INTRODUCTION: The aim of this review was to report an update of a previous meta-analysis (by another group) of a possible relationship between serotonin transporter (5-HTTLPR) genotype and smoking behavior, and extend previous work by factoring in some demographic parameters (age, gender, and ethnicity) in a multiple regression model to examine the relationship between these demographic factors and the effect of 5-HTTLPR polymorphism on smoking behavior. METHODS: Effect sizes were calculated for each study selected for meta-analysis and were pooled using the random-effects model, which assumes within-study sampling and between-study variance and provides wider confidence intervals. Effect sizes calculated in each study were used to evaluate the correlations with participant data for age, gender, and ethnicity (moderating variable) by multiple regression analysis. RESULTS: Meta-analysis indicated a relationship between smoking rate and the 5-HTTLPR genotype, but not smoking initiation and persistence, which was consistent with that of the previous review. Publication bias was not indicated for smoking initiation or persistence. Multiple regression analysis revealed that mean participant age significantly affected effect sizes for smoking initiation and persistence of each study. The proportion of Caucasians may have been partially influenced by the difference in effect sizes for smoking persistence among the studies. CONCLUSION: A significant relationship stratified by ancestry was observed between the 5-HTTLPR genotype and smoking rate, but not between the 5-HTTLPR genotype and smoking initiation and persistence. Regression analysis detected effects of age and/or ethnicity as moderating factors on smoking initiation and persistence.
INTRODUCTION: The aim of this review was to report an update of a previous meta-analysis (by another group) of a possible relationship between serotonin transporter (5-HTTLPR) genotype and smoking behavior, and extend previous work by factoring in some demographic parameters (age, gender, and ethnicity) in a multiple regression model to examine the relationship between these demographic factors and the effect of 5-HTTLPR polymorphism on smoking behavior. METHODS: Effect sizes were calculated for each study selected for meta-analysis and were pooled using the random-effects model, which assumes within-study sampling and between-study variance and provides wider confidence intervals. Effect sizes calculated in each study were used to evaluate the correlations with participant data for age, gender, and ethnicity (moderating variable) by multiple regression analysis. RESULTS: Meta-analysis indicated a relationship between smoking rate and the 5-HTTLPR genotype, but not smoking initiation and persistence, which was consistent with that of the previous review. Publication bias was not indicated for smoking initiation or persistence. Multiple regression analysis revealed that mean participant age significantly affected effect sizes for smoking initiation and persistence of each study. The proportion of Caucasians may have been partially influenced by the difference in effect sizes for smoking persistence among the studies. CONCLUSION: A significant relationship stratified by ancestry was observed between the 5-HTTLPR genotype and smoking rate, but not between the 5-HTTLPR genotype and smoking initiation and persistence. Regression analysis detected effects of age and/or ethnicity as moderating factors on smoking initiation and persistence.