Thus, the 142 remaining children fulfilled the criteria for the overweight/Ob category (ie, BMI z score, > 1.20). Of these, 111 children (78.2%) had polysomnographic evidence of OSA, and 46 fulfilled the serologic criteria for FLD (32.4%). Of note, most of the children had isolated elevations of ALT, with only 13 of the 46 children also having increased AST levels. Therefore, most of the analyses regarding FLD were conducted using ALT concentrations only. There were no significant relationships between BMI z score and ALT levels for either the whole cohort or for those children with BMI z scores of > 1.20. Among the 46 Ob children with FLD, 34 were evaluated and had no serologic evidence of chronic liver disease. Twelve of the 46 children with suspected FLD also underwent ultrasonographic liver evaluations, and evidence for steatosis was found in all but 1 child. Of the 46 Ob children with FLD, 42 children (91.3%) had polysomnographic evidence of OSA (Table 3
). In addition, I/G ratios of > 0.30 were present in 26 children with FLD (56.5%). In contrast, OSA was present in only 69 of the 96 Ob children without elevated LFT levels (71.8%), and elevated I/G ratios occurred in 22 of 96 children without FLD (22.9%; p < 0.0001 [vs FLD]). Thus, the relative risk for OSA was significantly higher among Ob children compared to NOb children (OR, 3.05; 95% CI, 1.78 to 5.25; p < 0.0001); furthermore, the risk for OSA among children with FLD was also markedly higher compared to those children without FLD (OR, 4.11; 95% CI, 1.25 to 14.95; p < 0.01), even after controlling for age, gender, and race. Similarly, the odds for insulin resistance FLD were significantly higher in the presence of FLD, and mean cholesterol and triglyceride levels were also higher among those children with FLD (p < 0.0001) [Table 3]. Of note, there were no significant relationships between ALT or AST level and AHI, nadir Spo2, or any other polysomnographic measure. Stepwise logistic regression analysis for the entire cohort indicated that age, ethnicity, fasting I/G ratio, BMI z score of > 1.2, AHI of > 5 events per hour of TST, and nadir Spo2 of < 85% all significantly accounted for the variance in ALT levels (r2 = 0.68; p < 0.0001).