OBJECTIVES: Diabetes is known to be associated with a considerable decline in life expectancy (LE). The aim of this study was to use a modelling approach to assess LE in low, intermediate and high-risk type-2-diabetes (T2D) populations and to compare these to observations from the Canadian-Chronic-Disease-Surveillance-System (CCDSS).  METHODS: This study used the IMS-Core-Diabetes-Model (CDM), a validated diabetes simulation model, to project the LE of T2D individuals with a low-risk (age=55, diabetes duration=5, no CVD history), intermediate-risk (age=65, diabetes duration=15, moderate CVD history) and high-risk profile (age=80, diabetes duration=30, advanced CVD history). LE was predicted utilising three alternative mortality risk prediction models (RPMs) from the UKPDS 68 study (UK68), the UKPDS 82 study (UK82) and a risk equation based on Western Australia (WA) administrative data. Life-years-lost (LYL) in diabetes vs. no-diabetes populations was estimated based on the difference in age matching LE obtained from  UK-national-life-tables subtracted by CDM projected-LE. Results were finally contrasted to LE and LYL estimations from the CCDSS study. RESULTS: When UK68 mortality RPMs were applied, LE projected was 23.29, 15.94 and 7.78 years for the low, intermediate and high risk cohort. This compared to 22.16, 14.88 and 7.29 years utilising UK82 RPMs and 25.94, 18.11 and 9.05years when utilising the WA RPMs. Based on UK life table data, LYL in diabetes vs. no-diabetes populations were 4.76, 3.61 and 1.11 (UK68), 5.89, 4.67 and 1.60 (UK82) and 2.11, 1.44 and -0.15 (WA) years. The CCDSS study reported outcomes for the low risk (age 55) and high risk (age 80) profile at 24.5 and 8.3 years (LE) and 5.5 and 2.25 years (LYL), respectively. CONCLUSIONS: UKPDS based models predicted LE and LYL very closely to CCDSS study findings. The Western Australian RPM seems not to be applicable to a UK and Canadian population.