Abstract Background Vascular endothelial growth factor D (VEGF-D) is a secreted glycoprotein that can induce lymphangiogenesis and angiogenesis. We recently demonstrated that serum levels of VEGF-D are independently associated with all-cause mortality in patients with suspected or known coronary artery disease (CAD). However, the impact of diabetes mellitus (DM) on the association between VEGF-D and cardiovascular (CV) mortality in those patients is unclear. Methods Serum VEGF-D levels were measured in 2418 patients with suspected or known CAD undergoing elective coronary angiography. The primary outcome was CV death. The secondary outcomes were all-cause death, major adverse CV events (MACE) defined as a composite of CV death, nonfatal myocardial infarction, and nonfatal stroke, and heart failure (HF) hospitalization. Patients were divided into 2 groups according to the presence (DM, n=1087) or absence (non-DM, n=1331) of DM, and followed up over a 6-year period. Results During the follow-up, 166 (81 DM and 85 non-DM) patients died from CV disease, 536 (284 DM and 252 non-DM) patients died from any cause, 297 (156 DM and 141 non-DM) patients developed MACE, and 268 (131 DM and 137 non-DM) patients developed HF hospitalization. After adjustment for potential clinical confounders and established CV biomarkers (i.e., N-terminal pro-brain natriuretic peptide, high-sensitivity cardiac troponin I, and high-sensitivity C-reactive protein), VEGF-D levels were significantly associated with all-cause death (hazard ratio [HR] for 1-SD increase, 1.20; 95% confidence interval [CI], 1.08–1.33) and HF hospitalization (HR, 1.23; 95% CI, 1.08–1.40), but not with CV death (HR, 1.18; 95% CI, 0.98–1.43) or MACE (HR, 1.11; 95% CI, 0.96–1.28), in DM patients, while VEGF-D levels were significantly associated with CV death (HR, 1.24; 95% CI, 1.08–1.43) and MACE (HR, 1.16; 95% CI, 1.03–1.32), but not with all-cause death (HR, 1.10; 95% CI, 0.99–1.22) or HF hospitalization (HR, 1.13; 95% CI, 0.99–1.28), in non-DM patients. The addition of VEGF-D levels to the model with potential clinical confounders and established CV biomarkers did not improve the prediction of CV death (P=0.92 for continuous net reclassification improvement [NRI], P= 0.32 for integrated discrimination improvement [IDI]), all-cause death (P=0.098 for NRI, P=0.12 for IDI), MACE (P=0.80 for NRI, P=0.53 for IDI) or HF hospitalization (P=0.96 for NRI, P=0.39 for IDI) in DM patients, whereas the addition of VEGF-D levels significantly improved the prediction of CV death (P=0.002 for NRI, P=0.03 for IDI) and all-cause death (P<0.001 for NRI, P=0.02 for IDI), but not that of MACE (P=0.11 for NRI, P= 0.09 for IDI) or HF hospitalization (P=0.89 for NRI, P=0.83 for IDI), in non-DM patients. Conclusions In patients with suspected or known CAD, the VEGF-D level independently predicted CV mortality in non-DM, but not in DM patients. The association between VEGF-D and CV mortality was attenuated in the presence of DM.