Tobacco smoking has been associated with low cardiovascular fitness (7,20,30) and an impaired heart rate response to exercise (i.e., chronotropic incompetence) (9,16,31), which are both important predictors of all-cause mortality (2,17,32).
The first shortcoming in almost all previous studies on smoking in relationship to cardiovascular fitness and heart rate response to exercise is that they have been conducted in studies with a cross-sectional design. With a cross-sectional design, however, nothing can be said about possible changes over time. It seems logical to assume that the adverse effects of smoking become stronger with increasing age, as it probably takes some time before detrimental effects of smoking on the cardiovascular system become measurable. The second shortcoming in previous studies on smoking and cardiovascular fitness and heart rate response to exercise is that only very few studies have included women (26,31) or subjects younger than 18 yr (4,21). From a public health perspective, it is interesting to know to what extent young healthy smokers already show a reduced cardiovascular fitness and an impaired heart rate response to exercise. From this same perspective, it is also interesting to know to what extent the effects of smoking are reversible.
In the Amsterdam Growth And Health Longitudinal Study (AGAHLS), smoking behavior, cardiovascular fitness, and heart rate response to exercise were measured one to nine times between ages 13 and 36, in a cohort of 298 men and 331 women. This gave us the opportunity to study smoking in relation to cardiovascular fitness and heart rate response to exercise longitudinally. The reversibility of smoking was investigated by studying smoking status (i.e., never smoker, short-term ex-smoker, long-term ex-smoker, and current smoker) in relation to cardiovascular fitness and heart rate response to exercise.
Subjects and design.
All subjects were participants in the AGAHLS that started in 1977 with 298 boys and 334 girls from the first and second grade of two secondary schools in Amsterdam and Purmerend, The Netherlands (13). Measurements were performed at calendar age 13, 14, 15, 16, 21, 27, 29, 32, and 36. Although the number of measurements on each subject varied between one and nine measurements, all pupils were taken into account in the longitudinal analyses. Dropout analyses were performed to investigate whether or not tobacco consumption, fitness parameters, lifestyle parameters, and body composition parameters at all measurements predicted dropout at the next measurement. No dropout effects were found between age 13 and 16 (14). From the age of 27 onward, single dropout effects were found with HRrest, body weight, Slopemax, and tobacco consumption. These dropout effects were not consistent over time, were not found in both sexes and, could be due to multiple testing. The AGAHLS was approved by the Medical Ethical Committee of the VU University Medical Center. Written informed consent was provided by the parents when the subjects were 13–16 yr of age. At age 21–36, subjects provided written consent themselves. Subjects who reported to use [beta]-blockers (N = 5) were excluded from the analyses in that particular year of measurement.
At each measurement year, subjects filled out a questionnaire about their use of cigarettes, own-rolled tobacco, cigars/cigarillos, and pipe tobacco. From this data, tobacco smoking was expressed in total gram of tobacco smoked per week (1 cigarette = 1 gram, 1 package of own-rolled tobacco = 40 g, 1 cigar/cigarillo = 3 gram, 1 package of pipe tobacco = 50 g). Subjects were counted as smokers if they smoked minimally 1 g of tobacco (i.e., 1 cigarette) per day. Each measurement year, subjects were categorized as nonsmokers (<1 g tobacco per day), light smokers (1 <= g tobacco per day < 10), or moderate to heavy smokers (>=10 g tobacco per day). The cut-off value of 10 g of tobacco per day was chosen arbitrarily. At the age of 36, subjects filled out an extensive questionnaire about tobacco smoking in the past in order to obtain a rough indication of lifetime tobacco smoking. Lifetime tobacco smoking was expressed in pack-years by multiplying smoking duration (starting at first onset) with daily tobacco consumption. One pack-year is defined as one packet, or 20 g tobacco, smoked each day over a course of 1 yr (27). At the age of 36, ex-smokers were defined as subjects smoking less than 1 g of tobacco per day with pack-years > 0. We discriminated between short-term ex-smokers and long-term ex-smokers based on the reported years of abstinence. Ex-smokers who had quit smoking for more than 4 yr were defined as long-term ex-smokers, whereas ex-smokers who had quit for 4 yr or less were defined as short-term ex-smokers. The cut-off value of 4 yr was chosen because there were 4 yr between the measurements at age 32 and the measurements at age 36. Never smokers at the age of 36 were defined as subjects who had never smoked at least 7 g of tobacco per day.
Pretest smoking condition.
All participants were requested not to smoke on the measurement day. Because all participants were not left unattended during the hour preceding the treadmill running, and none of the participants reported using a nicotine patch on the measurement day, we were able to rule out short-term effects of smoking.
Cardiovascular fitness heart rate response to exercise.
Three different measures of cardiovascular fitness were used. The first of these measures was absolute maximal oxygen uptake ([latin capital V with dot above]O2max), expressed in liters per minute, which was measured directly with the Ergoanalyzer (Jaeger, Bunnik, The Netherlands) during a maximal treadmill running test (15). This measurement of gas exchange was validated against the classical method of collecting expired air in Douglas bags and the Scholander technique to analyze the carbon dioxide and oxygen content of the expired air (15). Absolute [latin capital V with dot above]O2max is considered as a valid estimate of true cardiovascular fitness, as it is highly correlated with cardiac output (1,29). The slope of the treadmill was increased every 2 min until exhaustion while the running speed was kept constant at 8 km·h-1. During the treadmill running test, the ECG was monitored telemetrically and continuously and recorded at the end of each minute from the final 15 R-R intervals. The second measure of cardiovascular fitness was the maximum slope reached during the treadmill running test (Slopemax). The third measure of cardiovascular fitness was resting heart rate (HRrest), measured in a seated position before the treadmill running test.
Heart rate response to exercise (i.e., chronotropic response) was measured during the treadmill running test by: 1) heart rate at a treadmill slope of 5% (HRsubmax), 2) maximum heart rate (HRmax), and 3) heart rate reserve (HRR), the change in heart rate from rest to maximal exercise calculated by HRmax - HRrest.
To study smoking in relation to [latin capital V with dot above]O2max, Slopemax, HRrest, HRsubmax, HRmax, and HRR longitudinally, generalized estimating equation (GEE) analyses (35) were carried out with the Statistical Package for Interactive Data Analysis (8). In all the longitudinal analyses, light smokers and moderate to heavy smokers were compared with nonsmokers. With GEE analysis, the relation between two longitudinally measured variables can be studied using all longitudinal data simultaneously and correcting for differences in time interval between measurement periods and for within person correlations caused by the repeated measurement on each subject (33). To investigate whether or not the relationships between smoking and cardiovascular fitness and heart rate response to exercise were time dependent, we studied the interaction between smoking and time of measurement. Furthermore, with data from all time points included in the longitudinal model, the difference in cardiovascular fitness between smoking groups at age 36 was predicted. This predicted difference was based on the longitudinal model to which interactions with time were added (assuming a linear interaction). The longitudinal analysis to predict differences between smoking groups at age 36 differs from a cross-sectional analysis, because data of all time points were used instead of only at one time point.
A simplification of the longitudinal model is presented below:
t indicates that smoking status was time dependent; [beta]0 = intercept; [beta]1 = difference in outcome variable between light smokers and nonsmokers at t = 0; [beta]2 = difference in outcome variable between moderate/heavy smokers and nonsmokers at t = 0; [beta]3 = change in outcome variable per year; [beta]4 = change in the difference in outcome variable between light smokers and nonsmokers per year; [beta]5 = change in the difference in outcome variable between moderate to heavy smokers and nonsmokers per year; smoking1 = light tobacco smoking; smoking2 = moderate to heavy tobacco smoking; and time = time (or year) of measurement and varies between t = 0 (age 13 - 1) and t = 24 (age 36).
The expected difference in [latin capital V with dot above]O2max between light smokers and nonsmokers at the age of 36 (i.e., t = 24) was calculated by [beta]1 + (24 × [beta]4). Similarly, the expected difference in [latin capital V with dot above]O2max between moderate to heavy smokers and nonsmokers was calculated by [beta]2 + (24 × [beta]5).
To investigate the reversibility of smoking effects, linear regression analyses were used in which smoking status was categorized into never smokers, short-term ex-smokers, long-term ex-smokers, or current smokers and related to cardiovascular fitness and heart rate response to exercise in 36-yr-old men and women. Never smokers were used as the reference group.
All relationships were studied using an adjusted and unadjusted model. Covariables in the adjusted model were body height, physical activity, and body weight. In the relationship between smoking and HRsubmax, Slopemax was added to the covariables of the adjusted model in order to take differences in exercise capacity into account. Daily physical activity was estimated by a standardized activity interview and expressed in a total weighted metabolic activity score (METs·wk-1) (22). The activity interview that was used at age 36 differed slightly from the interview that was used in earlier years in that it covered more activities.
Table 1 presents the number of subjects that participated in the AGAHLS and the prevalence of smoking at each measurement year. Table 2 presents the characteristics of the participants at age 13 and 36. At the age of 36, 85 males and 89 females reported to have ever smoked tobacco (pack-year > 0). In these ever-smokers, the median number of pack-years and range was 7.43 [0.02–45.00] in males and 5.71 [0.01–37.70] in females.
Smoking and cardiovascular fitness.
[latin capital V with dot above]O2max. In both men and women, a negative relationship was found between moderate to heavy smoking and [latin capital V with dot above]O2max, which was stronger in men than in women (Table 3). The [beta] regression coefficient in males from the adjusted longitudinal model indicates that the [latin capital V with dot above]O2max of moderate to heavy smokers was overall (i.e. when data from all ages were included) 0.19 L·min-1 lower than in nonsmokers. The [latin capital V with dot above]O2max of moderate to heavy smoking females was overall 0.08 L·min-1 lower than in nonsmokers. In addition, the significant negative interaction between smoking and time of measurement in men (Table 3) indicates that the negative association between smoking and [latin capital V with dot above]O2max became stronger with increasing age.
Slopemax. In men, moderate to heavy smoking was negatively related to Slopemax (Table 3). The [beta] regression coefficient in men from the adjusted longitudinal model indicates that moderate to heavy smokers reached an overall Slopemax that was 0.87% lower than in nonsmokers. Furthermore, a significant negative interaction between smoking and time of measurement was found in both light and moderate to heavy smokers. This means that the negative association between smoking and Slopemax became stronger with increasing age in all male smokers. In contrast to the results in men, smoking was not related to Slopemax in women (Table 3).
HRrest. Only in men, a negative relationship was found between light smoking and HRrest, whereas no relationship was found between moderate to heavy smoking and HRrest (Table 3). Light tobacco smoking males reached an overall HRrest that was 2.73 bpm lower than in nonsmokers. No significant interaction between smoking and time was found.
Smoking and heart rate response to exercise.
HRsubmax. In both men and women, smoking was negatively related to HRsubmax, which was more pronounced in moderate to heavy smokers than in light smokers and more pronounced in men than in women (Table 4). In men, light smokers reached an overall HRsubmax that was 3.36 bpm lower than in nonsmokers, whereas moderate to heavy smokers reached an overall HRsubmax that was 6.04 bpm lower than in nonsmokers. No significant interaction between smoking and time was found.
HRmax. A negative longitudinal relationship was found between moderate to heavy smoking and HRmax, which was quite similar in men and women. The overall HRmax in moderate to heavy smoking men was 2.47 bpm lower than in nonsmokers versus 2.07 bpm in moderate to heavy smoking females. In men, a negative interaction between moderate to heavy smoking and time of measurement was found (Table 5), indicating that the negative association between moderate to heavy smoking and HRmax in males became stronger with increasing age.
HRR. In both sexes, no significant longitudinal association was found between smoking and HRR (Table 5). The association between smoking and HRR was similar at all ages, as no interaction was found between smoking and time of measurement.
Predicted differences at age 36.
The predicted difference between smokers and nonsmokers in cardiovascular fitness at the age of 36 is presented in Table 3. The expected difference between smokers and nonsmokers in heart rate response to exercise at the age of 36 is presented in Tables 4 and 5. Most striking are the expected reductions in [latin capital V with dot above]O2max (i.e., 5.7% and 8.7%) and Slopemax (13.7% and 15.4%) in male smokers as compared with male nonsmokers.
Reversibility of smoking effects.
Table 6 presents the results from the cross-sectional linear regression analyses between smoking status (i.e., long-term ex-smokers, short-term ex-smokers, and current smokers vs never smokers) and cardiovascular fitness and heart rate response to exercise at age 36. In 36-yr-old males, current smokers (N = 37) reached a lower [latin capital V with dot above]O2max, Slopemax, HRsubmax, HRmax, and HRR than never smokers. Both short-term ex-smokers (N = 25) and long-term ex-smokers (N = 23), on the contrary, reached a [latin capital V with dot above]O2max, Slopemax, and HRR similar as in never smokers. However, with regard to HRmax and HRsubmax, short-term ex-smokers mimicked the results found in current smokers, whereas long-term ex-smokers mimicked the results of never smokers. In 36-yr-old females, no relationship was found between smoking status and cardiovascular fitness and heart rate response to exercise, except for HRsubmax. Current smokers (N = 35) reached a lower HRsubmax than never smokers, whereas both short-term ex-smokers (N = 20) and long-term ex-smokers (N = 34) did not differ from nonsmokers (Table 6).
The results from our longitudinal analyses indicate that smoking is negatively related to cardiovascular fitness and heart rate response to exercise in men and women between age 13 and 36. The strongest associations were found with [latin capital V with dot above]O2max and Slopemax, as indicators of cardiovascular fitness, and with HRsubmax, as an indictor of heart rate response to exercise. The associations were stronger in moderate to heavy smokers than in light smokers and stronger in males than in females. In line with the expectations, the negative relationship between smoking and [latin capital V with dot above]O2max, Slopemax, and HRmax became stronger over time in males. Cross-sectional analyses suggest that the adverse effects of smoking were reversible in 36-yr-old males.
Smoking and cardiovascular fitness.
When we only consider [latin capital V with dot above]O2max and Slopemax as indicators of cardiovascular fitness, we found a negative (longitudinal) relationship between smoking and cardiovascular fitness, which is in line with prior studies (3,5,6,21,28). The maximum level of exhaustion was similar in smokers and nonsmokers, as no differences were found in RER at the end of the treadmill running test (i.e., mean RER in smokers was 111.4 SD 8.52 and in nonsmokers 112.4 SD 9.19). In contrast to prior studies (9,19,31), HRrest was lower in light smoking males than in nonsmoking males. An explanation might be that only long-term effects of smoking were studied in the present study, as the short-term effects could be ruled out. Although it is commonly known that HRrest rises directly after smoking due to increased sympathetic drive and reduced vagal modulation (11,19,23,25), less is known about the long-term effects of smoking on HRrest. Hayano et al. (11) found a blunted heart rate response to postural change in heavy smokers. In the present study, HRrest was measured before the maximal exercise test in a sitting position. The sitting position and psychological stress due to the coming treadmill running test, might have increased HRrest in nonsmokers to a higher extent than in light smokers, due to a blunted heart rate response in light smokers. In other studies where HRrest was measured before the maximal exercise test, the results were contradictive. Several studies found no association between long-term smoking and HRrest (3,6,23,26), whereas others found a positive (9,31) or a negative association (16,30). Unfortunately, this cannot explain why the lower HRrest was not found in moderate to heavy smokers.
Smoking and heart rate response to exercise.
Our study shows that the reduced heart rate response to exercise in smokers found by others (9,16,31) was already present in young and healthy men and women. Although the lower HRrest in smokers can explain part of the lower HRsubmax, the differences in HRrest between smokers and nonsmokers were smaller than in HRsubmax. This indicates that the lower heart rate response to exercise cannot be fully explained by the lower HRrest, especially not in moderate to heavy smokers.
Reversibility of smoking effects.
In accordance with our findings, Sandvik et al. (30) found the adverse effects of smoking on physical fitness to be reversible in men. Hirsch et al. (12) even found abstinence from smoking for one day, to increase [latin capital V with dot above]O2max. Hashizume et al. (10), on the other hand, found a prolonged exercise time after 6 or 7 d of abstinence from smoking but no change in [latin capital V with dot above]O2max.
Very few studies have described the association between smoking and cardiovascular fitness in women (4). In the present study, the association between smoking and cardiovascular fitness in women was weaker than in men, similar as in Boreham et al. (4). Prior studies on smoking and HRrest in women are contradictive. Whereas Sidney et al. (31) found no difference in HRrest between female smokers and nonsmokers, Lauer et al. (16) found female current smokers to reach a lower HRrest than female never smokers. The differences between men and women in the present study, regarding the association between smoking and fitness parameters, might be explained by the lower median number of pack-years in women. Nevertheless, only very small differences were found between men and women in the association between smoking and heart rate response to exercise, which is in line with other studies (16,31,32). In contrast to the males, no reversibility of smoking effects was found in females with the exception of HRsubmax. Additional reversibility analyses with and without correction for physical activity showed similar results, suggesting that differences in physical activity level between males and females could not explain differences in reversibility. Furthermore, no significant interaction was found between smoking status and physical activity level, suggesting that the reversibility of smoking effects was not dependent on physical activity level. Other explanations, such as gender differences in metabolism or weight management, do not seem likely. Although several studies reported weight gain after smoking cessation, most studies reported no differences between males and females (34). The most plausible explanation for the observed gender differences seems to be that the associations between smoking and cardiovascular fitness parameters in females were simply not strong enough to show any reversibility.
Mechanism smoking and heart rate response to exercise.
There is no agreement about the mechanisms by which smoking reduces heart rate response to exercise. Smoking increases the sympathetic drive to the heart and muscle blood vessels (24) and reduces vagal cardiac control (11). Narkiewicz et al. (24) found that baroreflexes, which respond to increasing blood pressure as a result of smoking, play an important role in reducing the sympathetic nerve traffic and limiting the increase in heart rate in young healthy subjects. However, when baroreflex function is impaired, smoking increases sympathetic tone, which might lead to a reduced heart rate response to exercise via downregulation of [beta]-adrenergic receptors (18).
Limitations of the study.
Due to the limitations of cross-sectional analyses, the results on the reversibility of smoking effects should be interpreted with caution. The results from the cross-sectional analyses do not prove that the effects of smoking are reversible but only suggest its reversibility. However, the finding that the results in short-term ex-smokers were mostly in between the results of long-term ex-smokers and current smokers, further strengthens the idea of reversibility. Caution should also be taken when interpreting the expected differences at age 36, which are based on the interaction between smoking and time of measurement and the initial differences in cardiovascular fitness and heart rate response to exercise between smoking groups. The most important reasons are that the initial differences between smoking groups are based on a few subjects and on the assumption that the interaction between smoking and time of measurement is linear. Finally, converting tobacco consumption from a continuous variable (grams per day) into a discrete measure (i.e., nonsmokers, light smokers, or moderate/heavy smokers) brings about some disadvantages. The most important disadvantage is that smokers consuming 1 g of tobacco per day are considered to be similar to those consuming 9 g of tobacco per day, whereas smokers consuming 9 g of tobacco per day are considered to be different from those who consume 10 g of tobacco per day. However, by using a discrete measure for tobacco consumption, we were able to show that some of the associations between smoking and cardiovascular fitness were nonlinear.
Even in young healthy men and women, smoking is negatively related to cardiovascular fitness and heart rate response to exercise, which are both predictors of all-cause mortality. In men, the negative relationship between smoking and cardiovascular fitness and heart rate response to exercise becomes stronger over time.
This study was financially supported by the Dutch Heart Foundation (grant 76051-79051), the Dutch Prevention Fund (grants 28-189a, 28-1106, and 28-1106-1), the Dutch Ministry of Well Being and Public Health (grant 90-170), the Dairy Foundation on Nutrition and Health, the Dutch Olympic Committee/Netherlands Sports Federation, Heineken Inc., and the Dutch Scientific Board Smoking and Health.
1. Astrand, P. O. Children and adolescents: performance, measurement, education. In: Pediatric Work Physiology, J. Coudert and E. Van Praagh (Eds.). Paris: Masson, 1992, pp. 3-7.
2. Blair S. N., J. B. Kampert, and H. W. Kohl III. Influences of cardiorespiratory fitness and other precursors on cardiovascular disease and all-cause mortality in men and women. JAMA
276: 205-210, 1996.
3. Bolinder, G., A. Norén, J. Wahren, and U. de Faire. Long-term use of smokeless tobacco and physical performance in middle-aged men. Eur. J. Clin. Invest.
27: 427-433, 1997.
4. Boreham, C., J. Twisk, W. van Mechelen, M. Savage, J. Strain, and G. Cran. Relationships between the development of biological risk factors for coronary heart disease and lifestyle parameters during adolescence: The Northern Ireland Young Hearts Project. Public Health
113: 7-12, 1999.
5. Brown, T. E., W. S. Myles, and C. L. Allen. The relationship between aerobic fitness and certain cardiovascular risk factors. Aviat. Space Environ. Med.
54: 543-547, 1983.
6. Chatterjee, S., S. K. Dey, and S. K. Nag. Maximum oxygen uptake capacity of smokers of different age groups. Jpn. J. Physiol.
37: 837-850, 1987.
7. Cooper, K. H., G. O. Gey, and R. A. Bottenberg. Effects of cigarette smoking on endurance performance. JAMA
203: 189-192, 1968.
8. Gebski, V., O. Leung, D. Mcneil, and D. Lunn. SPIDA User Manual, Version 6. NSW Australia: Macquarie University, 1992, pp. 45-72.
9. Gordon, D. J., A. S. Leon, L. G. Ekelund, et al. Smoking, physical activity, and other predictors of endurance and heart rate response to exercise in asymptomatic hypercholesterolemic men: the Lipid Research Clinics Coronary Primary Prevention Trial. Am. J. Epidemiol.
125: 587-600, 1987.
10. Hashizume K., K. Yamaji, Y. Kusaka, and K. Kawahara. Effects of abstinence from cigarette smoking on the cardiorespiratory capacity. Med. Sci. Sports Exerc.
32: 386-391, 2000.
11. Hayano, J., M. Yamada, Y. Sakakibara, et al. Short- and long-term effects of cigarette smoking on heart rate variability. Am. J. Cardiol.
65: 84-88, 1990.
12. Hirsch, G. L., D. Y. Sue, K. Wasserman, T. E. Robinson, and J. E. Hansen. Immediate effects of cigarette smoking on cardiorespiratory responses to exercise. J. Appl. Physiol.
58: 1975-1981, 1985.
13. Kemper, H. C. G. (Ed.). Growth, health and fitness of teenagers: longitudinal research in international perspective. Med. and Sport Science,
Vol. 20. Basel: Karger, 1985, pp. 12-34.
14. Kemper, H. C. G. (Ed.). The Amsterdam Growth Study: a longitudinal analysis of health, fitness and lifestyle. In: HK Sports Sci. Monograph Series,
Vol. 6. Champaign, IL: Human Kinetics, 1995, pp. 8-11.
15. Kemper, H. C. G., and R. Verschuur. Longitudinal study of maximal aerobic power in teenagers. Ann. Hum. Biol.
14: 435-444, 1987.
16. Lauer, M. S., F. J. Pashkow, and M. G. Larson. Association of cigarette smoking with chronotropic incompetence and prognosis in the Framingham Heart Study. Circulation
96: 897-903, 1997.
17. Lauer, M. S., G. S. Francis, P. M. Okin, F. J. Pashkow, C. E. Snader, and T. H. Marwick. Impaired chronotropic response to exercise stress testing as a predictor of mortality. JAMA
281: 524-529, 1999.
18. Laustiola, K. E., R. Lassila, J. Kaprio, and M. Koskenvuo. Decreased beta-adrenergic receptor density and catecholamine response in male cigarette smokers: a study of monozygotic twin pairs discordant for smoking. Circulation
78: 1234-1240, 1988.
19. Lucini, D., F. Bertocchi, A. Malliani, and M. Pagani. A controlled study of the autonomic changes produced by habitual cigarette smoking in healthy subjects. Cardiovasc. Res.
31: 633-639, 1996.
20. Marti, B., T. Abelin, C. E. Minder, and J. P. Vader. Smoking, alcohol consumption, and endurance capacity: an analysis of 6,500 19-year-old conscripts and 4,100 joggers. Prev. Med.
17: 79-92, 1988.
21. Montoye, H. J., R. Gayle, and M. Higgins. Smoking habits, alcohol consumption and maximal oxygen uptake. Med. Sci. Sports Exerc.
12: 316-321, 1980.
22. Montoye H. J., H. C. G. Kemper, W. H. M. Saris, and R. A. Washburn. Measuring Physical Activity and Energy Expenditure
, appendix P. Champaign, IL: Human Kinetics, 1996, pp. 183-184.
23. Morton A. R., E. V. Holmik. The effects of cigarette smoking on maximal oxygen consumption and selected physiological responses of elite team sportsmen. Eur. J. Appl. Physiol.
53: 348-352, 1985.
24. Narkiewicz, K., P. J. Van de Borne, M. Hausberg, et al. Cigarette smoking increases sympathetic outflow in humans. Circulation
98: 528-534, 1998.
25. Niedermaier, O. N., M. L. Smith, L. A. Beightol, et al. Influence of cigarette smoking on human autonomic function. Circulation
88: 562-571, 1993.
26. Pederson, L. L., M. Poulin, N. M. Lefcoe, A. W. Donald, and J. S. Hill. Does cigarette smoking affect the fitness of young adults
? Rationale and protocol for future research. J. Sports Med. Phys. Fitness
32: 96-105, 1992.
27. Prignot, J. Quantification and chemical markers of tobacco-exposure. Eur. J. Respir. Dis.
70: 1-7, 1987.
28. Robbins, A. S., S. Y. Chao, V. P. Fonseca, M. R. Snedecor, and J. J. Knapik. Predictors of low physical fitness in a cohort of active-duty U.S. Air Force members. Am. J. Prev. Med.
20: 90-96, 2001.
29. Rowland, T. W., J. Staab, V. Unnithan, and S. Siconolfi. Maximal cardiac responses in prepubertal and adult males (Abstract). Med. Sci. Sports Exerc.
20: (Suppl.) S32, 1988.
30. Sandvik, L., G. Erikssen, and E. Thaulow. Long term effects of smoking on physical fitness and lung function: a longitudinal study of 1393 middle aged Norwegian men for seven years. Br. Med. J.
311: 715-718, 1995.
31. Sidney, S., B. Sternfeld, S. S. Gidding, et al. Cigarette smoking and submaximal exercise test duration in a biracial population of young adults
: the CARDIA study. Med. Sci. Sports Exerc.
25: 911-916, 1993.
32. Srivastava, R., E. H. Blackstone, and M. S. Lauer. Association of smoking with abnormal exercise heart rate responses and long-term prognosis in a healthy, population-based cohort. Am. J. Med.
109: 20-26, 2000.
33. Twisk, J. W., H. C. Kemper, D. J. Mellenbergh, and W. Van Mechelen. Factors influencing tracking of cholesterol and high-density lipoprotein: the Amsterdam Growth and Health Study. Prev. Med.
25: 355-364, 1996.
34. Williamson, D. F., J. Madans, R. F. Anda, J. C. Kleinman, G. A. Giovino, and T. Byers. Smoking cessation and severity of weight gain in a national cohort. N. Engl. J. Med.
324: 739-745, 1991.
35. Zeger, S. L., and K. Y. Liang. Longitudinal data analysis for discrete and continuous outcomes. Biometrics
42: 121-130, 1986.