Open access Original research Estimation of smoking-­related mortality and its contribution to educational inequalities in life expectancy in Spain: an observational study, 2016–2019 Bárbara Piñeiro ‍ ‍,1 Sergi Trias-­Llimós ‍ ‍,1 Jeroen J A Spijker ‍ ‍,1 Amand Blanes Llorens ‍ ‍,1 Iñaki Permanyer ‍ ‍1,2 To cite: Piñeiro B, Trias-­ ABSTRACT Llimós S, Spijker JJA, et al. Objective To estimate smoking-­related mortality and its STRENGTHS AND LIMITATIONS OF THIS STUDY Estimation of smoking-­related contribution to educational inequalities in life expectancy ⇒ We used cause-­specific smoking-­related mortality mortality and its contribution data for the period 2016–2019 retrieved from the in Spain. to educational inequalities Spanish National Statistics Institute. Design Nationwide, observational study from 2016 to in life expectancy in Spain: 2019. Population-­attributable fractions were used to ⇒ Smoking per se is not a cause of death within the an observational study, 2016–2019. BMJ Open estimate age, sex and education-­specific cause-­of-­death International Classification of Diseases. 2022;12:e059370. doi:10.1136/ smoking-­attributable mortality. Life table techniques ⇒ We estimated smoking-­attributable mortality using bmjopen-2021-059370 and decomposition methods were used to estimate the population attributable fractions method. potential gains in life expectancy at age 35 and the cause-­ ⇒ We estimated life expectancies at age 35 by educa- ► Prepublication history and tional attainment using life table calculations. additional supplemental material specific contributions of smoking-­related mortality to life expectancy differences across educational groups. ⇒ Estimates of smoking-­related mortality need to rely for this paper are available online. To view these files, Setting Spain. on indirect techniques and the results should be in- please visit the journal online Participants We use cause-­specific mortality data from terpreted with this caution in mind. (http://dx.doi.org/10.1136/​ population registers and smoking prevalence from the bmjopen-2021-059370). National and the European Health Survey for Spain from 2017 and 2019/2020, respectively. smoking epidemic varies considerably across Received 19 November 2021 Results We estimated 219 086 smoking-­related deaths countries.2 Yet, all countries with available Accepted 23 July 2022 during 2016–2019, equalling 13% of all deaths, 83.7% of data have this commonality: the smoking those in men. In the absence of smoking, potential gains in prevalence decline has tended to start earlier male life expectancy were higher among the low-­educated among those populations with higher levels than the high-­educated (3.1 vs 2.1 years). For women, of education, occupation and income. These educational differences were less and also in the opposite widening socioeconomic inequalities in direction (0.6 vs 0.9 years). The contribution of smoking smoking prevalence have resulted in a rise in to life expectancy differences between high-­educated and low-­educated groups accounted for 1.5 years among men, smoking-­related mortality inequalities.3–10 and −0.2 years among women. For men, the contribution In Spain, the smoking prevalence has of smoking to these differences was mostly driven by declined during the last 15 years, among cancer in middle age, cardiometabolic diseases at younger men from 31.2% in 2009, to 27.6% in 2014, ages and respiratory diseases at older ages. For women, and 23.3% in 2019/2020; and among women © Author(s) (or their the contribution to this gap, although negligible, was driven from 21.3% in 2009, to 18.6% in 2014, and employer(s)) 2022. Re-­use by cancer at older ages among the higher educated. to 16.4% in 2019/2020; in part due to the permitted under CC BY-­NC. No Conclusions Smoking remains a relevant preventable risk introduction of the tobacco control policies commercial re-­use. See rights and permissions. Published by factor of premature mortality in Spain, disproportionately in 2006 and 2011.11 However, smoking preva- BMJ. affecting life expectancy of low-­educated men. lence is still relatively high (19.8%) overall12 1 Centre d'Estudis Demogràfics, compared with other European countries.13 Centres de Recerca de Like in other European countries, the Catalunya (CERCA), Universitat BACKGROUND burden of smoking-­related diseases has always Autònoma de Barcelona, Bellaterra-­Barcelona, Catalonia, Even though smoking prevalence has been been more prevalent among men in Spain Spain declining over the last decades in many high-­ because women began to smoke in large 2 ICREA, Passeig de Lluís income countries, particularly among men, numbers several decades later, initially among Companys 23, Barcelona, Spain smoking remains an important preventable the higher educated. Although women never Correspondence to risk factor of premature death worldwide, reached the same peak in smoking prevalence Dr Bárbara Piñeiro; accounting for >7 million deaths per year.1 as men did, from about 1980 smoking preva- ​bpineiro@​ced.​uab.​es In Europe, the timing and impact of the lence slowly began to converge.14 As a result, Piñeiro B, et al. BMJ Open 2022;12:e059370. doi:10.1136/bmjopen-2021-059370 1 Open access the generation of Spanish women who have smoked the life expectancy (PGLE) by eliminating smoking-­related most are those born in the 1960s. It should be noted mortality according to age, sex and smoking-­ related that more than 50 years of educational expansion in cause of death. Spain have led to a complete reversal in the educational composition by age group/birth cohort. While most of the Spanish elderly population never completed more than primary level education and only a small and highly METHODS selected minority obtained a university degree, about Data 80% of today’s adults aged 35–44 have at least completed We use sex, age, education and cause-­specific mortality secondary school. Moreover, among younger genera- data for individuals aged 35 and over for the period 2016– tions, women have already overtaken men, and the lower-­ 2019 retrieved from the Spanish National Statistics Insti- educated the higher-­educated, thereby exacerbating the tute (INE). The total number of deaths used in this study already existing socioeconomic differences in health and was 1 667 741 and the person-­years at risk was 130 568 mortality.9 15–17 843. INE obtained the deaths according to educational There is evidence suggesting that using nationally repre- attainment through multiple data linkages, including sentative data and statistically controlling for sociodemo- municipal population registers (Padrón), the 2011 graphic characteristics such as age, sex and education, are census or data on official degrees issued by the Ministry essential for accurately estimating mortality attributable of education.25 Population estimates from 2018 by age, to smoking.18 Previous studies assessing social inequal- sex and educational attainment, also retrieved from INE, ities in smoking-­related mortality for European popula- were used as denominators to produce our mortality indi- tions did not consider Spain as a whole but only included cators. Based on both data sources, three International data for just a few Spanish regions (ie, Barcelona, Basque Standard Classification of Education (ISCED) categories Country and/or Madrid) with data that is over a decade have been distinguished: low (primary education or less, old.19 This is fundamentally due to the lack of national-­ ISCED 0–1), middle (completed lower-­level and upper-­ level databases and the fact that cause-­specific mortality level secondary education and intermediate vocational data according to educational level was not collected and training, ISCED 2–3) and high (postsecondary voca- made available to researchers until recently. As cause-­ tional training and university, ISCED 4+). Educational specific mortality data according to educational attain- attainment was unknown in 1.96% of the deaths. In those ment has recently become available in Spain, it provides cases, educational attainment was distributed following an excellent opportunity to analyse in detail sociodemo- the educational distribution of deaths for each specific graphic and socioeconomic characteristics in mortality cause, sex and age group. Figure 1 shows the educational attributed to smoking. Having an estimation about the composition of the population by sex and age group/ smoking-­attributable mortality is necessary to accurately birth cohort. monitor the smoking epidemic at the population level, We selected and grouped causes of death related to for example, estimating life expectancies, one of the most smoking according to the Report of the Surgeon General26 popular indicators to monitor population-­level health.20 following the 10th revision of the International Classifica- This information can be used to reinforce the impor- tion of Diseases (ICD-­10),27 namely cancer, cardiometa- tance of comprehensive tobacco control policies and bolic and respiratory diseases (see table 1). programmes, one of the priorities set out in the WHO Smoking prevalence by sex, age groups (35–54; 55–64; Framework Convention on Tobacco Control.21 In Spain, 65–74 and 75+) and education (low, medium, high) was a recent study has estimated the country wide smoking-­ required to estimate cause-­ of-­ death specific smoking-­ attributable mortality in Spain with data from 2016.22 attributable mortality. Smoking prevalence was derived This study suggests the presence of social inequalities in from combined estimates of the 2017 National Health smoking-­related mortality, and it is highest among the Survey and the 2019/2020 European Survey of Health in low-­educated men. However, it is unknown to what extent Spain (table 2). Individuals who either smoked daily or smoking contributes to the existent gender gaps in life occasionally were considered as ‘smokers’, and those who expectancy. While recent studies found notable educa- had smoked in the past as ‘former smokers’. The 0.1% tional differences in life expectancy in Spain,23 24 little that did not know or did not respond were removed from is yet known about the role of smoking in causing these the sample. differences. The relative risks (RRs) of cause-­specific mortality for The aim of this study is to estimate the contribution current and former smokers compared with lifetime non-­ of smoking to life expectancy by educational attain- smokers were derived from the follow-­up of five cohort ment, and the contribution of cause-­specific smoking-­ studies: the National Institutes of Health-­ AARP Diet related mortality to educational inequalities in life and Health Study, the American Cancer Society’s CPS-­II expectancy in Spain. To our knowledge, this is the first Nutrition Cohort, the Women’s Health Initiative (WHI), study conducted in Spain that will estimate smoking-­ the Nurses’ Health Study and the Health Professionals attributable mortality in recent years and will quantify Follow-­Up Study26 28 and were retrieved from Haeberer the educational differences in terms of potential gains in et al.22 2 Piñeiro B, et al. BMJ Open 2022;12:e059370. doi:10.1136/bmjopen-2021-059370 Open access Figure 1 Distribution of the Spanish population by sex, age and educational attainment, 2018. Data analysis the contribution of smoking to life expectancy differences We estimated smoking-­attributable mortality using popu- between high and low-­educated groups. To do so, we used lation attributable fractions (PAFs) by age groups, cause the approach developed by Andreev et al.30 In mortality of death, sex and educational level.26 For each subpopu- studies, decomposition analysis is based on the principle of lation group, the share of deaths attributable to smoking separating life expectancy into demographic components was estimated as follows: that contribute to their dynamics, for instance, to assess ( ) ( ) prevsmokers × RRsmokers −1 +prevformer × RRformer −1 the contribution that age and cause of death groups make PAF = ( ) ( ) ‍ prevsmokers × RRsmokers −1 +prevformer × RRformer −1 +1 ‍ to differences in life expectancy between populations Where ‍prevsmokers‍is the smoking prevalence, ‍prevformer ‍ (eg, between men and women, different time periods or is the prevalence of former smokers, ‍RRsmokers‍ are the educational attainment categories). The approach devel- relative risks in smokers compared with never smokers oped by Andreev et al30 consisted of: (i) decomposing the and ‍RRformer ‍ are the relative risks of former smokers differences in life expectancy between two populations compared with never smokers. (high and low educational groups in our case) into the The estimates of cause-­specific mortality attributable to contribution of each age-­specific group and (ii) decom- smoking was estimated for each subpopulation by multi- posing these age-­specific contributions (estimated in (i)) plying the PAFs by the corresponding death counts. We into the contribution of individual smoking-­related and assumed that the PAFs remain constant within age groups non-­smoking-­related causes of death (for both smoking (ie, we applied the fraction estimated for the age group and non-­smoking related mortality), which was done by 35–54 to all 5-­year age groups (35–39, 40–44, 45–49 and multiplying the age-­specific contributions by the relative 50–54). The intermediate results of PAFs and cause-­ importance each cause to total mortality differences in specific death counts by age, sex and educational level is each age group. attached as online supplemental material 1. Finally, we grouped the causes of death into three We estimated sex-­ specific and education-­ specific life groups of smoking-­ related causes (cancer, cardiometa- expectancies at age 35 using standard life table calcula- bolic diseases and respiratory diseases) and one group tions.29 We used single-­decrement life tables where it was for all non-­smoking-­related causes to visualise the contri- assumed that the time lived for those who died was iden- bution of age groups and groups of causes of death to the tical in each age interval (nax=n/2, where n is the length educational differences in life expectancy. of the age interval) in both the life table for all-­cause Patient and public involvement mortality and the life table for all-­cause mortality excluding There was no formal patient or public involvement in the smoking-­ related mortality. This implies that smoking-­ project design, data collection or analysis. related deaths are assumed to occur, on average, in the middle of the interval. We then examined the contribution RESULTS of smoking-­related mortality on life expectancy at age 35 Sample characteristics by calculating PGLE by re-­estimating our life tables after For the period 2016–2019, we estimated 219 086 smoking-­ eliminating smoking-­ attributable mortality. We studied related deaths among the individuals aged 35 years and Piñeiro B, et al. BMJ Open 2022;12:e059370. doi:10.1136/bmjopen-2021-059370 3 Open access among the low educated (3.1 years) and lowest among Table 1 Smoking-­related deaths coded according to the 10th revision of the International Classification of Diseases the high educated (2.1 years), for women, there were (ICD-­10) few educational differences in PGLE when smoking was eliminated. In fact, the PGLE were slightly higher among Diseases ICD-­10 code the high-­educated (0.9 years) compared with the low-­ Cancer   educated (0.6 years).  Malignant neoplasms of the lung, C33-­C34 Table 3 also shows that smoking has made a noticeable bronchus, trachea contribution to sex differences in life expectancy. At age  Lip, oral cavity, pharynx C00-­C14 35, women outlived men by 5.2 years but that gap would  Oesophagus C15 have been only 3.2 in the absence of smoking.  Stomach C16 Contribution of cause-specific smoking-related mortality to  Pancreas C25 educational gaps in life expectancy  Larynx C32 Among men, out of the 5.3 years’ high-­low education gap  Cervix uteri C53 in life expectancy at age 35, smoking-­related causes of death accounted for 1.5 years (28.2%) (figure 2). In the  Kidney, renal pelvis C64-­C6 case of women, the corresponding contribution was −0.2  Urinary bladder C67 years out of the 2.9 years (−5.5%). These smoking contri-  Acute myeloid leukaemia C92.0 butions to educational differences in life expectancy  Colon and rectum C18-­C20 were found to vary across age and cause of death. Among  Liver C22 men, the contributions of smoking showed an inverse U-­shape pattern peaking at around age 60–64 and with Cardiometabolic   low contributions at young and old age groups. Smoking-­  Ischaemic heart diseases I20-­I25 related cancer deaths were most prominent among 50–74  Other heart diseases I00-­I09, I26-­I28, years old, and showed overall the highest contributions. I30-­I51 Cardiometabolic causes attributed to smoking played  Cerebrovascular diseases I60-­I69 an important role in middle-­ aged (35–64), whereas  Atherosclerosis I70 smoking-­related respiratory diseases were proportionally  Aortic aneurysm I71 most important above age 75. Among women, contribu- tions from smoking-­related causes were rather small at  Other vascular diseases I72-­I78 ages under 60 years, and they were negative after age 60  Diabetes mellitus E10-­E14 as smoking and smoking-­attributable mortality was higher Respiratory   among higher educated groups compared with their low  Influenza J09-­J11 educated counterparts. For additional analyses of high-­  Pneumonia J12-­J18 medium and medium-­low educated groups (see online supplemental figures S1 and S2).  Tuberculosis A15-­A19  Chronic obstructive pulmonary disease J40-­J44 DISCUSSION older, equalling 13.0% of all deaths. Of those who died Summary of the main results from smoking-­ attributable mortality, 83.7% were men, With this study, we aimed to estimate the contribution 51.6% were only aged 35–74 years and 48.5% had low of smoking to life expectancy by educational attainment educational attainment. and the contribution of cause-­specific smoking-­related mortality to educational inequalities in life expectancy PGLE if smoking is eliminated among Spanish men and women. Our findings show that The educational differences in the excess number of life expectancy would have been around 2.6 years higher smoking-­related deaths also translated to differences in among men and 0.6 years among women if smoking was life expectancy. Consistent with their higher levels of eliminated. This means that smoking translates in lower smoking-­attributable mortality, less educated men would life expectancy for both sexes but the contribution is experience the greatest life expectancy gains if smoking noticeably larger among men. was eliminated (table 3). Overall, life expectancy at age Our findings also reveal that life expectancy was lower 35 equalled 46.1 years for men and 51.3 years for women among the group with low levels of education than in 2016–2019 but was 5.3 years higher among the high-­ those with high levels of education, which goes in line educated than among the low-­educated in the case of with previous studies.23 24 31 However, life expectancy men, while the difference was 2.9 years among women. and gaps differed between sex when smoking was taken Life expectancy for both sexes and all three educa- into consideration. While for men smoking had a higher tional groups would have been higher in the absence of contribution on life expectancy among low-­educated (3.1 smoking. However, while for men PGLE were greatest years) and a lower contribution among high-­educated 4 Piñeiro B, et al. BMJ Open 2022;12:e059370. doi:10.1136/bmjopen-2021-059370 Open access Table 2 Proportions of current and former smokers by age groups, sex and education level Men Women Education Education Smokers Age Low Medium High Total Low Medium High Total Current Prop N Prop N Prop N Prop N Prop N Prop N Prop N Prop N 35–54 0.43 399 0.36 1555 0.21 623 0.32 2577 0.25 204 0.32 1275 0.20 742 0.26 2221 55–64 0.33 341 0.30 556 0.22 202 0.29 1099 0.22 254 0.25 520 0.21 234 0.23 1008 65–74 0.17 251 0.18 184 0.14 92 0.17 527 0.06 132 0.13 176 0.11 62 0.09 370 75+ 0.08 149 0.10 62 0.08 23 0.09 234 0.01 41 0.05 25 0.07 14 0.02 80 Total 0.25 1140 0.33 2357 0.21 940 0.28 4437 0.10 631 0.27 1996 0.19 1052 0.20 3679 Former 35–54 0.21 184 0.26 1056 0.26 706 0.25 1946 0.13 101 0.22 805 0.26 916 0.22 1822 55–64 0.42 416 0.45 799 0.41 417 0.43 1632 0.18 186 0.28 572 0.38 382 0.28 1140 65–74 0.54 795 0.52 524 0.55 315 0.53 1634 0.12 244 0.21 258 0.31 139 0.17 641 75+ 0.54 982 0.55 302 0.57 190 0.54 1474 0.04 153 0.11 77 0.20 66 0.06 296 Total 0.42 2377 0.31 2681 0.30 1628 0.33 6686 0.10 684 0.21 1712 0.25 1503 0.19 3899 Never 35–54 0.36 315 0.37 1471 0.53 1501 0.43 3287 0.63 477 0.46 1628 0.54 1844 0.51 3949 55–64 0.26 259 0.25 453 0.37 335 0.28 1047 0.60 669 0.47 852 0.40 388 0.49 1909 65–74 0.29 441 0.31 314 0.31 183 0.30 938 0.82 1665 0.66 761 0.58 241 0.74 2667 75+ 0.38 739 0.35 180 0.35 121 0.37 1040 0.95 3460 0.84 608 0.74 194 0.92 4262 Total 0.33 1754 0.35 2418 0.50 2140 0.39 6312 0.80 6271 0.52 3849 0.56 2667 0.61 12 787 51 cases out of 37 851 (0.1%) did not have information on smoking status and were removed from the database. To correct for sample bias, sample weights that were included in the datasets were applied to obtain the smoking status prevalences. Source: National Health Survey 2017 and European Survey of Health in Spain 2019/2020. (2.1 years), the contribution of smoking among women role in the stagnation of life-­expectancy increases among resulted in narrower differences across educational men in many North-­Western European countries in the groups. The higher PGLE values among men also suggest 1950s and 1960s, and in other European countries and that they (especially lower educated men) have the most among women in more recent decades.4 6 32 to gain from quitting smoking. Our findings also show an important educational gradient with a high contribution of smoking to educa- Comparison and explanation of results tional inequalities in life expectancy among men, but Our findings are consistent with previous studies on the not among women. The contribution of smoking on impact of smoking-­attributable mortality on life expec- the life expectancy gap between the highest and the tancy that have found that smoking played an important lowest educated groups was substantial among men as it accounted for 28% of the gap, but among women, the Table 3 Life expectancy (LE) at age 35, LE excluding gap was negligible (−6%). Regarding specific smoking-­ smoking-­attributable mortality, potential gains in life related causes of death, our results show that cancer is expectancy (PGLE) by eliminating smoking-­related mortality the group of causes with the greatest burden of mortality. by sex and educational groups, Spain 2016–2019 Yet, at young ages, cardiometabolic diseases predominate LE LE without smoking PGLE within smoking-­related mortality, whereas at older ages the most prevalent smoking-­related causes contributing Men All 46.1 48.7 2.6 to educational gaps in life expectancy were respiratory High 48.3 50.4 2.1 diseases. Our findings also show that among women the Medium 46.3 48.9 2.6 contributions of smoking-­related causes are rather small Low 43.0 46.1 3.1 at ages under 60 years and they are negative after age 60 Women All 51.3 51.9 0.6 as smoking was more common among higher educated High 52.7 53.5 0.9 elderly. These inverse educational gradients in smoking-­ related mortality at ages 60 and over represent a unique Medium 51.6 52.3 0.7 feature of smoking in Spain.14–17 Therefore, our results Low 49.7 50.3 0.6 need to be interpreted in the context of differences in Piñeiro B, et al. BMJ Open 2022;12:e059370. doi:10.1136/bmjopen-2021-059370 5 Open access Figure 2 Age and cause-­specific contributions of smoking-­related mortality and non-­smoking related mortality to educational inequalities in life expectancy at age 35 between high-­low educational level groups by sex, Spain 2016–2019. the smoking epidemic model.32 33 In Spain, women inequality dimensions such as education, social class, started to smoke 30 years later than women in other sex and race and ethnicity—may be a promising frame- high-­income countries (eg, the UK or the USA) mainly work for addressing health inequities in smoking cessa- due to contextual factors, that is, political, social and tion among individuals characterised by poor cessation economic transformations including women’s participa- outcomes (ie, individuals with low socioeconomic status tion in the workforce and access to tertiary education did or marginalised groups). not really accelerate until after the Franco dictatorship Continually monitoring smoking-­attributable mortality (1939–1975). For instance, when a decline in smoking is vital to keep a country’s population healthy. Further among college students started in the USA around research on monitoring smoking-­ attributable mortality the 1970s, higher educated women in Spain started to more regularly must be a health priority in Spain since smoke. Our findings are consistent with Haeberer et al22 women picked up smoking later. This information can who found that Spanish women aged 55 years and over then be used to reinforce the importance of compre- with high socioeconomic status showed higher smoking-­ hensive tobacco control programmes. Hence, we believe related mortality. As Wensink et al34 point out, the contin- that our findings are informative for Spanish health uous increase in smoking-­attributable mortality in older authorities who develop health policies and tobacco women cohorts remains a cause for serious concern. cessation programmes. In addition, given that men and Implications and future research women show different patterns in smoking prevalence The educational disparity we have found in smoking-­ and mortality, and that they are in different stages of the related mortality among men represents one example smoking epidemic model, sex and educational perspec- of how smoking can lead to socioeconomic inequali- tive in tobacco control policies needs to be taken in ties in health and mortality. This pattern of educational consideration when designing public health interven- inequality among men can be reduced by increasing tions (see also Bilal et al14). the prevalence of never smokers or quitters among less-­ We are also aware that smoking-­attributable mortality is educated individuals. To do so, tobacco control campaigns not the only factor that affects life expectancy. There are should target this specific and vulnerable population other behavioural risk factors such as alcohol consump- by investing more efforts and resources dedicated to tion, diet, physical inactivity, that have an impact on find evidence-­ based smoking cessation interventions mortality in the Spanish population. Yet, the impact of and approaches that target this specific population who smoking on life expectancy we estimated (2.6 years for continue to suffer the most from nicotine dependence. males and 0.6 years for females) is larger as compared with As Potter et al have recently reported,35 intersectionality— the one from alcohol using multiple causes of death esti- or the interaction between multiple sociodemographic mates (0.15 years for males and 0.03 years for females).36 6 Piñeiro B, et al. BMJ Open 2022;12:e059370. doi:10.1136/bmjopen-2021-059370 Open access Future research should therefore also look into the Particularly the low-­educated population should there- contribution of other behavioural risks factors such as fore be targeted. Furthermore, the increase in smoking-­ diet or physical inactivity, and particularly on the joint attributable mortality in older high-­educated women contribution of multiple lifestyles to social inequalities in remains a cause for serious concern that should be health and mortality, in line with recent research.31 37 monitored. Limitations of this study Contributors BP wrote the first draft of the manuscript. JS, ABL and ST-­L Certain limitations should be acknowledged when inter- contributed to data acquisition. ST-­L ran the analyses with input from BP, JS and IP. All authors contributed to the study design, interpretation of analyses, critically preting the results. First of all, we analyse mortality reviewed and revised the manuscript and approved the final version. All the authors among the population aged 35 and over because accept full responsibility for the work and/or the conduct of the study, had access to smoking-­related deaths rarely occur before the age of the data, and controlled the decision to publish. 35. Likewise, we also assume that the educational attain- Funding The study is supported by the European Research Council (ERC-­ ment of the population analysed is completed as moving 2019-­COG agreement no. 864616, HEALIN; PI: I. Permanyer)—a project on healthy lifespan inequality; the Juan de la Cierva-­Formación program of the up educational categories after age 35 is considered rare Spanish Ministry of Science and Innovation (FJC-­2019-­039314-­I; PI: S. Trias-­ (even beyond age 25).38 Second, estimating smoking-­ Llimós); the Ramón y Cajal program of the Spanish Ministry of Science and attributable mortality remains a challenge. Smoking per Innovation (RYC-­2013-­14851; PI: J. Spijker); the National R&D Plan (grant no. se is not a cause of death within the ICD. Therefore, esti- PID2020-­113934RB-­I00; PI: J. Spijker)—a project on gender and SE inequalities in comorbidity and multiple causes of death; and the Catalan Government under the mates of smoking-­related mortality need to rely on indi- CERCA programme. rect techniques and the results should be interpreted Competing interests None declared. with this caution in mind. We have opted for using PAF techniques that rely on smoking prevalence and RRs to Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research. estimate smoking-­attributable mortality. The main advan- tage of these indicators is that they provide cause-­specific Patient consent for publication Not applicable. mortality estimates as compared with other methods Ethics approval The results presented in this paper are based on the analysis of secondary data provided by the Spanish National Statistics Institute (INE), which that only provide overall smoking-­related mortality esti- has been properly anonymised. Thus, patient consent for publication and ethical mates.39 40 For instance, Janssen et al37 estimated a PGLE approval was not required. of 2.4 years among men and 0.3 years among women for Provenance and peer review Not commissioned; externally peer reviewed. Spain in 2014 using the Peto-­Lopez method.39 Their esti- Data availability statement All data relevant to the study are included in the mates are slightly higher among men and slightly lower article or uploaded as supplementary information. among women compared with our more recent estimates Supplemental material This content has been supplied by the author(s). It has for 2016–2019 for the total population (2.6 years among not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been men and 0.6 years among women), but are in line with peer-­reviewed. Any opinions or recommendations discussed are solely those the expected time trends. Third, as mentioned before, of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content PAFs rely on smoking prevalence. As we have used prev- includes any translated material, BMJ does not warrant the accuracy and reliability alence of smoking that comes from self-­reported survey of the translations (including but not limited to local regulations, clinical guidelines, data, there might be some degree of underestimation. terminology, drug names and drug dosages), and is not responsible for any error Moreover, the RRs of mortality between smokers and and/or omissions arising from translation and adaptation or otherwise. non-­smokers were derived from follow-­up cohort studies Open access This is an open access article distributed in accordance with the that had been conducted in the USA as Spain lacks this Creative Commons Attribution Non Commercial (CC BY-­NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-­commercially, kind of data. We, therefore, had to assume that the RRs in and license their derivative works on different terms, provided the original work is the USA also apply in Spain, although this is a common properly cited, appropriate credit is given, any changes made indicated, and the use practice in studies in the field.22 37 41 Despite the above-­ is non-­commercial. See: http://creativecommons.org/licenses/by-nc/4.0/. mentioned limitations, the consistency of our results with ORCID iDs other studies22 37 seems to validate the accuracy of our Bárbara Piñeiro http://orcid.org/0000-0003-4225-3500 overall smoking-­attributable mortality estimates. Sergi Trias-­Llimós http://orcid.org/0000-0002-8052-6736 Jeroen J A Spijker http://orcid.org/0000-0002-3957-9553 Amand Blanes Llorens http://orcid.org/0000-0001-8411-9567 Iñaki Permanyer http://orcid.org/0000-0002-7051-5144 CONCLUSIONS Our study demonstrates that smoking remains a major preventable risk factor of premature mortality in Spain, disproportionately affecting life expectancy of low-­ educated men. This educational disparity in the contribu- REFERENCES 1 World Health Organization. Who report on the global tobacco tion of smoking to life expectancy among men represents epidemic, 2017: monitoring tobacco use and prevention policies. an example of socioeconomic inequalities in health and Geneva, Switzerland, 2017. 2 Janssen F. The role of smoking in country differences in life mortality in Spain, therefore a public health priority. expectancy across Europe, 1985-­2014. Nicotine Tob Res Tobacco control policies remain vital for addressing 2021;23:152–60. 3 Gregoraci G, van Lenthe FJ, Artnik B, et al. Contribution of smoking socioeconomic inequalities in health because educa- to socioeconomic inequalities in mortality: a study of 14 European tional differences in smoking prevalence are widening. countries, 1990-­2004. Tob Control 2017;26:260–8. Piñeiro B, et al. BMJ Open 2022;12:e059370. doi:10.1136/bmjopen-2021-059370 7 Open access 4 Martikainen P, Ho JY, Preston S, et al. The changing contribution 23 Blanes A, Trias-­Llimós S. Shorter lives with POR health: the toll of smoking to educational differences in life expectancy: indirect on Spain's less educated population. Perspectives Demogr fiques estimates for Finnish men and women from 1971 to 2010. J 2021;24. Epidemiol Community Health 2013;67:219–24. 24 Permanyer I, Spijker J, Blanes A, et al. Longevity and lifespan 5 Östergren O, Martikainen P, Lundberg O. The contribution of variation by educational attainment in Spain: 1960-­2015. alcohol consumption and smoking to educational inequalities in life Demography 2018;55:2045–70. expectancy among Swedish men and women during 1991-­2008. Int 25 INE. Estadística de Defunciones. Asignación de nivel educativo J Public Health 2018;63:41–8. a ficheros de defunciones de 2018. Método de obtención y 6 Östergren O, Martikainen P, Tarkiainen L, et al. Contribution of advertencias a usuarios [Internet], 2019. Available: https://www.​ smoking and alcohol consumption to income differences in life ine.es/metodologia/t20/t2030306_niveduc.pdf [Accessed Aug expectancy: evidence using Danish, Finnish, Norwegian and Swedish 2021]. register data. J Epidemiol Community Health 2019;73:334–9. 26 U.S. Department of Health and Human Services. The health 7 Teng A, Blakely T, Atkinson J, et al. Changing social inequalities consequences of Smoking-­50 years of progress: a report of the in smoking, obesity and cause-­specific mortality: cross-­ surgeon General. Atlanta (GA, 2014. national comparisons using COMPASS typology. PLoS One 27 World Health Organization. International statistical classification of 2020;15:e0232971. diseases and related health problems. Vol.1. Geneva: World Health 8 Giskes K, Kunst AE, Benach J, et al. Trends in smoking behaviour Organization, 2004. between 1985 and 2000 in nine European countries by education. J 28 Thun MJ, Lopez AD, Hartge P. Smoking-­Related mortality in the Epidemiol Community Health 2005;59:395–401. United States. N Engl J Med 2013;368:1753. 9 Huisman M, Kunst AE, Mackenbach JP. Educational inequalities 29 Preston S, Heuveline P, Guillot M. Demography: measuring and in smoking among men and women aged 16 years and older in 11 modeling population processes. Malden, MA: Blackwell Publishers, European countries. Tob Control 2005;14:106–13. 2001. 10 Long D, Mackenbach J, Martikainen P, et al. Smoking and 30 Andreev E, Shkolnikov V, Begun AZ. Algorithm for decomposition inequalities in mortality in 11 European countries: a birth cohort of differences between aggregate demographic measures and analysis. Popul Health Metr 2021;19:3. its application to life expectancies, healthy life expectancies, 11 Perez-­Rios M, Fernandez E, Schiaffino A, et al. Changes in the parity-­progression ratios and total fertility rates. Demogr Res prevalence of tobacco consumption and the profile of Spanish 2002;7:499–522. smokers after a comprehensive smoke-­free policy. PLoS One 31 Mackenbach JP, Valverde JR, Bopp M, et al. Determinants of 2015;10:e0128305. inequalities in life expectancy: an international comparative study of 12 INE. European Survey of Health in Spain (ESHS), 2020. Available: eight risk factors. Lancet Public Health 2019;4:e529–37. https://www.ine.es/en/prensa/eese_2020_en.pdf [Accessed Aug 32 Janssen F. Similarities and differences between sexes and countries 2021]. in the mortality imprint of the smoking epidemic in 34 low-­mortality 13 Eurostat. Tobacco consumption statistics, 2021. Available: https://​ countries, 1950-­2014. Nicotine Tob Res 2020;22:1210–20. ec.europa.eu/eurostat/statistics-explained/index.php?title=Tobacco_​ 33 Thun M, Peto R, Boreham J, et al. Stages of the cigarette epidemic consumption_statistics [Accessed Oct 2021]. on entering its second century. Tob Control 2012;21:96–101. 14 Bilal U, Beltrán P, Fernández E, et al. Gender equality and smoking: a 34 Wensink M, Alvarez J-­A, Rizzi S, et al. Progression of the smoking theory-­driven approach to smoking gender differences in Spain. Tob epidemic in high-­income regions and its effects on male-­female Control 2016;25:295–300. survival differences: a cohort-­by-­age analysis of 17 countries. BMC 15 La Cruz A. Smoking in Spain. In: Lillard DR, Christopoulou R, eds. Public Health 2020;20:39. Life-­Course smoking behavior: patterns and national context in ten 35 Potter LN, Lam CY, Cinciripini PM, et al. Intersectionality and countries. Oxford University Press: Oxford, 2015: 87–99. smoking cessation: exploring various approaches for understanding 16 Schiaffino A, Fernandez E, Borrell C, et al. Gender and educational health inequities. Nicotine Tob Res 2021;23:115–23. differences in smoking initiation rates in Spain from 1948 to 1992. 36 Trias-­Llimós S, Spijker JJA. Educational differences in alcohol-­related Eur J Public Health 2003;13:56–60. mortality and their impact on life expectancy and lifespan variation in 17 Villalbí JR, Suelves JM, Martínez C, et al. [Smoking control in Spain (2016-­2018): a cross-­sectional analysis using multiple causes Spain: current situation and priorities]. Rev Esp Salud Publica of death. BMJ Open 2022;12:e053205. 2019;93:e201907044. 37 Janssen F, Trias-­Llimós S, Kunst AE. The combined impact of 18 Lariscy JT, Hummer RA, Rogers RG. Cigarette smoking and all-­cause smoking, obesity and alcohol on life-­expectancy trends in Europe. Int and cause-­specific adult mortality in the United States. Demography J Epidemiol 2021;50:931–41. 2018;55:1855–85. 38 Lim SS, Updike RL, Kaldjian AS, et al. Measuring human capital: 19 Regidor E, Kunst AE, Rodríguez-­Artalejo F, et al. Small socio-­ a systematic analysis of 195 countries and territories, 1990-­2016. economic differences in mortality in Spanish older people. Eur J Lancet 2018;392:1217–34. Public Health 2012;22:80–5. 39 Peto R, Lopez AD, Boreham J, et al. Mortality from tobacco in 20 Luy M, Di Giulio P, Di Lego V, et al. Life expectancy: frequently used, developed countries: indirect estimation from national vital statistics. but hardly understood. Gerontology 2020;66:95–104. Lancet 1992;339:1268–78. 21 World Health Organization. Who framework convention on tobacco 40 Preston SH, Glei DA, Wilmoth JR. A new method for estimating control 2003. Geneva: World Health Organization, 2003. https://www.​ smoking-­attributable mortality in high-­income countries. Int J who.int/fctc/text_download/en/ Epidemiol 2010;39:430–8. 22 Haeberer M, León-­Gómez I, Pérez-­Gómez B, et al. Social inequalities 41 Rey-­Brandariz J, Pérez-­Ríos M, Santiago-­Pérez MI, et al. Smoking-­ in tobacco-­attributable mortality in Spain. The intersection between attributable mortality in Spain: a systematic review. Adicciones. In age, sex and educational level. PLoS One 2020;15:e0239866. Press 2021:1619. 8 Piñeiro B, et al. BMJ Open 2022;12:e059370. doi:10.1136/bmjopen-2021-059370