previously suggested (Hughes et al. 2003), and those individuals still smoking by the time of the interview were treated as censored. We excluded subjects (N = 70) abstinent for less than 6 months prior to the interview because their cessation status cannot be deduced. Further, we excluded 55 subjects who reported quitting due to health reasons, as genetic background may have little or no effect on quitting success in such a situation. The basic characteristics of the data used in the analyses are listed in Table 2.
