Respiration 33: 236-244 (1976)
Chronic Cough in Young Adults in Relation to Smoking Habits, Childhood Environment and Chest Illness K. E. K iernan, J. R. T. C olley, J. W. B. D ouglas and D. D. R eid MRC Unit on Environmental Factors in Mental and Physical Illness, London School of Economics (Director: J. W. B. D ouglas), and London School of Hygiene and Tropical Medicine, London
Key Words. Cohort study • Pollution • Social class • Smoking • Childhood chest illness • Adult chronic cough Abstract. A survey of the respiratory symptoms and smoking habits of a population of 20-year-olds - followed since birth - was repeated when they reached the age of 25. The association of cough prevalence with current smoking habits and with lower respiratory tract illness in childhood found in the survey at age 20 was confirmed and in each instance appears to have increased in strength over the 5 years. At age 25, however, the prevalence of cough was associated at a statistically significant level with fathers’ occupation; and this association with social class of origin could not be explained by persisting differences in social status based on the educational levels attained by early adult life. The association with exposure to air pollution in childhood, although more obvious than before, could be due to chance. The prevalence of cough increased between the ages of 20 and 25 among those who smoked throughout or who started to smoke during this period. It declined for those who never smoked and for those who were smoking at 20 but had given up by 25. The implications of these changing patterns of respiratory disease behaviour at a crucial stage between adolescence and adult life are briefly discussed.
Introduction
Received: March 21, 1975; accepted: April 12, 1975.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/18/2018 12:56:39 AM
In a previous paper [Colley et al., 1973] it was shown that the prevalence of cough during the day or night among 20-year-old men and women was related to their current cigarette smoking habits and to whether or not there was a history of lower respiratory tract illness during the first 2 years of life. The occupation of their fathers and the level of air pollution of the area in
K iernan/C olley/D ouglas/R eid
237
which they lived as children showed no such association with respiratory symptoms. The same group of young people, who are members of a 1946 birth cohort which forms the National Survey of Health and Development, have provided further information on their smoking habits and respiratory symptoms at the age of 25. The object of the present paper is to re-examine the relation of the same four factors (i.e., smoking, early lower respiratory tract illness, occupational group and air pollution) to chronic cough in winter at this later age and to determine whether, as the cohort has become older, the nature of the relationships between these factors and respiratory disease has altered.
Information on current smoking habits and respiratory symptoms was collected in 1971 when the survey members, then aged 25, were sent a postal questionnaire which included questions on respiratory symptoms approved by the Medical Research Council’s com mittee on research into chronic bronchitis (1960 version) and questions on smoking. The questions on respiratory symptoms were the same as those used in the previous study of the same group, but the questions on smoking were somewhat different: there was a more rigorous definition of ‘regular smokers' and a number of differences in wording. By com paring in detail the answers given by the same persons to these two sets of questions we have reached equivalent definitions of smokers, past smokers, and non-smokers on the two occasions, based on all types of smoking habit, i.e., cigarette, pipe and cigar smokers. Those who failed to reply to the postal questionnaire after two reminders were inter viewed by health visitors on our behalf. There was also a group, approximately 20% of the sample, who were either of low literacy or known from earlier experience to be unlikely to return postal questionnaires. These were not sent forms but were interviewed by health visitors. Between March and September 1971, questions on smoking and respiratory symptoms were answered by 3,245 survey members. Of these 327 had not been included in the 20-year study and so had to be excluded from this analysis. For the main analysis, a further 830 were excluded for the following reasons or combinations of reasons: (a) incomplete in formation on early childhood chest illness; (b) incomplete information on air pollution; or (c) movement during childhood between areas that varied in the level of air pollution. The definition of air pollution levels has been changed for this analysis. In the earlier paper the levels were the sum of 7 estimates spread over 11 years. Some individuals over these 11 years changed their areas of residence and air pollution exposure. To minimise any effect of changing from one level of air pollution to another on respiratory experience, only those who had lived throughout the first 11 years of their lives in the same area, or had moved to areas of the same level of pollution were taken. This has resulted in the loss of 419 men and women for whom all the other required information was available. The final sample used for the main analysis was 2,088 men and women for whom complete information on childhood chest illness, air pollution, occupation of father, smoking at 20 and 25, and respiratory symptoms at 20 and 25 was available. This group is
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/18/2018 12:56:39 AM
Methods
238
K iernan/C olley/D ouglas/R eid
Table /. Effect of losses and exclusions on factor levels Factors
Childhood chest illness (0-2 years), % High level of pollution (2-11 years), % Manual working class2, % Regular smokers at 20, % Cough day or night at 20, %
Information obtained at earliest age
at 20 years
at 20 and 25 years
25.11 48.5 62.3 44.6 10.0
24.8 47.0 57.2 44.6 10.0
25.1 46.3 60.5 43.6 9.9
1 E.g. 25.1% of children in the sample known at 0-2 years had chest illness and this compares with 24.8% of those contacted at 20 years and 25.1% of those with complete information on all the factors, at age 20 and 25 years. 2 Based on occupation of father when survey members were aged eleven years, or if no information at this age, on the occupation of father at the time of birth of the survey member.
Results Smoking In the earlier study, smoking was found to be particularly relevant to the prevalence of cough. Table II compares the prevalence rates according to the history of smoking behaviour either in those who did not change their habit or gave up or took up smoking between the ages of 20 and 25 years. In all groups, many who reported cough during the day or at night in the winter at one age did not report it at the other. Nevertheless, the highest rates are seen in those smoking at both ages; and, among them, the pre valence rose significantly (p
Chronic Cough in Young Adults in Relation to Smoking Habits, Childhood Environment and Chest Illness K. E. K iernan, J. R. T. C olley, J. W. B. D ouglas and D. D. R eid MRC Unit on Environmental Factors in Mental and Physical Illness, London School of Economics (Director: J. W. B. D ouglas), and London School of Hygiene and Tropical Medicine, London
Key Words. Cohort study • Pollution • Social class • Smoking • Childhood chest illness • Adult chronic cough Abstract. A survey of the respiratory symptoms and smoking habits of a population of 20-year-olds - followed since birth - was repeated when they reached the age of 25. The association of cough prevalence with current smoking habits and with lower respiratory tract illness in childhood found in the survey at age 20 was confirmed and in each instance appears to have increased in strength over the 5 years. At age 25, however, the prevalence of cough was associated at a statistically significant level with fathers’ occupation; and this association with social class of origin could not be explained by persisting differences in social status based on the educational levels attained by early adult life. The association with exposure to air pollution in childhood, although more obvious than before, could be due to chance. The prevalence of cough increased between the ages of 20 and 25 among those who smoked throughout or who started to smoke during this period. It declined for those who never smoked and for those who were smoking at 20 but had given up by 25. The implications of these changing patterns of respiratory disease behaviour at a crucial stage between adolescence and adult life are briefly discussed.
Introduction
Received: March 21, 1975; accepted: April 12, 1975.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/18/2018 12:56:39 AM
In a previous paper [Colley et al., 1973] it was shown that the prevalence of cough during the day or night among 20-year-old men and women was related to their current cigarette smoking habits and to whether or not there was a history of lower respiratory tract illness during the first 2 years of life. The occupation of their fathers and the level of air pollution of the area in
K iernan/C olley/D ouglas/R eid
237
which they lived as children showed no such association with respiratory symptoms. The same group of young people, who are members of a 1946 birth cohort which forms the National Survey of Health and Development, have provided further information on their smoking habits and respiratory symptoms at the age of 25. The object of the present paper is to re-examine the relation of the same four factors (i.e., smoking, early lower respiratory tract illness, occupational group and air pollution) to chronic cough in winter at this later age and to determine whether, as the cohort has become older, the nature of the relationships between these factors and respiratory disease has altered.
Information on current smoking habits and respiratory symptoms was collected in 1971 when the survey members, then aged 25, were sent a postal questionnaire which included questions on respiratory symptoms approved by the Medical Research Council’s com mittee on research into chronic bronchitis (1960 version) and questions on smoking. The questions on respiratory symptoms were the same as those used in the previous study of the same group, but the questions on smoking were somewhat different: there was a more rigorous definition of ‘regular smokers' and a number of differences in wording. By com paring in detail the answers given by the same persons to these two sets of questions we have reached equivalent definitions of smokers, past smokers, and non-smokers on the two occasions, based on all types of smoking habit, i.e., cigarette, pipe and cigar smokers. Those who failed to reply to the postal questionnaire after two reminders were inter viewed by health visitors on our behalf. There was also a group, approximately 20% of the sample, who were either of low literacy or known from earlier experience to be unlikely to return postal questionnaires. These were not sent forms but were interviewed by health visitors. Between March and September 1971, questions on smoking and respiratory symptoms were answered by 3,245 survey members. Of these 327 had not been included in the 20-year study and so had to be excluded from this analysis. For the main analysis, a further 830 were excluded for the following reasons or combinations of reasons: (a) incomplete in formation on early childhood chest illness; (b) incomplete information on air pollution; or (c) movement during childhood between areas that varied in the level of air pollution. The definition of air pollution levels has been changed for this analysis. In the earlier paper the levels were the sum of 7 estimates spread over 11 years. Some individuals over these 11 years changed their areas of residence and air pollution exposure. To minimise any effect of changing from one level of air pollution to another on respiratory experience, only those who had lived throughout the first 11 years of their lives in the same area, or had moved to areas of the same level of pollution were taken. This has resulted in the loss of 419 men and women for whom all the other required information was available. The final sample used for the main analysis was 2,088 men and women for whom complete information on childhood chest illness, air pollution, occupation of father, smoking at 20 and 25, and respiratory symptoms at 20 and 25 was available. This group is
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/18/2018 12:56:39 AM
Methods
238
K iernan/C olley/D ouglas/R eid
Table /. Effect of losses and exclusions on factor levels Factors
Childhood chest illness (0-2 years), % High level of pollution (2-11 years), % Manual working class2, % Regular smokers at 20, % Cough day or night at 20, %
Information obtained at earliest age
at 20 years
at 20 and 25 years
25.11 48.5 62.3 44.6 10.0
24.8 47.0 57.2 44.6 10.0
25.1 46.3 60.5 43.6 9.9
1 E.g. 25.1% of children in the sample known at 0-2 years had chest illness and this compares with 24.8% of those contacted at 20 years and 25.1% of those with complete information on all the factors, at age 20 and 25 years. 2 Based on occupation of father when survey members were aged eleven years, or if no information at this age, on the occupation of father at the time of birth of the survey member.
Results Smoking In the earlier study, smoking was found to be particularly relevant to the prevalence of cough. Table II compares the prevalence rates according to the history of smoking behaviour either in those who did not change their habit or gave up or took up smoking between the ages of 20 and 25 years. In all groups, many who reported cough during the day or at night in the winter at one age did not report it at the other. Nevertheless, the highest rates are seen in those smoking at both ages; and, among them, the pre valence rose significantly (p
