Key words: COPD; diagnosis; epidemiology; gender; incidence; prevalence; respiratory symptoms; risk factors

Ten-Year Cumulative Incidence of COPD and Risk Factors for Incident Disease in a Symptomatic Cohort* Anne Lindberg, MD; Ann-Christin Jonsson, SRN; Eva Rönmark, PhD; Rune Lundgren, MD, PhD, FCCP; Lars-Gunnar Larsson, MD, PhD; and Bo Lundbäck, MD, PhD Study objectives: To determine the 10-year cumulative incidence of COPD in a cohort of subjects with respiratory symptoms (Global Initiative for Chronic Obstructive Lung Disease [GOLD] stage 0) using the British Thoracic Society (BTS) and GOLD spirometric criteria. Furthermore, we sought to evaluate risk and gender factors for incident COPD. Design and setting: A postal questionnaire was administered in 1986 to all 6,610 subjects in eight areas of northern Sweden who had been born in 1919 to 1920 (group 1), 1934 to 1935 (group 2), and 1949 to 1950 (group 3). The response rate was 86%. All of the subjects reporting respiratory symptoms were invited to participate in a structured interview and pulmonary function test (PFT), and 1,506 (91%) participated. In 1996, 90% could be traced for follow-up, of whom 1,165 (86%) of the invited subjects participated and 1,109 subjects (534 women) were able to perform technically adequate PFTs in both 1986 and 1996. Results: The 10-year cumulative incidence of COPD was estimated at 8.2% (using BTS criteria) and 13.5% (using GOLD criteria). Significant risk factors for incident COPD (using BTS and GOLD criteria) in a multivariate analysis were higher age (group 1 odds ratio [OR]: BTS criteria, 3.49; GOLD criteria, 3.37; group 2 OR: BTS criteria, 4.50; GOLD criteria, 5.70) and smoking (OR: BTS criteria, 5.37; GOLD criteria, 4.56), but not gender or heredity. Respiratory symptoms were significantly associated with incident COPD when added to the same model. In analogous analyses that were conducted separately for men and women, smoking yielded an OR of 8.52 among women (95% confidence interval [CI], 3.43 to 21.2) compared with 3.14 among men (95% CI, 1.26 to 7.84). The symptoms cough, sputum production, and chronic productive cough reached statistical significance in women, while dyspnea and wheeze did so in men. Conclusion: In this cohort, the 10-year cumulative incidence of COPD was 8.2% (using BTS criteria) and 13.5% (using GOLD criteria). Increasing age, smoking, and bronchitic symptoms, but not gender, were risk factors for incident COPD. GOLD stage 0 therefore appears to identify subjects who are at risk of COPD, but men and women presented different risk profiles. (CHEST 2005; 127:1544 1552) Key words: COPD; diagnosis; epidemiology; gender; incidence; prevalence; respiratory symptoms; risk factors Abbreviations: BTS British Thoracic Society; CI confidence interval; GOLD Global Initiative for Chronic Obstructive Lung Disease; OLD obstructive lung disease; OR odds ratio; PFT pulmonary function test; SEI socioeconomic index; VC vital capacity COPD is recognized as a major cause of morbidity and mortality worldwide. There have been reports 1 5 on the prevalence of COPD in the general *From the Obstructive Lung Disease in Northern Sweden Studies (Drs. Lindberg, Rönmark, and Larsson, and Ms. Jonsson), Sunderby Central Hospital of Norrbotten, Luleå; the Department of Respiratory Medicine and Allergy (Dr. Lundgren), Umeå University, Umeå; and Lung and Allergy Research (Dr. Lundbäck), National Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden. This research was supported by the Swedish Heart-Lung Foundation, Norrbotten Health Care Authority. Manuscript received January 15, 2004; revision accepted October 29, 2004. population ranging predominantly from 4 to 10%. By 2020, it is predicted that COPD will be the third leading cause of morbidity and the sixth leading cause of mortality. 6,7 The most widely recognized risk factors for COPD are smoking and increasing age. 8,9 Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. org/misc/reprints.shtml). Correspondence to: Anne Lindberg, MD, Senior Consultant, Division of Respiratory Medicine and Allergy, Department of Medicine, Sunderby Central Hospital of Norrbotten, SE 971 80 Luleå, Sweden; e-mail: anne.lindberg@nll.se 1544 Clinical Investigations

Epidemiologic data on COPD are important, since the progressive disease causes disability for the patient and imposes increasing demands on the health-care system and a growing economic burden on society. Despite this, there are still only a few reports on the incidence of COPD. 10 12 The Global Initiative for Chronic Obstructive Lung Disease (GOLD) guidelines classify subjects with respiratory symptoms and normal lung function as being at risk for COPD (GOLD stage 0). 13 However, the data on the symptoms and early clinical profile of patients who are at risk of developing COPD are incomplete. It is not known whether this subpopulation will develop COPD. 12 A greater knowledge of the incidence of COPD and the risk factors for incident disease is needed to help predict the future need for health-care resources and to focus on preventive action for subjects who are at risk of developing the disease. In previously published studies, different age distributions and smoking habits in the populations studied, and different spirometric definitions of COPD interfere when comparing the results. Several national and international guidelines 8,9,13,14 relating to the diagnosis and treatment of COPD have recently been presented, but the diagnostic and spirometric criteria for COPD differ among them. So, there are no generally accepted criteria for COPD, and nonuniform criteria will result in different prevalence rates. 15,16 The main aim of this study was to determine the 10-year cumulative incidence of COPD in a cohort of subjects reporting respiratory symptoms and to compare the differences that result from using the spirometric criteria of the British Thoracic Society (BTS) and GOLD. Other aims were to evaluate reported respiratory symptoms among subjects prior to the development of COPD, to evaluate the risk factors for incident COPD, and to identify gender differences relating to the risk factor pattern for incident COPD. Study Population Materials and Methods In 1986, a postal questionnaire relating to respiratory symptoms and diseases was sent to all 6,610 subjects born in 1919 to 1920, 1934 to 1935, and 1949 to 1950 in eight geographic areas in Norrbotten, which is the northernmost province of Sweden. The response rate was 86% (5,698 completed questionnaires). All of the subjects (n 1,655) giving an affirmative answer to any of the questions relating to respiratory symptoms in the questionnaire were identified and were invited to take part in a structured interview and pulmonary function test (PFT). In all, 1,506 subjects (91%) were interviewed and examined. A 10-year follow-up survey was performed in 1996, when 90% of the 1,506 subjects could be traced and were invited to participate in a second interview and a PFT. Of these, 1,165 subjects participated (86% of those invited to participate), and 1,109 subjects (74% of the original 1,506 participants) were able to perform adequate PFTs in both 1986 and 1996 (Table 1). The main reason for nonparticipation was death. Baseline Characteristics The study population and basic characteristics are shown in Tables 1 and 2. The symptoms reported in 1986 overlapped in this symptomatic cohort, with the greatest overlap existing be- Table 1 Basic Characteristics of Study Population, Age Group, Lung Function, Smoking Categories, and Symptoms* Variables Category Men Women All Age group, No. Born 1949 1950 189 187 376 Born 1934 1935 230 202 432 Born 1919 1920 156 145 301 Lung function 1986 FEV 1 % predicted 92.74 16.1 91.99 16.7 92.38 16.4 FEV 1 /FVC ratio 0.77 0.09 0.78 0.08 0.77 0.08 1996 FEV 1 % predicted 91.21 18.9 91.87 18.1 91.53 18.5 FEV 1 /FVC ratio 0.74 0.10 0.76 0.09 0.75 0.10 Smoking categories Persistent nonsmoker 126 (21.9) 252 (47.2) 378 (34.1) Persistent ex-smoker 198 (34.4) 94 (17.6) 292 (26.3) Quitter 80 (13.9) 55 (10.3) 135 (12.2) Restarter 9 (1.6) 13 (2.4) 22 (2.0) Persistent smoker 151 (26.3) 116 (21.7) 267 (24.1) Symptoms 1986 Cough 342 (59.5) 304 (56.9) 649 (58.5) Sputum production 316 (55.0) 269 (50.4) 585 (52.8) Chronic productive cough 258 (44.9) 207 (38.8) 465 (41.9) Dyspnea 188 (33.0) 239 (44.8) 427 (38.5) Wheeze 389 (67.7) 324 (60.7) 731 (65.9) All 575 534 1,109 *Values given as mean SD or No. (%), unless otherwise indicated. www.chestjournal.org CHEST / 127 / 5/ MAY, 2005 1545

Table 2 Socioeconomic Classification of the Study Population* tween cough and sputum production. The majority of the subjects who reported cough (83.6%) also reported sputum production, and 93.7% of subjects who reported sputum production also reported cough. Dyspnea was reported significantly more often by women (p 0.001), while chronic productive cough was more commonly reported in men (p 0.040). There were no gender differences when it came to reported cough, sputum, and wheeze. Of all the subjects, 35.7% reported that they had or at some time had had respiratory diseases other than obstructive lung disease (OLD). The vast majority of these reports were of previous respiratory tract infections, including pneumonia and acute bronchitis, and childhood diseases, including pertussis. Only 21 subjects (1.9%) reported a chronic lung disease other than OLD, and, of those, 9 subjects had a history of tuberculosis. Only five subjects reported having a respiratory disease that affected lung function, and they were excluded from the population at risk. Methods Socioeconomic Class Men Women All Professionals and executives 112 (19.5) 71 (13.3) 183 (16.5) Assistant nonmanual 76 (13.2) 106 (19.9) 182 (16.4) Manual industry 228 (20.6) 22 (4.1) 250 (22.5) Manual service 80 (13.9) 238 (44.6) 318 (28.7) Self-employed nonprofessional 37 (6.4) 15 (2.8) 52 (4.7) Housewife 0 (0) 31 (5.8) 31 (2.8) Unknown 42 (7.3) 51 (9.6) 93 (8.4) All 575 534 1,109 *Data are presented as No. of subjects (%). The postal questionnaire and the questionnaire used at the structured interview, together with their origins, have previously been described in detail. 17 23 They included questions about respiratory symptoms, factors provoking these symptoms, the use of antiasthmatic drugs, diagnoses of obstructive airway diseases, other lung diseases, contact with health care, smoking habits, and occupation. A dry volume spirometer (Vicatest 5; Mijnhardt; Bunnik, the Netherlands) was used for spirometry, and the test procedure followed the American Thoracic Society recommendations. 24 Vital capacity (VC) was derived from the best value of FVC and slow VC. Specially trained nurses performed the PFTs and the structured interviews. Swedish reference values 25 were used for normal values, which conform well with the symptom-free population of northern Sweden. 26 The study was approved by the Regional Committee for Ethics at the Umeå University. Definitions The criteria used for COPD were as follows: 1. BTS criteria: FEV 1 /VC ratio, 0.70; and FEV 1, 80% predicted 9 ; 2. GOLD criteria: FEV 1 /FVC ratio, 0.70. 13 Neither the BTS nor the GOLD make a clear exception for asthma with chronic airway obstruction, and, as a result, asthma has not been eliminated as a cause of chronic obstruction. Cumulative Incidence The proportion of a population at risk (in this case, those who were free from COPD at the start of the study) in whom the disease developed during a defined time period. An incident case was a subject in whom the disease (in this case, COPD) developed during the defined observation period. Smoking Categories At the interviews in 1986 and 1996, the participants were classified as nonsmokers, ex-smokers (ie, stopped smoking for at least 1 year), and current smokers. Based on these answers, they were classified as being in the following smoking categories (1986 status/1996 status): nonsmoker (nonsmoker/nonsmoker); persistent ex-smoker (ex-smoker/ex-smoker); quitter (smoker/ex-smoker); restarter (ex-smoker/smoker); persistent smoker (smoker/ smoker); and other (difficult to classify). Interview Key Questions Regarding Respiratory Symptoms Cough: Do you usually cough in the morning? or Do you usually cough at other times of the day? Sputum Production: Do you usually have phlegm when coughing? Chronic Productive Cough (Symptoms Required for the Diagnosis of Chronic Bronchitis): Do you usually have phlegm when coughing, or have phlegm which is difficult to bring up, most days for periods of at least three months, during at least the last two years? Dyspnea: Do you get shortness of breath during exertion? Recurrent Wheeze: Do you usually have wheezing, whistling or a noisy sound in your chest when breathing? In the following text, the term bronchitic symptoms is used as a synonym for the presence of cough and/or sputum production and/or chronic bronchitis. Socioeconomic Classification We used the socioeconomic classification system based on occupation (socioeconomic index [SEI]). 27 The following groupings were used: class 1, professionals and executives; class 2, assistant nonmanual employees; class 3, manual workers in industry; class 4, manual workers in service; class 5, selfemployed nonprofessionals; class 6, housewives; and class 7, occupation unknown (Table 2). The classification system also reflects the level of education. The level of education was subdivided as follows: classes 1 and 2, higher education; and classes 3 to 7, lower education. Statistical Analysis The prevalent cases of COPD according to the GOLD criteria in 1986 were excluded, and the 10-year cumulative incidence for COPD according to the GOLD criteria was calculated for the period from 1986 to 1996 in the remaining population (ie, the population at risk). The corresponding calculations were made for estimations of the 10-year cumulative incidence of COPD according to the BTS criteria. Subjects with pulmonary diseases other than obstructive disease affecting lung function were also excluded from the population at risk. Statistical calculations were made using a statistical software package (SPSS; SPSS; Chicago, IL). The 2 test was used for univariate and bivariate analyses, and for tests for trends. The symptoms cough, sputum production, chronic productive cough, dyspnea, and wheeze were individually analyzed as risk factors for incident COPD according to the BTS criteria. Socioeconomic class based on occupation was also analyzed as a risk factor for incident COPD. A multiple logistic regression model was created with the independent variables gender, family history of OLD, age group, and smoking 1546 Clinical Investigations

categories. The two definitions of COPD were used as dependent variables. Tests for independent association were performed in the same model for the symptoms cough, sputum production, chronic productive cough (ie, chronic bronchitis), dyspnea, and wheeze. Analogous analyses were performed separately for men and women. Furthermore, socioeconomic group was added to the model, and was tested for all seven SEI classes and by level of education (SEI classes 1 and 2 compared with SEI classes 3 to 7. The 95% significance level (p 0.05) was used. Results Smoking Categories In the total study population, women were significantly more frequently nonsmokers than men in both 1986 and 1996. However, the smoking pattern differed between age groups. In the youngest age group (ie, those born from 1949 to 1950), there were no gender differences in smoking habits, while nonsmoking was significantly more common among women compared with men in the oldest age group (ie, those born from 1919 to 1920). Smoking categories by gender are shown in Table 1. Mortality Data Of the 1,506 participants in 1986, 165 had died at the end of the follow-up period. The mortality rate was significantly higher among men compared with that among women (13.7% vs 8.0%, respectively; p 0.001). Mortality rates were also associated with higher age (birth in 1919 to 1920, 25.0%; birth in 1934 to 1935, 6.4%; and birth in 1949 to 1950, 1.1%; p 0.001) and smoking (nonsmokers, 8.0%; exsmokers, 13.5%; smokers, 11.3%; p 0.025). The prevalence of COPD in 1986 was significantly higher among those who had died at the time of follow-up compared with the survivors (using GOLD criteria, 38.5% vs 15.9%, respectively; using BTS criteria, 28.8% vs 9.6%, respectively; both p 0.001). Lung function in 1986 was also significantly lower (mean [ SD] FEV 1, 78.44 23.85% predicted vs 92.38 16.43% predicted; p 0.001) for those who had died in 1996 compared with survivors. Even when the prevalent cases of COPD in 1986 were excluded, a significant difference remained (mean FEV 1 when excluding prevalent GOLD criteria, 88.99 18.11% predicted vs 95.71 13.97% predicted, respectively [p 0.001]; mean FEV 1 when excluding prevalent BTS criteria, 89.11 16.93 vs 95.37 13.69, respectively [p 0.001]). Incidence of COPD In the population at risk, 83 subjects fulfilled the spirometric criteria for COPD according to the BTS criteria in 1996, while 127 subjects fulfilled the spirometric criteria for COPD using the GOLD criteria. The incident cases of COPD reported significantly more symptoms in 1986 compared with subjects who did not develop COPD during the following 10-year period, regardless of the spirometric criteria that were used (Table 3). Of the incident cases determined according to the BTS criteria, 65.1% had a chronic productive cough at the beginning of the observation period in 1986, while the corresponding figure for those subjects using the GOLD criteria was 53.5%. The 10-year cumulative incidence of COPD was estimated at 8.2% and 13.5%, respectively, using BTS and GOLD criteria. Increasing age and smoking were significantly associated with the incidence of COPD (Table 4). The cumulative incidence of COPD among persistent smokers was close to three times the incidence among persistent nonsmokers (using BTS criteria, 16.7% vs 4.8%, respectively; using GOLD criteria, 24.5% vs 9.4%, respectively), and it was less than half in the youngest age group compared with the two older age groups. Table 3 Comparing the Proportion of Subjects Reporting Respiratory Symptoms in 1986 Among Incident Cases of COPD and Subjects Not Developing COPD* Incident COPD Symptoms BTS Criteria GOLD Criteria Yes No p Value Yes No p Value Cough 79.5 (66) 55.0 (507) 0.001 70.1 (89) 54.1 (437) 0.001 Sputum production 72.3 (60) 49.0 (453) 0.001 63.0 (80) 48.1 (390) 0.007 Chronic productive cough 65.1 (54) 37.6 (348) 0.001 53.5 (68) 36.6 (297) 0.001 Dyspnea 59.0 (49) 34.8 (316) 0.001 47.6 (60) 34.8 (277) 0.005 Recurrent wheeze 86.7 (72) 62.1 (572) 0.001 74.8 (95) 60.8 (748) 0.002 *Data are presented as % (No. of subjects). Comparing subjects developing COPD (incident cases) with subjects not developing COPD using BTS criteria. Comparing subjects developing COPD (incident cases) with subjects not developing COPD using GOLD criteria. www.chestjournal.org CHEST / 127 / 5/ MAY, 2005 1547

Table 4 The 10-Year Cumulative Incidence of COPD According to BTS and GOLD Spirometric Criteria* Variables BTS Criteria GOLD Criteria Gender Men 9.0 15.3 Women 7.5 11.8 p value 0.389 0.125 Age group Born 1949 1950 4.1 6.9 Born 1934 1935 11.0 16.5 Born 1919 1920 9.8 18.9 p value 0.002 0.001 Smoking categories Persistent nonsmoker 4.8 9.4 Persistent ex-smoker 5.0 10.9 Quitter 9.0 10.3 Restarter 10.0 20.0 Persistent smoker 16.7 24.5 p value 0.001 0.001 All 8.2 13.5 *Values given as %, unless otherwise indicated. Test for trend. Univariate and Multivariate Relationships Univariate analysis of the individual symptoms of cough, sputum production, chronic productive cough, dyspnea, and wheeze, as risk factors for incident COPD according to the BTS criteria, all showed significant odds ratios (ORs) of 3.18, 2.72, 3.09, 2.70, and 3.99, respectively. The risk factors for incident COPD identified by the multiple logistic regression model were higher age (ie, birth in 1919 to 1920 and 1934 to 1935) and smoking, when adjusted for gender and family history of OLD (Table 5). When added singly to the multivariate analysis, every symptom was associated with an increased risk for COPD, when defined according to the BTS criteria, as follows: cough: adjusted OR, 1.98; 95% confidence interval (CI), 1.10 to 3.56; sputum production: adjusted OR, 1.77; 95% CI, 1.04 to 3.03; chronic productive cough: adjusted OR, 1.97; 95% CI, 1.19 to 3.26; dyspnea: adjusted OR, 2.67; 95% CI, 1.64 to 4.33; wheeze: adjusted OR, 3.20; 95% CI, 1.63 to 6.27. Manual workers in industry yielded an OR of 1.78 (95% CI, 0.80 to 3.97) as a risk factor for incident COPD, when added to the multiple logistic model using professionals and executives as the reference group. The ORs for all of the other groups were also nonsignificant ( 1.5). Subjects classified as having a low education level had a higher risk compared with those with a higher education (OR, 1.90; 95% CI, 1.11 to 3.25) in a univariate analysis. When analyzed in the multiple logistic model, a low education level had an adjusted OR of 1.73 (95% CI, 0.98 to 3.04), which was just below the threshold for significance. Table 5 Risk Factor for Incident COPD Analyzed in a Multiple Logistic Regression Model With Gender, Family History of OLD, Age Group, and Smoking Categories as Independent Variables, and Incident Case of COPD According to BTS and GOLD Criteria as Dependent Variables Variables BTS Criteria GOLD Criteria OR 95% CI OR 95% CI Gender Women 1 1 Men 1.03 0.63 1.69 1.29 0.85 1.96 Family history of OLD No 1 1 Yes 1.00 0.60 1.66 1.22 0.80 1.87 Age group Born 1949 1950 1 1 Born 1934 1935 4.50 2.21 9.16 5.70 3.15 10.30 Born 1919 1920 3.49 1.87 6.49 3.37 2.00 5.67 Smoking Persistent nonsmoker 1 1 Persistent ex-smoker 1.01 0.46 2.17 1.03 0.58 1.83 Quitter 2.32 1.02 5.30 1.34 0.63 2.86 Restarter 3.13 0.65 15.9 3.57 1.06 12.00 Persistent smoker 5.37 2.82 10.26 4.56 2.67 7.79 Gender Aspects There was no significant gender difference in the 10-year cumulative incidence of COPD either in the whole study population (Table 4) or when divided by groups for smoking habits. Further analysis of the three largest smoking categories (ie, nonsmoker, persistent ex-smoker, and persistent smoker) by age group and gender was not able to demonstrate significant gender differences. However, there was a trend for the 10-year cumulative incidence among persistent smokers to be higher in women compared with men, although not significantly so (Fig 1). Male and female incident cases of COPD reported respiratory symptoms to the same extent at the beginning of the observation period. In the multivariate analysis for the group as a whole, gender was not a significant risk factor for incident COPD when adjusted for a family history of OLD, age group, and smoking. However, when separate analyses were performed for men and women using the multivariate model, women had a higher adjusted OR for persistent smoking as a risk factor for incident COPD compared with men (women: OR, 8.52 [95% CI, 3.43 to 21.2]; men: OR, 3.14 [95% CI, 1.26 to 7.84]) using the BTS criteria. When added to the model, one at a time, the symptoms cough, sputum production, and chronic productive cough were all significant risk factors for COPD in women but not in men. The ORs among women were as follows: cough: OR, 3.60 (95% CI, 1.28 to 10.12); sputum production: OR, 2.97 (95% CI, 1.25 to 7.09); chronic 1548 Clinical Investigations

Figure 1. The 10-year cumulative incidence of COPD according to the BTS criteria by smoking habits, age group (year of birth), and gender. productive cough: OR, 2.93 (95% CI, 1.34 to 6.41). The corresponding for men were as follows: cough: OR, 1.38 (95% CI, 0.65 to 2.89); sputum production, 1.24 (95% CI, 0.62 to 2.51); chronic productive cough: OR, 1.47 (95% CI, 0.75 to 2.88). In contrast, we found that the symptoms dyspnea and wheeze were significant risk factors only among men. The adjusted OR for dyspnea was 5.27 (95% CI, 1.82 to 15.27), and that for wheeze was 3.28 (95% CI, 1.73 to 6.23). Among women, the OR for dyspnea was 2.11 (95% CI, 0.84 to 5.29), and that for wheeze was 2.06 (95% CI, 0.99 to 4.31). Discussion The cohort of subjects reporting respiratory symptoms in our study was derived from a large population from three age groups in representative areas from northern Sweden. The participation rates in the first survey 18,26 and the follow-up survey 5 were high ( 85% in both surveys). The cohort is regarded as being representative for subjects with respiratory symptoms in the general population in current age groups in which COPD is common. The GOLD guidelines give attention to subjects at risk of COPD (ie, subjects with respiratory symptoms and normal lung function [GOLD stage 0]). We have found only one report 12 on the development of COPD in subjects classified as GOLD stage 0. According to the Copenhagen City Heart Study, GOLD stage 0 did not predict the development of COPD. 12 More detailed information about the GOLD stage 0 population, or about subjects with respiratory symptoms and normal lung function, is necessary in order to evaluate the extent to which respiratory symptoms precede the development of COPD. In this cohort of middle-aged and elderly subjects reporting respiratory symptoms, the 10-year cumulative incidence of COPD was 8.2% and 13.5%, respectively, according to the BTS and GOLD spirometric criteria. As expected, we found that the cumulative incidence was significantly related to increasing age and smoking. 2 5 Persistent smokers yielded the highest 10-year cumulative incidence of COPD (using BTS criteria, 16.7%; using GOLD criteria, 24.5%) compared to those among nonsmokers (using BTS criteria, 4.8%; using GOLD criteria, 9.4%). The two definitions of COPD (BTS and GOLD) reflect different levels of disease severity (the GOLD definition also includes less severe disease [subjects with FEV 1 80% predicted]). This is the main reason for the higher incidence of COPD according to GOLD criteria compared with the BTS criteria. The severity of COPD is related to the cost of www.chestjournal.org CHEST / 127 / 5/ MAY, 2005 1549

COPD, 28 and the level of FEV 1 is related to mortality. 29 Changes in smoking habits, age distribution, and the definition of COPD will influence the incidence of COPD. However, as the different spirometric criteria of COPD also reflect the disease severity, both the health economic impact and the mortality rate will vary depending on the criteria used for definitions of disease. This must be taken into account when planning for future health-care resources. Only a few studies 10 12 about the incidence of COPD can be found in the literature. There are two reports from the 1980s 10,11 using lower quotients of the FEV 1 /FVC ratio ( 0.6 and 0.65, respectively) to define COPD compared with modern guidelines, giving an average annual incidence of 0.2% and 0.5%, respectively. The Copenhagen City Heart Study 12 reported COPD developed in 13.2% and 20.5%, respectively, of subjects in the GOLD stage 0 population who had been smoking at baseline at follow-ups after 5 and 15 years. In the Danish study, 12 smoking habits at baseline were used for analysis in contrast to our smoking categories. Furthermore, the age distributions differed. In the Danish study, 12 subjects 20 years of age were included, while the youngest age group in our study was 35 years at the start of the study. These differences make direct comparisons of the risks difficult, even though the observation times overlapped. Furthermore, in our study, low lung function was associated with increased mortality even after the exclusion of the prevalent cases of COPD in 1986. As a result, the 10-year cumulative incidence of COPD may be somewhat underestimated in our study. Despite these viewpoints, the cumulative incidence in both our study and the Danish study 12 were fairly similar. In contrast to the results of the Danish study, 12 which concluded that GOLD stage 0 does not predict the development of COPD, we found that bronchitic symptoms, including cough, sputum production, and chronic productive cough (symptoms of chronic bronchitis), at the beginning of the observation period were related to the development of COPD. Each bronchitic symptom, as well as the symptoms dyspnea and wheeze, was a significant risk factor for the incidence of COPD, according to the BTS criteria after adjustment for possible confounders. GOLD stage II disease (ie, FEV 1 /FVC ratio, 0.70; and FEV 1, 80% predicted) is similar to the BTS criteria and has been suggested as a practical definition for identifying symptomatic subjects with COPD. 16 Thus, we conclude that bronchitic symptoms, dyspnea, and wheeze increase the risk of developing COPD that may be clinically relevant. There have been several reports 30 34 showing that respiratory symptoms such as chronic cough and chronic phlegm are associated with a more rapid decline in FEV 1, but there are currently no reports of any association between respiratory symptoms and incident COPD. However, the clinical implication of our study is that the early identification of COPD is possible by performing repeated PFTs among subjects with respiratory symptoms (GOLD stage 0). According to our results, a family history of OLD was not a risk factor for incident COPD, although there have been reports 5,18,35,36 that it is a risk factor for prevalent disease. Prevalent disease has no specific time aspect and may include chronic asthma from childhood or the development of lung function impairment due to less efficacious pharmacologic treatment from the time before inhaled corticosteroids were introduced. Incident cases of COPD in our study were adults with normal lung function at the start of the observation period. Incident COPD was associated with smoking and was less likely to be of asthmatic origin, even if some cases of misclassification and the coexistence of chronic asthma could be ruled out. Our hypothesis is that a family history of OLD may include asthma with a known hereditary component, which may be important as a risk factor for OLD but may be less important as a risk factor for incident COPD among middle-aged and elderly subjects. Further studies are needed to answer the question of whether there is any genetic predisposition, in addition to the known 1 -antitrypsin deficiency, for developing COPD. 35 A low level of education and manual work in industry both tended to be risk factors for COPD, although our results did not reach statistical significance. These risk factors may be indicators of possible airborne occupational exposure and/or lifestyle factors. 2,30 Future studies focusing on work environment and airborne occupational exposure are necessary in order to increase our knowledge of their relationship with COPD. Moreover, an improved understanding of the environmental effects of outdoor air pollution, tobacco, including environmental tobacco exposure, and indoor air pollution, including cooking fuel exposure in some countries, will help to focus and direct future studies designed to identify risk factors that have not been addressed in this study. Gender was not a significant risk factor for COPD when adjusted for possible confounders in a multivariate model. Multivariate analyses performed separately in women and men revealed higher ORs in women compared with those in men for smoking, but the results are difficult to interpret. There was, however, a trend toward a higher cumulative incidence of COPD among persistent female smokers compared with persistent male smokers, and this 1550 Clinical Investigations

finding is in accordance with other reports of women being more susceptible to tobacco smoke. 36 38 Among nonsmokers, the opposite trend has been seen (higher incidence of COPD among men), possibly due to airborne occupational exposure. In the gender-specific multivariate analysis, bronchitic symptoms were found to be significant risk factors for incident COPD in women, while dyspnea and wheeze were significant risk factors in men (weaker association in women). Other studies 39 have found that women report more respiratory symptoms compared with men with the same lung function. Our results indicate that women and men may present different clinical risk profiles preceding the development of COPD. In conclusion, the 10-year cumulative incidence of COPD in this cohort of subjects with respiratory symptoms was estimated at 8.2% and 13.5%, respectively, according to the BTS and GOLD criteria. The incidence was strongly related to increasing age and smoking. Bronchitic symptoms, dyspnea, and wheeze were significant risk factors for developing COPD, and they persisted after adjustment for possible confounders. Thus, GOLD stage 0 disease identifies subjects who are at risk for developing COPD, and it appears that men and women may present different risk profiles. References 1 Lange P, Groth S, Nyboe GJ, et al. Effects of smoking and changes in smoking habits on the decline of FEV 1 : Copenhagen City Heart Study. Eur Respir J 1989; 2:811 816 2 Bakke PS, Baste V, Hanoa R, et al. Prevalence of obstructive lung disease in a general population: relation to occupational title and exposure to some airborne agents. Thorax 1991; 46:863 870 3 Sobradillo V, Miravitlles M, Jimenez CA, et al. Epidemiological study of chronic obstructive pulmonary disease in Spain (IBERPOC): prevalence of chronic respiratory symptoms and airflow limitation. Arch Bronconeumol 1999; 35:159 166 4 Mannino DM, Gahnon RC, Petty TL, et al. Obstructive lung disease and low lung function in adults in the United States: data from the National Health and Nutrition Examination Survey, 1988 1994. Arch Intern Med 2000; 160:1683 1989 5 Lundbäck B, Lindberg A, Lindström M, et al. Not 15 but 50% of smokers develop COPD? Report from the Obstructive Lung Disease in Northern Sweden Studies. Respir Med 2003; 97:115 122 6 Murray CJL, Lopez AD. Alternative visions of the future: projecting mortality and disability, 1990 2020. In: Murray CJL, Lopez AD, eds. The global burden of disease. Cambridge MA: Harvard University Press, 1996; 361 395 7 Gulsvik A. The global burden and impact of chronic obstructive pulmonary disease worldwide. Monaldi Arch Chest Dis 2001; 56:261 264 8 Siafkas NM, Vermiere P, Pride NB, et al. ERS consensus statement: optimal assessment and management of chronic obstructive pulmonary disease (COPD). Eur Respir J 1995; 8:1398 1420 9 British Thoracic Society. BTS guidelines for the management of chronic obstructive pulmonary disease. Thorax 1997; 52(suppl): 1S 28S 10 Huhti E, Ikkala J. A follow-up study on respiratory symptoms and ventilatory function in a middle-aged rural population. Eur J Respir Dis 1980; 61:33 45 11 Krzyzanowski M, Jedrychowski W, Wysocki M. Factors associated with the change in ventilatory function and the development of chronic obstructive pulmonary disease in a 13-year follow up of the Cracow study. Am Rev Respir Dis 1986; 134:1011 1019 12 Vestbo J, Lange P. Can GOLD stage 0 provide information of prognostic value in Chronic Obstructive Lung Disease? AJR Am J Roentgenol 2002; 166:329 332 13 Pauwels RA, Buist AS, Ma P, et al. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: National Heart Lung, and Blood Institute and World Health Organization Global Initiative for Chronic Obstructive Lung Disease (GOLD); executive summary. Am J Respir Crit Care Med 2001; 46:798 825 14 American Thoracic Society. Standards for the diagnosis and care of patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1995; 152(suppl):77S 120S 15 Viegi G, Pedreschi M, Pistelli F, et al. Prevalence of airways obstruction in a general population: European Respiratory Society vs American Thoracic Society definition. Chest 2000; 117(suppl):339S 345S 16 Celli BR, Halbert RJ, Isonaka S, et al. Population impact of different definitions of airway obstruction. Eur Respir J 2003; 22:268 273 17 Lundbäck B, Nystrom L, Rosenhall L, et al. Obstructive lung disease in northern Sweden: respiratory symptoms assessed in a postal survey. Eur Respir J 1991; 4:257 266 18 Lindström M, Kotaniemi J, Jönsson E, et al. Smoking, respiratory symptoms, and diseases: a comparative study between northern Sweden and northern Finland; a report from the FinEsS study. Chest 2001; 119:852 861 19 Larsson LG, Lindberg A, Franklin KA, et al. Gender differences in symptoms related to sleep apnea in a general population and in relation to referral to sleep clinic. Chest 2003; 124:204 211 20 Pallasaho P, Lundbäck B, Laspa SL, et al. Increasing prevalence of asthma but not of chronic bronchitis in Finland? Report from the FinEsS-Helsinki Study. Respir Med 1999; 93:798 809 21 Medical Research Council Committee on Aetiology of Chronic Bronchitis. Standardised questionnaires on respiratory symptoms. BMJ 1960; ii:1665 22 Lebowitz MD, Knudson RJ, Burrows B. Tucson epidemiologic study of obstructive lung diseases: I. Methodology and prevalence of disease. Am J Epidemiol 1975; 102:137 152 23 Burney PG, Laitinen LA, Perdrizet S, et al. Validity and repeatability of the IUATLD (1984) Bronchial Symptoms Questionnaire: an international comparison. Eur Respir J 1989; 2:940 945 24 American Thoracic Society. Standards for the diagnosis and care of patients with chronic obstructive pulmonary disease (COPD) and asthma. Am Rev Respir Dis 1987; 136:225 244 25 Berglund E, Birath G, Grimby G, et al. Spirometric studies in normal subjects, forced expirograms in subjects between 7 and 70 years of age. Acta Med Scand 1963; 173:185 192 26 Lundbäck B, Stjernberg N, Nyström L, et al. Epidemiology of respiratory symptoms, lung function and important determinants: report from the Obstructive Lung Disease in Northern Sweden Project. Tuber Lung Dis 1994; 75:116 126 27 Statistics Sweden The Socioeconomic classification of occupations. Stockholm, Sweden: Statistics Sweden, 1982 28 Jansson SA, Andersson F, Borg S, et al. Costs of COPD in Sweden www.chestjournal.org CHEST / 127 / 5/ MAY, 2005 1551

according to disease severity. Chest 2002; 122:1994 2002 29 Hansen EF, Phanareth K, Laursen L, et al. Reversible and irreversible airflow obstruction as predictor of overall mortality in asthma and chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1999; 159:1267 1271 30 Heederik D, Kromhout H, Burema J, et al. Occupational exposure and 25-year incidence rate of non-specific lung disease: the Zuphten Study. Int J Epidemiol 1990; 19:945 952 31 Annesi I, Kauffmann F. Is respiratory mucus hypersecretion really an innocent disorder? A 22-year mortality survey of 1,061 working men. Am Rev Respir Dis 1986; 134:688 693 32 Krzyzanowski M, Camilli AE, Lebowitz MD, et al. Relationships between pulmonary function and changes in chronic respiratory symptoms: comparison of Tucson and Cracow longitudinal studies. Chest 1990; 98:62 70 33 Sherman CB, Xu X, Speizer FE, et al. Longitudinal lung function decline in subjects with respiratory symptoms. Am Rev Dis 1992; 146:855 859 34 Vestbo J, Prescott E, Lange P. Association of chronic mucus hypersecretion with FEV 1 decline and chronic obstructive pulmonary disease morbidity: Copenhagen City Heart Study Group. AJR Am J Roentgenol 1996; 153:1530 1535 35 Sandford AJ, Weir TD, Pare PD. Genetic risk factors for chronic obstructive pulmonary disease. Eur Respir J 1997; 10:1380 1391 36 Viegi G, Vellutini M, Di Pe F, et al. Comparison of lung function in the Italian and Swedish population samples. Eur Respir Rev 2001; 11:80, 65 73 37 Xu X, Weiss ST, Rijken B, et al. Smoking, changes in smoking habits, and rate of decline in FEV 1 : new insight into gender differences. Eur Respir J 1994; 7:1056 1061 38 Prescott E, Bjerg AM, Andersen PK, et al. Gender difference in smoking effects on lung function and risk of hospitalization for COPD: results from a Danish longitudinal population study. Eur Respir J 1997; 10:822 827 39 Langhammer A, Johnsen R, Gulsvik A, et al. Sex differences in lung vulnerability to tobacco smoking. Eur Respir J 2003; 21:1017 1023 1552 Clinical Investigations