Prevalence, clinical characteristics and morbidity of the Asthma-COPD overlap in a general population sample

ABSTRACT

Objective: Although asthma and chronic obstructive pulmonary disease (COPD) have been regarded as distinct conditions, emerging literature suggests that overlapping phenotypes, called asthma-COPD overlap (ACO), exists. The aim of this study was to describe prevalence, patient characteristics and morbidity of ACO. Methods: From a cross-sectional population sample, the West Sweden Asthma Study, subjects with suspected asthma, chronic bronchitis or COPD, and a random sample, were invited to clinical examinations. ACO was defined as doctor-diagnosed asthma, or clear clinical signs of asthma at examination, with a FEV₁/FVC < 0.7. Results: Subjects were categorized as ACO (N = 181), COPD only (N = 89), asthma only (N = 651) or healthy (n = 1036) based on clinical examinations. Prevalence of ACO was 3.4% in the random sample (N = 1172) and 18.1% among asthmatics (N = 138) in the random sample. Subjects with ACO (mean age 59 years, 54% women) had an age and gender distribution in between asthma only (45 years, 63% women) and COPD only (62 years, 41% women). Ever-smoking was reported by 71%, 48% and 74% in the ACO, asthma only and COPD only groups, respectively. Subjects with ACO had worse lung function (mean FEV₁% of predicted normal 76%) than asthma only (100%) and COPD only (87%) and reported more respiratory symptoms. Also respiratory related emergency visits were more common in ACO compared to asthma only and COPD only, respectively. Conclusions: ACO is present in 3.4% of the population and common among subjects with both asthma and COPD. Subjects with ACO had worse lung function and more symptoms than subjects with asthma or COPD only.

KEYWORDS:

• Epidemiology
• risk factors
• smoking
• lung function

Introduction

Asthma and chronic obstructive pulmonary disease (COPD) have historically been regarded as distinct conditions, but it is now recognized that they sometimes occur concomitantly. Asthma is characterized by episodes of reversible airflow obstruction accompanied by breathlessness, cough, wheeze and chest pain or tightness [1]. COPD is typically caused by tobacco smoking or exposure to other noxious particulates [1, 2]. COPD generally develops later in life and is characterized by incomplete reversible airflow limitation resulting in progressive decline in lung function [2]. In recent years, there has been an increased focus on the heterogeneity of COPD [3], and an emerging literature suggests that overlapping subsets of different phenotypes exists [1, 4]. In fact, the 2014 Global Initiative for Asthma-Global initiative for Obstructive Lung Disease (GINA-GOLD) guidelines suggested that this group of patients should be considered as a subclass of respiratory conditions, either COPD or asthma, termed Asthma-COPD-Overlap Syndrome (ACOS) [5]. Lately, the term ACO has been more widely accepted.

Recognizing the different phenotypes within ACO, as well as their respective prevalence, is important to understand the underlying disease processes and may have clinical implications [6, 7]. However, data is scarce and prevalence^{Figure 1} estimates have varied widely across studies [6, 8], where current estimations correspond to a population level prevalence of about 1–5% in ages over 40 years. The phenotypes are clinically relevant due to the potential difference in response to therapeutic interventions, affecting the choice of treatment approach. Updated Finnish guidelines now recommend that the treatment of ACO should cover both asthma and COPD, thus including inhaled corticosteroids together with a long-acting bronchodilator [9]. However, ACO patients are often excluded from clinical trials, and their response to medications used for asthma or COPD is, therefore, not well characterized [1].

Current evidence suggests that compared with those who have an asthma or COPD diagnosis alone, individuals with ACO express a more rapid decline in lung function as well as more frequent exacerbations, increased use of healthcare resources, worse quality of life and a higher mortality [5]. In a population-based study by Miravitlles et al. [3], COPD patients with an ACO phenotype demonstrated more dyspnoea, wheezing, exacerbations, worse respiratory related quality of life and reduced levels of physical activity. In a Finnish registry study, subjects with ACO used more medication and had increased number of hospitalizations [10]. Also in the COPDGene Study, it was demonstrated that patients with ACO experienced a greater disease burden than those with asthma or COPD only [4, 11].

However, there is scarce knowledge regarding the demographic and clinical characteristics of ACO patients. It is, therefore, important to understand more thoroughly the clinical and therapeutic implications of this group of patients. The aim of this study was to describe prevalence, patient characteristics and morbidity of ACO by use of the West Sweden Asthma Study database.

Methods

Study population

This cross-sectional study includes participants from the population-based West Sweden Asthma Study (WSAS) performed in Västra Götaland in West Sweden. WSAS was initiated in 2008 when 30 000 randomly selected subjects, aged 16–75 years, received a postal questionnaire to which 18 087 responded. The study population and selection procedure have been described in detail elsewhere [12–16]. The subjects included in the current study were selected from these responders to the postal questionnaire: from the 18 087 responders, a random sample of 2000 were invited to clinical examination of which 1172 (59%) participated. All subjects with suspected asthma not already included in random sample (N = 1524) were also invited to the clinical examinations, of which 834 (55%) participated. Furthermore, additional subjects with suspected COPD were invited (N = 561), of which 209 (37%) participated. In total, 2215 subjects participated in the clinical examinations and were categorized as healthy (N = 1036), ACO (N = 181), asthma only (N = 651), and COPD only (N = 89 or remaining (N = 258, not fulfilling the criteria of the other groups) based on the results from the interview and clinical examinations (Figure 1). The study was approved by the regional ethics board in Gothenburg.

Figure 1. Flow chart of the West Sweden Asthma Study.

Clinical examinations

The clinical examinations (performed during 2009–2012) included dynamic spirometry using a MasterScope spirometer (Jaeger, Höchberg, Germany) according to the ERS/ATS guidelines [17] using the Global Lung Initiative normal equation for reference values [18] to calculate Forced Expiratory Volume after 1 second percent of predicted normal (FEV₁% predicted). Reported values are post bronchodilator administration. The reversibility testing was performed using 4 × 0.1 mg salbutamol through a spacer and lung function was measured after 15 minutes. COPD was defined using a fixed ratio of Forced Expiratory Volume after 1 second/Forced Vital Capacity (FEV₁/FVC) <0.7 post-bronchodilator, and disease severity was assessed using the GOLD criteria [19]. The diffusion capacity of carbon dioxide (DLCO) was measured with a single breath method using MasterScreen PFT (Jaeger, Höchberg, Germany). Methacholine challenge was performed using a shortened protocol with a highest administered cumulative dose of 1.96 mg using Spira equipment (Spira Respiratory Care Center Ltd, Hämeenlinna, Finland).

Blood cell counts and haemoglobin measurements were performed according to routine procedures at the Sahlgrenska University Hospital. Sensitization was assessed using specific IgE to 11 aeroallergens (Thermo Fisher, Uppsala, Sweden), with a value above 0.35 kUA/l considered as positive. The fraction of exhaled NO was assessed using a NiOX (Aerocrine, Morrisville, NC, USA).

The clinical examination also included detailed structured interviews and questionnaires on respiratory diseases and symptoms, various comorbidities, health care utilization and potential risk factors.

Definitions of ACO, asthma and COPD

ACO was defined as having a post-bronchodilator FEV₁/FVC < 0.7; in addition, subjects should fulfil at least 1 of 4 criteria: 1) reporting a physician-diagnosed asthma with respiratory symptoms or use of asthma medication during the last 12 months; 2) a positive reversibility test with an increase in FEV₁ of >12% and >200 ml; 3) reporting ever asthma, and a positive methacholine challenge with asthma symptoms during the last 12 months; or 4) reporting ever asthma and asthma symptoms during the last 12 months with increased blood eosinophils (>0.4 ^* 10⁹/l).

ACO ≥ 2 was defined as having ACO but also fulfilling spirometric criteria for GOLD grade ≥2, i.e. an FEV₁% predicted <80%.

Asthma only was defined as having a post-bronchodilator FEV₁/FVC ≥ 0.7 and fulfilling any of the 4 criteria above.

For COPD only, the definition was having a post-bronchodilator FEV₁/FVC < 0.7 and none of the asthma criteria defined above.

Healthy subjects, i.e. free from respiratory disease, should not fulfil any of the criteria for ACO, asthma only or COPD only or report ever having had asthma. Remaining subjects, who did not fulfil any of the above criteria, mainly subjects reporting ever asthma but with no current symptoms, were excluded from further analyses.

Other definitions

Ever asthma: Reported ever having had asthma but not a physician-diagnosed asthma MMRC dyspnoea grade 2.

“On level ground, I walk slower than people of the same age because of breathlessness, or have to stop for breath when walking at my own pace.”

Persistent wheeze: “Do you have whistling or wheezing in your chest most days of the week.”

Exacerbation: Reporting any one of the following; hospitalization, visit to emergency unit, oral steroids, sick leave for at least 3 days due to breathing problems, visit general practitioner due to breathing problems or other of equal value during the last 12 months.

Smoking was assessed using two parameters 1) Current smoking status: Smokers reported smoking during the year preceding the survey; Former smokers reported having quit smoking at least 12 months preceding the survey; Non-smokers reported neither smoking nor former smoking, and 0 pack-years.

Comorbid cardiovascular disease was physician-diagnosed self-reported angina, heart attack, heart failure and arrhythmia.

Statistical analyses

Comparisons of prevalence were analysed using two-sided Fisher's exact test for statistical significance. Comparisons of means are presented with standard deviations and were analysed using the independent Student's t-tests. p values < 0.05 were considered statistically significant. Unadjusted logistic regression was used to calculate odds ratios (ORs) with 95% confidence intervals (CIs). Adjusted logistic regressions, including statistically significant variables representing different domains, were used to calculate adjusted OR with 95% CI. All analyses were performed using SPSS v 22.0 (IBM Corp, New York, USA). Prevalence of disease was investigated in the random sample (N = 1172), while comparisons between ACO, asthma only and COPD only were done using subjects from the entire sample (N = 2215).

Results

Prevalence in the random sample

In the random sample (N = 1172), the prevalence of ACO was 3.4% and the prevalence of subjects fulfilling criteria for ACO and GOLD grade ≥2 was 2.1% (ACO ≥ 2). In the population older than 50 years, the prevalence of ACO and ACO ≥ 2 was higher, 5.3% and 3.4%, respectively (Figure 2). Among^{Table 1} asthmatic subjects with current symptoms or use of asthma medication, the prevalence of ACO was 18.1%, and among subjects with ever asthma, the prevalence was slightly lower, 16.3%. The prevalence of asthma only was 9.6%, and in addition, 3.2% reported ever asthma but did not report symptoms or medication during the last 12 months. Controlled asthma according to GINA was less common in ACO than in asthma only (26% vs. 40%, p = 0.04). The prevalence of COPD only was 4.8%. Prevalence of ACO among subjects with COPD increased with GOLD grade. In GOLD grade, 1, 24.6% fulfilled the criteria for ACO, among GOLD grade 2, the proportion of patients with ACO was 67.7%, and among subjects with GOLD grade 3, all subjects met the ACO criteria. In total, 41.5% of subjects with a FEV₁/FVC < 0.7 fulfilled the criteria for ACO, and 10% of these had GOLD ≥ 2.

Figure 2. Prevalence of ACO and ACO with GOLD grade ≥ 2 (ACO ≥ 2) in the random sample and among subjects 50 years or older in the random sample.

Demographics and exposures

Subjects with ACO had a mean age and gender distribution in between asthma only (45 years, 63% women) and COPD only (62 years, 41% women) with a mean age of 59 years and 54% were women (Table 1). Having a family history of asthma or allergy was more common among subjects with ACO than among subjects with COPD only, while there was no difference in a family history of chronic bronchitis or COPD (Table 1). Having a family history of allergy was more common in subjects with asthma only than with ACO, otherwise there were no differences.

Table 1. Demographic features for subjects with Asthma-COPD Overlap, asthma only, COPD only and healthy.

							p-values
		ACO	ACO ≥ 2^^	Asthma only	COPD only	Healthy	ACO vs asthma	ACO vs COPD	ACO vs healthy
		N = 181	N = 98	N = 651	N = 89	N = 1036
Age	Mean (±SD)	59 (12.5)	62.2 (11.2)	45.9 (15.3)	62.4 (11.2)	50.5 (15.4)	<0.001	0.027	<0.001
		%	%	%	%	%
Gender	Female	54.1%	54.1%	62.9%	41.6%	53.8%	0.038	0.070	0.928
Family history	Asthma	31.5%	33.7%	37.0%	14.6%	16.1%	0.236	0.002	<0.001
	Bronchitis^¤	21.5%	22.4%	16.6%	16.9%	8.5%	0.114	0.294	<0.001
	Allergy	37.6%	32.7%	48.7%	16.9%	27.7%	0.017	<0.001	<0.001
	OAD^#	39.2%	40.8%	44.4%	23.6%	21.1%	0.245	0.006	0.000
Occupational exposure		35.9%	35.7%	28.6%	32.6%	23.6%	0.062	0.567	<0.001
Smoking	Smoker	21.0%	26.5%	12.6%	24.7%	11.9%	<0.001^*	<0.001^*	<0.001^*
	Former smoker	50.3%	51.0%	35.0%	49.4%	36.5%
	Non-smoker	28.7%	22.4%	52.4%	25.8%	51.6%
Pack-years	0	28.7%	22.4%	52.4%	25.8%	51.6%
	<10	18.2%	11.2%	21.0%	11.2%	21.2%	0.570^*	0.238^*	0.569^*
	10–20	11.0%	11.2%	11.2%	19.1%	11.9%
	21–39	24.9%	30.6%	9.1%	33.7%	8.8%
	≥40	13.3%	20.4%	2.5%	9.0%	2.6%
Pack-years	Mean (±SD)	24.5 (16.4)	29.2 (15.9)	14.5 (13.9)	23.9 (16.2)	14.6 (12.9)	<0.001	0.811	<0.001

^ ACO≥2 includes subjects with ACO and GOLD stage 2, all these subjects are included also in ACO.

^* Test for trend.

^¤ Includes chronic bronchitis, emphysema and COPD.

^# Obstructive airway disease.

Bold values = p < 0.05.

There were no significant differences in occupational exposure to gas, dust or fumes between the groups. Overall prevalence of current or former smoking was similar among subjects with ACO and COPD only, 71.3% and 74.1%, respectively. Among subjects with asthma only, 47.6% reported either current or former smoking (Table 1).

Clinical characteristics

Of^{Table 2} the 181 subjects with ACO, 153 had physician-diagnosed asthma, 85 a positive reversibility test, 42 had ever asthma with symptoms and a positive methacholine challenge and 59 ever asthma with symptoms and a high methacholine challenge. Most cases of ACO were included on the criteria of a FEV₁/FVC < 0.7 and a physician-diagnosed asthma (153/181). Of the remaining 28 individuals with ACO, 27 were included because of ever asthma and a positive reversibility test and one based on ever asthma combined with symptoms and a high blood eosinophil count.

The mean FEV₁% predicted was lowest in the ACO group (mean 76%) compared with asthma only (mean 100%) and COPD only (87%). An eosinophil count ≥ 0.4 ^* 10⁹/l was found in approximately one fifth of all subjects with ACO and asthma only and among approximately one tenth of subjects with COPD only. The mean eosinophil count was highest in the ACO group at 0.27 ^* 10⁹/l and lowest in COPD only 0.19 ^* 10⁹/l (p = 0.011, Table 2). Also, the neutrophil count was higher in ACO at 4.23 ^* 10⁹/l than in asthma only and COPD only (3.84 and 3.95 ^* 10⁹/l, p = 0.002 and 0.215, respectively). Sensitisation to any of the 11 common aeroallergens was found in 44% of subjects with ACO, 61% of subjects with asthma only and 16% of subjects with COPD only (Table 2). No differences were noted in exhaled NO between disease groups. There were no statistical differences in regards to body mass index (BMI) between ACO and asthma only (p = 0.500), while subjects with COPD only had a lower BMI (p = 0.002). Subjects with ACO had an older age of asthma onset compared with asthma only, by mean 26.5 years vs. 20.3 years (Table 2). In general, subjects with ACO reported more respiratory, symptoms compared with subjects with asthma only or COPD only (Figure 3).

Table 2. Clinical characteristics for subjects with Asthma-COPD Overlap, asthma only, COPD only and healthy.

							p-values
		ACO	ACO ≥ 2^*	Asthma only	COPD only	Healthy	ACO vs asthma	ACO vs COPD	ACO vs healthy
BMI	Mean (±SD)	27.1 (4.6)	27.7 (4.9)	27.4 (5)	25.4 (4)	26.2 (4.2)	0.500	0.002	<0.001
Age of onset	Mean (±SD)	26.5 (21.6)	29.4 (22.4)	20.3 (16.1)			0.006
FeNO 50	Mean	24.5 (24.8)	20.3 (15.9)	26.9 (23.6)	19.8 (11.6)	19.4 (12.2)	0.289	0.073	0.020
FEV1% predicted	Mean (±SD)	76.2 (18.1)	63.0 (12.9)	100.4 (12.8)	86.8 (15.6)	104.4 (12.7)	<0.001	<0.001	<0.001
Specific IgE	% positive	43.6%	35.4%	61.0%	16.1%	23.2%	<0.001	<0.001	<0.001
Blood eosinophils	% ≥ 0.4^*10^9/L	19.3%	21.4%	17.4%	9.1%	7.4%	0.581	0.034	<0.001
Blood eosinophils	^*10^9/L, mean (±SD)	0.27 (0.27)	0.27 (0.25)	0.23 (0.16)	0.19 (0.14)	0.18 (0.15)	0.011	0.011	<0.001
Blood neutrophils	^*10^9/L, mean (±SD)	4.23 (1.64)	4.45 (1.57)	3.84 (1.47)	3.95 (1.77)	3.61 (1.35)	0.002	0.215	<0.001

^* ACO2 includes subjects with ACO and GOLD stage 2; all these subjects are included in ACO as well.

Bold values = p < 0.05.

Figure 3. Prevalence of respiratory symptoms in subjects with ACO, asthma only, COPD only and healthy subjects.

Subjects with ACO reported more frequent visits to an emergency department due to breathing problems during the last 12 months. Emergency department visits due to breathing problems was reported by 16% of subjects with ACO and 24% of subjects with ACO ≥ 2 compared with 9% of subjects with asthma only and 7% of subjects with COPD only (Table 3). Except for heart failure, which was more common among subjects with ACO, there were no statistical differences in prevalence of cardiac comorbidities between the disease groups. Asthma control, as assessed by GINA, showed that 80% of subjects with ACO were either uncontrolled or partly controlled, compared with 54% among subjects with asthma only (p < 0.001). The occurrence of an exacerbation during the last 12 months was 27% in ACO and 21% asthma only (p = 0.101) and more common in ACO than among subjects with COPD only (14%, p = 0.007).

Table 3. Health care contacts and heart comorbidities in subjects with Asthma-COPD Overlap, asthma only, COPD only and healthy.

							p-values
		ACO (%)	ACO ≥ 2^* (%)	Asthma only (%)	COPD only (%)	Healthy (%)	ACO vs asthma	ACO vs COPD	ACO vs healthy
Health care contact last 12 months		38.1	49.0	35.3	16.9	11.7	0.485	0.000	<0.001
Hospitalized last 12 months		17.0	9.2	7.9	10.1	3.2	0.106	1.000	0.022
Emergency visit last 12 months		16.0	24.5	9.1	6.7	1.2	0.009	0.035	<0.001
Heart comorbidity last 12 months	Any	16.0	21.4	11.2	18.0	11.8	0.095	0.729	0.113
	Angina	4.4	5.1	2.5	1.1	1.8	0.205	0.279	0.049
	Heart attack	1.1	1.0	1.5	2.2	2.3	1.000	0.601	0.410
	Heart failure	3.3	6.1	0.8	0.0	0.4	0.017	n/a	0.001
	Arrhythmia	8.8	13.3	5.1	11.2	5.2	0.073	0.519	0.058

Bold values = p < 0.05.

Risk factors

In unadjusted logistic regression analyses, the strongest risk factor for ACO compared to asthma only was belonging to the oldest age group (≥61 years) and having 40 or more pack-years (OR 9.6, 95% CI 6.0–15.6 and OR 9.8 95% CI 4.9–19.7, respectively) but also occupational exposure to gas dust and fumes (OR 1.4, 95% CI 1.0–2.0) and low educational level (OR 2.2, 95% CI 1.4–3.4, Table 4). However, in an adjusted model comparing asthma only and ACO, only having ≥20 pack-years and age ≥ 45 years remained significant (Table 5).

Table 4. Unadjusted logistic regression comparing Asthma-COPD overlap with asthma only, COPD only and healthy.

		OR (95% CI)	OR (95% CI)	OR (95% CI)
		ACO vs Asthma only	ACO vs COPD only	ACO vs Healthy
Female gender		0.7 (0.50–0.97)	1.66 (1.00–2.77)	1.01 (0.74–1.39)
Occupational exposure		1.39 (1.00–1.98)	1.17 (0.68–2.01)	1.84 (1.31–2.57)
Family history of OAD^*		0.83 (0.59–1.17)	2.24 (1.26–4.00)	2.58 (1.84–3.63)
Smoking	Former smoker	2.62 (1.79–3.83)	0.91 (0.50–1.68)	2.48 (1.72–3.57)
	Smoker	3.04 (1.88–4.92)	0.76 (0.37–1.57)	3.18 (2.00–5.04)
Pack-years	<10	1.58 (0.98–2.55)	1.46 (0.62–3.45)	1.54 (0.97–2.45)
	10–20	1.80 (1.01–3.19)	0.52 (0.23–1.17)	1.67 (0.96–2.90)
	21–39	5.00 (3.08–8.13)	0.66 (0.34–1.30)	5.09 (3.22–8.03)
	≥40	9.84 (4.9–19.74)	1.33 (0.52–3.39)	9.15 (4.92–16.99)
Education	High school	1.19 (0.81–1.75)	1.33 (0.72–2.46)	1.41 (0.97–2.04)
	Elementary	2.16 (1.39–3.35)	0.96 (0.51–1.80)	3.02 (1.98–4.61)
Age	46–60	4.25 (2.60–6.95)	0.91 (0.36–2.31)	3.05 (1.88–4.93)
	≥61	9.64 (5.96–15.60)	0.48 (0.20–1.12)	4.26 (2.69–6.73)

^* OAD, obstructive airway disease.

Reference categories were male gender, no occupational exposure, no family history of OAD, non-smoker, no pack-years, education at university level, age ≤ 45 years.

Table 5. Adjusted logistic regression comparing Asthma-COPD overlap with asthma only, COPD only and healthy.

		OR (95% CI)	OR (95% CI)	OR (95% CI)
		ACO vs Asthma only	ACO vs COPD only	ACO vs Healthy
Female gender		0.74 (0.50–1.09)	1.95 (1.07–3.53)	1.17 (0.81–1.69)
Occupational exposure		1.08 (0.71–1.64)	1.33 (0.70–2.52)	1.71 (1.16–2.52)
Family history of OAD		0.93 (0.64–1.37)	2.36 (1.27–4.36)	2.47 (1.71–3.57)
Pack-years	<10	1.52 (0.91–2.53)	1.71 (0.69–4.24)	1.31 (0.81–2.11)
	10–20	1.34 (0.72–2.48)	0.51 (0.20–1.25)	1.18 (0.66–2.11)
	21–39	3.13 (1.82–5.39)	0.69 (0.32–1.47)	3.01 (1.83–4.95)
	≥40	5.53 (2.54–12.04)	2.02 (0.70–5.83)	5.04 (2.56–9.92)
Education	High school	0.86 (0.55–1.33)	1.40 (0.71–2.78)	1.24 (0.83–1.85)
	Elementary	0.80 (0.47–1.37)	0.95 (0.44–2.04)	1.43 (0.88–2.31)
Age	46–60	3.85 (2.27–6.53)	1.06 (0.38–3.01)	2.55 (1.50–4.33)
	≥61	7.84 (4.60–13.37)	0.54 (0.20–1.42)	3.15 (1.87–5.31)

^* OAD, obstructive airway disease.

Reference categories were male gender, no occupational exposure, no family history of OAD, no pack-years, education at university level, age ≤ 45 years.

The only significant risk factors according to unadjusted analysis for ACO compared to COPD only were having a family history of obstructive airway disease and female gender (OR 2.2, 95% CI 1.3–4.0, and OR 1.7, 95% CI 1.0–2.8 respectively, Table 4). These risk factors remained statistically significant in an adjusted regression model with ORs of 2.4 and 2.0, respectively (Table 5).

Discussion

The results from our study show that ACO is a fairly common condition in the population. The present study indicates that the majority of subjects with ACO had a previous asthma diagnosis that developed a chronic airway obstruction, and smoking and age were important contributing factors for the development of ACO.

Subjects with ACO had worse lung function, and reported more respiratory symptoms and more emergency visits compared with subjects with only asthma or COPD, possibly reflecting undertreatment.

Prevalence estimates of ACO in the general population based on population samples are still rare. The precise definition of ACO varies, and there is still no consensus of how to define this phenotype in the best way [20, 21]. Prevalence estimate varies, both depending on the used definition of ACO, asthma and COPD but also the diagnostic labelling by the doctors if the data is based on health care labels, which often is the case. Prevalence estimates of 1–5% of the adult general population have been suggested [6, 8], and thus our result of about five percent in age >50 years is in line with the higher level of previously reported results and higher than we anticipated. Our study further demonstrates that as many as 20% of the asthmatics and 40% of the subjects with a FEV₁/FVC < 0.7 also fulfilled the ACO criteria utilized in the present study. The latter result is lower than the 55% found by in the referred study from New Zealand [6], but is in contrast to result from a study in the USA where only 13% of 915 subjects with spirometrically defined COPD also reported physician-diagnosed asthma [11], and to a study in Spain where 17% of subjects with spirometrically defined COPD had self-reported asthma [22]. Our result with about 20% of asthmatics having FEV₁/FVC < 0.7 is fairly in line with the result of the study from New Zealand [6].

Register studies generally have yielded considerably lower estimates of ACO than those found by us and by the referred studies as well [6, 8, 11, 22]. A large register study from Spain found the prevalence of ACO to be 0.55% and of COPD 3.0% [23]. In Finland, with a population of 5.35 million in 2009, slightly more than 100 000 subjects aged ≥ 35 years had been discharged from hospital for either asthma or COPD between year 2000 and 2009, corresponding to about 4% of the population aged ≥ 35 years [10].

Our study also confirmed the assumption made by GINA and GOLD [5] that disease severity is greater in subjects suffering from ACO, as compared to asthma only or COPD only. Several symptoms common in asthma, such as recurrent and persistent wheeze, and symptoms primarily associated with bronchitis and COPD, such as sputum production, were more common in ACO compared with COPD only or asthma only. This result is also in line with findings by other research groups [21, 22, 24]. While the majority of subjects with ACO and COPD only has a mild disease, fulfilling criteria for GOLD 1 or GOLD 2, it is clear that the subjects with ACO has a worse lung function. In the previously referred study from New Zealand, a condition similar to ACO was identified by using cluster analysis and was found to be the by far most severe phenotype of the five with obstructive airway diseases [25]. Morbidity expressed as utilization of health care resources, which in our study corresponded to emergency department visits, was considerably higher in subjects with ACO compared to COPD only or asthma only, also a result confirming previous findings [10, 11, 24, 26]. It has also been shown that ACO patients with a late asthma diagnosis have a worse prognosis [27].

Of the publications about ACO, only few discuss risk factors including comorbidities [23, 24, 28]. Not surprisingly, the prevalence of ACO increased by age in our study. This effect could at least in part be due to use of the fixed ratio for defining COPD. The risk factors identified by the above referred studies are in line with what can be expected for a condition fulfilling criteria for both asthma and COPD. It clearly seems that asthmatics who smoke develop irreversible airflow obstruction with increasing number of pack-years as found by the large follow-up in 1999–2001 of the initial European Community Respiratory Health Survey (ECRHS) from 1991 to 1993 [28]. We found ORs up to 5 compared with asthma only but with no or few differences of smoking as risk factor for ACO compared with COPD. Having a family history of asthma was not a risk factor for ACO vs asthma only, while it showed to be a significant risk factor compared with COPD only with an OR of 2.4. This indicates a stronger hereditary factor for ACO compared to COPD, strengthening the hypothesis that it is primarily asthmatics that develop COPD who constitute the ACO group. In line with our results, the ECRHS follow-up found a family history of asthma to be present in significantly more ACO subjects compared with subjects with COPD [28]. No difference in occupational exposure between ACO and asthma only or COPD only was found in our study in the adjusted risk analysis. However, in contrast to our result, the ECRHS follow-up found occupational exposure to be slightly but significantly associated with COPD only. As occupational exposures have been identified as a risk factor for both asthma and COPD [29, 30], this also strengthens the theory that ACO is an overlap between asthma and COPD.

In our study, heart failure was more common in subjects with ACO, while no other differences in cardiac comorbidities or other diseases were found between the groups. Others have reported that anxiety, allergic rhinitis, osteoporosis and gastro-esophageal reflux were slightly but independently associated with ACO [23].

The current study confirmed that an overlap of different phenotypes exists in both asthma and COPD. Exhaled NO and BMI reflect a disease that has features of both asthma and COPD, with similar FeNO levels in ACO and asthma only and a lower BMI among subjects with COPD only. The gender distribution in our study seems to be in between those traditionally seen for asthma and COPD with more men having ACO than asthma, and more women having ACO than COPD. Overall, the prevalence of ACO was higher among men, also a result in line with the ECRHS follow-up survey [28], in contrast to the study from Spain [23].

Although ACO clearly seems to be the simultaneous existence of two disease entities, asthma and COPD, the question regarding a new taxonomy for airway diseases has been initiated [31, 32]. Noteworthy, the taxonomy of obstructive airway diseases as to a large extent remained unchanged since Scadding in the late 1950s and the CIBA Guest Symposium in 1959 defined asthma, chronic bronchitis and emphysema [33, 34], and the only real new term, COPD, was launched in early 1960s [35] and became more clearly defined and commonly used in the middle of the 1990s [36, 37].

The main strength of our study is the population-based sampling procedure with a randomly selected sample allowing prevalence to be assessed [12, 15, 16]. The cross-sectional design is a limitation and does not allow causality or long-term outcomes to be evaluated. In the present study, all subjects who fulfilled GOLD grade 3 criteria also fulfilled criteria for ACO, which could have resulted from the cohort primarily being enriched for asthma and not COPD and should not be interpreted that the most severe COPD cases also have ACO. Another possible weakness is a misdiagnosis of COPD as asthma by the physician doing the original diagnosis. Approximately 27% of subjects with ACO and COPD only were never smokers which might seem high; however, it is in line with other areas of Sweden [29] and might be caused by other exposures such as passive smoking and occupational exposures.

In conclusion, the prevalence of ACO in the general adult population was 3% and increased with age. The development of ACO seems in large parts dependent on smoking and environmental exposures on subjects with asthma, leading to chronic airway obstruction. The subjects with ACO had a more severe disease with more symptoms and exacerbations than in asthma only and COPD only.

Declaration of interest

Dr. Ekerljung report grants from Herman Krefting Foundation against Asthma and Allergy Research and a grant from AstraZeneca NordicBaltic, during the conduct of the study. Prof. Lötvall has no relevant disclosures in relation to this work. He is currently employed by a biotech company, Codiak Biosciences, which is developing therapeutics for diseases, but not currently for respiratory disease. Dr. Stratelis is a full-time employee of AstraZeneca, the sponsor of the study. Dr. Telg is a full-time employee of AstraZeneca, the sponsor of the study. Dr. Bjerg has nothing to disclose. Dr. Hagstad reports personal fees from Boehringer Ingelheim, outside the submitted work. Dr. Mincheva has nothing to disclose.

Additional information

Funding

The work was supported by unconditional grants from Herman Krefting's Foundation for Asthma and Allergy Research and AstraZeneca NordicBaltic.