Adipose tissue morphology predicts improved insulin sensitivity following moderate or pronounced weight loss

Cross-sectional studies show that white adipose tissue hypertrophy (few, large adipocytes), in contrast to hyperplasia (many, small adipocytes), associates with insulin resistance and increased risk of developing type 2 diabetes. We investigated if baseline adipose cellularity could predict improvements in insulin sensitivity following weight loss.

Plasma samples and subcutaneous abdominal adipose biopsies were examined in 100 overweight or obese individuals before and 10 weeks after a hypocaloric diet (7±3% weight loss) and in 61 obese subjects before and 2 years after gastric by-pass surgery (33±9% weight loss). The degree of adipose tissue hypertrophy or hyperplasia (termed the morphology value) in each individual was calculated on the basis of the relationship between fat cell volume and total fat mass. Insulin sensitivity was determined by homeostasis model assessment-estimated insulin resistance (HOMAIR).

In both cohorts at baseline, subjects with hypertrophy displayed significantly higher fasting plasma insulin and HOMAIR values than subjects with hyperplasia (P<0.0001), despite similar total fat mass. Plasma insulin and HOMAIR were normalized in both cohorts following weight loss. The improvement (delta insulin or delta HOMAIR) was more pronounced in individuals with hypertrophy, irrespective of whether adipose morphology was used as a continuous (P=0.0002-0.027) or nominal variable (P=0.002-0.047). Absolute adipocyte size associated (although weaker than morphology) with HOMAIR improvement only in the surgery cohort. Anthropometric measures at baseline (fat mass, body mass index, waist-to-hip ratio or waist circumference) showed no significant association with delta insulin or delta HOMAIR.

In contrast to anthropometric variables or fat cell size, subcutaneous adipose morphology predicts improvement in insulin sensitivity following both moderate and pronounced weight loss in overweight/obese subjects.