Insulin sensitivity (euglycemic clamp, insulin infusion rate: 40 mU.m(-2).m
in(-1)) was studied in 30 subjects with biopsy-proven nonalcoholic fatty li
ver disease (NAFLD), normal glucose tolerance, and a BMI < 30 kg/m(2). Of t
hose 30 subjects, 9 had pure fatty liver and 21 had evidence of steatohepat
itis. In addition, 10 patients with type 2 diabetes under good metabolic co
ntrol and 10 healthy subjects were studied.-Most NAFLD patients had central
fat accumulation, increased triglycerides and uric acid, and low HDL chole
sterol, irrespective of BMI. Glucose disposal during the clamp was reduced
by nearly 50% in NAFLD patients, as well as in patients with normal body we
ight, to an extent similar to that of the type 2 diabetic patients. Basal f
ree fatty acids were increased, whereas insulin-mediated suppression of lip
olysis was less effective (-69% in NAFLD vs. -84% in control subjects; P =
0.003). Postabsorptive hepatic glucose production (HGP), measured by [6,6-H
-2(2)]glucose, was normal. In response to insulin infusion, HGP decreased b
y only 63% of basal in NAFLD vs. 84% in control subjects (P = 0.002). Compa
red with type 2 diabetic patients, NAFLD patients were characterized by low
er basal HGP, but with similarly reduced insulin-mediated suppression of HG
P. There was laboratory evidence of iron overload in many NAFLD patients, b
ut clinical, histological, and biochemical data (including insulin sensitiv
ity) were not correlated with iron status. Four subjects were heterozygous
for mutation His63Asp of the HFE gene of familiar hemochromatosis. We concl
uded that NAFLD, in the presence of normoglycemia and normal or moderately
increased body weight, is characterized by clinical and laboratory data sim
ilar to those found in diabetes and obesity. NAFLD may be considered an add
itional feature of the metabolic syndrome, with specific hepatic insulin re
sistance.