Abstract
The association between cigarettes per day (CPD) and CPD-calibrated polygenic risk scores (CPD-PGS) is positive, however, the shape of the association is unknown. CPD measurement is inconsistent across studies and different measurement can lead to different results. The pattern shape may also patterns may change over time, given differences in genetic influence on smoking. This study examines the dynamic pattern between number of cigarettes smoked and PGS-CPD over adolescence and young adulthood. A time-varying effects model in which CPD was the continuous dynamic variable, was estimated for ever-smokers in a nationally representative study tracking partiicpants over adolescence and young adulthood. Participants were genotyped and a CPD-PGS score was created using results from a large genome-wide study meta-analysis. Results indicated that the association between CPD and CPD-PGS changed over CPD. Low CPD-PGS related to higher odds of not smoking or smoking very low CPD. A flat positive association at 5-12 CPD suggeted a "low risk" group. The association peaked around 20 CPD, indicating that a high-risk score applied best to those smoking approximately 1 pack of cigarettes per day. Age also moderated the effect of CPD-PGS at specific CPD rates, such that CPD-PGS was stronger at later ages and higher CPD. The association between CPD and CPD-PGS should not be assumed to be linear or static over age; there seem to be CPD-PGS thresholds corresponding to significant CPD risk, although the effect of CPD-PGS also varies over age. More care and attention to measurement can improve behavior genetic addiction science.