Risk factors and prediction of hypoglycaemia using the Hypo-RESOLVE cohort: a secondary analysis of pooled data from insulin clinical trials
Authors' conclusion
Although hypoglycaemia rates show large variation according to sociodemographic and clinical characteristics and treatment history, looking at a 6 week period of hypoglycaemia events and glucose measurements predicts future hypoglycaemia risk.
Study context
- Population
- People with T1D or T2D in pooled insulin clinical trials across pharmaceutical and device companies (Hypo-RESOLVE)
- Participants
- 46,254 people90 trials
- Comparison
- T1D vs T2D
- Primary endpoint
- Prediction (AUC) of incident hypoglycaemia events
- Follow-up
- Whole trial duration (mean 34.8 weeks T1D, 25.3 weeks T2D)
Methodological detailspooled individual participant data from 90 trials, Poisson mixed models, XGBoost prediction models
- Pooled individual participant data from 90 trials
- Poisson mixed models
- XGBoost prediction models
Primary result
As reported in the abstract
Prediction of level 1 or worse hypoglycaemia · whole follow-up period
| Group | AUC |
|---|---|
| T1D | 0.835 (95% CI 0.826, 0.844) |
| T2D | 0.840 (95% CI 0.831, 0.848) |
AUC
T1D0.835 (95% CI 0.826, 0.844)
T2D0.840 (95% CI 0.831, 0.848)
Further findings
| Outcome | T1D | T2D |
|---|---|---|
| Prediction of level 3 hypoglycaemia (requiring assistance) · whole follow-up period | ||
| AUC | 0.689 (95% CI 0.667, 0.712) | 0.705 (95% CI 0.662, 0.748) |
Factors associated with more frequent hypoglycaemia (T1D and T2D)
- Female sex, white ethnicity, longer DM duration, human vs analogue-only insulin, higher glucose variability, higher 6-week baseline hypoglycaemia score, lower BP, lower lipid levels and psychoactive drugs were associated with more frequent hypoglycaemia
Sources of improvement in prediction over baseline models
- Almost all improvement captured by hypoglycaemia history, glucose variability and blood glucose over a 6-week baseline period
Original publication
Risk factors and prediction of hypoglycaemia using the Hypo-RESOLVE cohort: a secondary analysis of pooled data from insulin clinical trials. Diabetologia. Published 25 May 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →