<p><b>OBJECTIVE</b>To detect the plasma level of growth differentiation factor-15 (GDF-15) in patients of type 2 diabetic nephropathy and assess its value in diagnosis and evaluation of type 2 diabetic nephropathy.</p><p><b>METHODS</b>Thirty type 2 diabetic patients with normoalbuminuria, 20 with microalbuminuria, and 30 with macroalbuminuria, diagnosed according to Mogensen's criteria, were examined for plasma GDF-15 level using enzyme-linked immunosorbent assay.</p><p><b>RESULTS</b>The patients with macroalbuminuria had significantly higher plasma GDF-15 level than those with microalbuminuria and normoalbuminuria [1773.9 (1099.1-2357.4) pg/ml vs 864.0 (636.1-994.3) pg/ml and 704.5 (548.8-975.8) pg/ml, respectively, P<0.01], and patients with microalbuminuria had higher GDF-15 level than those with normoalbuminuria (P>0.05). Plasma GDF-15 level was found to increase early in the stage of mild renal dysfunction (60≤GFR<90 ml·min(-1)·1.73 m(-2)) with a median level of 999.5 (769.2-1372.1) pg/ml. Partial correlation analysis showed that plasma GDF-15 level was positively correlated with diabetic durations, mAlb, BUN and sCr (r=0.246, 0.493, 0.390, and 0.471, respectively, P<0.05), and negatively with eGFR (r=-0.438) and Alb (r=-0.397) (P<0.01). Multivariate linear regression analysis showed that a high plasma GDF-15 level was an independent risk factor for increased mAlb. In the diagnosis of renal dysfunction (eGFR<90 ml·min(-1)·1.73 m(-2)), the area under the receiver-operating characteristic curve (AUC) of GDF-15 was 0.801, significantly higher than that of mAlb (0.717, P<0.05). With the cut-off value of 733.78 pg/ml, plasma GDF-15 level had a sensitivity of 88.1% and a specificity of 58.1% for renal dysfunction diagnosis.</p><p><b>CONCLUSION</b>In patients with type 2 diabetic nephropathy, plasma GDF-15 level increases with the Mogensen stage, and as a independent risk factor for increased mAlb, it is significantly correlated with mAlb and eGFR, and serves, suggesting its value in early diagnosis, evaluation and prediction of the outcomes of type 2 diabetic nephropathy.</p>