Reply:We thank Drs. Manco and Nobili for reading our work and finding it interesting. Given the cross-sectional nature of the study we are unable to discern the sequence of metabolic events that might lead ultimately to insulin resistance and its complications. Due to the strong colinearity between visceral fat and intrahepatic fat accumulation, it is difficult to fully determine their relationship with glucose dysregulation, because both are strongly linked to gluconeogenesis. 1,2 Nevertheless, using multivariate analysis we found that hepatic fat content predicts the 2-hour plasma glucose independent of visceral fat (r 2 0.20, P Ͻ 0.04). Thus, hepatic fat content is a better predictor of prediabetes than visceral fat in obese adolescents.We agree that measurement of hepatic fat content could be a useful clinical tool to identify obese patients who are at increased risks for diabetes and cardiometabolic disease. Across a wide range of hepatic fat content, we found increasing worsening of insulin resistance, without an obvious threshold effect. Although hepatic insulin sensitivity was not directly measured in this study, other researchers have shown a monotonic increase in hepatic insulin resistance across varying degrees of hepatic steatosis. 2 Accurate assessment of hepatic fat content is essential in the clinical evaluation of the obese child presenting with cardiometabolic risks factors. Unfortunately, however, none of the available techniques are able to diagnose the presence of steatohepatitis and fibrosis, which can only be detected by liver biopsy. To the Editor:The recently published article by Korman et al. 1 suggests that the sequential use of two scores (alkaline phosphatase/total bilirubin ratio Ͻ 4 [94% sensitivity, 96% specificity] and aspartate aminotransferase/alanine aminotransferase ratio Ͼ 2.2 [94% sensitivity, 86% specificity]) in a population with acute liver failure can identify those with Wilson's disease with 100% sensitivity and 100% specificity. This is a statistical impossibility.Use of the first score will correctly identify 94% of patients with Wilson's disease and 96% of those without Wilson's disease. Applying the second score to the group with an alkaline phosphatase/total bilirubin ratio Ͻ 4 may improve the specificity, but the sensitivity (within the total population) cannot be greater than 94%. Applying the second score to that group with an alkaline phosphatase/total bilirubin ratio Ͼ 4 may improve the sensitivity, but the specificity (within the total population) cannot be greater than 86%. Applying the second score to both groups selected by the first score (i.e., the whole population) will result in 94% sensitivity and 86% specificity.
DR. DAMIAN DOWLING