- Email: [email protected]
Frequent and occasional physical activity in the elderly
Frequent and Occasional Physical Activity in the Elderly A 12-Year Follow-up Study of Mortality Kristina Sundquist, MD, PhD, Jan Qvist, PhL, Jan Sundquist, MD, PhD, Sven-Erik Johansson, PhD Background: The positive health effects of physical activity are well known. However, there are few studies of the association between different levels of physical activity and all-cause mortality among elderly people. Methods:
A national random sample of 3206 women and men aged ⱖ65 were interviewed in 1988 and 1989 and followed until December 31, 2000, for all-cause mortality. Cox regression was used to analyze the association between five different levels of physical activity and all-cause mortality, after adjustment for gender, age, education, smoking habits, body mass index, diabetes, hypertension, and self-rated health. All analyses were conducted in 2003.
Results:
For elderly people who were physically active occasionally, the risk of all-cause mortality was 28% lower than for those who were physically inactive (hazard ratio [HR]⫽0.72; confidence interval [CI]⫽0.64 – 0.81), after adjustment for all explanatory variables. For those who were physically active once a week, the risk of all-cause mortality was 40% lower than for those who were physically inactive (HR⫽0.60; CI⫽0.50 – 0.71). For those who were physically active more frequently, the reduction in all-cause mortality risk was about the same as for those who were physically active once a week. Diabetes, hypertension, and daily smoking were, as expected, significant risk factors for all-cause mortality.
Conclusions: Physical activity, even occasionally, decreases the risk of all-cause mortality among elderly people. Preventive resources among the elderly should include moderate exercise such as walking. (Am J Prev Med 2004;27(1):22–27) © 2004 American Journal of Preventive Medicine
Introduction
L
ittle is known about the relationship between different levels of physical activity and mortality among the elderly. Past studies of middle-aged women and men have demonstrated an association between physical activity and increased longevity.1,2 Moreover, individual changes over time to a more sedentary lifestyle were associated with increased allcause mortality.3 However, past research into physical activity and mortality among the elderly is sparse4,5 and inconsistent. For example, some studies have even shown an inverse effect between no physical activity and all-cause mortality among the elderly.6,7 An explanation for this inconsistency may be that some of the previous studies focusing on elderly people have not adjusted for self-rated health. Thus, these studies do not permit
From the Karolinska Institutet, Family Medicine (K. Sundquist, J. Sundquist, Johansson), and Statistics Sweden (Qvist), Stockholm, Sweden Address correspondence and reprint requests to: Kristina Sundquist, MD, Family Medicine Stockholm, Karolinska Institutet, Alfred Nobels alle´ 12, SE-141 83, Huddinge, Sweden. E-mail: [email protected].
22
evaluation of physical inactivity that is not a result of disease processes. Moreover, the beneficial effect of vigorous versus moderate physical activity on longevity is controversial.8,9 The purpose of this 12-year follow-up study was to analyze the association between five levels of physical activity and all-cause mortality for people aged ⱖ65 using a random sample representative of the entire Swedish population. The first aim of this study was to examine to what extent different levels of physical activity are associated with all-cause mortality in people aged ⱖ65. The second aim was to examine whether the hypothesized association between physical activity and all-cause mortality remains after adjustment for confounders such as age, educational status, smoking habits, body mass index (BMI), diabetes, hypertension, and self-rated health.
Methods A simple, random sample from the Swedish Annual Level-ofLiving Survey (SALLS), a national survey conducted by Statistics Sweden of the non-institutionalized population was used in this study. The sample consisted of 1792 women and
Am J Prev Med 2004;27(1) © 2004 American Journal of Preventive Medicine • Published by Elsevier Inc.
0749-3797/04/$–see front matter doi:10.1016/j.amepre.2004.03.011
1414 men, aged ⱖ65, surveyed in 1988 and 1989. The sample represented 792,000 women and 616,000 men of the Swedish population aged ⱖ65. The participants were interviewed face-to-face by trained interviewers about their living conditions, including questions about leisure-time physical activity, education, smoking habits, weight, height, and health status. Only those who were interviewed directly by the interviewer were included in the sample. Participants who were interviewed through others, such as relatives, were excluded. Participants were followed from the date of interview to death from all causes or until the end of the study on December 31, 2000 (mean follow-up time was 11.7 years). All analyses were conducted in 2003. The outcome, all-cause mortality, was obtained by linking SALLS to the Cause-of-Death Register. The nonresponse rate for SALLS was 22.1%. Independent variables, all of which were obtained from SALLS, are listed below: Gender. Age was categorized into the following groups: 65 to 69, 70 to 74, 75 to 79, 80 to 89, and ⱖ90. Educational status was classified into two categories according to the duration of school attendance: ⱕ9 years (compulsory school or less); and ⱖ10 years (at least 1 year of high school). Leisure-time physical activity level was based on the individual’s response to the following item. (1) I get practically no exercise at all; (2) I exercise occasionally (e.g., 1-hour walks, skiing a couple of times every year, swimming, picking mushrooms); (3) I exercise about once a week (e.g., fast walks, skiing, swimming, jogging, cycling); (4) I exercise about twice a week (e.g., fast walks, skiing, swimming, jogging, cycling); (5) I exercise vigorously at least twice a week (e.g., skiing, swimming, running, cycling for quite a while, ball games). Smoking habits were grouped as (1) never smokers; (2) former smokers; and (3) current smokers. BMI was calculated as kilograms/meters squared, according to the World Health Organization’s recommendations, and comprised four categories for both women and men: (1) ⱕ18.5 (underweight); (2) ⬎18.6 to ⬍25.0 (normal weight); (3) 25.0 to ⬍30 (overweight); and (4) ⱖ30 (obesity). Weight and height were self-reported. Cases lacking information about weight and/or height were excluded from the analysis. Diabetes and hypertension were based on responses to the following questions: “Do you suffer from any long-standing disease?” “If yes, what kind of disease?” “Do you suffer from diabetes/hypertension?” “Do you take medicine for your diabetes/hypertension?” If diabetes and/or hypertension were reported as a disease, or medicine was taken for diabetes and/or hypertension, the respondent was judged as having diabetes and/or hypertension. All others were judged as not having diabetes or hypertension. Self-rated health was based on responses to the question, “How would you describe your general health?” Response alternatives included “good,” “bad,” and “something in between.” Those who answered that their general health was bad or something in between were considered as having a poor self-rated health status. A poor self-rated
Table 1. Age-adjusted mortality rates (per 10,000 personyears) by explanatory variables and gender (n⫽3206) Variable Total Age
Education Physical activity
Smoking habits Body mass index
Diabetes Hypertension Self-rated health
Level 65–69 70–74 75–79 80–89 ⱖ90 1 (lowest) 2 (highest) 1 (none) 2 (occasionally) 3 (once a week) 4 (twice a week) 5 (vigorously at least twice a week) 1 (never smoker) 2 (former smoker) 3 (current smoker) 1 (underweight) 2 (normal weight) 3 (overweight) 4 (obesity) Yes No Yes No Poor Good
Men
Women
794 340 564 1016 1567 2777 804 641 1100 744 555 495 573
516 198 327 551 1211 2252 573 491 776 480 393 331 400
669 705 992 1621 731 706 1021 1172 706 931 695 1010 572
533 510 674 765 547 477 742 1100 513 603 524 710 409
health status can include other morbid or co-morbid conditions such as chronic lung disease.
Statistical Analysis The SAS software package was used in the statistical analyses (SAS Institute Inc., Version 6, Cary NC, 1989). Gender- and age-adjusted all-cause mortality rates (per 10,000 persons per year) were calculated between 1988 –1989 and December 31, 2000, by indirect standardization.10 A Cox regression model11 was used to estimate the hazard ratio (HR) of all-cause mortality for the variables. The results are shown as HRs with 95% confidence intervals (CIs). Risk time was calculated from the interview until death. By including an interaction between each independent variable and time, the proportional hazards assumption was tested. All variables except self-rated health met the assumption. A possible rationale for this finding might be that ill persons at baseline have a higher mortality risk. Therefore, self-rated health was included as a stratum variable in the Cox regression analysis. This means that all analyses are also adjusted for self-rated health by taking the lack of proportionality in self-rated health into consideration. It is not possible to calculate a hazard ratio for self-rated health when it is included as a stratum.
Results There were 925 deaths among women and 881 deaths among men. Age-adjusted all-cause mortality rates by explanatory variables are shown in Table 1. For both Am J Prev Med 2004;27(1)
23
Table 2. Relative age-adjusted mortality risks by explanatory variables and sex (self-rated health as stratum variable) Variable
Level
Men HR (CI)
Women HR (CI)
Age
65–69 70–74 75–79 80–89 ⱖ90 1 (lowest) 2 (highest) 1 (none) 2 (occasionally) 3 (once a week) 4 (twice a week) 5 (vigorously at least twice a week) 1 (never smoker) 2 (former smoker) 3 (current smoker) 1 (underweight) 2 (normal weight) 3 (overweight) 4 (obesity) Yes No Yes No
1 (Reference) 1.63 (1.31–2.01) 3.02 (2.45–3.73) 4.76 (3.85–5.88) 10.1 (7.38–13.8) 1.17 (1.01–1.35) 1 (Reference) 1 (Reference) 0.74 (0.62–0.87) 0.57 (0.44–0.73) 0.51 (0.41–0.64) 0.60 (0.44–0.82)
1 (Reference) 1.62 (1.25–2.08) 2.75 (2.16–3.49) 6.46 (5.18–8.06) 13.6 (10.1–18.5) 1.13 (0.97–1.32) 1 (Reference) 1 (Reference) 0.70 (0.59–0.82) 0.59 (0.46–0.77) 0.47 (0.35–0.62) 0.54 (0.31–0.94)
1 (Reference) 1.03 (0.89–1.21) 1.49 (1.24–1.80) 2.74 (1.82–4.13) 1 (Reference) 0.94 (0.81–1.08) 1.14 (0.87–1.48) 1.52 (1.23–1.88) 1 (Reference) 1.23(1.05–1.44) 1 (Reference)
1 (Reference) 0.88 (0.71–1.10) 1.45 (1.16–1.82) 1.29 (0.99–1.69) 1 (Reference) 0.85 (0.73–0.99) 1.31 (1.03–1.67) 2.01 (1.65–2.45) 1 (Reference) 1.07 (0.93–1.23) 1 (Reference)
Education Physical activity
Smoking habits Body mass index
Diabetes Hypertension
CI, 95% confidence interval; HR, hazard ratios.
men and women, the highest mortality rates were observed among those who reported low educational status, physical inactivity, current smoking, underweight, obesity, diabetes, hypertension, or poor selfrated health. Table 2 shows age-adjusted HRs with 95% CIs for all-cause mortality by explanatory variables, separated by gender. Self-rated health is included in the model as a stratum, which means that all models are also adjusted for self-rated health. Both men and women who were physically active occasionally or more often (levels 2 to 5) had a significantly decreased mortality risk compared to those who were physically inactive (level 1). For example, for those who were physically active once a week, the mortality risk was 43% lower for men (HR⫽0.57; CI⫽0.44 – 0.73) and 41% lower for women (HR⫽0.59; CI⫽0.46 – 0.77) than for those who were physically inactive. For those who were physically active more than once a week, that is, about twice a week or vigorously at least twice a week, there was no further decrease in mortality risk compared to those who were physically active once a week. Education level seemed to be of minor importance for the mortality risk among elderly people. Current smokers had a roughly 50% higher mortality risk in the 12-year follow-up period than never smokers. Underweight status among men and obesity among women were associated with higher mortality. Self-reported diabetes in men and women was associated with a 52% and 101% higher mortality, respectively, compared to men and women without 24
diabetes. Self-reported hypertension was associated with an increased mortality risk among men, but not among women. Table 3 shows the HRs for all-cause mortality after stepwise inclusion of the explanatory variables. Since women and men showed a similar risk pattern on every level of physical activity and the other explanatory variables, they were analyzed together in the final models. Self-rated health is included as a stratum variable in all the models, which allowed us to adjust for this variable. All three models are also adjusted for sex, age, and educational status. Model 2 is also adjusted for physical activity, smoking habits, and BMI. Model 3 is adjusted for all variables in Model 2 and also diabetes and hypertension. After adjustment for all the explanatory variables, results for the variable physical activity remained. For those who were physically active even occasionally, the mortality risk was lower than for those who were physically inactive (HR⫽0.72; CI⫽0.64 – 0.81). For those who were physically active once a week, the mortality risk decreased even further. Compared with those who were physically inactive, mortality risk decreased by 40% (HR⫽0.60; CI⫽0.50 – 0.71). There was no statistically significant difference in mortality risk between physical activity once a week and twice a week or vigorously at least twice a week. Diabetes, hypertension, and current smoking were associated with increased mortality risk, while obesity was no longer a risk factor for mortality in the final model.
American Journal of Preventive Medicine, Volume 27, Number 1
Table 3. Relative mortality risks with stepwise inclusion of explanatory variables (self-rated health as stratum variable) Variable
Level
Gender
Men Women 65–69 70–74 75–79 80–89 ⱖ90 1 (lowest) 2 (highest) 1 (none) 2 (occasionally) 3 (once a week) 4 (twice a week) 5 (vigorously at least twice a week) 1 (never smoker) 2 (former smoker) 3 (current smoker) 1 (underweight) 2 (normal weight) 3 (overweight) 4 (obesity) Yes No Yes No
Age (years)
Education Physical activity
Smoking habits Body mass index
Diabetes Hypertension
Model 1a HR (CI)
Model 2b HR (CI)
Model 3c HR (CI)
1.70 (1.55–1.87) 1 (Reference) 1 (Reference) 1.61 (1.37–1.90) 2.84 (2.42–3.32) 5.54 (4.76–6.45) 11.7 (9.38–14.5) 1.15 (1.03–1.27) 1 (Reference)
1.79 (1.61–2.00) 1 (Reference) 1 (Reference) 1.57 (1.33–1.85) 2.69 (2.30–3.16) 4.92 (4.20–5.76) 9.56 (7.61–12.0) 1.12 (1.01–1.25) 1 (Reference) 1 (Reference) 0.72 (0.64–0.81) 0.58 (0.49–0.70) 0.50 (0.42–0.60) 0.59 (0.45–0.77)
1.78 (1.60–1.99) 1 (Reference) 1 (Reference) 1.55 (1.31–1.83) 2.71 (2.31–3.18) 4.95 (4.22–5.79) 9.86 (7.84–12.4) 1.09 (0.98–1.21) 1 (Reference) 1 (Reference) 0.72 (0.64–0.81) 0.60 (0.50–0.71) 0.50 (0.42–0.59) 0.60 (0.46–0.79)
1 (Reference) 0.99 (0.88–1.12) 1.38 (1.20–1.59) 1.45 (1.16–1.81) 1 (Reference) 0.87 (0.78–0.96) 1.09 (0.91–1.30)
1 (Reference) 0.99 (0.89–1.13) 1.44 (1.25–1.66) 1.50 (1.20–1.87) 1 (Reference) 0.86 (0.77–0.95) 1.02 (0.85–1.22) 1.75 (1.51–2.03) 1 (Reference) 1.13 (1.02–1.26) 1 (Reference)
Note: In all three models, self-rated health is included as a stratum variable. a Model 1 is adjusted for gender, age, and education. b Model 2 is also adjusted for physical activity, smoking habits, and body mass index. c Model 3 is adjusted for all explanatory variables. CI, 95% confidence interval; HR, hazard ratios.
Discussion The main finding of this study is that leisure-time physical activity decreases all-cause mortality among men and women aged ⱖ65. Even occasional physical activity was associated with decreased mortality among the elderly. For those who were physically active about twice a week or vigorously at least twice a week the reduction in the risk of all-cause mortality was about the same as for those who were physically active once a week after adjustment for all the explanatory variables, including self-rated health. Diabetes, current smoking, and hypertension were significant risk factors for allcause mortality. The findings of this study agreed with other studies from the United States, which found that physical activity among older adults is associated with decreased all-cause mortality.12,13 Regular walking among older men participating in the Honolulu Heart Program was also associated with decreased all-cause mortality.8 Moreover, this study demonstrated that even occasional physical activity is associated with decreased all-cause mortality. In contrast, data from the Seventh-Day Adventist study revealed only a small effect of physical activity on mortality for men aged 65 to 74 at the start of the study.14 Cardiovascular disease (CVD) accounts
for most of the mortality in the Western countries despite the decline in CVD during the last few decades.15 The findings of an association between physical activity and decreased all-cause mortality are probably related mostly to decreased CVD mortality. A recent systematic review of previous research on older adults found a cardio-protective effect of physical activity among older men.16 The mechanisms between physical activity and CVD have been partly elucidated; for example, physical activity has a lowering effect on blood pressure,12,17 a lowering effect on plasma fibrinogen,17–19 and a lowering effect on plasma viscosity.18 Moreover, a growing body of research has demonstrated the important role of physical activity in the prevention of type 2 diabetes mellitus, a strong risk factor for CVD.20 –22 Other studies have shown the benefits of physical activity for the prevention of diabetes among individuals with impaired glucose tolerance.23,24 In addition, the Center for Prevention at the Danish National Board of Health recently published a report on the beneficial effect of physical activity in the prevention and treatment of several fatal and nonfatal diseases.25 Obesity was no longer a risk factor for mortality in the final model, which is in agreement with a study from the United States that showed that BMI Am J Prev Med 2004;27(1)
25
was more strongly associated with CHD risk among middle-aged men than among older men.26
Limitations and Strengths There were some limitations to this study. For example, questionnaires that do not take non–leisure-time physical activities, such as gardening or household-related activities, into consideration might serve to bias results. The total amount of physical activity would then be underestimated among certain groups. Self-report might also constitute an important bias. However, self-reported measures used in the SALLS, such as weight and height, have been found to be reliable in other studies,27–29 even though a recent British study found that older obese women tended to underestimate their weight.30 Finally, another possible limitation involves the survival-cohort effect. However, there is no reason to believe that the survival cohort effect is different among various subgroups. Another limitation is nonresponse (22.1%). Of the nonrespondents, approximately 70% were refusals, 20% could not be located, and 10% were too ill to participate. This possible response bias was examined by including both nonrespondents and respondents in a proportional hazard model that adjusted for gender, age, marital status, and region with all-cause mortality as the outcome. Those who refused had the same mortality risk as the respondents, but the other two groups had significantly higher mortality risk. Despite these limitations, there are several strengths in the present study. The SALLS survey is based on a representative sample of the Swedish population; therefore, people from various socioeconomic groups are included in the study. Response rates for SALLS were high (almost 80%), and missing data were minimized. Reliability of survey questions was estimated by reinterviewing a sample of the participants (test–retest method). The kappa coefficients were 0.64 for selfrated health and 0.58 for physical activity.31 The sampling procedure is simpler in Sweden than in other countries because of the Swedish population register and personal identification number, which enable tracking of every person. The completeness of data in the Cause of Death Register is high, with less than 0.5% missing. The set of variables also included several confounders, such as education, smoking habits, BMI, diabetes, and hypertension. Unlike many other studies of physical activity among the elderly, this study was adjusted for self-rated health, making it possible to evaluate whether disease processes confound the association between physical activity and mortality. Moreover, many previous studies focused only on men, whereas this study focused on both women and men aged ⱖ65. 26
Conclusions These findings provide some evidence that all-cause mortality is reduced when elderly men and women are physically active. Even occasional physical activity seems to be of great value for longevity. Preventive resources among elderly should provide more opportunities for physical activity. For example, senior centers could offer retirees moderate exercise such as walking in groups or bicycling, and healthcare professionals should encourage elderly people to be physically active, even occasionally. This work was supported by grants from the National Institutes of Health (1 R01 HL71084-01), the Swedish Council for Working Life and Social Research (2001-2373), the Swedish Research Council (K2001-27X-11651-06C), the Knut and Alice Wallenberg Foundation, and the Stockholm County Council. We are grateful to Sanna Sundquist, student at Foothill College, California, for technical assistance.
References 1. Pekkanen J, Marti B, Nissinen A, Tuomilehto J, Punsar S, Karvonen MJ. Reduction of premature mortality by high physical activity: a 20-year follow-up of middle-aged Finnish men. Lancet 1987;1:1473–7. 2. Paffenbarger RS Jr, Hyde RT, Wing AL, Hsieh CC. Physical activity, all-cause mortality, and longevity of college alumni. N Engl J Med 1986; 314:605–13. 3. Johansson SE, Sundquist J. Change in lifestyle factors and their influence on health status and all-cause mortality. Int J Epidemiol 1999;28:1073–80. 4. Kaplan GA, Seeman TE, Cohen RD, Knudsen LP, Guralnik J. Mortality among the elderly in the Alameda County Study: behavioral and demographic risk factors. Am J Public Health 1987;77:307–12. 5. Donahue RP, Abbott RD, Reed DM, Yano K. Physical activity and coronary heart disease in middle-aged and elderly men: the Honolulu Heart Program. Am J Public Health 1988;78:683–5. 6. Fraser GE, Shavlik DJ. Risk factors for all-cause and coronary heart disease mortality in the oldest-old. The Adventist Health Study. Arch Intern Med 1997;157:2249 –58. 7. Lee IM, Sesso HD, Paffenbarger RS Jr. Physical activity and coronary heart disease risk in men: does the duration of exercise episodes predict risk? Circulation 2000;102:981–6. 8. Hakim AA, Petrovitch H, Burchfiel CM, et al. Effects of walking on mortality among nonsmoking retired men. N Engl J Med 1998;338:94 –9 (see comments). 9. Rodriguez BL, Curb JD, Burchfiel CM, et al. Physical activity and 23-year incidence of coronary heart disease morbidity and mortality among middle-aged men. The Honolulu Heart Program. Circulation 1994;89:2540 –4. 10. Breslow NE, Day NE. Statistical methods in cancer research. Volume II. The design and analysis of cohort studies. Lyon: International Agency for Research on Cancer, 1987. 11. Kleinbaum D. Survival analysis. New York: Springer, 1995. 12. Simonsick EM, Lafferty ME, Phillips CL, et al. Risk due to inactivity in physically capable older adults. Am J Public Health 1993;83:1443–50. 13. Fried LP, Kronmal RA, Newman AB, et al. Risk factors for 5-year mortality in older adults: the Cardiovascular Health Study. JAMA 1998;279:585–92. 14. Lindsted KD, Tonstad S, Kuzma JW. Self-report of physical activity and patterns of mortality in Seventh-Day Adventist men. J Clin Epidemiol 1991;44:355–64. 15. Peltonen M, Asplund K. Age-period-cohort effects on ischaemic heart disease mortality in Sweden from 1969 to 1993, and forecasts up to 2003. Eur Heart J 1997;18:1307–12. 16. Batty GD. Physical activity and coronary heart disease in older adults. A systematic review of epidemiological studies. Eur J Public Health 2002;12: 171–6.
American Journal of Preventive Medicine, Volume 27, Number 1
17. MacAuley D, McCrum EE, Stott G, et al. Physical activity, physical fitness, blood pressure, and fibrinogen in the Northern Ireland health and activity survey. J Epidemiol Community Health 1996;50:258 –63. 18. Koenig W, Sund M, Doring A, Ernst E. Leisure-time physical activity but not work-related physical activity is associated with decreased plasma viscosity. Results from a large population sample. Circulation 1997;95:335–41 (see comments). 19. Greendale GA, Bodin-Dunn L, Ingles S, Haile R, Barrett-Connor E. Leisure, home, and occupational physical activity and cardiovascular risk factors in postmenopausal women. The Postmenopausal Estrogens/Progestins Intervention (PEPI) Study. Arch Intern Med 1996;156:418 –24. 20. Burchfiel CM, Sharp DS, Curb JD, et al. Physical activity and incidence of diabetes: the Honolulu Heart Program. Am J Epidemiol 1995;141:360 –8. 21. Perry IJ, Wannamethee SG, Walker MK, Thomson AG, Whincup PH, Shaper AG. Prospective study of risk factors for development of non-insulin dependent diabetes in middle aged British men. BMJ 1995;310:560 –4 (see comments). 22. Manson JE, Nathan DM, Krolewski AS, Stampfer MJ, Willett WC, Hennekens CH. A prospective study of exercise and incidence of diabetes among US male physicians. JAMA 1992;268:63–7. 23. Eriksson KF, Lindgarde F. Prevention of type 2 (non–insulin-dependent) diabetes mellitus by diet and physical exercise. The 6-year Malmo feasibility study. Diabetologia 1991;34:891–8.
24. Pan XR, Li GW, Hu YH, et al. Effects of diet and exercise in preventing NIDDM in people with impaired glucose tolerance. The Da Qing IGT and Diabetes Study. Diabetes Care 1997;20:537–44. 25. Fysisk aktivitet-håndbog om forbyggelse og behandling [Physical activity: handbook on prevention and treatment. Denmark: National Board of Health, Center for Prevention, 2003. 26. Rexrode KM, Buring JE, Manson JE. Abdominal and total adiposity and risk of coronary heart disease in men. Int J Obes Relat Metab Disord 2001;25: 1047–56. 27. Stewart AL. The reliability and validity of self-reported weight and height. J Chronic Dis 1982;35:295–309. 28. Stunkard AJ, Albaum JM. The accuracy of self-reported weights. Am J Clin Nutr 1981;34:1593–9. 29. Kuskowska-Wolk A, Karlsson P, Stolt M, Rossner S. The predictive validity of body mass index based on self-reported weight and height. Int J Obes 1989;13:441–53. 30. Lawlor DA, Bedford C, Taylor M, Ebrahim S. Agreement between measured and self-reported weight in older women. Results from the British Women’s Heart and Health Study. Age Ageing 2002;31:169 –74. 31. Wa¨ rneryd B. Levnadsfo¨ rhållanden. Återintervjustudie i underso¨ kningen av levnadsfo¨ rhållanden 1989 (ULF). [Living conditions. Reinterview in ULF 1989. Appendix 12.] Stockholm: Statistics Sweden, 1991.
Am J Prev Med 2004;27(1)
27
A national random sample of 3206 women and men aged ⱖ65 were interviewed in 1988 and 1989 and followed until December 31, 2000, for all-cause mortality. Cox regression was used to analyze the association between five different levels of physical activity and all-cause mortality, after adjustment for gender, age, education, smoking habits, body mass index, diabetes, hypertension, and self-rated health. All analyses were conducted in 2003.
Results:
For elderly people who were physically active occasionally, the risk of all-cause mortality was 28% lower than for those who were physically inactive (hazard ratio [HR]⫽0.72; confidence interval [CI]⫽0.64 – 0.81), after adjustment for all explanatory variables. For those who were physically active once a week, the risk of all-cause mortality was 40% lower than for those who were physically inactive (HR⫽0.60; CI⫽0.50 – 0.71). For those who were physically active more frequently, the reduction in all-cause mortality risk was about the same as for those who were physically active once a week. Diabetes, hypertension, and daily smoking were, as expected, significant risk factors for all-cause mortality.
Conclusions: Physical activity, even occasionally, decreases the risk of all-cause mortality among elderly people. Preventive resources among the elderly should include moderate exercise such as walking. (Am J Prev Med 2004;27(1):22–27) © 2004 American Journal of Preventive Medicine
Introduction
L
ittle is known about the relationship between different levels of physical activity and mortality among the elderly. Past studies of middle-aged women and men have demonstrated an association between physical activity and increased longevity.1,2 Moreover, individual changes over time to a more sedentary lifestyle were associated with increased allcause mortality.3 However, past research into physical activity and mortality among the elderly is sparse4,5 and inconsistent. For example, some studies have even shown an inverse effect between no physical activity and all-cause mortality among the elderly.6,7 An explanation for this inconsistency may be that some of the previous studies focusing on elderly people have not adjusted for self-rated health. Thus, these studies do not permit
From the Karolinska Institutet, Family Medicine (K. Sundquist, J. Sundquist, Johansson), and Statistics Sweden (Qvist), Stockholm, Sweden Address correspondence and reprint requests to: Kristina Sundquist, MD, Family Medicine Stockholm, Karolinska Institutet, Alfred Nobels alle´ 12, SE-141 83, Huddinge, Sweden. E-mail: [email protected].
22
evaluation of physical inactivity that is not a result of disease processes. Moreover, the beneficial effect of vigorous versus moderate physical activity on longevity is controversial.8,9 The purpose of this 12-year follow-up study was to analyze the association between five levels of physical activity and all-cause mortality for people aged ⱖ65 using a random sample representative of the entire Swedish population. The first aim of this study was to examine to what extent different levels of physical activity are associated with all-cause mortality in people aged ⱖ65. The second aim was to examine whether the hypothesized association between physical activity and all-cause mortality remains after adjustment for confounders such as age, educational status, smoking habits, body mass index (BMI), diabetes, hypertension, and self-rated health.
Methods A simple, random sample from the Swedish Annual Level-ofLiving Survey (SALLS), a national survey conducted by Statistics Sweden of the non-institutionalized population was used in this study. The sample consisted of 1792 women and
Am J Prev Med 2004;27(1) © 2004 American Journal of Preventive Medicine • Published by Elsevier Inc.
0749-3797/04/$–see front matter doi:10.1016/j.amepre.2004.03.011
1414 men, aged ⱖ65, surveyed in 1988 and 1989. The sample represented 792,000 women and 616,000 men of the Swedish population aged ⱖ65. The participants were interviewed face-to-face by trained interviewers about their living conditions, including questions about leisure-time physical activity, education, smoking habits, weight, height, and health status. Only those who were interviewed directly by the interviewer were included in the sample. Participants who were interviewed through others, such as relatives, were excluded. Participants were followed from the date of interview to death from all causes or until the end of the study on December 31, 2000 (mean follow-up time was 11.7 years). All analyses were conducted in 2003. The outcome, all-cause mortality, was obtained by linking SALLS to the Cause-of-Death Register. The nonresponse rate for SALLS was 22.1%. Independent variables, all of which were obtained from SALLS, are listed below: Gender. Age was categorized into the following groups: 65 to 69, 70 to 74, 75 to 79, 80 to 89, and ⱖ90. Educational status was classified into two categories according to the duration of school attendance: ⱕ9 years (compulsory school or less); and ⱖ10 years (at least 1 year of high school). Leisure-time physical activity level was based on the individual’s response to the following item. (1) I get practically no exercise at all; (2) I exercise occasionally (e.g., 1-hour walks, skiing a couple of times every year, swimming, picking mushrooms); (3) I exercise about once a week (e.g., fast walks, skiing, swimming, jogging, cycling); (4) I exercise about twice a week (e.g., fast walks, skiing, swimming, jogging, cycling); (5) I exercise vigorously at least twice a week (e.g., skiing, swimming, running, cycling for quite a while, ball games). Smoking habits were grouped as (1) never smokers; (2) former smokers; and (3) current smokers. BMI was calculated as kilograms/meters squared, according to the World Health Organization’s recommendations, and comprised four categories for both women and men: (1) ⱕ18.5 (underweight); (2) ⬎18.6 to ⬍25.0 (normal weight); (3) 25.0 to ⬍30 (overweight); and (4) ⱖ30 (obesity). Weight and height were self-reported. Cases lacking information about weight and/or height were excluded from the analysis. Diabetes and hypertension were based on responses to the following questions: “Do you suffer from any long-standing disease?” “If yes, what kind of disease?” “Do you suffer from diabetes/hypertension?” “Do you take medicine for your diabetes/hypertension?” If diabetes and/or hypertension were reported as a disease, or medicine was taken for diabetes and/or hypertension, the respondent was judged as having diabetes and/or hypertension. All others were judged as not having diabetes or hypertension. Self-rated health was based on responses to the question, “How would you describe your general health?” Response alternatives included “good,” “bad,” and “something in between.” Those who answered that their general health was bad or something in between were considered as having a poor self-rated health status. A poor self-rated
Table 1. Age-adjusted mortality rates (per 10,000 personyears) by explanatory variables and gender (n⫽3206) Variable Total Age
Education Physical activity
Smoking habits Body mass index
Diabetes Hypertension Self-rated health
Level 65–69 70–74 75–79 80–89 ⱖ90 1 (lowest) 2 (highest) 1 (none) 2 (occasionally) 3 (once a week) 4 (twice a week) 5 (vigorously at least twice a week) 1 (never smoker) 2 (former smoker) 3 (current smoker) 1 (underweight) 2 (normal weight) 3 (overweight) 4 (obesity) Yes No Yes No Poor Good
Men
Women
794 340 564 1016 1567 2777 804 641 1100 744 555 495 573
516 198 327 551 1211 2252 573 491 776 480 393 331 400
669 705 992 1621 731 706 1021 1172 706 931 695 1010 572
533 510 674 765 547 477 742 1100 513 603 524 710 409
health status can include other morbid or co-morbid conditions such as chronic lung disease.
Statistical Analysis The SAS software package was used in the statistical analyses (SAS Institute Inc., Version 6, Cary NC, 1989). Gender- and age-adjusted all-cause mortality rates (per 10,000 persons per year) were calculated between 1988 –1989 and December 31, 2000, by indirect standardization.10 A Cox regression model11 was used to estimate the hazard ratio (HR) of all-cause mortality for the variables. The results are shown as HRs with 95% confidence intervals (CIs). Risk time was calculated from the interview until death. By including an interaction between each independent variable and time, the proportional hazards assumption was tested. All variables except self-rated health met the assumption. A possible rationale for this finding might be that ill persons at baseline have a higher mortality risk. Therefore, self-rated health was included as a stratum variable in the Cox regression analysis. This means that all analyses are also adjusted for self-rated health by taking the lack of proportionality in self-rated health into consideration. It is not possible to calculate a hazard ratio for self-rated health when it is included as a stratum.
Results There were 925 deaths among women and 881 deaths among men. Age-adjusted all-cause mortality rates by explanatory variables are shown in Table 1. For both Am J Prev Med 2004;27(1)
23
Table 2. Relative age-adjusted mortality risks by explanatory variables and sex (self-rated health as stratum variable) Variable
Level
Men HR (CI)
Women HR (CI)
Age
65–69 70–74 75–79 80–89 ⱖ90 1 (lowest) 2 (highest) 1 (none) 2 (occasionally) 3 (once a week) 4 (twice a week) 5 (vigorously at least twice a week) 1 (never smoker) 2 (former smoker) 3 (current smoker) 1 (underweight) 2 (normal weight) 3 (overweight) 4 (obesity) Yes No Yes No
1 (Reference) 1.63 (1.31–2.01) 3.02 (2.45–3.73) 4.76 (3.85–5.88) 10.1 (7.38–13.8) 1.17 (1.01–1.35) 1 (Reference) 1 (Reference) 0.74 (0.62–0.87) 0.57 (0.44–0.73) 0.51 (0.41–0.64) 0.60 (0.44–0.82)
1 (Reference) 1.62 (1.25–2.08) 2.75 (2.16–3.49) 6.46 (5.18–8.06) 13.6 (10.1–18.5) 1.13 (0.97–1.32) 1 (Reference) 1 (Reference) 0.70 (0.59–0.82) 0.59 (0.46–0.77) 0.47 (0.35–0.62) 0.54 (0.31–0.94)
1 (Reference) 1.03 (0.89–1.21) 1.49 (1.24–1.80) 2.74 (1.82–4.13) 1 (Reference) 0.94 (0.81–1.08) 1.14 (0.87–1.48) 1.52 (1.23–1.88) 1 (Reference) 1.23(1.05–1.44) 1 (Reference)
1 (Reference) 0.88 (0.71–1.10) 1.45 (1.16–1.82) 1.29 (0.99–1.69) 1 (Reference) 0.85 (0.73–0.99) 1.31 (1.03–1.67) 2.01 (1.65–2.45) 1 (Reference) 1.07 (0.93–1.23) 1 (Reference)
Education Physical activity
Smoking habits Body mass index
Diabetes Hypertension
CI, 95% confidence interval; HR, hazard ratios.
men and women, the highest mortality rates were observed among those who reported low educational status, physical inactivity, current smoking, underweight, obesity, diabetes, hypertension, or poor selfrated health. Table 2 shows age-adjusted HRs with 95% CIs for all-cause mortality by explanatory variables, separated by gender. Self-rated health is included in the model as a stratum, which means that all models are also adjusted for self-rated health. Both men and women who were physically active occasionally or more often (levels 2 to 5) had a significantly decreased mortality risk compared to those who were physically inactive (level 1). For example, for those who were physically active once a week, the mortality risk was 43% lower for men (HR⫽0.57; CI⫽0.44 – 0.73) and 41% lower for women (HR⫽0.59; CI⫽0.46 – 0.77) than for those who were physically inactive. For those who were physically active more than once a week, that is, about twice a week or vigorously at least twice a week, there was no further decrease in mortality risk compared to those who were physically active once a week. Education level seemed to be of minor importance for the mortality risk among elderly people. Current smokers had a roughly 50% higher mortality risk in the 12-year follow-up period than never smokers. Underweight status among men and obesity among women were associated with higher mortality. Self-reported diabetes in men and women was associated with a 52% and 101% higher mortality, respectively, compared to men and women without 24
diabetes. Self-reported hypertension was associated with an increased mortality risk among men, but not among women. Table 3 shows the HRs for all-cause mortality after stepwise inclusion of the explanatory variables. Since women and men showed a similar risk pattern on every level of physical activity and the other explanatory variables, they were analyzed together in the final models. Self-rated health is included as a stratum variable in all the models, which allowed us to adjust for this variable. All three models are also adjusted for sex, age, and educational status. Model 2 is also adjusted for physical activity, smoking habits, and BMI. Model 3 is adjusted for all variables in Model 2 and also diabetes and hypertension. After adjustment for all the explanatory variables, results for the variable physical activity remained. For those who were physically active even occasionally, the mortality risk was lower than for those who were physically inactive (HR⫽0.72; CI⫽0.64 – 0.81). For those who were physically active once a week, the mortality risk decreased even further. Compared with those who were physically inactive, mortality risk decreased by 40% (HR⫽0.60; CI⫽0.50 – 0.71). There was no statistically significant difference in mortality risk between physical activity once a week and twice a week or vigorously at least twice a week. Diabetes, hypertension, and current smoking were associated with increased mortality risk, while obesity was no longer a risk factor for mortality in the final model.
American Journal of Preventive Medicine, Volume 27, Number 1
Table 3. Relative mortality risks with stepwise inclusion of explanatory variables (self-rated health as stratum variable) Variable
Level
Gender
Men Women 65–69 70–74 75–79 80–89 ⱖ90 1 (lowest) 2 (highest) 1 (none) 2 (occasionally) 3 (once a week) 4 (twice a week) 5 (vigorously at least twice a week) 1 (never smoker) 2 (former smoker) 3 (current smoker) 1 (underweight) 2 (normal weight) 3 (overweight) 4 (obesity) Yes No Yes No
Age (years)
Education Physical activity
Smoking habits Body mass index
Diabetes Hypertension
Model 1a HR (CI)
Model 2b HR (CI)
Model 3c HR (CI)
1.70 (1.55–1.87) 1 (Reference) 1 (Reference) 1.61 (1.37–1.90) 2.84 (2.42–3.32) 5.54 (4.76–6.45) 11.7 (9.38–14.5) 1.15 (1.03–1.27) 1 (Reference)
1.79 (1.61–2.00) 1 (Reference) 1 (Reference) 1.57 (1.33–1.85) 2.69 (2.30–3.16) 4.92 (4.20–5.76) 9.56 (7.61–12.0) 1.12 (1.01–1.25) 1 (Reference) 1 (Reference) 0.72 (0.64–0.81) 0.58 (0.49–0.70) 0.50 (0.42–0.60) 0.59 (0.45–0.77)
1.78 (1.60–1.99) 1 (Reference) 1 (Reference) 1.55 (1.31–1.83) 2.71 (2.31–3.18) 4.95 (4.22–5.79) 9.86 (7.84–12.4) 1.09 (0.98–1.21) 1 (Reference) 1 (Reference) 0.72 (0.64–0.81) 0.60 (0.50–0.71) 0.50 (0.42–0.59) 0.60 (0.46–0.79)
1 (Reference) 0.99 (0.88–1.12) 1.38 (1.20–1.59) 1.45 (1.16–1.81) 1 (Reference) 0.87 (0.78–0.96) 1.09 (0.91–1.30)
1 (Reference) 0.99 (0.89–1.13) 1.44 (1.25–1.66) 1.50 (1.20–1.87) 1 (Reference) 0.86 (0.77–0.95) 1.02 (0.85–1.22) 1.75 (1.51–2.03) 1 (Reference) 1.13 (1.02–1.26) 1 (Reference)
Note: In all three models, self-rated health is included as a stratum variable. a Model 1 is adjusted for gender, age, and education. b Model 2 is also adjusted for physical activity, smoking habits, and body mass index. c Model 3 is adjusted for all explanatory variables. CI, 95% confidence interval; HR, hazard ratios.
Discussion The main finding of this study is that leisure-time physical activity decreases all-cause mortality among men and women aged ⱖ65. Even occasional physical activity was associated with decreased mortality among the elderly. For those who were physically active about twice a week or vigorously at least twice a week the reduction in the risk of all-cause mortality was about the same as for those who were physically active once a week after adjustment for all the explanatory variables, including self-rated health. Diabetes, current smoking, and hypertension were significant risk factors for allcause mortality. The findings of this study agreed with other studies from the United States, which found that physical activity among older adults is associated with decreased all-cause mortality.12,13 Regular walking among older men participating in the Honolulu Heart Program was also associated with decreased all-cause mortality.8 Moreover, this study demonstrated that even occasional physical activity is associated with decreased all-cause mortality. In contrast, data from the Seventh-Day Adventist study revealed only a small effect of physical activity on mortality for men aged 65 to 74 at the start of the study.14 Cardiovascular disease (CVD) accounts
for most of the mortality in the Western countries despite the decline in CVD during the last few decades.15 The findings of an association between physical activity and decreased all-cause mortality are probably related mostly to decreased CVD mortality. A recent systematic review of previous research on older adults found a cardio-protective effect of physical activity among older men.16 The mechanisms between physical activity and CVD have been partly elucidated; for example, physical activity has a lowering effect on blood pressure,12,17 a lowering effect on plasma fibrinogen,17–19 and a lowering effect on plasma viscosity.18 Moreover, a growing body of research has demonstrated the important role of physical activity in the prevention of type 2 diabetes mellitus, a strong risk factor for CVD.20 –22 Other studies have shown the benefits of physical activity for the prevention of diabetes among individuals with impaired glucose tolerance.23,24 In addition, the Center for Prevention at the Danish National Board of Health recently published a report on the beneficial effect of physical activity in the prevention and treatment of several fatal and nonfatal diseases.25 Obesity was no longer a risk factor for mortality in the final model, which is in agreement with a study from the United States that showed that BMI Am J Prev Med 2004;27(1)
25
was more strongly associated with CHD risk among middle-aged men than among older men.26
Limitations and Strengths There were some limitations to this study. For example, questionnaires that do not take non–leisure-time physical activities, such as gardening or household-related activities, into consideration might serve to bias results. The total amount of physical activity would then be underestimated among certain groups. Self-report might also constitute an important bias. However, self-reported measures used in the SALLS, such as weight and height, have been found to be reliable in other studies,27–29 even though a recent British study found that older obese women tended to underestimate their weight.30 Finally, another possible limitation involves the survival-cohort effect. However, there is no reason to believe that the survival cohort effect is different among various subgroups. Another limitation is nonresponse (22.1%). Of the nonrespondents, approximately 70% were refusals, 20% could not be located, and 10% were too ill to participate. This possible response bias was examined by including both nonrespondents and respondents in a proportional hazard model that adjusted for gender, age, marital status, and region with all-cause mortality as the outcome. Those who refused had the same mortality risk as the respondents, but the other two groups had significantly higher mortality risk. Despite these limitations, there are several strengths in the present study. The SALLS survey is based on a representative sample of the Swedish population; therefore, people from various socioeconomic groups are included in the study. Response rates for SALLS were high (almost 80%), and missing data were minimized. Reliability of survey questions was estimated by reinterviewing a sample of the participants (test–retest method). The kappa coefficients were 0.64 for selfrated health and 0.58 for physical activity.31 The sampling procedure is simpler in Sweden than in other countries because of the Swedish population register and personal identification number, which enable tracking of every person. The completeness of data in the Cause of Death Register is high, with less than 0.5% missing. The set of variables also included several confounders, such as education, smoking habits, BMI, diabetes, and hypertension. Unlike many other studies of physical activity among the elderly, this study was adjusted for self-rated health, making it possible to evaluate whether disease processes confound the association between physical activity and mortality. Moreover, many previous studies focused only on men, whereas this study focused on both women and men aged ⱖ65. 26
Conclusions These findings provide some evidence that all-cause mortality is reduced when elderly men and women are physically active. Even occasional physical activity seems to be of great value for longevity. Preventive resources among elderly should provide more opportunities for physical activity. For example, senior centers could offer retirees moderate exercise such as walking in groups or bicycling, and healthcare professionals should encourage elderly people to be physically active, even occasionally. This work was supported by grants from the National Institutes of Health (1 R01 HL71084-01), the Swedish Council for Working Life and Social Research (2001-2373), the Swedish Research Council (K2001-27X-11651-06C), the Knut and Alice Wallenberg Foundation, and the Stockholm County Council. We are grateful to Sanna Sundquist, student at Foothill College, California, for technical assistance.
References 1. Pekkanen J, Marti B, Nissinen A, Tuomilehto J, Punsar S, Karvonen MJ. Reduction of premature mortality by high physical activity: a 20-year follow-up of middle-aged Finnish men. Lancet 1987;1:1473–7. 2. Paffenbarger RS Jr, Hyde RT, Wing AL, Hsieh CC. Physical activity, all-cause mortality, and longevity of college alumni. N Engl J Med 1986; 314:605–13. 3. Johansson SE, Sundquist J. Change in lifestyle factors and their influence on health status and all-cause mortality. Int J Epidemiol 1999;28:1073–80. 4. Kaplan GA, Seeman TE, Cohen RD, Knudsen LP, Guralnik J. Mortality among the elderly in the Alameda County Study: behavioral and demographic risk factors. Am J Public Health 1987;77:307–12. 5. Donahue RP, Abbott RD, Reed DM, Yano K. Physical activity and coronary heart disease in middle-aged and elderly men: the Honolulu Heart Program. Am J Public Health 1988;78:683–5. 6. Fraser GE, Shavlik DJ. Risk factors for all-cause and coronary heart disease mortality in the oldest-old. The Adventist Health Study. Arch Intern Med 1997;157:2249 –58. 7. Lee IM, Sesso HD, Paffenbarger RS Jr. Physical activity and coronary heart disease risk in men: does the duration of exercise episodes predict risk? Circulation 2000;102:981–6. 8. Hakim AA, Petrovitch H, Burchfiel CM, et al. Effects of walking on mortality among nonsmoking retired men. N Engl J Med 1998;338:94 –9 (see comments). 9. Rodriguez BL, Curb JD, Burchfiel CM, et al. Physical activity and 23-year incidence of coronary heart disease morbidity and mortality among middle-aged men. The Honolulu Heart Program. Circulation 1994;89:2540 –4. 10. Breslow NE, Day NE. Statistical methods in cancer research. Volume II. The design and analysis of cohort studies. Lyon: International Agency for Research on Cancer, 1987. 11. Kleinbaum D. Survival analysis. New York: Springer, 1995. 12. Simonsick EM, Lafferty ME, Phillips CL, et al. Risk due to inactivity in physically capable older adults. Am J Public Health 1993;83:1443–50. 13. Fried LP, Kronmal RA, Newman AB, et al. Risk factors for 5-year mortality in older adults: the Cardiovascular Health Study. JAMA 1998;279:585–92. 14. Lindsted KD, Tonstad S, Kuzma JW. Self-report of physical activity and patterns of mortality in Seventh-Day Adventist men. J Clin Epidemiol 1991;44:355–64. 15. Peltonen M, Asplund K. Age-period-cohort effects on ischaemic heart disease mortality in Sweden from 1969 to 1993, and forecasts up to 2003. Eur Heart J 1997;18:1307–12. 16. Batty GD. Physical activity and coronary heart disease in older adults. A systematic review of epidemiological studies. Eur J Public Health 2002;12: 171–6.
American Journal of Preventive Medicine, Volume 27, Number 1
17. MacAuley D, McCrum EE, Stott G, et al. Physical activity, physical fitness, blood pressure, and fibrinogen in the Northern Ireland health and activity survey. J Epidemiol Community Health 1996;50:258 –63. 18. Koenig W, Sund M, Doring A, Ernst E. Leisure-time physical activity but not work-related physical activity is associated with decreased plasma viscosity. Results from a large population sample. Circulation 1997;95:335–41 (see comments). 19. Greendale GA, Bodin-Dunn L, Ingles S, Haile R, Barrett-Connor E. Leisure, home, and occupational physical activity and cardiovascular risk factors in postmenopausal women. The Postmenopausal Estrogens/Progestins Intervention (PEPI) Study. Arch Intern Med 1996;156:418 –24. 20. Burchfiel CM, Sharp DS, Curb JD, et al. Physical activity and incidence of diabetes: the Honolulu Heart Program. Am J Epidemiol 1995;141:360 –8. 21. Perry IJ, Wannamethee SG, Walker MK, Thomson AG, Whincup PH, Shaper AG. Prospective study of risk factors for development of non-insulin dependent diabetes in middle aged British men. BMJ 1995;310:560 –4 (see comments). 22. Manson JE, Nathan DM, Krolewski AS, Stampfer MJ, Willett WC, Hennekens CH. A prospective study of exercise and incidence of diabetes among US male physicians. JAMA 1992;268:63–7. 23. Eriksson KF, Lindgarde F. Prevention of type 2 (non–insulin-dependent) diabetes mellitus by diet and physical exercise. The 6-year Malmo feasibility study. Diabetologia 1991;34:891–8.
24. Pan XR, Li GW, Hu YH, et al. Effects of diet and exercise in preventing NIDDM in people with impaired glucose tolerance. The Da Qing IGT and Diabetes Study. Diabetes Care 1997;20:537–44. 25. Fysisk aktivitet-håndbog om forbyggelse og behandling [Physical activity: handbook on prevention and treatment. Denmark: National Board of Health, Center for Prevention, 2003. 26. Rexrode KM, Buring JE, Manson JE. Abdominal and total adiposity and risk of coronary heart disease in men. Int J Obes Relat Metab Disord 2001;25: 1047–56. 27. Stewart AL. The reliability and validity of self-reported weight and height. J Chronic Dis 1982;35:295–309. 28. Stunkard AJ, Albaum JM. The accuracy of self-reported weights. Am J Clin Nutr 1981;34:1593–9. 29. Kuskowska-Wolk A, Karlsson P, Stolt M, Rossner S. The predictive validity of body mass index based on self-reported weight and height. Int J Obes 1989;13:441–53. 30. Lawlor DA, Bedford C, Taylor M, Ebrahim S. Agreement between measured and self-reported weight in older women. Results from the British Women’s Heart and Health Study. Age Ageing 2002;31:169 –74. 31. Wa¨ rneryd B. Levnadsfo¨ rhållanden. Återintervjustudie i underso¨ kningen av levnadsfo¨ rhållanden 1989 (ULF). [Living conditions. Reinterview in ULF 1989. Appendix 12.] Stockholm: Statistics Sweden, 1991.
Am J Prev Med 2004;27(1)
27