Growth differentiation factor-15 and the association between type 2 diabetes and liver fibrosis in NAFLD

Type 2 diabetes mellitus (T2DM) is a strong risk factor for liver fibrosis in non-alcoholic fatty liver disease (NAFLD). It remains uncertain why T2DM increases the risk of liver fibrosis. It has been suggested that growth differentiation factor-15 (GDF-15) concentrations increase the risk of liver fibrosis. We aimed to investigate (a) whether GDF-15 concentrations were associated with liver fibrosis and involved in the relationship between T2DM and liver fibrosis and (b) what factors linked with T2DM are associated with increased GDF-15 concentrations.

GDF-15 concentrations are a predictor of liver fibrosis and potentially involved in the association between T2DM and liver fibrosis in NAFLD. HbA1c concentrations explain a large proportion of the variance in GDF-15 concentrations.

Ninety-nine patients with NAFLD (61% men, 42.4% T2DM) were studied. Serum GDF-15 concentrations were measured by electro-chemiluminescence immunoassay. Vibration-controlled transient elastography (VCTE)-validated thresholds were used to assess liver fibrosis. Regression modelling, receiver operator characteristic curve analysis and Sobel test statistics were used to test associations, risk predictors and the involvement of GDF-15 in the relationship between T2DM and liver fibrosis, respectively.

Patients with NAFLD and T2DM (n = 42) had higher serum GDF-15 concentrations [mean (SD): 1271.0 (902.1) vs. 640.3 (332.5) pg/ml, p < 0.0001], and a higher proportion had VCTE assessed ≥F2 fibrosis (48.8 vs. 23.2%, p = 0.01) than those without T2DM. GDF-15 was independently associated with liver fibrosis (p = 0.001), and GDF-15 was the most important single factor predicting ≥F2 or ≥F3 fibrosis (≥F2 fibrosis AUROC 0.75, (95% CI 0.63-0.86), p < 0.001, with sensitivity, specificity, positive predictive (PPV) and negative predictive (NPV) values of 56.3%, 86.9%, 69.2% and 79.1%, respectively). GDF-15 was involved in the association between T2DM and ≥F2 fibrosis (Sobel test statistic 2.90, p = 0.004). Other factors associated with T2DM explained 60% of the variance in GDF-15 concentrations (p < 0.0001). HbA1c concentrations alone explained 30% of the variance (p < 0.0001). Conclusions: GDF-15 concentrations are a predictor of liver fibrosis and potentially involved in the association between T2DM and liver fibrosis in NAFLD. HbA1c concentrations explain a large proportion of the variance in GDF-15 concentrations.