Development and validation of SCORE2-Diabetes, a European 10-year CVD risk model for T2DM
SCORE2-Diabetes: 10-year cardiovascular risk estimation in type 2 diabetes in Europe
Introduction and methods
Several CVD risk prediction models for patients with T2DM include DM-related information, such as levels of HbA1c and kidney function markers [1-4]. However, these DM-specific models have potential limitations for use in Europe, particularly as they do not reflect the wide regional variation in CVD incidence across Europe [4-6]. To address these limitations, the European Society of Cardiology has summoned an effort to extend the regionally “recalibrated” (i.e., statistically adapted) European SCORE2 10-year risk models .
Aim of the study
The study aim was to develop, validate, and illustrate SCORE2-Diabetes, a recalibrated prediction model to estimate the 10-year CVD risk in T2DM patients in Europe.
SCORE2-Diabetes was developed based on the original SCORE2 algorithms, which include conventional risk factors (i.e., age, sex, smoking status, systolic blood pressure (SBP), total cholesterol, and HDL-c). These models were adapted for use in T2DM patients by adding DM-related variables (i.e., age at T2DM diagnosis, HbA1c level, and creatinine-based eGFR). For this purpose, individual-participant data were extracted from 4 large-scale population datasets across 7 countries (England, Wales, Scotland, France, Germany, Italy, and the US), comprising 229,460 patients aged >40 years with T2DM but with no previous CVD (43,706 CVD events).
The derived risk models were recalibrated to 4 European risk regions, defined according to CVD mortality rates (low, moderate, high, and very high risk). External validation was performed using 4 data registries from Sweden, Spain, Malta, and Croatia, comprising 217,036 individuals (38,602 CVD events). The authors illustrated the variation in CVD risk in T2DM patients across European regions by applying the recalibrated models to data from contemporary populations in each risk region.
The primary endpoint was a composite outcome of CVD events (CV death, nonfatal MI, and nonfatal stroke).
Internal and external validations and recalibration
- In the internal validation, the C index ranged from 0.666 (95%CI: 0.653–0.678) to 0.733 (95%CI: 0.727–0.739), depending on the dataset used.
- In the external validation, SCORE2-Diabetes showed improved risk discrimination over SCORE2, with a C-index increase ranging from 0.009 to 0.031.
- After recalibration, the SCORE2-Diabetes predicted risks showed good agreement with the expected 10-year CVD incidence in each European risk region. It also showed improved calibration over SCORE2.
Illustration of CVD risk distribution in European risk regions
- The SCORE2-Diabetes estimated absolute CVD risk for a particular age and combination of conventional CVD risk factors differed substantially according to a patient’s levels of the DM-related variables.
- For example, in the moderate-risk region, the estimated 10-year CVD risk was 11.0% for a 60-year-old, non-smoking male who was newly diagnosed with T2DM, had average levels of conventional risk factors (i.e., SBP of 140 mmHg, total cholesterol of 5.5 mmol/L, and HDL-c of 1.3 mmol/L), HbA1c of 50 mmol/mol and eGFR of 90 mL/min per 1.73 m². A woman with the same characteristics would have an estimated risk of 7.9%.
- In contrast, the estimated 10-year CVD risk in a similar man with less favorable DM-related factors (i.e., HbA1c of 70 mmol/mol and eGFR of 60 mL/min per 1.73 m²) and age at T2DM diagnosis of 50 years was 17.2%. For a woman with the same characteristics, the estimated risk was 12.7%.
- When the recalibrated SCORE2-Diabetes models were applied to simulated data representing populations from each risk region, the proportion of individuals aged 40–79 years with an estimated CVD risk ≥10% varied substantially by region, ranging from 61% in the low-risk region to 96% in the very-high-risk region in men and from 51% to 94%, respectively, in women. As expected, proportions increased with age.
This study detailed the development, recalibration, and validation of SCORE2-Diabetes, an extension of the SCORE2 risk models that was tailored to predict 10-year CVD risk in patients with T2DM across 4 distinct European regions. SCORE2-Diabetes had good ability to discriminate and provide individual risk estimates for T2DM patients, taking into account their specific risk factors such as age at T2DM diagnosis, HbA1c level, and kidney function.
The authors believe their prediction model can “be used to help guide clinicians and individual patients for considering the intensity of existing treatment (such as lipid-lowering therapies), as well as additional interventions to prevent CVD (such as [SGLT2] inhibitors or [GLP-1RAs]).”