The aim of this study was to check the reliability of sonography in measuring small variations in quantities of subcutaneous and intra-abdominal fat. Twenty-six obese women (BMI 39 +/- 6) underwent a 15 day very low calorie diet. The study included, both before and after very low calorie diet, computed tomography measurements of total (AT), visceral (VAT) and subcutaneous (SAT) adipose tissue areas, visceral/subcutaneous area ratio (V/S), waist/hip circumference ratio measurements (W/H), and ultrasound measurements of abdominal subcutaneous skin-muscle thickness, intra-abdominal muscle-aorta thickness and intra-abdominal/subcutaneous thickness ratio. Weight reduction was from 101 +/- 17 to 95 +/- 16 kg (P less than 0.001). W/H dropped from 0.83 +/- 0.06 to 0.82 +/- 0.07 (n.s.). VAT dropped from 158 +/- 72 to 134 +/- 61 cm2 (P less than 0.005), SAT from 572 +/- 151 to 566 +/- 164 cm2 (n.s.) and V/S from 0.29 +/- 0.15 to 0.25 +/- 0.11 (P less than 0.01). Abdominal subcutaneous fat thickness decreased from 36 +/- 8 to 35 +/- 10 mm (n.s.), intra-abdominal thickness from 39 +/- 25 to 20 +/- 20 mm (P less than 0.001) and intra-abdominal/subcutaneous from 1.1 +/- 0.7 to 0.8 +/- 0.6 (P less than 0.005). VAT measurement accurately identified small intra-abdominal fat variations. W/H could not evaluate visceral fat loss, because of simultaneous decreases in waist and hip circumferences. Ultrasound was able to measure small reductions in intra-abdominal fat.