Distribution patterns of intramyocellular and extramyocellular fat by magnetic resonance imaging in subjects with diabetes, prediabetes and normoglycaemic controls

• Published in: Diabetes, obesity & metabolism Vol. 23; no. 8; pp. 1868 - 1878
• Main Authors: Kiefer, Lena S., Fabian, Jana, Rospleszcz, Susanne, Lorbeer, Roberto, Machann, Jürgen, Kraus, Mareen S., Roemer, Frank, Rathmann, Wolfgang, Meisinger, Christa, Heier, Margit, Nikolaou, Konstantin, Peters, Annette, Storz, Corinna, Diallo, Thierno D., Schlett, Christopher L., Bamberg, Fabian
• Format: Journal Article
• Language: English
• Published: Oxford, UK Blackwell Publishing Ltd 01.08.2021; Wiley Subscription Services, Inc
• Subjects: body composition; Body mass index; Cohort analysis; cohort study; Diabetes; Diabetes mellitus (non-insulin dependent); Glucose metabolism; Image processing; Lipids; Magnetic resonance imaging; Metabolism; Musculoskeletal system; Obesity; Segmentation; Skeletal muscle; type 2 diabetes; body composition; cohort study; type 2 diabetes
• ISSN: 1462-8902; 1463-1326; 1463-1326
• DOI: 10.1111/dom.14413

Aim To evaluate the distribution of intramyocellular lipids (IMCLs) and extramyocellular lipids (EMCLs) as well as total fat content in abdominal skeletal muscle by magnetic resonance imaging (MRI) using a dedicated segmentation algorithm in subjects with type 2 diabetes (T2D), prediabetes and normoglycaemic controls. Materials and Methods Subjects from a population‐based cohort were classified with T2D, prediabetes or as normoglycaemic controls. Total myosteatosis, IMCLs and EMCLs were quantified by multiecho Dixon MRI as proton‐density fat‐fraction (in %) in abdominal skeletal muscle. Results Among 337 included subjects (median age 56.0 [IQR: 49.0‐64.0] years, 56.4% males, median body mass index [BMI]: 27.2 kg/m2), 129 (38.3%) were classified with an impaired glucose metabolism (T2D: 49 [14.5%]; prediabetes: 80 [23.7%]). IMCLs were significantly higher than EMCLs in subjects without obesity (5.7% [IQR: 4.8%‐7.0%] vs. 4.1% [IQR: 2.7%‐5.8%], P < .001), whereas the amounts of IMCLs and EMCLs were shown to be equal and significantly higher in subjects with obesity (both 6.7%, P < .001). Subjects with prediabetes and T2D had significantly higher amounts of IMCLs and EMCLs compared with normoglycaemic controls (P < .001). In univariable analysis, prediabetes and T2D were significantly associated with both IMCLs (prediabetes: β: 0.76, 95% CI: 0.28‐1.24, P = .002; T2D: β: 1.56, 95% CI: 0.66‐2.47, P < .001) and EMCLs (prediabetes: β: 1.54, 95% CI: 0.56‐2.51, P = .002; T2D: β: 2.15, 95% CI: 1.33‐2.96, P < .001). After adjustment for age and gender, the association of IMCLs with prediabetes attenuated (P = 0.06), whereas for T2D, both IMCLs and EMCLs remained significantly and positively associated (P < .02). Conclusion There are significant differences in the amount and distribution ratio of IMCLs and EMCLs between subjects with T2D, prediabetes and normoglycaemic controls. Therefore, these patterns of intramuscular fat distribution by MRI might serve as imaging biomarkers in both normal and impaired glucose metabolism.