Association of Vitamin-D Receptor (Vdr), Gc-globulin and Cyp2r1 Polymorphism With Progression of Hbv Related Liver Disease Along With an Exclusive Svm Based Disease Prediction Model

Abstract Background: Polymorphism in genes associated with Vitamin-D metabolism and its receptor molecule altogether modulates the level of active serum Vitamin-D level and as such serves as potent immune modulator. Impact of this polymorphism on outcome of HBV related liver disease is poorly understood till date. Aim: This study focus on analysing the VDR polymorphism (TaqI, ApaI, and BsmI) and its associated molecules GC-Globulin and CYP2R1 among HBV infected patients and its probable correlation with disease progression. Method: Three hundred forty HBV infected patients belonging to three distinct clinical groups, Acute Viral hepatitis (AVH, n=205), Chronic Hepatitis (CH, n=84) and hepatocellular carcinoma (HCC, n = 51) along with 102 healthy control were included in the study. VDR, GC and CYP2R1 polymorphism were genotyped by PCR-RFLP method followed by Sanger sequencing. Haplotype distribution was analysed using SHEsis software. Logistic regression analysis was performed to find the association between different genotype and VDR haplotype with progression of liver disease. Further, an SVM based prediction model has also been described in detection of different disease stage. Results and Conclusion: Apa-I CC genotype was more frequently observed among chronic HBV patients and HCC than the acute cases among North east Indian population (OR = 2.241, C.I = 0.504- 5.558. p= 0.001). The occurrence of bAt haplotype was also found to be significantly higher among Chronic (p=0.004) and HCC (p=0.001<0.05) cases in comparison to acute cases. Further, logistic regression analysis after adjusting covariates like age, gender, serum AST/ALT level, serum albumin and platelet count, showed that Apa-I CC genotype and bAt haplotype were independent predictors for advancement of acute to chronic HBV infection and further. The SVM based prediction model using similar covariates also predicts the different disease stage with 90% accuracy.