Advertisement

30-year trends in asthma and the trends in relation to hospitalization and mortality

Open ArchivePublished:July 19, 2018DOI:https://doi.org/10.1016/j.rmed.2018.07.012

      Highlights

      • During the past three decades the asthma prevalence has increased, especially in women.
      • Simultaneously, there was a decreasing trend in hospitalization, especially in asthmatics.
      • However, in asthmatics hospitalization was significantly higher than in non-asthmatics.
      • Additionally, in asthmatics compared with non-asthmatics mortality was higher during the study.

      Abstract

      Introduction

      The present study examines how trends in the prevalence of asthma during the past three decades associate with hospitalization and mortality during the same period.

      Methods

      Altogether 54 320 subjects aged 25–74 years were examined in seven independent cross-sectional population surveys repeated every five years between 1982 and 2012 in Finland. The study protocol included a standardized questionnaire on self-reported asthma, smoking habits and other risk factors, and clinical measurements at the study site. Data on hospitalizations were obtained from the Care Register for Health Care, and data on mortality from the National Causes of Death register.

      Results

      During the study, the prevalence of asthma increased - especially in women. In asthmatic compared with non-asthmatic subjects, hospitalization was significantly higher for all causes, respiratory causes, cardiovascular causes and lung cancer. In addition, particularly in asthmatic subjects, mean yearly hospital days in the 5-year periods after each survey diminished. In asthmatic subjects, the decrease in yearly all-cause hospital days was from 4.45 (between 1982 and 1987) to 1.11 (between 2012 and 2015) and in subjects without asthma the corresponding decrease was from 1.77 to 0.60 (p < 0.001). Similarly between 1982 and 2015, COPD hospitalization decreased more in asthmatic than in non-asthmatic subjects. Generally in the present study, all-cause mortality decreased between 1982 and 2015, though mortality in asthmatic subjects compared with non-asthmatics was higher from all causes, respiratory causes and lung cancer.

      Conclusion

      There was an increasing trend in the prevalence of asthma and a declining trend in hospitalization, especially in asthmatic subjects.

      Keywords

      1. Introduction

      In recent decades the prevalence of asthma has been increasing [
      • Backman H.
      • Räisänen P.
      • Hedman L.
      • et al.
      Increased prevalence of allergic asthma from 1996 to 2006 and further to 2016-results from three population surveys.
      ,
      • de Marco R.
      • Cappa V.
      • Accordini S.
      • et al.
      Trends in the prevalence of asthma and allergic rhinitis in Italy between 1991 and 2010.
      ,
      • Laatikainen T.
      • von Hertzen L.
      • Koskinen J.P.
      • et al.
      Allergy gap between Finnish and Russian Karelia on increase.
      ,
      • Gershon A.S.
      • Guan J.
      • Wang C.
      • To T.
      Trends in asthma prevalence and incidence in Ontario, Canada, 1996-2005: a population study.
      ]. In Sweden, the prevalence of asthma in adults increased from 8.4% to 10.9% between 1996 and 2016 [
      • Backman H.
      • Räisänen P.
      • Hedman L.
      • et al.
      Increased prevalence of allergic asthma from 1996 to 2006 and further to 2016-results from three population surveys.
      ]. In Canada, the age and sex-standardized asthma prevalence increased from 8.5% to 13.3% between 1996 and 2005 [
      • Gershon A.S.
      • Guan J.
      • Wang C.
      • To T.
      Trends in asthma prevalence and incidence in Ontario, Canada, 1996-2005: a population study.
      ]. In Finnish Karelia between 1997 and 2007, the prevalence of self-reported asthma increased from 5.5% to 8.1% in adults [
      • Laatikainen T.
      • von Hertzen L.
      • Koskinen J.P.
      • et al.
      Allergy gap between Finnish and Russian Karelia on increase.
      ].
      In Finland between 1972 and 1986, also asthma-related hospitalizations increased [
      • Keistinen T.
      • Tuuponen T.
      • Kivelä S.L.
      Asthma related hospital treatment in Finland: 1972-86.
      ]. However later, the number of hospital days due to asthma fell 54% between 1993 and 2003 [
      • Haahtela T.
      • Tuomisto L.E.
      • Pietinalho A.
      • et al.
      A 10 year asthma programme in Finland: major change for the better.
      ]. Between 2000 and 2010, the reduction in asthma related hospital days was 65% in relation to the number of asthmatics in Finland [
      • Kauppi P.
      • Linna M.
      • Martikainen J.
      • Mäkelä M.J.
      • Haahtela T.
      Follow-up of the Finnish Asthma Programme 2000-2010: reduction of hospital burden needs risk group rethinking.
      ]. Asthma hospitalizations have declined in other countries too since the 1980s and 1990s [
      • Moorman J.E.
      • Akinbami L.J.
      • Bailey C.M.
      • et al.
      National surveillance of asthma: United States, 2001-2010.
      ,
      • Wilson D.H.
      • Tucker G.
      • Frith P.
      • Appleton S.
      • Ruffin R.E.
      • Adams R.J.
      Trends in hospital admissions and mortality from asthma and chronic obstructive pulmonary disease in Australia, 1993-2003.
      ]. Data on the effect of asthma on all-cause hospitalization is lacking. In addition, asthmatic subjects have had increased all-cause mortality [
      • To T.
      • Simatovic J.
      • Zhu J.
      • et al.
      Asthma deaths in a large provincial system. A 10-year population-based study.
      ,
      • Ali Z.
      • Dirks C.G.
      • Suppli C.
      Long-term mortality among adults with asthma. A 25-year follow-up of 1075 outpatients with asthma.
      ] and smokers and ex-smokers with asthma have had higher mortality than never-smokers without asthma [
      • Colak Y.
      • Afzal S.
      • Nordestgaard B.G.
      • Lange P.
      Characteristics and prognosis of never-smokers and smokers with asthma in the copenhagen general population study.
      ]. The present study examines the 30-year trends in the prevalence of self-reported asthma. Concurrent hospitalization and mortality were also assessed.

      2. Methods

      2.1 Subjects

      The National Institute for Health and Welfare has conducted cross-sectional risk factor surveys every fifth year in Finland since 1972 [
      • Vartiainen E.
      • Laatikainen T.
      • Peltonen M.
      • et al.
      Thirty-five-year trends in cardiovascular risk factors in Finland.
      ,
      • Pelkonen M.K.
      • Notkola I.L.
      • Laatikainen T.K.
      • Koskela H.O.
      Twenty-five year trends in prevalence of chronic bronchitis and the trends in relation to smoking.
      ]. At each survey, an independent random population sample was drawn from the population register in five geographical areas [
      • Vartiainen E.
      • Laatikainen T.
      • Peltonen M.
      • et al.
      Thirty-five-year trends in cardiovascular risk factors in Finland.
      ,
      • Pelkonen M.K.
      • Notkola I.L.
      • Laatikainen T.K.
      • Koskela H.O.
      Twenty-five year trends in prevalence of chronic bronchitis and the trends in relation to smoking.
      ]:the provinces of North Karelia and Kuopio in eastern Finland since 1972, the Turku-Loimaa region in south-western Finland since 1982, the capital area since 1992 and the Oulu province in north-western Finland since 1997 and Lapland since 2002. The sampling and methods complied with the protocol of the World Health Organisation Multinational MONItoring of trends and determinants in CArdiovascular disease (MONICA) project [
      • WHO MONICA
      Project Principal Investigators. The World Health Organization MONICA Project (monitoring trends and determinants of cardiovascular disease): a major international collaboration.
      ], and since 2002, the later recommendations of the European Health Risk Monitoring project (EHRM) [
      • Tolonen H.
      • Kuulasmaa K.
      • Laatikainen T.
      Wolf H and the European Health Risk Monitoring Project. Recommendation for indicators, international collaboration, protocol and manual of operations for chronic disease risk factor surveys.
      ].
      The present study population consisted of altogether 54 320 subjects who participated in the cross-sectional surveys between 1982 and 2012 (Table 1). Subjects who had no data on the presence of asthma were excluded (n = 1000), as well as those without data on smoking habits and smokers who had quit smoking less than one month before the examination and because of a short abstinence could not be recorded as ex-smokers (Table 1).
      Table 1Description of the study population by the survey year.
      nThe survey year
      1982198719921997200220072012Total
      Invited11 3957932792711 50013 49812 00010 00074 252
      Examined (men/women)4615/47323109/33702849/32024253/41934482/50983740/42533041/338326 089/28 231
      Non-responders (men/women)1212/836853/6001116/7601747/13072267/16512260/17471959/161711 414/8518
      Participation rate (%)8282767371676473
      Subjects with hospitalizations between survey and the end of 2015862958324978643857283109161336 327
      Subjects with hospitalizations during 5 years after each survey3950283024793522338922411613
      Data of hospitalization between 2012 and the end of 2015.
      20 024
      Number of deaths between 1982 and the end of 20153937191610461552728296839558
      Self-reported asthma
      Diagnosed or treated within the past year.
      1571571793974754314802276
      Smoking status
       never smokers396228042596375242943634303424 076
       ex-smokers228916751622231724902425196214 780
       current smokers279118051811211727251855134514 449
       excluded
      Without data on smoking, and ex-smokers who had quit smoking less than one month ago (and due to a short abstinence not recorded as ex-smokers).
      305195222607179831015
      a Diagnosed or treated within the past year.
      b Data of hospitalization between 2012 and the end of 2015.
      c Without data on smoking, and ex-smokers who had quit smoking less than one month ago (and due to a short abstinence not recorded as ex-smokers).

      2.2 Measurement of self-reported asthma, smoking status and other variables

      At each survey, the participants completed a self-administered questionnaire [
      • Pelkonen M.K.
      • Notkola I.L.
      • Laatikainen T.K.
      • Koskela H.O.
      Twenty-five year trends in prevalence of chronic bronchitis and the trends in relation to smoking.
      ]. Definition of self-reported (physician-diagnosed) asthma was based on a positive response to the same question ‘Has a doctor diagnosed you with asthma or treated you for asthma during the past year (last 12 months)? (The National FINRISK Study questionnaire has been presented earlier [
      • Pelkonen M.K.
      • Notkola I.L.
      • Laatikainen T.K.
      • Koskela H.O.
      Twenty-five year trends in prevalence of chronic bronchitis and the trends in relation to smoking.
      ]). Smoking status, marital status, education, occupation, a history of diagnosed hypertension during the preceding 12 months and a history of diagnosed myocardial infarction and the use of asthma medication were asked with standardized questions in the same questionnaire. Body mass index (BMI) (kg/m2) was calculated from height and weight measured at the examination.
      Smoking status was classified into the three categories: never-, ex- and current smokers (Table 1). Smokers had smoked regularly (cigarettes, cigars, or pipe) at least a year and during the preceding month. Ex-smokers had stopped smoking at least a month before the survey. The educational background was classified into four categories: elementary school, vocational school, upper secondary school/college and an academic degree. Occupation was classified into seven categories: agriculture and dairy farming, factory, mine and construction work, office work (or suchlike work, e.g. service activity), students, housewives, pensioners and unemployed. Marital status was classified into four categories: married or cohabitation without marriage, unmarried, divorced and widowed. The area of residence was classified into four categories: North Karelia and Kuopio, south-western Finland (Turku-Loimaa area), Oulu and Lapland, and the capital area.

      2.3 Assessment of hospitalization and mortality

      The annual hospital days during five years after each survey were calculated by summing up the hospital days in five years after each survey and then dividing this sum by five (or by the time until death if the subject died before the end of the five-year follow-up). For those examined in 2012 the sum was divided by 3.83 (the mean duration of the follow-up for this cohort). The hospitalization data was obtained from the Care Register for Health Care [
      • Pelkonen M.K.
      • Notkola I.K.
      • Laatikainen T.K.
      • Jousilahti P.
      Chronic bronchitis in relation to hospitalization and mortality over three decades.
      ].
      Between 1982 and 2015, there were three different revisions of the International Classification of Diseases (ICD) in Finland (i.e. ICD-8 until 1986, ICD-9 between 1987 and 1995 and ICD-10 since 1996). In the present study, the concordance table was used for bridging the three revisions of ICD [
      • Janssen F.
      • Kunst A.E.
      ICD coding changes and discontinuities in trends in cause-specific mortality in six European countries, 1950-99.
      ], and the first-listed (main) discharge diagnoses were classified into the following four major categories [
      • Pelkonen M.K.
      • Notkola I.K.
      • Laatikainen T.K.
      • Jousilahti P.
      Chronic bronchitis in relation to hospitalization and mortality over three decades.
      ]: respiratory causes (the listing of the corresponding ICD-8, ICD-9 and ICD-10 codes have been presented earlier [
      • Pelkonen M.K.
      • Notkola I.K.
      • Laatikainen T.K.
      • Jousilahti P.
      Chronic bronchitis in relation to hospitalization and mortality over three decades.
      ]), cardiovascular disease (=CVD), cancer and other causes. COPD (= chronic obstructive pulmonary disease) and lung cancer were also reanalysed separately.
      The data on the length of the hospital visit was missing in 3890 cases, and the main discharge diagnosis was missing in 35 hospitalizations. Altogether, there were 20 024 subjects with hospitalizations during five years after each survey (Table 1).
      Data on the underlying causes of death were obtained from the National Causes of Death register. There were 9558 deaths between 1982 and the end of 2015. For the present study, the underlying causes of death were classified into the same four (major) categories as the hospital diagnoses. Altogether ten subjects had a missing cause of death.

      2.4 Statistical methods

      The area-adjusted age-specific prevalence rates for asthma were calculated from the results of the logistic regression models, which were fitted for each age group and gender separately using the survey year and the area as independent categorical variables, using the following formula [
      • Pelkonen M.K.
      • Notkola I.L.
      • Laatikainen T.K.
      • Koskela H.O.
      Twenty-five year trends in prevalence of chronic bronchitis and the trends in relation to smoking.
      ]: Adjusted prevalence = EXP(Constant + Bi_Year)/(1 + EXP(Constant + B i_Year)).
      The age-specific multivariable estimated prevalence odds ratios (ORs) for asthma were calculated by logistic regression analyses. The following independent categorical variables were included into the model: smoking status, occupation, the area of residence and education, and the year of examination and body mass index as continuous variables. The interaction between the gender and the yearly trend was evaluated by adding the term gender*the yearly trend into to the model.
      To assess if the distribution of the annual hospital days was different in asthmatic and non-asthmatic subjects, at first a nonparametric test was performed (Mann-Whitney U Test). Then analysis of covariance (ANCOVA) was used for the multivariable analyses (n = 48 694, and n = 44 056 in subjects aged 25–64 years). Though our data did not follow a normal distribution, because of the large study population the ANCOVA analysis gives a robust estimate on the significance of the differences [
      • Pelkonen M.K.
      • Notkola I.K.
      • Laatikainen T.K.
      • Jousilahti P.
      Chronic bronchitis in relation to hospitalization and mortality over three decades.
      ,
      • Harwell M.R.
      • Rubinstein E.N.
      • Hayes W.S.
      • Olds C.C.
      Summarizing Monte Carlo results in methodological research: the one- and two-factor fixed effects ANOVA cases.
      ]. The following independent categorical variables were included into the ANCOVA analyses: gender, the 10-year age-group, smoking status, education and marital status and the area of residence, the presence of hypertension and a history of diagnosed myocardial infarction as binary variables, and the survey year and body mass index as continuous variables. When studying the interaction effect between asthma and the study year (or smoking status), the interaction term asthma*the study year (asthma*smoking status) was added into the ANCOVA analysis. The χ [
      • de Marco R.
      • Cappa V.
      • Accordini S.
      • et al.
      Trends in the prevalence of asthma and allergic rhinitis in Italy between 1991 and 2010.
      ] test was used to compare the use of asthma medication (as a binary variable: yes/never or over a year ago) with the presence of asthma. Data of the use of asthma medication was available for 97.5% and 98.6% of subjects without and with asthma.
      Cox proportional hazards regression models were used to analyse the relation between asthma and mortality. Among study subjects with variables for multivariable analyses 1581 subjects had a duplicate case between 1982 and 2012. Among these subjects, the values of the first survey were used in the mortality analyses (n = 47 103). The mortality analyses were adjusted for the same co-variables as the ANCOVA analyses. When studying the interaction between asthma and the study year, the interaction term asthma*the study year was added into the Cox analysis.
      Statistical analyses were performed by SPSS 22 for Windows. The study was approved in 2001 b y the ethics committee of the National Institute for Health and Welfare in Helsinki.

      3. Results

      During the study asthma was more prevalent in women than in men (p < 0.001) (Fig. 1). In women aged 25–64 years the prevalence of asthma (past year) increased from 1.7% (95% CI 1.3–2.3%) in 1982 to 8.0% (95% CI 6.2–10.2%) in 2017 (Fig. 1). In men aged 25–64 years, the corresponding increase was from 1.6% (95% CI 1.2–2.1%) in 1982 to 5.5% (95% CI 4.1–7.3%) in 2017. In women aged 65–74 years, the prevalence of asthma increased from 6.1% (95% CI 3.9–9.5%) to 10.7% (95% CI 7.0–16.0%), (Fig. 1). The multivariable estimated prevalence ORs for asthma (Table 2) increased during 1982–2012 significantly in all age-groups, except in men aged 65–74 years. There was an interaction (p = 0.020) between the gender and the yearly trend, so that in women asthma prevalence increased more than in men (Table 2).
      Fig. 1
      Fig. 1The adjusted prevalence of self-reported asthma (diagnosed or treated within the past year) by 10-year age group and the survey year. The 95% confidence intervals are omitted because of clarity. Adjusted for the area of residence. p for the trend in parenthesis.
      Table 2Multivariable adjusted odds ratios (OR)
      OR is presenting the yearly change in the prevalence of asthma during 1982–2012.
      for the trend in self-reported asthma during 1982–2012 b y ten-year age groups.
      nOR (95% CI) during 1982–2012p-value
      Men
       Age group
      From a logistic regression model, adjusted for survey year, smoking status, occupation, education, body mass index and area of residence.
       25–3445651.054 (1.026–1.084)< 0.001
       35–4451951.061 (1.038–1.084)< 0.001
       45–5454011.024 (1.005–1.043)0.012
       55–6455901.017 (1.001–1.033)0.032
       65–7425191.007 (0.978–1.036)0.638
       All
      From a logistic regression model, adjusted for age-group, survey year, smoking status, occupation, education, body mass index and area of residence.
      23 2701.030 (1.020–1.039)< 0.001
      Women
       Age group
      From a logistic regression model, adjusted for survey year, smoking status, occupation, education, body mass index and area of residence.
       25-3453611.074 (1.054–1.095)< 0.001
       35-4458041.047 (1.028–1.067)< 0.001
       45-5459971.037 (1.020–1.054)< 0.001
       55-6460471.045 (1.031–1.059)< 0.001
       65-7421581.039 (1.008–1.071)0.012
       All
      From a logistic regression model, adjusted for survey year, smoking status, occupation, education, body mass index and area of residence.
      From a logistic regression model, adjusted for age-group, survey year, smoking status, occupation, education, body mass index and area of residence.
      25 3671.048 (1.040–1.056)< 0.001
      All (men + women)
      From a logistic regression, adjusted for gender, age-group, survey year, smoking status, occupation, education, body mass index and area of residence.
      48 6371.040 (1.034–1.046)< 0.001
      a OR is presenting the yearly change in the prevalence of asthma during 1982–2012.
      b From a logistic regression model, adjusted for survey year, smoking status, occupation, education, body mass index and area of residence.
      c From a logistic regression model, adjusted for age-group, survey year, smoking status, occupation, education, body mass index and area of residence.
      d From a logistic regression, adjusted for gender, age-group, survey year, smoking status, occupation, education, body mass index and area of residence.
      Subjects with asthma had higher all-cause hospitalization compared with subjects without asthma (Table 3). In multivariable analysis (p-value2), there was an interaction (p = 0.005) between asthma and smoking status on all-cause hospitalization so that in asthmatic never smokers all cause hospitalization was not significantly increased (Table 3). The cause-specific hospitalization (multivariable) was increased for respiratory causes and separately COPD, and for cardiovascular causes and lung cancer. When multivariable analyses were performed separately by gender (not shown in tables), in female asthmatics hospitalization was increased from all causes (p = 0.016), from respiratory causes (p < 0.001) and separately from COPD (p < 0.001), and in male asthmatics hospitalization was increased from all causes (p < 0.001), and cause-specially from respiratory causes (p < 0.001) and separately from COPD (p < 0.001), from CVD (p = 0.025) and from lung cancer (p < 0.001).
      Table 3Means (SD = standard deviation) for the yearly all-cause and cause-specific hospital days in five years after each survey
      Survey in 1982/1987/1992/1997/2002/2007/2012, 5-year hospitalization thereafter (except 4-year hospitalization with 2012 survey).
      by the presence of self-reported asthma and smoking.
      Cause of hospitalizationNo AsthmaAsthmap-value1
      P-value1 from the Mann-Whitney U test for each cause separately. When stratified by smoking status for each smoking category separately.
      p-value2
      P-value2 from ANCOVA for each cause separately. When stratified by smoking status for each smoking category separately. ANCOVA analysis adjusted for gender, age group, smoking status, body mass index, survey year, education, marital status, history of myocardial infarction, presence of hypertension and area of residence.
      n in analysisn with hospitalizationsmean hospital days/year (SD)
      Mean (SD) calculated for all non-asthmatics and asthmatics in analyses.
      n in analysisn with hospitalizationsmean hospital days/year (SD)
      Mean (SD) calculated for all non-asthmatics and asthmatics in analyses.
      All cause hospitalization
       Smoking group
       Current smokers12 65747931.59 (7.85)4292343.24 (10.55)< 0.0010.003
       Ex-smokers12 96549351.33 (6.06)6293382.92 (11.86)< 0.001< 0.001
       Never smokers21 06278711.23 (6.52)9524671.75 (7.01)< 0.0010.152
       All46 68417 5991.35 (6.79)201010392.44 (9.56)< 0.001< 0.001
      Cause specific hospitalization46 6842010
      Respiratory
      Including also COPD hospitalization.
      14650.06 (0.97)3130.44 (3.02)< 0.001< 0.001
       COPD
      Including ICD-8: 490–493, 518; ICD-9: 490–494, 496; ICD-10 J40-J47.
      2780.02 (0.54)2020.30 (2.18)< 0.001< 0.001
      Cardiovascular23990.20 (2.81)1890.46 (3.85)< 0.0010.018
      Cancer
      Including also lung cancer hospitalization.
      12030.18 (2.97)750.23 (3.54)0.0010.813
       Lung cancer800.03 (1.21)90.14 (3.42)0.004< 0.001
      Other15 7440.91 (4.76)8541.30 (5.74)< 0.0010.013
      a Survey in 1982/1987/1992/1997/2002/2007/2012, 5-year hospitalization thereafter (except 4-year hospitalization with 2012 survey).
      b Mean (SD) calculated for all non-asthmatics and asthmatics in analyses.
      c P-value1 from the Mann-Whitney U test for each cause separately. When stratified by smoking status for each smoking category separately.
      d P-value2 from ANCOVA for each cause separately. When stratified by smoking status for each smoking category separately. ANCOVA analysis adjusted for gender, age group, smoking status, body mass index, survey year, education, marital status, history of myocardial infarction, presence of hypertension and area of residence.
      e Including also COPD hospitalization.
      f Including ICD-8: 490–493, 518; ICD-9: 490–494, 496; ICD-10 J40-J47.
      g Including also lung cancer hospitalization.
      During the follow-up (Fig. 2), there was a decreasing trend (p < 0.001) in the yearly hospital days. In addition, there was an interaction between the presence of asthma and the decreasing trend on yearly hospital days for all-cause (p = 0.004) and COPD (p < 0.001) hospitalization so that hospitalization decreased more in asthmatics than in non-asthmatics. There was also a similar interaction for respiratory (p < 0.001) and cardiovascular (p = 0.047) hospitalization (not shown). In female asthmatics, COPD (and respiratory) hospitalization decreased during the study more (p < 0.001) than in female non-asthmatics (not shown). In males too, all cause (p = 0.003) and COPD (and respiratory) (p < 0.001) hospitalization decreased more in asthmatics than in non-asthmatics (not shown). From 1997 onwards, altogether 94.2% of asthmatics reported using asthma medication (past year) compared with 3.5% in those without asthma (p < 0.001) (not shown).
      Fig. 2
      Fig. 2Mean yearly hospital days for five-year periods after each survey (for four-year period with 2012 survey) by presence of asthma in subjects aged 24–64 years. Multivariable adjusted p-values from ANCOVA analyses for all cause and COPD hospitalization separately, adjusted for gender, age group, smoking status, body mass index, survey year, education, marital status, history of myocardial infarction, presence of hypertension and area of residence. In analyses n = 42 438 and 1618, respectively, without and with asthma. The number of subjects with all cause hospitalizations was, respectively, 15 551 and 801 without and with asthma. The corresponding number of subjects with COPD hospitalizations was 239 and 166, respectively, without and with asthma.
      Subjects with asthma had higher all-cause mortality than non-asthmatic subjects (Table 4). Cause-specifically mortality was increased from respiratory causes, and separately from COPD, and lung cancer. When mortality analyses were stratified by smoking, in asthmatics COPD mortality was higher in all smoking groups, respiratory mortality in smokers and ex-smokers, and lung cancer mortality in ex-smokers (not shown in tables). When mortality analyses were stratified by gender (not shown in tables), in female asthmatics mortality was increased from all-cause (hazard ratio (HR) = 1.25 (1.07–1.46), p = 0.006), from respiratory causes (HR = 2.59 (1.32–5.05), p = 0.005) and separately from COPD (HR = 6.07 (2.93–12.55), p < 0.001), and in male asthmatics from respiratory causes (HR = 4.08 (2.83–5.86), p < 0.001) and separately form COPD (HR = 7.80 (5.17–11.76), p < 0.001) and from lung cancer (HR = 1.71 (1.11–2.66), p = 0.016). However, generally in study subjects there was a decreasing (p < 0.001) trend in all-cause mortality during 1982 and 2015 (not shown in tables).
      Table 4Number of deaths in subjects without and with self-reported asthma and multivariable adjusted hazard ratios (95% CI)
      From Cox proportional hazards regression model, for each cause separately and when stratified by smoking for each smoking category separately.
      for all cause and cause-specific mortality during 1982–2015.
      Cause of deathNo asthmaAsthmaHazard ratio (95% CI)p-value
      n in analysisn of deathsn in analysisn of deaths
      All cause death
      Smoking group
      Current smokers12 35928344121151.31 (1.08–1.58)0.005
      Ex-smokers12 51022126041311.14 (0.96–1.36)0.146
      Never smokers20 30530929131481.09 (0.92–1.29)0.328
      All45 174813819293941.16 (1.05–1.29)0.004
      Cause specific death45 1741929
      Respiratory
      Including also COPD deaths.
      320463.71 (2.70–5.10)< 0.001
       COPD
      Including ICD-8: 490–493, 518; ICD-9: 490–494, 496; ICD-10 J40-J47.
      161427.73 (5.41–11.04)< 0.001
      Cardiovascular31421491.02 (0.86–1.21)0.816
      Cancer
      Including also lung cancer deaths.
      2136971.17 (0.95–1.43)0.146
       Lung cancer446281.71 (1.16–2.52)0.007
      Other causes25401021.04 (0.85–1.28)0.676
      Adjusted for gender, age group, smoking status, body mass index, survey year, education, marital status, history of myocardial infarction, presence of hypertension and area of residence. Subjects without asthma as a reference group. 95% CI = 95% confidence interval.
      a From Cox proportional hazards regression model, for each cause separately and when stratified by smoking for each smoking category separately.
      b Including also COPD deaths.
      c Including ICD-8: 490–493, 518; ICD-9: 490–494, 496; ICD-10 J40-J47.
      d Including also lung cancer deaths.

      4. Discussion

      In the present study, during the last three decades the prevalence of asthma has increased and yearly hospital days has decreased more in asthmatics than in non-asthmatics.
      Previously, risk factors for asthma have been, e.g. family history of asthma, smoking and female gender [
      • Lundbäck B.
      • Rönmark E.
      • Jönsson E.
      • Larsson K.
      • Sandström T.
      Incidence of physician-diagnosed asthma in adults – a real incidence or result of increased awareness? Report from the obstructive lung disease in northern Sweden studies.
      ,
      • Eder W.
      • Ege M.J.
      • von Mutius E.
      The asthma epidemic.
      ]. The OR for developing asthma has been 2.0–2.6 with continued smoking [
      • Godtfredsen N.S.
      • Lange P.
      • Prescott E.
      • Osler M.
      • Vestbo J.
      Changes in smoking habits and risk of asthma: a longitudinal population based study.
      ]. In the present study during the follow-up, male smoking decreased but smoking in middle-aged women increased [
      • Pelkonen M.K.
      • Notkola I.L.
      • Laatikainen T.K.
      • Koskela H.O.
      Twenty-five year trends in prevalence of chronic bronchitis and the trends in relation to smoking.
      ], probably explaining a part of the increase in asthma prevalence in women. In parallel with the present results, in an earlier paper from the FINRISK Study [
      • Jousilahti P.
      • Haahtela T.
      • Laatikainen T.
      • Mäkelä M.
      • Vartiainen E.
      Asthma and respiratory allergy prevalence is still increasing among Finnish young adults.
      ] the amount of those reporting ever having asthma increased between 1997 and 2012. In an earlier study in Finnish Karelia, sensitization rates for birch, timothy and cat allergens nearly doubled between 1997 and 2007 and the prevalence of asthma substantially increased [
      • Laatikainen T.
      • von Hertzen L.
      • Koskinen J.P.
      • et al.
      Allergy gap between Finnish and Russian Karelia on increase.
      ]. The increase in atopy was suggested to arise due to mass movements from rural to urban areas [
      • Laatikainen T.
      • von Hertzen L.
      • Koskinen J.P.
      • et al.
      Allergy gap between Finnish and Russian Karelia on increase.
      ]. Similarly in an American study, the prevalence of asthma and allergic sensitization was lower in children whose family followed traditional farming practices [
      • Stein M.M.
      • Hrusch C.L.
      • Gozdz J.
      • et al.
      Innate immunity and asthma risk in amish and hutterite farm children.
      ].
      Generally, increasing awareness of asthma may lead to increased diagnostic activity [
      • Lundbäck B.
      • Rönmark E.
      • Jönsson E.
      • Larsson K.
      • Sandström T.
      Incidence of physician-diagnosed asthma in adults – a real incidence or result of increased awareness? Report from the obstructive lung disease in northern Sweden studies.
      ,
      • Eder W.
      • Ege M.J.
      • von Mutius E.
      The asthma epidemic.
      ]. In children, the percentage of subjects with physician-diagnosed asthma has also increased steadily between 1985 and 2005 [
      • Kälvesten L.
      • Bråbäck L.
      Time trends for the prevalence of asthma among school children in a Swedish district in 1985-2005.
      ]. However, in some child populations asthma prevalence has recently reached a plateau [
      • Wennergren G.
      The prevalence of asthma has reached a plateau.
      ].
      In the present study, the mean yearly all-cause hospital days were approximately 1.9 fold higher in subjects with asthma than in those without it. Previously, the all-cause hospitalization risk was about 2.1 in adult asthmatics without chronic bronchitis [
      • Accordini S.
      • Corsico A.G.
      • Calciano L.
      • et al.
      The impact of asthma, chronic bronchitis and allergic rhinitis on all-acuse hospitalizations and limitations in daily activities: a population-based observational study.
      ]. Generally, there has been a decrease in average length of stay at hospital after 1970 [
      • Fuchs V.R.
      Major trends in the U.S. Health economy since 1950.
      ]. In the present study, the mean yearly hospital days after each survey decreased more in subjects with asthma than in those without it. Decreasing COPD hospitalizations in asthmatic subjects may imply that the occurrence of severe asthma has not increased.
      Earlier, smokers have had an increased risk for hospitalization, and the excess risk for COPD hospitalization has decreased within 5 years after smoking cessation [
      • Tran B.
      • Falster M.O.
      • Douglas K.
      • Blyth F.
      • Jorm L.R.
      Smoking and potentially preventable hospitalisation: the benefit of smoking cessation in older ages.
      ]. Also environmental tobacco smoke exposure has been associated with a greater risk for admission to hospital for asthma [
      • Eisner M.D.
      • Klein J.
      • Hammond S.K.
      • Koren G.
      • Lactao G.
      • Iribarren C.
      Directly measured second hand smoke exposure and asthma health outcomes.
      ], and the Smoke-Free Air Law (USA) has been associated with a reduction in adult asthma hospitalization rates [
      • Marchese M.E.
      • Shamo F.
      • Miller C.E.
      • Wahl R.L.
      • Li Y.
      Racial disparities in asthma hospitalizations following implementation of the smoke-free Air Law, Michigan, 2002-2012.
      ]. In Finland, environmental tobacco smoke exposure has decreased markedly from the 1980s’ because of legislative smoking restrictions [
      • Pelkonen M.K.
      • Notkola I.L.
      • Laatikainen T.K.
      • Koskela H.O.
      Twenty-five year trends in prevalence of chronic bronchitis and the trends in relation to smoking.
      ].
      During the 1990s, increased sales of asthma drugs (especially inhaled corticosteroids) have been associated with declining rates of hospital admissions for asthma in several countries [
      • Gupta R.
      • Anderson H.R.
      • Strachan D.P.
      • Maier W.
      • Watson L.
      International trends in admissions and drug sales for asthma.
      ]. Furthermore, use of inhaled corticosteroids reduced all-cause mortality and the risk of rehospitalisation due to asthma [
      • Sin D.D.
      • Tu J.V.
      Inhaled corticosteroid therapy reduces the risk of rehospitalization and all-cause mortality in elderly asthmatics.
      ]. On the other hand, reductions in hospitalizations may partly reflect better management of asthma and better compliance on the part of patients. In Finland, the National Asthma Programme to improve the asthma management was undertaken between 1994 and 2004 and the pharmacy programme was launched in 1997 with pharmacists for example checking the inhalation technique [
      • Haahtela T.
      • Tuomisto L.E.
      • Pietinalho A.
      • et al.
      A 10 year asthma programme in Finland: major change for the better.
      ]. The Asthma Programme resulted in a reduction of yearly costs for asthma by one-third, even though the prevalence of asthma increased during the period [
      • Haahtela T.
      • Tuomisto L.E.
      • Pietinalho A.
      • et al.
      A 10 year asthma programme in Finland: major change for the better.
      ]. On the contrary, in a Canadian study, the total direct cost of asthma has increased since 2002 because of a rise in asthma prevalence and cost of medication, despite the reductions in asthma hospitalization [
      • Bedouch P.
      • Marra C.A.
      • Fitzgerald J.M.
      • Lynd L.D.
      • Sadatsafavi M.
      Trends in asthma-related direct medical costs from 2002 to 2007 in British Columbia, Canada: a population based-cohort study.
      ].
      In the present study, in asthmatics compared with non-asthmatics mortality was higher (especially from COPD). Earlier, in one study asthmatics without airflow limitation did not have higher mortality [
      • Huang S.
      • Vasquez M.M.
      • Halonen M.
      • Martinez F.D.
      • Guerra S.
      Asthma, airflow limitation and mortality risk in the general population.
      ] but in another study such asthmatics had a higher risk for respiratory mortality [
      • Lange P.
      • Colak Y.
      • Ingebrigtsen T.S.
      • Vestbo J.
      • Marott J.L.
      Long-term prognosis of asthma, chronic obstructive pulmonary disease, and asthma-chronic obstructive pulmonary disease overlap in the Copenhagen City Heart study: a prospective population-based analysis.
      ]. Between 1999 and 2008, respiratory mortality was almost three times higher in asthmatics than in the general population [
      • To T.
      • Simatovic J.
      • Zhu J.
      • et al.
      Asthma deaths in a large provincial system. A 10-year population-based study.
      ]. In one study, never-smokers with asthma had an increased risk for COPD exacerbations and smokers with asthma also an increased risk for lung cancer, cardiovascular disease and death [
      • Colak Y.
      • Afzal S.
      • Nordestgaard B.G.
      • Lange P.
      Characteristics and prognosis of never-smokers and smokers with asthma in the copenhagen general population study.
      ].
      In epidemiological studies, questionnaires are often used to define if a subject has symptoms of asthma or has a physician-diagnosed asthma [
      • Lundbäck B.
      • Rönmark E.
      • Jönsson E.
      • Larsson K.
      • Sandström T.
      Incidence of physician-diagnosed asthma in adults – a real incidence or result of increased awareness? Report from the obstructive lung disease in northern Sweden studies.
      ,
      • Eder W.
      • Ege M.J.
      • von Mutius E.
      The asthma epidemic.
      ]. A self-report of physician-diagnosed asthma has been shown to have sufficient sensitivity and high specificity [
      • Torėn K.
      • Brisman J.
      • Järvholm B.
      Asthma and asthma-like symptoms in adults assessed by questionnaires. A literature review.
      ]. Unfortunately in the present study pulmonary function was not measured. In the present study, 94% of asthmatics reported using asthma medication during past year form 1997 onwards. In a ten-year population study, 90% of those who had been diagnosed asthma during the follow-up reported using asthma medicines [
      • Lundbäck B.
      • Rönmark E.
      • Jönsson E.
      • Larsson K.
      • Sandström T.
      Incidence of physician-diagnosed asthma in adults – a real incidence or result of increased awareness? Report from the obstructive lung disease in northern Sweden studies.
      ].
      The strength of the present study was a large study sample with a long follow-up and comprehensive hospitalization and mortality data. The participation rates decreased in the present study from older to younger cohorts and possibly some of those with higher morbidity are among non-participants. However, in an earlier questionnaire study, non-responders, evaluated by a telephone interview, have not reported more respiratory symptoms than responders [
      • Rönmark E.P.
      • Ekerljung L.
      • Lötvall J.
      • Torén K.
      • Rönmark E.
      • Lundbäck B.
      Large scale questionnaire survey on respiratory health in Sweden: effects of late- and non-response.
      ]. In the present study, three different revisions of the International Classification of Diseases were used. Thus, there may have been changes in reporting of the causes of hospitalizations and deaths [
      • Pelkonen M.K.
      • Notkola I.K.
      • Laatikainen T.K.
      • Jousilahti P.
      Chronic bronchitis in relation to hospitalization and mortality over three decades.
      ]. However, these changes in reporting were independent of the presence of asthma at each survey. Though the recording of the hospitalization diagnoses of the inpatients visits to the HILMO register has been similar in the course of the present study, it is possible that the access to the health care has become a bit easier since 1982. However, this would only attenuate the observed decrease in hospitalization. In addition, Finland has had a primary health care system based on health centres since 1972 and a comprehensive hospital network was built already between 1950s and 1970s [
      • Saarivirta T.
      • Consoli D.
      • Dhondt P.
      The evolution of the Finnish health-care system early 19th century and onwards.
      ].
      In conclusion in spite of the increasing prevalence in asthma, hospitalization in asthmatic subjects has decreased. Though all-cause mortality has decreased, mortality in asthmatic subjects compared with non-asthmatics is still higher.

      Funding

      The VTR Funding of the Kuopio University Hospital , the Finnish Medical Foundation .

      Conflicts of interest

      The material has not been published and is not under consideration for publication elsewhere. No author has any conflict of interest. We all have read the paper and approved the submission as well as approved our names in this paper and thus sign consent to publication.

      References

        • Backman H.
        • Räisänen P.
        • Hedman L.
        • et al.
        Increased prevalence of allergic asthma from 1996 to 2006 and further to 2016-results from three population surveys.
        Clin. Exp. Allergy. 2017; 13 (doi:10.1)
        • de Marco R.
        • Cappa V.
        • Accordini S.
        • et al.
        Trends in the prevalence of asthma and allergic rhinitis in Italy between 1991 and 2010.
        Eur. Respir. J. 2012; 39: 883-892
        • Laatikainen T.
        • von Hertzen L.
        • Koskinen J.P.
        • et al.
        Allergy gap between Finnish and Russian Karelia on increase.
        Allergy. 2011; 66: 886-892
        • Gershon A.S.
        • Guan J.
        • Wang C.
        • To T.
        Trends in asthma prevalence and incidence in Ontario, Canada, 1996-2005: a population study.
        Am. J. Epidemiol. 2010; 172: 728-736
        • Keistinen T.
        • Tuuponen T.
        • Kivelä S.L.
        Asthma related hospital treatment in Finland: 1972-86.
        Thorax. 1993; 48: 44-47
        • Haahtela T.
        • Tuomisto L.E.
        • Pietinalho A.
        • et al.
        A 10 year asthma programme in Finland: major change for the better.
        Thorax. 2006; 61: 663-670
        • Kauppi P.
        • Linna M.
        • Martikainen J.
        • Mäkelä M.J.
        • Haahtela T.
        Follow-up of the Finnish Asthma Programme 2000-2010: reduction of hospital burden needs risk group rethinking.
        Thorax. 2013; 68: 292-293
        • Moorman J.E.
        • Akinbami L.J.
        • Bailey C.M.
        • et al.
        National surveillance of asthma: United States, 2001-2010.
        Vital Health Stat. 2012; 35: 1-58
        • Wilson D.H.
        • Tucker G.
        • Frith P.
        • Appleton S.
        • Ruffin R.E.
        • Adams R.J.
        Trends in hospital admissions and mortality from asthma and chronic obstructive pulmonary disease in Australia, 1993-2003.
        Med. J. Aust. 2007; 186: 408-411
        • To T.
        • Simatovic J.
        • Zhu J.
        • et al.
        Asthma deaths in a large provincial system. A 10-year population-based study.
        AnnalsATS. 2014; 11: 1210-1217
        • Ali Z.
        • Dirks C.G.
        • Suppli C.
        Long-term mortality among adults with asthma. A 25-year follow-up of 1075 outpatients with asthma.
        Chest. 2013; 143: 1649-1655
        • Colak Y.
        • Afzal S.
        • Nordestgaard B.G.
        • Lange P.
        Characteristics and prognosis of never-smokers and smokers with asthma in the copenhagen general population study.
        Am. J. Respir. Crit. Care Med. 2015; 192: 172-181
        • Vartiainen E.
        • Laatikainen T.
        • Peltonen M.
        • et al.
        Thirty-five-year trends in cardiovascular risk factors in Finland.
        Int. J. Epidemiol. 2010; 39: 504-518
        • Pelkonen M.K.
        • Notkola I.L.
        • Laatikainen T.K.
        • Koskela H.O.
        Twenty-five year trends in prevalence of chronic bronchitis and the trends in relation to smoking.
        Respir. Med. 2014; 108: 1633-1640
        • WHO MONICA
        Project Principal Investigators. The World Health Organization MONICA Project (monitoring trends and determinants of cardiovascular disease): a major international collaboration.
        J. Clin. Epidemiol. 1988; 41: 105-114
        • Tolonen H.
        • Kuulasmaa K.
        • Laatikainen T.
        Wolf H and the European Health Risk Monitoring Project. Recommendation for indicators, international collaboration, protocol and manual of operations for chronic disease risk factor surveys.
        (www-publications from the EHRM Project. URL)
        http://www.ktl.fi/ehrm
        Date: October 2002
        • Janssen F.
        • Kunst A.E.
        ICD coding changes and discontinuities in trends in cause-specific mortality in six European countries, 1950-99.
        Bull. World Health Organ. 2004; 82: 904-913
        • Pelkonen M.K.
        • Notkola I.K.
        • Laatikainen T.K.
        • Jousilahti P.
        Chronic bronchitis in relation to hospitalization and mortality over three decades.
        Respir. Med. 2017; 123: 87-93
        • Harwell M.R.
        • Rubinstein E.N.
        • Hayes W.S.
        • Olds C.C.
        Summarizing Monte Carlo results in methodological research: the one- and two-factor fixed effects ANOVA cases.
        J. Educ. Behav. Stat. 1992; 7: 315-339
        • Lundbäck B.
        • Rönmark E.
        • Jönsson E.
        • Larsson K.
        • Sandström T.
        Incidence of physician-diagnosed asthma in adults – a real incidence or result of increased awareness? Report from the obstructive lung disease in northern Sweden studies.
        Respir. Med. 2001; 95: 685-690
        • Eder W.
        • Ege M.J.
        • von Mutius E.
        The asthma epidemic.
        N. Engl. J. Med. 2006; 355: 2226-2235
        • Godtfredsen N.S.
        • Lange P.
        • Prescott E.
        • Osler M.
        • Vestbo J.
        Changes in smoking habits and risk of asthma: a longitudinal population based study.
        Eur. Respir. J. 2001; 18: 549-554
        • Jousilahti P.
        • Haahtela T.
        • Laatikainen T.
        • Mäkelä M.
        • Vartiainen E.
        Asthma and respiratory allergy prevalence is still increasing among Finnish young adults.
        Eur. Respir. J. 2016; 47: 985-987
        • Stein M.M.
        • Hrusch C.L.
        • Gozdz J.
        • et al.
        Innate immunity and asthma risk in amish and hutterite farm children.
        N. Engl. J. Med. 2016; 375: 411-421
        • Kälvesten L.
        • Bråbäck L.
        Time trends for the prevalence of asthma among school children in a Swedish district in 1985-2005.
        Acta Paediatr. 2008; 97: 454-458
        • Wennergren G.
        The prevalence of asthma has reached a plateau.
        Acta Paediatr. 2011; 100: 938-939
        • Accordini S.
        • Corsico A.G.
        • Calciano L.
        • et al.
        The impact of asthma, chronic bronchitis and allergic rhinitis on all-acuse hospitalizations and limitations in daily activities: a population-based observational study.
        BMC Pulm. Med. 2015; 15: 10
        • Fuchs V.R.
        Major trends in the U.S. Health economy since 1950.
        N. Engl. J. Med. 2012; 366: 973-977
        • Tran B.
        • Falster M.O.
        • Douglas K.
        • Blyth F.
        • Jorm L.R.
        Smoking and potentially preventable hospitalisation: the benefit of smoking cessation in older ages.
        Drug Alcohol Depend. 2015; 150: 973-977
        • Eisner M.D.
        • Klein J.
        • Hammond S.K.
        • Koren G.
        • Lactao G.
        • Iribarren C.
        Directly measured second hand smoke exposure and asthma health outcomes.
        Thorax. 2005; 60: 814-821
        • Marchese M.E.
        • Shamo F.
        • Miller C.E.
        • Wahl R.L.
        • Li Y.
        Racial disparities in asthma hospitalizations following implementation of the smoke-free Air Law, Michigan, 2002-2012.
        Prev. Chronic Dis. 2015; 12 (E201)
        • Gupta R.
        • Anderson H.R.
        • Strachan D.P.
        • Maier W.
        • Watson L.
        International trends in admissions and drug sales for asthma.
        Int. J. Tubercul. Lung Dis. 2006; 10: 138-145
        • Sin D.D.
        • Tu J.V.
        Inhaled corticosteroid therapy reduces the risk of rehospitalization and all-cause mortality in elderly asthmatics.
        Eur. Rep. J. 2001; 17: 380-385
        • Bedouch P.
        • Marra C.A.
        • Fitzgerald J.M.
        • Lynd L.D.
        • Sadatsafavi M.
        Trends in asthma-related direct medical costs from 2002 to 2007 in British Columbia, Canada: a population based-cohort study.
        PLoS One. 2013; 8 (doi:10.1371)
        • Huang S.
        • Vasquez M.M.
        • Halonen M.
        • Martinez F.D.
        • Guerra S.
        Asthma, airflow limitation and mortality risk in the general population.
        Eur. Respir. J. 2015; 45: 338-346
        • Lange P.
        • Colak Y.
        • Ingebrigtsen T.S.
        • Vestbo J.
        • Marott J.L.
        Long-term prognosis of asthma, chronic obstructive pulmonary disease, and asthma-chronic obstructive pulmonary disease overlap in the Copenhagen City Heart study: a prospective population-based analysis.
        Lancet Respir Med. 2016; 4: 454-462
        • Torėn K.
        • Brisman J.
        • Järvholm B.
        Asthma and asthma-like symptoms in adults assessed by questionnaires. A literature review.
        Chest. 1993; 104: 600-608
        • Rönmark E.P.
        • Ekerljung L.
        • Lötvall J.
        • Torén K.
        • Rönmark E.
        • Lundbäck B.
        Large scale questionnaire survey on respiratory health in Sweden: effects of late- and non-response.
        Respir. Med. 2009; 103: 1807-1815
        • Saarivirta T.
        • Consoli D.
        • Dhondt P.
        The evolution of the Finnish health-care system early 19th century and onwards.
        Int. J. Bus. Soc. Sci. 2012; 3: 243-257