Background: Liver fat is an important determinant of insulin requirements during insulin therapy. PPAR agonists reduce liver fat. We therefore hypothesized that type 2 diabetic patients using exceptionally high doses of insulin might respond well to addition of a PPAR agonist.

Methods: We determined the effect of the PPAR agonist rosiglitazone on liver fat and directly measured hepatic insulin sensitivity in 14 patients with type 2 diabetes (age 51±3 yrs, BMI 36.7±1.1 kg/m²), who were poorly controlled (HbA_1c 8.9±0.4%) despite using high doses of insulin (218±22 IU/day) in combination with metformin. Liver fat content (¹H-MRS), hepatic insulin sensitivity (6-h hyperinsulinemic euglycemic clamp [insulin 0.3 mU/kg·min] combined with of [3-³H]glucose), body composition (MRI), substrate oxidation rates (indirect calorimetry), clinical parameters, and liver enzymes were measured before and after rosiglitazone treatment (8 mg/day) for 8 months.

Results: During rosiglitazone, HbA_1c decreased from 8.9±0.4% to 7.8±0.3% (p=0.007) and insulin requirements from 218±22 to 129±20 IU/day (p=0.002). Liver fat content decreased by 46±9% from 20±3% to 11±3% (p=0.0002). Hepatic insulin sensitivity, measured from the % suppression of endogenous glucose production by insulin, increased from -40±7% to -89±12% (p=0.001). The % change in liver fat correlated with the % decrease in HbA_1c (r=0.53, p=0.06), insulin dose (r=0.66, p=0.014), and suppression of endogenous glucose production (r=0.76, p=0.003).

Conclusions: Our results suggest that rosiglitazone may be particularly effective in type 2 diabetic patients who are poorly controlled despite using high insulin doses. The mechanism is likely to involve reduced liver fat and enhanced hepatic insulin sensitivity.