Trends in smoking prevalence and implication for chronic diseases in China: serial national cross-sectional surveys from 2003 to 2013

Articles

Trends in smoking prevalence and implication for chronic diseases in China: serial national cross-sectional surveys from 2003 to 2013 Minghuan Wang*, Xiang Luo*, Shabei Xu*, Wenhua Liu, Fengfei Ding, Xiaoxiang Zhang, Liang Wang, Jian Liu, Jianping Hu, Wei Wang*

Summary

Background China is the world’s largest consumer of tobacco and has a large smoking-related chronic disease burden. In this nationwide study, we aimed to evaluate smoking prevalence and its implication on chronic diseases in the Chinese population.

Lancet Respir Med 2018 Published Online October 25, 2018 http://dx.doi.org/10.1016/ S2213-2600(18)30432-6

Methods We collected data from serial cross-sectional National Health Service Surveys done in China in 2003, 2008, *Contributed equally and 2013. These surveys cover all 31 provinces, autonomous regions, and municipalities in mainland China, and use Department of Neurology multistage stratified cluster sampling. We divided mainland China into east, central, and west regions and then (M Wang MD, Prof X Luo MD, sampled counties from each region stratified by urban and rural areas. All respondents aged 15 years or older in the S Xu MD, F Ding MD, selected households were eligible. We analysed the variation in smoking prevalence from 2003 to 2013, further Prof W Wang MD), Department identified risk factors for smoking, and assessed the association between smoking and chronic diseases by using of Scientific Research Management (W Liu MPH), multiple logistic regression. Computer Center Findings The number of individuals interviewed and involved in the study was 153 450 in 2003, 145 223 in 2008, and 229 676 in 2013. The standardised smoking prevalence in China was consistently high, with a proportion of current smokers of 26·0% (95% CI 25·8–26·2) in 2003, 24·9% (24·8–25·1) in 2008, and 25·2% (25·1–25·4) in 2013 (p value for trend 0·5062). For men, prevalence was 48·4% (48·1–48·7) in 2003, 47·0% (46·6–47·4) in 2008, and 47·2% (46·9–47·5) in 2013. For women, prevalence was 3·1% (3·0–3·2) in 2003, 2·3% (2·2–2·5) in 2008, and 2·7% (2·6–2·8) in 2013. Smoking prevalence varied in different regions, and we identified four major patterns. While a consistently high proportion of Chinese men smoked, the standardised smoking prevalence in women younger than 40 years increased from 1·0% in 2003 to 1·6% in 2013. Moreover, the smoking prevalence among adolescent smokers aged 15–24 years increased from 8·3% in 2003 to 12·5% in 2013. Alcohol consumption was closely linked to smoking in adolescents (odds ratio 7·5, 95% CI 6·9–8·1). Risk factors for adolescent smoking were having older family members who smoke (1·9, 1·8–1·9) and low level education (1·3, 1·2–1·4). Increased risks of chronic diseases were related to smoking (1·1, 1·0–1·1), with higher risks related to early smoking initiation (1·1, 1·0–1·1) and long-term smoking (1·2, 1·2–1·3).

(X Zhang MPH, L Wang PhD), and Translational Medicine Center (J Liu PhD), Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China; and Centre for Health Statistics Information, National Health Commission of the People’s Republic of China, Beijing, China (J Hu MPH) Correspondence to: Prof Wei Wang, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology. Wuhan 430030, China [email protected]

Interpretation The implementation of tobacco control policies in China since the signing of the WHO Framework Convention on Tobacco Control in 2003 has not been effective in reducing smoking prevalence. Smoking prevalence among adolescents of both genders has increased substantially and there has been a steady increase among young women. More practical and effective policies targeting adolescents and women are urgently needed. Action is needed to prevent the large and growing smoking-related chronic disease burden further increasing as China’s population ages. Funding National Health Commission of the People’s Republic of China. Copyright © 2018 Elsevier Ltd. All rights reserved.

Introduction In 2003, WHO launched a worldwide tobacco control campaign—the Framework Convention on Tobacco Control (FCTC)—with the aim of creating a tobacco-free world.1 Some high-income countries such as the USA and the UK have shown a remarkable decline in smoking prevalence in the past decade, whereas recent estimates suggest that many low-income and middle-income countries (LMICs) are not on track to achieve tobacco control targets. 2,3 The Global Youth Tobacco Survey (GYTS)—a school-based survey of smoking prevalence in teenagers aged 12–15 years—has shown a prevalence of

tobacco use of 13·6% and a prevalence of second-hand smoke exposure of 55·9% for adolescents in LMICs.4 China is the largest consumer of tobacco as well as the most populous LMIC in the world.5 Since the Chinese Government signed the WHO FCTC in 2003, China has issued a series of national and local rules and regulations for tobacco control such as banning smoking in some public places, health warning labelling, and raising tobacco taxes.6 The efficacy of these policies in China remains to be elucidated. Several reports based on a single cross-sectional national survey have shown high smoking prevalence in China between 2004 and 2015.5,7–11

www.thelancet.com/respiratory Published online October 25, 2018 http://dx.doi.org/10.1016/S2213-2600(18)30432-6

1

Articles

Research in context Evidence before this study We searched PubMed for articles published up to May, 2018, using the terms “smoking” OR “tobacco use” AND “chronic diseases” AND “trend” OR “prevalence” and retrieved a total of 14 324 papers. Among them, 866 papers included data from China. None of these papers reported nationwide trends of tobacco use and its association with chronic diseases in China after 2003. We also searched PubMed using (“smoking”[MeSH Major Topic] AND prevalence [Title/Abstract]) AND China [Title/ Abstract]) and retrieved 242 articles. We identified six national surveys for smoking prevalence after 2003. Four surveys of Chronic Disease Risk Factor Surveillance in China were carried out in 2004, 2007, 2010, and 2013. Three of them surveyed individuals aged 18 years or older whereas only one surveyed individuals aged 15–69 years, meaning the trends of smoking prevalence in adolescents aged 15–18 years could not be well studied in these surveys. Two national surveys were carried out in 2010 and 2015 with a relatively smaller sample size of around 15 000 people, using the same survey method of the Global Adult Tobacco Survey (GATS). All surveys we found used an individual-based approach whereas our study used a household-based approach. We estimated familial aggregation by using multistage household cluster sampling. Only the 2010 GATS was reported on a global scale, whereas the other five surveys have not been reported globally. Researchers in the Global Burden of Diseases, Injuries, and Risk Factors Study 2015 estimated daily smoking prevalence by sex and age group for 195 countries from 1990 to 2015. Additionally, Bilano and colleagues investigated smoking prevalence and cigarette consumption in 173 countries for men and 178 countries for women from 1980 to 2012. However, none of these studies

See Online for appendix

2

However, no study based on serial national surveys has conducted a comprehensive evaluation of the efficacy of tobacco control policies during this period and provided solid data. Tobacco use is one of the most important preventable causes of chronic non-communicable diseases (NCDs) and premature deaths worldwide.12 The United Nation’s Sustainable Development Goals aim to reduce mortality from NCDs-attributed premature death by a third by 2030.13 NCDs accounted for more than 80% of total deaths in 2013 in China.14 A nationwide study15 has shown that about 99·9 million Chinese adults aged 20 years or older have chronic obstructive pulmonary disease (COPD).15 Controlling smoking prevalence is considered to be one of the most effective ways for the reduction of four major NCDs, including ischaemic heart disease, cancer, chronic respiratory diseases, and diabetes.16 However, the association between smoking and chronic diseases prevalent in China has yet to be reported. The Chinese Government aims to reduce smoking prevalence to 20% in the Chinese population aged 15 years or older by 2030.17 To achieve this goal, it is essential to evaluate

explored the most recent trends of smoking prevalence and chronic disease burden in China. Moreover, all of them were either a single cross-sectional study or on the global scale instead of focusing on China. Added value of this study Our national study, covering all the provinces in mainland China and spanning from 2003 to 2013, presents the most recent patterns of smoking prevalence and its relationship with NCDs in China since the WHO FCTC launched globally. Our results show that smoking prevalence in China remained at a high plateau during the studied decade. The smoking prevalence in some regions was significantly higher than the national average. In particular, we show a rapid increase in smoking prevalence in adolescents in China. Moreover, our study indicated an increase in smoking prevalence in women younger than 40 years. We identified the number of older family members who smoke— especially when the family members are female—as a risk factor for smoking in adolescents. We show a non-linear, negative correlation between chronic diseases and smoking initiation age, suggesting that individuals who started smoking earlier had a higher risk of chronic diseases. Implications of all the available evidence China has failed to fulfil the WHO FCTC fully in the studied decade. The increase in smoking prevalence among adolescents and women, combined with the association of female family members who smoke with increased adolescent smoking prevalence, implies that the Chinese Government should establish more stringent policies and enforce the law more strictly, with a strong focus on stopping and reversing the increased prevalence in adolescents and women.

the efficacy and implementation of tobacco control strategies periodically. Here, we present the trends in smoking prevalence from 2003 to 2013 and an assessment of the association between smoking and chronic NCDs in China.

Methods

Data sources and sampling We based our study on the latest three rounds (done in 2003, 2008, and 2013) of the National Health Services Survey (NHSS), which have been organised by the National Health Commission (NHC) of the People’s Republic of China at 5-year intervals since 1993. The NHSS is a serial cross-sectional survey covering all 31 provinces, autonomous regions, and municipalities in the mainland of China that aimed to provide information about Chinese individuals’ health to aid policy makers (see appendix p 3 for details). The NHSS uses multistage stratified cluster sampling; the methods have been described in detail elsewhere.18 We divided the mainland of China into east, central, and west regions and then sampled counties from each region stratified by urban

www.thelancet.com/respiratory Published online October 25, 2018 http://dx.doi.org/10.1016/S2213-2600(18)30432-6

Articles

and rural areas. To cover 0·02% of the population of the whole country and taking into account a 2% nonrespondent rate, we needed to sample at least 90 counties and 600 households for each county. 95 of a possible 2861 counties were randomly selected in 2003, 94 of which were chosen again in the 2008 survey (one county was excluded because its administrative division had changed). In 2013, we randomly selected another 52 counties from urban areas and ten counties from rural areas in addition to those counties involved in the 2008 survey (a total of 156 counties sampled), to better represent the increased urbanisation of China. Five streets (in urban areas) or townships (in rural areas) were sampled in each selected county, and within these two administrative communities (in urban areas) or villages (in rural areas) were sampled. Finally, we randomly sampled 60 households in each administrative community or village (each of which would typically contain 500–3000 households). Additionally, we randomly selected ten standby households in each administrative community or village; if we failed to interview one of the original 60 households after three visits in three different days, we would move onto one of ten standby households. All respondents aged 15 years or older in the selected household were eligible in this investigation. The NHSS was approved by the institutional review board of the Chinese National Bureau of Statistics. All respondents provided their oral consent to participate in the survey before the interview.

Procedures Details of the interview procedures have been reported previously.18 Briefly, local health-care workers were trained to carry out face-to-face interviews based on a structured questionnaire. The 26 questions related to our study were extracted from the results, including information about demographics, socioeconomic status, and health-related behaviour of participants in each round of the survey (appendix pp 4–6). We first asked the respondents whether they had smoked at least 100 cigarettes or not in their lifetime. If they had, we asked whether they were current or former smokers. We defined former smokers as those who had previously smoked at least 100 cigarettes but no longer smoked at the time of the survey and current smokers as those who had smoked at least 100 cigarettes and continued to smoke at the time of the survey. Both former and current smokers were counted as eversmokers in the analyses. Current smokers were asked the age they started smoking, which was used to calculate the duration of smoking (current age minus the age of smoking initiation to the nearest year). Alcohol consumption was defined as having had an alcoholic drink in the 12 months prior to the survey. Physical activity was defined as having done physical activity (including jogging, tai chi, aerobics, dancing, apparatus exercise, ball games, and swimming) at least once in the 6 months prior to the survey (appendix p 6).

Smoking-related NCDs were identified according to previously reported12,19 risk–outcome pairs (as classified by the International Classification of Diseases 10) including ischaemic heart disease (I20–25), digestive system cancer (C15–16), trachea or lung cancer (C33–34), COPD (J44), peptic ulcer (K25–28), rheumatoid arthritis (M05–06), hypertension (I10), cerebrovascular disease (I60–69), dental disease (K00–14), and diabetes (E10–14).12,19 History of these conditions was collected via self-report at each round of the survey (appendix pp 7–10). Medical records or prescriptions from doctors were required as evidence of diagnosis. Inclusion of these diagnoses in the survey results was supervised by doctors from township hospitals or higher-level health institutions, and the disease code was recorded in the questionnaire by the investigator (appendix pp 7–10).

Statistical analysis All data were documented on a written questionnaire and then double entered by two different people into an online system provided by the NHC of the People’s Republic of China. We calculated the proportion of current smokers in the entire population surveyed and calculated smoking cessation as the proportion of ever-smokers who could be classified as former smokers in the survey year. The raw prevalence of smoking in different survey years was calculated and standardised by age and gender according to the 2010 national census (appendix p 11) for the overall population and subgroups stratified by geographical regions (east, central, or west) and urbanisation (urban or rural), as well as provinces. We used Pearson χ² tests to compare the prevalence of smoking across subgroups. The duration of smoking was rounded to 10-year intervals. We applied a joinpoint regression model to analyse the percentage change in smoking prevalence over the different ages of the respondents. We constructed multiple logistic regression models to explore the association between smoking (ever-smoker vs non-smoker) and all NCDs as well as the association between smoking and each NCD individually. The regression models were adjusted for geographical regions, urbanisation, gender, age, marital status, education, employment, alcohol consumption, income, and physical activity. Results are presented as odds ratios (ORs) and 95% CIs. We used restricted cubic splines to explore linear and non-linear relationships between NCDs and the age at which respondents started smoking (appendix p 12). Because less than 5% of included households were selected repeatedly in different survey rounds, there was little correlation among the households involved in the survey. We analysed the associations between smoking and NCDs from all pooled data of the three surveys. Individuals with missing values in smoking history and status were not included in the final analysis and missing values of other indicators were not imputed. We used Microsoft Access to construct our database and did all analyses with SAS version 9.4.

www.thelancet.com/respiratory Published online October 25, 2018 http://dx.doi.org/10.1016/S2213-2600(18)30432-6

3

Articles

Role of the funding source The sponsor of this study had no role in study design, data collection, data analysis, data interpretation and writing of this study. The corresponding author had access to the data and was responsible for the decision to submit the manuscript for publication.

Results We sampled 57 023 households from 28 counties in urban areas and 67 counties in rural areas in 2003, 56 456 households from 28 urban counties and 66 rural counties in 2008, and 93 613 households from 78 urban counties and 78 rural counties in 2013. The number of individuals interviewed and involved in the study was 153 450 in 2003, 145 223 in 2008, and 229 676 in 2013 (figure 1). Individuals were comparable for most 2003

demo­graphic and socioeconomic variables and healthrelated behaviours across the three survey years (appendix pp 15–16). The standardised smoking prevalence in China was consistently high, with a proportion of current smokers of 26·0% (95% CI 25·8–26·2) in 2003, 24·9% (24·8–25·1) in 2008, and 25·2% (25·1–25·4) in 2013 (p value for trend 0·5062; table 1). In 2003, smoking prevalence was much higher in men (48·4%, 95% CI 48·1–48·7) than in women (3·1%, 3·0–3·2; p<0·0001) and remained similar for each gender across each survey (p value for trend among men 0·4086; p value for trend among women 0·6940; table 1). Standardised smoking cessation among ever-smokers increased steadily, from 5·3% (5·1–5·5) in 2003 to 7·6% (7·3–8·0) in 2008 and 9·1% (8·8–9·4) in 2013 (p<0·0001). The east region showed a decline in smoking prevalence (p value for

2008 95 of 2861 counties surveyed

2013 94 of 2859 counties surveyed

156 of 2853 counties surveyed

32 sampled in east region 11 urban 21 rural

28 sampled in central region 9 urban 19 rural

35 sampled in west region 8 urban 27 rural

32 sampled in east region 11 urban 21 rural

28 sampled in central region 9 urban 19 rural

34 sampled in west region 8 urban 26 rural

52 sampled in east region 26 urban 26 rural

52 sampled in central region 26 urban 26 rural

52 sampled in west region 26 urban 26 rural

Urban 110 communities sampled in 55 streets

Urban 90 communities sampled in 45 streets

Urban 80 communities sampled in 40 streets

Urban 110 communities sampled in 55 streets

Urban 90 communities sampled in 45 streets

Urban 80 communities sampled in 40 streets

Urban 260 communities sampled in 130 streets

Urban 260 communities sampled in 130 streets

Urban 260 communities sampled in 130 streets

Rural 210 villages sampled in 105 townships

Rural 190 villages sampled in 95 townships

Rural 270 villages sampled in 135 townships

Rural 210 villages sampled in 105 townships

Rural 190 villages sampled in 95 townships

Rural 260 villages sampled in 130 townships

Rural 260 villages sampled in 130 townships

Rural 260 villages sampled in 130 townships

Rural 260 villages sampled in 130 townships

Urban 20 201 individuals in 6594 households

Urban 15 654 individuals in 5397 households

Urban 13 843 individuals in 4820 households

Urban 18 785 individuals in 6601 households

Urban 14 773 individuals in 5400 households

Urban 12 952 individuals in 4801 households

Urban 44 499 individuals in 15 601 households

Urban 44 774 individuals in 15 600 households

Urban 44 120 individuals in 15 601 households

Rural 43 083 individuals in 12 598 households

Rural 38 614 individuals in 11 382 households

Rural 62 294 individuals in 16 232 households

Rural 40 204 individuals in 12 597 households

Rural 35 444 individuals in 11 399 households

Rural 55 343 individuals in 15 658 households

Rural 45 875 individuals in 15 606 households

Rural 44 883 individuals in 15 595 households

Rural 49 537 individuals in 15 610 households

40 239 individuals excluded 39 612 aged <15 years 627 missing smoking data

153 450 individuals included 42 683 urban 110 767 rural

32 278 individuals excluded 31 270 aged <15 years 1008 missing smoking data

145 223 individuals included 40 728 urban 104 495 rural

44 012 individuals excluded 43 621 aged <15 years 391 missing smoking data

229 676 individuals included 114 908 urban 114 768 rural

528 349 included in final analysis 198 319 urban 330 030 rural

Figure 1: Study profile One county was selected in the 2003 National Health Service Surveys in China but not in 2008 and 2013 because of a change of administrative division.

4

www.thelancet.com/respiratory Published online October 25, 2018 http://dx.doi.org/10.1016/S2213-2600(18)30432-6

Table 1: The crude and standardised proportions of current smokers in China

11 531

Adjusted proportions were standardised according to the age and gender structure of the 2010 National Population Census.

··

·· 19·5% (19·1–20·0) 4972 (19·8%) 25 084 ·· 20·0% (19·4–20·6) 2714 (20·1%) 13 476 ·· 23·2% (22·5–23·9)

··

≥70

2683 (23·3%)

27·3% (26·9–27·7) 9609 (26·8%) 35 921 ·· 27·4% (26·8–28·0) 4351 (27·3%) 15 936 ·· 28·8% (28·2–29·4) 14 663 60–69

4276 (29·2%)

·· 31·5% (31·2–31·8)

31·9% (31·6–32·3) 13 754 (30·7%)

14 905 (30·1%) 49 550

44 858 ··

·· 32·1% (31·7–32·6)

31·5% (31·0–31·9) 8735 (31·0%)

9146 (31·2%) 29 350

28 206 ··

33·8% (33·3–34·2)

32·6% (32·2–33·1) 24 029 50–59

7736 (32·2%)

30 863 40–49

10 264 (33·3%)

··

··

·· 17·0% (16·6–17·3)

28·0% (27·6–28·4) 8874 (26·8%)

6830 (16·6%) 41 142

33 121 ··

·· 13·1% (12·8–13·5)

29·3% (28·9–29·8) 7287 (27·7%)

4151 (13·0%) 31 994

26 261 ··

·· 13·3% (13·0–13·6)

30·1% (29·7–30·5) 34 257

5055 (13·3%) 38 107 15–29

30–39

9890 (28·9%)

··

<0·0001 ·· ·· ·· <0·0001 ·· ·· ·· <0·0001 ·· ·· Age, years

··

·· 47·2% (46·9–47·5)

2·7% (2·6–2·8) 3662 (3·1%)

55 282 (49·3%) 112 083

117 593 ··

·· 47·0% (46·6–47·4)

2·3% (2·2–2·5) 1946 (2·6%)

34 428 (48·0%) 71 687

73 497 ·· 3·1% (3·0–3·2) 2460 (3·2%)

·· 48·4% (48·1–48·7) 37 444 (48·9%) 76 579

76 868

Male

Female

··

<0·0001 ·· ·· ·· <0·0001 ·· ·· ·· <0·0001 ·· Gender

··

··

·· 24·5% (24·3–24·7)

26·0% (25·8–26·2) 30 982 (27·0%)

27 962 (24·3%) 114 908

114 768 ··

·· 22·6% (22·3–23·0)

25·9% (25·7–26·1) 27 211 (26·0%)

9173 (22·5%) 40 728

104 495 ··

·· 23·9% (23·5–24·2)

26·9% (26·7–27·2) 29 704 (26·8%)

10 200 (23·9%) 42 683

110 767

Urban

Rural

··

<0·0001 ··

26·2% (26·0–26·5) 20 400 (26·5%)

·· ··

76 984 ··

<0·0001 ··

25·2% (24·9–25·5) 13 516 (24·9%)

·· ··

54 183 ··

<0·0001 ··

26·3% (26·0–26·6)

··

15 137 (26·0%)

··

58 261 West

www.thelancet.com/respiratory Published online October 25, 2018 http://dx.doi.org/10.1016/S2213-2600(18)30432-6

Residence

··

·· 23·5% (23·3–23·8)

25·8% (25·5–26·1) 19 628 (26·3%)

18 916 (24·2%) 78 037

74 655 ··

·· 24·5% (24·2–24·8)

25·1% (24·8–25·5) 10 546 (25·8%)

12 322 (24·6%) 50 104

40 936 ·· 11 471 (26·8%)

·· 25·2% (24·9–25·5)

26·5% (26·2–26·9)

13 296 (25·4%) 52 322

42 867

East

Central

··

<0·0001 ··

25·2% (25·1–25·4) 58 944 (25·7%)

·· ··

229 676 ··

<0·0001 ··

24·9% (24·8–25·1) 36 384 (25·1%)

·· ··

145 223 ··

<0·0001 ··

26·0% (25·8–26·2)

··

39 904 (26·0%) 153 450

··

p value Standardised proportion (95% CI) Crude proportion (%) Standardised proportion (95% CI)

p value

Total Current smokers population Current smokers

Crude proportion (%)

Individuals

Region

Crude proportion (%)

Current smokers Total population

2013 2008 2003

Total population

trend <0·0001), whereas the prevalence in west or central regions remained similarly high across all years (table 1). The highest smoking prevalence in 2013 was in the west region of China (table 1; appendix pp 17–20). Compared with the urban population, the rural population showed a significantly higher smoking prevalence in all survey years (table 1). The standardised smoking prevalence varied in dif­ ferent provinces of China, ranging from 15·4% (95% CI 14·5–16·3) in the Xinjiang Uygur Autonomous Region (Xinjiang) to 35·9% (34·7–37·1) in the Inner Mongolia Autonomous Region in 2003, 14·7% (13·7–15·6) in Xinjiang to 32·9% (31·9–34·0) in Guizhou in 2008, and 16·2% (15·4–16·9) in Xinjiang to 34·6% (33·7–35·5) in Guizhou in 2013. We identified four major patterns of change in smoking prevalence in different provinces and autonomous regions between the years 2003 to 2013 (figure 2; appendix p 21). The first pattern was character­ ised by a consistently high smoking prevalence above the national average level, such as the prevalence seen in Yunnan (ranging from a high of 30·2% [95% CI 29·3–31·1] in 2003 to a low of 28·2% [27·2–29·2] in 2008) and Jilin (ranging from a high of 28·8% [27·6–29·9] in 2003 to a low of 25·5% [24·6–26·5] in 2008). The second pattern included provinces with smoking prevalence consistently below the national average of China, such as Hainan (from 18·3% [17·2–19·4] in 2003 to 20·9% [19·7–22·1] in 2013) and Xinjiang (from 15·4% [14·5–16·3] in 2003 to 16·2% [15·4–16·9] in 2013). The third pattern was a strong decrease in smoking prevalence, such as that seen in Liaoning in the northeast, decreasing from 31·3% (30·0–32·6) in 2003 to 22·9% (22·0–23·8) in 2013 (appendix pp 22–23). The last pattern showed the opposite trend whereby smoking prevalence increased significantly, as seen in Gansu where it increased from 20·3% (19·4–21·2) in 2003 to 27·6% (26·8–28·4) in 2013 (appendix pp 22–23). In the surveys undertaken, smoking prevalence was 16 times higher in men than in women in 2003, increasing to 20 times higher in 2008 and 17 times higher in 2013. In all three surveys, smoking prevalence increased sharply from 15 years to 24 years (defined as adolescents in our study), peaked in those aged 40–50 years, and declined after age 50 years (figure 3). Although smoking cessation in the overall population of ever-smokers was low during the surveyed years, this proportion was higher in people aged 50 years or older than in those younger than 50 years (p<0·0001; appendix p 27). Gender-standardised smoking prevalence in adolescents increased from 8·3% (95% CI 8·0–8·6) in 2003 to 12·5% (12·2–12·9) in 2013. Among male adolescents, prevalence increased from 16·0% (2160/13 519) in 2003 to 23·5% (3021/12 848) in 2013. Among female adolescents, it increased from 0·4% (48/12 795) in 2003 to 1·1% (138/12 339) in 2013. The smoking prevalence for non-student adolescents was more than ten times higher than that for student adolescents: prevalence among student adolescents ranged from 0·5%

Standardised p value proportion (95% CI)

Articles

5

Articles

2003

2008

2013

Prevalance of current smoking <20% 20–25% >25%

Figure 2: The geographical distribution of current smoking in China in 2003, 2008, and 2013 Age-standardised and gender-standardised prevalence of current smoking was calculated for each province in each survey according to the 2010 National Population Census. Hong Kong, Macao, and Taiwan were not included in this survey. The province-level prevalence estimated from the sampled counties in each individual province and the geographical distribution of the county-level prevalence are provided in the appendix (p 21). 2003 2008 2013

Male Total population Female

Prevalence of smoking (%)

80

Total population

25 20 15 10 5 0

0

20

15

25

30

40 Prevalence of smoking (%)

Prevalence of smoking (%)

60

30

20

0 0

15

20

25

30

35

40

45

50

55

60

Age (years)

65

70

75

80

85

3

Female

2 1 0

0

15

20

25

30

35

40

Age (years)

Figure 3: Age-specific prevalence of current smoking in China in 2003, 2008, and 2013 The two magnifications show age-specific prevalence of current smoking for the total population younger than 30 years and for the female population younger than 40 years.

in 2003 to 1·9% in 2013 whereas among non-student adolescents it ranged from 13·1% in 2003 to 19·9% in 2013 (appendix p 24). 77·9% of current smokers started smoking in adolescence (appendix p 28). Individuals aged 15–19 years showed the highest percentage increase of smoking prevalence in 2013 of all age groups (60·2%), followed by individuals aged 19–24 years (18·1%; appendix pp 25–26). Taken together, these results indicate that the smoking prevalence in adolescents increased substantially in China between 2003 and 2013. 6

We observed a different pattern of smoking prevalence in women compared with men. Smoking prevalence in women increased steadily with age and women aged 40 years or older showed a much higher smoking prevalence than those younger than 40 years (5·1% [95% CI 4·8–5·3] vs 1·0% [0·9–1·1] in 2003, 3·7% [3·5–3·9] vs 1·0% [0·8–1·1] in 2008, and 3·7% [3·5–3·9] vs 1·6% [1·5–1·7] in 2013, with p<0·0001 for all surveyed years). Although the proportion of women who smoke in the overall smoking population was very low (ranging

www.thelancet.com/respiratory Published online October 25, 2018 http://dx.doi.org/10.1016/S2213-2600(18)30432-6

Articles

Total population

Male

Odds ratio (95% CI) p value

Odds ratio (95% CI)

p value

Female

Adolescent (aged 15–24 years)

Odds ratio (95% CI) p value

Odds ratio (95% CI) p value

Alcohol consumption No

1 (ref)

Yes

3·1 (3·0–3·1)

·· <0·0001

1 (ref) 2·9 (2·9–3·0)

·· <0·0001

1 (ref) 5·1 (4·8–5·5)

·· <0·0001

1 (ref) 7·5 (6·9–8·1)

·· <0·0001

Physical activity Yes

1 (ref)

No

1·4 (1·4–1·4)

·· <0·0001

1 (ref) 1·4 (1·4–1·4)

·· <0·0001

1 (ref) 1·3 (1·3–1·4)

·· <0·0001

1 (ref) 1·1 (1·0–1·2)

·· 0·0370

Older family member who smokes None

··

··

··

··

··

··

1 (ref)

Any older family member

··

··

··

··

··

··

1·9 (1·8–1·9)

<0·0001

··

Male older family member

··

··

··

··

··

··

1·7 (1·6–1·8)

<0·0001

Female older family member

··

··

··

··

··

··

3·3 (2·9–3·7)

<0·0001

Marital status Married

1 (ref)

Unmarried

0·7 (0·7–0·7)

·· <0·0001

1 (ref)

··

0·7 (0·6–0·7)

<0·0001 <0·0001 0·0100

Divorced

1·3 (1·2–1·4)

<0·0001

1·2 (1·1–1·3)

Widowed

1·3 (1·2–1·3)

<0·0001

1·06 (1·01–1·11)

1 (ref) 1·3 (1·1–1·5)

··

1 (ref)

··

0·0010

0·5 (0·5–0·5)

<0·0001

1·6 (1·3–1·9)

<0·0001

0·6 (0·4–1·0)

0·0710

1·2 (1·1–1·3)

<0·0001

1·4 (0·6–3·5)

0·4920

Education level High

1 (ref)

Low

1·3 (1·3–1·4)

·· <0·0001

1 (ref) 1·3 (1·3–1·4)

·· <0·0001

1 (ref) 1·4 (1·3–1·5)

·· <0·0001

1 (ref) 1·3 (1·2–1·4)

·· <0·0001

Income Not low

1 (ref)

Low

1·07 (1·05–1·10)

·· <0·0001

1 (ref) 1·05 (1·02–1·08)

·· <0·0001

1 (ref) 1·2 (1·1–1·3)

·· <0·0001

1 (ref) 0·9 (0·8–1·0)

·· 0·0070

Occupation Employed

1 (ref)

Retired

0·7 (0·7–0·7)

<0·0001

··

Student

0·1 (0·1–0·1)

<0·0001

Unemployed

1·0 (1·0–1·0)

0·1930

1 (ref)

··

1 (ref)

··

1 (ref)

<0·0001

1·0 (0·9–1·1)

0·5470

0·1 (0·1–0·1)

<0·0001

0·4 (0·3–0·5)

<0·0001

0·1 (0·1–0·1)

<0·0001

0·0030

1·1 (1·0–1·1)

0·0500

0·8 (0·8–0·9)

0·0020

0·96 (0·93–0·98)

··

··

0·7 (0·7–0·7)

··

Models were adjusted for geographical regions and urbanisation using the pooled data of the three round surveys. For adolescents, regression model included the number of female older smokers and the number of male older smokers in their family in addition to other factors.

Table 2: Risk factors associated with current smoking in total population and subgroup populations

from 2·3% to 3·1%) in China, the smoking prevalence in women younger than 40 years increased significantly from 1·0% in 2003 to 1·6% in 2013 (Cochran-Armitage trend test p<0·0001; figure 3). When considering the risk factors associated with smoking (appendix pp 6–8), we found increased odds of smoking prevalence in people who consume alcohol compared with those who do not (table 2). People with lower education levels (junior high school or less) had higher smoking prevalence than those with higher education levels (senior high school or above; table 2). People who were divorced and had not remarried had a higher smoking prevalence than the married population (table 2). Furthermore, we found that the smoking prev­ alence in adolescents was positively correlated with the number of older family members who smoked, especially the number of female older family members (table 2). Our survey data show that the prevalence of chronic diseases rose sharply over time in both male and female

populations (appendix p 29), increasing from 8·3% (95% CI 8·2–8·5) in 2003 to 13·6% (13·5–13·8) in 2013 for women and from 7·3% (7·1–7·5) to 12·7% (12·5–12·8) for men (appendix p 29). The prevalence of NCDs increased steadily with age in both men and women in all surveyed years (p value for trend <0·0001). We found that ever-smokers had a higher risk of smoking-related NCDs than did nonsmokers (table 3), with increased risk associated with early smoking initiation (OR 1·1, 95% CI 1·0–1·1; non-linearity test p<0·0001). We observed the highest OR for peptic ulcer in ever-smokers, followed by trachea or lung cancer, COPD, and dental disease (table 3). The lowest increase in risk was for rheumatoid arthritis and ischaemic heart disease. The NCDs for which male ever-smokers had the highest risk were peptic ulcer, trachea or lung cancer, COPD, dental disease, and rheumatoid arthritis, whereas for female ever-smokers the highest OR was for peptic ulcer, followed by COPD, ischaemic heart disease, and cerebrovascular disease (table 3). A non-linear decrease in

www.thelancet.com/respiratory Published online October 25, 2018 http://dx.doi.org/10.1016/S2213-2600(18)30432-6

7

Articles

Total population Odds ratio (95% CI)

Male p value

Odds ratio (95% CI)

Female p value

Odds ratio (95% CI)

p value

Smoking* Smoking-related NCDs

1·1 (1·0–1·1)

<0·0001

1·0 (1·0–1·1)

0·0020

1·1 (1·1–1·2)

<0·0001

Peptic ulcer

1·7 (1·5–1·9)

<0·0001

1·6 (1·4–1·9)

<0·0001

1·9 (1·5–2·5)

<0·0001

Trachea or lung cancer

1·5 (1·1–2·0)

0·0160

1·4 (1·0–2·0)

0·0400

1·5 (0·7–3·4)

0·2790

Chronic obstructive pulmonary disease

1·4 (1·3–1·5)

<0·0001

1·4 (1·3–1·5)

<0·0001

1·5 (1·3–1·8)

<0·0001

Dental disease

1·2 (1·1–1·3)

<0·0001

1·2 (1·1–1·3)

0·0020

1·2 (1·0–1·4)

0·0670

Rheumatoid arthritis

1·1 (1·1–1·2)

0·0010

1·2 (1·1–1·3)

0·0010

1·1 (0·9–1·2)

0·3940 <0·0001

Ischaemic heart disease

1·1 (1·0–1·2)

0·0210

1·0 (0·9–1·1)

0·9010

1·4 (1·2–1·6)

Hypertension

1·0 (1·0–1·0)

0·3150

1·0 (1·0–1·0)

0·6600

0·9 (0·9–1·0)

0·0690

Diabetes

1·0 (1·0–1·1)

0·3260

1·0 (1·0–1·1)

0·4580

1·0 (0·9–1·1)

0·7380

Cerebrovascular disease

1·0 (1·0–1·1)

0·4730

1·0 (0·9–1·0)

0·2920

1·4 (1·2–1·6)

<0·0001

Digestive system cancer

0·9 (0·8–1·1)

0·4300

0·9 (0·7–1·1)

0·2410

1·3 (0·8–2·1)

0·3870

Duration of smoking† Smoking related NCDs

1·2 (1·2–1·3)

<0·0001

1·3 (1·2–1·3)

<0·0001

1·0 (1·0–1·1)

0·1170

Digestive system cancer

1·7 (1·3–2·4)

0·0010

1·8 (1·3–2·6)

0·0010

1·3 (0·6–2·9)

0·5420

Chronic obstructive pulmonary disease

1·4 (1·3–1·6)

<0·0001

1·4 (1·3–1·6)

<0·0001

1·5 (1·2–1·8)

<0·0001

Cerebrovascular disease

1·4 (1·3–1·5)

<0·0001

1·4 (1·3–1·5)

<0·0001

1·2 (1·1–1·5)

0·0080

Ischaemic heart disease

1·3 (1·1–1·4)

<0·0001

1·3 (1·2–1·5)

<0·0001

1·1 (0·9–1·3)

0·2440

Hypertension

1·2 (1·2–1·3)

<0·0001

1·3 (1·2–1·3)

<0·0001

1·0 (1·0–1·1)

0·3660

Diabetes

1·2 (1·1–1·3)

<0·0001

1·3 (1·2–1·4)

<0·0001

0·9 (0·8–1·1)

0·3180

Dental disease

1·1 (1·0–1·2)

0·0060

1·1 (1·0–1·2)

0·0170

1·1 (0·9–1·4)

0·2580

Rheumatoid arthritis

1·1 (1·0–1·2)

0·0170

1·1 (1·0–1·2)

0·0060

1·0 (0·8–1·1)

0·7640

Trachea or lung cancer

1·6 (0·8–3·1)

0·1670

2·0 (0·9–4·4)

0·0690

0·8 (0·3–2·2)

0·6700

Peptic ulcer

1·1 (1·0–1·2)

0·1210

1·2 (1·0–1·3)

0·0280

0·8 (0·6–1·1)

0·2260

Multiple logistic regression models were adjusted for geographical regions, urbanisation, income status, gender, age, marital status, education, employment, drinking, and exercise using the pooled data. NCDs=non-communicable chronic diseases. *All smokers were involved in this analysis, including former smokers and current smokers. †Odds ratios represent change in odds per 10 years of smoking.

Table 3: Association between chronic non-communicable diseases and smoking

chronic diseases and smoking initiation age was evidenced by restricted cubic spline plots (appendix p 30). We also noted that significantly higher risk of chronic disease was associated with long-term smoking (table 3). The risk factors with the greatest association with longterm smoking were digestive system cancer, COPD, and cerebrovascular disease (table 3). 84·5% of current smokers had smoked for more than 10 years and 74·8% had smoked for more than 15 years. The prevalence of major NCDs in the population stratified by the years of smoking showed a significant increase with the duration of smoking (appendix pp 30–32).

Discussion Our serial surveys indicate the temporal trends of smoking prevalence in China since the WHO FCTC was launched worldwide. Although health care improved substantially in China during these years,18 smoking prevalence has remained consistently high, at 25·2% in 2013, and no trend of decline was observed over the period 2003–13. Furthermore, gender-standardised smoking prevalence in adolescents increased from 8·3% in 2003 to 12·5% in 2013; increases among 8

adolescents were observed in both the student and nonstudent groups, as well as in both male and female adolescents. Without decisive and immediate action, it will be difficult to accomplish the government’s goal of reducing smoking prevalence among people aged 15 years or older to 20% in 2030. In fact, China has made some progress in recent years. In 2013, the Party School of the Central Committee of Communist Party of China (CPC), the think-tank of the Communist Party, issued a comprehensive report on tobacco control.20 Subsequently, the General Office of the CPC Central Committee and the General Office of the State Council jointly released a circular requiring both party and government officials to take the lead in banning smoking in all public places. Tobacco clauses were adopted in the Advertising Law in 2015 and Charity Law in 2016. Some cities, including three largest tier 1 cities (Beijing, Shanghai, and Shenzhen), have adopted local smoke-free regulations, signs of progress in China’s effort to curb tobacco use. However, our study has shown that smoking prevalence has remained high, suggesting that the implementation of the WHO FCTC in China has so far not been effective

www.thelancet.com/respiratory Published online October 25, 2018 http://dx.doi.org/10.1016/S2213-2600(18)30432-6

Articles

enough to reduce the future burden of tobacco-related chronic diseases. The failure to successfully implement the treaty might have been due to the once dominant development ideology of prioritising the country’s gross domestic product. It is possible that the Chinese Government previously focused on the potential negative economic effects of tobacco control policies while neglecting the long-term economic damage and serious health hazards caused by widespread smoking. A comprehensive national smoke-free law prohibiting tobacco smoking in all public places and workplaces is urgently needed to effectively protect all of China’s population. The draft Regulations on Smoking Control in Public Places was submitted to the State Council in 2014 but has not yet been approved. Although some cities have adopted local smoke-free regulations, most have struggled to consistently enforce them.21 China does not have a plain packaging policy for cigarettes. The Chinese text-only warnings cover 35% of the bottom of the pack and do not provide information on specific health risks of smoking, whereas in more than 100 countries or regions worldwide, including Taiwan, Hong Kong, Macao, and Singapore, graphic warnings on cigarette packaging are required to cover 50% of the pack or more. In these Asian regions that share a similar culture to China, smoking prevalence is less than 20% and has been halved over the past 30 years.22 Taxes on tobacco products increased slightly after 2009 and reached 50·8% of the retail price in 2016;22 a huge gap still exists with regard to the FCTC’s requirement of 70% taxation.23 The China National Tobacco Corporation is a state-owned profitmaking monopoly. Tobacco control policies would inevitably be influenced by the tobacco industry because of their strong lobbying power.6 Strong actions such as raising tobacco taxes, passing nationwide smoke-free laws, and plain packaging with graphic warnings should be taken by the government to fulfil its commitment to tobacco control. Of particular concern in our results were the sharp rise in smoking prevalence among adolescents in student and non-student populations and the steady increase in prevalence among young women. To our knowledge, this is the first longitudinal study to show these findings. China still lacks enforceable legislation to protect adolescents from smoking. The Law on Protection of Minors prohibits selling tobacco products to minors and requires identification checks. However, in real life, this law is not implemented consistently. Our data showed that more than three quarters of the smoking population started smoking during their adolescence, with a low smoking cessation rate. In this situation, it is imperative to take more practical and effective actions to prevent smoking initiation and to increase smoking cessation substantially. Plain packaging policy and raising tobacco taxes—neither of which are well implemented in China— have been shown to be effective strategies in preventing smoking initiation and increase smoking cessation.

To our knowledge, this study is the first to note that smoking prevalence in adolescents is associated with the number of older family members who smoke, especially when these family members are female—a result that was partially confirmed by the GYTS, which found a correlation between maternal smoking and adolescent smoking.4 Because older family members can provide role models for adolescents, the presence of older family members who smoke might encourage adolescents to do the same. Several other social factors might contribute to the increasing smoking prevalence in adolescents in China. Sharing and gifting of cigarettes is considered acceptable for almost all social events in China; adolescents in particular are victims of this norm because it might entice them to start smoking. Cigarettes are widely available and relatively cheap in China. Moreover, low education was shown to be a risk factor for adolescent smoking and smoking prevalence in non-student adolescents was more than ten times higher than in student adolescents. A low smoking prevalence in youth in Japan has also been shown to be associated with continuation of anti-smoking education.24 All of these results suggest that making full use of education, which possibly has a role in dissuading smoking behaviours, and encouraging families to prevent adolescents from smoking might provide a more practical solution for smoking control. Our results show that smoking was related to increased risks of chronic diseases, and even higher risks were related to early smoking initiation and long-term smoking. Importantly, the rapid ageing of China’s population will compound these negative effects in the future. Our results showed that smoking was more closely tied to alcohol consumption in adolescents than in other age groups. This association has also been observed in other LMICs.25 Adolescents who consume alcohol and smoke are prone to a wide range of substance abuse, which is a concerning health issue for adolescents in many high-income countries such as the USA and the UK.26 The Chinese Government and health professionals should be vigilant against the emergence and spread of adolescent substance abuse. In China, smoking prevalence in women was much lower than the world average of 11·3% (95% CI 10·7–12·0) among women found in the 2010 Global Adult Tobacco Survey;5 this result might be associated with the traditional Chinese concept of smoking in women being regarded negatively. However, our results showed that prevalence has increased steadily with age in China, and women younger than 40 years also showed an upward trend of smoking prevalence during the 10 years or so following the introduction of the FCTC. The market share of slim cigarettes has also risen sharply in recent years. Because such cigarettes were originally designed for women, this increase suggests women might be the new target population for tobacco companies. The incorrect yet widely held notion that fine cigarettes are less harmful

www.thelancet.com/respiratory Published online October 25, 2018 http://dx.doi.org/10.1016/S2213-2600(18)30432-6

9

Articles

and low-tar cigarettes have no risks needs to be dispelled to protect the public from being misled. Smoking prevalence varied widely across provinces, which suggests that effectiveness of tobacco control could be affected by the implementation of tobacco policies, local culture and customs, and tobacco production in different regions. The phenomenon of higher smoking prevalence in rural areas than in urban areas was probably because of urban–rural disparities in enforcement of tobacco control policies.27 The smoking prevalence in Xinjiang in the western region remained well below the national average during the survey period, which might be related to local culture and customs.28,29 By contrast, Yunnan and Guizhou, which were the two provinces with the highest production of tobacco, showed consistently higher smoking prevalence than the national average during the survey period. This result indicates that economic conflicts of interest can affect the effectiveness of tobacco control, which might cause a barrier to the full implementation of the FCTC. However, potential conflicts of interest might be lessened because the leadership of FCTC implementation was transferred to the NHC in 2018. China can expect to achieve ambitious goals in tobacco control only if the country implements strong actions and steps up in anti-tobacco leadership. Our study also provided evidence of an association between smoking and the main NCDs in the Chinese population. The smoking population had a higher risk of chronic diseases including COPD, peptic ulcer, ischaemic heart disease, and trachea or lung cancer than did nonsmokers, which was in line with previous studies.15,30–32 Additionally, cerebrovascular diseases, diabetes, and hypertension were found to be associated with long-term smoking (10 years or more). The ORs in our study for the association between smoking and the prevalence of NCDs such as diabetes, hypertension, and cerebro­ vascular diseases were lower than those given in other studies.33 There are several possible reasons for our underestimation. The data collected in this study were selfreported without screening tests for NCDs such as diabetes and hypertension,34,35 which could cause relatively low diagnosis rates. Moreover, the risks of NCDs associated with smoking in our study were calculated on the whole surveyed population who were older than 15 years, whereas most NCDs are associated with people in later decades of life. Data collection might also be affected by survivor bias, especially for people older than 60 years. Our study has some other limitations. First, this is a cross-sectional study. For smoking-related NCDs, a timelag exists between smoking initiation and onset of the NCD. Therefore, the temporal relationship between smoking and a specific chronic disease cannot be de­ termined directly. However, in our survey, 84·5% of the current smokers had smoked for more than 10 years and 74·8% had smoked for more than 15 years. Because most smokers had had a very long exposure time to tobacco, the association between tobacco use and occurrence of 10

NCDs would not be significantly biased. Additionally, the prevalence of major NCDs increased with the duration of smoking, which also suggested a correlation between smoking and NCDs. Second, the sampling design aimed to represent the national population instead of the provincial-level population. However, in each province, we surveyed 8869–26  519 individuals belonging to 1196–5400 households within two to nine counties. This large sample size enabled our surveys to represent the provincial-level population to a certain extent. Finally, the latest round of survey was done 5 years ago so the data presented here cannot reveal trends in prevalence after 2013. However, a national survey done in 2015 showed an even higher smoking prevalence of 27·7%.11 Additionally, data from the China National Tobacco Corporation shows that national cigarette sales peaked in 2014, experienced a decline in 2015 and 2016, and increased again in 2017.36 This trend in cigarette sales might partly reflect the fact that smoking prevalence in China has not decreased substantially since 2013.11 These data support our results that tobacco prevalence has remained at a high plateau in China. In summary, we found that between 2003 and 2013, smoking prevalence in China remained consistently high and increased sharply in adolescents and females younger than 40 years. In response to the rapidly growing burden of smoking-related chronic diseases, the Chinese Government should adopt and implement key WHO FCTC demand-reduction measures to reduce smoking initiation. China will have a hopeful future only if the country implements strong actions against tobacco. Contributors WW and JH conceived the idea for the study and managed the project. MW, XL, SX, JH, and WW designed the study. MW, XL, WL, XZ, LW, FD, and JL collected the data, did the statistical analyses, and wrote the statistical analysis plan. MW, XL, and SX wrote the manuscript. WW and JH had full access to all the data in the study. WW took full responsibility for the integrity of the data and the accuracy of the analyses. All authors read and approved the final manuscript. Declaration of interests We declare no competing interests. Acknowledgments We thank the Centre for Health Statistics Information, National Health Commission of the People’s Republic of China, which provided outstanding support in the data collection and analysis of this study. We thank Meian He (Department of Occupational and Environmental Health School of Public Health, Tongji Medical College, Huazhong University of Science and Technology) for her helpful advice. References 1 Nikogosian H. WHO Framework Convention on Tobacco Control: a key milestone. Bull World Health Organ 2010; 88: 83. 2 Bilano V, Gilmour S, Moffiet T, et al. Global trends and projections for tobacco use, 1990–2025: an analysis of smoking indicators from the WHO Comprehensive Information Systems for Tobacco Control. Lancet 2015; 385: 966–76. 3 Ng M, Freeman MK, Fleming TD, et al. Smoking prevalence and cigarette consumption in 187 countries, 1980–2012. JAMA 2014; 311: 183–92. 4 Xi B, Liang Y, Liu Y, et al. Tobacco use and second-hand smoke exposure in young adolescents aged 12–15 years: data from 68 low-income and middle-income countries. Lancet Glob Health 2016; 4: e795–805.

www.thelancet.com/respiratory Published online October 25, 2018 http://dx.doi.org/10.1016/S2213-2600(18)30432-6

Articles

5 6 7

8

9

10

11 12

13

14 15

16 17

18 19

Giovino GA, Mirza SA, Samet JM, et al. Tobacco use in 3 billion individuals from 16 countries: an analysis of nationally representative cross-sectional household surveys. Lancet 2012; 380: 668–79. Yang G, Wang Y, Wu Y, Yang J, Wan X. The road to effective tobacco control in China. Lancet 2015; 385: 1019–28. Chinese Center For Disease Control And Prevention, The National Center for Chronic and Noncommunicable Disease Control and Prevention, under the direct jurisdiction of the Chinese Center for Disease Control and Prevention. Report on chronic disease risk factor surveillance in China in 2010. Beijing: Military Medical Science Press, 2012. Chinese Center For Disease Control And Prevention, The National Center for Chronic and Noncommunicable Disease Control and Prevention, under the direct jurisdiction of the Chinese Center for Disease Control and Prevention. Report on chronic disease risk factor surveillance in China 2004. Beijing: Military Medical Science Press, 2009. Chinese Center For Disease Control And Prevention, The National Center for Chronic and Noncommunicable Disease Control and Prevention, under the direct jurisdiction of the Chinese Center for Disease Control and Prevention. Report on chronic disease risk factor surveillance in China 2007. Beijing: Military Medical Science Press, 2010. Chinese Center For Disease Control And Prevention, The National Center for Chronic and Noncommunicable Disease Control and Prevention, under the direct jurisdiction of the Chinese Center for Disease Control and Prevention. Report on chronic disease risk factor surveillance in China 2013. Beijing: Military Medical Science Press, 2016. Liang X. China adult tobacco survey report 2016. Beijing: People’s Medical Publishing House, 2016. GBD 2016 Risk Factors Collaborators. Global, regional, and national comparative risk assessment of 84 behavioural, environmental and occupational, and metabolic risks or clusters of risks, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet 2017; 390: 1345–422. United Nations. Progress towards the Sustainable Development Goals. Report of the Secretary-General. May 11, 2017. http://www.un.org/ga/search/view_doc.asp?symbol=E/2017/66&Lang=E (accessed Aug 18, 2018). Zhou M, Wang H, Zhu J, et al. Cause-specific mortality for 240 causes in China during 1990–2013: a systematic subnational analysis for the Global Burden of Disease Study 2013. Lancet 2016; 387: 251–72. Wang C, Xu J, Yang L, et al. Prevalence and risk factors of chronic obstructive pulmonary disease in China (the China Pulmonary Health [CPH] study): a national cross-sectional study. Lancet 2018; 391: 1706–17. Li Y, Zeng X, Liu J, et al. Can China achieve a one-third reduction in premature mortality from non-communicable diseases by 2030? BMC Med 2017; 15: 132. The Central Committee of Communist Party of China and the State Council. The “Healthy China 2030” blueprint. Oct 25, 2016. http:// english.gov.cn/policies/latest_releases/2016/10/25/ content_281475475062678.htm (accessed Oct 21, 2018). Meng Q, Xu L, Zhang Y, et al. Trends in access to health services and financial protection in China between 2003 and 2011: a cross-sectional study. Lancet 2012; 379: 805–14. Ezzati M, Lopez AD. Estimates of global mortality attributable to smoking in 2000. Lancet 2003; 362: 847–52.

20 Topic Group of Party School of the Central Committee of CPC. Tobacco control: international experience and China strategy. Beijing: Central Party School, 2013. 21 Chinese Association on Tobacco Control. Report on application for government information disclosure in 18 smoke-free cities in China (2017). Beijing: Chinese Association on Tobacco Control, 2018. 22 WHO. WHO technical manual on tobacco tax administration. Geneva: World Health Organization, 2010. 23 WHO. WHO framework convention on tobacco control. Geneva: World Health Organization, 2005. 24 Kota Katanoda, Yuan Jiang, Sohee Park, Min Kyung Lim, You-Lin Qiao, Manami Inoue. Tobacco control challenges in East Asia: proposals for change in the world’s largest epidemic region. Tob Control 2014; 23: 359–368. 25 Katulanda P, Ranasinghe C, Rathnapala A, Karunaratne N, Sheriff R, Matthews D. Prevalence, patterns and correlates of alcohol consumption and its association with tobacco smoking among Sri Lankan adults: a cross-sectional study. BMC Public Health 2014; 14: 612. 26 United Nations Office on Drugs and Crime. Drugs and associated issues among young people and older people. Drugs and age. World Drug Report 2018. Vienna: United Nations, 2018. 27 Chinese Association on Tobacco Control. Local tobacco control policies in different regions. 2018. http://www.catcprc.org.cn/index. aspx?menuid=19&type=article&lanmuid=189&language=cn (accessed june 25, 2018). 28 Whooley MA, Boyd AL, Gardin JM, Williams DR. Religious involvement and cigarette smoking in young adults: the CARDIA study (Coronary Artery Risk Development in Young Adults) study. Arch Intern Med 2002; 162: 1604–10. 29 Bush J, White M, Kai J, Rankin J, Bhopal R. Understanding influences on smoking in Bangladeshi and Pakistani adults: community based, qualitative study. BMJ 2003; 326: 962. 30 China Kadoorie Biobank Collaborative Group. CKB publications. http://www.ckbiobank.org/site/Research/CKB+Publications (accessed Aug 16, 2018). 31 Rosenstock S, Jorgensen T, Bonnevie O, Andersen L. Risk factors for peptic ulcer disease: a population based prospective cohort study comprising 2416 Danish adults. Gut 2003; 52: 186–93. 32 Kallberg H, Ding B, Padyukov L, et al. Smoking is a major preventable risk factor for rheumatoid arthritis: estimations of risks after various exposures to cigarette smoke. Ann Rheum Dis 2011; 70: 508–11. 33 Liu X, Bragg F, Yang L, et al. Smoking and smoking cessation in relation to risk of diabetes in Chinese men and women: a 9-year prospective study of 0·5 million people. Lancet Public Health 2018; 3: e167-e76. 34 Wang L, Gao P, Zhang M, et al. Prevalence and ethnic pattern of diabetes and prediabetes in China in 2013. JAMA 2017; 317: 2515–23. 35 Lu J, Lu Y, Wang X, et al. Prevalence, awareness, treatment, and control of hypertension in China: data from 1.7 million adults in a population-based screening study (China PEACE Million Persons Project). Lancet 2017; 390: 2549–58. 36 China Tobacco. Analysis of the national cigarette market in 2017. Beijing: China Tobacco, 2018.

www.thelancet.com/respiratory Published online October 25, 2018 http://dx.doi.org/10.1016/S2213-2600(18)30432-6

11