Table 3. Characteristics of the variables including in the optimized logistic model

The logistic regression model was optimized using the backward stepwise approach. In this model, age, smoking status, single nucleotide polymorphysm (SNP) rs3753394 and rs1061170 were included. Missing data of smoking status was excluded from the model. Therefore, the model was built with 343 samples (153 cases and 190 controls) with full data set. Asterisk (*) represents the p value obtained from the model was comparing category 1 with baseline (category 0) in cases and controls. Vertical arrow (+) represents the p value obtained from the model was comparing category 2 with baseline (category 0) in cases and controls. Double plus symbol (++) represents the SNPs rs800292 and rs1329428 were not included in the optimized model by backward stepwise approach. Here shows the coefficients and odds ratios in other models using "Enter" approach, adjusted for age, smoking and rs1061170.