This was a cross-sectional study conducted in urban populations in the 2 largest islands of the archipelago: Gran Canaria and Tenerife (Hospital Universitario de Gran Canaria Dr. Negrín and Hospital Universitario de Canarias). The participating cities in Gran Canaria were Las Palmas de Gran Canaria, Arucas and Gáldar, with a population of 444526 inhabitants, while in Tenerife, the city studied was La Laguna (152222 inhabitants). The total sample comprised of 596478 subjects, representing 27.2% of the population of the Canary Islands (2118519 inhabitants).11 The sample size was calculated using an estimated COPD prevalence of 10%, with a sample error of ±1.91% (95% confidence interval [CI]) with an estimated non-response rate of 30%. The minimum number of subjects for inclusion was 947. The first sample unit was constructed from the randomized selection of census tracts from the above-mentioned urban municipalities or metropolitan areas. A total of 69 census tracts (36 in Gran Canaria and 22 in La Laguna, Tenerife) were selected. Since each census tract covers a limited geographical area, the second sampling unit was defined as randomized routes within the limits of each census tract, with a randomized starting point within each of the tracts. Finally, all subjects between 40 and 70 years of age residing in the dwelling and who met the inclusion criteria were selected. If the selected subject(s) were not at home at the time of the first visit, the visit was repeated up to 3 times. The exclusion criteria included mental illness, incapacity to complete the questionnaire or spirometry and institutionalization in healthcare facilities or prisons. The study was approved by the ethics committees of both the Hospital Universitario de Gran Canaria Dr. Negrín and the Hospital Universitario de Canarias, and all participants signed informed consent before starting the study.

Subjects who agreed to participate completed an extensive questionnaire in their homes. This questionnaire included questions on demography, smoking habit, biomass exposure, education, comorbidities, respiratory symptoms and quality of life. The same questionnaire was used in the PLATINO study (The Latin American Project for the Investigation of Obstructive Lung Disease) and is an amalgam of several previously validated questionnaires.12 Interviewers were trained professionals not involved in healthcare activities, and their activity was submitted to quality control checks by qualified inspectors who reviewed all the interviews carried out. Subjects who refused to participate in the study were asked to complete a short questionnaire that included questions on demography (age, sex) and smoking habit.

Subjects who completed the long questionnaire were given an appointment for spirometry within the next 15 days in one of the participating hospitals. If for any reason the subject could not attend the hospital, spirometry was performed at home. Spirometries were carried out by a pulmonologist or by nurses with over 15 years’ experience in performing lung function tests. The device used in both the hospital and in the patients’ homes was a portable spirometer (Easyone® NDD, Medical Technologies, Chelmsford, Massachusetts and Zurich, Switzerland). This same spirometer was used in both the PLATINO study and in the BOLD study (The Burden of Obstructive Lung Disease), and has been internationally validated.13,14 The ERS/ATS recommendations for standardized spirometry and the ATS/ERS 2004 recommendations (coinciding with those from the GOLD document)15,16 were used for the diagnosis and classification of the severity of obstruction. All spirometries were reviewed and interpreted by 2 pulmonologists participating in the study.

COPD was diagnosed when the FEV1/FVC ratio after the bronchodilator test was less than 0.7. The bronchodilator test was carried out with a spacer device administering 200μg salbutamol, and the spirometry was repeated after 15 minutes. COPD patients who had never been previously diagnosed with emphysema, chronic bronchitis or COPD were considered underdiagnosed. Undertreatment was defined as treatment for subjects with a diagnosis of COPD who had never been advised to quit smoking or given medication for their disease. An ex-smoker was defined as someone who had given up smoking at least 6 months before the interview, and never smokers were those who had smoked less than 0.1 pack years.

COPD prevalences were calculated as percentages with 95% CI and were standardized for age and sex with the reference Canary Island population.11 This was subsequently standardized with the world population.17 Calculations were adjusted taking into account that a lower percentage of smokers participated in the spirometry. Similarly, prevalence was estimated according to the categorical variables of sex, age groups, body mass index (BMI), smoking habit, geographical area, social stratum, respiratory diseases in childhood and adolescence, family history of respiratory diseases, educational level and environmental exposure (other than tobacco smoke). COPD risk factors were analyzed using a logistic regression model. A P-value of <.05 was considered significant. The SPSS 15.0 statistical software package was used (SPSS Inc., Chicago, U.S.).

The final sample included 1353 subjects, of whom 1001 completed the long questionnaire and performed spirometry correctly (Fig. 1). The total non-response rate was 26% (31% in Tenerife and 28.1% in Gran Canaria, P<.05). Regarding spirometries, 95% were carried out in the hospitals (52 tests were carried out in the home by the same nurses, using the same spirometer and the same technique). Of the 33 invalid study spirometries, only 5 were carried out at home. The mean age of the subjects who refused to participate in the study was lower than that of the subjects who performed the spirometry (53.5±8.5 vs 55.1±8.6 years, P<.05). The characteristics of the subjects who refused to participate and of those who completed the questionnaire and successfully performed spirometry are summarized in Table 1.

Fig. 1.

Study flow chart, showing the stages between the sample of all eligible subjects and the final sample in which the prevalence of COPD was established.

(0.11MB).

The proportion of smokers was 29.4% (95% CI: 25.4–33.1), while 26.9% of the subjects were ex-smokers (95% CI: 23.4–29.2) and 43.8% were never smokers (95% CI: 40.4–49.2). Significant differences between males and females were observed in all groups, whether smokers (35.1% vs 25.4%, P<.001) or ex-smokers and never smokers (P<.001). Of the subjects who refused to participate, 68.2% were smokers or ex-smokers, while 56.2% of smokers or ex-smokers finally carried out the spirometry.

The raw prevalence of COPD was 7.4% (8.4% in Gran Canaria [95% CI: 6.4–10.6] and 6.1% in Tenerife [95% CI: 3.8–88], P=.162). Prevalence adjusted for age, sex and smoking status was 7.3% (95% CI: 5.5–9.5). The proportion of males diagnosed with COPD was greater than females (8.7% [95% CI: 5.8–12.7] vs 6.3% [95% CI: 4.7–8.4], P=.134), but no significant differences were observed when a logistic regression model analysis adjusted for age and smoking habit was made (Table 2). Prevalence adjusted to the world population was 6.4% (95% CI: 4.8–8.6). Of the 73 subjects with COPD, 16 had a positive bronchodilator test without full reversal of the obstruction (22% of those diagnosed). A previous diagnosis of asthma was found in 5% of the smoker/ex-smoker group and in 50% of the never smokers. Prevalence varied depending on smoking status, from 2.5% (95% CI: 1.2–5.2) in never smokers up to 14.3% (95% CI: 9.7–20.6) in current smokers; the prevalence also increased in line with increases in the pack year index (PYI): from 1.7% (95% CI: 0.9–3.2) in subjects with PYI<0.1 up to 20.5% (95% CI: 15.1–27.1) in those with PYI>30.

Diagnosis of COPD is clearly associated with age and prevalence increases in line with age, in both men and women, irrespective of smoking status. Among subjects with a diagnosis of COPD who had never smoked, only one was younger than 60 years. Most diagnosed subjects had grade iiobstruction according to the GOLD criteria (69.9%; 95% CI: 57.1–80.2), while the figures were lower for obstruction grades iii and iv (10.4%; 95% CI: 5.5–18.7, and 3.3%; 95% CI: 0.6–17.0, respectively). There were no non-smokers in this final group (Table 3). There is an inverse relation between prevalence of COPD and BMI (P=.001), irrespective of PYI, with the highest prevalence of COPD being observed in subjects with a BMI of between 20 and 24.9, and the lowest in those with a BMI greater than 30 (10.5%; 95% CI: 6.5–16.6 vs 3.0%; 95% CI: 1.5–5.9; P<.05). No subgroup analysis could be made in the subgroup with a BMI of less than 20 due to the small number of subjects. However, the trend clearly suggests that COPD is more prevalent in this group. Both age and PYI are significant risk factors (Table 4).

The percentage of underdiagnosis of COPD is below 71.6%, and is more common in the milder obstruction grades (GOLD 2010): 81.8% in stage I vs 33% in stage IV, P<.001). Undertreatment was found in 63.5%, with a greater proportion of more severe patients being treated than less severe patients (81% underdiagnosis in stage I vs 33% in stage IV, P<.001).