Table 6).

Discussion
NAFLD is a major cause of end-stage liver disease, HCC, and liver transplantation throughout the world. The pathogenesis of non-alcoholic fatty liver disease has not been exactly explained, but insulin resistance is known to have a key role in the development of fatty liver disease and potentially, steatohepatitis 29. . Significantly elevated free fatty acids levels have been reported in NAFLD and type 2 diabetes mellitus cases, compared ro those with type 2 diabetes without NAFLD 30 . Exenatide has been shown to improve hepatic steatosis by increasing lipid use in adipose tissue in non-diabetic rats 31 . Although the mechanism is still not understood by which GLP-1 analogs improve hepatic steatosis, alterations in hepatic lipid metabolism have been thought to be a contributory factor to these effects 32 . Irrespective of insulin resistance, visceral fat has also been related to liver inflammation and fibrosis in NASH patients, and it is thought that this effect may be mediated by interleukin-6 33 . Hepatic lipid metabolism is affected by intrahepatic lipogenesis and lipolysis, and changes in the use of lipids in skeletal muscle and adipose tissue. Conventional ultrasonography is low cost, safe, and accessible, and the most commonly used imaging technique for screening for fatty liver 34 . In the current study, USG imaging revealed grade 1 hepatosteatosis in 5 (10%) patients, grade 2 in 34 (68%) patients, and grade 3 in 11 (%22) patients. A meta-analysis showed that liver ultrasonography allows for reliable and accurate detection of moderate-severe fatty liver compared to histology with sensitivity and specificity of 84.8% and 93.6%, respectively 35 . Moreover, European guidelines recommend using ultrasonography as first-choice imaging for identify risk of NAFLD in adults 3 .
In the management of NAFLD, it is important to identify patients with a higher risk of NASH and advanced fibrosis. Liver biopsy is still required to identify fibrosis status and patients with NASH. However, liver biopsy is invasive, costly, less suitable for population-level screening, and shows inter-observer variability 28 . Noninvasive evaluation of fibrosis severity is important in management of NAFLD patients, because of the fibrosis stage is a determinant of mortality. Therefore, to identify NASH in patients with NAFLD, various clinical and  www.nature.com/scientificreports/ biochemical models have been used.There has been shown to be an association between aminotransferases, and some histological parameters such as inflammation and steatosis 36 . Patients with NAFLD may have mild or moderate elevations in AST and ALT, although normal aminotransferase levels do not exclude NAFLD. Elevated AST and ALT values have been shown to usually be 2-5 times the upper limit of normal, giving an AST to ALT ratio of < 1, which differs from alcoholic fatty liver disease. However, the extent of liver transaminases elevation does not predict the extent of hepatic inflammation or fibrosis, and an ALTvalue within normal limits does not exclude histological injury of clinical importance. Thus, aminotransferase levels do not correlate with the degree of fibrosis [37][38][39][40] . In the current study patients, the initial AST and ALT values were within normal limits (25.6 U/l, 29.4 U/l, respectively). However, there was a significant decrease in mean AST and ALT levels in the 6th month of exenatide treatment. Buse et al. demonstrated that 2 years of exenatide therapy was associated with significant improvement in aminotransferase levels 41 . In previous studies, new normal ALT cut-off values (30 U/l for males and19 U/l for females) have been found to be associated with NASH. In the current study, 78% of the patients were female. In this case, the initial ALT level is at an acceptable level above normal. ALP may be elevated to two to three times the upper limit of normal. In the current study, ALP was normal and there was no significant change in the ALP level after exenatide treatment. GGT levels were observed to decrease significantly after exenatide treatment. Studies have shown that the serum GGT level in particular is a marker of oxidative stress rather than a specific marker of NAFLD-related liver disease. Although the key mechanism leading to NAFLD and perhaps NASH is insulin resistance, additional oxidative injury is required for the necrosis and inflammatory process in steatohepatitis 42 . In a retrospective case-series study, once weekly dulaglutide (0.75 mg) administration for 12 weeks was able to reduced serum transaminases levels and liver stiffness. 29 weeks treatment with lixisenatide increased the proportion of patients with normalisation of ALT in overweight/obese NAFLD subjects. A small randomized clinical trials showed that, liraglutide (3 mg daily) administration for 26 weeks reduced serum ALT and AST on obese NAFLD subjects without T2DM. According to studies using different formulations of GLP-1 analogues on NAFLD and NAHS obtained similar efficacy, a class effect on improving liver function 43 . After the 12-week treatment, ALT level (< 40 U/l) was in the exenatide group was significantly lower than in the metformin group in a study 25 . In the other study, levels of the hepatic injury biomarkers ALT, AST and GGT were significantly decreased in the exenatide group versus ınsulin group after 12 weeks. Exenatide can reduce apoptosis, oxidative stress and endoplasmic reticulum stress response and act directly on the liver, protecting it from various stress responses 24 . GLP-1 RA has significant effects on glycemic control and weight loss. LEAN trial demonstrated that patients who used liraglutide underwent end-of-treatment liver biopsy had histological resolution in non-alcoholic steatohepatitis compared with placebo group 44 . In the phase 2 study, semaglutide at a daily dose of 0.4 mg significantly induced resolution of NASH without worsening of fibrosis after 72 weeks of treatment, but it did not induce an improvement of fibrosis stage > 1 45,46 .
For most patients with NAFLD, weight loss is the primary therapy. In a previous study, it was reported that steatosis was improved in 10 (20.4%) of 49 NAFLD patients with type 2 DM through treatment with exenatide plus modifications to lifestyle, although the rate of improvement showed no difference from that observed with treatment of metformin plus lifestyle modification 25 . In that study, BMI was < 30 kg/m 2 . In the current study, although BMI was > 40 kg/m 2 after treatment, a decrease was observed in fibrosis scores with the use of exenatide. In another study of 30 NAFLD patients with elevated liver enzymes, obesity and type 2 DM, exenatide plus insulin was seen to improve steatosis from baseline in 28 (93.3%) patients, which was a significantly higher improvement rate than that observed with intensive insulin therapy 24 . Therefore, although initial therapy for type 2 DM is typically with metformin, the presence of NAFLD can be decisive in the choice of glucose lowering therapy.
Non-invasive fibrosis scores help distinguish patients with NAFLD who are at increased risk of liver-related complications or death. There are now several non-invasive methods to detect fibrosis in patients with liver disease. One of the scores, the NFS, is specific to NAFLD. Studies have suggested that higher NFS may be associated with increased mortality from cardiovascular disease 47 . In some studies have been shown that the Fib-4 score have good predictive accuracy for advanced fibrosis in patients with chronic hepatitis C virus (HCV) infection 48 . In another study, it performed better than other serological markers for predicting advanced fibrosis in patients with NAFLD. The APRI has firstly been evaluated in patients with HCV and human immunodeficiency virus or alcoholic liver disease. The ability of the APRI to predict outcomes in patients with NAFLD was examined in a retrospective series with 320 patients and it was determined that only the high-risk group was at greater risk of death or liver transplantation 49 .
The results of this study showed that the NFS, and APRI score decreased significantly after 6 months of exenatide treatment, but there was no significant change in FIB-4 score. According to the NFS score, the number of high and intermediate-risk patients decreased significantly compared to the baseline values. These non-invasive scoring systems, including NFS, FIB-4 index, and APRI index yield high negative predictive values but poor positive predictive values. Although NFS and FIB-4 display good diagnostic efficacy, many factors such as age, diabetes, may influence their diagnostic performance. Previous studies have shown that high fibrosis scores are associated with an increased risk of cardiovascular and liver-related mortality. Therefore, these scoring systems are used to exclude subjects without advanced fibrosis, thereby prevented unnecessary liver biopsies 28 .
In this clinical trial, there was seen to be a mean reduction of 8 kg in weight, and 3 kg/m 2 in BMI after 6 months of exenatide treatment. The GLP-1 analogue may lead to improvements in fatty liver and insulin resistance. Although exenatide may not directly an insulin sensitizer agent, it can resulting in an insulin-sensitizing effect by significant weight loss. One possible hypothesis to explain this is that GLP-1 may have a direct effect on hepatocytes and affect the expression of genes such as PPAR-α and PPAR-gene connected to β-oxidation and insulin sensitivity of fatty acids 25 . HbA1c, FPG, PPG, and the mean serum levels of triglyceride significantly decreased from the baseline values after 6 months. In a study which a non-diabetic, HFD-induced rat NAFLD model was used the researchers showed that exenatide reduced both lipid deposition in the liver and adipose www.nature.com/scientificreports/ tissue and decreased the size of adipocytes 31 . Tushuizen et al. examined the effect of exenatide treatment on hepatosteatosis with Proton Magnetic Resonance Spectroscopy, and reported that after 44 weeks of treatment was reduction in a liver fat from 16.0 to 5.4% 50 . In another case series of 8 patients with type 2 DM and biopsyproven NAFLD, after 28 weeks of exenatide therapy, the liver histology improved in 3 patients 23 .
The main limitation of this study was the retrospective design, unequal number of patients by gender, and that the liver biopsy could not be evaluated. Although there are studies of exenatide and NAFLD in literature, this is the first study to have evaluated the relationship between exenatide treatment and fibrosis scores in diabetic and obese patients.
In conclusion, exenatide may not only reduce blood glucose and body weight, but also improve fibrosis scores, reduce aminotransferase, and extenuate NAFLD in diabetic patients. The improvement of liver histology is the gold standard in the evaluation of therapeutic efficacy in patients with NAFLD. Therefore, it should be supported by studies with larger sample sizes in patients with biopsy-proven NASH.