High circulating fibroblast growth factor-21 levels as a screening marker in fatty pancreas patients

Study population

The study was performed at the Affiliated Hospital of Yangzhou University in Yangzhou, China. We selected 99 fatty pancreas subjects and 16 healthy normal control subjects from the physical examination center of Affiliated Hospital of Yangzhou University from August 2018 to June 2019. Fatty pancreas subjects who met the following criteria were excluded: (1) subjects aged <18 years or >65 years; (2) subjects with acute or chronic inflammatory diseases; (3) subjects with a previous diagnosis of chronic pancreatic, liver or kidney disease; (4) subjects with severe immune system disorders or pregnancy; (5) subjects with incomplete information or refusal to provide clinical blood samples. Besides, the inclusion criteria of normal control subjects were as follows: (1) subjects aged 18–65 years; (2) subjects without fatty pancreas; (3) subjects without any acute or chronic diseases; (4) subjects without previous histories of acute or chronic diseases; (5) subjects without severe immune system disorders or pregnancy or cancers; (6) subjects without incomplete information or refusal to provide clinical blood samples. The exclusion criteria were those who did not meet the inclusion criteria.

This study conformed to the ethical principles of the Declaration of Helsinki. The study was approved by Ethics Committee of Affiliated Hospital of Yangzhou University (Ethical Application Ref: 2018-YKL11-27- topic1). All participants agreed to participate in the study, and written informed consent was obtained from each subject.

Diagnosis of fatty pancreas and fatty liver

As previously described, all subjects underwent transabdominal ultrasonography to diagnose fatty pancreas and fatty liver (Wang et al., 2018). Operations were performed by skilled surgeons with more than 10 years of experience using transabdominal ultrasonography (with an abdominal convex array probe, frequency: 3.5–5 MHz, LOGIQ E9, GE, USA). The ultrasound diagnostic criteria for fatty liver were as follows: the anterior echo of the liver was enhanced while the posterior echo was weakened, and the tubular structure of the liver could not be clearly displayed (Hamaguchi et al., 2007). The characteristics of the ultrasonic image of fatty pancreas were as follows: diffuse strong echoes of pancreatic parenchyma, normal or slightly increased volume, similar or slightly higher echogenicity compared to the adipose tissue in the area of the superior mesenteric artery (Wang et al., 2018; Smereczynski & Kołaczyk, 2016).

Anthropometric and biochemical findings

Clinical information, including the subject’s identity, age, gender, body weight, height, blood pressure, past medical history, drug history, history of smoking and alcohol intake, was recorded using a standardized questionnaire. Height and body weight were assessed using standardized and calibrated scales. Additionally, the body mass index (BMI) was calculated for each subject (BMI = body weight (kg)/square of height (m²)). The subjects included “continuous smokers” (continuous or cumulative smoking for 6 months or more in a lifetime) and “nonsmokers” (including those who quit smoking ≥1 year ago) (Wang et al., 2018). The subjects also included “continuous drinkers” (drinking volume ≥20 g/d; drinking duration ≥2 years) and “nondrinkers” (individuals who had not consumed alcohol for ≥half a year) (Wang et al., 2018).

All subjects fasted for at least 8 h the night before the visit, and blood samples were collected with the participants in a quiet state on an empty stomach the next morning. Blood samples were sent to the laboratory for uniform testing (dry chemical method), and the remaining unused serum was frozen in an ultralow temperature freezer for analysis using an enzyme linked immunosorbent assay (ELISA). Diabetes was defined as a fasting blood glucose (FBG) level ≥ 7.0 mmol/L or a previous diagnosis by a doctor (Jia et al., 2019).

Dyslipidemia was defined as meeting any of the following: high serum total cholesterol (CHO) (≥5.17 mmol/L), high triglyceride (TG) levels (≥1.7 mmol/L), decreased high density lipoprotein (HDL) levels (<1.03 mmol/L), high low density lipoprotein (LDL) levels (≥4.1 mmol/L), or a previous diagnosis of dyslipidemia by a doctor (Wang et al., 2018).

Measurement of FGF-21 and adipocytokine levels in human serum

Levels of serum FGF-21, adiponectin, leptin, resistin and tumor necrosis factor-alpha (TNF-α) were quantified using ELISA kits (USCN Kit Inc., Wuhan, China). All operations were carried out in strict accordance with the kit instructions. The quantification of the results was calculated according to the instructions. Averaged the readings for standards, controls, and samples and subtracted the average zero standard relative light unit. Created the standard curve on log–log graph paper, with FGF-21 or adipocytokines concentration on the y-axis and the relative light unit value on the x-axis. Drawn the best fit straight line through the standard points and it could be determined by regression analysis. If samples had been diluted, the concentration read from the standard curve must be multiplied by the dilution factor.

Statistical analysis

Statistical analysis was performed with IBM SPSS 19.0 software. Normality and homogeneity of variance tests were performed for each group of data by Shapiro–Wilk and Levene’s test. Continuous measurement data that were normally distributed are presented as the mean ± standard deviation (Mean ± SD). Continuous variables that were not normally distributed are presented as medians (25th and 75th percentiles), and categorical variables are presented as percentages (n%). The differences between the fatty pancreas (FP group) and normal control (NC group) groups were determined using Student’s t tests and Mann–Whitney U tests. The qualitative data were compared using the chi-square test. Pearson analysis was used to evaluate the correlation between FGF-21 and other factors. GraphPad Prism 7 was used to generate receiver operating characteristic (ROC) curve. A two-sided p value <0.05 was used to indicate statistically significant differences.

Basic and clinical characteristics of the study population

In this study, 99 fatty pancreas patients (FP group) were enrolled, and 16 healthy subjects were enrolled as the normal control group (NC group). As shown in Table 1, the proportion of males in the FP group and NC group was 58.6% (58/99) and 50% (8/16), respectively (P = 0.524). The average age and BMI in the FP group were significantly higher than those in the NC group (P = 0.003, P < 0.001). There was no difference in the proportion of smokers or drinkers between the two groups. Besides, the proportion of fatty liver, diabetes, hypertension, obesity, dyslipidemia in FP group were significantly increase than that in NC group (All P < 0.001). This also suggested that fatty pancreas may have an inseparable relationship with metabolic diseases.

Then, clinical characteristics were compared between the FP and NC groups. As shown in Table 2, it is self-evident that some key metabolic indicators, including serum uric acid (UA), glucose, CHO and TG were significantly higher in the FP group than in the NC group (All P < 0.05), with the corresponding serum HDL were decreased (P < 0.001).

Comparison of serum FGF-21 and adipocytokine levels

We measured the levels of serum FGF-21 and adipocytokines (adiponectin, resistin, leptin and TNF-α) in the two groups. As shown in Fig. 1, the serum levels of FGF-21 and resistin in the FP group were significantly higher than that in the NC group, and adiponectin showed the opposite trend (All P < 0.05). In addition, the serum leptin and TNF-α levels in the FP group showed the upward trend in the FP group without a significant difference. Interestingly, in Table S1, we found that the levels of serum FGF-21 in male of FP group were significantly higher than that in female of FP group (P = 0.004). However, the levels of serum leptin and TNF-α in female of FP group were significantly higher than that in male (All P < 0.01). There was no significant difference between the two groups in serum resistin and adiponectin levels (Table S1).

It is remarkable that serum FGF-21 levels were negatively correlated with leptin (r = −0.261, P = 0.005) in the Pearson correlation analysis. In addition, there was no clear correlation with adipocytokines (adiponectin, resistin and TNF-α) or key metabolic clinical indicators (Glucose, CHO, LDL, TG and HDL) (Table 3).

Fatty pancreas prediction analysis

Serum FGF-21 levels showed significantly greater discriminative ability for FP. Area under ROC curves (AUC) of FGF-21 were remarkably larger than adipocytokines (adiponectin, resistin, leptin and TNF-α) (Fig. 2). To determine the critical value of FGF-21 for predicting FP, ROC curves were assessed and are presented in Table 4. The results showed that 171 pg/mL was the best critical value (AUC 0.744, P = 0.002, 95% Confidence Interval (95% CI) 0.636−0.852). The corresponding sensitivity was 80.8%, and the specificity was 68.8% (Table 4).