A Multivariate Logistic Model to Predict Grades of Non-Alcoholic Fatty Liver Disease in the Pakistani Population

Abstract

Introduction: Non-alcoholic fatty liver disease has a high global burden. The patients are usually diagnosed with abdominal ultrasound and are then graded based on the extent of the disease.

Objective: Our aim is to make a multivariate logistic model that can efficiently predict grades of NAFLD.

Methodology: An analytical cross-sectional study was conducted at the Radiology departments of Allied hospitals of Rawalpindi Medical University on a sample of 138 patients. Data was collected using a proforma. Demographic data and data of all the lab values were collected. Informed consent was taken from all the participants.

Results: The mean age of the patients was 44.14±13.193 years of which 52.2%(n=72) were males. 24% of the patients (n=33) were diabetics. Upon correlation analysis, ALT, AST, age, duration of diabetes, and BMI were found to be significantly correlated (p<0.05) with grades of BMI. All the correlated variables were included in the univariate and multivariate logistic regression analysis. On multivariate logistic regression analysis, BMI (AOR=5.66, p<0.05), Age (AOR=1.98, p<0.05), ALT (AOR=1.92, p<0.05), and AST (AOR=1.16, p<0.05) were found to be significant predictors of grades of NAFLD.

 Conclusion: We have successfully identified the predictors of the grades of NAFLD in the Pakistani population and have proposed a model based on these factors. However, our model still needs to be clinically validated.