LAUSR

BACKGROUND AND AIMS Graphic warning labels (GWLs) on cigarette packs have been adopted by many jurisdictions worldwide. In the US, the introduction of GWLs has been delayed by claims that their high level of negative emotional arousal unnecessarily infringed on the tobacco manufacturers' free speech. This study aimed to provide experimental data on the contribution of emotional arousal to GWL efficacy. DESIGN Observational study using long-term naturalistic exposure and functional magnetic resonance imaging. SETTING Research university in Philadelphia, PA, USA. PARTICIPANTS 168 adult smokers. MEASUREMENTS For 4 weeks, participants received cigarettes in packs that carried either high-arousal or low-arousal GWLs (N=84 vs. 84). Smoking behavior, quitting-related cognitions and GWL-induced brain response were measured before and after the 4-week exposure. The amygdala and medial prefrontal cortex served as regions of interest. FINDINGS Compared with the high-arousal group, the low-arousal group smoked fewer cigarettes (log10-transformed, 1.076 vs. 1.019; difference=0.056, 95% confidence interval=[0.027,0.085], χ2 (1)=14.21, p<0.001) and showed stronger intention to quit (2.527 vs. 2.810; difference=-0.283 [-0.468,-0.098], χ2 (1)=8.921, p=0.007) and endorsement of the GWLs' textual component (4.805 vs. 5.503; difference=-0.698 [-1.016,-0.380], χ2 (1)=18.47, p<0.001). High-arousal GWLs induced greater amygdala response than low-arousal GWLs (0.157 vs. 0.052; difference=0.105 [0.049,0.161], χ2 (1)=23.52, p<0.001), although the response to high-arousal GWLs declined over time (slope=-0.087 vs. 0.016; difference=-0.103 [-0.198,-0.009], χ2 (1)=6.370, p=0.046). Greater baseline amygdala response was associated with more smoking at 4 weeks in the high-arousal group, but less smoking in the low-arousal group (slope=0.179 vs. -0.122; difference=0.287 [0.076,0.498], χ2 (1)=7.086, p=0.008). Medial prefrontal response did not significantly differ between groups. CONCLUSIONS High-arousal cigarette graphic warning labels (GWLs) appear to be less efficacious than low-arousal GWLs. The high emotional reaction that high-arousal GWLs elicit wanes over time. Baseline amygdala response negatively predicts efficacy of high-arousal GWLs and positively predicts efficacy of low-arousal GWLs. High emotional arousal may not be required for GWLs' sustained efficacy.