GRAPHICAL ABSTRACT

INTRODUCTION

Globally, the number of smokers is declining. Nonetheless, mortality resulting from smoking continues to be a pressing concern worldwide. It is projected that by 2030, nearly 6 million deaths each year will be attributable to direct smoking, with an additional 1 million due to exposure to secondhand smoke [1]. In Korea, smoking significantly contributes to lung cancer, the leading cause of death. This preventable risk factor is the main contributor to premature mortality. Beyond mortality, the effects of smoking extend to healthcare costs and productivity losses. In 2013, the socioeconomic cost of smoking was substantial, estimated at 7.1 trillion Korean won (5.34 billion US dollars).

The Framework Convention on Tobacco Control of the World Health Organization mandates that signatory countries report smoking-induced deaths biennially through national reports. Collating data on these deaths is pivotal for understanding the long-term repercussions of historical smoking behaviors on today’s population, including the degree of harm, and has policy implications. This is also crucial for anticipating and curtailing potential risks associated with current smoking habits. Monitoring smoking-induced fatalities provides public health officials with vital information for mitigating tobacco-induced harm and protecting the well-being of the population.

The smoking prevalence in Korea was notably high in 1997, at 66.2% for men and 6.5% for women. However, these rates have steadily declined over the years. Given that chronic illnesses induced by smoking often cause mortality after 20-30 years [2], current deaths attributed to such ailments likely reflect the high smoking rates of the past. Korean studies assessing smoking-induced fatalities typically employ the population-attributable fraction (PAF) to calculate smoking-attributable mortality (SAM). In earlier research, due to the absence of Korean cohorts, mortality risk data from Japan was adopted [3]. Subsequent research was based on a Korean cohort encompassing civil servants, private school employees, and their health insurance beneficiaries and families [4]. Some cohort studies had relatively brief follow-up durations, often less than 2 decades [4,5], or were restricted to particular diseases [3,4]. This highlights the need for smoking-related death risk estimates and SAM grounded in a large-scale cohort that spans over 20 years and represents the Korean population.

This study aims to investigate SAM in Korea in 2020. We estimated the risk of disease-specific deaths, the PAF across all fatalities, and the number of smoking-related deaths in 2020 using data from the Korean population.

METHODS

This study employed a prevalence-based approach to determine SAM.

Data Source

This study used multiple data sources to estimate smokingrelated deaths in Korea. Initially, 4 databases were utilized to calculate hazard ratios (HRs) for SAM (Table 1). The Korea National Health Examination Baseline (KNHEB) cohort includes individuals who participated in the National General Health Screening Program (GHSP) during 2002-2003 [6]. The National Health Insurance Service-Korea Cancer Prevention Study (NHISKCPS) is a cohort comprising civil servants, employees of private schools, and their health insurance beneficiaries and families from 1992 to 1995 [4]. KCPS-II includes individuals who underwent general health examinations at one of 18 institutions in the country, with 15 in Seoul and Gyeonggi Province and 3 in other regions, during 1994-2013 [7]. The Korean Metabolic Syndrome Mortality Study (KMSMS) is a retrospective cohort study involving individuals who had health check-ups at 18 general examination centers nationwide from 1994 to 2004 [8]. The smoking-related survey questions used in each database are included in the Supplemental Material 1. All 4 cohorts have informed cause of death statistics issued by Statistics Korea up to 2020.

To calculate SAM, we utilized smoking prevalence data from 1995. Given the 20-year to 30-year time lag between smoking initiation and subsequent mortality, we employed data from the 1995 National Health and Health Behavior Survey conducted by the Korea Institute for Health and Social Affairs. These figures were used by the Korea Disease Control and Prevention Agency to assess smoking-related risks in 2019. We referred to smoking rates for individuals aged 30-69 years, as the survey did not include those aged 70 years or older. In 1995, the smoking rates for men were as follows: current smokers, 65.6%; former smokers, 16.5%; and non-smokers, 17.9%. For women, the respective rates were 6.4%, 1.6%, and 92.0% (Supplemental Material 2).

To calculate SAM in 2020, disease-specific death data were used from Statistics Korea’s cause of death statistics [9]. Statistics Korea compiles nationwide death reports, including the year of death, age, and cause of death based on International Classification of Diseases, 10th revision codes. These reports are made available to the Korean public.

Data Extraction

The analysis was restricted to men and women adults aged 30 years and older. Based on smoking data obtained at the start of the cohort study, participants were categorized as current smokers, former smokers, or non-smokers. In total, 41 diseases strongly correlated with smoking, as identified in the Surgeon General’s Report [10] and prior Korean studies, were included in the analysis. Since the KNHEB cohort is extremely large, to avoid duplication of data from the NHIS-KCPS—which includes civil servants, private school staff, and their health insurance beneficiaries and families registered under the National Health Insurance Service—individuals within the NHISKCPS cohort were excluded from the KNHEB dataset. This revised dataset was designated as KNHEB-A, representing the analytical subset of the KNHEB cohort.

Statistical Analysis

To determine SAM, we utilized a prevalence-based approach employing the PAF. Initially, a Cox proportional hazards model was used to derive HRs for mortality for 41 smoking-related diseases. The HRs were adjusted for age and alcohol consumption and were calculated separately for current and former smokers, with non-smokers as the baseline. HRs for each disease were derived from the 4 databases individually. A metaanalysis was performed using the HR values obtained from each database to calculate the integrated estimates. Because the study populations in the included cohorts were not identical, and given our intention to generalize findings to the wider population, a random-effects model was used. Heterogeneity between studies was evaluated using the Higgins I² test [11].

The pooled HRs and the 1995 smoking prevalence data were used to estimate the PAFs as follows [12]:

PAF=pcurrent smorker×HRcurrent smoker+pformer smoker×HRformer smoker+pnever smoker-1pcurrent smorker×HRcurrent smoker+pformer smoker×HRformer smoker+pnever smoker

where p represents the prevalence of current smokers, former smokers, and never smokers. To obtain the final SAM, the disease-specific PAFs were multiplied by the respective disease-specific mortality figures from 2020. For sensitivity analysis, SAMs were computed using gender-specific smoking rates from 2001, 2005, and 2010 [13], with time lags of 19 years, 15 years, and 10 years, respectively. A comparative examination was conducted to assess the impact of these different periods on the analysis. The prevalence rate for 2001 was employed because it was the closest available data point to the year 2000.

Ethics Statement

The Institutional Review Board of Seoul National University approved KNHEB (IRB No. E2405/002-011). The Institutional Review Board of Yonsei University granted approval for NHIS-KCPS, KCPS-II, KMSMS (NHIS-KCPS IRB No. 4-2001-0029, KCPS-II IRB No. 4-2011-0277, KMSMS IRB No. 4-2007-0065).

RESULTS

Cohort Characteristics

The smoking-related attributes of the participants included in the HR calculations for each cohort are presented in Table 2. In the KNHEB cohort, the current smoking rate was 30.7%, while the former smoker rate was 9.7%. Among current smokers, the majority consumed 0.5-1.0 pack per day, and the most prevalent smoking duration was 10-19 years. For former smokers, the most common smoking duration was also 10-19 years. Within the NHIS-KCPS, the current smoking rate was 38.9% among men, and the rate of former smokers was 14.0%. The majority of current smokers smoked 0.5-1.0 pack daily, and 10-19 years was the most frequent smoking duration. In the KCPS-II cohort, the current smoking rate was 28.7%, and the former smoking rate was 18.8%. Among current smokers, the largest group smoked 1-2 packs daily, and the most common smoking duration was 10-19 years. For former smokers, the most common smoking duration was also 10-19 years. In the KMSMS cohort, the current smoking rate was 31.6%, and the former smoking rate was 17.7%. Among current smokers, the largest group smoked 1-2 packs per day, and 20-29 years was the most common smoking duration. Among former smokers, the largest proportion had a smoking duration of 10-19 years. The 4 cohorts showed differences in general characteristics, as detailed in the Supplemental Material 3.

Results of Meta-analysis

The pooled HRs, estimated using each HRs from 4 cohorts by disease (Supplemental Materials 4 and 5), for all-cause mortality exceeded 1.5 for current smokers and 1.0 for former smokers. Among men, current smokers exhibited an elevated overall risk of cancer, as well as an increased risk of esophageal cancer, lung cancer, stroke, atherosclerosis, coronary obstructive pulmonary disease (COPD), and pneumonia. Former smokers had a higher risk of all-cause death, as well as a higher overall cancer risk and increased risk of laryngeal cancer, lung cancer, ischemic heart disease (IHD), and COPD. In women, current smokers had higher HRs for cancer overall, as well as for cancers of the esophagus, larynx, lung, and bladder, IHD, atherosclerosis, COPD, and pneumonia. Former smokers showed increased HRs for all-cancer, cancers of the esophagus and lung, IHD, atherosclerosis, COPD, and pneumonia. Heterogeneity in pooled HR was observed in some diseases in both genders (Tables 3 and 4).

Population-attributable Fraction

The PAF, calculated using pooled HRs and smoking prevalence data from 1995, was 33.2% for men and 4.5% for women for all-cause death. For men, the PAF exceeded 50% for esophageal cancer, laryngeal cancer, lung cancer, and COPD, and surpassed 30% for all-cause death, all cancers, oropharyngeal cancer, stomach cancer, bladder cancer, IHD, and atherosclerosis. For women, the PAFs for laryngeal cancer, esophageal cancer, lung cancer, atherosclerosis, COPD, and ulcers were greater than 10% (Table 5).

Smoking-attributable Mortality

In 2020, smoking was responsible for 60 213 deaths (53 930 in men and 6283 in women). For men, lung cancer was the predominant cause of SAM, accounting for approximately 10 000 deaths, followed by pneumonia, IHD, stroke, and liver cancer (each with more than 2000 deaths). Among women, lung cancer was the leading cause of SAM (over 500 deaths), with pneumonia, stroke, and IHD each resulting in more than 400 deaths (Table 5).

Applying smoking rates for 2001, 2005, and 2010 resulted in SAM values of 56 839 (51 461 men, 5378 women), 54 684 (47 488 men, 7196 women), and 54 081 (45 902 men, 8179 women), respectively (Supplemental Materials 6 and 7).

DISCUSSION

This is the first study to integrate risk estimates and calculate SAM for smoking-associated diseases using 4 large Korean cohorts. There were 60 213 smoking-induced deaths in 2020 (53 930 men and 6283 women). The disease with the highest mortality rate was lung cancer, with 10 499 deaths (9921 men and 578 women), followed by pneumonia, IHD, and stroke.

To calculate HRs, this study utilized 4 databases, as previously noted. The NHIS-KCPS, established in 1992, has certain limitations as it predominantly consists of civil service personnel, educators, and their beneficiaries. To address these limitations, data from 3 additional databases were used: the KNHEB cohort, which includes more than 10.4 million individuals who underwent the National Health Insurance Service GHSP in 2002-2003 [14], and the KCPS-II and KMSMS databases. Although the latter 2 databases are more limited in scope, they provide greater temporal coverage, thereby helping to mitigate the limitations of the NHIS-KCPS. The representativeness of the HRs derived from these extensive cohorts was confirmed through a thorough meta-analysis.

In the study of Jung et al. [15] that was published in 2013, which used PAFs to determine SAM for Koreans, a meta-analysis of the HRs for 41 diseases was conducted based on both the NHIS-KCPS and the KMSMS. For current men smokers, the HRs for death were as follows: all-cause death, 1.75 (95% confidence interval [CI], 1.43 to 2.12); all cancers, 1.98 (95% CI, 1.60 to 2.44); esophageal cancer, 2.59 (95% CI, 2.19 to 3.01); lung cancer, 4.80 (95% CI, 4.02 to 5.73); IHD, 2.04 (95% CI, 1.55 to 2.68); stroke, 1.57 (95% CI, 1.22 to 2.03); and pneumonia, 1.51 (95% CI, 1.34 to 1.71). These figures align closely with this study’s findings. However, the HR for atherosclerosis in this study (2.38; 95% CI, 1.85 to 3.07) was lower than in Jung et al. [15] (3.55; 95% CI, 2.66 to 4.73), while that for COPD (3.34; 95% CI, 1.88 to 5.93) was higher than in Jung et al. (2.46; 95% CI, 2.16 to 2.22). For current women smokers, the HRs were mostly consistent for all-cause mortality, lung cancer, and IHD, but differences were noted in esophageal cancer, COPD, and pneumonia HRs. Although there were discrepancies in HRs across diseases, in most cases the difference was <1.0. The consistent outcomes might be due to the similar follow-up durations, of roughly 20 years, between the cohort in the research of Jung et al. [15] and the KNHEB cohort analyzed in this study.

In Katanoda et al. [16]’s 2008 study, which determined the HRs for smoking-associated disease mortality among the Japanese population, a meta-analysis of 3 large-scale Japanese cohort studies was performed. The age-adjusted HRs for current men smokers were as follows: all-cause mortality, 1.63 (95% CI, 1.56 to 1.70); all cancer, 1.97 (95% CI, 1.83 to 2.13); and lung cancer, 4.79 (95% CI, 3.88 to 5.92). These HRs closely align with those in our study. However, for conditions such as laryngeal cancer (HR, 5.47; 95% CI, 1.29 to 23.11), bladder cancer (HR, 5.35; 95% CI, 2.47 to 11.57), and IHD (HR, 2.18; 95% CI, 1.79 to 2.66), the HRs were slightly lower in this study. For women, excluding COPD (HR, 3.55; 95% CI, 1.53 to 8.21), the HRs were similar or slightly higher in Katanoda et al. [16]’s study. Their cohorts, predominantly from 1983 to 1993, had follow-up durations of 8.5-10.4 years, which is shorter than the follow-up in our study. Despite this, the HRs reported by Katanoda et al. [16] were higher. This could be due to a change in smoking intensity. Between 1950 and 1980, while Japan experienced an increase in smoking intensity, there was a decline in smoking prevalence, leading to an increase in per capita tobacco use [17]. This might explain the persistently high risk of smoking-related mortality even within shorter follow-up intervals.

In the present study, a high level of heterogeneity was found in some pooled HRs. This heterogeneity could be attributed to several factors.

First, differences in the baseline survey years and follow-up periods across cohorts might have contributed to the observed heterogeneity. Notably, although the KNHEB cohort was the largest in scale, it used smoking data from a survey conducted in 2002-2003 with a maximum observation period of 18 years. In contrast, the NHIS-KCPS, KCPS-II, and KMSMS cohorts had baseline survey years spanning 1992-1999, 1994-2013, and 1996-2004, respectively, with maximum observation periods of 28 years, 26 years, and 24 years, respectively. Each of these periods exceeded 20 years. In the KNHEB cohort, the HRs showed an overall tendency to be smaller than those observed for the other 3 cohorts.

Second, smoking quantity might have affected the HRs. As the smoking quantity increases among current smokers, the risk of death from cancer, cardiovascular diseases, and other smoking-related diseases also rises [18]. The proportion of smokers consuming more than 1 pack a day was higher in NHIS-KCPS, KCPS-II, and KMSMS at 29.3%, 43.9%, and 51.9%, respectively, compared to 21.3% in KNHEB. This suggests that differences in HR values could be due to this disparity in smoking amounts.

Third, considering the significant reduction in the risk of lung cancer incidence in response to decreased smoking and increased smoking cessation [19], changes in smoking habits after the baseline survey in each cohort may have contributed to differences in HRs due to the cumulative risk of smoking-related mortality.

Fourth, advancements in medical technology in Korea have significantly improved survival rates, which increased from 42.9% in 1993-1995 to 70.3% in 2014-2018 [20]. The more favorable prognosis in recent years may have contributed to the decreased HRs in the latest cohort.

However, compared with other studies with shorter observation periods, smaller sample sizes or specific population groups [5,15,21], this study derived more representative HRs from large Korean cohorts observed over more than a decade. Additionally, the random-effect model assumes variability in exposure estimates across studies. Therefore, the characteristics identified as sources of heterogeneity were appropriately addressed.

In the 2013 study of Jung et al. [15], which used pooled HRs similar to our approach and analyzed 1985 smoking rates, the PAFs were higher than in our study. However, the changes in smoking rates in Korea—with decreases from 71.1% in 1985 to 65.6% in 1995 in men and from 10.7% to 6.4% in women—had a greater impact than the differences in the HRs.

In Katanoda et al. [16]’s 2008 study, the PAF for all-cause deaths was 27.8% for men and 6.7% for women. For all cancer deaths, the PAF was 38.6% for men and 5.2% for women. For lung cancer, the PAF was 69.2% for men and 19.8% for women. For IHD, the PAF was 44.1% for men and 15.2% for women. These figures were slightly lower than those for men in our study but slightly higher than those for women. The differences in PAFs might be attributed to the differing smoking rates in Japan, rather than the higher HRs in Katanoda et al. [16]’s study. Specifically, the current smoking rate in Japan was reported as 54.4% for men and 8.1% for women, while the past smoking rates were 25.1% and 2.4%, respectively. The rates for Japanese men smokers are lower than those for their Korean counterparts, but the reverse is true for Japanese women smokers.

Ma et al. [22] calculated PAFs and SAM for individual United States states and reported PAFs of 20.6% and 15.1% for adult (aged >35 years) men and women, respectively. The rate for men was lower than that in our study; however, the reverse was true for women. They analyzed relatively recent (2014) data from the Behavior Risk Factor Surveillance System, which could explain the differences.

When calculating the PAF, it is crucial to consider the proportions of current, former, and non-smokers, as the chosen smoking prevalence at a specific time point significantly influences the SAM. Some studies have applied smoking prevalence from the year of interest for mortality [23] or have used smoking rates at the time of enrollment from a single cohort. However, some Korean studies typically do not derive from cohorts and use smoking rates not from the time of death but from more than 20 years in the past [4,15,24]. In Korea, smoking rates have sharply decreased after 2000 compared to the 1980s and 1990s, suggesting that using smoking rates from recently established cohorts might underestimate the impact of smoking. This study employed 4 cohorts with different start years, highlighting the need to select a consistent point in time for smoking rates. Given the latency period between smoking and the onset of disease, some argue that a 10-year lag time is appropriate [25]. Furthermore, it has been reported that the contribution of smoking to the incidence of lung cancer and overall mortality is most fully seen after an average of 30 years [2,26]. Therefore, this study utilized the smoking prevalence from 1995, which is the closest year to 30 years before 2020, to ensure a representative measure of smoking rates.

According to the most recent findings reported by Jung et al. [15], the SAM in 2012 was 58 155 for all-cause deaths (49 704 in men and 8451 in women) and 20 581 for all cancers (19 184 in men and 1397 in women). These numbers were lower than the estimates in our study. Regarding specific diseases, in men, the SAM was notably higher for lung cancer (8881 deaths), stroke (3563 deaths), and IHD (3256 deaths). In women, the respective numbers were 887, 585, and 602. Given that our study shares similarities with that of Jung et al. [15], including the disease categories and methods, it can be inferred that SAM increased after 2013. This increase is likely attributable to a growing annual number of deaths, despite the lower PAF estimates in our study. Thus, a larger population exposed to the effects of smoking likely contributed to an increase in smoking-related deaths.

This study is not without limitations. The primary limitation is the reliance on self-reported smoking status, which might have resulted in underreporting of current smoking among women due to socio-cultural factors [27] or the misclassification of smoking status. Second, we did not account for smoking history at the time of enrollment, including smoking frequency, quantity, and duration. Third, we did not consider cessation among current smokers or resumption among former smokers, which could potentially affect HRs. Smoking cessation reduces the risks of cancer and cardiovascular diseases [28,29], while resuming smoking increases the risk of cancer [30]. Given the diversity of individuals’ smoking behaviors, future studies should consider these characteristics. Fourth, additional confounding factors may influence the mortality risk associated with smoking-related diseases. However, adjustments for health behaviors or demographic characteristics do not significantly impact the estimates of mortality risk or SAM associated with these diseases [31,32].

In conclusion, this study conducted a meta-analysis of the risk of death from smoking-related diseases using data from 4 large Korean cohorts, with PAF and SAM values calculated for 2020. The PAF data indicated that smoking contributed to 33.2% of all-cause men deaths and 4.6% of all-cause women deaths. Regarding SAM, an estimated total of 60 213 deaths (53 930 in men and 6283 in women) due to smoking was projected for 2020, highlighting smoking as a significant public health issue. Given the historically high smoking prevalence in Korea, despite the current gradual decline, the harmful effects of smoking are likely to persist, with SAM expected to rise. To address these concerns, ongoing monitoring and prioritization of tobacco control and harm prevention policies are essential.

Supplemental Materials

Notes

Conflict of Interest

The authors have no conflicts of interest associated with the material presented in this paper.

Funding

This work was supported by the Korea Disease Control and Prevention Agency (2022-12-104).

Author Contributions

Conceptualization: Cho SI, Jee SH, Hwang J, Jo S, Park SY. Data curation: Hwang J, Jo S, Cheon E, Yang YS, Jung KJ, Jee SH, Lee S, Park SY. Formal analysis: Jo S, Cheon E, Yang YS. Funding acquisition: Na K, Cho SI, Jee SH. Methodology: Cho SI, Hwang J, Cheon E, Jung KJ, Jee SH. Project administration: Hwang J, Yang YS. Visualization: Cheon E. Writing – original draft: Cheon E, Cho SI. Writing – review & editing: Cheon E, Cho SI, Hwang J, Jo S, Jee SH, Jung KJ, Yang YS, Lee S, Park SY, Na K, Kim S.

Acknowledgements

REFERENCES

• 1. World Health Organization. World No Tobacco Day 2017 - tobacco: a threat to development; 2017 [cited 2023 Aug 27]. Available from: https://www.who.int/news-room/events/detail/2017/05/31/default-calendar/world-no-tobacco-day2017
• 2. Jee SH, Samet JM, Ohrr H, Kim JH, Kim IS. Smoking and cancer risk in Korean men and women. Cancer Causes Control 2004;15(4):341-348. https://doi.org/10.1023/B:CACO.0000027481.48153.97ArticlePubMed
• 3. Meng KH. Smoking-attributable mortality among Korean adults. Korean J Epidemiol 1988;10(2):138-145. (Korean)
• 4. Jee SH, Lee JK, Kim IS. Smoking-attributable mortality among Korean adults: 1981-2003. Korean J Epidemiol 2006;28(1):92-99. (Korean)
• 5. Jee SH, Yun JE, Park JY, Sull JW, Kim IS. Smoking and cause of death in Korea: 11 years follow-up prospective study. Korean J Epidemiol 2005;27(1):182-190. (Korean)
• 6. Cho SI, Chung W, Hwang J, Jee SH, Jung KJ, Lim MK, et al. Impact of smoking on morbidity, mortality, and economic cost. 2022. Cheongju: Korea Disease Control and Prevention Agency (Korean)
• 7. Jee YH, Emberson J, Jung KJ, Lee SJ, Lee S, Back JH, et al. Cohort profile: the Korean Cancer Prevention Study-II (KCPS-II) Biobank. Int J Epidemiol 2018;47(2):385-386f. https://doi.org/10.1093/ije/dyx226ArticlePubMed
• 8. Mok Y, Jeon C, Lee GJ, Jee SH. Physical activity level and colorectal cancer mortality. Asia Pac J Public Health 2016;28(7):638-647. https://doi.org/10.1177/1010539516661761ArticlePubMed
• 9. Statistics Korea. Microdata integrated service. 2023 [cited 2023 Aug 27]. Available from: https://mdis.kostat.go.kr/index.do (Korean)
• 10. U.S. Department of Health and Human Services. The health consequences of smoking—50 years of progress: a report of the Surgeon General; 2014 [cited 2023 Aug 27]. Available from: https://www.hhs.gov/sites/default/files/consequences-smoking-exec-summary.pdf
• 11. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327(7414):557-560. https://doi.org/10.1136/bmj.327.7414.557ArticlePubMedPMC
• 12. Levin ML. The occurrence of lung cancer in man. Acta Unio Int Contra Cancrum 1953;9(3):531-541PubMed
• 13. Korea Disease Control and Prevention Agency. Korea National Health and Nutrition Examination Survey, 2001, 2005, 2010 [cited 2023 Aug 27]. Available from: https://knhanes.kdca.go.kr/knhanes/main.do (Korean)
• 14. Choi E, Kim D, Lee W, Hwang I. Establishment of an efficient management system for advanced outcome of the National Health Screening Policy. Seoul: Korea Institute for Health and Social Affairs; 2009. p. 53-90 (Korean)
• 15. Jung KJ, Yun YD, Baek SJ, Jee SH, Kim IS. Smoking-attributable mortality among Korean adults, 2012. J Korea Soc Health Inform Stat 2013;38(2):36-48. (Korean)
• 16. Katanoda K, Marugame T, Saika K, Satoh H, Tajima K, Suzuki T, et al. Population attributable fraction of mortality associated with tobacco smoking in Japan: a pooled analysis of three large-scale cohort studies. J Epidemiol 2008;18(6):251-264. https://doi.org/10.2188/jea.JE2007429ArticlePubMedPMC
• 17. Tominaga S. Smoking and cancer patterns and trends in Japan. IARC Sci Publ 1986;(74):103-113
• 18. Jacobs DR Jr, Adachi H, Mulder I, Kromhout D, Menotti A, Nissinen A, et al. Cigarette smoking and mortality risk: twenty-five-year follow-up of the Seven Countries Study. Arch Intern Med 1999;159(7):733-740. https://doi.org/10.1001/archinte.159.7.733ArticlePubMed
• 19. Godtfredsen NS, Prescott E, Osler M. Effect of smoking reduction on lung cancer risk. JAMA 2005;294(12):1505-1510. https://doi.org/10.1001/jama.294.12.1505ArticlePubMed
• 20. Hong S, Won YJ, Lee JJ, Jung KW, Kong HJ, Im JS, et al. Cancer statistics in Korea: incidence, mortality, survival, and prevalence in 2018. Cancer Res Treat 2021;53(2):301-315. https://doi.org/10.4143/crt.2021.291ArticlePubMedPMC
• 21. Lee EH, Park SK, Ko KP, Cho IS, Chang SH, Shin HR, et al. Cigarette smoking and mortality in the Korean Multi-center Cancer Cohort (KMCC) study. J Prev Med Public Health 2010;43(2):151-158. (Korean). https://doi.org/10.3961/jpmph.2010.43.2.151ArticlePubMed
• 22. Ma J, Siegel RL, Jacobs EJ, Jemal A. Smoking-attributable mortality by state in 2014, U.S. Am J Prev Med 2018;54(5):661-670. https://doi.org/10.1016/j.amepre.2018.01.038ArticlePubMed
• 23. Pérez-Ríos M, Schiaffino A, Montes A, Fernández E, López MJ, Martínez-Sánchez JM, et al. Smoking-attributable mortality in Spain in 2016. Arch Bronconeumol 2020;56(9):559-563. https://doi.org/10.1016/j.arbr.2020.07.005ArticlePubMed
• 24. Park S, Jee SH, Shin HR, Park EH, Shin A, Jung KW, et al. Attributable fraction of tobacco smoking on cancer using population-based nationwide cancer incidence and mortality data in Korea. BMC Cancer 2014;14: 406. https://doi.org/10.1186/1471-2407-14-406ArticlePubMedPMC
• 25. Kong KA, Jung-Choi KH, Lim D, Lee HA, Lee WK, Baik SJ, et al. Comparison of prevalence- and smoking impact ratio-based methods of estimating smoking-attributable fractions of deaths. J Epidemiol 2016;26(3):145-154. https://doi.org/10.2188/jea.JE20150058ArticlePubMedPMC
• 26. Lopez AD, Collishaw NE, Piha T. A descriptive model of the cigarette epidemic in developed countries. Tob Control 1994;3(3):242-247. https://doi.org/10.1136/tc.3.3.242ArticlePMC
• 27. Park MB, Kim CB, Nam EW, Hong KS. Does South Korea have hidden female smokers: discrepancies in smoking rates between self-reports and urinary cotinine level. BMC Womens Health 2014;14: 156. https://doi.org/10.1186/s12905-014-0156-zArticlePubMedPMC
• 28. Duncan MS, Freiberg MS, Greevy RA Jr, Kundu S, Vasan RS, Tindle HA. Association of smoking cessation with subsequent risk of cardiovascular disease. JAMA 2019;322(7):642-650. https://doi.org/10.1001/jama.2019.10298ArticlePubMedPMC
• 29. Choi S, Chang J, Kim K, Park SM, Lee K. Effect of smoking cessation and reduction on the risk of cancer in Korean men: a population based study. Cancer Res Treat 2018;50(4):1114-1120. https://doi.org/10.4143/crt.2017.326ArticlePubMedPMC
• 30. Yoo JE, Han K, Shin DW, Jung W, Kim D, Lee CM, et al. Effect of smoking reduction, cessation, and resumption on cancer risk: a nationwide cohort study. Cancer 2022;128(11):2126-2137. https://doi.org/10.1002/cncr.34172ArticlePubMed
• 31. Malarcher AM, Schulman J, Epstein LA, Thun MJ, Mowery P, Pierce B, et al. Methodological issues in estimating smoking-attributable mortality in the United States. Am J Epidemiol 2000;152(6):573-584. https://doi.org/10.1093/aje/152.6.573ArticlePubMed
• 32. Thun MJ, Apicella LF, Henley SJ. Smoking vs other risk factors as the cause of smoking-attributable deaths: confounding in the courtroom. JAMA 2000;284(6):706-712. https://doi.org/10.1001/jama.284.6.706ArticlePubMed

Figure & Data

References

Citations

Citations to this article as recorded by  

• The economic cost of direct smoking in South Korea
  Kristine Namhee Kwon, Kangyeon Lee, Wankyo Chung
  Preventive Medicine Reports.2024; 46: 102865.     CrossRef