This UK Biobank cohort study (n=19,255 European-ancestry adults, median age 65) examined whether MRI-derived pancreatic proton density fat fraction (PDFF), categorised by 2026 international consensus thresholds (normal <6%, mild 6–<16%, moderate-to-severe ≥16%), was independently associated with prevalent and incident T2D, CKD, and MACE over a median ~4.8-year follow-up.
Moderate-to-severe fatty pancreas (≥16% PDFF) was strongly linked to prevalent and incident T2D (OR 3.25; HR 2.72), incident CKD (HR 1.82), and both prevalent and incident MACE (OR 1.26; HR 1.30). Mild fatty pancreas (6–<16%) also raised incident T2D risk (HR 2.19) and incident MACE (HR 1.29). Each 5% PDFF increment raised T2D odds by 16% and hazard by 17%.
- Slice-based MRI measurement (not whole-organ volumetric), potentially missing regional fat heterogeneity; single time-point precludes tracking PDFF changes. - Outcomes from administrative ICD-10/OPCS-4 codes may underascertain T2D and CKD; HbA1c and eGFR were unavailable at imaging visit. - Restricted to European-ancestry participants; healthy volunteer bias may underestimate true population associations; results may not generalise to younger or more diverse groups.
When patients undergo abdominal MRI, reporting pancreatic PDFF using the 2026 consensus thresholds can flag elevated cardiometabolic risk—particularly for T2D and CKD—even after accounting for BMI and visceral fat. Pancreatic PDFF should complement, not replace, standard risk tools, and may help identify patients who warrant closer metabolic monitoring.
Explore related topics
