To the Editor:

Cigarette smoking decreases corticosteroid sensitivity in patients with asthma and worsens their symptoms and exacerbation frequency (1, 2). Inhaled corticosteroids (ICS) are the cornerstone of asthma treatment. However, the optimal therapies for smokers with asthma are not well defined because smokers (and ex-smokers with a >10 pack-year history) are usually excluded from clinical trials (3, 4).

Exposure to even low levels of cigarette smoke is known to induce small airway inflammation (5, 6), which is associated with worse asthma control (5, 7). Therefore, ICS deposition in small airways, which increases with small- versus standard-size–particle ICS (8, 9), might be an important determinant of ICS effectiveness in smokers with asthma.

To explore this hypothesis, a historical matched cohort study was designed to compare the effectiveness of representative small-particle and standard-size–particle ICS for current and ex-smokers with asthma and to investigate any differential effect compared with effects in nonsmokers. Some of the results of this study have been previously reported in the form of abstracts (10, 11).

We examined anonymized medical record data for patients with asthma in two large UK electronic data sets, the General Practice Research Database (now in the Clinical Practice Research Datalink) and the Optimum Patient Care Research Database (12, 13). We studied patients aged 30–70 years when they were prescribed a step-up in asthma therapy of 50% or more increase in ICS dose, using hydrofluoroalkane beclomethasone dipropionate (Qvar; Teva Respiratory, LLC, Sellersville, PA; particles of median mass aerodynamic diameter, 1.1 μm) or fluticasone propionate (Flixotide; GlaxoSmithKline Australia Pty Ltd, Melbourne, Australia; median mass aerodynamic diameter, 2.4–3.2 μm) by pressurized metered-dose inhaler (14). Eligible patients had 2 consecutive years of data within the study period (January 1999–2010), including 1 baseline year before and 1 outcome year after the step-up index date, and remained at the stepped-up ICS dose during the outcome year, with at least one additional asthma prescription. Exclusion criteria were any chronic respiratory disease other than asthma, maintenance oral corticosteroid therapy during baseline, multiple ICS prescriptions on the index date, and baseline or index date prescription of fixed-dose combination ICS plus long-acting β-agonist. Moreover, ex-smoking status had to be recorded at age 30 years or older to ensure long-term smoking. Patients stepped-up to small-particle ICS from 2005 onward were matched 1:1 to patients prescribed standard-size–particle ICS from 2000 onward to maximize patient numbers.

The primary effectiveness measure in this study was the rate of severe exacerbations, defined according to expert working group criteria (15) as an asthma-related emergency hospital attendance/admission or an oral corticosteroid pulse. Other effectiveness measures were the rate of acute respiratory events, defined as a severe exacerbation or general practice consultation for lower respiratory tract infection, and risk-domain asthma control, defined as the absence of any acute respiratory event.

Patients in small-particle and standard-size–particle ICS cohorts were matched sequentially in a 1:1 ratio for demographic characteristics and baseline markers of asthma severity, including sex, age (±5 yr within subgroups of ≥30–60 and >60–70 yr), smoking status, and baseline mean daily ICS dose, mean daily short-acting β-agonist dose, number of severe exacerbations, number of asthma consultations without oral corticosteroid prescription (i.e., without exacerbation), asthma control status, and index date year (±5 yr).

Outcome year exacerbation and acute respiratory event rates were compared using conditional Poisson regression models. The adjusted odds of achieving risk-domain asthma control were compared using a conditional binary logistic regression model. Potential confounding variables considered were those that differed between cohorts at baseline (P < 0.10) and those that were predictive (P < 0.05) of each outcome variable in multivariable analyses. All models included treatment, smoking status (current/ex-smoker vs. nonsmoker), and interaction between treatment and smoking status to determine any difference in treatment effect by smoking status. A P value less than 0.10 was considered significant for the exploratory interaction analysis. Analyses were performed using SPSS Statistics version 19 (SPSS Statistics, IBM, Somers, NY), SAS version 9.3 (SAS Institute, Cary, NC), and Microsoft Excel 2007 (Microsoft, Bellevue, WA).

After matching, there were 889 patients in each cohort, including 314 (35%) current or ex-smokers (Table 1). Nonsmokers in the standard-size–particle ICS cohort were more likely to have concomitant rhinitis or receive rhinitis therapy than nonsmokers in the small-particle ICS cohort. Patients in the standard-size–particle ICS cohorts were more likely to have been prescribed a long-acting β-agonist during the baseline year. Other differences between cohorts after matching were not clinically important.

On the index date, small-particle ICS was prescribed at significantly lower doses than standard-size–particle ICS (median [interquartile range (IQR)], 400 [200–400] vs. 500 [500–1,000] μg/d; P < 0.001). During the outcome year, ICS dose exposure was also significantly lower in the small-particle ICS cohort (median [IQR], 301 [164–438] μg/d vs. 397 [238–658] μg/d; P < 0.001). The short-acting β-agonist dose was similar in the two cohorts (median [IQR], 219 [110–438] μg/d).

Adjusted rates of severe exacerbations during outcome were significantly lower for current/ex-smokers and nonsmokers in the small-particle than in the standard-size–particle ICS cohort, with no interaction between treatment and smoking status (P = 0.31; Figure 1). In contrast, for current and ex-smokers, adjusted rates of acute respiratory events were significantly lower, and adjusted odds of risk-domain asthma control significantly higher, with small-particle than with standard-size–particle ICS; for nonsmokers, no difference was found. The interaction between treatment and smoking status was significant at the 10% level for both of these outcomes (Figure 1).

Figure 1. Adjusted outcome results for current/ex-smokers and nonsmokers prescribed small-particle ICS or standard-size–particle ICS, matched cohorts. P values are given for the interaction analysis. Asterisk indicates rate of severe exacerbations, adjusted for cardiac disease diagnosis, number of primary care consultations, and adherence to ICS therapy during baseline. Dagger indicates rate of acute respiratory events, adjusted for cardiac disease diagnosis, rhinitis diagnosis and/or therapy, number of primary care consultations, number of prescriptions for SABA, and LRTI consultations resulting in a prescription for antibiotics. Double dagger indicates odds of risk-domain asthma control, adjusted for rhinitis diagnosis and/or therapy, number of primary care consultations, adherence to ICS therapy, and LRTI consultations resulting in a prescription for antibiotics. CI = confidence interval; ICS = inhaled corticosteroid; LRTI = lower respiratory tract infection; SABA = short-acting β-agonist; standard SP = standard-size particle.

Download Figure | Download Powerpoint

[More] [Minimize]

When we examined results separately for smokers and ex-smokers, the same trends were observed for each group (as compared with nonsmokers): for risk-domain asthma control, the adjusted odds ratio (95% confidence interval) for small-particle ICS as compared with standard-size–particle ICS was 1.85 (0.96–3.58) for smokers, 1.88 (1.07–3.28) for ex-smokers, and 1.17 (0.86–1.60) for nonsmokers. Others have reported similarities in inflammation and treatment response between smokers and ex-smokers (16).

Another explanation for the similar findings for smokers and ex-smokers in this study could be the so-called “healthy smoker effect,” whereby patients with fewer respiratory symptoms tend to keep smoking and those with more symptoms tend to quit (17). Finally, an interaction between effects of age and smoking can be hypothesized because ex-smokers were older on average (mean [SD] age, 53 [10.2] yr) than smokers (43 [9.0]) and nonsmokers (50 [10.8]), and thus may have had more pack-years’ exposure than the current smokers.

These results confirm previous studies showing that in asthma, small-particle ICSs are at least as effective as standard-size–particle ICS despite being administered at significantly lower doses (18, 19). They also suggest a possible differential treatment effect with regard to smoking status and particle size of ICS, with a clear trend for greater beneficial effects of small-particle ICS for current/ex-smokers in terms of acute respiratory events and asthma control. This finding could be explained by increased small airway disease induced by smoking and improved targeting of small airways by smaller ICS particles. Thus, small-particle ICS may be a preferred treatment choice for patients who are current or ex-smokers.

In any observational study, confounding (although limited by the matching process) and partial characterization of patients, most of whom lacked recorded lung function, must be considered. Another study limitation is the lack of detailed information on smoking history, which prevented us from examining the influence of smoke-free duration for ex-smokers and of pack-years for smokers and ex-smokers. As a result, the findings of this study should be considered exploratory. Nonetheless, we believe further investigation of a possible differential effect is warranted.

References

Section:

ChooseTop of pageReferences <<CITING ARTICLES