Body fat topography as an independent predictor of fatty liver

Metabolism. 1993 May;42(5):548-51. doi: 10.1016/0026-0495(93)90210-f.

Abstract

Abdominal (truncal) fat distribution reflected by an elevated waist to hip ratio (WHR) predicts metabolic abnormalities such as diabetes and dyslipidemia as well as hypertension and stroke, all of which are associated with obesity. The pathogenesis is not known, although elevated splanchnic serum free fatty acid levels and reduced hepatic insulin clearance have been implicated. WHR and body fat (BF) by 40K-counting and 3H2O were measured before liver biopsy during antiobesity surgery in 68 severely obese women (body mass index [BMI], 48.9 +/- 1.1 SEM) and 15 men (BMI, 49.0 +/- 3.1) without histories of liver disease, diabetes, or hepatotoxic exposure. Biopsies were graded for fat content semiquantitatively (0 to 4+) by the hepatologist who was blinded to the patients' clinical characteristics. All 15 men had fatty infiltration (score, 2.5 +/- 0.3 v 1.4 +/- 0.1 in women; P < .001). The correlation between WHR and liver fat was .44 (P < .0005), while BF (-.16), weight (.15), or BMI (.04) did not correlate significantly with steatosis (all NS). As expected, percentage body fat (BF%) was greater in women than in men (40.3 +/- 0.8 kg v 33.9 +/- 2.0, P < .007), and accordingly liver fat was inversely related to BF% (r = -.32, P < .002). Steatosis was significantly greater in 14 men (2.5 +/- 0.3) than in 20 women (1.7 +/- 0.3, P < .04) matched for BF%. In multiple regression analysis R2 = .49, P < .0001), WHR and sex accounted for the variance in liver fat content without any further contribution from weight, BMI, BF, or BF%.(ABSTRACT TRUNCATED AT 250 WORDS)

MeSH terms

  • Adipose Tissue / pathology*
  • Adolescent
  • Adult
  • Biopsy
  • Body Composition*
  • Body Mass Index
  • Fatty Liver / complications
  • Fatty Liver / pathology*
  • Female
  • Forecasting
  • Humans
  • Male
  • Middle Aged
  • Obesity, Morbid / complications
  • Obesity, Morbid / pathology*
  • Reference Values
  • Regression Analysis