Twelve blood proteins improve diabetes risk prediction
In 52,743 participants, adding the panel to a standard risk score improved ten-year diabetes prediction, raising a performance measure from 0.736 to 0.805.
Limit: Diabetes could cause the fat dysfunction used to identify the protein panel
Edited by Boris Djordjevic · AgingResearch.News
In an observational study, researchers analyzed fat tissue from human donors in the GTEx project. They examined fat under the skin from 168 donors with diabetes and 478 without, and fat around internal organs from 129 with diabetes and 407 without. Diabetes was associated with altered activity of 1,140 genes under the skin and 314 around internal organs. Under-skin fat cells were larger in donors with diabetes, but only in overweight individuals.
In 52,743 UK Biobank participants, 38 blood proteins differed in abundance in the same direction as corresponding genes’ activity in under-skin fat from overweight donors with diabetes. A panel of 12 proteins improved ten-year diabetes prediction beyond the Leicester Diabetes Risk Score. The proteins were also associated with heart failure, heart attacks, liver fat buildup, and atrial fibrillation, an irregular heartbeat.
Why it matters
The authors suggested that the protein panel may enable targeted metabolic intervention, potentially giving diabetes risk assessment a role in guiding prevention.
Caveats
The observational donor-tissue comparison could not show whether fat dysfunction preceded or followed diabetes. The findings were reported in a conference abstract.
- Machine learning
- ER unfolded protein response
- Extracellular matrix remodeling
- Diabetes
- Leicester Diabetes Risk Score
- Humans
