INTRODUCTION: To examine the long-term impact of graphic health-warning labels (GHWL) on adolescents' cognitive processing of warning labels and cigarette pack perceptions. METHODS: Cross-sectional school-based surveys of students aged 13-17 years residing in urban centers, conducted prior to GHWL introduction (2005) and 6 months (2006), 2 years (2008), and 5 years (2011) post-GHWL introduction. Students who had seen a cigarette pack in the previous 6 months or in 2006, who had seen GHWL were included in analyses (2005 n = 2,560; 2006 n = 1,306; 2008 n = 2,303; 2011 n = 2,716). Smoking stage, reported exposure to cigarette packs, cognitive processing of GHWL, and positive and negative perceptions of pack image were assessed. RESULTS: While cognitive processing of GHWL in 2006 and 2008 was greater than 2005 (p <.01), by 2011 scores had returned to 2005 levels. This pattern of change was consistent across smoking status groups. Pack image perceptions became more negative over time among all students, irrespective of smoking experience. While positive pack image ratings were lower in all subsequent years than 2005, the 2008 rating was higher than 2006 (p <.01). A significant interaction between survey time and smoking status (p <.01) showed that significant increases in positive pack ratings after 2006 only occurred among current and experimental smokers. CONCLUSIONS: When novel, GHWL on cigarette packs increase cognitive processing among adolescents. However, this effect diminishes after 5 years, suggesting more regular message refreshment is needed. Australia's adoption of plain packaging is intended to undermine positive pack appeal and increase warning salience.