Purpose Only a few studies have examined the effect of public smoking bans on respiratory conditions. These showed reduced admission rates for different respiratory diseases.
Objective The objective of the present study was to evaluate the effect of the public smoking ban implemented in Graubünden, Switzerland, on the incidence of acute hospital admissions for acute exacerbated chronic obstructive pulmonary disease (AECOPD).
Methods We searched a database, including all nationwide hospitalisations in Switzerland, for AECOPD and analysed incidence rates before and after introduction of the smoking ban using Poisson regression and incidence rate ratios (IRRs).
Results After introduction of the smoking ban, we observed a significant 22.4% decrease in the incidence of AECOPD hospitalisations in Graubünden (IRR=0.78 (0.68 to 0.88), p<0.001). In the same period, the incidence of AECOPD hospitalisations only slightly decreased by 7.0% in the rest of Switzerland (IRR=0.93 (0.91 to 0.95), p<0.001). The observed reduction in AECOPD hospitalisation incidence was significantly greater in GR than in the rest of CH (p=0.008).
Conclusions Our study supports the limited body of evidence demonstrating that a reduction of secondhand smoke by legislated bans on smoking is associated with reduced rates of admission to hospital for respiratory conditions, hereby shown for AECOPD, in addition to the meanwhile well-documented impact on cardiovascular disease.
- Harm Reduction
- Public policy
- Secondhand smoke
- Global health
Statistics from Altmetric.com
Secondhand smoke (SHS), also called environmental tobacco smoke, defined as unintended exposure to diluted, side-stream cigarette smoke and exhaled smoke from smokers,1 has been shown to be an important reason for preventable poor health and death in the developed world, for example, due to lung cancer or heart disease, mainly coronary artery disease.2 ,3
A growing body of evidence has emerged in the last years to show significant declines of hospital admissions for acute coronary syndrome after implementation of smoking bans in public places. For example, Trachsel and Bonetti showed a reduction of acute myocardial infarction hospitalisations of 21% after the implementation of a public smoking ban in Graubünden (GR), Switzerland.4 ,5
Only a few studies have examined the effect of smoking bans on admissions for respiratory conditions. A Canadian study found a significant decrease of admission rates for asthma, chronic obstructive pulmonary disease (COPD), pneumonia and bronchitis of 33% after implementing smoking bans in restaurants in Ontario.6 Another study found a significant reduction of admission rates for asthma in Arizona, USA.7 The fact that a relevant percentage of COPD patients are never-smokers (ranging from 17% to 69%, depending on the investigated population) raises suspicion of SHS as a possible risk factor for the development of COPD in never-smokers/non-smokers.8–11
The objective of the present study was to evaluate the effect of anti-smoking legislation—in this case the public smoking ban (affecting public buildings, eg, restaurants, bars and cafés) implemented in the canton of Graubünden, Switzerland, on 1 March 2008—on the incidence of acute hospital admissions for acute exacerbated chronic obstructive pulmonary disease (AECOPD).
Patients and methods
For our analysis, we used a nation-wide database, including all hospitalisations in Switzerland (CH), provided by the Swiss Federal Office for Statistics, Neuchâtel, Switzerland.
We searched this database for AECOPD by using the International Classification of Diseases (ICD) -10-codes, which had been used for coding the primary diagnosis leading to hospitalisation. These were J40–44, including all subgroups for the diagnosis of AECOPD. Only acute/unplanned hospitalisations were taken into the analysis, whereas planned admissions, for example, for rehabilitation were excluded. Outpatients were not available in the database and could therefore not be analysed. The criteria for hospitalisation or treatment guidelines for AECOPD did not change during the study period.
We compared inhabitants of the canton of GR with inhabitants of the rest of CH except the canton of Ticino, no matter where they had been hospitalised, because we postulated that the implementation of public smoking bans mainly affects the people constantly living in this defined area by defending them from the regular exposure to SHS. In Ticino, a public smoking ban was introduced in 2007; therefore, we excluded this canton from our analysis. In the following, the term ‘rest of CH’ refers as Switzerland without GR and TI.
We analysed the total numbers of admissions and the corresponding incidences (respecting the yearly fluctuating number of inhabitants of the compared groups, table 1) for the described conditions in the 5 years before and the 2 years after implementation of a public smoking ban in the canton of GR (analysed time period: 1 March 2003 until 28 February 2010), which took place on 1 March 2008. All other cantons implemented different levels of smoking bans later, either in the second half of 2009 or in 2010 and on 1 May 2010 a nation-wide anti-smoking legislation became effective, thus, no more affecting our analysis. The evolution of the incidence rates of AECOPD was analysed using Poisson regression. The number of inhabitants living in Graubünden and the rest of Switzerland available on a yearly basis (denominator of the incidence rate) was used as offset values in the Poisson regression analysis in order to account for population size variations. The smoking ban effect (before vs after implementation) together with the regional effect (GR vs rest of Switzerland) as well as their interaction was tested. The results are reported as incidence rate ratio (IRR) and 95% CIs. Statistical analyses were performed using the R statistical software.12
From March 2003 to February 2010, the population increased from ∼6.87×106 to ∼7.27×106 in CH (without cantons Graubünden and Ticino) and from ∼187.000 to ∼192.000 in GR.13
The total number and incidence of hospital admissions due to AECOPD are demonstrated in table 1.
The age distribution of all hospitalised AECOPD patients between the rest of CH and GR was similar (median 73 years, IQR 64–80 years). Before the implementation of the smoking ban in GR, we observed a higher incidence of hospitalisations due to AECOPD in GR in comparison with the rest of CH. After the introduction of the smoking ban and despite clear seasonal variations, we observed a significant 22.4% decrease in the incidence of AECOPD hospitalisations in GR (IRR=0.78 (0.68 to 0.88), p<0.001). In the same period, the incidence of AECOPD hospitalisations only slightly decreased by 7.0% in the rest of CH (IRR=0.93 (0.91 to 0.95), p<0.001, table 1 and figure 1). The observed reduction in AECOPD hospitalisation incidence was significantly greater in GR than in the rest of CH (p=0.008). As a result in the second year after implementation of the ban, the incidences of hospitalisations due to AECOPD in GR and the rest of CH became similar (66 vs 61 per 100 000, table 1). We found no difference in the percentage of COPD patients with a concomitant asthma diagnosis before and after the smoking ban in GR (2.9% and 2.8%, respectively; Pearson’s χ2 test: p=1). This indicates that the decrease in admission rate for AECOPD did not occur primarily among asthmatics.
In Graubünden, there was no significant decrease in the incidence of hospitalisation for asthma, acute bronchitis or pneumonia after introduction of the smoking ban (data not shown).
We could demonstrate a significant and markedly greater decline in the incidence of AECOPD hospitalisations in GR as compared with the rest of CH (−22.4% vs −7.0%; p=0.008) after introduction of the smoking ban in GR. Before the smoking ban, there was a markedly higher incidence of AECOPD hospitalisations in GR compared with the rest of CH. After the smoking ban, the incidences in GR and CH became similar (table 1). The decrease in admission rate for AECOPD did not occur primarily among asthmatics.
In a Canadian study, admissions due to composite respiratory conditions (asthma, COPD and lung infection) decreased by 33% after implementation of a smoking ban in restaurants.6 A recent Irish study detected an overall reduction of acute hospital admissions for asthma after implementation of a national workplace smoking ban.14 This was also found after implementation of a state-wide public smoking ban in Arizona, USA, in May 20077 and the association between SHS and the risk of asthma and COPD has already been described before.15 ,16 It is well documented that SHS increases the risk for bronchitis and bronchiolitis in infants,17 whereas the evidence concerning adults is very limited.
Jones et al investigated the occupational exposure to SHS in bar and nightclub employees and showed markedly higher hair nicotine concentrations in employees working in places without smoking bans. This was true for smokers and non-smokers, emphasising the need for legislation measures to ensure protection from SHS in these occupations.18
Matt et al19 proposed an ineffectiveness of partial smoking bans in Californian hotels by showing that non-smoking guests staying in non-smoking rooms had higher levels of finger nicotine and urine cotinine than those staying in hotels with complete smoking bans.
Although we document a clear temporal association between the smoking ban and the decrease in the incidence of AECOPD hospitalisations, it is not possible to conclude a causal relationship. Other exogenous factors (eg, seasonal respiratory infections, geography, climate or air pollution) might influence the incidence of AECOPD hospitalisations, but were not investigated in the frame of the current study. One could speculate on different reasons for the observed higher baseline incidence of AECOPD hospitalisation incidence in GR as compared with the rest of CH. Possibly climate could play a role with lower average temperatures and longer and colder winters than the rest of CH due to the higher average altitude in this canton. This could lead to a higher indoor smoking activity and a higher incidence of respiratory infections as trigger factors for AECOPD hospitalisation.
Our results are in line with other studies supporting the hypothesis that policies aiming at the reduction of SHS can affect hospitalisation rates due to AECOPD.20 Although the main goal of smoking bans is to protect non-smokers from exposure to SHS and consecutive health deterioration, behavioural effects on smokers have also been shown. A Dutch study evaluated these effects on smokers and could detect the development of a higher degree of self-awareness and feelings of increased social exclusion after the implementation of a smoke-free legislation. The authors describe different types of reactions, including subsets of smokers actively trying to quit or to reduce their smoking consumption.21 A recent European study showed that national smoke-free legislations were associated with a higher proportion of smokers introducing smoking bans in their homes.22
We hypothesise that it probably takes time, possibly even more than 2 years, before the full effect of smoking bans can be seen. We understand this as a learning process, in which the adherence to the given legislation and the associated behavioural changes increase over time.
The fact that the hospitalisation incidences due to asthma and acute bronchitis did not drop after introduction of the smoking ban could possibly be due to the fact that these conditions are mostly and efficiently treated in an ambulatory setting and therefore were not available for this analysis.
Regarding community-acquired pneumonia (CAP), one could speculate on different reasons, including ageing population, increasing use of immunosuppressive therapies or increasing numbers of concomitant chronic illnesses, which potentially could mask a beneficial effect of the smoking ban.
A limitation of our study includes the fact that only hospitalisation cases could be identified. Therefore, we cannot exclude some kind of patient ‘shift’ from hospital to general practitioner. But this seems rather unlikely, because in general, many patients with AECOPD are primarily treated as outpatients and are hospitalised only in severe cases when this is necessary. This strategy is unlikely to have changed after introduction of the smoking ban. However, the Swiss coding system provides a nationwide coverage and high data completeness enabling valid comparisons across regions.23 Further, possible cases of mistakes in ICD-coding by the treating physicians cannot be excluded. Smoking status was unfortunately not available in our database, so that it was not possible to discriminate between current smokers, ex-smokers and never-smokers.
In conclusion, our study supports the hitherto very limited body of evidence demonstrating that legislated bans on smoking are associated with reduced rates of admission to hospital for respiratory conditions, hereby shown for acute exacerbated COPD, in addition to the meanwhile well-documented impact on cardiovascular disease. Our study represents, thus, a further documentation of the overall general positive impact of a smoking ban in public places on public health. This growing evidence of beneficial effects on public health should legitimise and encourage extending efforts to further reduce exposure to environmental smoke.
Prospective studies taking smoking status into account, which was not available in our database, could be of interest to further assess the relevance of SHS for the development of COPD in non-smokers.
What this article adds
Only a few studies have examined the effect of public smoking bans on respiratory conditions and showed reduced admission rates for different respiratory diseases.
Our study supports this and shows an association of the introduction of a public smoking ban and reduced rates of admission to hospital for acute exacerbated chronic obstructive pulmonary disease.
We thank the Swiss Federal Office for Statistics, Neuchâtel, Switzerland, for providing the data set of hospitalisations in Switzerland.
FB and MHB share the same working place address.
Contributors FD drafted the manuscript. Statistical analysis was conducted by FB and FD. All three authors developed the study question and critically revised the manuscript.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.