Advertisement

Poor lung function and tonsillectomy in childhood are associated with mortality from age 18 to 44

  • D. Mészáros
    Affiliations
    Menzies Research Institute, University of Tasmania, Private Bag 23, Hobart, Tasmania 7001, Australia
    Search for articles by this author
  • S.C. Dharmage
    Correspondence
    Corresponding author. Tel.: +61 3 8344 0737; fax: +61 3 9349 5815.
    Affiliations
    Centre for Molecular Environmental Genetic and Analytic Epidemiology, School of Population Health, University of Melbourne, Level 2, 723 Swanston Street, Carlton Victoria 3052, Australia
    Search for articles by this author
  • M.C. Matheson
    Affiliations
    Centre for Molecular Environmental Genetic and Analytic Epidemiology, School of Population Health, University of Melbourne, Level 2, 723 Swanston Street, Carlton Victoria 3052, Australia
    Search for articles by this author
  • A. Venn
    Affiliations
    Menzies Research Institute, University of Tasmania, Private Bag 23, Hobart, Tasmania 7001, Australia
    Search for articles by this author
  • C.L. Wharton
    Affiliations
    Centre for Molecular Environmental Genetic and Analytic Epidemiology, School of Population Health, University of Melbourne, Level 2, 723 Swanston Street, Carlton Victoria 3052, Australia
    Search for articles by this author
  • D.P. Johns
    Affiliations
    Menzies Research Institute, University of Tasmania, Private Bag 23, Hobart, Tasmania 7001, Australia
    Search for articles by this author
  • M.J. Abramson
    Affiliations
    Department of Epidemiology and Preventive Medicine, Monash University, Central & Eastern Clinical School, Melbourne Victoria 3004, Australia
    Search for articles by this author
  • G.G. Giles
    Affiliations
    Cancer Epidemiology Centre, The Cancer Council of Victoria, 1 Rathdowne 18 Street, Carlton Victoria 3053, Australia
    Search for articles by this author
  • J.L. Hopper
    Affiliations
    Centre for Molecular Environmental Genetic and Analytic Epidemiology, School of Population Health, University of Melbourne, Level 2, 723 Swanston Street, Carlton Victoria 3052, Australia
    Search for articles by this author
  • E.H. Walters
    Affiliations
    Menzies Research Institute, University of Tasmania, Private Bag 23, Hobart, Tasmania 7001, Australia

    Centre for Molecular Environmental Genetic and Analytic Epidemiology, School of Population Health, University of Melbourne, Level 2, 723 Swanston Street, Carlton Victoria 3052, Australia

    Department of Epidemiology and Preventive Medicine, Monash University, Central & Eastern Clinical School, Melbourne Victoria 3004, Australia
    Search for articles by this author
Open ArchivePublished:January 15, 2010DOI:https://doi.org/10.1016/j.rmed.2009.12.001

      Summary

      Background

      The aim of this analysis was to examine associations between lung health in childhood and mortality between ages 18 and 44 years in the Tasmanian Longitudinal Health Study (TAHS).

      Methods

      The 1961 Tasmanian birth cohort who attended school in 1968 (n=8583) were linked to the Australian National Death Index (NDI) to identify deaths. Additional deaths were notified by families through a 37 year follow-up postal questionnaire. Information on lung health at age 7 years and on potential confounders was obtained from the original 1968 TAHS survey and school medical records. Cox proportional hazards modelling was used to assess determinants of mortality.

      Results

      A total of 264 (3%) deaths were identified. The principal causes of death were external injury (56.1%, n=97) and cancer (17.9%, n=31). Males were more likely than females to have died (p=<0.1). Only two (1.1%) participants had died from respiratory conditions. Having an FEV1<80% predicted at 7 years of age was associated with a 2-fold increased incidence of death. Tonsillectomy before age 7 years was associated with a 1.5-fold increase in mortality (p=0.05); being male with a 3.6-fold increase in mortality (p=0.0001); and repeated chest illnesses at age 7 years causing >30 days confinement in the last year, was associated with a 2.2-fold increase in mortality (p=0.03).

      Conclusions

      Childhood lung health appears to be associated with increased mortality in adulthood, perhaps by affecting the ability to survive trauma, major illnesses and other physical stresses.

      Keywords

      Introduction

      Lung health in early life has been shown to influence morbidity in adulthood.
      • Marossy A.E.
      • Strachan D.P.
      • Rudnicka A.R.
      • Anderson H.R.
      Childhood chest illness and the rate of decline of adult lung function between ages 35 and 45 years.
      However, links between lung health in childhood and premature mortality up to middle age remain unclear.
      Population-based prospective cohort studies have predominantly focused on impaired lung function in middle aged adults and all cause-mortality.
      • Knuiman M.W.
      • James A.L.
      • Divitini M.L.
      • Ryan G.
      • Bartholomew H.C.
      • Musk A.W.
      Lung function, respiratory symptoms, and mortality: results from the Busselton Health Study.
      It has been demonstrated that impaired lung function in adults without existing lung disease, such as decline in FEV1 (forced expiratory volume in 1 s)
      • Ryan G.
      • Knuiman M.W.
      • Divitini M.L.
      • James A.
      • Musk A.W.
      • Bartholomew H.C.
      Decline in lung function and mortality: the Busselton Health Study.
      and diffusing capacity,
      • Neas L.M.
      • Schwartz J.
      Pulmonary function levels as predictors of mortality in a national sample of US adults.
      is a risk marker for early mortality in middle age from a wide range of diseases. Similarly, studies examining the risk of death for adults with established respiratory disease have found that the survival from chronic obstructive pulmonary disease (COPD) and lung cancer is worse for those with adult asthma.
      • Mannino D.M.
      • Buist A.S.
      • Petty T.L.
      • Enright P.L.
      • Redd S.C.
      Lung function and mortality in the United States: data from the First National Health and Nutrition Examination Survey follow up study.
      Childhood lung health and premature death in young to middle aged adults has received less research attention. Population-based prospective longitudinal studies that span from childhood to adulthood, and include measurements of lung health and lung function, are required to accurately investigate this association. There is increasing evidence that, to provide valid inference, such associations should be examined while controlling for potential confounders operating in childhood such as gender, socio-economic status, parental occupations, and rural residence, which are known to influence adult mortality.
      • Power C.
      • Hypponen E.
      • Smith G.D.
      Socioeconomic position in childhood and early adult life and risk of mortality: a prospective study of the mothers of the 1958 British birth cohort.
      The three most important studies assessing connections between early lung health and health outcomes in later adult life are: the National Child Development Study (NCDS) in the United Kingdom, the Tasmanian Longitudinal Health Study (TAHS) and the Melbourne Epidemiological Study (MES) of childhood asthma.
      • Power C.
      A review of child health in the 1958 birth cohort: National Child Development Study.
      • Phelan P.D.
      • Robertson C.F.
      • Olinsky A.
      The Melbourne Asthma Study: 1964-1999.
      • Martinez F.D.
      Links between pediatric and adult asthma.
      TAHS recruited a total population-based birth cohort from a single island state of Australia and was specifically designed to assess the respiratory health of children who were born in 1961 and attending school in Tasmania at age 7 years. In 1968, baseline characteristics were collected including lung function, general lung health, allergic diseases, having a tonsillectomy (as a marker for upper respiratory tract disease), and parental smoking. Since then four follow-up surveys have been undertaken, two of which were clinical sub-studies including repeat lung function testing.
      • Gibson H.B.
      • Silverstone H.
      • Gandevia B.
      Respiratory disorders in seven-year-old children in Tasmania: aims, methods and administration of the survey.
      • Giles G.G.
      • Lickiss N.
      • Gibson H.B.
      Respiratory symptoms in Tasmanian adolescents: a follow up of the 1961 birth cohort.
      • Jenkins M.A.
      • Hopper J.L.
      • Flander L.B.
      • et al.
      The associations between childhood asthma and atopy, and parental asthma, hay fever and smoking.
      • Jenkins M.A.
      • Hopper J.L.
      • Bowes G.
      • et al.
      Factors in childhood as predictors of asthma in adult life.
      We have completed a 37-year follow-up of TAHS participants, who are now middle-aged. During the tracing process for this follow-up we linked the TAHS cohort to the National Death Index (NDI). This linkage provided an opportunity to investigate childhood lung health and subsequent mortality into middle-age. As the NDI was only established in 1980, there was a lack of comprehensive death records for Australia prior to this time. Our analysis, therefore, is limited to examining associations between early lung health and the risk of mortality from age 18 to 44 years.

      Methods

      Study population

      The TAHS commenced in 1968 when parents of a cohort of 8583 Tasmanian children (participants) born in 1961 and attending a school in Tasmania completed a questionnaire on the respiratory health of their children. These children also had medical examinations and lung function measurements. Subsequently four follow-up surveys have been carried out at the ages of 13 (in 1974), 20 (in 1981), 31 (in 1992) and 44 years (in 2004) on either the total or sub-groups of the participants. The methodology and some results of previous surveys have been reported elsewhere.
      • Nakajima K.
      • Dharmage S.C.
      • Carlin J.B.
      • et al.
      Is childhood immunisation associated with atopic disease from age 7 to 32 years.
      The 37 year follow-up of the TAHS participants started in 2001 and has involved tracing the original participants from 1968. Of the participants 7312 (85.2%) were traced to an address and sent a detailed respiratory questionnaire. The methodology of tracing participants has been reported elsewhere.
      • Wharton C.L.
      • Dharmage S.C.
      • Jenkins M.A.
      • et al.
      Tracing 8,600 participants 36 years after recruitment at age 7 for the Tasmanian Asthma Study.
      Of these, 78.4% (n=5729) returned the questionnaire.

      National Death Index Matching

      The NDI contains death records for Australia since 1 January 1980, when participants were 18 years old. NDI linkage was carried out by the Australian Institute of Health and Welfare (AIHW) using first name, middle name or initial, surname, date of birth and gender. Probabilistic linkage rules were used where multiple passes grouped the data based on various characteristics and applied a weighting for the type of match. The linkage factored in minor alternatives of spelling and phonetic variations of names. All possible matches and their weights were reviewed and the most likely match decided. Data received from the NDI included each deceased participant's name, date of birth, date of death and underlying cause of death (COD) if available. Other deaths not recorded by the NDI were notified by families as a result of direct contact through the follow-up postal questionnaire.

      Exposures, confounders and definitions

      Data on exposures and confounders were extracted from the multiple TAHS follow up surveys and the school medical records which were retrieved from the Tasmanian Archive Office. Lung function (FEV1 and FVC [forced vital capacity]) was measured in the 1968 clinical assessment with a wedge bellows spirometer (Vitalograph Limited, Maids Moreton, UK). FEV1/FVC ratio was computed using raw FEV1 and FVC values. Predicted values for FEV1 and FVC were calculated from height and gender using equations by Zapletal et al.
      • Zapletal A.
      • Paul T.
      • Samanek M.
      Die Bedeutung heutiger Methoden der Lungenfunktionsdiagnostik zur Feststellungeiner Obstruktion der Atemwege bei Kindern und Jugendlichen (Significance of contemporary methods of lung function testing for the detection of airway obstruction in children and adolescents).
      Impaired lung function at age seven years was defined in two ways; (1) as an FEV1 less than 80% of predicted and (2) as an FEV1/FVC ratio less than 75%.
      Tonsillectomy was ascertained from the 1968 questionnaire by an affirmative response to the question ‘Has he/she had the tonsil removed’?
      Tonsillitis in the last 12 months was defined by the question ‘Has he/she had more than 2 sore throats or attacks of tonsillitis in the past 12 months’?
      Food allergies was defined by ‘Have you been told by a doctor that he/she is allergic to any foods or medicines’?
      Asthma in 1968 was defined by an affirmative response to the question ‘Has he/she at any time in his/her life suffered from attacks of asthma or wheezy breathing’?
      Bronchitis in 1968 was defined by an affirmative response to the question ‘Has he/she at any time in his/her life suffered from attacks of bronchitis or attacks of cough with sputum (phlegm) in the chest (“loose” or rattly” cough)’?
      Pneumonia/ pleurisy in 1968 was defined by an affirmative response to the question ‘Have you ever been told by a doctor that he/she had pneumonia or pleurisy’?
      Hay fever in 1968 was defined from by an affirmative answer to the question ‘Does he/she get attacks of “hay fever” (that is, sneezing, running or blocked nose, sometimes with itchy eyes or nose)’?
      Repeated chest illness over the last 12 months was ascertained from the 1968 questionnaire by the question ‘For how much time in the last 12 months has the child been confined to the house because of chest illness?, with parents giving one of the following responses ‘not at all’, ‘1–7 days’, ‘8–30 days’, or ‘>30 days’.
      Maternal and paternal smoking was ascertained from the 1968 questionnaire with an affirmative response to the question “Do you smoke every day (six out of seven)?”. Smoking intensity was defined by one of the following response to the question “How much do you smoke'?, ‘Less than 6 cigarettes a day’, ‘6–20 cigarettes a day’, and ‘more than 20 cigarettes’.
      Region of Residence at age 7 years was defined using postcodes of the school the child was attending in 1968. These postcodes were grouped into three regions, (1) South, (2) North, and (3) North-West, as the regional centres of the three discrete population areas in Tasmania.
      In addition, the paternal and maternal occupational class in 1968 was extracted from the archived school medical records available for each child. This was used as a measure of socio-economic status (SES) and coded according to the Australian Standard Classification of Occupations (ASCO) four-digit codes.
      • McLennan W.
      For paternal occupation these codes were then grouped into four major skill groups: (1) Managers/Professionals; (2) Tradespersons and Advanced Clerical; (3) Intermediate clerical and production; (4) Elementary clerical, labourers and related workers. The majority of women were classified as “house duties” and the classifications for maternal occupation were (1) Housewife; (2) Professional; (3) Tradespersons and other workers.

      Statistical analyses

      The significance of the univariate associations between vital status and selected characteristics were evaluated using chi-square tests and by computing crude risk ratios (RR) and corresponding 95% confidence intervals (CI). Multiple logistic regression was conducted to examine the independent associations between these childhood factors and death beyond the age of 18 years. Selected variables were included in this multivariate model if they were significant at p0.1 in the univariate analysis.
      Mortality was estimated by dividing the number of deaths by the total number of person-years at risk, from the commencement of the NDI to the age of death for deceased subjects, or to the age at the time of the NDI linkage for non-deceased subjects. To study the associations between mortality and risk factors, Cox proportional hazards models were fitted using the age at death as the dependent variable. The variables included in the model were gender, parental smoking (father and mother separately), region of residence at age 7 years, tonsillectomy and FEV1<80% predicted in 1968 as potential predictors of mortality. Results are presented as hazard ratios (HRs) with 95% CIs. All tests were two tailed and, following convention, p<0.05 was considered to be the threshold for statistical significance. All analyses were performed using Stata (StataCorp. Stata Statistical Software, Release 9 College Station, TX).

      Ethical considerations

      The study was approved by the Human Research Ethics Committees at the Universities of Tasmania and Melbourne and Australian Institute of Health and Welfare.

      Results

      A total of 264 participants were found to have died. The NDI identified 211, with the remaining 53 identified by family members. Age at death was available for 211 participants, with a mean of 32.9 years (SD 8.6). Underlying cause of death was available for 173 participants. The principal cause of death was external trauma (56.1%, n=97). Cancer (17.9%, n=31) and circulatory system diseases (9.8%, n=17) were the next most common causes of death. Only two (1.1%) participants died from respiratory conditions (pulmonary sepsis) (Fig. 1). Forty of these 173 participants had a tonsillectomy, and the principal cause of death for these individuals was cancer (14.5%, n=25), followed by endocrine disorders (5.7%, n=10), mental and respiratory disorders (1.4%, n=2)
      Figure thumbnail gr1
      Figure 1Cause specific mortality of the 173 deceased participants for whom cause of death was available.
      Table 1 shows that being male, having repeated chest illness causing >30 days confinement over a 12 month period, maternal smoking and having a tonsillectomy by age 7 years were associated with an increased risk of death. There were also modest associations with the region of residence (those living in the Burnie region of NW Tasmania having the greatest risk) and maternal occupational class in 1968 (all p<0.1). No other characteristics (i.e. food allergies, asthma, bronchitis, hay fever or father smoking) were found to be associated with an increased risk of death.
      Table 1Selected respiratory and other characteristics at age 7 years by vital status, with crude and adjusted risk ratios comparing those who were deceased to those who were not.
      Alive n=8321 (%)Deceased n=264 (%)Unadjusted risk ratios (95% CI)Adjusted
      Mutually adjusted for all variables in the table with a p value<0.1.
      risk ratios (95% CI)
      Sex – male4122 (49.5)195 (73.9)2.70 (2.06,3.54)*3.07 (2.19,4.31)
      Tonsillectomy1238 (15.3)50 (19.2)1.30 (0.96,1.76)*1.42 (1.00,2.03)
      Tonsillitis in the last 12 months1843 (23.6)65 (26.0)1.13 (0.86,1.49)
      Food Allergies547 (6.8)23 (8.9)1.33 (0.87,2.02)
      Asthma in 19681348 (16.7)51 (19.8)1.22 (0.91,1.65)
      Bronchitis3831 (47.4)129 (49.8)1.10 (0.86,1.39)
      Pneumonia/ pleurisy 19686856 (85.8)214 (83.3)1.21 (0.87,1.67)
      Hayfever1026 (12.8)35 (13.6)1.07 (0.75,1.52)
      Repeated chest illness over last 12 months
       Confined to house – not at all5166 (65.3)155 (62.3)1.01.0
       Confined to house – 1–7 days2012 (25.4)63 (25.3)1.04 (0.78,1.39)0.92 (0.64,1.34)
       Confined to house – 8–30 days528 (6.7)16 (6.4)1.01 (0.61,1.68)0.82 (0.42,1.60)
       Confined to house – >30 days202 (2.6)15 (6.0)2.37 (1.42,3.96)*2.06 (1.12,3.80)
       FEV1/FVC less than 75% at 7 yrs108 (1.3)6 (2.3)1.73 (0.78,3.80)
       FEV1 less than 80% at 7 yrs226 (2.7)13 (4.9)1.81 (1.05,3.11)1.98 (1.04,3.75)
      Mother smoking – yes2925 (37.6)108 (43.4)1.26 (0.99,1.62)*1.30 (0.97,1.75)
      Mother smoking intensity
       (1) None4863 (62.6)141 (56.6)1.0
       (2) Less than 6 cigarettes a day501 (6.5)20 (8.0)1.36 (0.86,2.16)
       (3) 6–20 Cigarettes a day1946 (25.1)71 (28.5)1.25 (0.94,1.65)
       (4) More than 20 cigarettes457 (5.9)17 (6.8)1.27 (0.78,2.09)
      Father smoking – yes4665 (61.4)154 (63.1)1.07 (0.83,1.39)*0.94 (0.70,1.75)
      Father smoking intensity
       (1) None2930 (39.6)90 (38.3)1.0
       (2) Less than 6 cigarettes a day1542 (20.8)53 (22.6)1.09 (0.78,1.52)
       (3) 6–20 cigarettes a day2670 (36.1)89 (37.9)1.07 (0.81,1.43)
       (4) More than 20 cigarettes256 (3.5)3 (1.3)0.35 (0.11,1.11)
      Region of residence at age 7
       (1) Hobart3696 (45.1)104 (41.3)1.01.0
       (2) Launceston2329 (28.4)66 (26.2)1.01 (0.74,1.36)0.95 (0.67,1.35)
       (3) Burnie2169 (26.5)82 (32.5)1.33 (1.00,1.77)*1.17 (0.83,1.66)
      Father occupational class in 1968
       Managers/Professionals2126 (27.4)57 (24.5)1.01.0
       Tradespersons/Advanced Clerical2297 (29.6)72 (30.9)1.16 (0.83,1.64)1.26 (0.86,1.85)
       Inter clerical & production2209 (28.5)68 (29.2)1.14 (0.81,1.62)1.10 (0.74,1.64)
       Elementary Clerical & Labourers1122 (14.5)36 (15.5)1.19 (0.79,1.80)1.26 (0.78,2.05)
      Mother occupational class in 1968
       Housewife6721 (91.7)195 (88.2)1.01.0
       Professional276 (3.8)14 (6.3)1.71 (1.01,2.91)*1.63 (0.92,2.87)
       Tradespersons & other workers330 (4.5)12 (5.4)1.24 (0.70,2.21)1.38 (0.74,2.59)
      *p<0.1.
      a Mutually adjusted for all variables in the table with a p value<0.1.
      Table 2 shows factors associated with an increased incidence of death. The multivariate analysis found a significant 2.4-fold increased incidence of death in children with an FEV1<80% predicted at age 7. Tonsillectomy before age 7 years was associated with a 1.5-fold increased mortality (p=0.046); being male with a 3.7-fold increased mortality (p=0.0001); and repeated chest illness by age 7 years causing >30 days confinement in the last year, with 2.2-fold increased mortality (p=0.027).
      Table 2Mortality and hazard ratios by selected respiratory and other characteristics.
      CharacteristicPerson-yearsDeathsMortality per /1000/yr (95% CI)HR (95%CI)p Value
      Male190,1681620.85 (0.73, 0.99)3.75 (2.53, 5.56)0.0001
      Female182,464470.26 (0.19, 0.34)1.0
      Tonsillectomy at age 7
       Yes57,232430.75 (0.56, 1.01)1.50 (1.01, 2.23)0.05
       No313,3471650.53 (0.45, 0.61)1.0
      FEV1 less than 80% at 7 yrs
       Yes10,328.68111.06 (0.59, 1.92)2.38 (1.21, 4.69)0.01
       No362,304.701980.55 (0.48, 0.63)1.0
      Mother smoking
       Yes134,495850.63 (0.51, 0.78)1.26(0.90,1.76)0.18
       No222,6901120.50 (0.42, 0.61)1.0
      Father smoking
       Yes214,0051210.57 (0.47, 0.68)0.96 (0.68, 1.35)0.81
       No134,364730.54 (0.43, 0.68)1.0
      Repeated chest illness over last 12 months
       Confined to house – not at all236,8241280.54 (0.45, 0.64)1.0
       Confined to house – 1–7 days92,167510.55 (0.42, 0.73)1.01 (0.70, 1.46)0.97
       Confined to house – 8–30 days24,100100.41 (0.22, 0.77)0.83 (0.42, 1.65)0.60
       Confined to house – >30 days9497111.16 (0.64, 2.09)2.17 (1.09, 4.32)0.03
      Region
       (1) South166,117820.49 (0.40, 0.61)1.0
       (2) North103,461530.51 (0.39, 0.67)0.98 (0.66, 1.46)0.94
       (3) North West97,206660.68 (0.53, 0.86)1.27 (0.86, 1.87)0.23
      Adjusted for gender, maternal smoking, region of residence, tonsillectomy by age 7, FEV1<80% at age 7, father's and mother's occupational class in 1968.

      Discussion

      We observed being male, having airflow limitation and having a tonsillectomy before the age of 7 years to be strongly associated with an increased risk of premature adult death in a cohort of followed from the age of 7 years to middle age. Furthermore, having low lung function at age 7, a mother who smoked and having severe repeated childhood chest illness by age 7 were also implicated, but a history of asthma and other allergic diseases were not related to mortality.
      The principal COD was external injury, followed by malignant neoplasms and diseases of the circulatory system. Only two participants died from diseases of the respiratory system. Australian Bureau of Statistics (ABS) data confirm that the relative contribution of various underlying causes of death differs with age, and nationally the most common COD for the age groups 15–24 years and 25–44 years were identical to the causes of mortality for our cohort.
      • Australia's Health
      The tenth biennial health report of the Australian Institute of Health and Welfare.
      The mean age of death for our participants was 32.9 years, so trauma would be expected to be the leading COD. We did not expect chronic respiratory diseases (such as emphysema and COPD) to be a significant COD up to this age, as they are relatively uncommon in younger age groups.
      It is known that airway obstruction and impaired FEV1 in adults are risk factors for mortality from a variety of diseases.
      • Hole D.J.
      • Watt G.C.
      • Davey-Smith G.
      • Hart C.L.
      • Gillis C.R.
      • Hawthorne V.M.
      Impaired lung function and mortality risk in men and women: findings from the Renfrew and Paisley prospective population study.
      We observed an FEV1<80% predicted at the age of 7 years to be independently associated with an increased risk of death. An FEV1 less than 80% is a measure of airflow limitation due to airway obstruction. We do not have adult measurements for these individuals, but we presume that lung function would also have been relatively impaired in adulthood. There is limited information on paediatric lung function and its association with ventilatory impairment and mortality in adulthood.
      • Barker D.J.
      • Godfrey K.M.
      • Fall C.
      • Osmond C.
      • Winter P.D.
      • Shaheen S.O.
      Relation of birth weight and childhood respiratory infection to adult lung function and death from chronic obstructive airways disease.
      Burrows et al.
      • Burrows B.
      • Knudson R.J.
      • Lebowitz M.D.
      The relationship of childhood respiratory illness to adult obstructive airway disease.
      suggested that lower FEV1/FVC ratio in childhood indicated an underlying respiratory illness, which could increase susceptibility to the adverse effects of a variety of bronchial irritants and infectious agents.
      It has also been found that the relative risk of all cause mortality in adults increased with the number of respiratory symptoms reported.
      • Frostad A.
      • Søyseth V.
      • Andersen A.
      • Gulsvik A.
      Respiratory symptoms as predictors of all-cause mortality in an urban community: a 30-year follow-up.
      Our data suggest that this increased risk starts early in life. Repeated and prolonged childhood chest illness was associated with mortality and this association was independent of lung function at age 7. The reasons for this association could be multifactorial and unfortunately we do not have more detailed information on the exact nature of these chest illnesses.
      As the principal COD was external injury, it is possible that poor lung function may impair the ability to survive trauma. What is unclear is whether it is the incidence or type of trauma that was different in these individuals or their ability for survival. It is known that having a pre-existing disease can affect post-operative complications, inability to survive prolonged lengths of stay in hospital, and mortality.
      • Bochiccio G.V.
      • Joshi M.
      • Bochiccio K.
      • Shih D.
      • Meyer W.
      • Scalea T.M.
      Incidence and impact of risk factors in critically ill trauma patients.
      Published data on the incidence of risk factors and their effects in trauma patients, have failed to examine the possible link between trauma and impaired lung function, particularly in middle-aged adults.
      Another finding was that having had a tonsillectomy by the age of 7 years was associated with an increased risk of death. Tonsillectomy is an indicator of chronic and recurrent tonsillitis and pharyngitis as well as failure of antibiotic treatment for tonsillitis. A large number of participants in our cohort had a tonsillectomy as they were born at a time when tonsillectomies were a very common procedure performed on children. In a study undertaken by Witucki
      • Witucki J.
      Tonsillectomy analysis of the data for the 25-year period.
      the great majority of tonsillectomies performed between 1961 and 1985 were on children aged 3–7 years, and the rate of tonsillectomies performed during this time frame reduced by 32% (from 14.5 to 9.8 for 10,000). Specific long-term follow-up studies of the consequences of tonsillectomies are lacking, but several population-based studies have attempted to assess the long-term health outcomes for patients who underwent a tonsillectomy in childhood. They suffered more respiratory tract infections,
      • Johansson E.
      • Hultcrantz E.
      Tonsillectomy – clinical consequences twenty years after surgery.
      and increased risk of Hodgkin's disease and lymphocytic (but not myeloid) leukaemia in adulthood.
      It would seem that tonsillectomy may have permanent adverse effects on the immune system. Alternatively, the need for a tonsillectomy may identify a group of already immunocompromised individuals who were susceptible to developing respiratory infections. Certainly TAHS subjects who had undergone a tonsillectomy reported higher frequencies of other infections (data not shown). Another possibility is that tonsillectomy is a marker of some other health issue which we have not measured. Whatever the precise link, tonsillectomy (like chest illness and impaired lung function) may be highlighting a group less able to cope with trauma or other severe physiological stress and its consequences such as intubation and infection.
      An alternative explanation for the association between tonsillectomy, low FEV1 and survival may be related to residual confounding due to socioeconomic status (SES). Our measures of SES were limited to relatively crude measures of parental occupational class, and region of residence. We did not have any information on parental education, maternal diet, smoking during pregnancy, housing, diet in childhood or income levels in 1968. The Australian Bureau of Statistics now produces a socioeconomic status index based on postcode of residence, but no index was available for 1968. While we acknowledge some limitations in our measures of SES, the association between mortality, low FEV1 and tonsillectomy strengthened after adjusting for the SES measures we had available suggesting that there was negative confounding by these SES measures. These results are consistent with other published studies.
      • Kanner R.E.
      • Renzetti Jr., A.D.
      • Stanish W.M.
      • Barkman Jr., H.W.
      • Klauber M.R.
      Predictors of survival in subjects with chronic airflow limitation.
      We have confirmed that our tracing of deceased individuals was likely to be comprehensive using NDI linkage and through family contact with the 37 year follow-up postal questionnaire. The potential for selection bias in this study would have been small given the excellent participation achieved in the 1968 study The NDI has been identified as an accurate and sensitive database. Nagle et al.
      • Nagle C.M.
      • Purdie D.M.
      • Webb P.M.
      • Green A.C.
      • Bain C.J.
      Searching for cancer deaths in Australia: National Death Index vs. cancer registries.
      and Magliano et al.
      • Magliano D.
      • Liew D.
      • Pater H.
      • et al.
      Accuracy of the Australian national Death Index: comparison with adjudicated fatal outcomes among Australian participants in the long term intervention with Pravastatin in Ischaemic Disease (LIPID) study.
      found that the NDI had a sensitivity of 93% and specificity of 100%. In spite of this we were surprised to find fifty three participants (20% of deaths) who were not listed in the NDI. These individuals were identified though family members while conducting the TAS 37 year follow-up postal survey. The AIHW estimates that approximately 1% of deaths will not be recorded in the National Death Index, but we found this to be much higher in our cohort. In some instances, after a notification of death, the registrars of Births Deaths and Marriages (BDM) do not receive enough information to fully complete a death certificate. Consequently, they delay sending the information to the AIHW until they have obtained the missing details. This seemed to be a particular problem with Coroners' cases which would have predominated in these young traumatic deaths. Deaths that occurred overseas would also have been missed. There were also limitations to the causes of death analysis as only 173 participants had a recorded cause of death in the NDI. Two rounds of linkage to the NDI at different times (December 2002 and May 2006) obtained the same number of deaths in each disease category. Unfortunately, in some instances the COD is not recorded on the NDI. This is especially the case for Coroners' cases where an underlying cause of death was not always able to be determined.
      There may have been a potential for selection bias in this study because of missing data, particularly those relating to lung function. Access to Tasmanian death records to find the missing data was not possible due to restrictions in access to recent records. However we believe that respiratory diseases like Cystic Fibrosis would not have been a significant contributor to deaths after the age of 21, because it is unlikely that such individuals would have survived beyond childhood in this cohort.
      Since the National Death Index was established only in 1980, we were only able to examine the association between early lung health and the risk of death from 18 to 44 years of age. Deaths in the age group between 7-17 years would be expected to include those attributable to lethal childhood conditions e.g. inborn errors of metabolism, congenital malformations, but these were not of primary interest to our study. We were unable to examine adult lung function and compare it with childhood lung function in this analysis because lung function data from the TAHS 1974 and 1981 clinical sub-study were only undertaken when the probands were 13 and 20 years of age, respectively. In addition, lung function data from the 1981 sub-study were only available for 250 subjects.
      In conclusion, low FEV1 in childhood is associated with subsequent mortality in adulthood, although the mechanism is unclear. Airway obstruction at age 7 years was associated with a 2-fold increased risk of death. Whilst further large-scale longitudinal studies are required to investigate this issue more extensively, the findings from our study suggest that routine screening of lung function could be considered as part of health and risk assessment of children. Parental smoking had a long lasting negative effect, as may also be the case for severe childhood chest illness, and both need to be considered during prognostic risk assessment for severely ill young adults.

      Funding

      This work was supported by grants from the National Health and Medical Research Council, Clifford Craig Medical Research Trust, The Victorian, Tasmanian, and Queensland Asthma Foundations.

      Conflict of interest statement

      The authors have no conflicts of interest to report.

      Acknowledgements

      The authors would like to thank the Australian Institute of Health and Welfare, and John Harding for their assistance with linkage to the National Death Index and Tasmanian State Archives and Jillian Waters for their assistance with linkage to the School Medical Records. This work was supported by grants from the National Health and Medical Research Council, Clifford Craig Medical Research Trust, The Victorian, Tasmanian, and Queensland Asthma Foundations. Shyamali Dharmage, Melanie Matheson and John Hopper are supported by NHMRC. The study sponsors had no involvement in the study design, in the collection, analysis and interpretation of data, in the writing of the manuscript, and in the decision to submit the manuscript for publication.

      References

        • Marossy A.E.
        • Strachan D.P.
        • Rudnicka A.R.
        • Anderson H.R.
        Childhood chest illness and the rate of decline of adult lung function between ages 35 and 45 years.
        Am J Respir Crit Care Med. 2007; 175: 355-359
        • Knuiman M.W.
        • James A.L.
        • Divitini M.L.
        • Ryan G.
        • Bartholomew H.C.
        • Musk A.W.
        Lung function, respiratory symptoms, and mortality: results from the Busselton Health Study.
        Ann Epidemiol. 1999; 9: 297-306
        • Ryan G.
        • Knuiman M.W.
        • Divitini M.L.
        • James A.
        • Musk A.W.
        • Bartholomew H.C.
        Decline in lung function and mortality: the Busselton Health Study.
        J Epidemiol Community Health. 1999; 53: 230-234
        • Neas L.M.
        • Schwartz J.
        Pulmonary function levels as predictors of mortality in a national sample of US adults.
        Am J Epidemiol. 1998; 147: 1011-1018
        • Mannino D.M.
        • Buist A.S.
        • Petty T.L.
        • Enright P.L.
        • Redd S.C.
        Lung function and mortality in the United States: data from the First National Health and Nutrition Examination Survey follow up study.
        Thorax. 2003; 58: 388-393
        • Power C.
        • Hypponen E.
        • Smith G.D.
        Socioeconomic position in childhood and early adult life and risk of mortality: a prospective study of the mothers of the 1958 British birth cohort.
        Am J Public Health. 2005; 95: 1396-1402
        • Power C.
        A review of child health in the 1958 birth cohort: National Child Development Study.
        Paediatr Perinat Epidemiol. 1992; 6: 81-110
        • Phelan P.D.
        • Robertson C.F.
        • Olinsky A.
        The Melbourne Asthma Study: 1964-1999.
        J Allergy Clin Immunol. 2002; 109: 189-194
        • Martinez F.D.
        Links between pediatric and adult asthma.
        J Allergy Clin Immunol. 2001; 107: 449-455
        • Gibson H.B.
        • Silverstone H.
        • Gandevia B.
        Respiratory disorders in seven-year-old children in Tasmania: aims, methods and administration of the survey.
        Med J Aust. 1969; 2: 201-205
        • Giles G.G.
        • Lickiss N.
        • Gibson H.B.
        Respiratory symptoms in Tasmanian adolescents: a follow up of the 1961 birth cohort.
        Aust N Z J Med. 1984; 14: 631-637
        • Jenkins M.A.
        • Hopper J.L.
        • Flander L.B.
        • et al.
        The associations between childhood asthma and atopy, and parental asthma, hay fever and smoking.
        Paed Perinatal Epidemiol. 1993; 7: 67-76
        • Jenkins M.A.
        • Hopper J.L.
        • Bowes G.
        • et al.
        Factors in childhood as predictors of asthma in adult life.
        BMJ. 1994; 309: 90-93
        • Nakajima K.
        • Dharmage S.C.
        • Carlin J.B.
        • et al.
        Is childhood immunisation associated with atopic disease from age 7 to 32 years.
        Thorax. 2007; 62: 270-275
        • Wharton C.L.
        • Dharmage S.C.
        • Jenkins M.A.
        • et al.
        Tracing 8,600 participants 36 years after recruitment at age 7 for the Tasmanian Asthma Study.
        Australian and New Zealand Journal of Public Health. 2006; 30: 105-110
        • Zapletal A.
        • Paul T.
        • Samanek M.
        Die Bedeutung heutiger Methoden der Lungenfunktionsdiagnostik zur Feststellungeiner Obstruktion der Atemwege bei Kindern und Jugendlichen (Significance of contemporary methods of lung function testing for the detection of airway obstruction in children and adolescents).
        Z Erkr Atmungsorgane. 1977; 149: 343-371
        • McLennan W.
        ASCO: Australian Standard Classification of Occupations. 2nd ed. Australian Bureau of Statistics, 1997
        • Australia's Health
        The tenth biennial health report of the Australian Institute of Health and Welfare.
        AIHW, Canberra; AIHW2006 (cat. no. AUS73)
        • Hole D.J.
        • Watt G.C.
        • Davey-Smith G.
        • Hart C.L.
        • Gillis C.R.
        • Hawthorne V.M.
        Impaired lung function and mortality risk in men and women: findings from the Renfrew and Paisley prospective population study.
        BMJ. 1996; 313: 711-715
        • Barker D.J.
        • Godfrey K.M.
        • Fall C.
        • Osmond C.
        • Winter P.D.
        • Shaheen S.O.
        Relation of birth weight and childhood respiratory infection to adult lung function and death from chronic obstructive airways disease.
        BMJ. 1991; 303: 671-675
        • Burrows B.
        • Knudson R.J.
        • Lebowitz M.D.
        The relationship of childhood respiratory illness to adult obstructive airway disease.
        Am Rev Respir Dis. 1977; 115: 751-760
        • Frostad A.
        • Søyseth V.
        • Andersen A.
        • Gulsvik A.
        Respiratory symptoms as predictors of all-cause mortality in an urban community: a 30-year follow-up.
        J Intern Med. 2006; 259: 520-529
        • Bochiccio G.V.
        • Joshi M.
        • Bochiccio K.
        • Shih D.
        • Meyer W.
        • Scalea T.M.
        Incidence and impact of risk factors in critically ill trauma patients.
        World J Surgery. 2006; 30: 114-118
        • Witucki J.
        Tonsillectomy analysis of the data for the 25-year period.
        Otolaryngol Pol. 1990; 44: 267-275
        • Johansson E.
        • Hultcrantz E.
        Tonsillectomy – clinical consequences twenty years after surgery.
        Int J Pediatr Otorhinolaryngol. 2003; 67: 981-988
        • Kanner R.E.
        • Renzetti Jr., A.D.
        • Stanish W.M.
        • Barkman Jr., H.W.
        • Klauber M.R.
        Predictors of survival in subjects with chronic airflow limitation.
        Am J Med. 1983; 74: 249-255
        • Nagle C.M.
        • Purdie D.M.
        • Webb P.M.
        • Green A.C.
        • Bain C.J.
        Searching for cancer deaths in Australia: National Death Index vs. cancer registries.
        Asian Pac J Cancer Prev. 2006; 7: 41-45
        • Magliano D.
        • Liew D.
        • Pater H.
        • et al.
        Accuracy of the Australian national Death Index: comparison with adjudicated fatal outcomes among Australian participants in the long term intervention with Pravastatin in Ischaemic Disease (LIPID) study.
        Aust N Z J Public Health. 2003; 27: 649-653