MRI-derived liver fat measurements accurately predicted future diabetes risk in patients with obesity, according to findings published in Radiology.

In this retrospective study of 683 patients aged 18 to 65 years (mean age, 33 years; 18% men) who underwent MRI and an oral glucose tolerance test between September 2020 and December 2023. Ming Deng, MD, and colleagues of Zhongnan Hospital of Wuhan University identified a dose-dependent relationship between liver fat content and the likelihood of prediabetes or type 2 diabetes. Each 1% increase in liver proton density fat fraction (PDFF) was linked to a higher probability of both prediabetes and diabetes, and patients with severe steatosis (PDFF 25% or greater) had more than triple the likelihood of abnormal glucose metabolism compared with those without steatosis.

Participants were categorized by MRI-derived PDFF as nonfatty ( 5% or less), mild (5% to less than 15%), moderate (15% to less than 25%), or severe (25% or greater). PDFF values rose progressively with worsening glucose tolerance—14.5% among those with normal glucose tolerance, 19.2% in prediabetes, and 18.9% in type 2 diabetes. Elevated liver enzyme levels partially explained this relationship, with alanine aminotransferase and aspartate aminotransferase accounting for 39% and 27% of the association, respectively.

In an independent validation cohort of 100 patients, those with type 2 diabetes had higher mean PDFF values than those without (19.7% vs 12.9%). Greater liver fat content was also associated with higher fasting and 2-hour glucose, insulin, and C-peptide levels and greater insulin resistance, as measured by HOMA-IR.

Investigators noted that the cross-sectional design limited causal inference and that MRI-based PDFF was not confirmed histologically. Because all participants had obesity (BMI of 28 or greater), the results may not apply to leaner or more diverse populations. Although liver enzymes mediated part of the link between hepatic steatosis and dysglycemia, additional metabolic mechanisms may contribute, underscoring the need for longitudinal imaging studies to clarify causal pathways.

The study was funded by the National Key Research and Development Program of China, the Health Commission of Hubei Province, and institutional grants from Zhongnan Hospital of Wuhan University. The authors reported no relevant conflicts of interest.

Source: Radiology