Pancreatic Fat and the Risk of Future Dysglycemia: A Systematic Review and Meta-analysis

Abstract Context Pancreatic fat has emerged as a metabolically relevant ectopic fat depot, but its longitudinal association with future dysglycemic outcomes remains incompletely defined. Objective To evaluate the longitudinal association of pancreatic fat with incident type 2 diabetes (T2D) and glycemic progression, with exploratory narrative synthesis of evidence in lean populations. Data Sources PubMed, Embase, Web of Science, and the Cochrane Library were searched from inception to March 1, 2026. Study Selection Longitudinal observational studies assessing pancreatic fat at baseline and reporting subsequent incident T2D, glycemic progression, or both were included. Data Extraction Two reviewers independently screened studies and extracted data. Methodological quality was assessed using the Newcastle-Ottawa Scale. Data Synthesis Ten studies were included. In the primary binary meta-analysis (5 studies), higher pancreatic fat burden was associated with incident T2D (pooled effect estimate, 2.56; 95% CI, 1.27-5.14; I2 = 93.3%). Exclusion of 1 influential ultrasound-based study attenuated heterogeneity while preserving the association (pooled effect estimate, 1.45; 95% CI, 1.23-1.73; I2 = 10.2%). A separate continuous analysis (3 studies) also supported an association with incident T2D (1.17; 95% CI, 1.04-1.32; I2 = 79.2%), although exposure scales were not directly comparable. Pancreatic fat was also associated with glycemic progression (2 studies; 1.96; 95% CI, 1.10-3.48; I2 = 73.2%). Exploratory lean-population evidence was limited to 2 analytical contexts, synthesized narratively, and directionally consistent with the overall findings. Conclusions Pancreatic fat was associated with adverse future glycemic outcomes across multiple analytical contexts. These findings support pancreatic fat as a potentially meaningful imaging-derived marker of future dysglycemia, while suggesting that heterogeneity is partly explained by differences in exposure ascertainment.