CONTRASTING PREDICTIONS OF CARDIOVASCULAR INCIDENCE DERIVED FROM ALTERNATIVE RISK PREDICTION MODELS IN TYPE 1 DIABETES
Author(s)
McEwan P1, Foos V2, Lamotte M3
1Health Economics and Outcomes Research Ltd, Cardiff, UK, 2IMS Health, Basel, Switzerland, 3IMS Health, Vilvoorde, Belgium
OBJECTIVES: Cardiovascular disease (CVD) risk prediction models are available for the general population (Framingham) and for type-2-diabetes (T2D) (UKPDS 68 and 82) but may not be appropriate in type-1-diabetes (T1D). The IMS CORE Diabetes Model (CDM) uses Framingham and UKPDS risk equations (REs) to predict CVD incidence in T2D and has recently been updated to include two CVD risk prediction approaches specific to T1D populations based on data from the Epidemiology-of-Diabetes-Interventions-and-Complications-study (EDIC) and a novel RE from the Pittsburgh-Epidemiology-of-Diabetes-Complications-Study (PEDC). The objective of this study was to compare CVD incidence across T1D model projections utilizing UKPDS, EDIC and PEDC REs and compare these to published EDIC findings. METHODS: The CDM was applied to project the incidence of myocardial-infarction (MI), stroke, heart-failure (HF) and ischemic-heart-disease (IHD) utilizing three alternative CVD REs, the UKPDS 68 RE (UK68-RE), EDIC-RE and PEDC-RE. The risk profile of a newly diagnosed T1D population (age 21 years, HbA1c 7%, systolic-blood-pressure 114 mmHg, body-mass-index 32 Kg/m, high-density-lipoprotein 45 mg/dl and total-cholesterol 170 mg/dl) was projected over 30 years. The incidence of total CVD was estimated as the sum of the individual composites (%CVD=%MI+%stroke+IHD+HF) to enable comparison to published EDIC findings. RESULTS: When UK68-REs were applied, the 30-year cumulative incidence of CVD for a newly diagnosed T1D individual was projected at 2.66%, 0.27%, 3.88% and 0.72% for MI, stroke, IHD and HF, respectively. This compared to 4.10%, 0.66%, 3.36 and 0.58% utilizing EDIC-RE and 5.27%, 1.01%, 3.44 and 1.18% utilizing PEDC-RE. Total predicted CVD incidence added up to 7.53%, 8.70% and 10.90% for UK68-RE, EDIC-RE and PEDC-RE respectively, which compares to 8.70% incidence of CVD as observed during the EDIC study. CONCLUSIONS: As expected, the CDM reproduced the published EDIC CVD incidence when using the EDIC approach but demonstrated a slight underestimation utilizing UK68-RE and overestimation with PEDC-RE.
Conference/Value in Health Info
2015-11, ISPOR Europe 2015, Milan, Italy
Value in Health, Vol. 18, No. 7 (November 2015)
Code
PRM72
Topic
Methodological & Statistical Research
Topic Subcategory
Modeling and simulation
Disease
Diabetes/Endocrine/Metabolic Disorders