Asthma–COPD overlap syndrome (ACOS) in primary care of four Latin America countries: the PUMA study

The principal findings of this study are: first, ACOS prevalence depends on the asthma component definition and the population it is being evaluated in (e.g. general population, obstructive population or COPD population). The lowest prevalence was 2.3% when ACOS was defined as wheezing plus reversibility and FEV₁/FVC <0.70 in the total population, whereas the highest was 26.5% when ACOS was defined as previous medical diagnosis of asthma and FEV₁/FVC <0.70 in the COPD population. Second, after adjusting for confounding factors, ACOS defined as FEV₁/FVC <0.7 and wheezing plus reversibility was associated with a higher risk for exacerbations compared with those subjects with COPD.

Proposed definitions for ACOS vary widely and include: a) patients with COPD who have a previous diagnosis of asthma; b) patients with a spirometric COPD definition who have significant reversibility (FEV₁/FVC <0.70 and post-bronchodilator increase in FEV₁ or FVC of 200 mL and 12%); c) patients with asthma who have persistent airflow obstruction. It is important to recognise whether a patient has ACOS as it may influence the clinical course, long-term outcome, and response to therapy. Other documents such as the Global Initiative for Asthma (GINA)–Global Initiative for Chronic Obstructive Lung Disease (GOLD) consensus and the Spanish guideline for COPD have also proposed their own definitions [35, 36], however they have not been fully validated in large cohorts.

Using a definition similar to the present study, some authors have assessed ACOS prevalence in the general population. Marsh et al. found a prevalence of ACOS of 11% in the total population studied and 55% in the COPD population [18]. However, this study was conducted only in volunteers and had a small sample size. Using the population-based Spanish EPI-SCAN study data, Miravitlles et al. reported a prevalence of ACOS in the general population of 1.7%, and of 17.4% in the COPD patients using the previous asthma diagnosis definition [19].

The PLATINO study reports a prevalence based on previous asthma diagnosis and FEV₁/FVC <0.70 criteria in the total population of 2.9%, and a prevalence of 1.8% using criteria of wheezing plus reversibility and FEV₁/FVC <0.70 [3]. In the same study, in the obstructive population, the prevalence was 13% by previous asthma diagnosis and FEV₁/FVC <0.70 criteria, and 11.6% by wheezing plus reversibility and FEV₁/FVC <0.70 [3]. Other authors have assessed the prevalence of ACOS in selected COPD populations [6, 7, 22, 37]. The prevalence of ACOS in the COPDGene study was 12.6% using self-reported asthma criteria [7], and similar results have been reported elsewhere [6]. Recently, Cosio et al. reported an ACOS prevalence of 15% in a COPD Spanish cohort of over 800 patients using one major criterion for asthma definition (reversibility >400 mL and 15% plus medical history of asthma) or two minor criteria (blood eosinophils >5%, IgE >100 IU/mL, or two separate bronchodilator tests >200 mL and 12%) [22]. A higher prevalence (25%) was reported in the ECLIPSE cohort when using their primary study definition of COPD patients answering “yes” to the question “Have you ever had asthma?” [37].

Little information exists regards the prevalence of ACOS in the primary care setting. As expected, the prevalence varied depending on the method by which ACOS was defined. Barrechenguren et al. reported a prevalence of 5.4% using the previous diagnosis of asthma in newly diagnosed patients with COPD [38]. In a separate study, the same authors found a higher ACOS prevalence in COPD patients with a history of asthma (10.8%) [39]. Others have reported a prevalence of 5.5% using a history of asthma in the total study population, and 19.1% using a restrictive analysis (asthma defined by reversibility criteria) in the COPD population [40].

The findings of the present study are consistent with those reported in some general and selected COPD populations that have used the previous diagnosis of asthma plus spirometric COPD to define ACOS [3, 5, 7, 30]. The comparison with the PLATINO study deserves special consideration as this is another study from Latin America that use the same two ACOS definitions in the same population [3]. The most important difference between the two studies that needs to be highlighted is that the PLATINO study was a larger population-based (general population) study, whereas PUMA is a study in a primary care population at risk for COPD; as a result of these being two different populations, differences in the results are to be expected. The prevalence of ACOS by both definitions reported here in the PUMA study (population at risk for COPD) were slightly higher than those reported in the PLATINO study.

The above-mentioned findings support the concept that the criteria used to define ACOS, as well as the population used to calculate the prevalence, have a significant influence on prevalence; it is thus essential to know this information when interpreting the results of other studies. The discrepancies observed with the findings of primary care studies could be partially explained by the selection of participating patients (only newly diagnosed COPD patients and/or a younger population), and the ACOS definition used [38‐40]. However, when spirometric COPD diagnosis and asthma defined by reversibility was used to define ACOS elsewhere [40], the prevalence was similar to our results in the COPD population. Another important aspect to highlight is that the Latin American population has a very distinct characteristic of being exposed to biomass fuel. There is no literature on biomass exposure and ACOS. In the present study, more than a third of the patients in each group had biomass exposure and irrespective of the definition used, approximately 3% of patients with ACOS had no smoking history. The size of the PUMA sample does not allow us to analyse the characteristics of ACOS patients due to biomass exposure; therefore, futures studies in regions with high biomass exposure, such as Latin America, aimed at characterizing this group of patients are warranted.

Two recent systematic reviews and meta-analyses indicate that ACOS patients may have more symptoms, more frequent exacerbations and hospitalisations, worse HRQOL and higher healthcare costs than patients with only asthma or COPD [20, 21]. Similar to patients with COPD, ACOS patients appear to have a high occurrence of comorbidities, including diabetes. In agreement with the results of these systematic reviews, we found the spirometry plus symptom-based (wheezing plus reversibility and FEV₁/FVC <0.70) definition identify a clinical phenotype with more frequent exacerbations. Also, in agreement with other results, we did not find any difference in the number of comorbidities between the groups [39]. In the present study, the ACOS patients (defined by wheezing plus reversibility and FEV₁/FVC <0.70) had the lowest lung function measurements. These findings are consistent with other studies in population-based sample that reported lower level of lung function in the ACOS subjects compared with asthma and COPD groups [3, 41]. As has been mentioned previously, there is no universal definition for ACOS. However, this is a phenotype recognised as a different COPD subpopulation with important therapeutic implications. The GINA–GOLD consensus recommends the use of inhaled corticosteroids (ICS) in patients with suspected ACOS [34]. However, ICS therapy has been linked with increased risk of pneumonia in COPD patients [42, 43], so it is crucial to be as accurate as possible with the prevalence of ACOS as well as determining the most appropriate definition to avoid over-diagnosis and subsequent overuse of ICS in patients with COPD.

Finally, when considering and interpreting the current findings, it is important to be aware of the following study limitations: these results are not generalizable to all Latin American countries as the study was only performed in four countries; it is possible that some results did not reach statistical significance as a result of sample size and lack of power, despite the efforts made to ensure a representative sample. Nevertheless, the procedure used was the most reasonable in view of the operational possibilities in each country; this was a transversal study and so was only designed to evaluate the characteristics of the patients and not the follow-up; we did not assess any pathophysiological link among ACOS, COPD and asthma, or a pathway that could explain the characteristics of the ACOS patients. It is important to note that the PUMA centres were not randomised, so sites selection did not follow a representative sampling of national primary care practice. In addition, other limitations to consider are that the diagnosis of asthma was, in part, based on patient recall and this may influence the true “incidence” of ACOS, and “wheezing” was obtained from questionnaires and was not directly observed by a physician. Finally, it is important to highlight that although wheezing is a hallmark of asthma, it often occurs in COPD, especially during exacerbations. Hence, if a patient has wheezing there is a possibility that this symptom originated from a COPD exacerbation; therefore, is not entirely surprising that patients with “ACOS” had higher incidence of exacerbations. Another limitation is the lack of a variable that could indicate severity. We performed a sensitivity analysis including FEV₁ in the model as a proxy of severity, but the statistical model became unstable with variance inflation factors higher than 10. It should be highlighted that the direction of the association did not change adding FEV₁. Therefore, we opted to maintain the model with the best quality criteria for the adjustment and did not include FEV₁ as a possible proxy for severity of COPD in the model.