© 2018 Oshima et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Background According to studies by the National Kidney Foundation and Food and Drug Administration, 30% and 40% declines in estimated glomerular filtration rate (eGFR) could be used as surrogate endpoints of end-stage renal disease (ESRD). However, the benefits of using these endpoints in diabetic patients remain unclear. Methods This cohort study comprised Japanese patients with type 2 diabetes; those with repeated serum creatinine measurements during a baseline period of 2 years (n = 1868) or 3 years (n = 2001) were enrolled. Subsequent risks of ESRD following eGFR declines were assessed. Results In the 2-year baseline analysis, the cumulative prevalence of −20%, −30%, −40%, and −53% changes in eGFR were 23.9%, 11.1%, 6.8%, and 3.7%, respectively. There were 133 cases (7.1%) of subsequent ESRD during a median follow-up period of 6.5 years. In the 3-year baseline analysis, the corresponding proportions were 28.1%, 14.0%, 7.7%, and 3.9%, respectively, with 110 participants (5.5%) reaching ESRD during a median follow-up period of 5.5 years. The adjusted hazard ratios of subsequent ESRD following −53%, −40%, −30%, and −20% changes in eGFR during the 2-year baseline period were 22.9 (11.1–47.3), 12.8 (6.9–23.7), 8.2 (4.3–15.5), and 3.9 (2.2–7.0), respectively when compared with the no changes in eGFR. In the 3-year baseline analysis, the corresponding risks were 29.7 (10.8–81.9), 18.4 (7.6–44.7), 12.8 (5.2–32.2), and 5.4 (2.3–12.8), respectively. In the subgroup analysis, similar trends were observed in patients with macroalbuminuria at baseline. Conclusions Declines in eGFR were strongly associated with subsequent risk of ESRD in Japanese type 2 diabetic patients. In addition to 30% and 40% declines, a 20% decline in eGFR over 2 years could be considered as a candidate surrogate endpoint of ESRD in diabetic kidney disease.