High Prevalence of Severe Hepatic Fibrosis in Type 2 Diabetic Outpatients Screened for Non-Alcoholic Fatty Liver Disease

Background: Non-alcoholic fatty liver disease (NAFLD) is a highly frequent condition in patients with type 2 diabetes (T2D), but the identification of subjects at higher risk of developing the more severe forms remains elusive in clinical practice. The aim of this study was to evaluate the occurrence and severity of liver fibrosis and its predictive factors in T2D outpatients without a known history of chronic liver disease by using recommended non-invasive methods. Methods: Consecutive T2D outpatients underwent a set of measurements of clinical and laboratory parameters, FIB-4 score (Fibrosis-4 index), and liver stiffness with controlled attenuation-parameter (CAP) performed by transient elastography (FibroScan) after excluding previous causes of liver disease. Results: Among the 205 T2D outpatients enrolled in the study (median age: 64 years, diabetes duration: 11 years, HbA1c: 7.4%, and BMI: 29.6 kg/m2), 54% had high ALT and/or AST levels, 15.6% had liver stiffness value > 10.1 kPa (severe fibrosis), 55.1% had CAP values > 290 dB/m (severe steatosis), and FIB-4 score was >2 in 11.2% of subjects (>2.67 in 15 subjects). Moreover, 49 (23.9%) T2D patients had clinically meaningful liver harm, with either a FIB-4 score > 2 and/or FibroScan > 10.1 kPa. At regression analysis, BMI, HbA1c, creatinine, and triglycerides values were independent predictors of liver fibrosis. Conclusions: Liver fibrosis is a frequent finding in T2D outpatients without a known history of liver disease, especially in those with obesity, hypertriglyceridemia, worse glycemic control, and high creatinine levels.


Introduction
The term non-alcoholic fatty liver disease (NAFLD) identifies a broad spectrum of clinical conditions ranging from hepatic steatosis to non-alcoholic steatohepatitis (NASH) [1]. Its prevalence in the general population is about 25% [2], and, in recent decades, it has become a major cause of end-stage liver disease, including hepatocarcinoma (HCC) and liver transplantation [3].
NAFLD is so frequently associated with extra-hepatic metabolic disorders, including obesity and type 2 diabetes (T2D) [4], that it has been termed MAFLD (metabolic-associated fatty liver disease) [5]. Thus, NAFLD can be diagnosed in up to about 70-80% of T2D patients, although its presentation, clinical course, and outcomes are quite heterogeneous.
The coexistence of NAFLD in patients with T2D has deleterious consequences for the course of both conditions. In studies on T2D patients, NAFLD has been associated with

Patients and Methods
All patients with T2D consecutively attending the Diabetes Unit of University Hospital of Messina from 1 January to 30 March 2021, who have consented to participate, were enrolled in this single-center study.
Exclusion criteria were any previous diagnosis of liver disease (viral hepatitis, autoimmune hepatitis, hemochromatosis, primary biliary cholangitis, Wilson's disease, sclerosing cholangitis, biliary obstruction, and alpha-1 antitrypsin deficiency), including viral infection (HBV and/or HCV) and/or alcohol use disorder. Alcohol intake [23,24] was evaluated and quantified as the number of units of drinks per day [one drink or alcoholic unit (AU) = 12.5 g of pure ethanol contained in a glass of wine, a pint of beer, or a mini glass of spirits].
Moreover, in order to exclude cases of unknown viral infection, all participants underwent viral infection (HBV and/or HCV) markers measurements before entering the study: Australian antigen (HBsAg) and antibodies (HbsAb), Hepatitis B e-Antigen (HbeAg) and antibodies (HbeAb), and Hepatitis B core antibody (HBcAb) were dosed to exclude Hepatitis B virus infection, while anti-HCV serology was performed to rule out Hepatitis C virus infection. Further exclusion criteria were the following conditions: malignancies, heart failure (New York Heart Association class III-IV), pregnancy, end-stage renal disease, and the use of steatogenic drugs (e.g., estrogens, amiodarone, steroids, and tamoxifen).
Written informed consent was obtained from all patients, and the Institutional Review Board of district of Messina approved the study (protocol number 30/20).

Assessment of Chronic Diabetes Complications
Both micro-and macro-vascular complications of T2D were screened according to national and international diabetes guidelines [27,28].
Macrovascular disease: coronary heart disease was defined based on clinical documentation and of the reports of cardiologist specialists and/or hospital discharge (myocardial infarction, chronic ischemic heart disease, coronary heart by-pass, coronary angioplasty); a standard electrocardiogram and a cardiologist visit are performed annually in all T2D patients as part of the usual screening program. Cerebrovascular disease and peripheral arterial disease were assessed by color-Doppler ultrasonography by B-mode real-time ultrasound, as part of the periodic screening of macrovascular complications.
Microvascular disease: diabetic retinopathy was diagnosed based on direct ophthalmoscopy performed by an expert ophthalmologist and/or by fluorescein angiography within 1 year before the start of study. Diabetic kidney disease was assessed according to albuminuria measurement and estimation of Glomerular Filtration Rate (eGFR) by CKD-EPI formula [29].

Assessment of NAFLD/Liver Fibrosis
All T2D study subjects underwent evaluation of NAFLD and liver fibrosis through non-invasive methods, e.g., the measurements of serum ALT/AST and γ-GT, the calculation of FIB-4 score and the evaluation of liver stiffness through liver elastography, performed by FibroScan with controlled attenuation-parameter (CAP), for the quantification of intrahepatic fat [30,31]. Aminotransferase normal value and γ-GT were measured by standard laboratory methods; normal range of our laboratory was for serum AST and ALT values ≤ 40 UI/L and γ-GT values ≤ 50 UI/L. FIB-4 score was calculated to detect NAFLD by non-invasive scores before elastography (as indicated in Table 1): FIB-4 score values < 1.3 were considered not predictive of fibrosis, whereas FIB-4 score values > 2 were considered positive predictive of fibrosis, and FIB-4 score > 2.67 was considered a high positive predictive value of advanced fibrosis [32]. Liver Stiffness and CAP were measured by Echosens FibroScan 630 Expert, using M or XL probe. Liver stiffness values were considered as both categorical and binomial variables considering the cut-off of 10.1 kPa for severe fibrosis [33]. Liver CAP values were considered categorical variables, values < 238 dB/m were considered normal and severe steatosis values > 290 dB/m [34]. According to current guidelines [clinical practice guideline of the Italian Association for the Study of the Liver (AISF), the Italian Society of Diabetology (SID), and the Italian Society of Obesity (SIO), 2021], NAFLD diagnosis was assumed with a FIB-4 value > 2 and a liver stiffness > 10.1 kPa [13]. Three grades, from S1 to S3, were considered in order to define the severity of liver steatosis. Data are n, % and median and range. Abbreviations: AT (Aminotransferase), BMI (Body mass index), CAP (Controlled attenuation parameter), T2D (type 2 diabetes mellitus), GLP1-Ras (Glucagon-like peptide 1-receptor agonist). AST and ALT ≤ 40 UI/L, and γ-GT ≤ 50 UI/L were considered normal values. Study subjects were in treatment with insulin and/or oral hypoglycemic agents and/or GLP-1RAs or a combination of these drugs.

Statistical Analysis
The numerical data were expressed as median and range (minimum and maximum), and the categorical variables as number and percentage. The non-parametric approach was used because most of the numerical variables were not normally distributed, such as those verified by Kolmogorov-Smirnov test. Univariate and multivariate logistic regression models were estimated in order to identify significant predictive factors of hepatic fibrosis. Univariate logistic regression models were applied to the 3 independent variants FIB-4 > 2, CAP > 290 dB/m and liver Stiffness (FibroScan) > 10.1 kPa. Considering liver Stiffness (FibroScan >10.1 kPa), significative results were achieved for BMI (p value 0.000), HbA1c value (p value 0.002), and creatinine clearance value (p value 0.011): these results were confirmed by multivariate logistic regression models except for the creatinine clearance value (p value 0.779). The same univariate regression model was applied to CAP > 290 dB/m, and significant results were underlined for BMI (p value 0.000), oral diabetic therapy (p value 0.026), normal values of aminotransferase (p value 0.037), triglycerides (p value 0.010), and creatinine clearance value (p value 0.005); among them, just BMI value (0.006) and triglycerides (0.025) were confirmed by multivariate logistic regression model. Eventually, univariate regression model was used for FIB-4 > 2: we obtained significant results just for creatinine value (p value 0.045), a datum confirmed using the multivariate approach (p value 0.035). Statistical analyses were performed using SPSS 22.0 for Windows package. A p value lower than 0.05 was statistically significant.
A total of 125 also had also a previous diagnosis of arterial hypertension and 146 had a diagnosis of dyslipidemia. The amount of alcohol intake was more than 2 UI/die in 10% of the population.
As shown in Figure 1, altered aminotransferase levels (AST and ALT > 40 U/L; GGT > 50 U/L) were detected in 112 (54.6%) ( Figure 1A). The median FIB-4 score was 1.5 (range 0.00-12.76), and 11.2% of subjects had FIB-4 scores > 2 ( Figure 1B results just for creatinine value (p value 0.045), a datum confirmed using the multivariate approach (p value 0.035). Statistical analyses were performed using SPSS 22.0 for Windows package. A p value lower than 0.05 was statistically significant.
A total of 125 also had also a previous diagnosis of arterial hypertension and 146 had a diagnosis of dyslipidemia. The amount of alcohol intake was more than 2 UI/die in 10% of the population.

Discussion
NAFLD affects~30% of the general population [35] in developed countries, but it may be diagnosed in up to 70% of patients with T2D, with a higher risk of developing more severe forms and worse outcomes [36]. Despite such evidence, there is a lack of clear information on when and how to screen T2D outpatients in routine clinical practice.
In this study, in accordance with current Guidelines [13], we applied non-invasive methods in order to screen all consecutive T2D patients attending a single diabetes outpatient clinic after excluding other causes of liver disease.
In particular, the FIB-4 score was calculated to identify T2D patients with potentially metabolic liver disease, and a combination of US and transient elastography with CAP was performed in order to identify more severe forms and to address patients to further liver workouts, including liver biopsy, when necessary. Overall, our study showed that 55.6% of our T2D patients had unknown severe NAFLD based on CAP values. In these patients, the high risk of fibrosis should be considered to refer them to the correct clinical management.
Moreover, a serious warning came from our study since 23.9% of screened T2D patients had clinically meaningful liver harm, with a FIB-4 score > 2 and/or stiffness > 10.1 kPa. Similar results have been shown in a small study conducted in Morocco [37], where noninvasive hepatic biomarkers indicated the presence of significant fibrosis in 18.3% of the studied population.
The high prevalence of fibrosis observed in T2D outpatients without any known liver disease, and regularly attending a diabetes visit is particularly alarming since the presence of fibrosis has been associated with a higher risk of cirrhosis and mortality [38]. In addition, our data showed that obesity and poor glucose control were independent predictors of liver stiffness, and this information may be useful to individuate T2D individuals at higher fibrosis risk.
Although the pathophysiological mechanisms underlying the occurrence of liver fibrosis in patients with T2D are not fully identified yet, our data point to a relevant role for obesity and its corollaries in terms of insulin resistance, clarifying which patients should be considered at higher risk of the more severe forms of NAFLD.
NAFLD is considered the hepatic manifestation of visceral obesity and metabolic syndrome [39], and obesity and dyslipidemia have been proven to be involved in the initiation and progression of NAFLD in patients with T2D [40,41]. Moreover, NAFLD and T2D have common physio-pathological pathways represented by insulin resistance (IR), oxidative stress, systemic inflammation, and excessive production of free fatty acids (FFA) [42,43]. IR is considered the principal link between NAFLD and T2D for its effects on lipid metabolism [43], determining a deposit of lipid droplets in hepatocytes, which promote hepatocellular ballooning and lobular inflammation up to NASH development [44,45]. In line with this pathogenetic hypothesis, in our study, CAP levels > 290 dB/m were related to high BMI and triglycerides values, representing the triglycerides' accumulation > 5% of liver weight as the basis for NAFLD.
Moreover, our results showed that a higher creatinine value was associated with a FIB-4 score value > 2. High creatinine values are associated with a 3.8 times greater risk of having a FIB-4 value > 2, suggesting liver fibrosis. These results are in accordance with recent studies supporting a relationship between chronic kidney disease (CKD) and the prevalence and severity of NAFLD [46,47]. Although several mechanisms have been proposed to justify this correlation, including metabolic syndrome, fructose and uric acid accumulation, oxidative stress, intestinal dysbiosis, platelet activation, and genetic predispositions [47,48], the pathogenetic link between NAFLD and microvascular complications in T2D patients still needs to be further investigated.
Contrariwise, AST and ALT values did not enter the predictive model. Thus, it is important to underline that although ALT levels still represent the most used laboratory parameter to screen for NAFLD in routine clinical practice, they cannot be considered an optimal marker to predict NASH and advanced fibrosis, showing poor sensitivity in early NAFLD diagnosis and prognosis [49]. On the other hand, the interplay between T2D and liver damage is not limited to NAFLD [50].
In line with the recent literature data, our study also underlines the importance of using a combination of non-invasive methods in order to screen T2D patients for NAFLD and its more severe forms in the "real world setting" of diabetes clinics. Considering the alarming prevalence of the severe forms of liver disease among patients with T2D unaware of the liver damage, the effective screening of T2D patients with non-invasive methods may help to identify individuals at high risk of advanced liver disease. Additionally, considering the increased risk of cardiovascular events associated with NAFLD, this systematic screening may help to prevent MACE in T2D patients [51].
Besides the pathogenetic background, the finding of a higher risk of liver fibrosis in unaware T2D obese patients should prompt physicians to investigate liver damage in these patients, also in light of their correct management. Thus, although no medical treatment has been currently authorized for the management of NAFLD in patients with T2D, a growing body of evidence points to the beneficial effects of insulin-sensitivity drugs, such as pioglitazone, metformin, GLP-1RAs, and SGLT2i, irrespective of their beneficial effect on body weight [52][53][54]. In particular, a recent meta-analysis showed that treatment with GLP-1RAs was associated with significant reductions in the absolute percentage of liver fat content, serum liver enzyme levels, and a greater histological resolution of NASH without worsening liver fibrosis compared to placebo or reference therapies, indicating the great potential of this class of drugs for the treatment of T2D patients with NAFLD [55]. However, when we verified the impact of these classes of hypoglycemic drugs on CAP, liver stiffness, and FIB-4 score, none showed significant associations either at univariate or multivariate regression analysis. The relatively small sample size, especially when considering the single drug classes, may likely have played a role in these results.
Several limitations should be acknowledged when interpreting our results. The cross-sectional study design impedes assessing causality, and the small sample size and the monocentric recruitment of patients may partly prevent the generalizability of our results to other study populations. The systematic screening approach by using a combination of noninvasive methods, the unselected T2D population screened in the setting of an outpatient diabetes clinic, the careful exclusion of potential causes of liver disease, including unknown virus hepatitis infection, and the continued follow-up of patients for their future evaluation should conversely be acknowledged among the strengths of this study.
Currently, a hint for future research is represented by the assessment of a possible association between liver disease progression and T2D complications, including both microand macro-vascular ones. Additionally, after the implementation of the sample, further stratifications of the study population could be made according to the prescribed therapies, with special attention to GLP1-RAs that represent the future of NAFLD treatment.

Conclusions
In conclusion, liver fibrosis is a frequent finding in T2D outpatients without a known history of liver disease, especially those with obesity, hypertriglyceridemia, worse glycemic control, and high creatinine levels. The use of a combination of non-invasive methods should be implemented in routine clinical practice in order to identify high-risk subjects and referred them to the appropriate management.

Institutional Review Board Statement:
The study was performed in accordance with the principles of the Declaration of Helsinki and it was approved by the Ethics Committee of the Messina University Hospital (protocol number 30/20, date 29 June 2020).

Informed Consent Statement:
Written informed consent has been obtained from the patients to publish this paper.
Data Availability Statement: Data are unavaiable due to privacy.

Conflicts of Interest:
The authors declare no conflict of interest.