OP58Is the Excess Risk of Myocardial Infarction Among People with Diabetes Falling Over Time?

BackgroundIncidence of myocardial infarction (MI) in the UK general population has declined considerably in recent years. However it is unclear whether the decline in MI risk has occurred among people with diabetes. People with diabetes have an estimated two-fold excess risk of MI, compared to those without diabetes. A differential trend in MI incidence among diabetic patients could correspond to a rise or fall in this excess risk, which has implications for prognosis and management of diabetes. We compared recent trends in MI incidence among those with and without diabetes in a representative UK population sample, and estimated the excess risk of MI among diabetic patients in different calendar periods.MethodsThe population sample comprised 2,927,137 patients (49% men) aged 30 years and over, with no prior MI, from 434 general practices belonging to The Health Improvement Network (THIN) UK-wide primary care database in 1995-2008. Incidence of MI in 1995-1998 was compared with that 10 years later in 2005-2008. Rate ratios comparing incidence over these intervals were estimated from multi-level Poisson regression (patients nested in practices), with an indicator for time interval as a covariate, adjusting for age and gender, with practice as a random effect. An interaction between time interval and an indicator for diabetes was used to assess whether the rate ratios comparing intervals differed among those with and without diabetes, and equivalently whether the excess relative risk among diabetic patients has changed over time.ResultsIn 1995-1998 age-standardised incidence rates for MI (per 1000 person years) among those without and with diabetes were 3.22 (95% CI 3.10-3.34) and 9.56 (8.42-10.7) respectively. In 2005-2008, corresponding incidence rates were 1.47 (1.44-1.50) and 4.43 (4.23-4.65). Among people without diabetes, the rate ratio comparing incidence in 2005-2008 with 1995-1998 was 0.46 (0.44-0.48). Among people with diabetes, the corresponding rate ratio was 0.31 (0.28-0.35), indicating a greater decline in MI incidence over the period (p