BACKGROUND: There are inconsistent findings regarding associations between triglyceride levels and cardiovascular disease (CVD) in type 2 diabetes mellitus (T2DM). This study aimed to test whether the association between triglycerides and CVD depends upon duration of diabetes. METHODS: From April 1, 2012, to June 30, 2012, we conducted a cross-sectional survey of 223 612 patients with T2DM from 630 hospitals in China. Cardiovascular disease was defined as having either prior coronary heart disease or stroke, or diabetic foot. Binary logistic regression was used to estimate odds ratios of triglyceride for CVD. Relative excess risk due to interaction, attributable proportion due to interaction, and synergy index were used to estimate effect size of additive interaction between low triglyceride, ie, <1.7 mmol/L, and duration of diabetes, ie, ≥15 years. RESULTS: Among 223 612 T2DM patients, 31 898 (14.27%) suffered from CVD. A low level of triglyceride was associated with decreased risk of CVD (univariable OR, 0.91, 95% CI, 0.88-0.93; multivariable OR, 0.94, 95% CI, 0.92-0.97) among patients with <15 years of duration of diabetes but increased risk of CVD (univariable OR, 1.12, 95% CI, 1.04-1.21; multivariable OR, 1.18, 95% CI, 1.09-1.27) among those patients with 15 and more years of duration of diabetes with significant additive interactions (relative excess risk due to interaction, 0.39, 95% CI, 0.25-0.52; attributable proportion due to interaction, 0.20, 95% CI, 0.14-0.27; and synergy index, 1.80, 95% CI, 1.43-2.28). CONCLUSIONS: Whereas a high triglyceride level was associated with increased risk of CVD in short-term T2DM, low triglyceride was associated with increased CVD risk in long-term T2DM. Low triglyceride may be a marker of CVD risk in Chinese patients with long-term T2DM.