Aims/hypothesis We aimed to determine associations of regression to normal glucose tolerance (NGT), maintaining impaired glucose tolerance (IGT) or progression to diabetes with subsequent risks of CVD and microvascular disease among Chinese adults with IGT. Methods We conducted an observational study among 540 participants in the Da Qing Diabetes Prevention Study, a 6 year lifestyle intervention trial in people with IGT, defined by 1985 WHO criteria as fasting plasma glucose <7.8 mmol/l and 2 h post-load plasma glucose >= 7.8 and <11.1 mmol/l. At the end of the trial, the groups that had regressed to NGT, remained with IGT or progressed to diabetes were identified. Participants were then followed for 24 years after completion of the trial, during which we compared the incidence and hazard ratios for CVD and microvascular disease in each group and estimated the differences in their median time to onset from parametric Weibull distribution models. Results At the end of the 6 year trial, 252 (46.7%) participants had developed diabetes, 114 (21.1%) had remained with IGT and 174 (32.2%) had regressed to NGT. Compared with those who developed diabetes during the trial, the median time to onset of diabetes was delayed by 14.86 years (95% CI 12.49, 17.25) in the NGT and 9.87 years (95% CI 8.12, 11.68) in the IGT groups. After completion of the trial, among those with diabetes, IGT and NGT, the 24 year cumulative incidence of CVD was 64.5%, 48.5% and 45.1%, respectively, and 36.8%, 21.7% and 16.5% for microvascular diseases. Compared with participants who had progressed to diabetes during the trial, those who regressed to NGT had a 37% (HR 0.63; 95% CI 0.47, 0.85) reduction in CVD incidence and a median delay of 7.45 years (95% CI 1.91, 12.99) in onset, and those who remained with IGT had a 34% (HR 0.66; 95% CI 0.47, 0.91) lower CVD incidence with a median delay in onset of 5.69 years (95% CI 1.0, 10.38). Participants with NGT had a 66% (HR 0.34; 95% CI 0.20, 0.56) lower incidence of microvascular diseases and a median delay in the onset of 18.66 years (95% CI 6.08, 31.24), and those remaining with IGT had a 52% (HR 0.48; 95% CI 0.29, 0.81) lower incidence with a median delay of 12.56 years (95% CI 2.49, 22.63). Conclusions/interpretation People with IGT who reverted to NGT or remained with IGT at the end of the 6 year trial subsequently had significantly lower incidences of CVD and microvascular disease than those who had developed diabetes.