In this article, the authors assessed whether continuously scaled symptom parameters derived from growth models (T. M. Piasecki et al., 2003) are linked to smoking at long-term follow-up by using data from a large-scale clinical trial (N = 893). Results revealed that higher withdrawal intercepts, positive linear slopes, and greater volatility were all positively associated with relapse, and cigarette coefficients (indicating smoking-induced withdrawal reduction) were negatively related to relapse. In models keyed around the first lapse to smoking, those destined to lapse reported more severe withdrawal during abstinence, and withdrawal patterns discriminated groups defined according to lapse duration. The findings complement earlier heterogeneity studies in implicating the pattern of changing withdrawal symptoms over time as a factor strongly associated with smoking relapse.