MA 18.04 Changes in Smoking Behavior in the Early Cancer Detection Test Lung Cancer Scotland (ECLS) Study

Background: Lung cancer screening might be a ‘teachable moment’ for smoking cessation or conversely could provide a ‘license to smoke’. Such effects should be considered in the overall benefits and harms of screening. Existing evidence of the impact of screening on smoking is mixed. Method: A randomized controlled trial of a blood autoantibody test (EarlyCDT-Lung) for the early detection of lung cancer was conducted in 12,210 smokers and ex-smokers in Scotland, UK. The test allowed risk stratification for targeting of a chest X-ray and repeat CT scans. Sub-samples of positive test (n = 321), negative test (n = 361) and control (n = 350) participants completed questionnaires before screening, after receipt of blood test results and at 3, 6 and 12 months post-screening. They self-reported smoking point prevalence, attempts to quit, number of cigarettes smoked per day and the Heaviness of Smoking Index. Multi-level regression analyses, adjusted for confounders, explored differences in smoking over time between screened and control arms and between positive test, negative test and control groups. Result: Preliminary results show no statistically significant differences in smoking prevalence between the screened and control arms over time. There was a reduction in smoking prevalence of borderline statistical significance in the positive test group versus controls across all time points (OR 0.46, 95% CI 0.21-1.03). This difference reduced when assuming non-responding smokers were still smoking (OR 0.73, 95% CI 0.38-1.42). Significantly more smokers in the positive test group had recently attempted to stop smoking at 3 months compared to controls (OR 2.29, 95% CI 1.04-5.04). Positive test group smokers were significantly less likely to report smoking 20 or more cigarettes a day than controls across all time points (OR 0.32, 95% CI 0.14-0.69). Negative test group smokers were more likely to score moderate/high/very high on the Heaviness of Smoking Index compared to controls at 6 months. This difference was statistically significant before adjusting for confounders but the adjusted model was no longer significant (OR 2.51, 95% CI 0.90-6.97). Conclusion: There was no effect of lung cancer screening on smoking prevalence. The findings indicate a positive test result can be a teachable moment for smoking cessation. They also highlight the short term risk of heavier smoking after a negative test result. This is an important area for further research to ensure negative lung cancer screening test results do not inadvertently promote continued and heavier smoking.