The Portuguese National Asthma Survey (INAsma, Inquérito Nacional sobre Asma) was a nationwide, cross-sectional population-based telephone survey targeting people of all ages who were resident in Portugal and had a landline telephone at home. The study design has been described in detail elsewhere.20 Briefly, to obtain a representative sample of the general population a two-stage cluster and stratified sampling design was used, including two steps. First, we selected a simple random sample of households within each municipality (stratum) using a list of landline telephone numbers; overall, 24,412 telephone numbers were selected. Second, within each selected household, one eligible household resident was randomly selected and invited to participate in the study. We decided that the participant would be the last person celebrating his/her birthday; when the selected individual was younger than 15 years we interviewed the usual caregiver. Individuals unable to understand spoken Portuguese or with cognitive or physical conditions hampering the interview were excluded. Of the selected numbers, 17,698 were contacted and 11,695 excluded based on exclusion criteria (n=1569), refusal to participate (n=6028), no answer to the phone call (n=3059) or wrong numbers (n=1039).

The study was approved by the Ethics Committee of Centro Hospitalar São João.

All participants gave oral informed consent.

The INAsma was primarily designed to estimate the prevalence of asthma in the Portuguese population and was mainly based on the GA2LEN survey.21,22 It had 7 questions related to characterization of tobacco smoking, including tobacco use and cessation and ETS exposure. Questions were asked about current asthma symptoms and known diagnosis of asthma or heart disease.

Interviews were conducted between March and May of 2010 using computer-assisted telephone interview (CATI).

The definitions of the variables (smoking and disease statuses) are described in Table 1.

Home ETS exposure was reported by 26.6% of the participants (Table 3). More than half (53.2%) of the CS were exposed to ETS at home while only 14.1% of the ex-smokers and 23.2% of the non-smokers reported exposure (p<0.001). Children and young adults (<25 years old) were the most exposed (39.0% vs. 21.1% in the older population, p<0.001) (Figure S1).

The prevalence of exposure to ETS at home was highest in those with asthma (33.8% vs. 26.6% in the general population vs. 16.0% in those with self-reported heart disease) (Table 3).

In the adjusted multiple logistic regression model for exposure to ETS at home, in adults (Table S3), having a higher number of persons in the household (OR=1.58 [1.22–2.04] for 3 persons in the household and OR=2.31 [1.81–2.96] for ≥4 persons), being a current smoker (OR=7.29 [5.74–9.26]) or having current asthma (OR=2.06 [1.45–2.94]) were factors positively associated with ETS exposure. Being ≥65 years (vs. 18–64 years; OR=0.65 [0.49–0.88]) decreased the odds of being exposed. The Hosmer–Lemeshow test revealed good calibration (p=0.455) and the predictive power was good (AUC=0.765).

When analyzed by gender, the effect of current asthma was maintained only for female; being a current smoker presented the highest adjusted OR for both genders (Table 4).

Nineteen percent of the Portuguese population was CS and 17.2% were ex-smokers (Table 3). Regarding individuals with ≥18 years, 19.8% were CS. The highest proportion of CS was seen in the 35–44 age group (30.2%) with a progressive drop thereafter. In the age groups older than 44 years, the prevalence of ex-smokers exceeded the CS. Overall, there was a higher prevalence of CS in males compared to females (26.5% vs. 12.2%, respectively; p<0.001); this relation was consistent in all the age groups. In men, the 35–44 years age group was the only one with a higher proportion of CS than never-smokers (39.6% vs. 38.9%); in women, the prevalence of CS was always lower than the never-smokers.

We found a higher prevalence of CS in those with higher SES; when analyzed by gender this association was present only in females.

The mean (SD) starting age of smoking was 17.2 (4.2) years (17.2 (3.5) in men vs. 17.6 (4.8) in women, p<0.001). One in five (21.6%) started smoking before the age of 15 years.

Current smoking men, on average, smoked more cigarettes per day than women (14.9 (13.2) vs. 10.6 (7.4), p<0.001). Almost 10% (7.9%) of the CS smoked >20 cigarettes per day (10.7% of CS men, 2.5% of CS women, p<0.001).

No differences were found in CS prevalence when comparing individuals with asthma with the general population (17.8% vs. 19.0%, respectively). Those with heart disease had the lowest CS prevalence (7.8%) and the highest prevalence of ex-smokers (25.5% vs. 16.3% in those with asthma vs. 17.2 in the general population) (Table 3).

In the adjusted multiple logistic regression model for CS in adults (Table S4), the odds of being a CS was higher for the male gender, the more educated and those exposed to ETS at home (OR=7.32 [5.93–9.17]). Being ≥65 years (OR=0.41 [0.31–0.54]) decreased the odds of being CS. The Hosmer–Lemeshow test revealed good calibration (p=0.760) and the predictive power was high (AUC=0.839).

When analyzed by gender, the effect of school education was maintained only for female (Table 5).

The effect of the new tobacco control policies in the prevalence of ETS exposure at home cannot be accurately assessed as results of the available studies are difficult to compare: study questions and definitions of exposure are widely variable and most reports don’t present explicit data on home exposure.12,14,18,24 The National Health Survey 2005–2006,14 for example, refers to the exposure to ETS in closed spaces but not specifically at home. The Eurobarometer study held in 200618 reported that the prevalence of participants in whose homes smoking was not allowed was 29%, suggesting that about 71% of the population was potentially exposed to ETS at home; in the Eurobarometer 332 (from 2009),12 potential exposure was reduced to 34% of the population. To the best of our knowledge, the definition used in our study is similar to only one recent Portuguese report, ECOS201015; that study, carried out in 2010, estimated that 15.7% of the Portuguese population was exposed inside their homes, which is much lower than the prevalence we report. There are some methodological differences that might explain this discrepancy: the ECOS2010 study collected data from a sample of Portuguese families willing to be periodically contacted to answer health surveys (ECOS panel); only individuals with ≥18 years were asked to participate and there was no random selection inside each household.

Our results are in agreement with other studies25–28 in relation to a high prevalence of exposure among children and adolescents.

The prevalence of tobacco smoking in Portugal, in this study, was lower than in reports prior to the 2008 tobacco law (Fig. 1).

Figure 1.

Prevalence of tobacco smoking in the population aged 15 years or more in Portugal (data since 1994) and evolution of tobacco price (pack of premium brand); stratification by gender is also presented.12,15,18,24

(0.35MB).

Recent studies,12,14,15 conducted after the 2008 smoking bans, presented prevalence estimates of CS in Portugal that ranged between 16.4% and 25.5%; our estimate is in the same range. This wide range probably reflects different survey methods, different definitions and different samples sizes.12,14,24 The slight decreasing tendency of the prevalence of smoking, shown in Fig. 1, does not seem to be directly related with the increase in tobacco price or more restrictive smoking bans. The previously reported gender-specific trends13 are also supported by the data of this study: CS prevalence in females has been rising in the last 15 years while, in males, there was a gradual decrease (Fig. 1).

The higher CS prevalence in women with higher SES is in agreement with other data for Portugal29–31 but differs from other countries’.30–32 This discrepancy may be related to differences in the stage of the smoking epidemic,30,31 but the economic capacity to pay for cigarettes may also have a strong impact.29

Those exposed to home ETS presented the highest adjusted OR for smoking. This may suggest that ETS exposure is a risk factor for smoking initiation.25,33,34

As far as we know, in Portugal, no other population-based study has estimated the prevalence of smoking and home exposure to ETS in individuals with asthma or heart disease.

The low smoking prevalence in subjects with heart disease suggests that these individuals have some awareness of the harmful influence of tobacco use in heart pathologies.35

The consequences of smoking and ETS exposure on asthmatics36–39 and the beneficial effects of a smoke-free legislation (e.g. associated with a lower rate of hospitalizations for childhood asthma17) are increasingly recognized. In spite of this knowledge, we found a smoking prevalence in asthmatics that is similar to the general population; other population-based studies reported similar findings.39–45 To the best of our knowledge, the only Portuguese study46 reporting a CS prevalence in asthmatics was held in an asthma clinic; the prevalence was 8%. This low prevalence was explained by a possible selection bias in relation with regular follow-up by lung physicians.46

Several reports found a prevalence of ETS exposure at home in asthmatics similar to the general population.41,47,48 In our study, on the contrary, ETS exposure was higher in asthmatics; this is in agreement with the findings of Leech et al.,49 in Canada. In that study, it was suggested that this higher prevalence could be explained by an increased reporting related to greater awareness of smoke exposure in asthmatics.49 However, when interpreting these findings and considering the cross-sectional nature of our study, we cannot exclude the possibility of reverse causality (exposure to ETS at home may be a cause of asthma)

The sampling process, using landline telephone numbers, may be a limitation because of the increasing number of mobile-only households and resistance to telephone surveying50; it may also introduce a selection bias, with under-representation of some groups’, including CS.51 However, the use of the mobile telephone network as an alternative would not give a geographical stratification, would reduce the control of the sampling process and has lower response rates.51 The fact that the only available data at the time of the data collection were from the 2001 National Census may be another limitation because demographic changes in recent years could have influenced the sampling methods; however the data from the 2011 Census necessary for the weighting procedures had not yet been published.

Smoking/exposure status was self-reported which can lead to underreporting of smoking habits.52,53 The definition of current smoking used in this study did not include occasional smokers in CS or as an independent group; the inclusion of occasional smokers would be an added value for the study and its absence might lead to an under-estimation of the total current smoking prevalence. Moreover, only home exposure to ETS was questioned limiting the comparability with other reports.

The definition of diseases using questionnaires has been widely used; however, several combinations are accepted. The one we chose was associated with a higher specificity and lower asthma prevalence.20

Repeated population surveys using representative, randomly selected samples of sufficiently large size, with similar questionnaires and definitions is the only way to provide comparable estimates of smoking and ETS exposure prevalence.2 Methods to improve reporting accuracy (e.g. cotinine saliva) are reserved for smaller samples and may be part of a future study.

In conclusion, exposure to environmental tobacco smoke at home in Portugal is higher than in a previous report using the same definition. Our results suggest that children/adolescents and asthmatics, but not individuals with heart disease, may be at higher risk for home exposure to environmental tobacco smoke. This study supports a decreasing trend of the prevalence of tobacco smoking in Portuguese male, but a tendency to increase in female.