Preventive Medicine 123 (2019) 34–40 Contents lists available at ScienceDirect Preventive Medicine journal homepage: www.elsevier.com/locate/ypmed Burden of respiratory disease attributable to secondhand smoke exposure at T home in children in Spain (2015) Xavier Continentea,b,c, Teresa Arechavalaa,d, Esteve Fernàndeze,f,g, Mónica Pérez-Ríosh,i, Anna Schiaffinof,j, Joan B. Sorianok, Giulia Carrerasl, Ángel López-Nicolásm, Giuseppe Gorinil, Maria José Lópeza,b,c,d, ⁎ a Agència de Salut Pública de Barcelona (Public Health Agency, Barcelona), Pl. Lesseps, 1, PC 08023 Barcelona, Spain b CIBER en Epidemiología y Salud Pública (CIBERESP) (Biomedical Research Centre Network for Epidemiology and Public Health), Av. Monforte de Lemos, 3-5, Pabellón 11, Planta 0, PC 28029 Madrid, Spain c Institut d'Investigació Biomèdica de Sant Pau (IIB Sant Pau) (Institute of Biomedical Research, Barcelona), Sant Antoni Maria Claret, 167, PC 08025 Barcelona, Spain d Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra (UPF) (Experimental and Health Sciences Department, Pompeu Fabra University), Dr. Aiguader, 88, PC 08003 Barcelona, Spain e Tobacco Control Unit, Cancer Control and Prevention Program, Institut Català d'Oncologia (ICO), Granvia de L'Hospitalet, 199-203, PC 08908 L'Hospitalet de Llobregat, Spain f Cancer Prevention and Control Group, Institut d'Investigació Biomèdica de Bellvitge (IDIBELL), Granvia de L'Hospitalet, 199, PC 08908 L'Hospitalet de Llobregat, Spain g Department of Clinical Sciences, Campus de Bellvitge, School of Medicine and Health Sciences, Universitat de Barcelona, Feixa Llarga, s/n, PC 08907 L'Hospitalet de Llobregat, Spain h Epidemiology Unit, Galician Directorate for Public Health, Galician Health Authority, Xunta de Galicia, San Caetano, s/n, PC 15704 Santiago de Compostela, Spain i Department of Preventive Medicine and Public Health, School of Medicine, University of Santiago de Compostela, Praza do Obradoiro, s/n, PC 15782 Santiago de Compostela, Spain j Direcció General de Planificació en Salut, Departament de Salut, Generalitat de Catalunya, Travessera de les Corts, 131-159 (Pavelló Ave Maria), PC 08028 Barcelona, Spain k Fundación para la Investigación Biomédica del Hospital Universitario La Princesa (IISP) (The Biomedical Research Foundation of University Hospital La Princesa), Diego de León, 62 1st floor, PC 28006 Madrid, Spain l Istituto per lo Studio, la Prevenzione e la Rete Oncologica (ISPRO) (Oncological Network, Prevention & Research Institute), Via Cosimo Il Vecchio, 2, PC 50139 Firenze, Italy m Universidad Politécnica de Cartagena (UPCT) (Polytechnic University of Cartagena), Plaza Cronista Isidoro Valverde, s/n, PC 30202 Cartagena, Spain ARTICLE INFO ABSTRACT Keywords: This study aimed to estimate the number of incident cases and hospital admissions attributable to secondhand smoke Tobacco smoke pollution (SHS) exposure at home for asthma, otitis media (OM), and lower respiratory infections (LRI) in children in Spain. Child The burden of respiratory disease caused by SHS exposure was estimated in terms of incident cases and hospitalized Respiratory tract diseases cases for asthma, OM, and LRI. Estimates were calculated using the population attributable fraction. The age-specific Morbidity (0-1 year, 0-4 years, 5-11 years, and 0-11 years) prevalence of SHS exposure in children was estimated through a Hospitalization telephone survey performed in a representative sample of Spanish households with children in 2016. The risk es- timates for all diseases were selected from international meta-analyses. The number of hospitalized cases was ob- tained for each disease from the Hospital Minimum Data Set provided by the Ministry of Health of Spain. Incident cases were obtained from the Global Health Data Exchange. In 2015, SHS exposure caused an estimated total of 136,403 incident cases of the following respiratory diseases: 9058 (8.5%) cases of asthma, 120,248 (8.5%) of OM, and 7097 (13.5%) of LRI in children aged 0–14 years old in Spain. Likewise, SHS exposure caused a total of 3028 hospitalized cases, with 379 (8.5%) for asthma and 167 (8.5%) for OM in children 0–11 years old, and 2482 (11.6%) for LRI in children < 2 years old. The high burden of respiratory disease attributed to SHS exposure supports the need to improve protection of children against SHS exposure by extending smoke-free regulations to homes and cars. Abbreviations: CI, confidence interval; GHDx, Global Health Data Exchange; ICD-9-CM, International Classification of Diseases, ninth revision, Clinical Modification; ICD-10, International Classification of Diseases, tenth revision; LRI, lower respiratory infections; MDS, Hospital Minimum Data Set; OM, otitis media; OR, odds ratio; PAF, population attributable fraction; SHS, secondhand smoke ⁎ Corresponding author at: Servei d'Avaluació i Mètodes Intervenció, Agència de Salut Pública de Barcelona, Pl. Lesseps, 1, PC 08023, Spain. E-mail address: mjlopez@aspb.cat (M.J. López). https://doi.org/10.1016/j.ypmed.2019.02.028 Received 18 October 2018; Received in revised form 12 February 2019; Accepted 23 February 2019 Available online 25 February 2019 0091-7435/ © 2019 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/). X. Continente, et al. Preventive Medicine 123 (2019) 34–40 1. Introduction interview, which has been widely used in studies analyzing the asso- ciation between SHS exposure and health outcomes (The Royal College Secondhand smoke (SHS) exposure has been comprehensively as- of Physicians, 2010; Jones et al., 2012, 2011; Burke et al., 2012). The sociated with adverse health effects, especially on the cardiovascular prevalence of exposure was calculated for boys and girls separately and and respiratory systems (Öberg et al., 2010). Children are particularly for different age groups: 0 to 1 years old, 0 to 4 years old, 5 to 11 years vulnerable to the detrimental health effects of SHS, mainly due to their old, and 0 to 11 years old. less developed immune and respiratory systems and their faster breathing rate (Kerem, 1996), among other factors. Additionally, 2.2. Health outcomes and risk estimates younger children are usually very close to the source of SHS on account of their tendency to be held or to sit on their parents' laps (López and The health outcomes selected where those with sufficient and strong Continente, 2016). In children, SHS exposure has been causally asso- evidence to be causally linked to SHS exposure and with information ciated with asthma, lower respiratory infections (LRI), otitis media about incidence and hospitalizations available for Spain in 2015. The (OM), sudden infant death syndrome, and low birth weight, among health outcomes analyzed were asthma, OM and LRI, which includes other conditions (Office on Smoking and Health (US), 2006). bronchitis, bronchiolitis, and other LRI diseases. All measures were Smoke-free policies have been progressively introduced worldwide estimated for the year 2015. Risk estimates of disease from SHS ex- since the application of the World Health Organization Framework posure were selected from previous meta-analyses (Jones et al., 2012; Convention on Tobacco Control (2003) to protect people from SHS. In Jones et al., 2011; Tinuoye et al., 2013), which showed consistent Spain, a partial smoke-free law was implemented in 2006 (Law 28/ evidence of causality between SHS exposure and the health outcomes 2005) (López, 2010; BOE, 2005), which was subsequently modified and studied (Table 1). became more restrictive in 2011 (Law 42/2010) (BOE, 2010). The in- Incident cases were estimated through complex mathematical troduction of smoke-free laws reduced SHS exposure in venues directly models using different information sources according to data avail- affected by the law (enclosed workplaces, public places, and hospitality ability in each country by the Global Health Data Exchange (GHDx) venues) as well as in unregulated settings, such as homes (Fernández (Global Burden of Disease Study 2015, 2016) and were defined as the et al., 2017). After Law 42/2010 came into effect, a study found a 2% number of new cases of a given disease during the year of the study. We reduction in the proportion of Spanish households with smokers selected figures referring to incident cases of asthma, OM, and LRI in (García-Villar and López-Nicolás, 2015). However, homes are still one 2015 among children aged 0 to 14 years old in Spain, stratifying by sex of the main sources of SHS exposure among children (López et al., and age groups (0–4 years old for all diseases and 5–14 years old for 2018), who are highly exposed whenever there are smokers in the fa- asthma and OM). The codes used to define each health outcome were mily (Arechavala et al., 2018a). those used by the GHDx for the incidence of nonfatal diseases (Global Worldwide, 603,000 deaths were attributed to SHS exposure in Burden of Disease Study 2015, 2015): 1) asthma included diseases with 2004, including 166,221 deaths among children younger than 15 years tenth Revision of International Classification of Diseases (ICD-10) codes old (Öberg et al., 2011). The number of deaths caused by LRI attribu- J45-J45.5 and J45.8-J45.9; 2) OM included diseases with ICD-10 codes table to SHS exposure among children younger than 5 years old was the H65-H65.4 and H66-H66.2; and 3) LRI included diseases with ICD-10 second largest (165,000) after ischemic heart disease deaths in adults codes A48.1, J10-J10.2, J10.8-J10.9, J11-J11.2, J11.8, J12-J12.3, (379,000). In Europe, 6245 deaths caused by LRI were attributable to J12.8-J12.9, J15-J15.9, J16.0, J16.8, J17-J17.3, J17.8, J18-J18.2, SHS (Öberg et al., 2011). In children, estimates of the burden of disease J18.8-J18.9, J20-J20.9, J21-J21.1, J21.8-J21.9, J85.1, P23-P23.6, and in terms of mortality and disability-adjusted life years lost attributable P23.8-P23.9. to SHS have been published in countries such as the United Kingdom Hospitalizations were obtained from the Hospital Minimum Data (The Royal College of Physicians, 2010), Poland (Jarosińska et al., Set (MDS) provided by the Ministry of Health, Social Services and 2014), Canada (Simons et al., 2011), and New Zealand (Mason, 2016). Equality of Spain. The database gathers individual information on the These studies have mostly focused on asthma, LRI, and OM. main and secondary diagnosis by physicians in all Spanish hospitals. We In Spain, mortality in adults due to SHS-related causes was esti- subsequently selected the number of cases hospitalized, which was mated before (López et al., 2007) and after (López et al., 2016) the defined as those children aged 0 to 11 years old who had been admitted introduction of the smoke-free law. However, to the best of our at least once to a hospital with the selected health outcomes (asthma, knowledge, studies assessing the burden of SHS-related diseases attri- OM, bronchitis, bronchiolitis, and LRI) as the main admission diagnosis butable to SHS exposure in children in Spain are scarce, and none has in 2015 in Spain. The number of cases with any hospitalization was focused on incident cases and hospital admissions. The objective of this calculated according to sex and age groups depending on the health study was to estimate the number of incident cases and hospital ad- outcome (0–4 years and 5–11 years old for asthma and OM; 0–1 year missions attributable to SHS exposure at home for asthma, OM, and LRI old for bronchitis, bronchiolitis and LRI). The codes used to define each in children in Spain. health outcome were selected according to those employed in studies included in the meta-analysis used in the present study for risk esti- 2. Methods mates. The ICD-9-CM codes included for each health outcome were as follows: 1) asthma: code 493; 2) OM: codes 381.0–381.4 and 382; 3) 2.1. Exposure to secondhand smoke bronchitis: codes 466.0, 490, and 491.22; 4) bronchiolitis: codes 466.11 and 466.19; and 5) LRI: 073.0, 079.8, 466, 480–488, 490, 491.22, Data on the prevalence of SHS exposure in children were obtained 513.0, and 770.0. from a telephone survey administered in 2016 to a representative sample of 2411 Spanish households with children younger than 2.3. Estimation of burden of disease 12 years old (López et al., 2018). Households were randomly selected considering different strata according to geographical region, size of The burden of respiratory disease (for asthma, OM, bronchitis, municipality of residence and quota of sex and age of the youngest child bronchiolitis, and LRI) attributable to SHS exposure was estimated by living at home. The questionnaire gathered specific and extensive in- multiplying the population attributable fraction (PAF) to SHS by the formation related to SHS exposure in different settings and was ad- total burden of these diseases in terms of incident cases and hospitalized ministered to parents or legal tutors through landlines and mobile cases. phones. Exposure to SHS at home was defined by the existence of at The PAF, which is the proportion of a health outcome attributable to least one smoker among household members at the time of the a specific risk factor, was calculated by applying the classic 35 X. Continente, et al. Preventive Medicine 123 (2019) 34–40 Table 1 Effect estimates for the health outcomes related to SHS in children. Health outcome Odds ratio SHS exposure Age (y) Design of included Reference (95%CI) studies Asthma 1.32 (1.23–1.42) Self-reported SHS or biomarkers (cotinine) < 18 y 14 cross-sectional; 2 Tinuoye et al., 2013 case-control; 4 cohort Otitis media 1.32 (1.20–1.45) Self-reported SHS at home (any household member < 16 y Cross-sectional; case- Jones et al., 2012 smokes) control; cohort (38) Bronchitis 1.58 (1.27–1.98) Self-reported SHS at home and/or biochemical <2 y 3 cross-sectional; 4 Jones et al., 2011 validation (any household member smokes) cohort Bronchiolitis 2.51 (1.96–3.21) Self-reported SHS at home and/or biochemical <2 y 2 cross-sectional; 5 Jones et al., 2011 validation (any household member smokes) case-control Lower respiratory 1.54 (1.40–1.69) Self-reported SHS at home and/or biochemical <2 y 17 cross-sectional; 11 Jones et al., 2011 infection validation (any household member smokes) case-control; 9 cohort CI: confidence interval; SHS: secondhand smoke. epidemiological formula (Szklo and Nieto, 2014): asthma and 167 due to OM were attributable to SHS exposure. For both health outcomes, the number of cases hospitalized was slightly higher PAFSHS = (PSHS × (OR 1))/(PSHS × (OR 1) + 1) in children aged 0 to 4 years old (206 in asthma and 103 in OM) and where PSHS is the prevalence of children exposed to SHS and OR refers among boys (235 and 97, respectively). An estimated 2482 cases of to the odds ratio of morbidity related to SHS exposure. The PAF and the hospitalization due to LRI (1493 in boys and 998 in girls) were attrib- total burden of disease attributable to SHS was calculated by sex and uted to SHS exposure in infants 0 to 1 years old. When bronchiolitis was age groups for each health outcome. analyzed separately from other LRI, SHS exposure was responsible for an estimated 4036 of hospitalized cases, this number being 2415 in 2.4. Sensitivity analysis boys and 1634 in girls (Table 3). Estimates obtained from sensitivity analyses are shown in Table 4. We conducted a sensitivity analysis including alternative assump- When we used a more restrictive definition of SHS exposure (based on tions to assess different scenarios. Incident cases attributable to SHS reporting smoking inside the home), incident cases attributable to SHS were assessed according to the following alternative assumptions: 1) exposure decreased to 38,117 (a 72% decrease). On the other hand, defining SHS exposure at home as any household member usually when we assumed an overall SHS exposure including different private smoking inside the home (excluding outside areas such as balconies or and public settings, the total number of incident cases increased to terraces); 2) defining SHS exposure at home as any household member 299,239. When we used the lower and upper limits of the 95%CI of OR usually smoking either inside or outside (i.e., balcony, terrace) the and prevalence of SHS exposure, the total number of incident cases home; 3) defining SHS exposure as overall exposure in both private and ranged from 89,272 to 184,512. For hospitalizations, cases of asthma public settings; 4) using the lower 95% confidence interval (CI) of OR ranged from 105 to 831, cases of OM from 46 to 367, and cases of LRI estimates and prevalence of SHS exposure; and 5) using the upper from 660 to 5463 when we assumed different definitions of SHS ex- 95%CI of OR estimates and prevalence of SHS exposure. For hospita- posure and 95%CI of OR and prevalence of SHS exposure. When we lizations, we also tested the following assumptions: 6) for asthma, using assumed an estimate risk of hospital admissions due to asthma ex- the risk rate of hospital admissions due to asthma exacerbation on the acerbations instead of the incidence of disease, incident asthma cases basis of another recent meta-analysis (Wang et al., 2015); and 7) con- increased to 882 (an increase of 133%). Finally, when we considered sidering the main diagnosis and subsequent diagnosis as the cause of not only the main diagnosis but also secondary diagnosis, SHS exposure the hospitalizations for each disease. was responsible for 505 cases hospitalized with asthma (an increase of All sensitivity analyses were conducted for asthma, OM, and LRI 33.2%), 610 with OM (an increase of 265.3%), and 2782 with LRI (an without considering sex and age groups. The effect of changing as- increase of 12.1%). sumptions is presented in terms of both absolute and percentage changes in the number of incident cases and hospitalized cases attri- 4. Discussion butable to SHS with respect to the baseline assumption. SHS exposure caused almost 140,000 incident cases of asthma, OM 3. Results and LRI in Spanish children in the year 2015. By far the highest number of incident cases was caused by OM (120,248). Hospitalizations attri- In 2015, SHS exposure caused an estimated total of 136,403 in- butable to SHS exposure occurred mostly in infants aged 0–1 years old cident cases (65,860 cases in boys and 70,552 in girls) of the following and were due to LRI, especially to bronchiolitis (4036). respiratory conditions: 9058 of asthma, 120,248 of OM, and 7097 of Smoke-free laws introduced in Spain (Law 28/2005 and Law 42/ LRI in children aged 0 to 14 years old in Spain. The largest number of 2010) have been found to be successful in reducing SHS exposure cases attributed to SHS exposure was caused by OM (120,248). In (Fernández et al., 2017; López et al., 2013). These laws reduced SHS children 0 to 4 years old, 73,194 incident cases of OM, 7097 of LRI, and exposure not only in the public places covered by their content, but also 3560 of asthma were attributed to SHS exposure. In children 5 to in private transport and homes (Fernández et al., 2017; García-Villar 14 years old, this number was lower for OM (46647) and was higher for and López-Nicolás, 2015; Sureda et al., 2014). In some regions of Spain, asthma (5498). SHS exposure caused 57,326 incident cases of OM, 4159 there are also specific strategies implemented in primary health care of LRI and 4375 of asthma in boys. In girls, these numbers were 62,922, level focused on reducing SHS exposure in children (i.e. smoke-free 2948 and 4682, respectively (Table 2). childhood in Catalonia). Nevertheless, SHS exposure in homes remains In terms of hospitalizations in 2015, SHS exposure caused 3028 unacceptably high. In 2016 one out of every 3 children younger than cases with at least 1 hospitalization due to asthma, OM, or any LRI. In 12 years of age was exposed to SHS at home (López et al., 2018). children 0 to 11 years old, a total of 379 cases hospitalized due to Moreover, previous studies have found high nicotine concentrations in 36 X. Continente, et al. Preventive Medicine 123 (2019) 34–40 Table 2 Number of incident cases attributable to secondhand smoke exposure in children by sex. Spain, 2015. Health BOYS GIRLS TOTALa outcome Total Prevalence of PAFSHS Cases Total number Prevalence of PAFSHS Cases Total number Prevalence of PAFSHS Cases number of SHS exposureb attribut.c of incident SHS exposureb attribut.c of incident SHS exposureb attribut.c incident cases cases cases N % (95%CI) % n N % (95%CI) % n N % (95%CI) % n Asthma 0–4 y 21,594 29.4 8.6 1856 20,598 28.1 8.3 1699 42,192 28.8 8.4 3560 (25.5–33.6) (24.2–32.5) (26.0–31.8) 5–14 y 29,772 28.9 8.5 2520 34,370 29.8 8.7 2992 64,142 29.3 8.6 5498 (25.7–32.3) (26.5–33.2) (27.0–31.7) All ages 51,366 29.1 8.5 4375 54,968 29.1 8.5 4682 106,334 29.1 8.5 9058 (26.6–31.7) (26.6–31.8) (27,3–31.0) Otitis media 0–4 y 415,246 29.4 8.6 35,707 452,162 28.1 8.3 37,304 867,408 28.8 8.4 73,194 (25.5–33.6) (24.2–32.5) (26.0–31.8) 5–14 y 257,703 28.9 8.5 21,814 286,470 29.8 8.7 24,939 544,173 29.3 8.6 46,647 (25.7–32.3) (26.5–33.2) (27.0–31.7) All ages 672,949 29.1 8.5 57,326 738,632 29.1 8.5 62,922 1,411,581 29.1 8.5 120,248 (26.6–31.7) (26.6–31.8) (27.3–31.0) LRI 0–4 y 30,356 29.4 13.7 4159 22,381 28.1 13.2 2948 52,737 28.8 13.5 7097 (25.5–33.6) (24.2–32.5) (26.0–31.8) Totald 44,458 – – 65,860 1,609,581 – – 70,552 77,566 – – 136,403 CI: confidence interval; LRI: lower respiratory infection; PAF: population attributable fraction; SHS: secondhand smoke. a Totals were calculated directly using specific prevalences of SHS exposure for both girls and boys together. Totals do not exactly sum results for girls and boys. b Any household member is a smoker. SHS exposure was calculated according to the age groups 0–4 y, 5–11 y and 0–11 y. c Incident cases attributable to SHS have been calculated using all decimal digits obtained from the PAF classic formula. d Includes the sum of incident cases of asthma, OM, and LRI attributable to SHS exposure. homes with smokers. A study performed in homes with children showed (71.8%) to calculate the PAFSHS would increase the figures obtained by a median nicotine concentration of 1.04 μg/m3 in the living room and 120% (around 300,000 incident cases and almost 7000 cases of hos- 0.48 μg/m3 in the children's bedroom in homes with smokers who re- pitalized cases). ported smoking inside the home, these values being relevant to child Mortality due to asthma, OM, and LRI attributable to SHS exposure health (Arechavala et al., 2018a). among children from countries with similar mortality rates to Spain is There is strong evidence to infer a causal association between SHS lower than those found in African or Asian countries (Öberg et al., exposure and the occurrence and severity of respiratory diseases in 2011; Mason, 2016). However, morbidity caused by SHS exposure is children (Office on Smoking and Health (US), 2006). Our results not negligible and calls for further public health awareness. A study showed that the PAFSHS estimated for OM and asthma ranged from conducted in the United Kingdom estimated PAFSHS and the disease 8.3% to 8.7%, depending on sex and age group. The PAFSHS for LRI events caused by SHS exposure at home in children for 2008 and re- were higher, ranging from 11.1% to 27.6%. The markedly highest ported PAFSHS (11% of LRI, 7% of middle ear disease and 4–10% of PAFSHS were found for bronchiolitis in infants 0 to 1 years old, which asthma depending on age) similar to those shown in our study. In terms doubled those obtained for bronchitis and LRI in general. This differ- of morbidity, that study showed that 20,500 cases of LRI (children < ence was mainly due to the stronger causal relationship between SHS 3 years), 121,400 cases of OM (children 0–16 years), and 15,400 cases exposure and bronchiolitis (Jones et al., 2011). The risk estimate for LRI of asthma (children 3–14 years) were attributable to SHS exposure. was lower than for bronchiolitis, because the estimate was obtained After analyzing differences in population size (2-fold higher in the from a meta-analysis that included some other LRI (i.e., pneumonia, United Kingdom (The Royal College of Physicians, 2010) than in Spain acute respiratory infection) along with bronchiolitis, which have been (National Statistics Institute, 2016)) and similarities in PAFSHS, these found to be more weakly associated with SHS exposure. estimates seem to differ from our results (120,248 for OM and 9058 for In our study, SHS exposure was defined as living with at least one asthma in children 0–14 years and 7097 for LRI in children 0–4 years), smoker. The use of a more stringent definition of SHS exposure, such as being higher in our study for OM. However, comparisons between the children living in homes where smoking inside was reported, would two studies should be interpreted with caution, due to methodological have led to more intensive SHS exposure and therefore more con- differences such as the year when the study was conducted, the age servative criteria to calculate estimates of PAFSHS (a decrease of ap- groups analyzed, sources of data, differences in SHS exposure pre- proximately 70% according to results found in the sensitivity analysis). valences, disease incidences and differences in the ICD codes used to However, most studies analyzing the relationship between SHS ex- classify cases for the studied health outcomes. posure and health problems define exposure as reported in the present Our results show a large number of hospital admissions due to LRI. study (The Royal College of Physicians, 2010; Jones et al., 2012; Jones Around 2500 hospital admissions could have been prevented by in- et al., 2011; Burke et al., 2012). On the other hand, we could consider troducing smoke-free rules at home. When we estimated hospitaliza- this assumption as conservative, as we did not analyze global SHS ex- tions specifically due to bronchiolitis, this figure increased to > 4000. posure, taking into account exposure in other settings such as private An explanation for this huge difference could be that bronchiolitis by cars, relatives' households or outdoor settings where children have been itself is a disease with one of the strongest causal associations with SHS found to be exposed to SHS (López et al., 2018), which could harm their exposure. As previously stated, when we included other diseases of the health. As per our sensitivity analysis, use of global SHS exposure lower respiratory tract in the analysis, the risk estimate of morbidity 37 X. Continente, et al. Preventive Medicine 123 (2019) 34–40 Table 3 Number of hospitalized cases attributable to secondhand smoke exposure in children by sex. Spain, 2015. Health Boys Girls Totala outcome Total number Prevalence PAFSHS Hospitalized Total number Prevalence PAFSHS Hospitalized Total number Prevalence PAFSHS Hospital of of SHS cases of of SHS cases of of SHS admission hospitalized exposureb attribut.c hospitalized exposureb attribut.c hospitalized exposureb attribut.c cases cases cases N % (95%CI) % n N % (95%CI) % n N % (95%CI) % n Asthma 0–4 y 1563 29.4 8.6 134 888 28.1 8.3 73 2451 28.8 8.4 206 (25.5–33.6) (24.2–32.5) (26.0–31.8) 5–11 y 1202 28.9 8.5 101 798 29.8 8.7 69 2000 29.3 8.7 171 (25.7–32.3) (26.5–33.2) (27.0–31.7) All ages 2765 29.1 8.5 235 1686 29.1 8.5 143 4451 29.1 8.5 379 (26.6–31.7) (26.6–31.8) (27,3–31.0) Otitis media 0–4 y 731 29.4 8.6 62 495 28.1 8.3 40 1226 28.8 8.4 103 (25.5–33.6) (24.2–32.5) (26.0–31.8) 5–11 y 418 28.9 8.5 35 325 29.8 8.7 28 743 29.3 8.6 63 (25.7–32.3) (26.5–33.2) (27.0–31.7) All ages 1149 29.1 8.5 97 820 29.1 8.5 69 1969 29.1 8.5 167 (26.6–31.7) (26.6–31.8) (27,3–31.0) Bronchitis 0–1 y 1884 25.3 12.8 241 1169 23.2 11.9 138 3053 24.3 12.4 377 (19.3–32.4) (17.1–30.6) (19.9–29.3) Bronchiolitis 0–1 y 8738 25.3 27.6 2415 6301 23.2 25.9 1634 15,039 24.3 26.8 4036 (19.3–32.4) (17.1–30.6) (19.9–29.3) LRI 0–1 y 12,427 25.3 12.0 1493 8970 23.2 11.1 998 21,397 24.3 11.6 2482 (19.3–32.4) (17.1–30.6) (19.9–29.3) Totald 30,877 – – 1825 21,452 – – 1210 52,329 – – 3028 CI: confidence interval; LRI: lower respiratory infection; PAF: population attributable fraction; SHS: secondhand smoke. a Totals were calculated directly using specific prevalences of SHS exposure for both girls and boys together. Totals do not exactly sum results for girls and boys. b Any household member is a smoker. SHS exposure was calculated according to the age range analyzed. c Hospitalized cases attributable to SHS have been calculated using all decimal digits obtained from the PAF classic formula. d Includes the sum of hospitalized cases for asthma, OM, and LRI. due to SHS exposure for LRI decreased, since the relationship between differed much from then onward. Children born before the law came SHS and other LRI are weaker. Thus, estimates may change depending into effect may have been exposed to higher levels of SHS, and conse- on the diseases classified as LRI in the analysis. quently we may have underestimated the number of incident cases and Previous studies (Rayens et al., 2008; Rando-Matos et al., 2017) hospitalizations attributable to SHS exposure. Additionally, SHS ex- have shown a decrease in hospital admissions due to asthma in adults posure was defined on the basis of self-reported information, which after the introduction of smoke-free legislation. However, in children, may generate information bias. However, self-reported SHS exposure reductions have been less intense (Rando-Matos et al., 2017). These has been proven to have good validity (Arechavala et al., 2018b). discrepancies among adults and children might be due to differences in Moreover, the indicator used was the number of smokers at home, SHS exposure settings. Adults might be exposed in settings covered by which should not be that sensitive to this type of bias. Second, our smoke-free legislations such as hospitality venues and the workplace estimates are not based on relative risks for SHS exposure from the while the main venues of SHS exposure in children are homes and Spanish population, due to a lack of local and national evidence. private cars (Office on Smoking and Health (US), 2006), neither of Moreover, risk estimates used in this study did not take into con- which are covered by legislation in most countries, except for some, sideration some potential confounding factors such as indoor or out- such as the United Kingdom, France, Australia, the US, Canada (López door pollution. Nevertheless, we obtained risks estimates from meta- and Continente, 2016), or recently Italy in cars. Our results show fig- analyses, pooling data from different countries, used in previous studies ures for OM (167) and asthma (379) markedly lower than for LRI. to determine the burden of respiratory disease attributable to SHS ex- However, when considering the risk of hospital admissions for asthma posure in other countries (Jarosińska et al., 2014), which also made our exacerbations in the sensitivity analysis, morbidity attributable to results comparable with those of other published studies. Third, the hospitalizations due to asthma (Wang et al., 2015), which are probably number of incident cases for all diseases was not obtained from mi- due to more severe asthma exacerbations, increased by 133% to a total crodata, but from estimates published in the GHDx website, which of 882 cases. Therefore, these figures should not be disregarded since hampers the acquisition of further information about what this measure they are not trivial and might represent high health costs for children takes into account. However, incidence data for the diseases studied are and economic costs for healthcare services (Batscheider et al., 2012). not available in Spain. Finally, we analyzed heterogeneus age groups, This study has intrinsic limitations. First, we estimated the number depending on the health outcome, the measure studied, and the in- of incident cases and hospitalized cases caused by past SHS exposure by formation source used. Thus, the prevalence of SHS exposure was cal- using the most recent year (2016) exposure figures. However, since culated for children 0 to 11 years old, while data on incident cases were most children involved in this study were born after the introduction of for children 0 to 14, since incident cases were obtained from aggregated smoke-free laws, the prevalence of SHS exposure would not have data published in specific age groups instead of microdata; 38 X. Continente, et al. Preventive Medicine 123 (2019) 34–40 Table 4 Sensitivity analysis of incident cases and hospitalized cases attributable to SHS based on different alternative assumptions.a Alternative assumption Asthmab Otitis mediab LRIc TOTAL d d d n % of change n % of change n % of change n % of changed Incident cases SHS exposure based on household members smoking inside (7.6% in 0- 2524 −72.1% 33,514 −72.1% 2079 −70.7% 38,117 −72.1% 11y) SHS exposure based on household members smoking inside or outside 8109 −10.5% 107,652 −10.5% 6448 −9.1% 122,209 −10.4% (25.8% in 0-11y) Overall SHS exposure (71.8% in 0-11y)e 19,866 119.3% 263,730 119.3% 15,643 120.4% 299,239 119.4% SHS exposure and odds ratio obtained from lower 95%CI 6670 −26.4% 77,635 −35.4% 4967 −30.0% 89,272 −34.6% SHS exposure and odds ratio obtained from upper 95%CI 11,580 27.8% 163,443 35.9% 9489 33.7% 184,512 35.3% Hospitalizations SHS exposure based on household members smoking inside (7.6% in 0-11y 105 −72.3% 46 −72.5% 660 −73.4% 811 −73.2% and 5.9% in 0-1y) SHS exposure based on household members smoking inside or outside 339 −10.6% 150 −10.2% 2318 −6.6% 2807 −7.3% (25.8% in 0-11y and 22.5% in 0-1y) Overall SHS exposure (71.8% in 0-11y and 63.5% in 0-1y)d 831 119,3% 367 119,8% 5463 120,1% 6661 120.0% SHS exposure and odds ratio obtained from lower 95%CI 279 −26.4% 108 −35.3% 1895 −23.7% 2282 −24.6% SHS exposure and odds ratio obtained from upper 95%CI 484 27.7% 227 35.9% 3072 23.8% 3783 24.9% Hospital admissions according to the main or subsequent diagnosis 505 33.2% 610 265.3% 2782 12.1% 3897 28.7% Risk rate for hospital admissions due to asthma exacerbation (OR = 1.85 882 132.7% – – – – – – [95%CI:1.20–2.86]) (Wang et al., 2015) CI: confidence interval; LRI: lower respiratory infections; OR: odds ratio; SHS: secondhand smoke. a Risk estimates used are the ones showed in Table 1 except for the assumption “risk rate for hospital admissions due to asthma exacerbation”. b Children aged 0–14 years old for incident cases and children aged 0–11 years old for hospitalizations. c Children aged 0–4 years old for incident cases and children aged 0–1 years old for hospitalizations. d Percentage of change compared with the main assumption. e It includes homes, cars, public transport stations, school and nursery gates, hospitality venues, leisure centers, family or friends' houses and other places related to leisure time (Lopez et al., 2018). consequently, we were not able to disaggregate data for different age the Ministry of Health, Social Services and Equality of Spain for pro- groups. The inclusion of children aged 12–14 years old for some dis- viding us with the Hospital Minimum Data Set. eases might result in an overestimation of attributable incident cases among this group of age since some of them might already have ever Funding smoked, and be thus misclassified as passive smokers. Nonetheless, the mean age of starting tobacco consumption in Spain in 2016 was This work was supported by the Instituto de Salud Carlos III (PN I 14.1 years old (Plan Nacional sobre Drogas, 2018), minimizing thus this +D+I 2013–2016) and co-funded by the European Regional limitation. Development Fund (FEDER) (grant number PI13/02734). The To our knowledge, there are no published studies estimating the Evaluation and Intervention Methods Service of the Public Health burden of respiratory disease attributable to SHS exposure in children Agency of Barcelona and the Tobacco Control Unit of the Catalan in a southern European country. In addition, we used primary data from Institute of Oncology were also partially supported by the Ministry of a representative sample of Spanish children to determine the pre- Universities and Research of Catalonia (AGAUR) (grant numbers valence of SHS exposure, which provided specific prevalences of SHS 2014SGR1373 and 2014SGR319). This project has received funding exposure according to sex and age groups to calculate more reliable from the European Union's Horizon 2020 research and innovation estimates. We also used objective data of hospital admissions from the program under grant agreement No 681040. MDS to determine cases hospitalized due to the diseases studied, which allowed us to analyze data aggregated by using suitable age groups. Conflicts of interest 5. Conclusions The authors declare that there are no conflicts of interest. Exposure to SHS at home in 2015 caused about 140,000 incident References cases and 3000 cases hospitalized for preventable respiratory diseases such as asthma, OM, and LRI in children 0 to 14 years old. This relevant Arechavala, T., Continente, X., Pérez-Ríos, M., et al., 2018a. Second-hand smoke exposure in homes with children: assessment of airborne nicotine in the living room and burden of respiratory disease due to SHS exposure calls for improve- children's bedroom. Tob. Control. 27 (4), 399–406. https://doi.org/10.1136/ ments in the protection of children against SHS exposure through the tobaccocontrol-2017-053751. further implementation of existing legislation and the extension of Arechavala, T., Continente, X., Pérez-Ríos, M., et al., 2018b. Validity of self-reported smoke-free regulations. Regarding SHS exposure at home, educational indicators to assess secondhand smoke exposure in the home. Environ Res. 164, 340–345. https://doi.org/10.1016/j.envres.2018.03.014. campaigns and counseling about smoking cessation among household Batscheider, A., Zakrzewska, S., Heinrich, J., et al., 2012. Exposure to second-hand smoke members of families with children or alternatively interventions fo- and direct healthcare costs in children – results from two German birth cohorts, cused on promoting smoke-free homes would be needed. GINIplus and LISAplus. BMC Health Serv. Res. 12, 344. https://doi.org/10.1186/ 1472-6963-12-344. Ley 28/2005, de 26 de diciembre, de medidas sanitarias frente al tabaquismo y reg- Acknowledgments uladora de la venta, el suministro, el consumo y la publicidad de los productos del tabaco. BOE; 2005. https://www.boe.es/boe/dias/2005/12/27/pdfs/A42241- 42250.pdf. Accessed September 28, 2017. The authors would like to thank all participants of the survey to Ley 42/2010, de 30 de diciembre, por la que se modifica la Ley 28/2005, de 26 de assess SHS exposure. We also thank the Health Information Institute of 39 X. Continente, et al. Preventive Medicine 123 (2019) 34–40 diciembre, de medidas sanitarias frente al tabaquismo y reguladora de la venta, el Induc. Dis. 16, 14. https://doi.org/10.18332/tid/85717. suministro, el consumo y la publicidad de los productos del tabaco. BOE; 2010. Mason, K., 2016. Burden of disease from second-hand smoke exposure in New Zealand. N https://www.boe.es/boe/dias/2010/12/31/pdfs/BOE-A-2010-20138.pdf. Accessed Z Med J. 129 (1432), 16–25. September 4, 2017. National Statistics Institute. Resident population according to date, sex and age. Madrid. Burke, H., Leonardi-Bee, J., Hashim, A., et al., 2012. Prenatal and passive smoke exposure Spain; 2016. http://www.ine.es/jaxiT3/Datos.htm?t=9663. (Accessed October 4, and incidence of asthma and wheeze: systematic review and meta-analysis. 2017). Pediatrics. 129 (4), 735–744. https://doi.org/10.1542/peds.2011-2196. Öberg, M., Woodward, A., Jaakkola, M.S., Peruga, A., Prüss-Ustün, A., 2010. Global Fernández, E., Fu, M., Pérez-Ríos, M., Schiaffino, A., Sureda, X., López, M.J., 2017. Estimate of the Burden of Disease from Second-Hand Smoke. World Health Changes in secondhand smoke exposure after smoke-free legislation (Spain, 2006- Organization, Geneva. http://www.who.int/tobacco/publications/second_hand/ 2011). Nicotine Tob. Res. 19 (11), 1390–1394. https://doi.org/10.1093/ntr/ntx040. global_estimate_burden_disease/en/ (Accessed May 5, 2017). García-Villar, J., López-Nicolás, Á., 2015. Who is afraid of smoking bans? An evaluation Öberg, M., Jaakkola, M.S., Woodward, A., Peruga, A., Prüss-Ustün, A., 2011. Worldwide of the effects of the Spanish clean air law on expenditure at hospitality venues. Eur. J. burden of disease from exposure to second-hand smoke: a retrospective analysis of Health Econ. 16 (8), 813–834. https://doi.org/10.1007/s10198-014-0631-3. data from 192 countries. Lancet. 377 (9760), 139–146. https://doi.org/10.1016/ Global Burden of Disease Study 2015. Global Burden of Disease Study 2015 (GBD 2015) S0140-6736(10)61388-8. Causes of Death and Nonfatal Causes Mapped to ICD Codes. Seattle, United States: Office on Smoking and Health (US). The Health Consequences of Involuntary Exposure to Institute for Health Metrics and Evaluation (IHME); 2015. Tobacco Smoke: A Report of the Surgeon General. Atlanta (GA): Centers for Disease Global Burden of Disease Study 2015. Global Burden of Disease Study 2015 (GBD 2015) Control and Prevention (US); 2006. https://www.surgeongeneral.gov/library/ Results. Seattle, United States: Institute for Health Metrics and Evaluation (IHME), reports/secondhandsmoke/fullreport.pdf. (Accessed January 22, 2017). 2016. http://ghdx.healthdata.org/gbd-results-tool. Acessed October 2, 2017. Plan Nacional sobre Drogas. [Encuesta sobre uso de drogas en Enseñanzas Secundarias en Jarosińska, D., Polańska, K., Wojtyniak, B., Hanke, W., 2014. Towards estimating the España (ESTUDES 2016)]. Madrid: Ministerio de Sanidad, Servicios Sociales e burden of disease attributable to second-hand smoke exposure in Polish children. Int. Igualdad y Políticas Sociales; 2018. http://www.pnsd.mscbs.gob.es/profesionales/ J. Occup. Med. Environ. Health 27 (1), 38–49. https://doi.org/10.2478/s13382-014- sistemasInformacion/sistemaInformacion/pdf/ESTUDES_2016_Informe.pdf. 0223-6. Accessed January 30, 2019. Jones, L.L., Hashim, A., McKeever, T., Cook, D.G., Britton, J., Leonardi-Bee, J., 2011. Rando-Matos, Y., Pons-Vigués, M., López, M.J., et al., 2017. Smokefree legislation effects Parental and household smoking and the increased risk of bronchitis, bronchiolitis on respiratory and sensory disorders: a systematic review and meta-analysis. PLoS and other lower respiratory infections in infancy: systematic review and meta-ana- One 12 (7), e0181035. https://doi.org/10.1371/journal.pone.0181035. lysis. Respir. Res. 12, 5. https://doi.org/10.1186/1465-9921-12-5. Rayens, M.K., Burkhart, P.V., Zhang, M., et al., 2008. Reduction in asthma-related Jones, L.L., Hassanien, A., Cook, D.G., Britton, J., Leonardi-Bee, J., 2012. Parental emergency department visits after implementation of a smoke-free law. J. Allergy smoking and the risk of middle ear disease in children: a systematic review and meta- Clin. Immunol. 122 (3), 537–541.e3. https://doi.org/10.1016/j.jaci.2008.06.029. analysis. Arch Pediatr Adolesc Med. 166 (1), 18–27. https://doi.org/10.1001/ Simons, E., To, T., Dell, S., 2011. The population attributable fraction of asthma among archpediatrics.2011.158. Canadian children. Can J Public Health. 102 (1), 35–41. Kerem, E., 1996. Why do infants and small children breathe faster? Pediatr. Pulmonol. 21 Sureda, X., Martínez-Sánchez, J.M., Fu, M., et al., 2014. Impact of the Spanish smoke-free (1), 65–68. legislation on adult, non-smoker exposure to secondhand smoke: cross-sectional López, M.J., 2010. The Spanish smoking law: a model to be followed? Int J Public Health. surveys before (2004) and after (2012) legislation. PLoS One 9 (2), e89430. https:// 55 (3), 231–232. https://doi.org/10.1007/s00038-010-0129-4. doi.org/10.1371/journal.pone.0089430. López, M.J., Continente, X., 2016. Secondhand smoke exposure in children: it is time to Szklo, M., Nieto, F.J., 2014. Measuring associations between exposure and outcomes. In: move forward. Rev. Port. Pneumol. 22 (4), 187–188. https://doi.org/10.1016/j. Ty Field. Epidemiology: Beyond the Basics. Third Edition. Burlington, Massachusetts: rppnen.2016.05.002. Jones & Bartlett Learning, pp. 79–106. López, M.J., Pérez-Ríos, M., Schiaffino, A., et al., 2007. Mortality attributable to passive Tinuoye, O., Pell, J.P., Mackay, D.F., 2013. Meta-analysis of the association between smoking in Spain, 2002. Tob. Control. 16 (6), 373–377. https://doi.org/10.1136/tc. secondhand smoke exposure and physician-diagnosed childhood asthma. Nicotine 2006.019679. Tob. Res. 15 (9), 1475–1483. https://doi.org/10.1093/ntr/ntt033. López, M.J., Fernández, E., Pérez-Ríos, M., et al., 2013. Impact of the 2011 Spanish The Royal College of Physicians, 2010. Passive Smoking and Children. Tobacco Advisory smoking ban in hospitality venues: indoor secondhand smoke exposure and influence Group of the Royal College of Physicians, UK. of outdoor smoking. Nicotine Tob. Res. 15 (5), 992–996. https://doi.org/10.1093/ Wang, Z., May, S.M., Charoenlap, S., et al., 2015. Effects of secondhand smoke exposure ntr/nts218. on asthma morbidity and health care utilization in children: a systematic review and López, M.J., Pérez-Ríos, M., Schiaffino, A., Fernández, E., 2016. Mortality attributable to meta-analysis. Ann. Allergy Asthma Immunol. 115 (5), 396–401.e2. https://doi.org/ secondhand smoke exposure in Spain (2011). Nicotine Tob. Res. 18 (5), 1307–1310. 10.1016/j.anai.2015.08.005. https://doi.org/10.1093/ntr/ntv130. World Health Organization Framework Convention on Tobacco Control, 2003. Generva: López, M.J., Arechavala, T., Continente, X., Schiaffino, A., Pérez-Ríos, M., Fernández, E., World Health Organization. http://www.who.int/tobacco/framework/WHO_FCTC_ 2018. Social inequalities in secondhand smoke exposure in children in Spain. Tob. english.pdf (Accessed November 18, 2017). 40