Design, setting and participants: A cross-sectional study of people in the community aged ≥ 40 years, selected at random using electoral rolls, in six sites chosen to reflect the sociodemographic and geographic diversity of Australia, conducted between 2006 and 2010. Standardised questionnaires were administered by interview. Forced expiratory volume in 1 second (FEV1), forced vital capacity (FVC), and the FEV1/FVC ratio were measured by spirometry, before and after bronchodilator administration.
Results: Complete data were available for 1620 men (participation rate, 26%) and 1737 women (participation rate, 28%). The prevalence of GOLD Stage II or higher COPD (defined as post-bronchodilator FEV1/FVC ratio < 0.70 and FEV1 < 80% predicted) was 7.5% (95% CI, 5.7%–9.4%) among people aged ≥ 40 years, and 29.2% (95% CI, 18.1%–40.2%) among those aged ≥ 75 years. Among people aged ≥ 40 years, the prevalence of wheeze in the past 12 months was 30.0% (95% CI, 27.5%–32.5%), and prevalence of shortness of breath when hurrying on the level or climbing a slight hill was 25.2% (95% CI, 22.7%–27.6%).
Conclusions: Symptoms and spirometric evidence of COPD are common among people aged 40 years or older and increase with age. Further research is needed to better understand the diagnosis and management of COPD in Australia, along with continuing efforts to prevent the disease.
Several recent expert summaries have highlighted the health impacts of chronic obstructive pulmonary disease (COPD).1-4 In Australia, there were 4761 deaths (4% of all deaths) attributed to COPD in 2006,5 and 47 207 years of life were lost due to COPD in 2003.6 In 2006–07, there were 52 560 hospital separations in Australia attributed to COPD, with an average length of stay of 7 days.5 COPD has a substantial impact on mortality and health service use.
Data on the prevalence of COPD and related symptoms are limited, with estimates ranging from 1.4% to 6.9%, depending on the age group studied and the definitions used.5,7,8 COPD is usually not diagnosed until it is moderately advanced and begins to impair quality of life. Furthermore, due to poor utilisation of spirometry in primary care settings9-11 and the largely silent nature of the disease in its early stages, COPD is under-recognised by doctors and underreported by patients. Surveys that have used objective measurement of lung function to identify COPD have found a high proportion of previously undiagnosed cases.12
Valid estimation of the prevalence of COPD requires a comprehensive, nationwide, population-based survey, including high-quality post-bronchodilator spirometry, conducted in a representative sample of the population.4 In collaboration with the international Burden of Obstructive Lung Disease (BOLD) study,13 we conducted this research to describe the prevalence of obstructive lung disease, including symptoms, diagnoses and level of airflow obstruction, in people aged 40 years or older in Australia.
We conducted the survey in a representative sample of adults aged ≥ 40 years living in six locations around Australia. One centre with a large Indigenous population was included. Post-hoc weights were used to make inferences about prevalence in the entire Australian population aged ≥ 40 years.
The study commenced in Sydney in 2006, as part of the international BOLD study. The protocol13 and main results14 for Sydney have already been published, together with data from centres in 11 other countries. Data from Sydney are included here, along with data from the remaining five Australian locations. The protocol used in this study closely followed that used in the global BOLD study.
The study was approved by the Human Research Ethics Committee of the University of Sydney (ref. no. 12-2006/9724). Additionally, all sites obtained local ethics approval. All participants gave written informed consent.
Study participants were sampled from electoral rolls, after excluding those who were institutionalised, using a sex-stratified, simple random sample in all sites except Broome and Busselton (Appendix 1).
As many of the Aboriginal and Torres Strait Islander residents of Broome are not on the electoral roll, the sampling frame at this site was established using a household census. We then recruited a stratified random sample of Indigenous and non-Indigenous male and female residents aged ≥ 40 years from this sampling frame. In Busselton, we followed a two-stage sampling strategy. A sample was recruited from the electoral roll for the Busselton Health Study,15 and study participants were then selected using sex-stratified random sampling from among those who participated in the Busselton Health Study.
All selected individuals in study centres other than Broome initially received a letter inviting them to participate and, if interested, to telephone for an appointment. After 2 weeks, study staff telephoned those individuals who had not made contact and made several further attempts to contact them by telephone or mail.
We used the BOLD study questionnaire13,14 without modification in all study centres, except for Aboriginal and Torres Strait Islander participants in Broome, whose first language was often not English. The questionnaire was administered by interview.
We performed spirometry testing, using the EasyOne spirometer (ndd Medizintechnik), before and 15 minutes after administration of salbutamol 200 μg via metered dose inhaler and spacer.13 Participants were asked to refrain from using their bronchodilator inhaler during the 6–12 hours before testing.
All spirograms were reviewed by one of us (D P J) and assigned a quality score based on published acceptability and repeatability criteria.16
The highest recorded forced expiratory volume in 1 second (FEV1) and forced vital capacity (FVC) from acceptable trials16 were used in the analysis. Participants with a post-bronchodilator FEV1/FVC ratio < 0.70 were classified as having COPD based on Global Initiative for Chronic Obstructive Lung Disease (GOLD) spirometric criteria.4 Among these, participants with FEV1 ≥ 80% predicted17 were classified as having GOLD Stage I COPD; those with FEV1 ≥ 50% to < 80% predicted as GOLD Stage II; those with FEV1 ≥ 30% to < 50% predicted as GOLD Stage III, and those with FEV1 < 30% predicted as GOLD Stage IV. In addition, those with a post-bronchodilator increase in FEV1 that was both ≥ 12% of the pre-bronchodilator FEV1 and ≥ 200 mL were classified as having reversible spirometry consistent with asthma.
The sample size at each site of 300 men and 300 women was designed to achieve a 95% confidence interval half-width of between 2.5% and 5.3% for prevalence rates ranging from 6% to 33% within each study centre.13 In most centres, an initial sample of 2400 was drawn, and additional samples were drawn from the sampling frame until the required sample size was achieved.
As this was not a simple random sample of the Australian population aged ≥ 40 years, we used post-hoc weights to estimate prevalence rates for the Australian population after adjustment for age, sex, socioeconomic status (using Socio-Economic Indexes for Areas [SEIFA]18), remoteness (using the Accessibility/Remoteness Index of Australia [ARIA]19), and Indigenous status. SEIFA and ARIA categories were assigned to individuals by postcode of residence. Weights were calculated by comparing the distribution of the study participants according to these characteristics with the distribution of the Australian population aged ≥ 40 years, from the 2006 Australian Census of Population and Housing, obtained from the Australian Bureau of Statistics.
Prevalence estimates and associated 95% confidence intervals were calculated for participants who completed the core questionnaire and had acceptable post-bronchodilator spirometry data. Prevalence was estimated, taking account of the stratified nature of the sample and including the post-hoc weights, using Proc Surveyfreq (SAS 9.2; SAS Institute). Study site was the stratifying variable.
Overall, the initial study sample for all sites included 6328 men and 6198 women aged ≥ 40 years (Appendix 2). After excluding people who were known to have permanently left the study area, or who were deceased, aged < 40 years, institutionalised, or for whom a registered address did not exist, there were 5352 eligible men and 5408 eligible women. Core questionnaire and acceptable post-bronchodilator spirometry data were available for 1620 men (participation rate, 26%) and 1737 women (participation rate, 28%). After excluding individuals who were not contactable, the adjusted participation rates were 35% for men and 37% for women, for all sites except Broome. Data on age, respiratory diagnoses and smoking status were available for an additional 893 men (14% of initial population) and 1111 women (18% of initial population) who had declined to participate.
Participants who completed the core questionnaire and had acceptable post-bronchodilator spirometry data were younger and more likely to report a history of emphysema, asthma, asthmatic bronchitis, chronic bronchitis or COPD than those for whom only minimal data were available (Appendix 3). Men with complete data were less likely to have ever smoked than men with minimal data, but the reverse was the case among women. Compared with the Australian population aged ≥ 40 years, those with complete data were less likely to be aged ≥ 75 years (among women), more likely to live in the most socioeconomically disadvantaged areas and remote areas, and more likely to be Aboriginal or Torres Strait Islanders (Box 1).
The weighted prevalence of respiratory symptoms, reported clinical respiratory diagnoses and spirometric diagnoses in the study sample, classified by age group and sex, are shown in Box 2. Wheeze and shortness of breath when hurrying on the level or climbing a slight hill were common symptoms. The prevalence of cough was similar among the age groups, but the prevalence of sputum production, consistent with a diagnosis of chronic bronchitis, increased with age. There was a steep increase in the prevalence of shortness of breath in people aged ≥ 75 years. About half the study population reported ever having smoked cigarettes. A reported diagnosis of asthma or related illness was much more common than a reported diagnosis of COPD or related illness (18.8% v 5.2% among all people aged ≥ 40 years), even in the oldest age group. Among all people aged ≥ 40 years, the prevalence of GOLD Stage II or higher COPD was 7.5% and the prevalence of severe COPD (GOLD Stage III or higher) was 0.9%; prevalences were higher in the oldest age group.
We found that the prevalence in Australia of GOLD Stage II or higher COPD, defined by spirometric criteria, is 7.5% among people aged ≥ 40 years and 29.2% among people aged ≥ 75 years. Smaller proportions of people reported having previously received a diagnosis of COPD (including chronic bronchitis or emphysema), but much higher proportions reported breathlessness on exertion.
The prevalence of GOLD Stage II or higher COPD among 11 international sites ranged from 8.5% in Reykjavik, Iceland, to 22.2% in Cape Town, South Africa, in men; and from 3.7% in Hannover, Germany, to 16.7% in Cape Town in women.14 Our estimates for Australia, 6.9% and 8.1%, respectively, therefore lie at the lower end of the international range for men and in the middle for women.
These estimates are not directly comparable with other Australian estimates of the prevalence of COPD. The 2004–05 National Health Survey reported that 2.8% of Australians aged ≥ 18 years self-reported a diagnosis of COPD, chronic bronchitis or emphysema.5 We found that 5.2% of people aged ≥ 40 years reported having received these diagnoses. The difference may be attributable to the different ages of the survey populations. This may also explain why the prevalence in the Northwest Adelaide Health Study, which included people aged ≥ 18 years, was lower than in our study.8
The main strengths of our study are the nationwide study population, the large sample size leading to relatively precise estimates, the use of objective measures with careful attention to quality control of spirometry, and the application of standardised definitions. Harmonisation of the study protocol with a widely used international protocol allows comparison with many reference populations.
The main limitation was the poor overall response rate, which introduces the possibility of selection bias affecting the prevalence estimates. Participants who provided complete data, and hence contributed to the prevalence estimates, were slightly younger but were more likely to self-report a diagnosis of COPD than those who provided only minimal data. Although this implies that people with COPD may be overrepresented in the study sample, we do not have information on those who declined to provide any information, and we should not assume that they were similar to those who provided minimal data.
The study population was not a simple random sample of the Australian population because the six study centres were not randomly chosen. The study centres were deliberately selected to provide adequate representation of the sociodemographic and geographic diversity of Australia. Hence, rural, Indigenous and disadvantaged areas were deliberately oversampled. Post-hoc weights, based on the Australian census, were intended to simultaneously adjust for the non-random selection of sites and for bias attributable to non-response among eligible participants. This strategy adjusts for non-representativeness with respect to age, sex, socioeconomic disadvantage and remoteness. However, it is possible that there were other, unmeasured biases relevant to the prevalence of COPD.
We defined and classified COPD according to the internationally agreed GOLD guidelines4 that have also been adopted by the BOLD study.13 There has been substantial debate around these definitions.21-23 This is because spirometric function is measured on a continuous scale, and any binary classification based on cut-points is essentially arbitrary. Lung function declines with age, and there is debate about the extent to which this is “normal” and hence should be incorporated into reference equations, and consequently used to discount the increasing prevalence of COPD in older people. We have dealt with this issue by reporting the prevalence of a range of measures, both objective and subjective, to encapsulate the spectrum of obstructive lung disease and related illness in the community.
The relationship between asthma and COPD in older people is complex.24 In our study population, 5.8% of people had bronchodilator reversibility, which may be a feature more consistent with asthma than COPD, although it does not exclude the latter diagnosis. However, the prevalence of bronchodilator reversibility was higher in men than in women, which is the reverse of the difference in prevalence of self-reported asthma in Australia.25 While asthma and COPD may coexist in individuals, it is not possible to deduce any causal relationship between them in cross-sectional studies such as this.26
The findings of our Australian BOLD study have important implications for health service development in Australia. Only by accurately diagnosing COPD is it possible to offer the range of interventions that have been demonstrated to improve quality of life, reduce disability and limit health care use. The finding that many participants with confirmed airflow obstruction consistent with COPD did not have a pre-existing diagnosis suggests greater effort is needed in making high-quality spirometry available in all health care settings. The prevalence findings will also be integral in informing the ongoing development of treatment services for people living with COPD. This includes access to smoking cessation programs, pulmonary rehabilitation, lung volume reduction and lung transplantation.
COPD is a common and serious health problem, particularly among older people, with major impacts on health resources and personal and community costs. It is often not recognised, in part because the diagnosis requires spirometry, a procedure that is not widely used in primary care.25 There is a need for further research to better understand the extent to which COPD is optimally recognised and managed in Australia and for continuing efforts to prevent the disease by avoidance of smoking and improvements in ambient and occupational air quality.
1 Demographic characteristics of the study sample with complete data compared with the Australian population aged ≥ 40 years in the 2006 census
2 Weighted prevalence of respiratory symptoms and illnesses and spirometric diagnoses, by age group and sex*