Physical fitness is inversely associated with the risk of total and cardiovascular disease mortality(6,10,11,14,17,18,19,24-26). One study showed that men assessed to be physically unfit in each of two examinations several years apart had higher mortality rates than men who were fit at both exams or who improved their fitness over time(4). Most studies that have reported longitudinal change in physical fitness have been relatively small and not population-based(1,7,21); there are few available data for women, and none that we are aware of for blacks.
Symptom-limited maximal graded exercise treadmill testing was performed in a biracial (black and white) population of men and women, ages 18-30 yr old, at the baseline examination of the Cardiovascular Risk Factors in Young Adults(CARDIA) study in 1985-1986. Evaluation of these data showed that age was unrelated to treadmill test duration in three of the four race-gender groups and was positively related to a submaximal measure of performance in three of the four race-gender groups (22). Treadmill testing was repeated at the year 7 follow-up examination in 1992-1993. We report here on the longitudinal change in treadmill test performance over this 7-yr time period.
Study population. The study population was a subgroup of the 4,086 participants who underwent the year 7 follow-up examination for the CARDIA study in 1992-1993, representing 80% of the study population of 5,115 young adults who were recruited for the baseline examination in 1985-1986. CARDIA subjects were recruited from four geographical locations (Birmingham, AL; Chicago, IL; Minneapolis, MN; and Oakland, CA). Recruitment efforts were successful in achieving the aim of study population that was approximately balanced by age (45% aged 18-24, 55% aged 25-30), race (52% black, 48% white), sex (46% men, 54% women), and education (40% having completed ≤12 yr of education, 60% having completed >12 yr of education). Further details of the recruitment process and description of the study population have been published elsewhere (12,15). Informed consent was obtained from all subjects and IRB approval was obtained.
Examination of the treadmill test data subsequent to the year 7 exam showed that the mean test duration was substantially higher in the Minneapolis center(729 s in Minneapolis vs 538 s in the other three centers combined) and that 84 Minneapolis participants completed the 18-min exercise test protocol versus a total of 3 participants in the other centers. A total of 11 participants across all four centers completed the same exercise test protocol at baseline, 7 yr earlier. Although no protocol deviation could be documented, the consensus of a committee of study investigators and consultants with expertise in fitness assessment was that Minneapolis participants had probably been allowed to hold onto the treadmill railing (a protocol deviation) resulting in longer treadmill tests. This made assessment of longitudinal change in treadmill test performance of Minneapolis participants (N = 1,139) invalid. We also performed age-race-gender adjusted regression analysis of the association of variables known to be associated with fitness (HDL cholesterol, physical activity score, and body mass index (BMI) with exercise test duration and time to heart rate 130. There were only small differences in the association of these factors to treadmill test performance in Minneapolis versus the other centers; for example, a 1 U (kg·m-2) increase in BMI was associated with a 16.6-s decrease in treadmill test duration in Minneapolis, versus a 12.0-s decrease at all other centers combined. The CARDIA study steering committee therefore decided that although Minneapolis data were not valid for estimation of absolute performance level, participants were validly ranked and therefore the data could be used in the study of the cross-sectional associations of other measures with treadmill test performance.
Of the 2,947 non-Minneapolis year 7 exam participants, exclusions from analysis were performed hierarchically as follows: 446 because they did not have a treadmill test(s) at either the baseline and/or year 7 exam(s), 351 because either the baseline and/or year 7 treadmill test was terminated for a reason other than fatigue or shortness of breath, 45 because a heart rate of 130 beats·min-1 was not achieved on the baseline and/or year 7 exam(s), 140 because of medical conditions noted in a pretest questionnaire that may have impacted negatively on exercise test performance, and 3 because of a change in maximum heart rate of greater than 30 beats·min-1 and a change in the maximum rating of perceived exertion of greater than 4, leaving 1,962 participants (67% of non-Minneapolis year 7 exam participants) for analysis.
Graded exercise treadmill test protocol. The exercise test protocol was designed to assess maximal, symptom-limited performance and has been described elsewhere (22). Tests were performed after an overnight fast. Study participants with acute illness or injuries that might affect treadmill test performance and those taking cardiovascular medications were excluded from the treadmill test. Participants were instructed not to smoke on the morning of the examination. The test consisted of up to nine 2-min stages (up to 18 min total) of progressively increasing difficulty. The estimated rate of energy expenditure for the completion of each of the nine stages was reported in metabolic equivalents (MET; 1 MET is the rate of energy expenditure equal to oxygen consumption of 3.5 mL·kg-1·min-1). Because many exercise tests were not terminated exactly at the completion of a 2-min stage, the estimated maximal exercise capacity in MET was determined by linear interpolation, as described previously (22). The test protocol, including the estimated METs at the completion of each stage, is shown inTable 1.
Heart rate, blood pressure, and electrocardiogram were obtained on each subject at rest, at the end of each stage, at maximum exercise, and at 3 min after exercise completion. The Borg rating of perceived exertion (RPE) scale, ranging from 6 (very light) to 20 (very hard), was obtained near the end of each stage and at the end of exercise (5).
Assessment of other variables. Height and weight were measured with the subject lightly clothed without shoes (8). Body mass index was calculated as weight/height (kg·m-2)(8). Physical activity score was assessed by interviewer-administered questionnaire, which assessed the amount of time spent in 13 different activities (12 leisure, 1 occupational) during the past year (16,23). Cigarette smoking status (never, former, current) was assessed by the responses to the following questions from a self-administered questionnaire: “Have you ever smoked cigarettes regularly for at least 3 months? By `regularly' we mean at least five cigarettes per week, almost every week.” An additional question defined current use: “Do you still smoke regularly?”
Statistical methods. The Statistical Analysis System (SAS) was used for all statistical analyses, including the determination of means, standard deviation, standard error of the mean, t-tests, analysis of variance, and multiple linear regression (20). All analyses were race-gender specific. The workload 130 (WL130), or exercise test duration to a heart rate of 130 beats·min-1, was determined by linear interpolation between heart rates recorded at the end of adjacent test stages (22). t-Tests were used to assess differences between means, and multiple linear regression was used to assess the association of baseline and change in weight, physical activity, and smoking to test duration and WL130, adjusted for potential confounders such as age, center effects, and education. Before multivariate modeling, age-adjusted linear regression models were used to assess, separately, the relationship of baseline weight, baseline physical activity score, 7-yr weight change, and 7-yr change in physical activity score, all treated as continuous variables, to change in test duration and in WL130. Similarly, we examined the association between treadmill test performance and baseline smoking status(current, former, nonsmoker) and major patterns of longitudinal use(persistent current, former, or nonsmoker (i.e., the same category at both exams) and quitter (current use at baseline, former user at year 7)). A longitudinal categorization of education status, not shown in the presentation of results, was entered into the multivariate model as follows: ≤12 yr at both exams; between 12-15.9 yr at both exams, 16+ years at both exams; ≤12 yr at baseline, >12 yr at year 7, between 12 and 15.9 yr at baseline, 16+ years at year 7.
Mean 7-yr change in exercise test variables by race-gender. The mean exercise test duration declined significantly in all race-gender groups, with the declines ranging from 41 s (7.0%) in white women to 95 s (13.6%) in black men (Table 2). Although exercise test duration is a good measure of fitness, the interpretation of test duration is limited by the reliance on symptom limitation by fatigue or shortness of breath as the end points, measures which were subject to variability due to participant's perception of these end points and to the intensity of coaching by the test supervisor. Physiological assessment of maximal effort (i.e., pulmonary gas exchange or blood lactate level) was not performed.
Two indirect indicators of maximal effort were measured. The mean maximum heart rate decreased modestly but significantly over 7 yr in all race-gender groups, ranging from 0.9 beats·min-1 in white men to 3.7 beats·min-1 in black women. As maximum heart rate decreases with age, it is unclear whether the changes observed reflect true physiological effects of aging, small changes in effort, or both. The mean rating of perceived exertion changed slightly but significantly in whites, but it did not change significantly in blacks. In summary, these two indicators suggested that symptom-limited efforts were reasonably comparable although probably not identical at the two examinations.
Because of the limitation noted above in using symptom-limited test duration as a measure of fitness, we also used the duration of exercise needed to reach a fixed heart rate, specifically 130 beats·min-1, which was achieved by almost all participants, as another measure of fitness. The workload 130 (WL130), or amount of time it takes to reach a heart rate of 130 beats·min-1, is a measure of submaximal exercise test performance. In general, the longer the time to heart rate 130, the more fit the individual. The mean WL130 declined significantly in all race-gender groups, with the declines ranging from 14 s (6.1%) in white women to 60 s(16.9%) in black men.
Weight, change in weight, and change in exercise test performance. The mean 7-yr weight change in the study population was substantial, ranging from a 4.8-kg increase in white women to an 8.7-kg increase in black women (Table 2). Baseline body weight was unassociated with the change in exercise test duration for all race-gender groups and was significantly associated with change in time to heart rate 130 in white women only (positive association, P = 0.03). Change in body weight had a strong inverse association with both change in exercise test duration and change in WL130 (P < 0.001 in all race-gender groups) (Fig. 1). The mean test duration decreased 4.4% in men and 1.7% in women maintaining weight within 5 pounds, whereas it decreased 16.5% in men and 21.2% in women gaining 20 pounds or more. A small cell size (N = 17) may have contributed to the slightly larger decrease in mean test duration among black men losing 5 or more pounds than in those maintaining weight with 5 pounds. The mean WL130 decreased 9.1% in men and decreased 1.9% in women maintaining weight within 5 pounds, whereas it decreased 21.0% in men and 23.0% in women gaining 20 pounds or more.
Physical activity score, change in physical activity score, and change in exercise test performance. The mean physical activity score decreased over the 7-yr period in all race-gender groups(Table 2). The decreases ranged from 9.3% of the baseline level in black men to 25.3% in white women. Baseline physical activity score was unassociated with the change in exercise test duration and was significantly associated with change in WL130 in black women and white women only (inverse association, P < 0.05). Change in physical activity score was associated with change in exercise test duration and with change in WL130 in all race-gender groups except for black men (Fig. 2). The changes in exercise test performance between physical activity strata were generally more modest than those for weight change strata. The mean test duration decreased 12.6% in men and 10.4% in women in the lowest tertile of physical activity score change (tertile 3, largest decrease in physical activity score) and decreased 6.9% in men and 5.6% in women in the highest tertile of physical activity score change (tertile cutoff points were-165 and 92 for black men, -198 and 18 for white men, -130 and 32 for black women, and -73 and 8 for white women).
Smoking and change in exercise test performance. The percentage of current smokers decreased during the 7-yr period in all groups except for black men (Table 2). Baseline smoking status was associated with change in test duration among black men (P = 0.04) and with change in WL130 among black men (P = 0.02) and black women(P = 0.01).
Analysis that took into consideration both the baseline and year 7 smoking status was also performed. Over the 7-yr period, 59.7% of participants remained nonsmokers, 7.8% former smokers, and 17.3% current smokers, whereas 4.9% quit smoking and 1.6% initiated smoking (0.6% were unknown category and 8.1% were in all other categories). The longitudinal pattern of smoking was not significantly associated with change in either test duration or WL130 in any of the race-gender specific models.
Multivariate analyses. Change in weight and change in physical activity score were strongly associated with change in exercise test duration in each of the race-gender models (Table 3). Baseline smoking status was associated with test duration overall in black men(P = 0.03) and black women (P = 0.05), with current smokers having significantly greater decreases than nonsmokers among blacks of both genders. Age was associated with test duration in white men only(inversely). Education (not shown) was unassociated with test duration change. There were significant examination center effects. The Chicago center had significantly smaller decreases in test duration in all race-gender groups. The relatively smaller 7-yr change in maximum heart rate in Chicago participants (data not shown) suggests that these participants performed relatively harder on their year 7 exam treadmill tests than their counterparts in the other centers.
Changes in weight were also strongly associated with change in WL130 in all race-gender groups (Table 4). Physical activity score change was associated with WL130 change in whites only. Age was associated with change in WL130 in white men only (inverse association). Baseline smoking status was associated with WL130 change in black women, with current smokers having significantly greater decreases than nonsmokers. Examination center effects were evident only for white women. In contrast with the test duration change analysis, the white women from the Chicago center had a greater decrease in WL130 than women from the other centers.
Change in fitness of 25- to 30-yr olds between 1985-1986 and 1992-1993. There were 1,055 members of the study group who were 25-30 yr old in 1985-1986 and 716 members of the study group who were 25-30 yr old in 1992-1993. The mean test duration and time to heart rate 130 were substantially lower in the 1992-1993 group except among black women, who experienced virtually no difference in these measurements(Table 5). The lack of significant differences in maximum heart rate and in maximum RPE between the two examinations suggests that the similar levels of effort were achieved in the symptom-limited tests. Weight was considerably higher among 25-30-yr-olds in 1992-1993 than in 25- to 30-yr olds in 1985-1986, average weight gain ranging from 3.9 kg in black women to 5.8 kg in white women. Physical activity score was lower in 1992-1993 for all groups except for black women, significantly so for white men and white women. The prevalence of current smoking was lower in the 1992-1993 group for all race-gender groups except white women, with greater decreases in blacks than in whites.
The main finding from this study is that fitness, as measured by either symptom-limited or submaximal performance on a graded exercise treadmill test, declined in all race-gender groups. There was heterogeneity among the race-gender groups in the degree of decline, with black men exhibiting the largest and white women the smallest decline.
The change in fitness had a strong inverse association with weight change and a somewhat weaker direct association with change in physical activity. Weight loss or small changes in weight was associated with gains in or relatively small declines in fitness, whereas weight gain of 20 pounds or more was associated with fitness declines of approximately 20%. The constellation of fitness change, weight change, and physical activity change findings suggests that decreasing physical activity may have resulted in both decreased physical fitness and increased adiposity and weight. It is not possible to infer with certainty that decreasing physical activity was the underlying cause of these findings, in part because the technical difficulty in measuring physical activity results in relatively low correlations between physical activity score and measures of body mass and fitness(23).
The disparity between the cross-sectional findings at the baseline examination regarding age and exercise test performance noted in the introduction with the 7-yr longitudinal change findings has been noted by others and underlines the importance of longitudinal measurements in assessing the effects of age (9). The rates of fitness change, as estimated by the symptom-limited test results, were approximately 0.3 mL·kg-1·min-1·yr-1 in white women, 0.4 mL·kg-1·min-1·yr-1 in black women and white men, and 0.7 mL·kg-1·min-1·yr-1 in black men. The rate of fitness decline observed in men in the CARDIA cohort is in the range found in other studies, which have generally been performed in small groups of highly selected individuals (7). A small study of 35 female physical education teachers demonstrated an average decline of 0.44 mL·kg-1·min-1·yr-1 over a 21-yr period between tests (2). In the CARDIA study, changes in blacks were greater than changes in whites; as noted earlier, we do not know of any other reports of longitudinal fitness changes in blacks.
These data also suggest that, with the exception of black women, a secular decline in physical fitness has occurred in young adults ages 25-30 yr old in this study during the 7 yr between 1985-1986 and 1992-1993 concurrent with an increase in weight and a decrease in physical activity. The stability of fitness in black women was associated with stability in their mean physical activity score and a decrease in the prevalence of smoking. The decline of fitness in the other three race-gender groups was substantial, representing 10% or more of the 1985-1986 level whether assessed by test duration or by WL130.
In summary, both symptom-limited and submaximal exercise treadmill fitness measurements declined in all four race-gender groups, in association with weight gain and decreased physical activity. Physical training can increase aerobic fitness over the short term and decrease the rate of decline of fitness over the long term (3,13). With relatively recent evidence from longitudinal studies that decreasing physical fitness and/or physical fitness have adverse health effects, the CARDIA study findings should be viewed with concern. Minimal fitness change occurred in those maintaining body weight within 5 pounds. However, most participants decreased physical activity and gained weight. As a result, there was a substantial loss of fitness in this cohort.
This study was supported by NHLBI contracts NO1-HC-84047, NO1-HC-84048, NO1-HC-84049, and NO1-HC-84050. The authors acknowledge Marianne Sadler for computer programming and analysis and Diana Holt for technical assistance with manuscript preparation.
1. Andersen, L. B. and J. Haraldsdóttir. Tracking of cardiovascular disease risk factors including maximal oxygen uptake and physical activity from late teenage to adulthood: an 8-year follow-up study.J. Intern. Med.
2. Åstrand, I., P.-O. Åstrand, I. Hallbäck, and A. Kilbom. Reduction in maximal oxygen uptake with age. J. Appl. Physiol.
3. Åstrand, P.-O. and K. Rodahl. Textbook of Work Physiology: Physiological Bases of Exercise
. 3rd Ed., Chap. 12. New York: McGraw-Hill, 1986, pp.523-582.
4. Blair, S. N., H. W. Kohl III, C. E. Barlow, R. S. Paffenbarger, Jr., L. W. Gibbons, and C. A. Macera. Changes in physical fitness
and all-cause mortality: a prospective study of healthy and unhealthy men. J. A. M. A.
5. Borg, G. and H. Linderholm. Perceived exertion and pulse rate during graded exercise in various age groups. Acta Med. Scand.
6. Bruce, R. A., K. F. Hossack, T. A. Derouen, and V. Hofer. Enhanced risk assessment for primary coronary heart disease events by maximal exercise testing: 10 years' experience of Seattle Heart Watch. J. Am. Coll. Cardiol.
7. Buskirk, E. R. and J. L. Hodgson. Age and aerobic power: the rate of change in men and women. Fed. Proc.
8. Cutter, G. R., G. L. Burke, A. R. Dyer, G. D. Friedman, et al. Cardiovascular risk factors in young adults: the CARDIA baseline monograph. Controlled Clin. Trials
9. Dehn, M. M. and R. A. Bruce. Longitudinal variations in maximal oxygen intake with age and activity. J. Appl. Physiol.
10. Ekelund, L.-G., W. L. Haskell, J. L. Johnson, F. S. Whaley, M. H. Criqui, and D. S. Sheps. Physical fitness
as a predictor of cardiovascular mortality in asymptomatic North American men: the Lipid Research Clinics mortality follow-up study. N. Engl. J. Med.
11. Erikssen, J. Physical fitness
and coronary heart disease morbidity and mortality: a prospective study in apparently healthy, middleaged men. Acta Med. Scand. Suppl.
12. Friedman, G. D., G. R. Cutter, R. P. Donahue, et al. CARDIA: study design, recruitment, and some characteristics of the examined subjects. J. Clin. Epidemiol.
13. Hagberg, J. M. Effect of training on the decline of˙VO2max
with aging. Fed. Proc.
14. Hein, H. O., P. Suadicani, and F. Gyntelberg. Physical fitness
or physical activity as a predictor of ischaemic heart disease? a 17-year follow-up in the Copenhagen Male Study. J. Intern. Med.
15. Hughes, G. H., G. Cutter, R. Donahue, et al. Recruitment in the coronary artery disease risk development in young adults(CARDIA) study. Controlled Clin. Trials
16. Jacobs, D. R., L. P. Hahn, W. L. Haskell, P. Pirie, S. Sidney. Validity and reliability of short physical activity history: Cardia and the Minnesota Heart Health Program. J. Cardiopulmon. Rehabil.
17. Lie, H., R. Mundal, and J. Erikssen. Coronary risk factors and incidence of coronary death in relation to physical fitness
: seven-year follow-up study of middle-aged and elderly men. Eur. Heart J.
18. Peters, R. K., L. D. Cady, Jr., D. P. Bischoff, L. Bernstein, and M. C. Pike Physical fitness
and subsequent myocardial infarction in healthy workers. J. A. M. A.
19. Sandvik, L., J. Erikssen, E. Thaulow, G. Erikssen, R. Mundal, and K. Rodahl. Physical fitness
as a predictor of mortality among healthy, middle-aged Norwegian men. N. Engl. J. Med.
20. SAS Institute, Inc. SAS User's Guide: Basics, Version 5
. Cary, NC: SAS Institute, Inc., 1985.
21. Shea, S., C. E. Basch, B. Gutin, et al. The rate of increase in blood pressure in children 5 years of age is related to changes in aerobic fitness and body mass index. Pediatrics
22. Sidney, S., W. L. Haskell, R. Crow, et al. Symptom-limited graded treadmill exercise testings in young adults in the CARDIA study. Med. Sci. Sports Exerc.
23. Sidney, S., D. R. Jacobs, Jr., W. L. Haskell, et al. Comparison of two methods of assessing physical activity in the coronary artery risk development in young adults (CARDIA) study. Am. J. Epidemiol.
24. Slattery, M. L. and D. R. Jacobs, Jr. Physical fitness
and cardiovascular disease mortality: the US railroad study. Am. J. Epidemiol.
25. Sobolski, J., M. Kornitzer, G. De Backer, et al. Protection against ischemic heart disease rather than physical activity?Am. J. Epidemiol.
26. Wilhelmsen, L., J. Bjure, B. Ekström-Jodal, et al. Nine years' follow-up of a maximal exercise test in a random population sample of middle-aged men. Cardiology
68(Suppl. 2):1-8, 1981.