Associations between adolescent smoking habits and exercise, particularly participation in sports and lung function were studied.
All students aged 13–19 yrs in Nord-Trøndelag County, Norway, 1995–1997, were invited to join a cross-sectional study. Information on smoking habits and exercise was obtained by self-administered questionnaire. Spirometry was performed in accordance with American Thoracic Society standards.
Of the 6,811 students (aged 13–18 yrs, without asthma), 2,993 (44%) reported never-smoking, and 1,342 (20%) reported current smoking (10% daily). Frequency of physical exercise was inversely associated with smoking, but participants in individual sports with lesser endurance, especially body-building and fighting sports, were more likely to be daily smokers than nonparticipants. Both daily (53%) and occasional smokers (43%) were more likely to have quit sports than never-smokers (26%). Never-smokers showed a positive dose-response between physical exercise and lung function (forced vital capacity and forced expiratory volume in one second, adjusted for age and height). No similar significant association was observed in daily smokers.
These data suggest that smoking habits in different sports should be considered when promoting physical activity as smoking prevention, and sports organizations should include smoking prevention programmes. Adolescents with better lung function may self-select into sports; this possibility needs to be studied in a longitudinal design.
This study was supported by the Norwegian Research Council, the Norwegian State Dept of Social and Health Affairs, the County of Nord-Trøndelag, Norway and ASTRA ZENECA AS.
Low physical activity is associated with increased morbidity and mortality in adulthood 1, 2. An underlying premise for promotion of physical activity in youth is that it may persist through adulthood and may also reduce the risk for initiating unhealthy habits such as smoking 3.
Several cross-sectional studies have reported that physically active adolescents and those who participate in sports are less likely to be regular smokers 4–7 compared to sedentary youths. Encouraging participation in sports has been widely recommended for smoking cessation programmes 4, 8, 9. Aaron et al. 8 found female adolescents less likely to initiate smoking if they were more physically active or fit, but male adolescents showed a trend (not significant) towards being more active individuals and more likely to initiate cigarette use. Smoking habits, by different types of sports have received little attention.
Some studies report a positive association between physical activity and physical fitness and lung capacity 10–13, while others do not 14. Whether physical activity increases lung function remains uncertain. Adult-smoking sportsmen exhibit some degree of lung-function impairment compared to nonsmoking sportsmen, but still have better lung function than smokers who are not sportsmen 15. Young swimmers have been found to have better lung function than sedentary youths, as have participants of other sport activities 10, 16, 17. An impact of training on lung growth in young swimmers has been suggested 17, but others have found no effect of physical activity on lung growth 18, 19. Smoking in youth has been found to increase respiratory symptoms and decrease growth 4, 20, 21.
In this paper the cross-sectional association between smoking habit and physical exercise, participation in different types of sports, and lung function in adolescents is reported.
Material and methods
The population is from the YOUNG-HUNT study, the youth part of the large Nord-Trøndelag Health Study (HUNT), conducted from 1995–1997 in Nord-Trøndelag County, Norway. All students, aged 13–19 yrs, in the county were invited to join the study. A self-administered questionnaire was completed during one school hour, in a setting with no opportunity to look at other's papers. The questionnaire did not include participants' names. Subjects were linked to the questionnaire by a bar code of the 11-digit personal number with which all Norwegians are registered at birth. Each student put their completed questionnaire in a blank envelope and sealed it. Project nurses collected the envelopes.
Questions used in this paper are shown in the Appendix. Exercise of higher intensity, defined as being short of breath and/or sweaty, was assessed by using questions from the World Health Organization cross-national survey of health behaviour in schoolchildren 22 and was divided into three groups: “exercising 4 days a week or more”, “exercising 2 to 3 days a week” and “exercising 1 day a week or less”. Sports were classified as individual sports or team sports. Because sports questions do not capture individual level of effort, sports were classified based on expected average intensity as follows: “Individual sports with higher endurance ” including cross country skiing, cycling and running; “Individual sports with lesser endurance” including slalom skiing, horseback riding and gymnastics; “Swimming”; “Body-building and fight sports” including body-building, weightlifting, boxing, wrestling, Judo, Taekwondo and similar sports; and “Team sports” including football/soccer, handball, basketball and volleyball. In the text, physical exercise includes both organized and nonorganized activities unless otherwise specified. Specific sports questions refer to organized sport.
Current smokers were defined as those who answered “yes” to ever having tried smoking (at least one cigarette) and in addition answered “yes, I smoke daily” (daily smokers) or “yes, I smoke occasionally, but not daily” (occasional smokers) to the question: “Do you smoke now?” Smokers were compared to those who answered “no” to ever having tried smoking one cigarette (never-smokers). Passive smoking was defined as exposure to smoking at home by mothers, fathers or siblings.
Within a month after completing the questionnaire, the students underwent a clinical examination, performed in schools, including spirometry and measurement of height and weight. Spirometry was performed by specially trained nurses, in accordance with American Thoracic Society (ATS) standards 23, using computerized pneumotachograph (Jaeger Masterscope, software version 4.15, Jaeger Inc., Wurtzberg, Germany). The acceptability of the spirometry results was assessed both during testing and data analysis, and included review of the computerized ATS error codes reported from the Masterscopes, and visual inspection of volume/time and flow/volume graphics. Achieving end-of-test acceptability was confirmed either from the computerized error code regarding flow plateaus or visual inspection of spirometry displays during testing.
Forced vital capacity (FVC), forced expiratory volume in one second (FEV1), forced mid-expiratory flow (FEF50) and FEV1 per cent in relation to the maximal FVC (FEV1%FVC) were registered. FVC was defined as whichever was largest of either forced expiratory or forced inspiratory vital capacity from technically acceptable curves. Standing height was registered without shoes with standardized metre measures.
All 1,005 students reporting a history of asthma were excluded from these analyses.
Each student signed a written consent to participate in the study. Parents of students aged <16 yrs also gave written consent. The study was approved by the Regional Medicine Ethical Research Committee and the Norwegian Data Inspectorate Board.
Comparisons of age at onset of smoking and leaving active sports were made by unpaired t-tests, and comparisons of categorical variables by Chi-squared analysis. Comparisons between daily smokers and those who had never tried smoking were performed using logistic regression with age, exposure to passive smoking and physical exercise in the model. Amount of exercise was adjusted for when comparing different types of sports activities. Separate models were used for both females and males and for the different types of sports.
Analyses of lung function were performed by linear regression models with FVC, FEV1, FEV1%FVC and FEF50 as dependent variables. Because of heteroscedasticity, logarithmic (ln) transformation of lung function (Y) was used to fit model assumptions. Separate models were made for daily smokers and for those who had never tried smoking, adjusted for age, standing height, weight, passive smoking, physical activity, rhinitis and acute bronchitis with cough in both males and females. Analyses of variance were used for comparisons between mean values of lung-function measures by different levels of exercise with the same variables, described earlier, as covariates. Estimates and confidence intervals (CI) are expressed as per cent differences from the logarithmic scale, with those who reported lowest level of exercise (⩽1 day a week) as reference. Comparisons of height within each level of physical exercise were carried out at all ages (13–18 yrs) using one way analysis of variance. Females and males were analysed separately, 95% CI are shown.
Eighty-three per cent (8,305) completed both the questionnaire and spirometry. Included in these analyses were 6,811 students without asthma, aged 13–18 yrs, (50% males).
As shown in table 1, 2,993 students (45% of males and 43% of females) had never tried smoking and 1,342 students (18% of males and 22% of females) reported current smoking (10% daily) (table 1⇓). Daily smoking increased with age (p<0.001), and was marginally more common in girls (p=0.09). In daily smokers, mean age when smoking began was 13.9 yrs in both males and females. Overall, 52% males and 53% females reported being exposed to passive smoking at home.
Physical exercise and sports participation
Most students (68%) exercised at least 2 days a week outside school hours (table 2⇓). Males exercised more than females (p<0.001), and participated more in sports competitions (p<0.001). Compared to the younger age group (13–15 yrs), both males and females in the older age group (16–18 yrs) exercised less (p<0.001) and joined in less sports competitions (p<0.001). Compared to females, more males participated or had participated in “individual sports of higher endurance” (p<0.001) and “body-building and fight sports” (p<0.001), while females participated more in “individual sports with lesser endurance” (p<0.001) (table 2⇓). Overall, 31% males and 34% females reported having previously participated in sports but not currently.
Physical exercise and smoking
Current smoking was more prevalent in students with lower levels of physical exercise (fig. 1⇓). It was more common in “body-building and fighting sports”, “swimming”, and “individual sports of lesser endurance” compared to “team sports” and “individual sports of higher endurance”, especially in females (table 3⇓). More daily smokers (52% of males, 61% of females) and occasional smokers (43% of males, 39% of females) had dropped out of active sports compared to never-smokers (26% males, 27% females), p<0.001. Mean age for quitting sports was similar in daily smokers 15 yrs for both males and females and never-smokers (males 15 yrs, females 14 yrs). About 35% had started smoking before they quit sports; another 30% started smoking and quit sports in the same year.
Daily smokers reported less physical exercise and less participation in sports competitions than never-smokers (table 4⇓). Both males and females currently participating in team sports and sports competitions (regardless of what type of sport) were less likely to be daily smokers than those who did not. In males, daily smoking was positively associated with both current and previous participation in body-building and fight sports and with present participation in swimming. Females who had previously participated in these types of sports and lesser endurance sports were more often daily smokers than those who had not participated. Few daily-smoking females however, were participating in body-building/fight sports and swimming, (table 3⇓).
Lung function and physical exercise
A significant difference in mean height by exercise status was seen only between the highest and lowest level of exercise at age 13 yrs (mean difference 2.2 cm, p=0.01). Lung function was adjusted for age and height.
In never-smoking males and females there was a stepwise increase in levels of FVC and FEV1 with increasing level of exercise (fig. 2⇓), with a statistically significant and clinical relevant mean difference between the highest and lowest level of exercise. Mean differences were for FVC: males 195 mL, p<0.001, females 122 mL, p<0.001; and for FEV1: males 154 mL, p<0.001, females 119 mL, p<0.001. The same dose-response relationship was observed in daily smokers, but this trend was not statistically significant (fig. 2⇓). Daily-smoking males with medium level of exercise had significantly larger FEV1 compared to the lowest level of exercise. No significant associations were found between FEF50 and FEV1%FVC and frequency of exercise in never-smokers or daily smokers.
Both never-smoking males and females who participated in sports competitions (regardless of type of sport) had larger FVC, FEV1 and FEF50 compared to those who did not participate (adjusted for frequency of exercise). Mean differences in males: FVC 95 mL, p=0.01, FEV1 113 mL, p<0.001, FEF50 155 mL, p=0.03; and in females: FVC 65 mL, p=0.04, FEV1 70 mL, p=0.01, FEF50 156 mL, p=0.02. No such differences were found in daily smokers.
Lung function and different sports
Never-smoking males and females who presently participated in team sports had marginally higher FEF50 (mean difference: males 174 mL, p=0.05, females 15 mL, p=0.06) than those who did not. Females participating in team sports also had larger FEV1 (mean difference 80 mL, p=0.03) compared to nonparticipants. Daily-smoking males who presently participated in low-endurance sports had significantly lower FEV1 (p=0.05) and FEF50 (p=0.04) than nonparticipants. No other differences were found for other types of sports in never-smokers or daily smokers.
Never-smoking males and females who remained active in sports had better lung capacity compared to those who were no longer active. Mean difference in males: FVC: 135 mL, p<0.001, FEV1: 125 mL, p<0.001, and in females FVC: 69 mL, p=0.01, FEV1: 63 mL, p=0.01. No such difference was seen in daily smokers.
There were no important differences in the results when these analyses were repeated separately for the age groups 13–15 yrs and 16–18 yrs (data not shown).
In this study of a large general population of adolescents, with a high participation rate and carefully supervised spirometry testing, higher levels of physical exercise were associated with less daily smoking overall. But daily smoking was more freciated with less frequent participation in competitions. It is not clear whether these differences by type of sport are because smoking is perceived as less likely to impair performance or reflects peer influence, which plays an important part in adolescent smoking 4, 24, 25.
Difference in smoking habits among different types of sports has previously received little attention. If confirmed elsewhere, it suggests that type of sport should be considered when sports are recommended for smoking prevention or cessation. The importance of focusing on smoking prevention within sports organizations 26 is further supported by the observation that the age of smoking initiation either preceded or coincided with the age of quitting sports.
Whether physical exercise actually leads to better lung capacity cannot be determined in a cross-sectional study. In a prospective study, very young female competitive swimmers were found to increase their vital capacity and total lung capacity during a year of training 17, suggesting that the larger lung volumes in swimmers may be due to impact of training on lung growth. No impact of training on lung growth has been found in older swimmers or other youth 18, 19. Being good in sports may depend on a better lung capacity, representing self-selection of adolescents who both join and continue with sports. Self-selection is a plausible explanation why never-smoking students who had discontinued sports had a lower lung capacity compared to never-smokers who were still active.
In this (data not shown) and other studies 20, 21, 27, daily smokers with a light smoke burden were found to have a larger lung capacity (FVC) than nonsmokers. If daily smokers with good lung function self-select into physical activity, this would tend to diminish the observed “effects” of exercise. This possibility is supported by the finding that, among less active students, daily smokers had a larger FVC than never-smokers, while no significant difference in lung capacity was found between daily and never-smokers who were more physically active. Few daily smokers exercised at high levels, however, and the fact that the absent significant association reflects the small number of daily-smoking athletes, cannot be excluded.
In this student population frequency of physical exercise, not type of sport, was associated with better lung capacity; this has been reported in some studies of athletes, particularly in swimmers 10, 12, 16, 17. In the present study, swimmers did not have better lung capacity than nonswimmers, but there were few swimmers and none competed at a high level. The independent association between participation in sports competitions and larger lung function suggests that intensity of physical activity may be associated with larger lung function.
In spite of the computerized ATS error code warnings during testing, and careful assessment of the quality of the flow/volume curves by the nurses, a number of students did not meet the 1987 ATS criteria as judged by the ATS error-code messages from the Jaeger Masterscope. Because the ATS criteria in adolescents has been reported to be difficult 23–31, and excluding those not achieving ATS recommendations might exclude smokers, all students were included in the analysis. However, analyses were repeated excluding students not meeting the different Jaeger Masterscope ATS error codes with no significant differences in associations.
Rhinitis or bronchitis with cough was adjusted for in the regression analyses, as this might have had an impact on lung function, and all students who reported ever having asthma were excluded. The inclusion of adolescents reporting asthma in the analysis did not change the results (data not shown). Larger FEF50 and FEV1 values in never-smoking participants in team sports may reflect less airway dysfunction (not diagnosed as asthma) in students participating in these sports.
In self-reported physical exercise there is always a question of validity. The exercise questions used were chosen because they have been extensively used to describe and compare large groups in previously published population studies in many countries, including Norway 32, 33. The questions did not ask about seasonal exercise, but the majority of the adolescents who reported physical exercise were engaged in both winter and summer activities (data not shown). Substituting days per week with hours per week of exercise or using a combination of the two did not change the results (data not shown).
Smoking habits are also potentially subject to biased reporting. Confidentiality and consistency with other studies of adolescent smoking habits support the validity and generalizability of the results 4, 32. The long dark winters of northern Norway may impact exercise preferences in youths, but do not appear to have impacted smoking patterns.
More frequent physical exercise was associated with less daily smoking, but daily smoking was more common in youths performing some individual sports. Smokers quit sports more often than never-smokers. A significant dose/response relationship between lung capacity (FVC and FEV1) and levels of physical exercise was observed in never-smokers, but not in daily smokers.
These data provide indirect support for the promotion of physical activity as a smoking prevention strategy, and suggest that sports organizations should include smoking-prevention programmes. Initiation of sports programmes before the average age of smoking initiation (13 yrs in this study) could potentially yield long-term benefits in pulmonary function. Self-selection of adolescents with better lung function into sports is also suggested by these data, but causality will need to be studied prospectively.
APPENDIX: Questions and alternative answers used in this presentation from the YOUNG-HUNT study
The authors gratefully acknowledge the contributions of B. Dillan, G. Bratberg Ross, A. Enes, I. D. Holbø and the project nurses.
↵ For editorial comments see page 1.
- Received January 8, 2001.
- Accepted June 29, 2001.
- © ERS Journals Ltd