AJR:198, May 2012
Perry J. Pickhardt Young Jee Stacy D. O'Connor Alejandro Munoz del Rio
OBJECTIVE. Visceral adiposity and hepatic steatosis may correlate with the metabolic syndrome but are not currently among the diagnostic criteria. We evaluated these features at unenhanced MDCT.
MATERIALS AND METHODS. Semiautomated measurements of subcutaneous fat area, visceral fat area, and visceral fat percentage were obtained at the umbilical level at unenhanced MDCT of 474 adults (217 men, 257 women; mean age, 58.3 years) using a dedicated application (Fat Assessment Tool, EBW version 4.5). Unenhanced liver attenuation was also recorded. Metabolic syndrome was defined using the criteria proposed by the International Diabetes Federation in 2005.
RESULTS. The prevalence of metabolic syndrome was 35.0% (76/217) among men and 35.8% (92/257) among women. The area under the receiver operating characteristic curve (AUC) for visceral fat area was 0.830 (95% CI, 0.784-0.867) in men and 0.887 (0.848-0.918) in women (p = 0.162). The AUC for subcutaneous fat area was 0.865 (0.823-0.899) in men and 0.762 (0.711-0.806) in women (p = 0.024). The AUC for visceral fat percentage was 0.527 (0.472-0.581) in men and 0.820 (0.774-0.859) in women (p < 0.001). The AUC for liver attenuation was 0.706 (0.653-0.754). Thresholds of subcutaneous fat area greater than 204 cm2 in men, visceral fat area greater than 70 cm2 in women, and liver attenuation less than 50 HU yielded a sensitivity and specificity of 80.3% and 83.7%; 83.7% and 80.0%; and 22.0% and 96.7%, respectively. Visceral fat area was elevated in 55% of patients without metabolic syndrome (11/20) but with a documented cardiovascular event or complication and in 32.1% of patients with a body mass index of 30 kg/m2 or less.
CONCLUSION. Accumulation of visceral fat was the best predictor for metabolic syndrome in women. Unexpectedly, the percentage of visceral fat was a poor predictor for metabolic syndrome in men and subcutaneous fat area was best. Decreased liver attenuation was insensitive but was highly specific for metabolic syndrome. The implications of these sex-specific differences and the relationship of fat-based CT measures to cardiovascular risk warrant further investigation.