Introduction Chronic obstructive pulmonary disease (COPD) is a chronic respiratory disease characterized by a decline in lung function over time and accompanied by respiratory symptoms, primarily dyspnea, cough, and sputum production.1 Consequently, COPD is associated with a significant economic burden, including hospitalization, work absence, and disability.1 Current data suggest that COPD mortality is increasing, and by 2020, COPD is predicted to be the third-leading cause of death worldwide.2 The severity of COPD can be determined and classified by different methods. Incidence and prevalence estimates differ greatly, depending on the methods used for diagnosis and classification. It is important to understand the true epidemiology of COPD to monitor trends over time and to determine the effectiveness of potential treatments or preventive measures. The objectives of this study were to conduct a structured, comprehensive literature review to identify articles on the epidemiology of COPD in eleven developed countries (Australia, Canada, France, Germany, Italy, Japan, The Netherlands, Spain, Sweden, the United Kingdom, and the United States of America [USA]); quantify the burden of illness of COPD in terms of incidence, prevalence, and mortality; identify trends in these data over time; and identify any trends regarding age, sex, and/or disease severity. Methods A structured and comprehensive search of medical literature indexed in the electronic PubMed (http://www.ncbi.nlm.nih.gov/sites/entrez) and EMBASE (http://www.embase.com/info/accessing-embase) databases was conducted using a detailed search strategy with a combination of free-text search terms and medical subject headings. Search terms included terms related to COPD, chronic bronchitis, and pulmonary emphysema, and terms for epidemiology including incidence, prevalence, rate of mortality, and risk of dying (see Table S1). The search was restricted to articles in English published between January 2000 and September 2010. Articles identified from each literature search were screened in two phases by one reviewer using predefined inclusion and exclusion criteria. Phase 1 involved reviewing all titles and abstracts to determine whether to include or exclude them, and Phase 2 involved reviewing the full text of the articles identified in Phase 1 to determine their inclusion or exclusion for data extraction. Articles were included if they reported incidence, prevalence, and/or mortality in COPD, or trends in such data for at least one of the countries of interest (Australia, Canada, France, Germany, Italy, Japan, The Netherlands, Spain, Sweden, the UK, or the USA). Articles were excluded if they met at least one of the following exclusion criteria; that is, if the article: was a comment, an editorial, a letter, a case report, or a clinical trial; did not report data specifically for COPD; did not report data on incidence, prevalence, and/or mortality, or trends in such data; was not concerned with any of the countries of interest; focused on a limited population, including studies in small numbers of patients, patients in very limited sub-populations, such as patients who were hospitalized, and patients with an existing condition that increased their risk for COPD, or studies that investigated risk factors for COPD; reported a study conducted in a single site, clinic, hospital, or city; focused on comorbidities in patients with COPD; or reported incidence, prevalence, or mortality associated specifically with exacerbations of COPD, not COPD overall; reported incidence or prevalence estimates from a model (ie, the article was not the primary data source); reported on design of a study but did not report results; was a duplicate of an article that had been previously identified. Inclusion and exclusion processes were documented fully, and a Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart was completed.3 Relevant data were extracted from the included articles into evidence tables for each country. Quality-control checks verifying the summarized data against the source articles to confirm correct extraction were performed by an independent quality-control specialist on all extracted data. Results Summary of identified studies The PRISMA flow chart (Figure 1) presents the two-phase screening approach, and the number of articles included, and excluded at each phase. From the initial database searches, 2838 unique articles were identified of which 299 articles were retrieved for full-text evaluation. Of those, 133 were included for data extraction. Overall, the greatest number of relevant articles was identified for the USA (n = 49), Sweden (n = 19), and Canada (n = 12) (see Table S2). A total of 19 articles were identified that reported data for more than one country (“multicountry” studies). Most articles (80) focused on prevalence of COPD; another 15 articles reported incidence, and 58 reported mortality associated with COPD (Table S2). Twelve articles reported trends in incidence and/or prevalence, whereas 25 articles reported trends in mortality. Prevalence The reported prevalence of COPD ranged from 0.2% in Japan to 37% in the USA, but this varied widely across countries and populations, by diagnosis method, and by age group analyzed. Table 1 presents those studies that measured COPD by multiple methods within the same population to compare prevalence estimates resulting from different methods. Prevalence estimates varied according to the method of diagnosis and classification of COPD.4–7 When individuals were identified by spirometry, and classified using the 2001 Global Initiative for Chronic Obstructive Lung Disease (GOLD) criteria for COPD (forced expiratory volume in 1 second/forced vital capacity [FEV1/FVC] < 0.70), a greater COPD prevalence was reported than when using other classification methods such as the British Thoracic Society (BTS), European Respiratory Society (ERS), American Thoracic Society (ATS) spirometric, or ATS clinical criteria.4–6,8,9 This was supported by information from other studies that found that prevalence estimates by spirometry were higher than those estimated using methods based on symptoms (Table 1).5,6,10–16 Some multicountry studies reported similar findings when looking at data from several countries, reporting a greater prevalence of COPD diagnosed by spirometry compared with self-reporting (see Table 1). COPD was more commonly reported in older populations and was most prevalent in adults aged 75 years and older. Overall, the studies showed that the prevalence of COPD has increased over time, although the rate of increase has declined in recent years, particularly among men. Details of all studies providing prevalence data are given in Table S3 in the supplementary material. Incidence Table 2 presents a summary of the population-incidence data reported in the identified articles. The incidence of COPD varied greatly between countries, but it is difficult to compare estimates because they are reported in different units and over different lengths of time. In most of the studies, the incidence of COPD was greater in men than in women.17–21 The incidence of COPD was also greater in older individuals, particularly in those aged 75 years and older.15,21 Six articles reported trends in incidence over time for Australia, Canada, Sweden, and the USA.15,18,22–25 Although COPD incidence has increased over the last 20 years, within the last 10 years, there has been an overall decrease. Studies in Canada18 and the USA25 reported that trends in incidence over time were similar between men and women; however, in Australia, COPD incidence decreased in men between 1998 and 2003 but increased in women.22 Two articles, both conducted in Sweden as part of the Obstructive Lung Disease in Northern Sweden (OLIN) study, reported incidence rates in smokers (Table 2).20,26 These studies reported a two- to three-times greater incidence in smokers than nonsmokers when measured by spirometry, and assessed by GOLD or BTS criteria.20,26 One study also reported that COPD incidence in former smokers was more than double that in nonsmokers.26 Mortality The 58 articles that presented mortality associated with COPD varied in the way they reported the data. Twenty-four articles reported the mortality rate within a group of patients with COPD, 14 reported the proportion of all deaths that could be attributed to COPD, and 21 articles reported overall mortality from COPD within the whole population. Of the studies that reported mortality rates within patients with COPD, length of follow-up differed, which resulted in difficulties comparing studies. However, the one-year mortality rate of COPD (all severity stages) was reported in four studies and varied from 4.1% in patients aged 45 years and older, to 27.7% in patients aged 65–100 years in Canada,18,27,28 and to 5.1% in patients aged 41–83 years in Sweden.29 Between 2.3% and 8.4% of all deaths were caused by COPD, and this proportion was greater in men than women,30–32 and greatest in subjects aged 65–74 years.33 Measuring the number of COPD deaths per whole population provides a true picture of the burden of COPD mortality within the population. The overall mortality rate varied between countries, ranging from 3–9 deaths per 100,000 population in Japan to 7–111 deaths per 100,000 population in the USA. In almost all these studies, COPD mortality was greater within the male population than within the female population15,34–45 and was greatest in elderly adults aged 75 years and older.15,35–38,43 Two studies were identified that reported deaths due to COPD as a proportion of deaths attributable to smoking: numbers ranged from 12.8% across several industrialized countries46 to 20.9% in the USA.47 One study also reported that 19%–24% of all smoking-related deaths in women and 52%–54% of all smoking-related deaths in men resulted from COPD.48 One US study reported that mortality in a population of those who quit smoking was almost half of that in a population of individuals who switched from cigarette smoking to spit tobacco (49 versus 89 per 100,000 population).49 Trends in mortality A total of 25 articles reported COPD mortality over different years to allow trends to be observed, 14 of which reported the changes in COPD mortality within the overall population. These included studies conducted in Australia (2), Canada (1), France (1), and the USA (10) (Table 3). Our literature review did not identify any articles reporting trends in mortality in Germany, Italy, Japan, The Netherlands, Spain, Sweden, or the UK. In general, the studies reported an overall increase in COPD mortality rates within the last 30–40 years, with a much greater increase in mortality in women compared with men.15,34,35,38,40,42,45 Some studies have indicated that more recently (within the last 10 years) mortality rates have increased at a slower rate or have decreased, particularly in men.22,34,35,42,43,45 Some remarkable differences in COPD mortality exist between countries, particularly regarding the differences between men and women. In Australia, one study34 reported a decrease in COPD mortality in men between 1979 and 1997, whereas an increase was seen in women over the same period. In France, COPD mortality has increased in women over time, whereas a decrease has been reported in men.35 Data from several US studies show more heterogeneity. Data from two studies showed a clear increase in COPD mortality in women and only a slight increase in men between 1980 and 2000.15,45 Data from a later study43 suggested that COPD mortality decreased between 2000 and 2005 in men, with little change in women. Discussion We conducted a structured and comprehensive literature review to identify published data on the prevalence, incidence, and mortality in COPD, and/or trends in those data. The review identified a wealth of data on the prevalence of COPD in the eleven countries studied (Australia, Canada, France, Germany, Italy, Japan, The Netherlands, Spain, Sweden, the UK, and the USA). However, data on mortality and incidence were sparser. Only 15 articles reported incidence data, and six reported trends in incidence; 21 articles reported mortality from COPD within the whole population, and 14 of those reported trends in those data. Several other literature reviews have previously been conducted to identify prevalence and/or mortality data.50–53 One of these reported data only for the Asia-Pacific region and, of those countries investigated here, included only Japan.53 Results from the other three literature reviews can be compared with findings from our review. One review included articles published between 1962 and 2001 that were indexed on MEDLINE,51 one review included articles published between 1990 and 2004 that were indexed on PubMed, and also provided pooled estimates of prevalence by means of a meta-analysis,52 and the third review included articles reporting prevalence, and/or mortality in Europe published between 1991 and 2009 in the Science Citation Index database via the Web of Science.50 As with our study, all three published reviews reported substantial heterogeneity between studies, particularly in terms of the definition of COPD used, methods used (eg, self-report, spirometry), diagnostic criteria (eg, GOLD, ATS), populations studied, and year(s) of study.50–53 The estimates obtained from the multicountry studies in our review ranged from 3.6%–10.1%, which is in line with the estimates reported in two of the previous reviews (4%–10%,51 9%–10%52). When all studies in our review were taken into account, prevalence estimates ranged from 0.2%–37%, which was in line with the most recent published review (2.1%–26.1%50). Differences can be accounted for by the wider scope of our study, which identified 80 studies reporting prevalence estimates in Europe, the USA, Canada, Australia, and Japan compared with 32 studies reporting estimates for Europe only, as identified by Atsou et al.50 Our findings with respect to mortality were also similar to those reported in a recent literature review regarding both mortality within the overall population (3–111 per 100,000 [current review] versus 7.2–36.1 per 100,000 [review by Atsou et al50]) and the greater mortality rate in men compared with women.50 The slightly higher mortality rates identified in our studies again relate to the scope of the two reviews. The lowest and highest mortality estimates in our review were from Japan and the USA, respectively,38,54 which were not captured in the European-focused literature review.50 Therefore, it is likely that the inclusion of countries outside Europe led to the greater heterogeneity in estimates that were identified in our review. The current review also reported that, although COPD mortality rates have increased over time, rates have declined in more recent years, which suggests improvements in COPD management. However, several studies identified within the review also reported that the mortality rate in women with COPD has increased or stabilized, whereas it has decreased in men. The difference in these trends may be explained by trends in smoking prevalence in the countries of interest. A relationship between smoking and COPD mortality can be investigated by examining trends in smoking prevalence such as using data from the Organisation for Economic Co-operation and Development (OECD).55 We were specifically interested in those countries where a difference in COPD mortality trends was observed between men and women (ie, Australia, France, and the USA). These countries all showed an overall decline in smoking rates with the greatest prevalence in men.55 Recently, the discrepancy in smoking rate between men and women has reduced because the rate in men has declined at a much greater rate than in women. In Australia,34 COPD mortality between 1979 and 1997 followed a pattern similar to that observed in smoking prevalence between 1965 and 1980, with a decrease in men and an increase in women. The mortality data mirrored the smoking patterns with a delay of 15–20 years in women and 20–25 years in men. This “lag time” between smoking and COPD onset has been reported in previous literature.46 In France, both smoking prevalence and COPD mortality have increased over time, whereas a decrease in smoking prevalence and COPD mortality has been reported in men.35 Smoking prevalence data in France were not available from the OECD before 1981, which made it difficult to determine whether a lag time between smoking and COPD onset occurred. However, COPD mortality data from US studies show more heterogeneity; smoking prevalence substantially decreased over time in both men and women, whereas COPD mortality increased to a greater extent in women than men between 1980 and 2000, after which a decrease was observed in men, and a plateau in women between 2000 and 2005. Although smoking prevalence might explain some of the discrepancy between men and women in COPD mortality, other reasons must be considered as well. Recent evidence suggests that women younger than 55 years are significantly more susceptible to severe COPD than men.56 Furthermore, women tend to have smaller airways and lung volumes than men,57 and previous studies have shown that females are consequently more vulnerable to the adverse effects of smoking than men.58–60 As with all literature reviews, both the current review and the data identified had certain limitations. First, this review focused on only eleven countries of interest (Australia, Canada, France, Germany, Italy, Japan, The Netherlands, Spain, Sweden, the UK, and the USA). Although the literature search itself was not restricted to certain countries, articles related only to countries outside those of interest were excluded from the review during the screening process. Second, the search was limited to articles published in English, so we may not have identified relevant articles published in other languages, particularly those relating to the non–English-speaking countries of interest. Third, several articles did not report true population-based estimates of prevalence or incidence, but instead reported prevalence or incidence of COPD within a population at increased risk for the condition. Fourth, and as with similar reviews involving searches of literature databases, any articles that were not indexed in PubMed or EMBASE would not have been initially identified. Fifth, the studies varied widely in the ages of populations studied, so they were difficult to compare and to draw conclusions from overall. Finally, differences between countries in terms of COPD diagnosis and management will also lead to discrepancies and hinder meaningful comparisons across countries. However, our review has certain strengths when compared with other similar literature reviews in the epidemiology of COPD. Our review was a comprehensive literature review that identified literature from the MEDLINE and EMBASE databases. Furthermore, we investigated data on prevalence, incidence, and mortality as well as trends in prevalence, incidence, and mortality. Our review included more recent data (published from January 2000 to September 2010) compared with the previous reviews.51,52 Also, compared with the most recent review, which only reviewed data from countries in Europe,50 our review considered data from Australia, Canada, Japan, and the USA as well as from European countries. Consequently, we anticipate that our review contains more complete epidemiology data that present a current picture of the burden of COPD in major developed countries. Although our review reported an overall decrease in the burden of COPD, in incidence, prevalence, and mortality in certain countries in recent years,18,22,25,26,31,61,62 COPD remains a substantial health problem throughout the world. We found that several data gaps exist within the current literature on the epidemiology of COPD, particularly regarding studies reporting the incidence of COPD or trends in mortality data. Also, no studies were identified that reported incidence or trends in incidence in France, Germany, Italy, Spain, and The Netherlands, or trends in overall mortality in Germany, Italy, Japan, The Netherlands, Spain, Sweden, or the UK. A need exists for studies in these countries to examine trends in COPD incidence and mortality to fully understand the true burden of COPD in the population. There is also a need to continue to improve uniformity in definitions and methods of diagnosis to improve understanding of the burden of disease and aid in clearer evaluation of the patient response to treatment.