Respiratory Medicine (2011) 105, 1872e1884

ava i lab le at www.sc iencedi rect .com

journal homepage : www.e lsev ie r . com/ loca te / rmed

COPD in the general population: Prevalence,incidence and survival

Ana S.M. Afonso a,d, Katia M.C. Verhamme a,*,d,Miriam C.J.M. Sturkenboom a,b, Guy G.O. Brusselle c

aDepartment of Medical Informatics, Erasmus University Medical Center, PO Box 2040, 3000 CA Rotterdam,The NetherlandsbDepartment of Epidemiology, Erasmus University Medical Center, Rotterdam, The NetherlandscDepartment of Respiratory Medicine, Ghent University Hospital, Ghent, Belgium

Received 22 October 2010; accepted 26 June 2011Available online 17 August 2011

KEYWORDSChronic obstructivepulmonary disease;Epidemiology(pulmonary);Incidence;Prevalence;Survival analysis

Abbreviations: ATC, anatomical therainitiative for chronic obstructive lunggrated primary care information; IR,organization.* Corresponding author. Tel.: þ31 1E-mail address: k.verhamme@eras

d Both authors contributed equally t

0954-6111/$ - see front matter ª 201doi:10.1016/j.rmed.2011.06.012

Summary

Worldwide, COPD is a leading cause of chronic morbidity and mortality. Although its preva-lence is already well documented, very few studies have measured its incidence. We thereforeinvestigated the prevalence, incidence and lifetime risk of COPD in the general population.

In a population-based study including subjects � 40, with 12 months of history available inthe Dutch IPCI database, we identified COPD cases by a two-step validation algorithm.

Among 185,325 participants with 601,283 years of follow-up, 7308 subjects with COPD wereidentified, and 1713 had incident COPD. The overall IR of physician-diagnosed COPD was 2.92/1000PY (95%CI 2.78e3.06). The incidence of COPD was higher in men (3.54; 95%CI 3.33e3.77)than in women (2.34; 95%CI 2.17e2.52), and the overall baseline prevalence of COPD was3.02% (95%CI 2.94e3.10). For people who had entered the study free of COPD at the age of40, the risk of developing COPD within the next 40 years was 12.7% for men and 8.3% forwomen. In patients with very severe COPD, 26% died after 1 year of follow-up, whereas2.8% died among the non-COPD subjects.

In the general population in theNetherlands, three on 1000 subjectswere diagnosedwith COPDper year. The incidence increased rapidly with age and was higher in men than in women. One in

peutic chemical; CI, confidence interval; COPD, chronic obstructive pulmonary disease; GOLD, globaldisease; GPs, general practitioner(s); ICPC, international classification of primary care; IPCI, inte-incidence rate; PY, person-years; RR, relative risk; SD, standard deviation; WHO, world health

0 704 41 28; fax: þ31 10 704 47 22.musmc.nl (K.M.C. Verhamme).o this work.

1 Elsevier Ltd. All rights reserved.

COPD in the general population 1873

eight men and one in 12 women, being COPD free at the age of 40, will develop COPD during theirfurther life. Mortality rates differed substantially between COPD patients and non-COPD subjectsof the same age, underlining the burden of this disease.ª 2011 Elsevier Ltd. All rights reserved.

Introduction

Chronic obstructive pulmonary disease (COPD) is charac-terized by a largely irreversible obstruction of the airways,and encompasses both emphysema and chronic bronchitis.1

The obstruction of the airways is usually progressive, andoften associated with an abnormal inflammatory responseof the lungs to harmful particles or gases such as tobaccosmoke. COPD is a leading and still-increasing cause ofchronic morbidity and mortality worldwide,2 and accordingto the World Health Organization (WHO), it is the fifth mostcommon cause of death and the 10th most burdensomedisease.3 Chapman et al.4 and Mannino et al.5 projectedthat between 1990 and 2020, COPD will become the thirdmost common cause of death worldwide. A Dutch studypredicts that an increase of 76% in the prevalence of COPDcan be expected within approximately twenty years.6

Although the prevalence of COPD has been well stud-ied,7e11 there are less population-based studies11e19 whichhave investigated its incidence. Among the studies that havebeen conducted, a wide range in incidence rates can beobserved varying between 2e16/1000PY depending on theCOPD definition being used and the population being stud-ied.11e14,16e21 Little is known about trends in COPD preva-lence, incidence and all-cause mortality.20

COPD is one of the few chronic diseases that caused anincrease in mortality in recent years.22 In 1990, the WHOestimated the European standardized mortality rate ofCOPD to be 50 per 100,000 in males and 20 per 100,000 infemales.23,24 Data provided by the WHO in 1997, showedthat COPD was the cause of death in 4.7% of men and 2.4%of women23,24 in Europe, and the World Health Report of1998 stated that 2.9 million adults die each year of thisdisease.24,25 Death rates in patients with COPD are usuallylower among women than among men in all countries. Atpresent, there are few observational studies reportingmortality rates in COPD patients.20,25e28

The objectives of our study were to investigate theprevalence, incidence, mortality and lifetime risk of COPDas a function of age, sex and smoking status in the generalpopulation.

Methods

Setting

The study was conducted within the Integrated PrimaryCare Information (IPCI) database, a longitudinal observa-tional database (dynamic cohort), that started in 1992 andcontains data from computer based medical records ofgeneral practitioners (GPs) throughout the Netherlands.29

The database covers data from more than 400 GPs whocurrently capture data on more than 1 million patients.29 Inthe Dutch health care system, patients are registered with

a single GP who acts as a gatekeeper for and receiver ofinformation from secondary care.30 The GPs’ electronicmedical records contain coded and anonymous data onpatient demographics, narratives, patient reported symp-toms, signs, GP and specialist diagnoses (using the Inter-national Classification of Primary Care (ICPC) codes andfree-text31), as well as prescriptions, physical findings,laboratory values, and summaries of specialist letters.29,31

Downloads are made periodically and the information issent to the gatekeeper who de-identifies all informationbefore further access is provided.

Prescription data encompass product name, quantitydispensed, dosage regimens, strength and indication. TheNational Database of drugs, maintained by the Royal DutchAssociation for the advancement of Pharmacy, enables thecoding of prescriptions, according to the Anatomical Ther-apeutic Chemical (ATC) classification scheme recommendedby the WHO.32e34 The system complies with European Unionguidelines on the use of medical research and has beenproven valid for pharmaco-epidemiological research.34 Allobservational research using IPCI data is conductedaccording to good pharmaco-epidemiological guidelines.35

Source population

The source population comprised 185,325 men and womenaged 40 or older, with at least one-year of valid databasehistory meaning that the practice had been contributing datato the IPCI database for at least one-year and that the patienthad been registered with the GP for at least one-year. Thisone-year pre-enrollment period was required to guaranteesufficient medical history prior to study entry. The studystarted on 1 January 2000 or the date at which one-year ofvalid history was obtained. All patients were followed fromstudy entry until the date of diagnosis of (incident) COPD,death or 1 July 2007, whichever occurred first.

COPD cohort

A broad automated search was conducted on ICPC COPDdisease codes and narratives to identify all potential COPDpatients. The medical records of all potential cases werereviewed by a medical doctor and classified as: definiteCOPD e diagnosis by a specialist or a GP diagnosis confirmedby spirometry (GOLD definition of FEV1/FVC<70%); prob-able COPD e COPD diagnosed by the GP with at least tworecords of COPD within one-year of the first record ofCOPD. Doubtful COPD patients were further reviewed andclassified by a pulmonologist (GB).

All COPD patientswere further categorized into prevalentor incident COPD. Patients with a COPD diagnosis prior tostudy entry were classified as prevalent (pre-existing) COPD.If patients were disease free (no COPD specific respiratorysymptoms, no COPD disease codes, no respiratory treatmentfor COPD, and no specialist letters referring to a diagnosis of

1874 A.S.M. Afonso et al.

COPD) at cohort entry and later developed COPD, they wereconsidered as having incident (new onset) COPD.

COPD severity was assessed at the time of cohort entryand changes in COPD severity were captured during follow-up. If spirometry data were available, severity of COPD wasassessed according to the GOLD criteria. In all otherpatients, previously published algorithms for COPD severityassessment were used.36e38 For further details on COPDseverity assessment, see online supplement.

Statistical analysis

The baseline characteristics of the COPD cohort werecompared to the baseline characteristics of the populationvia the ManneWhitney U test for continuous variables, andthe Chi-square test for categorical variables. Age andgender-specific incidence rates of COPD (IRs) per 1000person-years (PY) and cumulative prevalence of COPD, allstratified by smoking status, as well as weighted (3 yearmean) annual incidence rates were calculated (See onlinesupplement). A direct and indirect standardizationmethod was used to compare the incidence rate to theCOPD incidence rate of a reference population.

Thecumulative incidenceand the lifetimeriskofCOPDwascalculated on the basis of a Cox regressionmodel adjusted forcompeting risk of death, as described by Rosthoj et al.39

Crude and adjusted hazard ratios of incident COPD werecalculated using a Cox regression analysis. In the finalmodel, we adjusted for all factors that, in the univariateanalysis, were associated with the outcome.

To compare the mortality rates in subjects with orwithout COPD, to each COPD patient, 4 controls weresampled from the source population, matched on age, sex,smoking status and calendar year (year of entry into thecohort). Mortality rates were calculated by dividing thenumber of deaths by the person years attributed. Survivalanalyses were conducted using KaplaneMeier analysis.40 Allstatistical analyses were conducted with SPSS/PC 15.0(SPSS Inc., Chicago, Ill), SAS and Episheet (by Rothman).

Results

Baseline characteristics

In the source population of 185,325 patients with at leastone-year of valid history (601,283 years of follow-up time),18,643 COPDpotential caseswere identified (Fig. E1e onlinesupplement). The mean follow-up time was 3.4 years (stan-dard deviation (SD),�3.2 years). After manual review of thecomplete medical record, 7308 patients were identified ashaving COPD of which 57% were male. Table 1 shows thebaseline characteristics of the total source population atstudy entry (n Z 185,325) and the patients with COPD(nZ 7308), both incident and prevalent. COPDpatientswereolder, more frequently male, and had more co-morbiditythan non-COPD patients at study entry (Table 1).

A total of 5595 patients (77%) in the COPD cohort hadprevalent COPD at baseline and 1713 (23%) were incidentCOPD patients. Incident COPD patients were slightlyyounger than the prevalent ones (mean age 64.7,SD � 11.8, versus 67.8, SD � 12.3), and more often smoked

or had smoked in the past (Table 1). The mean and medianage at first diagnosis did not change over time (data notshown). The majority of incident COPD patients had mildCOPD at the time of first diagnosis.

Co-morbidity was similar between incident and preva-lent COPD patients apart from a history of heart failure,which was more often recorded in prevalent COPD patients(12.4% versus 10.4%), and myocardial infarction, hyper-tension, lipid disorders and a history of depressive disor-ders, which were more common in incident COPD patients(Table E1 e online supplement).

Risk factors of incident COPD

Table 2 describes the risk factors of incident COPD. Malegender, increasing age and smoking history were strong riskfactors of COPD and these associations remained significantafter adjustment. Cardiovascular co-morbidity was stronglyrelated to the risk of incident COPD in the crude analysis,but these associations weakened upon adjustment. Amedical history of asthma and pneumonia in the one-yearprior to study start were strongly associated with incidentCOPD (Table E2 e online supplement).

Incidence, prevalence, risk and mortality

The overall incidence of COPD in persons 40 years and olderwas 2.92/1000PY (95%CI, 2.78e3.06) (Table E3a e onlinesupplement). The incidence of COPD was higher in menthan in women (with a relative risk (RR) of 1.5-fold higher inmen). The incidence increased almost 10-fold from 0.78/1000PY at age 40e44 to 6.82/1000PY at age 75e79 (Fig. 1).The incidence rate of COPD was much higher in smokerscompared to non-smokers and this for all age categoriesand in both sexes (Fig. 2) (Table E3 b e online supplement).

The prevalence of COPD at cohort entry was 3.02%, 95%CI (2.94e3.10). The age-specific cumulative prevalence ofCOPD increased with age until the age of 81 years in menand 83 years in women. After this age, the prevalence didno longer increase and even gradually declined, especiallyin men. The prevalence of COPD was, for all age categories,higher in men than in women (Fig. E2 e online supplement)and higher in ever-smokers than in non-smokers (Fig. E3 eonline supplement).

For a man free of COPD at the age of 40, the risk to bediagnosed with COPD was 0.8% (10 years), 2.8% (20 years),7.2% (30 years) and 12.7% (40 years) respectively. Fora woman of the same age, the risks were 0.8%, 2.3%, 5.0%and 8.3% (Fig. 3). Here as well, the cumulative incidence ofCOPD was much higher in ever-smokers than in non-smokers, both for men and women (Fig. 4).

Mortality rates were 2.0-fold higher for COPD patientsthan for the age and sex matched reference population.Mortality rates increased with COPD severity from 48.1/1000PY (95%CI, 42.5e54.3/1000PY) in mild to 262/1000PY(95%CI, 179e371/1000PY) in very severe COPD patients(Table 3). Survival is lower in COPD patients than non-COPDsubjects and is highly influenced by COPD severity (Fig. 5).The one-year risk of dying was 26% (95%CI 18.4e35.4%) inthe very severe COPD patients, while the risk in the sex andage matched reference group was nearly ten times lowernamely 2.8% (95%CI 0.92e8.17%).

Table 1 Patient characteristics of source population and COPD cohort.

Characteristics Source population n Z 185,325 Overall COPD Incident COPD Prevalent COPD

No. % No. % No. % No. %

Number of COPD patients 7308 3.9 7308 100.0 1713 23.4 5595 76.6Gender

Males 89,697 48.4 4153 57.3 995 58.1 3158 56.4Females 95,628 51.6 3155 42.7 718 41.9 2437 43.6

Age (years)40e59 119,798 64.6 1982 26.5 501 29.2 1481 26.560e69 30,930 16.7 1872 25.6 473 27.6 1399 25.0�70 34,597 18.7 3454 47.9 739 43.1 2715 48.5Mean age (�SD) 56.5 �13.3 67.8 �12.1 64.7 �11.8 67.8 �12.3

Ever smokinga 39,250 21.2 3468 47.4 1028 60.0 2440 43.6Smoking historya

Never smoker 506 6.9 130 7.6 376 6.7Current smoker 3009 41.2 854 49.9 2155 38.5Past smoker 459 6.3 174 10.2 285 5.1Smoking not specified 3334 45.6 555 32.4 2779 49.7

Severity at cohort entrya

Severity by spirometryMild (GOLD stage I) 639 9.0 243 14.2 393 7.0Moderate (GOLD stage II) 923 13.3 187 10.9 736 13.2Severe (GOLD stage III) 558 8.1 73 4.3 485 8.7Very Severe (GOLD stage IV) 21 0.3 3 0.2 18 0.3

Severity by proxyMild 2352 30.7 1115 65.1 1237 22.1Moderate 1968 26.7 40 2.3 1928 34.5Severe 768 10.8 50 2.9 718 12.8Very Severe 81 1.2 1 0.1 80 1.4

Mean time since first diagnosis (years) (�SD) 0 0 3.16 �4.47a At baseline for prevalent COPD patients or at the date of diagnosis for incident COPD patients.

COPD in the general population 1875

Discussion

In this large population-based cohort study of the generalDutch population of 40 years and older, the overall inci-dence rate of physician-diagnosed COPD was 2.92/1000PY.Based on these data the risk to be diagnosed with COPD in

Table 2 Risk factors of incident COPD.

Patients/Cohort No. HRcrude

COPD risk factorsGender

Female 718/92,473 ReferenceMale 995/85,544 1.51

Age at baseline40e59 592/117,825 Reference60e69 506/29,061 3.47�70 615/31,131 6.64

Smoking statusNever smoker 685/142,235 ReferenceEver smoker 1028/35,782 3.67

Patients with prevalent COPD (n Z 5595) excluded from this analysisa Adjusted for all factors that were univariately associated with thunderlying co-morbidity (history of stroke/TIA, myocardial infarctiondisease, diabetes mellitus, renal failure, pneumonia (previous year),

the coming 40 years was 12.7% for a 40-year-old male and8.3% for a 40-year-old female. The incidence increased withage, and was higher in men than in women. Known riskfactors of COPD were confirmed such as smoking status,male gender and increasing age. Mortality was high; espe-cially in very severe COPD patients with one-year mortality

95%CI HRadjusteda 95%CI

Reference Reference Reference1.37e1.67 1.65 1.49e1.82

Reference Reference Reference3.06e3.94 3.67 3.23e4.175.93e7.44 8.55 7.58e9.65

Reference Reference Reference3.33e4.04 3.89 3.52e4.31

.e outcome namely gender, smoking status, age at baseline and, heart failure, hypertension, angina pectoris, peripheral arteryasthma, depression and hyperlipidemia).

1876 A.S.M. Afonso et al.

risks nearly 10-fold that of non-COPD subjects of the sameage and gender.

Previous studies mainly focused on the prevalence ofCOPD as they often applied a cross sectional spirometrybased approach to obtain the true prevalence of COPD. Fewspirometry based incidence studies have beenreported.11e14,16,17,19 (Table 4) Studies reporting diagnosedCOPD, based on health care databases, were conducted inthe UK and in Canada (Table 4).18,20,21 The most recentincidence study, published in 2010, is a Canadian studyinvestigating trends in incidence, prevalence and mortalityof COPD over time.20 For this study, the Ontario HealthInsurance Plan database was used and COPD was definedbased on visits and hospitalization with ICD-9 or ICD-10 codesfor COPD.20 The incidence of COPD was high: 11.8/1000 in1996 and 8.5/1000 in 2007.20 The rates of incidence andprevalence were much higher than observed in our study(even after standardization), but the study was very inclu-sive and COPDwas not validated. COPDCodes included termssuch as chronic bronchitis and asthmatic bronchitis whichnot necessarily are synonyms of COPD.20 Our COPD definitionwas much stricter, requiring specialist diagnosed COPD or

Figure 1 Age- and gender-specific incidence rates of physician-calendar year-specific incidence of COPD (bottom).

COPD diagnosed by GPwith spirometry data and/or at least 2records of COPD within one-year, explaining our lower IRs.

van Durme et al. also studied the incidence of COPD ina prospective cohort of elderly in the Netherlands. COPDwas defined based on specialists/GPs (similar to ours)diagnosis as well as on tri-annual spirometry assessments.Our age standardized rate was 35% lower showing the levelof under-diagnosis.11 Using the indirect age standardiza-tion, the calculated standardized incidence rate was 0.49(95%CI 0.46e0.52). We thus identified only half of the COPDpatients as expected based on the incidence rate of theRotterdam study. Our IR is much lower than in the studyfrom van Durme et al., as in the latter, spirometry wasconducted on 44% of the study population, whereas in oursetting, spirometry was only performed if the subjectseeked medical attention because of respiratorycomplaints. Our IRs are more in line with the results ofa recent population-based cohort study, using primary caredata (UK GPRD database) as well and similar method-ology.18 The incidence rate of physician-diagnosed COPD in40e89 year-olds was 2.6/1000PY, which is almost identicalto the incidence rate we found (2.9/1000PY).18

diagnosed COPD (/1000PY) (top) and weighted (3 year mean)

COPD in the general population 1877

These differences in COPD incidence rate can beexplained by differences in COPD definition and/or byunder-diagnosis (the use of spirometry as a diagnostic testin symptomatic subjects seeking medical attention versusthe use of spirometry as screening tool for the detection ofCOPD). In addition, we applied a very strict COPD validationalgorithm, indeed, of all potential cases with COPD, only39% were withheld as having COPD. Since respiratorysymptoms are often absent in patients with COPD and sincenot all patients with respiratory symptoms seek medicalattention, the incidence of physician-diagnosed COPD (thisstudy) is lower than the incidence of spirometry definedCOPD.11,41

After the age of 81 years in men and 83 years in women,the COPD prevalence did no longer increase and evengradually declined. Although the exact mechanism isunknown, this could be explained by the “healthy survivoreffect”, where patients with COPD die before the age of 80.

Figure 2 Age- and gender-specific incidence rates of physician-calendar year-specific incidence of COPD (bottom) by smoking.

In our study, all-cause mortality rates were higher inCOPD patients than in patients without COPD of similar ageand sex. Our mortality rates are much in line with the datafrom Soriano et al.25 and Gershon et al.,20 who also studiedmortality in COPD patients. In their study, Gershon et al.describe mortality rates that vary between 5.7% in 1996 and4.3% in 2007,20 which is in line with the mortality rate of6.1% that we found in our COPD patients. Soriano et al.reported an overall death rate of 8.5% per year.25

Furthermore, in comparison with data of patients withCOPD followed by GPs,42 Lundback et al. refer to a bettersurvival among the subjects with COPD identified by anepidemiological population study, with a 5-year risk ofdying of approximately 7% in subjects without both chronicbronchitis and asthma-like phenotype, and approximately15% in subjects with chronic bronchitis.43 Our 5-year risk ofdying was somehow higher with 14% in non-COPD patientsand 27% in COPD patients, probably because Lundback

diagnosed COPD (/1000PY) (top) and weighted (3 year mean)

Figure 3 Age-related risk to develop COPD over the coming 10, 20, 30, and 40 years in men (top) and women (bottom).CI Z cumulative.

1878 A.S.M. Afonso et al.

et al. included younger patients, applied another COPDdefinition and used different in and exclusion criteria.

Similar to previous literatures,20,25e28 we found thatmortality in patients with COPD is higher in men than inwomen. A partial explanation could be that, on average,women with COPD still have less severe COPD or smoke lessthan men.

As for all observational studies, our study has strengthsand weaknesses. The strength of this study is thepopulation-based design; its large cohort size and thedetailed information on underlying co-morbidity. Further-more, the large population in this study is representative ofthe general population, which eliminates selection bias. Toavoid inclusion of false positive cases we followeda rigorous validation algorithm and used data fromspirometry and specialist referral letters, if available. Inaddition, we assessed COPD severity according to spirom-etry data or according to an algorithm already successfully

used by other research groups.36e38 As this is a population-based study, we did not have spirometry data on allsubjects. For this reason we believe that we probablyunderestimated the true incidence rate of COPD in thegeneral population, resulting in an under-diagnosis of COPDdue to lack of spirometry performed and/or recorded. Thisis one of the reasons why our COPD incidence rate is lowerthan in those studies where COPD was systematicallyassessed by spirometry. In our study, patients diagnosed tohave COPD are mainly the ones presenting themselves withrespiratory symptoms. Some patients with COPD neverdevelop respiratory symptoms and a considerable propor-tion of patients with respiratory symptoms get accustomedto these complaints and never consult their GP.41,44

Among patients with spirometry data, we used the GOLDdefinition of COPD (FEV1/FVC < 70%) instead of the use ofLLN values for FEV1/FVC. It is known that the FEV1/FVCratio declines with age, and thus the true lower limit of

0%

5%

10%

15%

20%

25%

30%

35%

41 45 50 55 60 65 70 75 80 85 90

1. Age at cohort entry men smokers

Pro

bab

ility o

f C

OP

D

10 year CI COPD 20 year CI COPD 30 year CI COPD 40 year CI COPD

0%

2%

4%

6%

8%

10%

12%

14%

41 45 50 55 60 65 70 75 80 85 90

3. Age at cohort entry women smokers

Pro

bab

ility o

f C

OP

D

10 year CI COPD 20 year CI COPD 30 year CI COPD 40 year CI COPD

0%

5%

10%

15%

20%

25%

30%

35%

41 45 50 55 60 65 70 75 80 85 90

2. Age at cohort entry men non-smokers

Pro

bab

ility o

f C

OP

D

10 year CI COPD 20 year CI COPD 30 year CI COPD 40 year CI COPD

0%

2%

4%

6%

8%

10%

12%

14%

41 45 50 55 60 65 70 75 80 85 90

4. Age at cohort entry women non-smokers

Pro

bab

ility o

f C

OP

D

10 year CI COPD 20 year CI COPD 30 year CI COPD 40 year CI COPD

Figure 4 Age-related risk to develop COPD over the coming 10, 20, 30, and 40 years in men (1 and 2) and women (3 an 4) bysmoking status. CI Z cumulative.

Table 3 Number of deaths and mortality rates (/1000PY) in patients with and without COPD.

COPD Non-COPDa

All Mild Moderate Severe Very severe

Total numberpatients, n (%)

7308 2989 (40.9) 2891 (39.6) 1326 (18.1) 102 (1.4) 24,722

Number of deaths 1371 393 523 402 53 2750Males 897 (65.4) 259 (65.9) 329 (62.9) 280 (69.6) 29 (54.7) 1886 (68.6)Females 474 (34.6) 134 (34.1) 194 (37.1) 122 (30.3) 24 (45.3) 864 (31.4)

Mean (SD) ageat death

78.2 (�9.3) 78.2 (�9.8) 78.9 (�8.8) 77.7 (�9.2) 73.6 (�9.7) 79.7 (�8.6)

Age at death,n (%)40e59 53 (3.9) 18 (4.6) 15 (2.9) 15 (3.7) 5 (9.5) 89 (3.3)60e69 195 (14.2) 54 (13.7) 64 (12.2) 64 (16.0) 13 (24.5) 261 (9.5)�70 1123 (81.9) 321 (81.7) 444 (84.9) 323 (80.3) 35 (66.0) 2400 (98.2)

Mortality rate(/1000PY),95%CI_ 70.4

(65.9e75.1)48.1(42.5e54.3)

63.9(57.3e71.1)

133.4(118.5e149.8)

262.0(179.1e370.9)

36.3(34.7e38.0)

\ 48.5(44.3e53.0)

33.5(28.2e39.6)

47.5(41.1e54.5)

76.8(64.0e91.3)

236.7(195.5e404.9)

21.7(20.3e23.2)

All 60.9(57.7e64.2)

41.9(37.9e46.2)

56.6(51.9e61.6)

109.0(98.7e120.1)

249.9(189.2e324.2)

30.0(28.9e31.1)

a Matched for age, sex, smoking status and calendar year (year of entry in the cohort); 4 controls were sampled from the sourcepopulation for each COPD patient.

COPD in the general population 1879

1880 A.S.M. Afonso et al.

normal (LLN) for FEV1/FVC also declines. This may result inmisclassification of disease, with an over-diagnosis in theelderly and under-diagnosis in the young age categories.45

The sex differences in incidence rates may also be influ-enced by this reasoning since the cross over point for a plotof the fixed ratio FEV1/FVC and the true LLN against age isat a later age in women.46 We are aware of this limitation,but as information on height and race is not systematicallyreported in the database, we were not able to calculate theLLN and had to use the fixed ratio of FEV1/FVC instead.47

Figure 5 Survival among (top) men and (bottom) women diagn(controls are COPD free).

Finally, in all observational research, misclassification ofdeath and cause of death could be a concern although veryunlikely in the context of the IPCI database. First, weretrieved all patients identified as being deceased from thedemography file in the IPCI database. In addition, we con-ducted a free-text search on “death” and reviewed thecomplete medical record (including discharge letters). Thecause of death, as adjudicated by the IPCI researchers, wasfurther verified with the GP in an initial sample of 10% of allrespiratory or cardiovascular related deaths. The positive

osed with COPD by severity status and reference population

Table 4 Overview of studies that investigated the incidence of COPD.

Author Source population Country COPDdefinitiona

Year ofstudy

Cohortsize

Follow-uptime

Number ofincidentCOPDpatientsb

Agerange inyears

IR COPDb

Spirometry based studiesvan Durme et al.11

Chest 2009Population based the Netherlands

(Rotterdam)Spirometry/dischargeletters

1990e2004 7983 11 yrs 648 �55 9.2/1000PY

Krzyzanowski M et al.12

Am R Res Dis 1986Longitudinal data,random sample; olderpersons and those withchest complaintsperformed lung functiontests

Poland (Cracow) Spirometry 1968e1981 4612 13 yrs 1864 19e70 5.0/1000PY

Huhti E et al.13

Eur J Res Dis 1980Non-selected populationand standardquestionnaire onrespiratory symptomscompleted for eachsubject

Finland(Harjavalta)

Spirometry 1961e1971 1476 10 yrs 1163 40e64 2.0/1000PY and10.0/1000PY forsmokers

de Marco R et al.17

AJRCCM 2007ECRHS on random sampleof young adults.Participants were invitedby questionnaire. Of theresponders, a randomsample of 20% wereinvited for spirometry

Belgium,Denmark, Estonia,France, Germany,Iceland, Italy, theNetherlands,Norway, Spain,Sweden,Switzerland, UK

Spirometry(FEV1/FVC < 70%)

ECRHS I:1991e1993ECRHS II:1999e2002

5002 8.9 yrs 123 20e44 2.8/1000PY10 yrs CI 2.8%

Lindberg A et al.16

Chest 2006Participants selected viapostal questionnaire anda random sample wasinvited to a structuredinterview and spirometryin 1996 and 2003

Sweden(Noorbotten)

Spirometry(GOLD)

85% (1009)participated in1996 and 2003

963 7 yrs 45 (Gold IIand more)91 (Gold Iand more)

46e77 6.7/1000PY Gold II(CI 4.9%)13.5/1000PY GoldI and more (CI11.0%)

Vestbo J et al.14 General populationsample without COPD atstudy entry, selected atrandom after agestratification.

Denmark(Copenhagen)

Spirometry 1976e17781981e19831992e1994

14,223(maximal)

9 yrs Notmentioned

�20years

19/1000PY (5 yrs)and 9/1000PY(15 yrs)5 yrs CI 9.7% and15 yrs CI 13.2%

(continued on next page)

COPDin

thege

neralpopulatio

n1881

Table 4 (continued)

Author Source population CountryCOPDdefinitiona

Year ofstudy

Cohortsize

Follow-uptime

Number ofincidentCOPDpatientsb

Agerange inyears IR COPDb

Johannessen Aet al.19

Participants selectedvia postal questionnaire(1985), a questionnaire(1987e8) and spirometry(1996e7)

Norway(Bergen)

Spirometry(GOLDcriteria)

1985, 1987e8(84% e1275p.), 1996e7(1123 p.)

908 12 yrs 40 18e74 7/1000PY (annualincidence)6.1% (9-year CI)

Diagnosis based studiesGarcia Rodriguezet al.18

Population based-GPRD UK Based on COPDspecific OXMISand Read Codes

1996 808,513 Notspecified

1927 40e89 2.6/1000PY

Huerta C et al.21 Population based-GPRD UK Presence ofrecordeddiagnosisof COPD

1993e1998 5000 6 yrs 206 60e85 7.2/1000PY

Gershon et al.20 Population based-Claimsdatabase

Canada ICD codes 1996e2007 5e6.4million

12 yrs 61,998e55,903

�35 11.8e8.5/1000PY

Abbreviations: CIZ Cumulative incidence, ECRHS Z European community respiratory health survey OXMIS Z Oxford medical information system, READ codes Z current UK coding systemfor diseases.a BTS criteria: FEV1/FVC ratio, <0.70; and FEV1, <80% predicted; GOLD criteria: FEV1/FVC ratio, <0.70.b GOLD: GOLD stage I to IV with FEV1/FVC ratio of <0.70; GOLD II: GOLD II and higher with FEV1/FVC ratio if <0.70 and FEV1<80%predicted.

1882A.S.M

.Afonso

etal.

Supplementary material

Supplementary data associated with this article can befound, in the online version, at doi:10.1016/j.rmed.2011.06.012.

COPD in the general population 1883

predictive value was high, as cause of death was confirmedin more than 90%.

In conclusion, in the Netherlands around 3 on 1000subjects are newly diagnosed with COPD per year; however,the true incidence of COPD, based on spirometry criteria,may be 30e40% higher. The lifetime risk that COPD wouldbe diagnosed over the coming 40 years, if diagnosticcriteria and processes remain the same, is 12.7% for a 40-year-old-man still free of COPD and 8.3% for a woman. Ifascertainment increases this is likely to be higher. Mortalityrates are substantially higher in patients with COPD than inthe general population, which emphasizes the need forbetter primary and secondary prevention in patients withCOPD.

Financial support

None.

Conflict of interest

AA received an unconditional research grant from Pfizerand Boehringer-Ingelheim. This grant was not related to thesubject of this study. KV has been involved as projectleader in analyses contracted by various pharmaceuticalcompanies and received unconditional research grants fromPfizer, Yamanouchi and Boehringer-Ingelheim; none ofwhich are related to the subject of this study. MS isscientific coordinator of the Integrated Primary CareInformation (IPCI) group which is partially funded throughunconditional research grants from the pharmaceuticalindustry namely: Pfizer, Merck, Astra Zeneca, Eli Lilly, GSKand Altana; she has been consultant to Pfizer, NovartisConsumer Health, Servier, Celgene and Lundbeck. GB has,within the last 5 years, received honoraria for lectures fromAstra Zeneca, Boehringer-Ingelheim, GlaxoSmithKline,MerckSharp&Dohme, Novartis, Pfizer and UCB; he isa member of advisory boards for Astra Zeneca, Glax-oSmithKline, Novartis and UCB.

Acknowledgments

We would like to thank the physicians contributing data tothe IPCI database. The principal investigator had fullaccess to all of the data in the study and takes responsi-bility for their integrity and the accuracy of the dataanalysis.

AA participated in the analysis and interpretation of thedata, statistical analysis, drafting of the manuscript.

KV participated in the concept and design, acquisition ofthe data, the validation of the COPD cases, interpretationof data, critical revision of the manuscript for importantintellectual content and supervision.

MS participated in the conception and design, acquisi-tion of data, critical revision of the manuscript for impor-tant intellectual content and supervision.

GB participated in the conception and design, acquisitionof data, second validation of COPD cases and critical revi-sion of the manuscript for important intellectual content.

References

1. Burney P, Suissa S, Soriano JB, Vollmer WM, Viegi G,Sullivan SD, et al. The pharmacoepidemiology of COPD: recentadvances and methodological discussion. Eur Respir J 2003Sep;43:1se44s.

2. GOLD. Global initiative for chronic obstructive lung disease.Global strategy for the diagnosis, management, and preventionof chronic obstructive pulmonary disease, Available from:http://www.goldcopd.com/ [cited July 27, 2009].

3. World Health Organization (WHO). Global surveillance,prevention and control of chronic respiratory diseases:a comprehensive approach. Geneva, Switzerland: World HealthOrganization; 2007.

4. Chapman KR, Mannino DM, Soriano JB, Vermeire PA, Buist AS,Thun MJ, et al. Epidemiology and costs of chronic obstructivepulmonary disease. Eur Respir J 2006 Jan;27(1):188e207.

5. Mannino DM, Gagnon RC, Petty TL, Lydick E. Obstructive lungdisease and low lung function in adults in the United States:data from the National Health and Nutrition ExaminationSurvey, 1988e1994. Arch Intern Med 2000 Jun 12;160(11):1683e9.

6. Rutten van-Molken MP, Feenstra TL. The burden of asthma andchronic obstructive pulmonary disease: data from TheNetherlands. Pharmaco Econ 2001;19(Suppl. 2):1e6.

7. Anto JM, Vermeire P, Vestbo J, Sunyer J. Epidemiology ofchronic obstructive pulmonary disease. Eur Respir J 2001 May;17(5):982e94.

8. BuistAS,VollmerWM,SullivanSD,WeissKB, LeeTA,MenezesAM,et al. The burden of obstructive lung disease initiative (BOLD):rationale and design. COPD 2005 Jun;2(2):277e83.

9. Groenewald P, Vos T, Norman R, Laubscher R, van Walbeek C,Saloojee Y, et al. Estimating the burden of disease attributableto smoking in South Africa in 2000. S Afr Med J (Suid-Afrikaansetydskrif vir geneeskunde) 2007 Aug;97(8 Pt 2):674e81.

10. Zhong N, Wang C, Yao W, Chen P, Kang J, Huang S, et al.Prevalence of chronic obstructive pulmonary disease in China:a large, population-based survey. Am J Respir Crit Care Med2007 Oct 15;176(8):753e60.

11. van Durme YM, Verhamme KM, Stijnen T, van Rooij FJ, VanPottelberge GR, Hofman A, et al. Prevalence, incidence, andlifetime risk for the development of COPD in the elderly: theRotterdam study. Chest 2009 Feb;135(2):368e77.

12. Krzyzanowski M, Jedrychowski W, Wysocki M. Factors associatedwith the change in ventilatory function and the development ofchronic obstructive pulmonary disease in a 13-year follow-up ofthe Cracow study. Risk of chronic obstructive pulmonary disease.Am Rev Respir Dis 1986 Nov;134(5):1011e9.

13. Huhti E, Ikkala J, Hakulinen T. Chronic respiratory disease,smoking and prognosis for life. An epidemiological study. ScandJ Respir Dis 1977 Jun;58(3):170e80.

14. Vestbo J, Lange P. Can GOLD Stage 0 provide information ofprognostic value in chronic obstructive pulmonary disease? AmJ Respir Crit Care Med 2002 Aug 1;166(3):329e32.

15. Lindberg A, Jonsson AC, Ronmark E, Lundgren R, Larsson LG,Lundback B. Ten-year cumulative incidence of COPD and riskfactors for incident disease in a symptomatic cohort. Chest2005 May;127(5):1544e52.

16. Lindberg A, Eriksson B, Larsson LG, Ronmark E, Sandstrom T,Lundback B. Seven-year cumulative incidence of COPD in an

1884 A.S.M. Afonso et al.

age-stratified general population sample. Chest 2006 Apr;129(4):879e85.

17. de Marco R, Accordini S, Cerveri I, Corsico A, Anto JM, Kunzli N,et al. Incidence of chronic obstructive pulmonary disease ina cohort of young adults according to the presence of chroniccough and phlegm. Am J Respir Crit Care Med 2007 Jan 1;175(1):32e9.

18. Garcia Rodriguez LA, Wallander MA, Tolosa LB, Johansson S.Chronic obstructive pulmonary disease in UK primary care:incidence and risk factors. COPD 2009 Oct;6(5):369e79.

19. Johannessen A, Omenaas E, Bakke P, Gulsvik A. Incidence ofGOLD-defined chronic obstructive pulmonary disease ina general adult population. Int J Tuberc Lung Dis 2005 Aug;9(8):926e32.

20. Gershon AS, Wang C, Wilton AS, Raut R, To T. Trends in chronicobstructive pulmonary disease prevalence, incidence, andmortality in ontario, Canada, 1996 to 2007: a population-basedstudy. Arch Intern Med Mar 22 2010;170(6):560e5.

21. Huerta C, Garcia Rodriguez LA, Moller CS, Arellano FM. Therisk of obstructive airways disease in a glaucoma pop-ulation. Pharmacoepidemiol Drug Saf 2001 MareApr;10(2):157e63.

22. Maclay JD, Rabinovich RA, MacNee W. Update in chronicobstructive pulmonary disease 2008. Am J Respir Crit Care Med2009 Apr 1;179(7):533e41.

23. Chapman KR. Chronic obstructive pulmonary disease: arewomen more susceptible than men? Clin Chest Med 2004 Jun;25(2):331e41.

24. World Health Organization (WHO). World Health Report. Lifein the 21st century. A vision for all. Geneva: World HealthOrganization; 1998.

25. Soriano JB, Maier WC, Egger P, Visick G, Thakrar B, Sykes J,et al. Recent trends in physician diagnosed COPD in women andmen in the UK. Thorax 2000 Sep;55(9):789e94.

26. Tan WC, Seale P, Ip M, Shim YS, Chiang CH, Ng TP, et al. Trendsin COPD mortality and hospitalizations in countries and regionsof Asia-Pacific. Respirology 2009 Jan;14(1):90e7.

27. Brown DW, Pleasants RA. Mortality from chronic obstructivepulmonary disease among adults aged 25 years or older inNorth Carolina. South Med J 2011 Jan;104(1):20e3.

28. Tovar Guzman VJ, Lopez Antunano FJ, Rodriguez Salgado N.Recent trends in mortality due to chronic obstructive pulmo-nary disease (COPD) in Mexico, 1980e2002. Arch Med Res 2005JaneFeb;36(1):65e9.

29. van der Lei J, Duisterhout JS, Westerhof HP, van der Does E,Cromme PV, Boon WM, et al. The introduction of computer-based patient records in The Netherlands. Ann Intern Med1993 Nov 15;119(10):1036e41.

30. Schrijvers A. Health and health care in the Netherlands:a critical self assessment of Dutch experts in medical andhealth sciences. Utrecht: De Tijdstroom; 1997.

31. Lamberts H, Wood M, Hofmans-Okkes IM. International primarycare classifications: the effect of fifteen years of evolution.Fam Pract 1992 Sep;9(3):330e9.

32. World Health Organization (WHO). Anatomical therapeuticalchemical index with defined daily dosages. Oslo, Norway:

World Health Organization Collaborating Center for DrugStatistics Methodology; 2002.

33. DeSmet P. The Dutch approach to computerized drug infor-mation: conceptual basis and realization. J Soc AdministrativePharm 1988;5:49e58.

34. Vlug AE, van der Lei J, Mosseveld BM, van Wijk MA, van derLinden PD, Sturkenboom MC, et al. Postmarketing surveillancebased on electronic patient records: the IPCI project. MethodsInf Med 1999 Dec;38(4e5):339e44.

35. Epstein M. Guidelines for good pharmacoepidemiology prac-tices (GPP). Pharmacoepidemiol Drug Saf 2005 Aug;14(8):589e95.

36. Curkendall SM, Lanes S, de Luise C, Stang MR, Jones JK, She D,et al. Chronic obstructive pulmonary disease severity andcardiovascular outcomes. Eur J Epidemiol 2006;21(11):803e13.

37. Eisner MD, Trupin L, Katz PP, Yelin EH, Earnest G, Balmes J,et al. Development and validation of a survey-based COPDseverity score. Chest 2005 Jun;127(6):1890e7.

38. Soriano JB, Maier WC, Visick G, Pride NB. Validation ofgeneral practitioner-diagnosed COPD in the UK general prac-tice research database. Eur J Epidemiol 2001;17(12):1075e80.

39. Rosthoj S, Andersen PK, Abildstrom SZ. SAS macros for esti-mation of the cumulative incidence functions based on a Coxregression model for competing risks survival data. ComputMethods Programs Biomed 2004 Apr;74(1):69e75.

40. Hosmer DLS, et al. Applied survival analysis: regressionmodelling of time to event data. Chichester: John Wiley &Sons; 1999.

41. Frank TL, Hazell ML, Linehan MF, Frank PI. The diagnosticaccuracies of chronic obstructive pulmonary disease (COPD) ingeneral practice: the results of the MAGIC (Manchester AirwaysGroup Identifying COPD) study. Prim Care Respir J 2006 Oct;15(5):286e93.

42. Soriano JB, Vestbo J, Pride NB, Kiri V, Maden C, Maier WC.Survival in COPD patients after regular use of fluticasonepropionate and salmeterol in general practice. Eur Respir J2002 Oct;20(4):819e25.

43. Lundback B, Eriksson B, Lindberg A, Ekerljung L, Muellerova H,Larsson LG, et al. A 20-year follow-up of a population study-based COPD cohort-report from the obstructive lung diseasein Northern Sweden studies. COPD 2009 Aug;6(4):263e71.

44. Tirimanna PR, van Schayck CP, den Otter JJ, van Weel C, vanHerwaarden CL, van den Boom G, et al. Prevalence of asthmaand COPD in general practice in 1992: has it changed since1977? Br J Gen Pract 1996 May;46(406):277e81.

45. Celli BR, Halbert RJ, Isonaka S, Schau B. Population impact ofdifferent definitions of airway obstruction. Eur Respir J 2003Aug;22(2):268e73.

46. Swanney MP, Ruppel G, Enright PL, Pedersen OF, Crapo RO,Miller MR, et al. Using the lower limit of normal for theFEV1/FVC ratio reduces the misclassification of airwayobstruction. Thorax 2008 Dec;63(12):1046e51.

47. Miller MR, Pedersen OF, Dirksen A. A new staging strategy forchronic obstructive pulmonary disease. Int J Chron ObstructPulmon Dis 2007;2(4):657e63.