Abstract
Rationale: The new Global Initiative for Obstructive Lung Disease (GOLD) stratification of chronic obstructive pulmonary disease (COPD) into categories A, B, C, and D is based on symptoms, level of lung function, and history of exacerbations.
Objectives: To investigate the abilities of this stratification to predict the clinical course of COPD.
Methods: Two similar population studies were performed in an area of Copenhagen including 6,628 individuals with COPD.
Measurements and Main Results: The patients were monitored for an average period of 4.3 years regarding COPD exacerbations, hospital admissions, and mortality. The percentages of individuals experiencing a COPD exacerbation during the first year of observation were 2.2% in group A, 5.8% in group B, 25.1% in group C, and 28.6% in group D. One- and 3-year mortality rates were 0.6 and 3.8%, respectively, in group A, 3.0 and 10.6% in group B, 0.7 and 8.2% in group C, and 3.4 and 20.1% in group D. Groups B and D, characterized by a higher degree of dyspnea than groups A and C, had five to eight times higher mortality from cardiovascular disease and cancer than did groups A and C.
Conclusions: The new stratification performs well by identifying individuals at risk of exacerbations. Surprisingly, subgroup B, characterized by more severe dyspnea, had significantly poorer survival than group C, in spite of a higher FEV1 level. This subgroup warrants special attention, as the poor prognosis could be caused by cardiovascular disease or cancer, requiring additional assessment and treatment.
The Global Initiative for Obstructive Lung Disease (GOLD) launched the 2011 revision of the strategy document for diagnosis and management of chronic obstructive pulmonary disease (COPD) in December 2011. It is not known how this profound change in the stratification strategy works regarding prediction of the clinical course of COPD.
First, we find that the new, multidimensional GOLD A–D classification provides better diagnostic separation as it identifies more individuals at high risk of exacerbations than a classification entirely based on the level of FEV1. Second, we also find that survival is better in the more severe COPD group C (low lung function but less dyspnea) than in the less severe group B (much better lung function but more dyspnea). We show that this is most likely caused by comorbidities, in particular heart disease.
Chronic obstructive pulmonary disease (COPD) has emerged as the most important respiratory disease globally, and its prevalence and impact on the society are expected to increase even further in the years to come. The Global Strategy for the Diagnosis, Management, and Prevention of Chronic Obstructive Pulmonary Disease document, produced by the Global Initiative for Chronic Obstructive Lung Disease (GOLD), is one of the key inspirations for local COPD guidelines and thus the ideas expressed in such document are followed by thousands of practicing clinicians worldwide (1).
Stratification of patients according to severity of COPD is important to initiate relevant treatment and is one of the most crucial aspects of a management guideline. In the 2011 update of the GOLD document on COPD, a profound change in the stratification of the patients has been introduced (1). In the previous version, the classification of severity was mainly based on the FEV1 level, whereas the new stratification also includes the level of daily symptoms, in particular dyspnea, and the history of exacerbations (Figure 1) (1, 2). In the present study, we wanted to compare the abilities of the old stratification (GOLD 2007) and the new stratification (GOLD 2011) regarding predicting exacerbations, hospital admissions, and mortality. We used the data from two large population-based studies in the area of Copenhagen, comprising in total more than 60,000 individuals and including more than 6,500 subjects with COPD. Some of the results have been previously reported in the form of an abstract (3).
Figure 1. Combined assessment of chronic obstructive pulmonary disease (COPD) using Global Initiative for Obstructive Lung Disease (GOLD) based on FEV1, modified Medical Research Council (mMRC), and exacerbation history. CAT = COPD Assessment Test. Reproduced by permission from Reference 1.
[More] [Minimize]
In the present study we pooled data from two similar but independent studies: the fourth examination of the Copenhagen City Heart Study (CCHS) in 2001–2003 and the examination of the Copenhagen General Population Study (CGPS) in 2003–2010. These studies were approved by institutional review boards and Danish ethics committees (KF100.2039/91, H-KF01-144/01), and were conducted according to the Declaration of Helsinki. Written informed consent was obtained from all participants.
The CCHS is a prospective epidemiologic study initiated in 1976–1978. A sample of 19,698 subjects aged 20 to 100 years was selected at random from the national Danish Civil Registration System, after age stratification in 5-year age groups, from residents of inner Copenhagen, and 14,223 subjects participated in the initial survey at Copenhagen University Hospital (4). They were all reinvited to participate in later surveys along with additional subjects in the youngest age groups. A total of 6,237 attended the fourth survey in 2001–2003 (response rate, 50%) (5).
The CGPS is a prospective epidemiologic study that aims to recruit more than 100,000 subjects representative of the general population and to collect genotypic and phenotypic data of relevance to a wide range of health-related problems. It is designed almost identically to the CCHS. Recruitment began in 2003 and is still ongoing (response rate, 45%) (6).
In both studies, the participants filled in an extensive questionnaire concerning lifestyle and health topics, including medication use at home. A physical examination was performed, in which basic variables such as weight and height were measured. A spirometer (Vitalograph, Maids Moreton, Buckinghamshire, UK) was used in the CCHS and for the first 14,624 participants in the CGPS. An EasyOne diagnostic spirometer (ndd Medizintechnik, Zurich, Switzerland) was used with the remaining participants in the CGPS. Spirometry was performed in the standing position without the use of a nose clip. Three sets of FEV1 and FVC values were obtained and as a criterion for correct performance of the procedure at least two measurements differing by less than 5% had to be produced together with the correct visual appearance of the spirometry tracings. The highest obtained values for every single participant of both FEV1 and FVC were used. Only prebronchodilator measurements were available.
Individuals with COPD were identified according to international COPD guidelines, based on an FEV1/FVC ratio less than 0.7. In accordance with the GOLD document we excluded individuals with self-reported asthma (1). In the present analyses, we included only individuals at least 40 years of age. Among 5,919 participants in the CCHS and 55,731 participants in the CGPS we identified 6,628 individuals fulfilling the above-mentioned COPD criteria. They were subdivided according to the two GOLD stratifications, using the FEV1 level, expressed as a percentage of the predicted value (FEV1%pred):
GOLD1: Mild COPD: FEV1%pred ≥ 80
GOLD2: Moderate COPD: 80 > FEV1%pred ≥ 50
GOLD3: Severe COPD: 50 > FEV1%pred ≥ 30
GOLD4: Very severe COPD: 30 > FEV1%pred
The GOLD 2011 classification into groups A–D was, in addition to FEV1%pred, also based on the self-reported severity of dyspnea (modified Medical Research Council [mMRC]) and on the frequency of COPD exacerbations in the last year. To define exacerbations, we used information regarding hospital admissions because of COPD and the use of oral corticosteroids and antibiotics related to treatment of exacerbations. For each individual, we defined an observation window of 1 year before the examination date and analyzed the use of inhaled medications and contacts with the region’s hospitals during this period. This was done using the unique personal civil registration number assigned to all Danish inhabitants and by linking our database to two national registries: the Danish National Patient Registry, covering all hospital contacts in Denmark, and the Danish Registry of Medicinal Product Statistics, which contains information on all prescriptions dispensed in all Danish pharmacies (7, 8). We identified the relevant medications using the Anatomic Therapeutic Chemical (ATC) code (9). From the Danish National Patient Registry, we identified hospital admissions with a discharge diagnosis of COPD (International Classification of Diseases 10th edition: DJ41–44). An exacerbation of COPD was defined as a short course of treatment with prednisolone (up to 3 wk) alone or in combination with an antibiotic or an acute admission to hospital because of COPD. There were only a few subjects (n = 4) who received continuous treatment with oral prednisolone for COPD, and in these subjects exacerbations were defined by usage of antibiotics only, as it was difficult to estimate short-term changes in their use of oral corticosteroids. No individuals received continuous antibiotic treatment for COPD.
We also performed analyses of the subgroups of the GOLD C and D categories, as patients can be stratified into these categories through different scenarios:
FEV1 < 50%pred and fewer than two exacerbations in the previous year (scenario 1: C1 or D1)
FEV1 ≥ 50%pred and two or more exacerbations in the previous year (scenario 2: C2 or D2)
FEV1 < 50%pred and two or more exacerbations in the previous year (scenario 3: C3 or D3)
Thus, we subdivided categories C and D into subgroups C1, C2, C3, D1, D2, and D3, defined above, depending on which scenario was causing the individuals to be in category C or D.
The presence of ischemic heart disease and previous myocardial infarction was determined by means of a questionnaire and by using information from the Danish National Patient Registry.
After the examination date, the participants were monitored by means of the registries up to 8.9 years, with an average of 4.3 years.
The following end points were defined:
COPD exacerbation (defined on the basis of the medication used and hospital admissions)
Admissions to hospital (due both to COPD and with other diagnoses)
All-causes mortality until August 17, 2010
Cause-specific mortality was determined according to the 10th classification of the International Classification of Diseases (ICD): respiratory deaths (J00–99), cardiovascular deaths (I00–99), cancer deaths (C00–D09) until December 31, 2009.
We also report the future prevalence of treatment with inhaled medications.
All analyses were performed with R version 2.13.1 (R Foundation for Statistical Computing, Vienna, Austria). For demographics, analysis of variance was used for continuous variables and chi-square tests were used for categorical variables. To account for censoring of data, the Kaplan-Meier estimator was used to estimate the percentage of events during follow-up. Hence, in Tables 3 and 5 all percentages are 100% minus the Kaplan-Meier estimate of being event-free. The log-rank test was used to compare the difference in exacerbations, hospital admissions, and mortality within GOLD subgroups 1–4 and A–D. Last, we tested the predictive accuracy of GOLD 2011 A–D stratification compared with GOLD 2007 1–4 stratification regarding the prediction of exacerbations. We decided to test the following strata: A + B versus C + D against 1 + 2 versus 3 + 4 for exacerbations during the first year of follow-up. We report Harrell’s C-statistic, which assesses the prognostic ability of a variable using a binary outcome (at least one exacerbation vs. no exacerbations). The data set was split in half: one-half for developing the models and the other half for validating the models.
A total of 6,628 individuals fulfilled our COPD inclusion criteria based on spirometry, age, and absence of asthma. Table 1 shows the shift from the 2007 1–4 GOLD classification to the 2011 A–D classification. The new classification results in more individuals belonging to the most severe group, group D, in the 2011 update (n = 292) in comparison with only 45 individuals belonging to the most severe group, group 4, in the 2007 classification.
| GOLD 2007/GOLD 2011 | GOLD A | GOLD B | GOLD C | GOLD D |
| GOLD 1, n = 3,306 | 2,961 (89.6%) | 299 (9.0%) | 37 (1.1%) | 9 (0.3%) |
| GOLD 2, n = 2,851 | 2,165 (75.9%) | 637 (22.3%) | 25 (0.9%) | 24 (0.8%) |
| GOLD 3, n = 426 | — | — | 201 (47.2%) | 225 (52.8%) |
| GOLD 4, n = 45 | — | — | 8 (17.8%) | 37 (82.2%) |
The general characteristics of the individuals with COPD according to the two classifications are shown in Table 2. The patients with more severe dyspnea (groups B and D) are older than those with less dyspnea (groups A and C) (P < 0.001). The percentage of individuals with exacerbations during the year before the examination is relatively low, reflecting that this is a cohort recruited from the general population and not in a hospital setting. In general, the prevalence of individuals receiving inhaled medications is low, and almost 50% of individuals in the most severe categories are not treated at all. The prevalence of ischemic heart disease and previous myocardial infarction is significantly higher in subgroups B and D than in subgroups A and C.
| GOLD 2007 | GOLD 2011 | |||||||||
| 1 | 2 | 3 | 4 | A | B | C | D | |||
| (n = 3,306) | (n = 2,851) | (n = 426) | (n = 45) | P Value | (n = 5,126) | (n = 936) | (n = 271) | (n = 292) | P Value | |
| Males, % | 46 | 48 | 52 | 60 | 0.012 | 48 | 44 | 49 | 50 | 0.152 |
| Age, yr (SD) | 66 (12) | 66 (11) | 70 (9) | 68 (9) | <0.001 | 65 (11) | 71 (11) | 69 (10) | 72 (9) | <0.001 |
| BMI, kg/m2 (SD) | 25.2 (3.9) | 25.8 (4.4) | 25.5 (4.9) | 25.7 (4.6) | <0.001 | 25.2 (3.9) | 27.3 (4.8) | 24.9 (3.9) | 26.1 (5.2) | <0.001 |
| FEV1, %pred (SD) | 94.7 (11.3) | 67.8 (8.1) | 42.6 (5.1) | 24.6 (4.6) | NA | 83.8 (16.3) | 74.2 (16.3) | 51.8 (19.5) | 43.3 (14.7) | NA |
| mMRC ≥ 2, % | 9.3 | 23.2 | 52.8 | 82.2 | <0.001 | 0 | 100 | 0 | 100 | NA |
| Never-smokers, % | 26.0 | 13.6 | 7.0 | 4.7 | <0.001 | 21.2 | 14.9 | 11.3 | 6.3 | <0.001 |
| Current smokers, % | 32.5 | 47.9 | 50.4 | 39.5 | <0.001 | 39.5 | 41.2 | 55.1 | 38.8 | <0.001 |
| Sedentary, % | 6.5 | 11.9 | 18.9 | 22.2 | <0.001 | 6.9 | 20.9 | 9.0 | 24.6 | <0.001 |
| No exacerbations in previous year, % | 97.5 | 96.6 | 87.8 | 80 | <0.001 | 98.7 | 97.8 | 72.0 | 72.9 | <0.001 |
| Two or more exacerbations in previous year, % | 1.4 | 1.7 | 6.8 | 11.1 | <0.001 | 0 | 0 | 25.1 | 20.7 | NA |
| Any inhaled medication, % | 4.2 | 12.1 | 36.6 | 55.6 | <0.001 | 5.4 | 19.6 | 18.8 | 52.5 | <0.001 |
| On LABA, LAMA, or ICS, % | 2.9 | 8.3 | 30.0 | 51.1 | <0.001 | 3.6 | 13.7 | 15.5 | 43.4 | <0.001 |
| Ischemic heart disease, % | 7.8 | 10.8 | 15.5 | 13.3 | <0.001 | 7.0 | 20.7 | 10.3 | 19.3 | <0.001 |
| Myocardial infarction, % | 3.1 | 5.3 | 8.7 | 8.9 | <0.001 | 3.3 | 8.2 | 6.3 | 10.2 | <0.001 |
Table 3 shows the 1- and 3-year prognosis according to the two GOLD stratifications. In general, the old classification 1–4 reflected prognosis well, with increasing prevalence of all investigated end points with increasing GOLD group. The new A–D classifications was a good predictor regarding exacerbations, with groups C and D experiencing a much higher incidence in the following years and a higher average number of exacerbations per year than groups A and B (P < 0.001). Harrell’s C-statistic for GOLD 2011 regarding predicting exacerbations during the first year of observation was 0.70 (95% confidence interval [CI], 0.66–0.74), which was significantly higher than the 0.59 (95% CI, 0.55–0.62) based on GOLD 2007 (P < 0.001). This indicates a better predictive performance of the 2011 revision regarding predicting future exacerbations.
| GOLD 2007 | GOLD 2011 | |||||||||
| 1 | 2 | 3 | 4 | A | B | C | D | |||
| (n = 3,306) | (n = 2,851) | (n = 426) | (n = 45) | P Value* | (n = 5,126) | (n = 936) | (n = 271) | (n = 292) | P Value* | |
| One-year prognosis | ||||||||||
| Exacerbations, n (%) | 88 (2.7) | 145 (5.1) | 66 (15.5) | 17 (37.8) | <0.001 | 111 (2.2) | 54 (5.8) | 68 (25.1) | 83 (28.3) | <0.001 |
| Average number of exacerbations per year | 0.1 | 0.1 | 0.4 | 1.1 | <0.001 | 0·0 | 0.1 | 0.7 | 0.8 | <0.001 |
| Hospital admission due to COPD, n (%) | 6 (0.2) | 38 (1.3) | 33 (7.8) | 11 (24.4) | <0.001 | 13 (0.3) | 28 (3.0) | 7 (2.6) | 40 (13.6) | <0.001 |
| Hospital admission due to all causes, n (%) | 720 (21.8) | 763 (26.8) | 144 (33.8) | 17 (37.8) | <0.001 | 1,088 (21.2) | 354 (37.8) | 69 (25.5) | 133 (45.1) | <0.001 |
| Death, n (%) | 23 (0.7) | 28 (1.0) | 10 (2.3) | 1 (2.2) | 0.007 | 31 (0.6) | 19(2.0) | 2 (0.7) | 10 (3.4) | <0.001 |
| Future treatment with inhaled medication, n (%) | 181 (5.5) | 525 (18.5) | 214 (50.3) | 33 (73.3) | <0.001 | 434 (8.5) | 249 (26.7) | 87 (32.1) | 183 (62.1) | <0.001 |
| Three-year prognosis | ||||||||||
| Exacerbations, n (%) | 186 (6.1) | 329 (12.4) | 138 (34.4) | 26 (66.0) | <0.001 | 297 (6.4) | 141 (17.0) | 94 (35.9) | 147 (53.1) | <0.001 |
| Average number of exacerbations per year | 0.1 | 0.1 | 0.5 | 1.4 | <0.001 | 0.0 | 0.2 | 0.6 | 0.9 | <0.001 |
| Hospital admission due to COPD, n (%) | 27 (0.9) | 120 (4.7) | 83 (20.9) | 18 (42.5) | <0.001 | 63 (1.4) | 73 (8.9) | 21 (8.5) | 91 (33.3) | <0.001 |
| Hospital admission due to all causes, n (%) | 1,489 (48.0) | 1,524 (55.5) | 258 (62.2) | 29 (67.3) | <0.001 | 2,336 (48.2) | 612 (67.8) | 139 (53.1) | 213 (74.4) | <0.001 |
| Death, n (%) | 118 (4.0) | 154 (5.9) | 57 (14.3) | 11 (25.8) | <0.001 | 173 (3.8) | 91 (10.6) | 20 (8.2) | 56 (20.1) | <0.001 |
| Cardiovascular death,† n (%) | 16 (0.6) | 34 (1.4) | 8 (2.2) | 2 (4.5) | <0.001 | 27 (0.6) | 22 (2.9) | 1 (0.5) | 10 (4.0) | <0.001 |
| Respiratory death,† n (%) | 2 (0.1) | 2 (0.1) | 8 (2.3) | 3 (10.1) | <0.001 | 2 (0.1) | 2 (0.3) | 1 (0.4) | 10 (4.4) | <0.001 |
| Cancer death,† n (%) | 16 (0.6) | 27 (1.1) | 8 (2.1) | 0 (0.0) | 0.01 | 25 (0.6) | 16 (2.1) | 1 (0.4) | 9 (3.5) | <0.001 |
| Future treatment with inhaled medication, n (%) | 299 (10.2) | 788 (29.5) | 269 (64.9) | 36 (81.8) | <0.001 | 709 (15.3) | 346 (40.1) | 123 (48.1) | 214 (74.1) | <0.001 |
However, regarding mortality, the trend was not the same, as group B had higher mortality than group C (log-rank test, P = 0.02), which is also shown in Figure 2 depicting survival according to both GOLD 1–4 and GOLD A–D. The difference in survival between groups B and C remained statistically significant after inclusion of age and sex in the model (Cox proportional hazards; P = 0.03). In particular, the mortality from cardiovascular diseases and cancer was significantly higher in group B compared with group A (log-rank test, P < 0.001 for both) and in group D compared with group C (log-rank test, P = 0.008 for cardiovascular deaths and P = 0.01 for cancer deaths). A similar trend was seen regarding hospital admissions due to all causes, where group B experienced a higher risk of admission than group C (log-rank test, P < 0.001).
Figure 2. Survival shown as Kaplan-Meier curves according to the GOLD 2007 classification (left) and GOLD 2011 classification (right).
[More] [Minimize]The probability of being treated with inhaled medication increased significantly after participation in our survey, especially in the most severe groups: after 3 years almost 82% in group 4 and 74% in group D were receiving inhaled medications for COPD (Table 3).
The distribution of the 271 individuals belonging to group C and of the 292 individuals belonging to group D according to the C1–D3 subgroups and their characteristics are shown in Table 4. The major criterion of placing the individuals in groups C and D was the low level of FEV1 (77%), but there were 95 individuals (17%) (C2 + D2: 62 + 33) stratified to group C or D because of frequent exacerbations only. The number of individuals with both low FEV1 and frequent exacerbations was only 34 (6%) and especially subgroup C3 (mMRC < 2 and FEV1 < 50%pred and two or more exacerbations in the previous year) was very small, comprising only six individuals. Groups C2 and D2 (FEV1 ≥ 50%pred and two or more exacerbations in the previous year) differed from the other subgroups by comprising a high prevalence of women with much higher prevalence of never-smokers, suggesting a specific phenotype.
| C Subgroups of GOLD 2011 | D Subgroups of GOLD 2011 | |||||||
| C1: FEV1 < 50%pred and < 2 exa (n = 203) | C2: FEV1 ≥ 50%pred and ≥ 2 exa (n = 62) | C3: FEV1 < 50%pred and ≥ 2 exa (n = 6) | C Subgroup P Value | D1: FEV1 < 50%pred and < 2 exa (n = 234) | D2: FEV1 ≥ 50%pred and ≥ 2 exa (n = 33) | D3: FEV1 < 50%pred and ≥ 2 exa (n = 28) | D Subgroup P Value | |
| Males, % | 56 | 27 | 50 | <0.001 | 51 | 42 | 50 | 0.644 |
| Age, yr (SD) | 68 (10) | 73 (9) | 73 (8) | <0.001 | 71 (9) | 74 (11) | 75 (7) | 0.024 |
| BMI, kg/m2 (SD) | 25.0 (4.1) | 24.9 (3.3) | 23.5 (2.7) | 0.670 | 26.0 (5.4) | 26.5 (4.6) | 26.0 (4.2) | 0.883 |
| FEV1, %pred (SD) | 42.8 (6.1) | 82.8 (16.7) | 37.7 (7.3) | <0.001 | 39.7 (7.8) | 73.1 (21.2) | 38.4 (8.5) | <0.001 |
| mMRC ≥ 2, % | 0.0 | 0.0 | 0.0 | NA | 100.0 | 100.0 | 100.0 | NA |
| Never-smokers, % | 8.6 | 21.0 | 0.0 | 0.034 | 5.3 | 12.9 | 7.1 | 0.204 |
| Current smokers, % | 61.4 | 37.1 | 33.3 | 0.002 | 43.6 | 25.8 | 14.3 | 0.002 |
| Sedentary, % | 9.5 | 6.6 | 16.7 | 0.477 | 27.3 | 6.7 | 21.4 | 0.031 |
| No exacerbations in previous year, % | 96.1 | 0.0 | 0.0 | NA | 91.9 | 0.0 | 0.0 | NA |
| Two or more exacerbations in previous year, % | 0.0 | 100.0 | 100.0 | NA | 0.0 | 100.0 | 100.0 | NA |
| Any inhaled medication, % | 18.7 | 16.1 | 50.0 | 0.146 | 49.6 | 45.5 | 85.7 | 0.001 |
| On LABA, LAMA, or ICS, % | 15.8 | 11.3 | 50.0 | 0.058 | 39.3 | 36.4 | 85.7 | <0.001 |
| Ischemic heart disease, % | 9.9 | 12.9 | 0.0 | 0.734 | 18.8 | 15.2 | 28.6 | 0.349 |
| Myocardial infarction, % | 6.9 | 4.8 | 0.0 | 0.844 | 11.1 | 9.1 | 3.6 | 0.567 |
The prognosis for the C1–D3 subgroups is given in Table 5 and showed considerable heterogeneity regarding future exacerbation: subgroups C1 and D1 (FEV1 < 50%pred and fewer than two exacerbations in the previous year) had a much lower risk of exacerbations than the other groups (log-rank test, P < 0.001). Regarding mortality, the overall trend was that low lung function and the presence of dyspnea were a stronger predictor of death than exacerbation history, although we did observe high mortality in subgroup C2 (mMRC < 2 and FEV1 ≥ 50%pred and two or more exacerbations in the previous year).
| C Subgroups of GOLD 2011 | D Subgroups of GOLD 2011 | |||||||
| C1: FEV1 < 50%pred and < 2 exa (n = 203) | C2: FEV1 ≥ 50%pred and ≥ 2 exa (n = 62) | C3: FEV1 < 50%pred and ≥ 2 exa (n = 6) | C Subgroup P Value* | D1: FEV1 < 50%pred and < 2 exa (n = 234) | D2: FEV1 ≥ 50%pred and ≥ 2 exa (n = 33) | D3: FEV1 < 50%pred and ≥ 2 exa (n = 28) | D Subgroup P Value* | |
| One-year prognosis | ||||||||
| Exacerbations, n (%) | 16 (7.9) | 47 (75.8) | 5 (83.3) | <0.001 | 43 (18.5) | 21 (63.6) | 19 (68.8) | <0.001 |
| Hospital admissions due to COPD, n (%) | 6 (3.0) | 1 (1.6) | 0 (0.0) | 0.78 | 28 (12.0) | 2 (6.1) | 10 (36.4) | <0.001 |
| Hospital admissions all causes, n (%) | 49 (24.1) | 18 (29.0) | 2 (33.3) | 0.55 | 96 (41.0) | 23 (69.7) | 14 (50.0) | 0.002 |
| Death, n (%) | 1 (0.5) | 1 (1.6) | 0 (0.0) | 0.65 | 8 (3.4) | 0 (0.0) | 2 (7.1) | 0.30 |
| Future treatment with inhaled medications, n (%) | 74 (36.5) | 9 (14.5) | 4 (66.7) | <0.001 | 146 (62.5) | 14 (42.4) | 23 (82.1) | 0.002 |
| Three-year prognosis | ||||||||
| Exacerbations, n (%) | 38 (20.1) | 51 (83.2) | 5 (83.3) | <0.001 | 98 (45.4) | 26 (79.8) | 23 (87.5) | <0.001 |
| Hospital admissions due to COPD, n (%) | 17 (9.2) | 3 (5.6) | 1 (16.7) | 0.48 | 67 (30.8) | 8 (27.8) | 16 (62.1) | <0.001 |
| Hospital admissions all causes, n (%) | 96 (48.6) | 39 (66.7) | 4 (66.7) | 0.03 | 161 (71.0) | 26 (81.3) | 26 (92.9) | 0.002 |
| Death, n (%) | 14 (7.6) | 5 (8.8) | 1 (20.0) | 0.65 | 43 (19.4) | 3 (10.3) | 10 (36.6) | 0.02 |
| Future treatment with inhaled medications, n (%) | 104 (53.6) | 15 (27.5) | 4 (66.7) | <0.001 | 172 (75.0) | 17 (54.4) | 25 (89.3) | <0.001 |
This is to our knowledge the first study focusing on the distribution and prognosis of individuals in the general population according to the novel GOLD 2011 classification. Compared with the older version, which was based mainly on the level of FEV1, the new multidimensional classification identifies a higher numbers of individuals at risk of exacerbations, thus providing a better prognostic separation as shown by the Harrell C-statistic.
The setting of our study in the general population is the most likely explanation for the observation that the majority of our individuals were treatment naive. In fact, only in group D were more than 50% of the participants receiving inhaled medication at the time of examination. This reflects clearly the underdiagnosis of COPD in the general population and is in line with studies focusing on the substantial need for an earlier diagnosis of COPD (10, 11). After participation in our survey, the individuals were informed of the results of their health examination and it is reassuring to observe that this resulted in more individuals starting relevant medical treatment (Table 3).
In our cohort, a history of repeated exacerbations (two or more in the previous year) was a good predictor of subsequent exacerbations. Thus, although the prevalence of “frequent exacerbators” was lower in our sample, we could reproduce the findings from the ECLIPSE (Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints) study in our general population sample (12). In fact, our figures describing the risk of an exacerbation in the year to come (groups C2 and C3 and groups D2 and D3; the groups with two or more exacerbations in the previous year) (Table 5) are quite comparable to the findings in the ECLIPSE cohort: approximately 70% of those who had at least two exacerbations in the previous year experienced a new exacerbation in the following year. Similarly, less than 20% of those with FEV1 < 50%pred and less than two exacerbations in the previous year experienced an exacerbation next year.
The main role of the new A–D stratification is to suggest first-line medical treatment for the patients within these categories. The patients belonging to groups C and D have been defined as those with a higher risk of disease progression and the treatments suggested for these individuals include inhaled corticosteroids and possibly Phosphodiesterase4 inhibitors (1). However, groups C and D consist of individuals with quite variable lung function and a variable history of exacerbations, which is mirrored by quite distinct prognosis regarding risk of exacerbations and mortality (Table 5). In the setting of the general population, which is more comparable to the situation in the office of a general practitioner than to a specialized COPD clinic, the majority of the patients in groups C and D are stratified into these groups because of low FEV1 level (77%; Table 3) and are thus at low risk of exacerbation, although they have more hospital admissions than individuals in groups A and B (Tables 3 and 5). Hence, although the new GOLD recommendations seem sound for groups C and D in general, our findings suggest that in the setting outside pulmonary clinics the vast majority of individuals in these groups are at low risk of COPD exacerbation in the future, and that this should be taken into account when choosing an appropriate treatment.
Our findings substantiate the importance of dyspnea as a predictor of poor survival (13–15). We believe that the presence of an underlying disease causing dyspnea is responsible for the quite surprising finding of the significantly poorer survival of group B compared with group C, despite a much lower FEV1%pred in the latter group. This is in line with a high prevalence of both diagnosed and unrecognized heart disease but also of cancer (5, 16–19). Our findings underline the importance of searching for comorbidities in individuals with COPD, a notion that has come into increasing focus and is also acknowledged in the new GOLD revision (1, 20–22). These comorbidities are not only present in the most severe disease category (group D) but also in the less severe disease category (group B). Also importantly, the presence of these comorbidities, in particular heart disease, results in the fact that using the A–D categories as a definite and precise guide to COPD therapy can be problematic, as dyspnea in some patients with COPD may benefit more from treatment of heart disease than from an uncritical increase in inhaled medications. We therefore suggest that in the next revision of GOLD, comorbidities should be included in the staging of COPD.
A limitation of our study is the fact that it was not especially designed to detect COPD exacerbations. This means that mild exacerbations not treated with oral corticosteroids and antibiotics were overlooked. On the other hand, we used the generally accepted definition of an exacerbation, which is employed in most of the clinical trials of COPD and is in line with the GOLD document (1). Our findings are also affected by “nonresponder bias,” and we are likely to underestimate of the prevalence of most severe COPD cases (23). Yet, as we are using prebronchodilator values to classify participants, we are at the same time likely to overestimate the prevalence of milder cases. On the basis of our previous studies of reversibility in a similar population, which were done in connection with a randomized trial, we do not think that lack of postbronchodilatory values for FEV1 has a significant influence on our results and conclusions (24). Last, we must acknowledge that the GOLD system is designed for use in clinical practice, whereas we have tested it in an epidemiological setting.
In conclusion, our analysis of a large sample of individuals with COPD selected from the general population shows that the new multidimensional GOLD A–D classification provides better diagnostic separation regarding predicting exacerbations than the GOLD 2007 1–4 stratification. Our findings also substantiate the role of dyspnea as the strong predictor of poor survival in COPD even among individuals with relatively preserved FEV1, and we suggest that such individuals should be examined for the presence of heart disease. Last, our study shows that GOLD groups C and D are heterogeneous, being composed of phenotypes with variable risk of future exacerbations, and we suggest that this should be taken into account when planning treatment for these individuals.
| 1. | Global Initiative for Chronic Obstructive Lung Disease (GOLD). Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: revised 2011. Available from: http://www.goldcopd.org/uploads/users/files/GOLD_Report_2011_Jan21.pdf |
| 2. | Rabe KF, Hurd S, Anzueto A, Barnes PJ, Buist SA, Calverley P, Fukuchi Y, Jenkins C, Rodriguez-Roisin R, van Weel C, et al.. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: GOLD executive summary. Am J Respir Crit Care Med 2007;176:532–555. |
| 3. | Lange P, Marott J, Dahl M, Ingebrigtsen TS, Vestbo J, Nordestgaard B. Predicting the natural course of COPD using the old and the new GOLD classification. Eur Respir J 2012 Suppl 56:278s. |
| 4. | Appleyard M, Hansen AT, Schnohr P, Jensen G, Nyboe J. The Copenhagen City Heart Study: a book of tables with data from the first examination (1976–78) and a five-year follow-up (1981–83). Scand J Soc Med 1989;170(Suppl 41):1–160. |
| 5. | Lange P, Mogelvang R, Marott JL, Vestbo J, Jensen JS. Cardiovascular morbidity in COPD: a study of the general population. COPD 2010;7:5–10. |
| 6. | Stender S, Frikke-Smidth R, Nordestgaard BG, Tybjærg-Hansen A. Sterol transporter adenosine triphosphate–binding cassette transporter G8, gallstones, and biliary cancer in 62,000 individuals from the general population. Hepatology 2011;53:640–648. |
| 7. | Danish National Patient Registry. Available from: http://www.sst.dk/Indberetning%20og%20statistik/Landspatientregisteret |
| 8. | Gaist D, Sørensen HT, Hallas J. The Danish prescription registries. Dan Med Bull 1997;44:445–448. |
| 9. | WHO Collaborating Centre for Drug Statistics Methodology. Norwegian Institute of Public Health, ATC/DDD Index 2010. Available from: http://www.whocc.no/atc_ddd_index/ |
| 10. | Lange P, Marott JL, Dahl M, Ingebrigtsen TS, Vestbo J, Nordestgaard BG. Substantial need for early diagnosis, rehabilitation and treatment of COPD. Dan Med J 2012;59:A4396. |
| 11. | Ulrik CS, Løkke A, Dahl R, Dollerup J, Hansen G, Cording PH, Andersen KK; TOP Study Group. Early diagnosis of COPD in general practice. Int J Chron Obstruct Pulmon Dis 2011;6:123–127. |
| 12. | Hurst J, Vestbo J, Anzueto A, Locantore N, Mülerova H, Tal-Singer R, Miller B, Lomas D, Agusti A, MacNee W, et al.. Susceptibility to exacerbation in chronic obstructive pulmonary disease. N Engl J Med 2010;363:1128–1138. |
| 13. | Celli BR, Cote CG, Marin JM, Casanova C, Montes de Oca M, Mendez RA, Pinto Plata V, Cabral HJ. The body-mass index, airflow obstruction, dyspnea, and exercise capacity index in chronic obstructive pulmonary disease. N Engl J Med 2004;350:1005–1012. |
| 14. | Vestbo J, Knudsen KM, Rasmussen FV. Should we continue using questionnaires on breathlessness in epidemiologic surveys? Am Rev Respir Dis 1988;137:1114–1118. |
| 15. | Nishimura K, Izumi T, Tsukino M, Oga T. Dyspnea is a better predictor of 5-yr survival than airway obstruction in patients with COPD. Chest 2002;121:1434–1440. |
| 16. | Rutten FH, Cramer MJ, Grobbee DE, Sachs AP, Kirkels JH, Lammers JW, Hoes AW. Unrecognized heart failure in elderly patients with stable chronic obstructive pulmonary disease. Eur Heart J 2005;26:1887–1894. |
| 17. | Brekke PH, Omland T, Smith P, Soyseth V. Underdiagnosis of myocardial infarction in COPD—Cardiac Infarction Injury Score (CIIS) in patients hospitalised for COPD exacerbation. Respir Med 2008;102:1243–1247. |
| 18. | McAllister DA, Maclay JD, Mills NL, Leitch A, Reid P, Carruthers R, O'Connor J, McAlpine L, Chalmers G, Newby DE, et al.. Diagnosis of myocardial infarction following hospitalisation for exacerbation of COPD. Eur Respir J 2012;39:1097–1103. |
| 19. | Sode BF, Dahl M, Nordestgaard BG. Myocardial infarction and other co-morbidities in patients with chronic obstructive pulmonary disease: a Danish nationwide study of 7.4 million individuals. Eur Heart J 2011;32:2365–2375. |
| 20. | Sin DD, Anthonisen NR, Soriano JB, Agusti AG. Mortality in COPD: role of comorbidities. Eur Respir J 2006;28:1245–1257. |
| 21. | Fabbri LM, Luppi F, Beghe B, Rabe KF. Complex chronic comorbidities of COPD. Eur Respir J 2008;31:204–212. |
| 22. | Mannino DM, Doherty DE, Buist SA. Global Initiative on Obstructive Lung Disease (GOLD) classification of lung disease and mortality: findings from the Atherosclerosis Risk in Communities (ARIC) Study. Respir Med 2006;100:115–122. |
| 23. | Jensen G. Epidemiology of chest pain and angina pectoris. Acta Med Scand 1983;214(Suppl 682):1–120. |
| 24. | Vestbo J, Sørensen T, Lange P, Torre P, Viskum K. Long-term effect of inhaled budesonide in mild and moderate chronic obstructive pulmonary disease: a randomised controlled trial. Lancet 1999;353:1819–1823. |
Supported by the Capital Region of Copenhagen, Danish Heart Foundation, Danish Lung Foundation, Velux Foundation, Herlev Hospital, GlaxoSmithKline (grant EPI 115882–EUPharmaLocal). The funding sources had no role in the design and conduct of the study; in the collection, management, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript.
Author Contributions: Study concept and design: P.L., J.L.M., J.V., M.D., B.G.N.; acquisition of data: P.L., B.G.N., M.D.; analysis and interpretation of data: P.L., J.L.M., K.R.O., J.V., T.S.I.; first drafting of the manuscript: P.L.; critical revision of the manuscript for important intellectual content: all authors; statistical analysis: J.L.M., P.L.; obtained funding: P.L., J.V., B.G.N.; administrative, technical, and material support: B.G.N., P.L.; study supervision: P.L., B.G.N. Data access and responsibility: P.L. and statistician J.L.M. had full access to all the data in the study, and take responsibility for the integrity of the data and the accuracy of the data analysis.
Originally Published in Press as DOI: 10.1164/rccm.201207-1299OC on September 20, 2012
Author disclosures are available with the text of this article at www.atsjournals.org.
