Introduction The association between liver disease and diabetes has been described, but less information is available regarding mortality from liver disease among people with diabetes from a population perspective.

Aim To compare mortality from liver disease among people with prevalent diabetes in 2001 and incident diabetes between 2001 and 2007 with that of the general population of Scotland.

Method We used a population-based diabetes register derived from primary and secondary care electronic records linked to death records to compare mortality from liver disease among people with prevalent diabetes in 2001 and incident diabetes between 2001 and 2007 with that among the general population of Scotland for people of 35–84 years for the period 2001–2007. There were just over 1 million person years of data for people with diabetes and almost 20 million person years of data for the general population. Mortality rates were estimated for liver disease as underlying (primary) cause of death on death certificates using conventional ICD-10 codes including those for hepatocellular carcinoma (HCC). Standardised mortality ratios (SMRs) were estimated adjusting for age, sex and quintile of an area-based measure of socio-economic status, the Scottish Index of Multiple Deprivation 2006 using the population of Scotland in 2004 as the standard.

Results There were 1267 and 10 100 records that mentioned liver disease as the primary underlying cause of death in the diabetes and general populations, giving crude mortality rates of 122.4 and 50.9/100 000 person years, respectively. The major single cause of liver-related death was alcoholic liver disease which accounted for 38% and 63% of liver disease deaths among people with diabetes and the general population respectively, with HCC accounting for 24% and 9% of liver disease deaths in these populations. SMRs (95% CI) for underlying (primary) cause of death for people with diabetes compared to the general population were 169 (160–178) for all liver disease, 116 (106–126) for alcoholic liver disease and 318 (283–356) for HCC. SMRs were similar for any mention of liver disease on death certificate. SMRs for women with diabetes were higher than men. Further analyses will stratify by type of diabetes.

Conclusion Analysis of a very large population based data set has shown that people with diabetes have a raised mortality rate from liver disease compared to the general population. Management of people with diabetes should include strategies to reduce risk of liver disease.