Lipidomic profiles and their relationship to incident diabetic retinopathy in type 2 diabetes: The Fremantle Diabetes Study Phase II.
Timothy M E Davis, Jocelyn J Drinkwater, Alanah Grant-St James et al.
✦ AI-curated · Sources linked
In 30 seconds
This exploratory longitudinal study examined whether baseline plasma lipid profiles predicted diabetic retinopathy (DR) in 762 adults with type 2 diabetes. DR was assessed at baseline and Year 4/6, and lipid profiles were measured using mass spectrometry. New or worsening DR occurred in 121 participants (16%); among 495 without baseline DR, 35 (7.0%) developed DR. Adding selected lipid biomarkers to conventional risk factors improved model performance for both outcomes, with a 28.3% net reclassification improvement for incident DR at a 5% risk cut-off.
Key findings
- New or worsening DR occurred in 121 participants (16%); incident DR developed in 35 of 495 participants without baseline DR (7.0%).
- For new or worsening DR, three lipid biomarkers improved prediction beyond conventional risk factors (likelihood ratio test, P = 0.00002); NRI gain at a 10% cut-off was 5.7% (SE 2.8%; P = 0.043).
- For incident DR, three lipid biomarkers alongside HbA1c improved model performance (likelihood ratio test, P = 0.00001); NRI gain at a 5% risk cut-off was 28.3% (P = 0.002).
Why it matters
DR risk prediction may be improved by lipid measures in addition to established clinical factors, potentially helping clarify metabolic pathways associated with retinal disease. The findings support further evaluation of lipid biomarkers in risk models, while the exploratory nature of the analysis means their predictive value and clinical utility require confirmation.
What to keep in mind
The study is described as exploratory; the abstract does not report external validation of the prediction models.