Associations of Changes in Exercise Level With ... - CiteSeerX

Journal of Gerontology: MEDICAL SCIENCES 2006, Vol. 61A, No. 1, 97–102

Copyright 2006 by The Gerontological Society of America

Associations of Changes in Exercise Level With Subsequent Disability Among Seniors: A 16-Year Longitudinal Study David R. Berk, Helen B. Hubert, and James F. Fries Department of Medicine, Stanford University School of Medicine, California. Background. The effect of changes in physical exercise on progression of musculoskeletal disability in seniors has rarely been studied. Methods. We studied a prospective cohort annually from 1984 to 2000 using the Health Assessment Questionnaire Disability Index (HAQ-DI). The cohort included 549 participants, 73% men, with average end-of-study age of 74 years. At baseline and at the end of the study, participants were classified as ‘‘High’’ or ‘‘Low’’ vigorous exercisers using a cutpoint of 60 min/wk. Four groups were formed: ‘‘Sedentary’’ (LowfiLow; N ¼ 71), ‘‘Exercise Increasers’’ (LowfiHigh; N ¼ 27), ‘‘Exercise Decreasers’’ (HighfiLow; N ¼ 73), and ‘‘Exercisers’’ (HighfiHigh; N ¼ 378). The primary dependent variable was change in HAQ-DI score (scored 0–3) from 1984 to 2000. Multivariate statistical adjustments using analysis of covariance included age, gender, and changes in three risk factors, body mass index, smoking status, and number of comorbid conditions. Participants also prospectively provided reasons for exercise changes. Results. At baseline, Sedentary and Increasers averaged little exercise (16 and 22 exercise min/wk), whereas Exercisers and Decreasers averaged over 10 times more (285 and 212 exercise min/wk; p , .001). All groups had low initial HAQ-DI scores, ranging from 0.03 to 0.08. Increasers and Exercisers achieved the smallest increments in HAQ-DI score (0.17 and 0.11) over 16 years, whereas Decreasers and Sedentary fared more poorly (increments 0.27 and 0.37). Changes in HAQ-DI score for Increasers compared to Sedentary were significantly more favorable ( p , .05) even after multivariate statistical adjustment. Conclusions. Inactive participants who increased exercise achieved excellent end-of-study values with increments in disability similar to those participants who were more active throughout. These results suggest a beneficial effect of exercise, even when begun later in life, on postponement of disability.

W

ITH aging of the population, the number of seniors with disability will likely continue to grow. However, the disability associated with aging is not an inevitable constant (1). The Compression of Morbidity hypothesis suggests that adopting healthy lifestyles may postpone disability onset until the few years immediately preceding death, reducing lifetime disability (2). Longitudinal comparison of cumulative disability among seniors in three health-risk strata supports this hypothesis (3). Moreover, mortality followback analysis demonstrates lower end-of-life disability with lower health risks (4). Recently, the rate of disability development among seniors has declined faster than has mortality, demonstrating that compression of morbidity is possible nationally (1,3). A 13-year longitudinal study of runners initially aged 50– 72 years demonstrated that vigorous exercise in seniors is associated with disability-free life and disability postponement of 8–12 years (5). Over time, however, initially active seniors may become sedentary, and initially sedentary seniors may become active. To make informed recommendations about lifestyle modifications at older age, studying the impact of changes in health risk factors among seniors is important. Moreover, examining reasons for exercise changes should inform targeted interventions. Examining changes in lifestyle risks rather than baseline risks has important advantages. Associations of outcomes

with baseline risk status may be related to self-selection bias, whereas associations of outcomes with risk status changes more likely suggest a causal relationship. Statistical adjustment procedures seldom remove all self-selection bias, and unobserved factors still may be causally related to outcomes. Although many studies have examined associations between changes in physical activity and mortality (6–18), fewer have examined older populations (6,16,17,19). Furthermore, data are limited regarding the effect of changes in exercise and other risk factors on disability or health-related quality of life (10,13,19), particularly in seniors (19). A 16-year longitudinal analysis was performed using data from a comparative study of vigorous exercisers and population controls (5). We hypothesized that (i) among participants originally self-selecting for inactivity, those who increased exercise would have slower disability progression; and (ii) among participants originally self-selecting for exercise, those who decreased activity would have more rapid disability progression, associated with increased comorbidity. PARTICIPANTS AND METHODS

Study Sample In 1984, participants were recruited from the Fifty-Plus Runner’s Association including predominantly Californians 97

98

BERK ET AL.

(runners), and a Stanford, California community-based sample originally recruited as controls for the Lipid Research Clinics Study (controls) (20). Initially, 1311 runners club and 2181 Stanford community members were invited to participate. Baseline questionnaires and consent forms were sent to the 654 runners club and 568 community members who expressed interest and satisfied the inclusion criteria of age
50 years, high school education, and English as primary language; 538 runners and 423 controls (N ¼ 961) consented. Participants annually completed a mailed questionnaire assessing self-reported sociodemographic characteristics, health risk factors (including body mass index (BMI), smoking status, and exercise), medical conditions, and health care utilization. They also completed the Health Assessment Questionnaire assessing disability (HAQ-DI). Details of sample selection and data collection are described elsewhere (21).

Assessment of Disability The HAQ-DI contains 21 questions covering eight physical function domains, including rising, dressing and/ or grooming, hygiene, eating, walking, reach, grip, and activities and/or chores. Each dimension is scored 0–3, with 0 ¼ no difficulty, 1 ¼ some difficulty, 2 ¼ much difficulty, and 3 ¼ inability to perform. The HAQ-DI score (scored 0–3) is the average of scores across domains. A score of 0.125 is minimal disability, equal to some difficulty in one domain; a score of 0.375 represents inability in one domain or lesser difficulty in two or three domains. Thus, small differences in HAQ-DI score describe significant differences in functioning. The HAQ-DI is a widely used, rigorously validated index of disability (22,23). Formation of Exercise Groups Vigorous exercise was assessed as minutes per week (min/wk) for specific types of activities. Examples of vigorous exercise that were queried included running, swimming, cycling, dancing, brisk walking, and racket sports. Participants reported whether and why they made substantial changes (increase, decrease, stop, or start) in any type of exercise. We studied 549 participants followed from 1984 to 2000 (completers), including 335 runners and 214 controls. Individuals were classified as ‘‘High’’ or ‘‘Low’’ exercisers at baseline and at end of study according to whether they averaged more than 60 min/wk of vigorous exercise in 1984–1986 and 1998–2000; 3-year periods were used to reduce year-to-year variability. Four groups were formed based on exercise during these two periods: ‘‘Sedentary’’ (LowfiLow; N ¼ 71): 60 exercise min/wk during both periods; ‘‘Exercise Increasers’’ (LowfiHigh; N ¼ 27): 60 exercise min/wk at baseline, but .60 min/wk at end of study; ‘‘Exercise Decreasers’’ (HighfiLow; N ¼ 73): .60 exercise min/wk at baseline, but 60 min/wk at end of study; and ‘‘Exercisers’’ (HighfiHigh; N ¼ 378): .60 exercise min/wk during both periods. Of the remaining 412 participants who discontinued participation, 106 died and 306 represented attrition. Overall attrition rates were only 2% per year. Dropout rates were similar among baseline Low and High exercisers. However,

baseline Low exercisers were more likely than were High exercisers to die during follow-up (1.3% vs 0.5% annually, p , .001). For both groups, those participants who died or dropped out were significantly older than were the completers, and had higher baseline disability. Among baseline Low exercisers, those who died or dropped out had more comorbid conditions (0.25 vs 0.05 for completers, p , .01). However, there were no significant differences between withdrawals and completers in baseline exercise minutes for either group.

Statistical Analysis Differences between those participants who remained in their exercise group and those who changed were tested using analysis of variance to contrast variable means and Kruskal– Wallis chi-square tests to compare proportions. Adjusted change in HAQ-DI score was estimated using analysis of covariance controlling for age, gender, changes in BMI and smoking status (yes/no) from 1984 to 2000, and change in number of comorbid conditions from 1989 to 2000. Comorbid conditions that were considered include cardiovascular conditions, pulmonary conditions, cancer, gastrointestinal conditions, diabetes, neurologic conditions, musculoskeletal conditions, and other. Not all factors thought to influence disability, such as cognitive function, could be included in our analysis due to lack of data in early study years. Other possible covariables such as dietary factors were also not included for the same reasons. Interaction terms were not tested due to small numbers in several groups. Statistical significance was considered (p , .05) using a two-tailed test. RESULTS Table 1 describes participants at baseline and end of study. Compared to Sedentary, Increasers were more likely to be men and to be slightly more educated. Compared to Exercisers, Decreasers were more likely to be women, smokers, slightly older, and have a higher BMI. Comorbidity data were similar at baseline across groups. However, by 2000, Decreasers averaged more comorbid conditions than Exercisers (1.4 vs 1.1, p , .05). Decreasers compared to Exercisers, and Sedentary compared to Increasers more often had arthritic, neurological, and cardiovascular comorbidities. The Sedentary cohort had the most pulmonary comorbidities. At baseline, Sedentary and Increasers (the original sedentary groups) averaged well below the 60 exercise min/ wk cutoff (16 and 22 min/wk), whereas Exercisers and Decreasers (the original exercising groups) averaged well above this cutoff (285 and 212 min/wk). The baseline difference between Exercisers and Decreasers, however, was significant ( p , .001). At end of study, exercise minutes were 22 for Sedentary compared to 134 for Increasers (p , .01), and 261 for Exercisers compared to 24 for Decreasers ( p , .001). By end of study, Decreasers increased BMI more than did Exercisers (1.6 vs 1.1, p , .05; Table 2). The percentage of smokers remained low. The increase in number of comorbidities ranged from 1.0 to 1.4 and showed no statistically significant difference between Decreasers and Exercisers.

EXERCISE EFFECTS ON DISABILITY

Table 1. Participant Characteristics at Baseline and End of Study

99

Table 2. Changes in Participant Characteristics From Baseline

Exercise Groups

Exercise Groups

Sedentary

Increasers

Decreasers

Exercisers

Participant Characteristics

LowfiLow (N ¼ 71)

LowfiHigh (N ¼ 27)

HighfiLow (N ¼ 73)

HighfiHigh (N ¼ 378)

Mean age, 1984, y Mean age, 2000, y Men, % Education, mean y Exercise min/wk, 1984, mean Exercise min/wk, 1984, median Exercise min/wk, 2000, mean Exercise min/wk, 2000, median BMI, 1984, mean kg/m2 BMI, 2000, mean kg/m2 % Smokers, 1984 % Smokers, 2000 No. of comorbid conditions, 1989, mean No. of comorbid conditions, 2000, mean

59.8 (0.7) 75.6 (0.7) 49.3 15.9 (0.3)

59.6 (1.1) 75.5 (1.1) 85.2y 17.0 (0.4)*

59.2 (0.8) 75.2 (0.8) 63.0 16.5 (0.3)

57.5 (0.3)* 73.4 (0.3)* 78.0y 16.8 (0.1)

15.9 (4.3)

21.5 (7.0)

212.2 (19.5)

285.2 (9.4)z

160.0

240.0

0.0 21.5 (4.3) 0.0

0.0 133.5 (17.5)y 140.0

24.2 (4.5) 0.0

261.1 (10.7)z 220.0 y

24.4 (0.4)

24.9 (0.4)

23.8 (0.4)

22.9 (0.1)

25.2 (0.5) 12.7 7.0

25.4 (0.5) 7.4 0.0

25.5 (0.6) 12.3 4.1

24.0 (0.1)z 1.1z 1.6

0.06 (0.03)

0.00 (0.00)

0.08 (0.03)

0.06 (0.01)

1.5 (0.1)

1.2 (0.2)

1.4 (0.1)

1.1 (0.1)*

Note: Standard errors are shown in parentheses. Values of p used for testing differences between Sedentary versus Increasers and Exercisers versus Decreasers: *p , .05. y p , .01. z p , .001. BMI ¼ body mass index.

The Sedentary group maintained exercise between 16 and 23 min/wk during the baseline (1984–1986) and final (1998– 2000) periods (Figure 1). Exercisers maintained exercise between 233 and 327 minutes throughout, whereas the exercise minutes of the Decreasers declined more than 8-fold. Activity change data indicated that among Decreasers, 75% of swimmers and 64% of runners diminished levels of their specific exercise. Lack of time and/or interest and joint pain were the most common reasons for declining activity among Decreasers. Orthopedic problems and injuries were more frequently reasons for decreasing running or walking than for decreasing swimming or cycling. Runners also were most likely to change forms of exercise (20% did so). Increasers most frequently increased walking and cycling. One third of those who initially cycled and 29% of those who initially walked increased their exercise, whereas only 15% of runners in this group increased their overall exercise min/wk. Doctors’ recommendations and personal desires for better fitness most often motivated Increasers’ positive change in exercise. Activities that Increasers most often started were aerobics (19%), cycling (19%), and weight lifting (11%). No Increasers started running. Table 3 indicates that all groups averaged low HAQ-DI scores in 1984, ranging from 0.03 to 0.08 disability units on the 0–3 scale, well below a score of 0.125 that reflects some

Sedentary Change in:

Increasers

LowfiLow LowfiHigh (N ¼ 71) (N ¼ 27)

Exercise min/wk, 1984–2000, mean 5.1 (6.4) BMI, 1984–2000, mean kg/m2 0.9 (0.3) No. of comorbid conditions, 1989–2000, mean 1.4 (0.1) % Smokers, 1984–2000 5.6

Decreasers

Exercisers

HighfiLow (N ¼ 73)

HighfiHigh (N ¼ 378)

112.0 (17.2)* 185.3 (20.1) 23.7 (11.4)y 0.4 (0.4)

1.6 (0.3)

1.1 (0.1)*

1.2 (0.2)

1.3 (0.1)

1.0 (0.1)

7.4

8.2

0.5y

Note: Standard errors are shown in parentheses. Values of p for testing differences between Sedentary versus Increasers and Exercisers versus Decreasers: *p , .05. y p , .001. BMI ¼ body mass index.

difficulty in 1 of 8 domains of function evaluated by the HAQ-DI. However, the baseline HAQ-DI score in Decreasers already was significantly higher than in Exercisers (0.08 vs 0.03). Sedentary recorded the greatest change and poorest end-of-study HAQ-DI scores, increasing 0.37 disability units to 0.42, equivalent to more than some difficulty in 3 domains of daily function or inability to perform 1 domain evaluated by the HAQ-DI. Decreasers showed poor end-of-study scores, increasing 0.28 to 0.36 units. Increasers achieved very good end-of-study scores, increasing only 0.17 to 0.20 units. Exercisers demonstrated the least change and the best end-of-study scores, increasing 0.11 to 0.14 units. Disability for Decreasers approaches that for Sedentary over time, whereas disability levels of Exercisers and Increasers remained relatively low through-

Figure 1. Mean exercise over time. Exercisers maintained high levels of exercise throughout the 16 years, whereas Sedentary remained relatively inactive. On average, Increasers achieved greater than a 6-fold increase in exercise min/ wk, whereas Decreasers experienced a greater than 8-fold decrease.

BERK ET AL.

100

Table 3. HAQ-DI Score (Standard Error) at Baseline, End of Study, and Change Over Time Exercise Groups

HAQ-DI Score Disability, 1984 Disability, 2000 Change in disability, 1984–2000 Change in disability, 1984–2000, adjusteda

Sedentary

Increasers

Decreasers

Exercisers

LowfiLow (N ¼ 71)

LowfiHigh (N ¼ 27)

HighfiLow (N ¼ 73)

HighfiHigh (N ¼ 378)

0.05 (0.01) 0.42 (0.07)

0.03 (0.01) 0.20 (0.07)y

0.08 (0.02) 0.36 (0.06)

0.03 (0.00)z 0.14 (0.02)z

0.37 (0.06)

0.17 (0.07)*

0.27 (0.05)

0.11 (0.01)z

0.29 (0.06)

0.12 (0.08)*

0.20 (0.06)

0.12 (0.06)§

Note: aAdjusted for age and gender in 1984, change in body mass index and change in smoking status (yes/no) from 1984 through 2000, and change in number of comorbid conditions from 1989 through 2000. Values of p for testing differences between Sedentary versus Increasers and Exercisers versus Decreasers: *p , .05. y p , .01. z p , .001. § p ¼ .06.

out (Figure 2). Final disability scores were more associated with the current than with the initial exercise group. The HAQ-DI score change from 1984 to 2000 differed significantly between groups, with Sedentary worse than Increasers and Exercisers better than Decreasers. Differences between groups regarding the HAQ-DI score change from 1984 to 2000, our main outcome measure, persisted after adjustments for age, gender, and changes in three risk factors, BMI, smoking status, and number of comorbidities. In fact, adjusted change in HAQ-DI score was the same for Increasers and Exercisers.

DISCUSSION Because those persons older than 65 years compose the fastest growing segment of the population in many nations (24), methods to maximize disability-free life now have critical medical, social, and economic importance (1,25). Despite medical advances, primary prevention and health risk reduction will remain important approaches to health improvement, especially with increasingly sedentary, overweight populations (26–29). This longitudinal study supports the thesis that exercising at high levels, even late in life, delays disability development. Additionally, the results examining effects of exercise change suggest that increasing exercise in seniors may still have important health effects. Individuals with high exercise levels throughout our study period (Exercisers, HighfiHigh) had the smallest adjusted increase in disability, consistent with a beneficial effect of exercise. Those participants with little exercise throughout our study period (Sedentary, LowfiLow) had the greatest adjusted increase in disability, consistent with a negative effect of more sedentary lifestyles. This analysis focused on those individuals who initially self-selected for inactivity, but later increased exercise. These Increasers had little progression of disability, similar

Figure 2. Mean disability over time. At baseline, all groups had very low levels of disability. The Sedentary group experienced the poorest end-of-study Health Assessment Questionnaire Disability Index (HAQ-DI) score, increasing 0.37 to 0.42. Exercisers achieved the best end-of-study HAQ-DI score, increasing only 0.11 to 0.14. Increasers achieved good end-of-study HAQ-DI score, increasing 0.17 to 0.20. Decreasers experienced poor end-of-study HAQ-DI score, increasing 0.28 to 0.36.

to Exercisers who initially self-selected for exercise and exercised at high levels throughout (Exercisers, Highfi High). Despite the small number of Increasers, comparisons with Sedentary were statistically significant in both unadjusted and adjusted analyses. These findings are consistent with a beneficial effect of vigorous exercise on disability development, even when exercise starts later in life. Through 74 years of age, continuation or increase in vigorous activity from a sedentary level appears beneficial. Although this analysis focused on those individuals who increased exercise (Increasers, LowfiHigh), it is interesting to explore why individuals who decreased exercise (Decreasers, HighfiLow) had poor outcomes. Decreasers had progression of disability almost as rapid as individuals who were Sedentary throughout. Either disability developed because exercise was discontinued, or some individuals developed comorbid conditions that caused the exercise decrease; the latter seems likely considering the higher level of arthritic, neurological, and cardiovascular comorbidities among Decreasers and their reported reasons for exercise reduction. When change in number of comorbid conditions was included in adjustments, the difference in change in disability between Exercisers and Decreasers reached borderline significance ( p ¼ .056), suggesting that the development of comorbid conditions explained part but not all of the difference in disability progression between the two groups. There are limitations to all observational studies. Although analyses were adjusted for BMI and smoking status changes, Increasers and Decreasers may have changed in other ways that were not ascertained. Although our approach was designed to diminish self-selection bias, it is impossible to eliminate all such bias in nonrandomized studies given that, in this case, groups were categorized by self-selected changes in exercise. These analyses did not

EXERCISE EFFECTS ON DISABILITY

capture individuals who attempted but could not maintain increases in exercise. Accordingly, we considered covariates differing between groups, including age, gender, and change in BMI, smoking status, and number of comorbidities. Validation studies in runners and nonrunners found no evidence to suggest reporting biases in such data (21). Although our cohort is a selective group that is socioeconomically homogeneous and not representative of most elderly populations, it may provide advantages for examining health outcomes in relation to lifestyle modifications such as exercise. A generally healthy, well-educated, highly selected population may allow us to more easily isolate the impact of other risk factors such as exercise without the confounding effects of other factors such as socioeconomic status. Moreover, there may be biases that could result in underestimation of the magnitude of intergroup differences. For example, baseline Low exercisers more frequently died during follow-up than did baseline High exercisers. Many deaths were undoubtedly among individuals who had greater changes in disability prior to death. Their exclusion would cause underestimation of accrued disability in Sedentary individuals. Additionally, not all factors known to influence disability, such as cognitive function, could be included in adjustments due to incomplete available data. The loss of participants who were not known to have died also may introduce bias, although we feel that our attrition rate is small for a study of its duration (overall attrition rate ¼ 2% annually). Within both baseline High and Low exercise groups, later withdrawals had higher baseline disability but equivalent exercise levels compared to ongoing participants. By identifying reasons for activity changes, these results may help target efforts at exercise enhancement. Cycling and aerobics were most frequently initiated. Cycling and walking were most frequently increased. Runners were most likely to change types of exercise. Thus, efforts to promote starting, increasing, or maintaining vigorous exercise among seniors might be most successful when emphasizing walking, cycling, or aerobics. Although most studies of older adults have used crosssectional (not longitudinal) designs (30–34), poor health, fear of injury, and perceived decreased activity needs have been cited as principal barriers to exercise (30,32). Environmental factors (access to facilities, weather, lack of exercise companion) and program-based factors (inconvenience, expense, lack of time and/or interest) have been less frequently reported (30,32). Seniors’ reasons for maintaining or increasing activity include their desire for better health and doctors’ recommendations (30,34). Prospective studies examining what motivates seniors’ activity changes are limited, but suggest similar interpretations (34). Orthopedic problems and lack of time and/or interest were the most common reasons for exercise declines in this study, suggesting that injury prevention may promote exercise maintenance. Few seniors who decreased vigorous exercise reported changing their type of exercise. The poor outcomes of decreasing exercise suggest the importance of sustainable exercise forms, even after comorbidity and chronic illness. For example, cycling and swimming represent low impact alternatives for runners with joint

101

difficulties. Weight training allows individuals to focus on uninvolved muscles after injuries develop. Isometric exercise isolates muscle groups and allows exercise without movement. Indeed, exercise for elderly individuals with joint disease or injuries is clearly beneficial (35). Reports of loss of time or interest are greater here than in other studies (30–33), but suggest that exercise should be made more interesting and of greater priority for seniors. For example, a stationary bicycle by the television may make exercise interesting. Exercise clubs or neighborhood walking groups can maintain seniors’ interest with social components. Doctors’ recommendations to maintain or improve health most often motivated increases in exercise. However, young adults more frequently exercise to improve physical appearance (36). Given more medical visits in seniors, physicians have more opportunity to educate and motivate (33). Improving Medicare reimbursement for visits for exercise promotion would encourage more active roles for physicians (37). Our data suggest that increasing vigorous exercise decreases disability progression. More sophisticated analyses of time order issues (individual exercise change before or after disability) is beyond the scope of this report, and could help clarify interpretation of associations described. These findings strongly suggest, however, that addressing modifiable health risks even late in life may promote healthier aging.

ACKNOWLEDGMENTS This study was supported by grant R01 AG15815 from the National Institute on Aging to Stanford University. We thank Miho Bennett, PhD, for technical assistance. This study was presented in part in abstract form at the American College of Rheumatology Annual Scientific Meeting, October 27, 2003; Orlando, Florida. Address correspondence to James F. Fries, MD, Professor of Medicine, Immunology and Rheumatology, 1000 Welch Rd. #203, Stanford, CA 94305-5755. E-mail: [email protected]

REFERENCES 1. Freedman VA, Martin LG, Schoeni RF. Recent trends in disability and functioning among older adults in the United States: a systematic review. JAMA. 2002;288:3137–3146. 2. Fries JF. Aging, natural death, and the compression of morbidity. N Engl J Med. 1980;303:130–135. 3. Vita AJ, Terry RB, Hubert HB, Fries JF. Aging, health risks, and cumulative disability. N Engl J Med. 1998;338:1035–1041. 4. Hubert HB, Bloch DA, Oehlert JW, Fries JF. Lifestyle habits and compression of morbidity. J Gerontol A Biol Sci Med Sci. 2002;57A: M347–M351. 5. Wang BW, Ramey DR, Schettler JD, Hubert HB, Fries JF. Postponed development of disability in elderly runners: a 13-year longitudinal study. Arch Intern Med. 2002;162:2285–2294. 6. Bijnen FC, Feskens EJ, Caspersen CJ, Nagelkerke N, Mosterd WL, Kromhout D. Baseline and previous physical activity in relation to mortality in elderly men: the Zutphen Elderly Study. Am J Epidemiol. 1999;150:1289–1296. 7. Wannamethee SG, Shaper AG, Walker M. Changes in physical activity, mortality, and incidence of coronary heart disease in older men. Lancet. 1998;351:1603–1608.

102

BERK ET AL.

8. Lissner L, Bengtsson C, Bjorkelund C, Wedel H. Physical activity levels and changes in relation to longevity. A prospective study of Swedish women. Am J Epidemiol. 1996;143:54–62. 9. Sherman SE, D’Agostino RB, Silbershatz H, Kannel WB. Comparison of past versus recent physical activity in the prevention of premature death and coronary artery disease. Am Heart J. 1999;138(5 pt 1): 900–907. 10. Johansson S, Sundquist J. Change in lifestyle factors and their influence on health status and all-cause mortality. Int J Epidemiol. 1999;28: 1073–1080. 11. Paffenbarger RS Jr, Hyde RT, Wing AL, Lee IM, Jung DL, Kampert JB. The association of changes in physical-activity level and other lifestyle characteristics with mortality among men. N Engl J Med. 1993; 328:538–545. 12. Paffenbarger RS Jr, Hyde RT, Wing AL, Steinmetz CH. A natural history of athleticism and cardiovascular health. JAMA. 1984;252: 491–495. 13. Paffenbarger RS Jr, Kampert JB, Lee IM, Hyde RT, Leung RW, Wing AL. Changes in physical activity and other lifeway patterns influencing longevity. Med Sci Sports Exerc. 1994;26:857–865. 14. Blair SN, Kohl HW III, Barlow CE, Paffenbarger RS Jr, Gibbons LW, Macera CA. Changes in physical fitness and all-cause mortality. A prospective study of healthy and unhealthy men. JAMA. 1995;273: 1093–1098. 15. Kaplan GA, Strawbridge WJ, Cohen RD, Hungerford LR. Natural history of leisure-time physical activity and its correlates: associations with mortality from all causes and cardiovascular disease over 28 years. Am J Epidemiol. 1996;144:793–797. 16. Byberg L, Zethelius B, McKeigue PM, Lithell HO. Changes in physical activity are associated with changes in metabolic cardiovascular risk factors. Diabetologia. 2001;44:2134–2139. 17. Gregg EW, Cauley JA, Stone K, et al. Relationship of changes in physical activity and mortality among older women. JAMA. 2003;289: 2379–2386. 18. Schnohr P, Scharling H, Jensen JS. Changes in leisure-time physical activity and risk of death: an observational study of 7,000 men and women. Am J Epidemiol. 2003;158:639–644. 19. Leinonen R, Heikkinen E, Jylha M. Predictors of decline in selfassessments of health among older people—a 5-year longitudinal study. Soc Sci Med. 2001;52:1329–1341. 20. The Lipid Research Clinics Coronary Primary Prevention Trial results. I. Reduction in incidence of coronary heart disease. JAMA. 1984; 251:351–364. 21. Lane NE, Bloch DA, Wood PD, Fries JF. Aging, long-distance running, and the development of musculoskeletal disability. A controlled study. Am J Med. 1987;82:772–780.

22. Fries JF, Spitz P, Kraines RG, Holman HR. Measurement of patient outcome in arthritis. Arthritis Rheum. 1980;23:137–145. 23. Bruce B, Fries JF. The Stanford Health Assessment Questionnaire: a review of its history, issues, progress, and documentation. J Rheumatol. 2003;30:167–178. 24. Manton KG, Vaupel JW. Survival after the age of 80 in the United States, Sweden, France, England, and Japan. N Engl J Med. 1995;333: 1232–1235. 25. Rowe JW. Geriatrics, prevention, and the remodeling of Medicare. N Engl J Med. 1999;340:720–721. 26. Kuczmarski RJ, Flegal KM, Campbell SM, Johnson CL. Increasing prevalence of overweight among US adults. The National Health and Nutrition Examination Surveys, 1960 to 1991. JAMA. 1994;272: 205–211. 27. Stephens T. The demography of physical activity. In: Bouchard C SR, Stephens T, eds. Physical Activity, Fitness and Health. Champaign, IL: Human Kinetics Publishers; 1994. 28. DiPietro L. Physical activity in aging: changes in patterns and their relationship to health and function. J Gerontol A Biol Sci Med Sci. 2001;56 Spec No 2:13–22. 29. Flegal KM, Carroll MD, Ogden CL, Johnson CL. Prevalence and trends in obesity among US adults, 1999–2000. JAMA. 2002;288: 1723–1727. 30. O’Neill K, Reid G. Perceived barriers to physical activity by older adults. Can J Public Health. 1991;82:392–396. 31. Dishman RK. Motivating older adults to exercise. South Med J. 1994; 87:S79–S82. 32. Rhodes RE, Martin AD, Taunton JE, Rhodes EC, Donnelly M, Elliot J. Factors associated with exercise adherence among older adults. An individual perspective. Sports Med. 1999;28:397–411. 33. King AC. Interventions to promote physical activity by older adults. J Gerontol A Biol Sci Med Sci. 2001;56A(Spec Iss No II):36–46. 34. Burton LC, Shapiro S, German PS. Determinants of physical activity initiation and maintenance among community-dwelling older persons. Prev Med. 1999;29:422–430. 35. O’Grady M, Fletcher J, Ortiz S. Therapeutic and physical fitness exercise prescription for older adults with joint disease: an evidencebased approach. Rheum Dis Clin North Am. 2000;26:617–646. 36. Stutts WC. Physical activity determinants in adults. Perceived benefits, barriers, and self efficacy. AAOHN J. 2002;50:499–507. 37. Nied RJ, Franklin B. Promoting and prescribing exercise for the elderly. Am Fam Physician. 2002;65:419–426. Received January 19, 2005 Accepted May 16, 2005 Decision Editor: Luigi Ferrucci, MD, PhD