External validation of SCORE2-Diabetes risk prediction model in Chinese patients with diabetes
Authors' conclusion
With the appropriate selection of the scaling factor, SCORE2-Diabetes shows satisfactory discrimination and calibration in Chinese, indicating its good generalisability.
Study context
- Population
- Chinese patients with diabetes aged 40–90 without prior cardiovascular disease (CVD)6,540 CVD events
- Participants
- 50,426 patients
- Comparison
- Original vs recalibrated SCORE2-Diabetes
- Outcome
- 10-year CVD risk prediction (discrimination and calibration); outcome defined consistently with SCORE2-Diabetes
- Follow-up
- Median 7.09 years
Methodological detailspopulation-based cohort, recalibration with region-specific scaling factors
- Population-based cohort
- Recalibration with region-specific scaling factors
Performance
As reported in the abstract
| Test or model | C-statistic | Original SCORE2-Diabetes calibration | Recalibrated (high-risk region scaling factor) calibration |
|---|---|---|---|
| Men | 0.702 (95%CI: 0.691-0.713) | underestimated risk by 30% | overestimating risk by only 4% |
| Women | 0.726 (0.716-0.736) | underestimated risk by 50% | underestimating by 7% |
C-statistic
Men0.702 (95%CI: 0.691-0.713)
Women0.726 (0.716-0.736)
Original SCORE2-Diabetes calibration
Menunderestimated risk by 30%
Womenunderestimated risk by 50%
Recalibrated (high-risk region scaling factor) calibration
Menoverestimating risk by only 4%
Womenunderestimating by 7%
Further findings
Risk category distribution
- Very high risk (≥20%)
- 14,501(28.76%)
- High risk (10-20%)
- 20,789(41.23%)
Baseline vs targeted LDL-C, very high-risk group
- Average baseline vs targeted LDL-C (significant difference)
- 2.85 vs. 1.40 mmol/L
Non-CVD deaths
- Non-CVD deaths
- 2,852
Original publication
External validation of SCORE2-Diabetes risk prediction model in Chinese patients with diabetes. Eur J Prev Cardiol. Published 17 April 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →