Our study is the first clinical cohort establishing the importance of cigarette smoking as a determinant of disease severity and control in allergic subjects who developed new onset asthma. Smoking status and smoking duration were markedly related in a dose-dependent fashion to the level of asthma severity and to poor asthma control. The demonstration of strong association and clear-cut dose-response relationship of smoking with asthma severity and control is in support of causality.
Previous surveys have used cross-sectional and case-control design, employed ill-defined asthma severity criteria, and mostly relied on questionnaires for the documentation of asthma symptoms and smoking status [4
]. Relying on questionnaires for the documentation of asthma symptoms may be unsatisfactory and cumulative exposure of tobacco measured by pack-years is more important than plain smoking status. Moreover, the possibility that treatment modalities (especially regular topical corticosteroids) might have altered the severity of the disease in the previous studies cannot be excluded. Lastly, our well characterized clinic cohort of allergic subjects at high risk for incident asthma represents an exclusive experimental model in which the effect of a common environmental risk factor (i.e. cigarette smoking) can be studied in relation to the progression of the natural history of the disease and circumvents the methodological limitations of previous surveys that have used cross-sectional and case-control design.
Using the GINA severity classification, data from our analyses show that smoking status and smoking duration are markedly related in a dose-dependent fashion to more severe asthma. The strongest association with more severe disease being observed in those who smoked more than 20 pack-years. Our findings largely agree with what has been illustrated in previous surveys [4
], but here we show for the first time that disease severity is associated with increased pack-years in a dose-response relationship.
Our cohort study of non-asthmatic adults with allergic rhinitis and followed up for 10 years shows that smoking can predict not just asthma incidence [13
], but also severity and level of control of the disease (this paper). We do not know the exact mechanism by which asthmatics who smoke have a more severe form of the disease, but it is likely that the inherent biologic intensity of the asthmatic inflammatory process is amplified by active smoking. Cigarette smoking may induce a neutrophil-predominant inflammation of the airways [26
], which may render patients less responsive to asthma treatment [9
]. Moreover, persistent exposure to cigarette smoke not only enhances allergic Th2-driven inflammation [28
], but also Th1-mediated inflammatory responses [26
]. Given that a mixed Th1/Th2 inflammatory response is a key event in the process of developing a more severe asthma phenotype [15
], development of a more severe disease may be anticipated in those allergic individuals who smoke regularly.
Smoking status and smoking duration are also markedly related in a dose-dependent fashion to poor asthma control, the strongest association with poor controlled disease being observed in those who smoked more than 10 pack-years. This is in agreement with recent population-based surveys of smoking status in asthma from Switzerland (30), UK (31), France [32
] and United States [33
]. The reason for asthmatics who smoke to have uncontrolled disease is not clear, but behavioral factors such as non-adherence and poor inhaler technique may play a role [34
]. In particular, non-adherence with antiasthma medications is common in asthmatic patients who are smokers [36
]. Additionally, poor asthma control can be due to the reduced therapeutic response to inhaled and oral corticosteroids in asthmatics who smoke [9
]. Another potential reason for apparent poor asthma control among smokers is misdiagnosis of chronic obstructive pulmonary disease (COPD) as asthma. Research suggests that up to one third of smokers over the age of forty with an asthma diagnosis may in fact have COPD [37
]. Although we cannot rule out completely a diagnosis of concomitant COPD in those who smoked, misdiagnosis of COPD in our study is unlikely as a result of the relatively young age of the study population entry criteria who had to be between the ages of 18 and 40 years.
A possible limitation of our study includes relying on medical records for the selection of the study subjects. However, all these subjects were examined and carefully diagnosed and documented in the clinic by experienced allergy specialists. Pack-years is a crude estimate of the amount and duration of smoking exposure, but this is universally used to address duration of exposure to tobacco smoke [38
], and in the present study allows us to demonstrate clear dose-related associations with disease severity and control. Another possible weakness of our study includes relying on a relatively small sample size to run multiple linear regression analyses on four class of asthma severity and three category of disease control for current smokers, former smokers, and never smokers. We minimized this problem by combining together class severity (Step I with II, and Step III with IV) and category of control ("uncontrolled" with "partly controlled"). Lastly, incomplete assessment of other important factors may limit our ability to define the relative importance of the key determinants of asthma control, as we did not collect information on compliance, socio-economical status, and education.
Our study has the advantage of the rigorous clinical assessment of asthma symptoms, medication use and lung function during the follow up visit at the same clinic. This is a substantial advancement compared with previous work in which self-report documentation of asthma symptoms, lack of objective measures for the diagnosis of asthma, and the cross-sectional design represented a severe limitation. Also, the fact that the all subjects examined were atopic (mostly sensitized to Parietaria judaica- the most prevalent allergen in Sicily) contributed to an important reduction in confounding factors for asthma severity/control. Furthermore, the possibility that regular nasal corticosteroids might have influenced study outcomes was addressed by excluding subjects using nasal corticosteroids for more than 6 weeks/year. Lastly, by examining asthma at ages when chronic obstructive pulmonary disease (COPD) is not prevalent, we have minimized this important confounders of poor asthma control.
The negative impact of smoking on asthma severity and control appears to be at least partially reversible in our study, as patients who had quit smoking reported significantly less severity and better asthma control than current smokers. This may have profound implications for clinical practice. If a modifiable determinant of asthma severity and poor control such as smoking can be easily identified in routine practice, it should be addressed in order to reduce asthma severity and to improve asthma control. Indeed, smoking cessation is associated with improvements in asthma symptoms, lung function quality of life scores, and BHR [39
] and most recent GINA guidelines recommend that smoking cessation should be an integral part of asthma treatment strategy.