INTRODUCTION Smoking behavior can change with time and lead to different health outcomes. This study explored the trajectory of smoking and its relationship with cancer incidence and mortality among Korean male adults. METHODS We used 2002–2018 data from the National Health Insurance Service (NHIS). Smoking status was repeatedly measured in four waves of general health examinations provided by the NHIS between 2002 and 2009. Cancer incidence and mortality were tracked from 2010 to 2018. Trajectory analysis was used to identify the patterns of smoking. The hazard ratio was calculated using Cox proportional regression models. RESULTS For the 2448548 men (≥20 years), 137788 cases of cancers and 41146 cancer deaths were found. We identified six trajectory groups: never smokers, former smokers, new current smokers, decreasing light smokers, steady moderate smokers, and steady heavy smokers. All smoking groups had an increased risk of cancer. The steady heavy smokers showed higher cancer incidence and mortality rate than the steady non-smokers (hazard ratio, HR=1.53; 95% CI: 1.49–1.58 and HR=2.64; 95% CI: 2.50–2.79, respectively). The cancer-specific analysis showed that the larynx and lung cancer incidence and mortality rate of the smoking group were higher than in never smokers. CONCLUSIONS Smoking, even at low doses, increases the risk of most cancers in men. Quitting or reducing smoking, especially at a young age, can lower cancer incidence and mortality. This study may provide more objective results on the relationship between smoking and cancer, because smoking behavior was examined at multiple time points.

No abstract

BackgroundAlcohol drinking behaviors change temporally and can lead to changes in related cancer risks; previous studies have been unable to identify the association between the two using a single-measurement approach. Thus, this study aimed to investigate the association of drinking trajectories with the cancer risk in Korean men. MethodsA trajectory analysis using group-based trajectory modeling was performed on 2,839,332 men using data on alcohol drinking levels collected thrice during the Korean National Health Insurance Service's general health screening program conducted between 2002 and 2007. Cox proportional hazards regression was performed to evaluate the associations between drinking trajectories and cancer incidence, after adjustments for age, income, body mass index, smoking status, physical activity, family history of cancer, and comorbidities. ResultsDuring 10.5 years of follow-up, 189,617 cancer cases were recorded. Six trajectories were determined: non-drinking, light, moderate, decreasing-heavy, increasing-heavy, and steady-heavy. Light-to-heavy alcohol consumption increased the risk for all cancers combined in a dose-dependent manner (adjusted hazards ratio, 95% confidence interval for trajectories = 1.

INTRODUCTION Previous research on post-diagnosis smoking among cancer survivors mainly relied on smoking status, which may not fully reflect the impact of changes in smoking levels. This study aimed to evaluate mortality risk according to smoking trajectories among Korean male cancer survivors, using a trajectory approach to comprehensively capture smoking patterns. METHODS The study included 110555 men diagnosed with cancer between 2002 and 2018 from the Korean National Health Information Database. Group-based trajectory modelling was used to identify post-diagnosis smoking trajectories among pre-diagnosis current smokers (n=45331). Cox hazards models were fitted to evaluate mortality risk according to smoking trajectories for pooled cancers, pooled smoking-related cancers, smoking-unrelated cancers, and gastric, colorectal, liver, and lung cancers. RESULTS Smoking trajectories included light-smoking quitters, heavy-smoking quitters, consistent moderate smokers, and decreasing heavy smokers. Smoking significantly increased all-cause and cancer mortality risks in cancer patients for pooled cancers, pooled smoking-related cancers, and pooled smoking-unrelated cancers. Compared to non-smokers, all-cause mortality risk for pooled cancers significantly increased according to smoking trajectories: (AHR=1.33; 95% CI: 1.27–1.40), (AHR=1.39; 95% CI: 1.34–1.44), (AHR=1.44; 95% CI: 1.34–1.54), and (AHR=1.47; 95% CI: 1.36–1.60), respectively. Smoking increased all-cause and cancer mortality risks in gastric and colorectal cancer patients and cancer-specific mortality in lung cancer patients. The significant associations of smoking trajectories with all-cause and cancer mortality risks were primarily observed in 5-year survivors but not in short-term survivors. Among heavy smokers, smoking cessation significantly reduced all-cause mortality risk in the long-term. CONCLUSIONS The post-diagnosis smoking trajectory independently predicts cancer prognosis among male cancer patients. Proactive cessation support should be strengthened, particularly for those who smoke heavily.