SAT356 Predicting Nonalcoholic Fatty Liver Disease in Adolescent Girls with Polycystic Ovary Syndrome and Obesity

Abstract

Abstract Disclosure: J.S. Lo: None. F. Bril: None. A. Taylor: None. N. Morelli: None. M.A. Ware: None. Y. Garcia-Reyes: None. M.J. McPhaul: Employee; Self; Quest Diagnostics. M.G. Cree: Consulting Fee; Self; Pollie, Inc. Research Investigator; Self; Amino Corp. Adolescents with obesity and polycystic ovary syndrome (PCOS) have significant insulin resistance and are at high risk of nonalcoholic fatty liver disease (NAFLD). This risk may be related to increased growth hormone and insulin resistance associated with puberty, in addition to excess testosterone. NAFLD is difficult to diagnose in obese individuals. Intact insulin (II), as measured by mass spectrometry, multiplied by alanine aminotransferase (ALT; II*ALT) has emerged as a strong predictor for NAFLD in adults (ROC 0.94 [0.89-0.99]). However, II*ALT has not been used in predict NAFLD in adolescents with PCOS. To assess the utility of II*ALT measurements in predicting NAFLD in the context of pubertal timing in a high-risk population of adolescents with PCOS and obesity. Participants were screened to confirm PCOS status per 2018 international guidelines. Liver fat fraction was measured with MRI Dixon, and fasting laboratory assessments were performed including traditional insulin (TI) and insulin growth factor-1 (IGF-1), measured with ELISA and ALT, AST. NAFLD was defined as a liver fat ≥5.0%. Mass spectrometry was used to measure intact molecules of II and C-peptide. Seventy-seven adolescent girls with PCOS and obesity who were Tanner 5 and at least 1-year post-menarche were enrolled (age 16.3 ± 1.7 years, BMI 34.9 ± 5.0 kg/m2). There was no difference in the prediction of NAFLD between II and TI (AUROC 0.72 [0.59-0.85] vs. 0.66 [0.53-0.79], p=0.39). Of note, the combination of II*ALT (ROC 0.83 [0.72-0.95], sensitivity 77%, specificity 65%) performed better in predicting NAFLD than II alone (ROC 0.72 [0.59-0.85]), ALT alone (ROC 0.81 [0.71-0.91]), and better than TI*ALT (ROC 0.75 [0.63-0.86]) for the entire cohort (p<0.01 for overall comparison). Patients were divided into 2 groups based on age as a surrogate for pubertal duration (Age <15: n=22, age 13.9 ± 0.8, BMI 32.7 ± 4.9 kg/m2, and Age >15: n=55, age 17.3 ± 0.9, BMI 35.8 ± 4.8 kg/m2). II*ALT had a higher ROC to predict NAFLD in the >15 group (ROC 0.90 [0.81-0.98], sensitivity 97%, specificity 60%) compared to the <15 group (ROC 0.71 [0.41-0.99] sensitivity 81%, specificity 83%), although it did not reach statistical significance. There was no effect of race and ethnicity on these relationships. The II*ALT index performs better than II and ALT alone for the detection of NAFLD and outperforms TI*ALT as well. It is a strong predictor for NAFLD in adolescents >15 years old with PCOS and obesity, with results very similar to those published in adults. However, it is not as robust in adolescents younger than 15 years old, perhaps due to the effect of pubertal hormones altering the relationship between liver fat and insulin concentrations. Presentation Date: Saturday, June 17, 2023