Different effects of smoking on atopic and non‐atopic adult‐onset asthma

Abstract Background Both tobacco smoking and atopy increase the risk of adult‐onset asthma. We studied if there are differences in the effects of smoking on the risks of atopic and non‐atopic adult‐onset asthma, and if gender modifies these effects. Methods The Finnish Environment and Asthma Study (FEAS) includes 521 incident cases of adult‐onset asthma and 932 population‐based controls, aged 21 to 63 years, recruited from a geographically defined area of Pirkanmaa, South Finland. Asthma was defined based on symptoms and lung function measurements, atopy by IgE antibodies to common aeroallergens and smoking by the study questionnaire. Results Altogether 212 cases were atopic, and 251 cases were non‐atopic. Regular smoking increased the risk of atopic asthma (adjusted OR 1.24, 95% CI 0.83–1.85), this effect was seen in women (aOR 1.77, 1.06–2.95) but not in men (aOR 0.75, 0.39–1.45). Among regular smokers, the amount smoked was lowest among women with atopic asthma. Recent quitting of smoking was related to increased risk of both atopic (aOR 4.91, 2.26–10.65) and non‐atopic (aOR 4.37, 1.87–10.21) asthma. Having quitted smoking over a year ago was related to increased risk of non‐atopic asthma (aOR 1.57, 1.08–2.28), mainly in men (aOR 2.03, 1.06–3.88). Conclusions In women, rather small amounts of regular smoking increase the risk of atopic asthma. However, for non‐atopic asthma, the smoking induced risk continues for longer after quitting, especially in men. In conclusion, the effects of smoking on the risks of atopic and non‐atopic asthma differ, and gender modifies these effects.

associated with increased airway hyperresponsiveness (AHR), severe asthma symptoms and decline in lung function. 2,4 We have previously reported from the Finnish Environment and Asthma Study (FEAS) that both current smoking (OR 1.33, 95% CI 1.00-1.77) and previous smoking (OR 1.49, 1.12-1.97) increase the risk of adult-onset asthma. 5 Several other studies have confirmed these findings.  Moreover, we observed in the FEAS that the risk of adult-onset asthma related to current or past smoking was about 2.4 times higher in women compared to never smoking men. 5 Thus, our previous results suggested that women are more susceptible to the harmful effects of smoking. Findings compatible with ours have been reported by others. 9,14,16,19,22 In addition, we have reported from the FEAS 28 along with other studies 6,9,10,14,17,18,21,24,27,[29][30][31][32] that atopy and allergic diseases are risk factors for adult-onset asthma. However, there are only a few previous studies that have investigated potentially different effects of smoking on the risk of atopic and non-atopic adult-onset asthma. 10,14,17,32 The results of these studies are somewhat inconsistent, and they did not address potential modifying effect by gender. Our objective was to elaborate potentially different effects of personal smoking on the risk of atopic and non-atopic adult-onset asthma subtypes, and to evaluate if such relations are modified by gender. Further information on the role of atopy and gender in smoking and asthma would support personalised approach to prevent taking up smoking and to motivate quitting.

| Study subjects
The detailed description of the recruitment of cases and controls has been published elsewhere. 5, 28 We systematically recruited all new cases of adult-onset asthma 21-63 years old in 1997-2000 in the Pirkanmaa Hospital District, South Finland. We selected as cases only those with no previously diagnosed asthma or previous longterm use of any asthma medication. A total of 521 cases participated (response rate 86%). The control subjects were randomly drawn from the source population. A total of 1016 control subjects participated (response rate 80%). After excluding those with previous or current asthma, those older than 63 years, and those returning incomplete questionnaire, our study population included 932 controls. The study was approved by the ethics committees of the Finnish Institute of Occupational Health and the Tampere University Hospital.

| Study design
This was a population-based incident case-control study of adult-onset asthma. 5,28 The source population consisted of adults 21-63 years of age living in a geographically defined area of Pirkanmaa in South Finland. The population of the study area was 447,051 in 2000.

| Questionnaire
When entering the study, the subjects answered a self-administered questionnaire enquiring about their personal characteristics, health information, active smoking, second-hand smoke exposure, occupation, work environment, home environment, and diet. 5,28 Smoking was categorised in the following way: regular, ≥1 cigarette or cigar a day or ≥25 g of pipe tobacco a month; occasional, <1 cigarette a day or <25 g of pipe tobacco a month; current: regular and occasional smokers; previous: quit smoking either <12 months ago (=recent quitting) or >1 year ago. The cumulative life-time consumption of tobacco was estimated as cigarette-years, calculated as the average smoking rate (cigarettes/cigars/pipefuls per day) � duration of smoking.

| Diagnosis of asthma, lung function measurements, and definition of atopy
The details of diagnosis of asthma and lung function measurements 5,28 as well as definition of atopy 28 have been published previously. Shortly, the diagnostic criteria for asthma were (1) occurrence of at least one asthma-related symptom, and (2) demonstration of airways obstruction with significant reversibility in lung function investigations. Spirometry and bronchodilatation test were recorded with a pneumotachograph spirometer using disposable flow transducer (Medikro 905, Medikro) according to the standards of the American Thoracic Society. 33 Presence of obstruction was defined based on the reference values derived from the Finnish population. 34 For this study, we excluded 26 individuals who were found to have asthma-COPD overlap syndrome (ACOS). 35 Atopy was defined based on presence of specific IgE antibodies to common aeroallergens, including birch, timothy grass, mugwort, cat, dog, horse, Dermatophagoides pteronyssinus, and/or Aspergillus fumigatus in serum. 28 The results were expressed as Phadiatop scores from 0 to 6.
A positive score (≥1) corresponded to the limit of ≥0.35 kU/L and was used to define atopy. The Phadiatop measurements were available for 463 cases, which is the final number of cases included in this substudy. There were no significant differences in the distribution of the main characteristics between the 463 cases included in this substudy and the 32 cases who did not give a blood sample and thus have no Phadiatop measurement (Table S1).

| Statistical analysis
We analysed potential effect of current and former smoking on the risk of atopic and non-atopic adult-onset asthma, with never smokers forming the reference category. Next, we considered the regularity of current smoking (occasional or regular) and among ex-smokers, the time since quitting smoking (i.e. ≤12 months ago or >1 year ago). We adjusted the analyses for gender, age, education (as an indicator of socioeconomic status), pets indoors, work exposures except for moulds, and mould exposure at work or at home as potential confounders. In addition, we studied if gender modifies these relations by performing gender-stratified analyses. We applied multinomial logistic regression analysis for all models. We also compared the years smoked, daily smoking rate, duration since smoking cessation, and cigarette-years across the categories of asthma subtypes and by gender applying Kruskal-Wallis test with Dwass, Steel, Critchlow-Fligner multiple comparisons post-hoc procedure. We used SAS 9.4 for all statistical analyses (SAS Institute Inc.).

| Characteristics
Those with atopic asthma reported more often pets indoors now or previously than those with non-atopic asthma or the controls (Table 1). In contrast, those with non-atopic asthma were more often women, were older and had a lower level of education than those with atopic asthma or the controls.

Characteristic
Controls, n (%) Non-atopic asthma, n (%) Atopic asthma, n (%)  3.2 | The effects of smoking on the risk of asthma subtypes

| Current smoking
Current smoking was found to have little effect on the risk of either asthma subtype in the whole study population, and in men ( Figure 1A,B, Tables 2 and 3). In women, the risk was somewhat increased for both atopic (OR 1.48, 0.92-2.38) and non-atopic (OR 1.44, 95% CI 0.92-2.27) asthma, but these were not statistically significant (NS) ( Figure 1C, Table 3).
When we considered the regularity of smoking, occasional smoking had no effect on atopic asthma but increased the risk of non-atopic asthma in the whole study population (OR 1.90, 1.05-3.43) and especially in women (OR 2.32, 1.14-4.71) ( Figure 1A-C, Tables 2 and   3). Regular smoking, on the other hand, increased the risk of atopic asthma in women (OR 1.77, 1.06-2.95). For non-atopic asthma in women, the effect estimate of regular smoking was somewhat elevated, but NS (OR 1.20, 0.71-2.03).

| Former smoking
In comparison to never smoking former smoking associated with somewhat elevated (OR 1.22, 0.82-1.82), but NS, risk of atopic asthma. Whereas the risk of non-atopic asthma was increased (OR 1.73, 95% CI 1.21-2.48) ( Figure 1A, Table 2). Analysing the genders separately did not reveal many differences ( Figure 1B,C, for non-atopic asthma ( Figure 1A, Table 2). Stratification by gender did not influence the results much, except that the risk of non-atopic asthma in men was NS ( Figure 1B,C, Table 3).
Having quitted smoking over a year ago did not affect the risk of atopic asthma neither in the whole study population (OR 0.89, 0.56-1.41), nor in the analyses by gender ( Figure 1A-C, Tables 2 and 3). In contrast, the risk of non-atopic asthma was increased in the whole study population (OR 1.57, 1.08-2.28), and in men (OR 2.03, 1.06-3.88) ( Figure 1A-C, Tables 2 and 3).

| Smoking history
We next analysed if there are differences in smoking habits that could explain the differences observed in the risk of the asthma subtypes (Table 4). Among those, who had quitted smoking over a year ago, men with non-atopic asthma had generally smoked more, that is, had higher median smoking rate and cigarette-years, than women with non-atopic or atopic asthma. Among recent quitters and occasional smokers, there were no statistically significant differences in smoking variables according to asthma subtype or gender. Among regular smokers, men with non-atopic asthma had a longer history of smoking (in years smoked) and had smoked more (in terms of both smoking rate and cigarette-years) than women with atopic asthma. In addition, women with non-atopic asthma had smoked more (in cigarette-years) than women with atopic asthma.

| DISCUSSION
In this population-based study we show evidence that smoking af-  Smoking missing for one control and one case with non-atopic asthma in men, and for one control and two cases with non-atopic asthma in women. b Adjusted for age, education, pets at home, work exposures and indoor moulds.

| Current smoking
In our pervious study, the risk of adult-onset asthma due to occasional smoking was elevated (OR 1.25, 95% CI 0.76-2.06), but NS. 5 Our present, more detailed analysis revealed that occasional smoking associated with an increased risk of non-atopic asthma, but not with atopic asthma. This effect was stronger in women. On the other hand, regular smoking associated with an increase in the risk of atopic asthma, and only in women. However, in women there was also some tendency for increased risk of non-atopic asthma. It is worth noticing that the regular smoking atopic asthmatics, and especially women, had smoked relatively small quantities when compared to the nonatopic regular smokers. Thus, our findings imply that women may develop atopic asthma even from relatively small quantities of regular smoking. In line with our finding, an interaction between pets and smoking on asthma was reported for Canadian women, but not for men. 16 On the other hand, in relation to occasional smoking the risk of non-atopic asthma increased, while no such increase was detected for atopic asthma. This could be due to an individual level interplay between the strength of nicotine dependency and the severity of smoking caused symptoms. Perhaps among the most sensitive individuals, only those with strong addiction keep on smoking, while others prefer to quit smoking. In contrast, those who have less severe symptoms may prefer occasional smoking over quitting. However, the number of occasional smokers in our study was small, and thus, the interpretation of those results needs to be done with some caution. In men, current smoking had no effect on the risk of either asthma subtype. This may be explained by the 'healthy smoker'-concept, 36 according to which only those, who do not develop smoking related symptoms remain smokers.
Further, in women ovarian hormone fluctuations may make total quitting of smoking more difficult. 37

| Previous smoking
The risk of new asthma in our study was the highest among those, who had quitted smoking recently, that is, during the previous 12 months. This effect was not markedly different between atopic and non-atopic asthma or dependent on gender. In line with this, we have previously demonstrated that recent quitting of smoking associates with decreased lung function among those with new asthma in the FEAS. 38 The most likely explanation for the increased risk of asthma in this smoking category is that people who start to experience asthma symptoms tend to quit smoking, but this may happen too late in relation to development of asthma. In line with our findings, other studies have also reported higher risk of asthma among recent quitters, when compared to never smokers. 7 After longer smoking cessation periods, our results show that the beneficial effect of quitting smoking begins to emerge, especially in relation to atopic asthma. However, among non-atopic asthmatics we observed that the risk remained somewhat elevated and was more pronounced in men. This somewhat unexpected finding of higher risk in men could be explained by higher amount of tobacco smoked among the non-atopic men.

| Possible mechanisms
Among those who develop non-atopic asthma, smoking may direct the immune responses towards T2-low type asthma that is considered to be more innate immunity mediated, is characterised by normal levels of eosinophils, low fractional exhaled nitric oxide (FeNO), and can be neutrophilic or pauci-granulocytic, 3,4 32 In addition, one study reported increased, but NS, risk among both non-atopic with OR 1.5 (0.9-2.5) and atopic with OR 1.9 (0.8-4.2) participants. 17 In summary, there was no consistency among the previous studies concerning the effect of smoking on atopic and non-atopic adult-onset asthma. Comparison of the results is difficult due to different definitions of exposures and outcomes being applied. Different smoking habits between the countries, for example, more common use of snus and less common cigarette smoking among Swedish versus Finnish men, 40 could play some role for the observed differences. To our knowledge, our study is the first one that examines the effects of occasional and regular smoking and the effect of quitting time on subtypes of adult-onset asthma, and studies genders separately. These specific details were found to be of importance, since if we had not considered them, we would have only found an increased risk of non-atopic asthma in former smokers.

| CONCLUSIONS
This study showed that tobacco smoking causes atopic and nonatopic adult-onset asthma with different dynamics. We show for the first time that gender modifies these effects. In women, relatively small amounts of regular smoking increase the risk of atopic asthma, and quitting reduces the asthma risk relatively fast. The effect of regular smoking on non-atopic asthma risk is weaker, but this effect continues for longer after quitting, especially in men. This raises new questions about the underlying mechanisms of adult-onset asthma and supports personalised approach in preventing smoking related damages.