Asthma-COPD Overlap Phenotypes and Smoking :Comparative features of asthma in smoking or non-smoking patients with an incomplete reversibility of airway obstruction

ABSTRACT

The development of COPD features, such as an incomplete reversibility of airway obstruction (IRAO), in smoking or non-smoking asthmatic patients, a condition often named Asthma-COPD Overlap (ACO), has been recognized for decades. However, there is a need to know more about the sub-phenotypes of this condition according to smoking.

This study aimed at comparing the clinical, physiological and inflammatory features of smoking and non-smoking asthmatic patients exhibiting IRAO.

In this cross-sectional study, patients with an IRAO with (ACO, ≥20 pack-years) or without (NS-IRAO, <5 pack-years) significant smoking history completed questionnaires about asthma control (ACQ, score 0–6, 6 = better score) and quality of life (AQLQ, score 1–7, 1 = better score) and performed expiratory flows, lung volume and carbon monoxide diffusion capacity measurements. Blood sampling and induced sputum were obtained for systemic and lower airway inflammation assessment.

A total of 115 asthmatic patients were included (75 ACO: age 61 ± 10 years, 60% women and 40 NS-IRAO: age 64 ± 9 years, 38% women). ACO patients had worse asthma control scores (1.8 ± 0.9 vs 1.4 ± 0.9, P = 0.02) and poorer asthma quality of life (5.3 ± 1.0 vs 5.9 ± 1.0, P = 0.003). In addition, ACO had higher residual volume (145 ± 45 vs 121 ± 29% predicted, P = 0.008) and a lower carbon monoxide diffusing capacity corrected for alveolar volume (90 ± 22 vs 108 ± 20% predicted, P = 0.0008). No significant differences were observed in systemic or lower airway inflammation.

In conclusion, in smokers and non-smokers, the presence of IRAO in asthmatics is associated with different phenotypes that reflect the addition of smoking-induced changes to asthma physiopathology.

KEYWORDS:

• Irreversible airway obstruction
• airway diseases
• lung function
• asthma control
• inflammation

ABBREVIATIONS LIST
ACO	=	Asthma-COPD overlap
ACQ	=	Asthma control questionnaire
ACSS	=	Asthma control scoring system
AQLQ	=	Asthma quality of life questionnaire
ATS	=	American thoracic society
BMI	=	Body mass index
CO	=	Carbon monoxide
COPD	=	Chronic obstructive pulmonary disease
CRP	=	C-reactive protein
ERS	=	European respiratory society
FeNO	=	Fractional exhaled nitric oxide
FEV1	=	Forced expiratory volume in one second
FVC	=	Forced vital capacity
GLI	=	Global lung initiative
ICS	=	Inhaled corticosteroids
LABA	=	Long-acting beta-agonist
LAMA	=	Long-acting muscarinic-antagonist
NS-IRAO	=	Non-smoking patients with incomplete reversibility of airway obstruction
QoL	=	Quality of life

Introduction

Asthma is a chronic condition of the airways characterized by recurrent symptoms of wheezing, breathlessness, chest tightness, and cough that are associated with variable airway obstruction (1). In most asthmatic patients, airway obstruction can resolve spontaneously or following treatment. However, in a subset of people with asthma, whatever they smoke or not, an incomplete reversibility of airway obstruction (IRAO) can be observed despite optimal corticosteroid treatment, a condition often called Asthma-Chronic Obstructive Pulmonary Disease (COPD) Overlap (ACO) (2–5).

Many factors can contribute to the development of IRAO although smoking history (6,7) is considered the most important one (8–11). Smoking influences asthma by changing the airway inflammatory phenotype to a more neutrophilic type, makes asthma more difficult to control, may impair perception of bronchoconstriction and is associated with a more rapid decline in lung function (12–15). Accordingly, we previously showed that smoking is associated with early signs of COPD in asthmatic people (16,17).

Asthma and smoking-induced COPD share some clinical features in adult patients, both being characterized by an airway inflammatory process and significant airway structural changes, although of a different nature and magnitude (18,19). Nevertheless, most studies on asthma have usually excluded smoking patients or patients with a significant smoking history and studies on COPD have generally excluded patients with asthma or asthma history. Studying such population can help identify mechanistic pathways leading to the development of ACO.

Although there is growing evidence that ACO represents a distinct clinical entity (20), the use of such label has been recently debated (21–23), mainly because there is no consensus about the definition of this condition (24). As a result, different criteria have been used to define ACO patients in the various studies, making comparison of study results difficult. Nevertheless, it becomes increasingly evident that this condition includes many phenotypes. However, these patients are frequently studied as a global population and sub-phenotypes need to be better defined. Whether they are called asthma with IRAO or ACO, to our knowledge, studies are needed to assess the impact of smoking history on the features of this condition in comparison to non-smokers. Smoking is considered as a mandatory criterion by some authors although not by others. This study aimed to compare the clinical, physiological and inflammatory characteristics of patients with IRAO associated or not with smoking.

Methods

Participants

Two groups of asthmatic patients who developed IRAO (see inclusion criteria below) were recruited between April 2014 and December 2016 from the asthma clinic of the Quebec Heart and Lung Institute-Laval University, a tertiary care center of the Quebec metropolitan area (Canada): 1) With a significant smoking history (ACO) and 2) without a significant smoking history (NS-IRAO). People in the ACO group were defined as current or ex-smokers with a ≥20 pack-years history of tobacco smoking. People in the NS-IRAO group were defined as never-smokers or ex-smokers with <5 pack-years smoking history, who had completely quit smoking ≥12 months before inclusion in the study.

To be included in the study, patients had to: a) be aged ≥40 years, b) have a previous diagnosis of asthma, characterised by a positive response to bronchodilator, as shown by a >200 mL and >12% increase from baseline forced expiratory volume in one second (FEV₁) and/or a positive methacholine bronchoprovocation test (<16 mg/mL), associated with history of respiratory symptoms (1), c) require treatment with inhaled corticosteroids (ICS) with or without additional asthma medication, and d) show IRAO, as defined by persistence of a post-bronchodilator forced expiratory volume in one second/forced vital capacity (FEV₁/FVC) ratio <0.7 in addition to a FEV₁ less than 80% of predicted value on at least two occasions while on a treatment considered as optimal by a respirologist (1). Patients were excluded if they had: a) any other respiratory conditions than asthma (this included previous diagnosis of COPD without a confirmed diagnosis of asthma), b) unstable respiratory or non-respiratory condition, c) previous bronchial thermoplasty, d) evidences of respiratory infection in the four weeks preceding entry in the study, and e) changes in respiratory medications in the previous four weeks.

Severity of asthma was determined according to the medication prescribed to keep asthma under control, as defined by current guidelines (1). Asthma was considered as moderate if treated with moderate doses (250 mcg/day < dose ≤500 mcg/day beclomethasone HFA or equivalent) of ICS, with or without additional therapy (long-acting inhaled β2-agonist (LABA) or leukotriene receptor antagonist (LTRA)). Asthma was considered severe if the patients needed high doses (>500 mcg/day beclomethasone HFA or equivalent) of ICS and additional pharmacotherapy (LABA, LTRA, and/or oral corticosteroids).

Study design

This was a cross-sectional study evaluating: a) asthma control, b) asthma-related quality of life (QoL), c) prevalence of exacerbations/healthcare use, d) reversibility of airway obstruction, e) lung volumes, f) lower airway inflammation, g) systemic inflammation.

The study was approved by the IUCPQ-UL Ethics Committee (CÉR 21047) and all subjects gave their written informed consent

Evaluation

Data regarding age, sex, body mass index (BMI), smoking history, duration of asthma, family history of asthma and atopy, medication use, and severity of the disease were recorded. Atopic status was assessed with allergy skin-prick tests and serum IgE levels.

Asthma control was assessed with the validated French version of the Asthma Control Questionnaire (ACQ) (25). The Asthma Control Scoring System (ACSS), which evaluates eosinophilic inflammation in addition to clinical asthma control criteria was also used to determine asthma control (26,27). QoL was assessed by the French version of the Mini Asthma Quality of Life Questionnaire (Mini-AQLQ) (28).

Unscheduled medical and emergency department visits, and hospitalizations were documented for the previous year. Asthma exacerbations were defined according to the American Thoracic Society/European Respiratory Society (ATS/ERS) statement (29).

Baseline FEV₁ and FVC were measured according to the ATS criteria (30). Predicted values were obtained from the ERS Global Lung Function Initiative (GLI-2012) (31,32). Reversibility of airway obstruction was measured after administration of 200–400 mcg of inhaled salbutamol (30). Lung volumes and carbon monoxide diffusion capacity (D_LCO) were measured as per ATS/ERS standards (33,34). Before pulmonary function tests, short-acting β2-agonists (SABA) and short-acting muscarinic-antagonists (SAMA) were withheld for at least 8 hours, LABA were withheld for at least 12 hours and long-acting muscarinic-antagonists (LAMA) were withheld for at least 24 hours when applicable. Current smokers were asked to refrain from smoking for at least 12 hours before study visit.

Exhaled nitric oxide (FeNO) measurements were performed according to ATS recommendations using a NiOXMino handheld analyzer (35). Sputum was induced by inhalation of hypertonic saline and processed using the method described by Pin et al. (36) and modified by Pizzichini et al. (37). Differential cell count was obtained.

Peripheral blood was sampled for complete blood count and for the following biomarkers: C-Reactive Protein (CRP), fibrinogen, and alpha-1 antitrypsine.

Statistical analyses

Data are expressed using means±SD for continuous variables and proportions for categorical variables. Phenotypes were determined by comparing the various characteristics between ACO and NS-IRAO patients. For variables for which normality assumption was not fulfilled, analyses were performed first on an appropriate transformation (log, square root). When no transformation was able to satisfy the assumptions, the Wilcoxon rank-sum test was used to complete statistical analyses. These data are presented as medians (25–75 percentile). Categorical variables were analysed using Chi-Square or Fisher's exact test. Continuous variables were analysed using one-way ANOVA to compare groups with heterogeneous variances and were tested whether the model could be reduced to a one-way analysis with the same variance between groups. When effect that specifies heterogeneity in the covariance structure was significant (heteroscedasticity) compared to the same variance between groups, the statistical analyses were performed using separate residual variance per group. The Satterthwaite's degree of freedom statement was added for unequal variance structures. Because a “sex effect” was observed between groups, all statistical analyses were adjusted for differences in female sex proportion and data were analysed using a two-way ANOVA model. The normality assumption was verified with the Shapiro-Wilk test after a Cholesky factorisation on residuals from the statistical model. The Brown and Forsythe's variation of Levene's test statistic was used to verify the homogeneity of variances. Correlations between variables were expressed using the Spearman's correlation coefficients. The results were considered statistically significant with P-values <0.05. The term “significant” will be used to refer to “statistically significant”. All analyses were conducted using the statistical packages SAS, version 9.4 (SAS Institute Inc, Cary, NC, U.S.A.) and R (R Core Team (2016), Foundation for Statistical Computing, Vienna, Austria.).

Results

Medical charts from 245 consecutive asthmatic patients were screened of which 115 with IRAO were included: 75 in the ACO group and 40 in the NS-IRAO group (Figure 1).

Figure 1. Flowchart of study participants.

Table 1 shows subjects' demographics. ACO included significantly more women and had a shorter duration of asthma as compared to NS-IRAO. ACO included 60 ex-smokers and 15 current smokers. Asthma severity and daily dose of ICS were similar between groups, although there were more patients taking LAMA in the ACO group. ACO reported a higher number of oral corticosteroid short courses in the year preceding the study compared to NS-IRAO. No differences were observed between the two groups for the other exacerbations' outcomes.

Table 1. Subjects' demographics.

	ACO	NS-IRAO	P
Number of subjects	75	40
Women/Men	47/28	15/25	0.01
Age (years), mean ± SD (range)	61 ± 10 (44–84)	64 ± 9 (44–78)	0.141
BMI (kg/m^2)	29 ± 6	28 ± 5	0.438
Smoking history, n (%)			<0.001
Never smokers	0	28 (70)
Ex-smokers	60 (80)	12 (30)
Smokers	15 (20)	0
Pack-years	36 ± 13	3 ± 2	<0.001
Asthma duration (years)	23 ± 17	35 ± 21	0.006
Asthma severity			0.862
Mild	3 (4)	1 (3)
Moderate	39 (52)	19 (48)
Severe	33 (44)	20 (50)
Dose of ICS (mcg/day, equivalent beclomethasone)	715 ± 362	801 ± 424	0.537
ICS+LABA	67 (89)	35 (88)	0.806
LABA	5 (7)	5 (0.1)	0.867
LTRA	14 (19)	11 (28)	0.173
Anti-IgE	0	0	1.0
LAMA	39 (52)	6 (15)	<0.001
SAMA	3 (4)	0	0.867
SABA^*	15 (20)	5 (13)	0.005
Prednisone	0	0	1.0
Acute care use in the previous year
mean ± SD/patient (total number of events)
Unscheduled medical visits related to respiratory problems	0.4 ± 0.8 (29)	0.3 ± 0.5 (10)	0.519
Emergency room visits (<24 hours)	0.5 ± 1.4 (39)	0.2 ±0.4 (7)	0.226
Hospitalisations (>24 hours)	0.1 ± 0.4 (8)	0.0 ± 0.2 (1)	0.315
Health care use
mean ± SD/patient (total number of events)
Oral corticosteroids in the previous year	0.8 ± 1.2 (64)	0.4 ± 0.6 (15)	0.04
Atopy (yes)	53 (71)	33 (83)	0.454
Family history of asthma (yes)	39 (52)	22 (55)	0.885
Family history of allergy (yes)	29 (39)	22 (55)	0.181

Results are presented as mean ± SD for continuous variables or as number (%) of patients for categorical variables unless stated otherwise.

^* Patients not having a SABA as rescue medication were using their combined ICS-LABA as rescue medication.

ACO: asthma-COPD overlap, ACSS: asthma control scoring system, BMI: body mass index, ICS: inhaled corticosteroid, LABA: long-acting beta-agonist, LAMA: long-acting muscarinic antagonist, LTRA: leukotriene receptor antagonist, NS-IRAO: non-smoking patients with incomplete reversibility of airway obstruction, SABA: short-acting beta-agonist, SD: standard deviation.

The ACQ, ACSS and Mini-AQLQ data are shown in Table 2. Mean ACQ score was significantly higher (poorer asthma control) in ACO than in NS-IRAO. As well, ACO reported significantly more clinical asthma symptoms than NS-IRAO, although the physiological and inflammatory scores were similar between groups, resulting in comparable global ACSS scores. QoL was also lower in ACO, with Mini-AQLQ total score as well as symptoms and activity limitations sub-scores significantly lower than NS-IRAO.

Table 2. Comparative scores of asthma control and asthma quality of life questionnaires.

	ACO	NS-IRAO	P
Asthma Control Questionnaire
Total	1.8 ± 0.9	1.4 ± 0.9	0.02
Asthma Control Scoring System
Clinical	76 ± 17	86 ± 17	<0.001
Daytime symptoms	17 ± 6	20 ±5	0.02
Nighttime symptoms	23 ± 4	23 ± 4	0.446
Rescue medication need	20 ± 7	22 ± 5	0.126
Activity limitation	15 ± 7	20 ± 7	<0.001
Physiologic	33 ± 16	37 ± 19	0.443
Inflammatory	59 ± 26	53 ± 30	0.518
Total	56 ± 13	60 ± 14	0.103
Asthma Quality of Life Questionnaire
Symptoms	5.4 ± 1.1	5.9 ± 1.1	0.01
Activity limitations	4.9 ± 1.2	5.8 ± 1.2	0.001
Emotions	5.3 ± 1.6	6.0 ± 1.4	0.080
Environment	5.5 ± 1.2	5.9 ± 1.3	0.091
Total	5.3 ± 1.0	5.9 ± 1.0	0.003

Results are presented as mean ± SD.

ACO: asthma-COPD overlap, NS-IRAO: non-smoking patients with incomplete reversibility of airway obstruction

Expiratory flows and bronchodilator reversibility were similar between the two groups (Table 3). However, compared to NS-IRAO, ACO had a significantly higher residual volume and significantly lower baseline and volume adjusted D_LCO.

Table 3. Pulmonary function in ACO and NS-IRAO.

	ACO	NS-IRAO	P
Spirometry
FEV1 (% predicted)	56 ± 13	61 ± 11	0.151
FVC (% predicted)	80 ± 14	81 ± 12	0.643
FEV1/FVC	0.56 ± 0.08	0.58 ± 0.06	0.164
Reversibility to BD (%)	14 ± 11	12 ± 10	0.827
Lung volumes
TLC (% predicted)	107 ± 16	101 ± 11	0.081
FRC (% predicted)	112 ± 25	105 ± 20	0.187
RV (% predicted)	145 ± 45	121 ± 29	0.008
DLCO (% predicted)	75 ± 19	89 ± 22	0.01
KCO (% predicted)	90 ± 22	108 ± 20	<0.001

Results are presented as mean ± SD.

ACO: asthma-COPD overlap, BD: bronchodilator, D_LCO: single-breath diffusing capacity for carbon monoxide, FEV₁: forced expiratory volume in one second, FRC: functional residual capacity, FVC: forced vital capacity, KCO: diffusing capacity for carbon monoxide corrected for alveolar volume, RV: residual volume, TLC: total lung capacity, NS-IRAO: non-smoking patients with incomplete reversibility of airway obstruction.

No significant differences were observed in lower airway or systemic inflammatory cells and in markers of inflammation between ACO and NS-IRAO (Table 4). There was, however, a trend towards higher FeNO levels in never smokers, as expected. Distribution of sputum inflammatory phenotypes showed no differences between proportions of patients with eosinophilic, neutrophilic, paucygranulocytic and mixed phenotypes (data not shown).

Table 4. Lower airway and systemic inflammation in ACO and NS-IRAO.

	ACO	NS-IRAO	P
Lower airway
FeNO (ppb)	17.9 ± 12.7	24.5 ± 19.0	0.086
Induced sputum
N	47	23
Total cell count (×10^6/mg)	9.2 (7.0–15.7)	6.1 (3.7–12.3)	0.122
Eosinophils (%)	2.3 (0.5–7.3)	4.0 (0.8–23.3)	0.245
Neutrophils (%)	54.8 (37.6–75.9)	51.2 (28.5–75.6)	0.421
Systemic
White blood cell (×10^9/L)	7.9 ± 1.7	7.5 ± 1.4	0.214
Hemoglobin (g/L)	139 ± 12	142 ± 13	0.984
Eosinophils (×10^9/L)	0.23 ± 0.17	0.26 ± 0.23	0.824
Neutrophils (×10^9/L)	5.0 ± 1.5	5.0 ± 1.2	0.928
Fibrinogen (g/L)	3.5 ± 0.9	3.5 ± 0.8	0.454
Total IgE (IU/mL)	96 (27–254)	127 (63–295)	0.475
CRP (mg/mL)	3.7 ± 3.9	2.8 ± 3.0	0.251
Alpha-1 antitrypsin (g/L)	1.1 ± 0.2	1.1 ± 0.2	0.323

Results are presented as mean ± SD except for induced sputum counts and total IgE which are presented as median (25–75 percentile).

ACO: asthma-COPD overlap, CRP: C-reactive protein, FeNO: exhaled nitric oxide, g: gram, IgE: immunoglobulin E, IU: international unit, L: liter, mg: milligram, mL: milliliter, NS-IRAO: non-smoking patients with incomplete reversibility of airway obstruction, ppb: part per billion, WBC: white blood cells.

Correlations between clinical, physiological and inflammatory parameters

There were positive correlations between pre- and post-BD FEV₁ percentage predicted and ACSS total score as well as inverse correlations with ACQ total score in both groups (Figure 2). There were no statistically significant correlations between D_LCO and either score of asthma control (ACSS: ACO: r_s = 0.20, P = 0.09, NS-IRAO: r_s = −0.13, P = 0.42; ACQ: ACO: r_s = −0.16, P = 0.18, NS-IRAO: r_s = −0.006, P = 0.97). Sputum eosinophils and neutrophils did not correlate with either FEV₁ percentage predicted, FEV₁/FVC, functional residual capacity (FRC) or D_LCO (all P > 0.05).

Figure 2. Correlations between pre- and post-BD FEV₁ (% predicted) and ACSS and ACQ in smoking IRAO and non-smoking IRAO.

ACSS: asthma control scoring system, ACQ: asthma control score, BD: bronchodilator, FEV₁: forced expiratory volume in one second, NS-IRAO: non-smoking patients with incomplete reversibility of airway obstruction.

Discussion

To our knowledge, our study is original as it compares the multiple clinical, physiological and inflammatory characteristics of patients with IRAO, according to tobacco exposure. Although several studies have shown that smoking strongly influences the clinical presentation and prognosis of asthma (12,16,17,38), our results confirm these effects, especially in regard to asthma control (15,39), QoL (40) and lung function, in a sample of patients with chronic airflow limitation and significant smoking history. Hence, there were two main sub-phenotypes of IRAO according to smoking history, with marked differences in clinical and physiological features. Indeed, when IRAO was associated with significant current or past smoking history (referred to as ACO), patients reported worse asthma control, worse QoL and a higher number of short courses of systemic corticosteroids in the year preceding the study, in parallel with more gas trapping and lower diffusion capacity in comparison to non-smoker asthmatics with IRAO.

This study was planned on the premise that it would be important to differentiate smoking from non-smoking patients with asthma who developed IRAO. Asthmatics may develop an IRAO early in the course of their disease, possibly as an airway remodelling process representing a sequelae of respiratory infections or uncontrolled inflammation, particularly in those with low lung function (smaller lungs) at an early age or with long-standing disease (41,42). Consistent with this notion, asthma duration was significantly longer in non-smokers, suggesting long-term effects of asthma mechanisms on the development of fixed airway obstruction. In smokers, however, IRAO may develop more quickly. Smoking-related lung parenchymal changes, with their associated influence on lung volumes and loss in elastic recoil, may further increase airway obstruction from their additive effects to the asthma-related increased contractile properties of the airways (43).

In the present study, there was a statistically significant difference between ACO and NS-IRAO in asthma QoL and control as shown by AQLQ and ACQ scores. Interestingly, this difference was also clinically significant for the AQLQ and showed a trend towards clinical significance in regard to ACQ. In previous studies, asthma patients who smoked had a poorer asthma control (15,44) and QoL (40) compared with non-smokers. Nevertheless, our results provide additional information in showing that even when patients had quit smoking for a few years, asthma is still globally more difficult to control.

Although expiratory flows were similar in the two groups, there were more evidences of airway trapping and reduced CO diffusion capacity in ACO, suggesting an evolution towards an emphysematous phenotype. This is in keeping with the observations from a recent retrospective study by Kitaguchi et al. who reported lower lung volumes and CO diffusion in asthmatic patients with IRAO and a significant smoking history (also referred to as ACO) compared to those with IRAO only (45). In our study, this did not seem to be related to the poorer asthma control and QoL observed in ACO, as no statistically significant correlations were observed between D_LCO and scores of asthma control or QoL.

Lower airway or systemic inflammatory cells and markers of inflammation were similar between ACO and NS-IRAO. In previous studies, mixed results have been reported. Indeed, increased neutrophils (16,46,47) and, in some cases, increased eosinophils as well (46,47) have been observed in smoking asthmatics. However, others observed similar sputum eosinophils in both groups, but lower blood eosinophils in smoking asthmatics compared to non-smoking ones (48). Nevertheless, many studies were done in current smokers and, in some cases, patients were not prohibited from smoking on the day of attendance (47). Furthermore, the proportion of current and ex-smokers was not always reported (48). Our study brings new information on these features by suggesting that neutrophilia is probably mainly observed in current smokers and that this feature diminishes with smoking cessation. Nevertheless, we compared cell percentages and systemic markers' concentrations between current and ex-smokers in the ACO group and found no statistically significant difference between them (data not shown).

This study has some limitations but some are in fact hypothesis generating. First, the ACO group mostly included ex-smokers who had quit smoking for variable time periods. Previous studies showed that smoking cessation was associated with improvement of symptoms and lung function, better asthma control and a decrease in sputum neutrophils (49). We compared results between current and ex-smokers and found no statistically significant differences in these various outcomes (data not shown), although we cannot exclude a lack of power due to the relatively small number of current smokers. In a recent study, Tommola et al. (50) reported that in patients with adult-onset asthma re-evaluated 12 years after diagnosis, ACO patients had lower diffusing capacity, higher blood neutrophil levels, and higher serum IL-6 levels as compared to never- and ex-smokers with <10 pack-years, or non-obstructive patients with ≥10 pack-years smoking history, respectively. ACO patients also showed reduced lung function, higher bronchial obstruction reversibility and a higher number of co-morbidities. This is in keeping with our results showing that in a sample of patients with a significant smoking history, even when smoking has been stopped for many years, there are still negative impacts on asthma control, QoL and lung function.

Second, duration of asthma was not similar between groups, NS-IRAO having been asthmatic for a significantly longer time. This suggests that IRAO in non-smokers is probably due to the long-term effects of asthma while it develops more quickly in smokers. Despite this, ACO showed a significantly more severe profile, as evidenced by worse air trapping and gas transfer capacity of the lungs, when compared to their non-smoking counter parts.

Third, the study was cross-sectional, using convenience sampling. This may result in a selection bias and impact the generalizability of the results. Nevertheless, although larger prospective studies are needed to compare the long-term progression and future risk of these two groups of patients with IRAO, this study describes key elements of phenotype differences between IRAO with or without a significant smoking history.

Fourth, since some COPD patients may show reversibility to bronchodilator, one may question the accurate initial diagnosis of asthma in ACO. Nonetheless, we performed a thorough selection of patients and asthma diagnosis was made by an experienced respirologist and based on objective measures (1), which should considerably minimize the possibility of including some ACO with purely COPD.

Fifth, our definition of fixed airway obstruction using a 0.7 cut-off value for FEV₁/FVC instead of using FEV₁/FVC ratio of less than the 5th percentile (LLN) may be questioned. Indeed, there is disagreement regarding the preferred FEV₁/FVC criteria to define airway obstruction, considering not only the methodological and mathematical aspects of the classification but also its association with clinically relevant outcomes (51). Although we acknowledge that using a fixed FEV₁/FVC ratio to define airflow limitation may lead to patient misclassification, we opted for this approach as it is the most widespread criteria in clinical studies and the reference value of international guides on clinical strategies for the management of asthma (24) and COPD (52). In the present study, 10 patients in the ACO group and 9 patients in the NS-IRAO group had an FEV₁/FVC ratio <0.7 but >LLN). However, based on the finding that these patients presented similar characteristics than those who had both an FEV₁/FVC <0.7 and <LLN (data not shown), we believe that our findings are valid. Importantly, IRAO was not solely defined based on the FEV₁/FVC ratio since all study participants were required to have a clinical history compatible with asthma and an FEV₁ <80% predicted. In a CanCOLD substudy, van Dijk et al. reported that although the use of the fixed FEV₁/FVC ratio alone may lead to misclassification of airway obstruction, when an FEV₁ <80% predicted was also required to define airway obstruction (as performed in the present study) both criteria for FEV₁/FVC, the fixed ratio or the LLN, performed equally well to predict clinical outcomes (53).

Finally, we showed that there were more women than men in the ACO group. This may suggest more marked effects of smoking on airways of women, possibly related to the smaller initial caliber of airways in women. The fact that this was not so evident in non-smokers suggests a lesser effect of asthma mechanisms compared to smoking.

In conclusion, this study provides additional key-information on ACO sub-phenotypes. Although it is undeniable that some asthmatic patients will develop, over time, features of COPD associated or not to smoking or environmental exposures, the concept of ACO is currently debated. Some consider that it is a specific entity, while others believe that this only reflects a specific phenotype of either asthma or COPD (54,55). Nonetheless, it is increasingly apparent that ACO represents in fact a multitude of sub-phenotypes, with probably different treatment responses and clinical outcomes (48). The present study adds to the current discussion about ACO by showing that non-smoking and smoking asthmatic patients with IRAO represent two distinct entities that should be differentiated when looking at asthma-COPD overlap characteristics and outcomes.

Declaration of interest

LPB considers having no conflict of interest but wishes to declare what can be perceived as potential conflicts of interest. Advisory Boards: GlaxoSmithKline, Novartis. Conferences (honoraria): AstraZeneca, GlaxoSmithKline, Merck, Novartis. Sponsorship for investigator-generated research: AstraZeneca, GlaxoSmithKline, Merck Frosst, Schering. Sponsorship for research funding for participating in multicenter studies: AllerGen, Altair, Amgen, Asmacure, AstraZeneca, Boehringer-Ingelheim, Genentech, GlaxoSmithKline, Novartis, Ono Pharma, Pharmaxis, Schering, Wyeth. Support for the production of educational materials: AstraZeneca, GlaxoSmithKline, Merck Frosst, Boehringer-Ingelheim, Novartis. Organizational: Chair of the Global Initiative for Asthma (GINA) Guidelines Dissemination and Implementation Committee, Knowledge Translation, Education and Prevention Chair in Respiratory and Cardiovascular Health, Vice-President of Interasma (Global Asthma Organization).

MEB, JLD, JM, JL and LB have no conflicts of interest to declare.

FM considers having no conflict of interest but wishes to declare what can be perceived as potential conflicts of interest. Received fees for speaking at conferences sponsored by Boehringer Ingelheim, Novartis and Grifols; research grants for participating in multicentre trials sponsored by GlaxoSmithKline, Boehringer Ingelheim, AstraZeneca, and Novartis; and unrestricted research grants from Boehringer Ingelheim, Novartis, and Grifols. F.M. holds a CIHR/GlaxoSmithKline research chair on COPD.

Authors' contribution

MEB, JLD, FM and LPB contributed to conception and design of the study. JM, JL, LB, FM and LPB contributed to collection of data. MEB, FM and LPB contributed to analysis and interpretation of data. LPB and MEB contributed to writing of the manuscript. All authors reviewed the manuscript and approved its final version. LPB is the guarantor of this study.

Acknowledgments

The authors thank Serge Simard for performing all statistical analyses, Manon Boisvert, Myriam Nadeau and Justine Veilleux for data entry. Finally, we thank all patients who agreed to participate to the study.

Additional information

Funding

This study received a non-restrictive investigator-initiated grant from AstraZeneca Canada Inc.