Internal Validation of the Metabo-Reno-Cardiovascular Disease Model: Cardiovascular Outcomes in Type 2 Diabetes

OBJECTIVES: Several generic disease models have been developed in cardiovascular disease (CVD), diabetes, obesity and kidney disease. These diseases have common comorbidities; thus, most of these models describe all of them to some extent, with some limitations. To the best of our knowledge, the metabo-reno-cardiovascular-disease-model (MRCDM) is the first model that fully integrates the impact of changes in all the appropriate risk factors on the different diseases, depending on the glycemic status and body weight. The aim of this abstract was to validate the prediction of CVD in individuals with type 2 diabetes (T2D).

METHODS: The MRCDM is a patient-level model. Individuals that enter the model can be with or without diabetes (type 1 or 2), obesity (defined by BMI), CVD and chronic kidney disease. To predict the risk of CVD in T2D in Western countries different sets of risk equations are available: UKPDS82, Framingham and the Swedish National Diabetes Registry (SweNDR). The model was populated with UKPDS specific baseline characteristics (age 53; HbA1c 7.1%). The UKPDS82 mortality and UKPDS90 progression of risk factors equations were used. Risk of CVD at 25 years was compared to the UKPDS study outcomes.

RESULTS: In the UKPDS study at 25 years, the cumulative incidence of MI, angina, stroke and heart failure were 29%, 23%, 15% and 11%, respectively. Using the MRCDM, applying the UKPDS82 CVD equations, the cumulative incidence of MI, angina, stroke and heart failure were 27%, 21%, 14% and 11%, respectively. With Framingham equation it was 20%, 15%, 10% and 17% and with SweNDR CVD equations 27%, 22%, 15% and 12%, respectively.

CONCLUSIONS: Using UKPDS82 CVD equations, MRCDM is predicting very similar outcomes to the ones reported in the source. Framingham and SweNDR equations are derived from different populations. Hence, predictions are different nevertheless, having face validity.