Validation of a Non-Laboratory Based Screening Tool for Predicting Non-Alcoholic Fatty Liver Disease in an Egyptian Setting

Abstract

BackgroundNon-alcoholic fatty liver disease (NAFLD) is a common chronic liver disorder in more than 20% of the general population worldwide. Several combinations of non-invasive factors and scoring models were investigated as indicators of NAFLD. This study aimed to validate and adapt an established fatty liver score, which allows the identification of NAFLD based on routinely available clinical and laboratory data. Materials and MethodsThe study cohort comprised 190 adults seeking health check-up at the out-patient clinic of a tertiary care hospital in Alexandria, Egypt. Anthropometric, clinical, and laboratory data were recorded and the status of fatty liver was diagnosed by abdominal ultrasound. A logistic regression model was built to determine the predictors of NAFLD. The performance of the derived risk scores was compared to other existing models. ResultsObesity (60.0%), metabolic syndrome (42.6%), and NAFLD (56.8%) were predominant features among the study population. Smoking [OR (95% CI) = 4.4 (0.9–21.4)], obesity [OR (95% CI) = 4.0 (1.7–9.7)], hypertension [OR (95% CI) = 2.4 (1.03–5.5)], elevated serum total cholesterol [OR (95% CI) = 4.8 (1.8–13.1)], triglycerides [OR (95% CI) = 11.8 (2.3–661.02)], and ALT [OR (95% CI) = 4.8 (1.8–13.1)] were multivariate predictors of NAFLD. A NAFLD screening questionnaire with values applicable for Egyptians was adapted from an existing model after validation. A total score ≥7 was suggestive of NAFLD [AUC = 0.810 (0.749–0.871); sensitivity = 87.0%; specificity = 62.2%; PPV = 75.2%; NPV = 78.5%]. ConclusionsNAFLD can be sufficiently predicted among apparently healthy Egyptians by a tempted simple and non-invasive scoring index although external validation is warranted.