Multiple logistic regression analysis predicts cancer risk among tobacco usage with glutathione S-transferase p1 genotyping in patients with head and neck cancer.

Abstract

Numerous studies have been investigated to understand the association between glutathione S-transferase P1 (GSTP1) polymorphism and risk of head and neck cancer (HNC) but yielded contradictory results, and no studies could confirm polymorphism in GSTP1 and that tobacco usage increases the risk of HNCs. Therefore, this study aimed to understand the association of GSTP1 Ile105Val polymorphism with or without tobacco usage in carcinogenesis and clinicopathological characteristics of patients with HNC. Binary logistic regression analysis was performed to predict HNC risk with tobacco use and GSTP1 genotyping. Five predictor variables such as gender, age, tobacco usage, familial, and GSTP1 genotypes were included in the model. The results of the logistic regression analysis show that the full model which considered all the five independent variables together was statistically significant, log-likelihood = -111.820, and all slopes are zero: G = 74.297, degree of freedom (DF) = 5, P = 0.000. The strongest predictor in this model is tobacco usage (odds ratio = Z = -5.16, P = 0.000). The study concludes that multiple logistic regression analysis model could predict the risk factors in case-control studies where control samples are compromised.