Abstract
Background and aim: To study the association between achievement of guideline-defined treatment targets on HbA1c, low-density lipoproteins (LDL-C), and blood pressure with the progression of diabetic complications in patients with type 1 diabetes (T1D). Methods: The study included 355 patients at baseline and 114 patients with follow-up data after 3–5 years. Outcome variables were the progression of diabetic kidney disease, retinopathy, or cardiovascular disease (CVD). We used logistic regression and other machine learning algorithms (MLA) to model the association of achievement of treatment targets and probability of progression of complications. Results: Achievement of the target blood pressure was associated with 96% lower odds of a new CVD event (0.04 (95% CI 0.00, 0.53), p = 0.016), and 72% lower odds of progression of any complication (0.28 (95% CI 0.09, 0.89), p = 0.027. Achievement of HbA1c target was associated with lower odds of composite complication progression by 82% (0.18 (95% CI 0.04, 0.88), p = 0.034.) None of the patients who achieved HbA1c target progressed in CVD. MLA demonstrated good accuracy for the prediction of progression of CVD (AUC 0.824), and lower accuracy for other complications. Conclusion: The achievement of blood pressure and HbA1c treatment targets is associated with lower odds of vascular complication of T1D in a real life study.
| Original language | English |
|---|---|
| Article number | 108072 |
| Pages (from-to) | 1-8 |
| Journal | Journal of Diabetes and its Complications |
| Volume | 35 |
| Issue number | 12 |
| DOIs | |
| Publication status | Published - Dec 2021 |
UN SDGs
This output contributes to the following UN Sustainable Development Goals (SDGs)
-
SDG 3 Good Health and Well-being
Keywords
- Complication
- Machine learning
- Treatment targets
- Type 1 diabetes
Fingerprint
Dive into the research topics of 'Achievement of treatment targets predicts progression of vascular complications in type 1 diabetes'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver