Table 4: Multiple regression analysis for annual BMI gain and variables in Japanese men.

Estimate SE value Pr Intercept 0.193 0.097 1.998 0.047* Age −0.005 0.002 −2.326 0.021* Alcohol intake 0.000 0.000 −0.649 0.517 Current exercise 0.002 0.027 0.067 0.947 Working style 0.001 0.038 0.015 0.988 Smoking status −0.053 0.028 −1.883 0.061 ADRB3 (Trp/Arg) 0.009 0.038 0.227 0.821 ADRB3 (Arg/Arg) 0.218 0.079 2.761 0.006** LDL 0.002 0.001 3.245 0.001**

From the stepwise multiple regression analysis, age, alcohol intake, current exercise, working style, smoking status, LDL, and the ADRB3 polymorphism were selected for the final model of the predictor of BMI gain, adjusted (F-statistic = 3.224, ).