Study sampling was conducted using a pre-selected list of the post codes closest to each hospital. A list of random telephone numbers was obtained, stratified according to these post codes and quotas for sex and age groups, all according to the latest EU directives on data protection and privacy.14,15

During the first telephone call, the subject was informed about confidentiality and data protection, and if they agreed to respond, they were asked questions about their cohabitants, confirmation of the post code for assignment of the nearest hospital, previous diagnoses of respiratory disease (chronic bronchitis, emphysema, COPD, or asthma), smoking habit (smoker and number of cigarettes, never-smoker, former smoker and number of cigarettes) and presence of cough or expectoration. During the second telephone call, conducted by the investigator from the hospital, a survey was administered with questions on the previous diagnosis of respiratory diseases, smoking habit, and the presence of other symptoms associated with COPD. The variables collected during the visit with the healthcare professional provided a comprehensive profile of both non-COPD individuals who were selected for the study visit, and, in particular, participants in the COPD group.

Information were collected on age, sex, level of education, family conditions, weight and height, and conventional use of tobacco (cigarettes, pipe, cigar) or use of other modes of delivery (electronic cigarette, chewing tobacco, etc.).

Other questionnaires included: COPD Assessment Test (CAT); Hospital anxiety and depression scale (HADS); Yale Physical Activity Questionnaire (YPAS); European Coal and Steel Community Questionnaire on respiratory symptoms; Questions on exposure to dust and fumes in the workplace; Mini-Mental State Questionnaire.(administered only to participants 60 years of age or older); Fagerström test; and Prochaska's Stages of Change.

Lung Function and Clinical Tests conducted for the entire sample included: Baseline pulse oximetry; Forced spirometry; and assessment of multicomponent indices and comorbidities.

Baseline pulse oximetry was determined using a Pulsox 300i (Konica-Minolta, Japan) pulsioxymeter and the fraction of carbon monoxide (CO) in exhaled air was determined with aco-oxymeter (MicroCO, Carefusion, UK).

Forced spirometry was performed using a pneumotachograph (Vyntus Spiro, Carefusion, Germany), according to standardized procedures as indicated by SEPAR guidance16 and Global Lung Function Initiative (GLI) equations were used as reference value.17 Bronchodilator test was conducted with the inhalation of 400μg salbutamol. According to the ATS/ERS guidelines,18 criteria for bronchodilation were an increase in FVC or FEV1>200ml and greater than 12% compared to the baseline value. Only spirometries with quality grades A-C were accepted for analysis.15 Airflow limitation was defined as a postbronchodilator FEV1/FVC ratio<0.7 (or alternatively also expressed as an FEV1/FVC<lower limit of normal (LLN)). Underdiagnosis of COPD by age, sex and area was defined as the percentage of those not reporting a previous diagnosis of COPD, chronic bronchitis or emphysema among those with airflow limitation.

In order to have a better characterization of the population, comorbidities were quantified by means of: the Charlson index (19 comorbidities)19; the COTE index (12 comorbidities)20; and previous diagnosis of other respiratory diseases. As multicomponent indices, BODE and BODEx were calculated.21

According to data from the 2011 Census of Population and Dwellings, published by the National Institute of Statistics,22 the population of Spain aged 40 years or more comprises 23,957,645 individuals. Taking into account the 10.2% prevalence of COPD found in the EPISCAN study,6 an a priori sample size calculation estimated with an accuracy of ±0.6% and a 10% dropout rate, that approximately 10,200 eligible individuals were needed to be included in the study. Therefore, between 300 and 600 participants (150–300 men and 150–300 women) were included in each site. The geographic information system inverse distance weighted (IDW) interpolation technique was used for mapping the spatial distribution of epidemiological variables.23 The statistical and analysis plan are available elsewhere.13 A level of significance of 0.05 was used for all statistical tests performed on the study variables.

As per the primary analysis, 1077 out of 9092 participants were considered COPD, while 8015 were not, with a prevalence of 11.8% (95% C.I. 11.2–12.5). By sex, prevalence was 14.6% (95% C.I. 13.5–15.7) in males and 9.4% (95% C.I. 8.6–10.2) in females.

Cases with COPD were a mean of seven years older, less frequently female (41.6% vs. 54.1%), of lower attained education, and were more frequently current- and former-smokers than controls (p<0.001), although the number of cigarettes and pack-years in non-COPD participants was substantial (Table 1); reported use of e-cigarettes, (7.0% vs. 5.5%) was significantly different (0.0450). There were also significant differences versus participants without COPD in all respiratory and other clinical variables, including living alone, previous respiratory diagnoses, respiratory symptoms, more comorbidities with the Charlson index, greater BODE and COTE scores, cognitive impairment, and depression HADS (all p<0.001), except for BMI (p=0.5482) and Anxiety HADS (p=0.4112).

As expected, COPD prevalence increased by age in both men and women, and the greatest prevalence was observed in those 80 years and older, with a prevalence of 26.1% (95% C.I. 20.9–31.9) in women and of 34.7% (95% C.I. 28.1–41.6) in men (Fig. 2). There was a high variability (2.4-fold) among the 17 regions, with a minimum prevalence of 7.1% in Asturias and a maximum of 17.3% in Catalonia (Table 2 and Fig. 3). Estimates of the prevalence of COPD according to the LLN were 6.0% (95% C.I. 5.5–6.5) overall, by sex being 7.1% (95% C.I 6.4–8.0) in males and 4.9% (95% C.I. 4.3–5.6) in females (Table 2). Underdiagnosis of COPD was 74.7%, higher in women than in men (80.6% vs. 70.4%, p<0.001), yet again with a high variability by region (Table 3). However, COPD overdiagnosis was overall low (2.6%). Underdiagnosis of COPD measured by the LLN was 61.0%, again higher in women than in men (78.4% vs. 68.6%, p<0.001), yet again with a high variability by region. However, according to the LLN there were no statistically significant differences in the autonomous communities compared to the total (Table 3).

Fig. 2.

Prevalence of COPD by age and sex according to the fixed ratio FEV1/FVC<0.7 or the lower limit of normal (LLN) in: A) Women and B) Men.

(0.23MB).

Fig. 3.

Map of the prevalence of airflow limitation in Spain by Autonomous Community, measured as a fixed ratio in (A) women; and (B) men.

(0.2MB).

Adhering to the compromise in global health agendas of an expanded emphasis on non-communicable diseases, sound evidence on trends and determinants of COPD at the national and international levels are essential. In here, EPISCAN II updates previous findings and includes new estimates and trends on COPD prevalence in Spain. The confirmation that COPD is one of the most prevalent conditions (more than one in ten) in the general population, found in all ages from young adults to the very elderly, and that is more frequent in men but with an increasing burden in women, yet with a regional variability (2.4-fold) of a similar magnitude than asthma,24 will set ground for further Public Health interventions. Note that a previous diagnosis of asthma was two-fold higher in those with COPD (16.9%) than without (6.7%) in this study. Regrettably our efforts to reduce COPD underdiagnosis, appear not enough, as three out of four individuals with objective airflow limitation compatible with COPD reported no medical diagnosis previously, therefore suffering unnecessary individual and population burden, as COPD is considered a preventable and treatable disease.2

The so-called shoe-leather epidemiology25 is essential to monitor trends of chronic diseases, as populations change, but diseases and risk factors also change.26 Direct comparison with previous estimates of spirometry in Spain will require of further analyses,27 as age thresholds differ, spirometers and guidelines have changed, and even interpretation of respiratory manoeuvres has evolved.28 However, compared with the first EPISCAN, an increase of COPD in women is seen in the overlapping age strata from age 50 years and onwards (3.2% vs 3.2% in 40–49 yrs., 4.5% vs 6.6% in 50–59 yrs., 7.6% vs 10.0% in 60–69 yrs., and 10.8% vs 14.5% in 70–79 yrs.), which is a new reminder of the growing toll of women catching up with men in cigarette smoking.29 Of note, Spain is considered to be in phase IIIb of the tobacco epidemic,30 but new forms of smoking, including the health effects of vaping and heat-not-burn products such as IQOS,31 should be actively assessed. Our finding of reported use of e-cigarettes, in 7.0% of COPD cases but also in 5.5% of non-COPD participants is worrying.32 Exploring the asthma and COPD overlap,33 or those with preserved ratio impaired spirometry (PRISm), alternatively known as restrictive, unclassified spirometry,34 are just a few of the pre-established analyses scheduled with EPISCAN II,12 all beyond the scope of this manuscript.

Our finding of a 74.7% underdiagnosis of COPD in Spain should be put into perspective, as it is no better than the 73% in 1997,5 and only marginally different than the 78% seen in 20076; these are unwelcome news in our global effort to reduce COPD unmet burden,4,35 where new strategies might be field tested and then implemented.36 The 80.6% underdiagnosis in Spanish women, ten points higher than the 74.6% observed in men, confirms previous findings of higher female underdiagnosis in Spain, which is at odds with observations elsewhere, where male COPD underdiagnosis prevails, and to date its reasons appear elusive.37 There were substantial differences in the variability of COPD underdiagnosis by Autonomous Community, total, and by sex, with percentages ranging from the 35% only in men in La Rioja and up to 100% in women in Castilla-La Mancha (Table 3). Given that a common protocol was identical in all sites, and that these findings might be outlier values but not errors, the determinants of these differences will be specifically explored in already planned analyses of EPISCAN II. However, for the record indeed La Rioja has an advanced system of COPD screening according to the National Strategy,38 and they apply a near systematic population case-finding and screening.

There are few studies that “medicalize” new COPD diagnoses (or airflow limitation) after a population spirometric study.39,40 Perhaps the most relevant is the one by Llordés et al.41 They conducted spirometry in 1738 population smokers aged 45 years or older. All those newly diagnosed with COPD, defined as post-BD FEV1/FVC<0.7, were tested with a 4-week treatment with formoterol and budesonide. The prevalence of COPD was 24.3% (95%, CI 22.3–26.4), with an overall underdiagnosis of 56.7%. After 4 weeks of treatment, 16% of initially obstructed patients had normal spirometry; in addition, 15.6% of individuals with a diagnosis of COPD did not sustain airflow obstruction, while 84.4% did.

EPISCAN II has several strengths, as it is consistent and even surpasses previous studies in Spain (IBERPOC and EPISCAN),5,6 namely: By assessing all 17 regions in Spain, we were able to produce a complete map of the distribution of COPD and its determinants for the first time. We did not establish an upper age limit for surveying, as population growth and ageing, particularly affecting women, is a global trend42; high quality spirometry was applied to nearly 10,000 individuals, that was centrally monitored for quality, and adhered to the strictest international guidance and protocols, as with all other measurements. Indeed, the oldest participant in EPISCAN II was a woman aged 98 years old from Extremadura, with spirometry quality graded B. Given the large sample size and high response rate, it can be considered that the final sample of participants is representative of the Spanish population older than 40 years.

However, a number of limitations are worth considering: Most areas surveyed were urban, and although 90%+ of the Spanish population live in cities, the unmet burden in rural areas is expected to be high but under-represented in here.2 Like all COPD epidemiological studies, diagnosis is based solely on spirometry (airflow limitation) and this may cause other obstructive diseases than COPD itself to be included here; same token, there is a high proportion of participants (7.9%) with a self-reported diagnosis of asthma. At the population level there are no studies that quantify the so-called ‘clinical COPD’, as the thresholds to determine which symptoms and which exposures define it are even more diffuse than those of airflow limitation. Therefore, for consistency with the design of other epidemiological studies on COPD internationally (BOLD,8 etc.) and the previous ones in Spain (IBERPOC,5 EPISCAN6) the same design is maintained.

Better defining and grading obstructive airway diseases, and COPD in particular, is a long quest.43 The GOLD initiative has been modifying definitions, thresholds of spirometry and staging classifications in its several iterations.2 New, recent evidence from COPDGene proposed the combination of four COPD domains: environmental exposure (cigarette smoking), clinical symptoms (dyspnoea and/or chronic bronchitis), chest CT imaging abnormalities, and abnormal spirometry; this new staging was strongly associated with spirometric disease progression (FEV1>350ml loss over 5 years), and all-cause mortality.44 We strongly believe that the COPDGene proposal of possible, probable, and definite COPD will help to advance COPD research and medicine, just as a similar proposal with symptoms, EKG, enzymes helped ischaemic heart disease previously.

We conclude that COPD remains a major, common cause of disease in the general population, yet with substantial variability by age, sex, and geography, among other determinants. There is significant unmet need of airflow limitation at the individual and population levels. Last but not least, efforts to reduce COPD underdiagnosis, and the toll of cigarettes and of other old and new ways of smoking, should be streamlined.