We performed post-hoc multivariate mediation analyses to evaluate the association of rs1051730 with nicotine dependence measures, nicotine dependence measure associations with 6MO abstinence, and rs1051730 direct effects on 6MO abstinence in the PLA and in the NRT PGs. We restricted these analyses to rs1051730 due to the significant effect sizes observed with this locus on 6MO abstinence with and without adjustment for multiple nicotine dependence measures. We observe a significant mediational path through the FTND score in the association of rs1051730 with 6MO abstinence in the NRT PG, but not in the PLA PG, perhaps due to sample size limitations (Fig. 2 and see Table, Supplemental Digital Content 9: Mediation of rs1051730 association with 6MO abstinence by nicotine dependence measures in individuals randomized to NRT and to PLA). The direct effect [OR (95%CI) P] of rs1051730 on abstinence with both FTND and CPD included in the mediation model is 2.73 (1.34–5.53) 0.005, and the pseudo-r2 is 0.083. rs1051730 is significantly and positively associated with CPD and with FTND score (p<0.001 and p=0.016, respectively). FTND score is significantly negatively associated with
