Abstract
BACKGROUND: Pancreas fat accumulation has been associated with impaired glucose metabolism and stimulation of inflammation. The associations of pancreatic fat fraction (PFF) with adverse outcomes remain understudied. METHODS: We estimated baseline predicted PFF (pPFF) from abdominal magnetic resonance imaging (MRI) in 29 447 UK Biobank participants. Since the initial imaging visit occurred 9.3 [interquartile range (IQR) 8.2-10.2] years after baseline visit, we calculated annual change in PFF using linear regression in 337 participants who underwent repeat imaging 2.8 (IQR 2.2-4.8) years after the initial scan. We back-calculated pPFF by adding this change to each participant based on the time between baseline visit and initial MRI visit assuming linear change in PFF over time. Multivariable-adjusted proportional hazards models tested associations of pPFF with incident chronic kidney disease (CKD), cardiovascular disease (CVD) and death. RESULTS: Participants had a mean ± standard deviation age of 56 ± 8 years, mean eGFR of 98 ± 13 mL/min/1.73 m(2), median (IQR) albuminuria of 5.7 (3.5-9.2) mg/g and median pPFF of 4.7 (1.6-10.3)%. During median follow-up of 13.8 years, there were 572 incident CKD, 1048 incident CVD and 468 death events. In final adjusted models, participants in the 4th, 3rd and 2nd quartiles had 1.58-, 1.42- and 1.76-fold higher risks of incident CKD compared with the 1st quartile, respectively. Participants in the 4th quartile had 1.27-fold higher risk of incident CVD compared with the lowest quartile. The associations of pPFF with incident CKD and CVD were each non-linear (each P < .001). pPFF was not associated with death. CONCLUSIONS: Higher pPFF levels are associated with higher risk of incident CKD and CVD but not death. Future studies should investigate whether pancreas fat depots should serve as a target of current or emerging therapies.