Physical activity and mortality in women in the Framingham Heart Study

Physical activity and mortality Framingham Heart Study

in women

in the

Men who are more active live longer, but it is not clear if the same is true for women. We monitored 1404 women aged 50 to 74 who were free of cardiovascular disease. We assessed physical activity levels and ranked subjects into quartiles. After 16 years, 319 (23%) women had died. The relative risk of mortality, compared to the least active quartile, was as follows: second quartile, 0.95 (95% confidence interval [Cl] 0.72 to 1.26); third quartile, 0.63 (95% Cl 0.46 to 0.86); most active quartile, 0.67 (95% Cl 0.48 to 0.92). The relative risks were not changed by adjustment for cardiac risk factors, chronic obstructive pulmonary disease, or cancer or by excluding all subjects who died in the first 6 years (to eliminate occult disease at baseline). There was no association between activity levels and cardiiscular morbtdity or mortality. We conclude that women who were more active lived longer; this effect was not the result of decreased cardiovascular disease. (AM HEART J 1994;126:879-84.)

Scott E. Sherman, MD, MPH, *$b Ralph B. D’Agostino, PhD,C Janet L. Cobb, MPH,C and William B. Kannel, MD, MPHd Sepulveda and Los Angeles, Calif., and Boston, Mass.

Exercise has a wide range of reported benefits, most notably reduced overall and cardiovascular mortality.rm3 It is also helpful in preventing hypertension, obesity, and coronary artery disease4 and an effective treatment for conditions ranging from hypertension5r 6 to osteoporosis 7y8 to depression. g, lo Most of these studies, however, especially those of the association between exercise and mortality, were done with middle-aged men. A few studies have examined the relation between exercise and mortality in women. Some found exercise to reduce mortalitylrl l2 or morbidity,13j r4 and others showed no effectI We monitored a group of women aged 50 to 74 for up to 16 years to see whether those who were more active lived longer. METHODS Study population.

The methods of the Framingham Study have been described in detail elsewhere.16 In brief From the “Pilot Ambulatory Care and Education Center, Veterans Affairs Medical Center, Sepulveda; the bDepartment of Medicine, University of California; CDepartment of Mathematics, Boston University; dSection of Preventive Medicine and Epidemiology, Boston University School of Medicine. Supported includes Pharma;

by the Framingham Heart Study Visiting Scholar’s support from the National Heart, Lung, and Blood Merck, Sharp, and Dohme; and Pfizer Inc.

Received

for publication

Reprint Affairs

requests: Scott Medical Center,

Copyright ‘9 1994 OOOZ-8703/94/$3.00

Dec.

13, 1993;

accepted

Dec.

E. Sherman, MD, PACE Center 16111 Plummer St., Sepulveda,

by Mosby-Year +0 4/l/58131

Book,

Inc.

Fund, which Institute; ICI

28, 1993. - OOPG, Veterans CA 91343.

Table

I. Assessmentof physical activity Activity state

Sleeping Sedentary Activity Light Moderate Heavy Total weighted

hours

Weight factor

Sample hours

Weighted hours

1.0 1.1

8 12

8.0 13.2

1.5 2.4 5.0

4 0 0

6.0 0 0 27.2

Physical activity questions were part of a 20-minute survey by a trained physician-interviewer. Subjects were asked to quantify number of hours in an average day spent in each of five activity levels. Physical activity index is calculated for sample subject.

the study was begun in 1948 with a general population sampleof 5209men and womenaged 30 to 62 living in the town of Framingham, Massachusetts.Every 2 years, subjects have had a detailed history and physical examination and a wide range of ancillary tests. At the 11th and 12th biennial examinations (1969 to 1973), physicians asked subjects how much time they spent sleeping, resting, or engagedin light, moderate, or heavy activity on a typical day. The primary goal of the Framingham Study hasbeen to determine the incidence of cardiovascular diseaseand the factors associatedwith its development. Women included in the study met three criteria: (1) they answered the physical activity questions on either the eleventh or twelfth biennial examination; (2) they were free of cardiovascular diseaseat that time; and (3) they were <75 years old when they answeredthe physical activity questions.The study sampleconsistedof 1404of the 2873 879

880

Sherman

et al.

American

November 1994 Heart Journal

Table II. Baseline characteristics, by level of physical activity Quartile Least No. of subjects Physical activity index Age (yd Systolic BP (mm Hg) Serum cholesterol (mg/dl) Cigarettes per day Metropolitan life insurance Glucose intolerance (X) LVH (%) COPD (%) Cancer ( % )

chart

weight

active

352 25-30 63.2 139.1 241.7 5.9 123.4 8.6 2.5 6.0 5.4

2nd

3rd

351 31-32 62.0 140.7 239.8 5.2 121.2 9.8 1.3 3.7 6.8

350 33-34 61.4 140.1 244.1 4.6 121.0 4.6 1.6 4.6 3.2

Values for blood pressure (Bp), cholesterol level, cigarettes/day, weight, and presence of glucose intolerance, obstructive pulmonary disease (COPD), and cancer are adjusted for subject’s age. NS, Not significant.

Table Ill. Age-adjusted rates of mortality and cardiovascular diseaseat 16 years by level of physical activity

Quartile Least 2nd 3rd Most p Value

active

active for trend

Death

CVD death

CVD

280 254 177 188 0.001

81 72 54 96 0.73

248 240 257 249 0.93

These figures are number of events per low-up. CVD, Cardiovascular disease.

1000

subjects during 16 years of fol-

women enrolled in the study in 1948. Subjects were excluded for the following reasons:507had already died at the time physical activity wasassessed; 468did not answer the physical activity questions(391of thesewomenmissed both the eleventh and twelfth examinations); 276were 275 years old; 15 were lost to follow-up; and 203 already had cardiovascular disease. Health outcome measures. The primary outcomemeasure was death from all causes.Secondary outcomeswere the incidence of cardiovascular diseaseand death from it. Cardiovascular diseaseincluded coronary heart disease (angina,coronary insufficiency, myocardial infarction, and sudden or nonsudden death resulting from coronary disease), congestive heart failure, stroke and transient ischemic attack, and intermittent claudication. Sudden coronary death wasdefined as death within 1 hour of the onsetof symptomsin the absenceof other identifiable causes. Three physicians reviewed all cardiovascular diagnoses and all deaths. Agreement of all three physicians was required. Additional detail on the determination of outcomesin the Framingham study is available elsewhere.17 Statistical analyses. The reported levels of physical activity were weighted to reflect metabolic expenditure, asin

Most

active 351 35-58 60.2 138.7 242.8 5.2 119.9 7.1 1.5 4.4 2.7

left ventricular

hypertrophy

p

Value

0.0001 NS NS NS NS 0.001 NS NS 0.001 (LVH),

chronic

a previous study by Kannel and Sorlie.15The measureof physical activity wasthe sum of the weighted values, with 24 being the absolute minimum. For example, a woman who usually slept 8 hours, wassedentary for 12hours, and did 4 hours of light activity would have a scoreof 27.2 (Table I). We then ranked the women by their level of physical activity and grouped them into quartiles. Our initial hypothesis was an inverse linear association between activity levels and mortality rates. We used the log-rank test to check for differences in morbidity and mortality among the quartiles of physical activity,rs and the Kaplan-Meier estimate at the specified length of follow-up for incidence and mortality rates. We analyzed the relation betweenphysical activity levels and the study end points with Cox proportional hazards regressionmodels.We usedthe beta coefficients of the indicator variables to calculate relative risks and 95% confidence intervals, with the most sedentary quartile asa reference.We entered the following variables into the Cox models: physical activity levels, age, systolic blood pressure, total serum cholesterol, Metropolitan life insurance chart weight, cigarettes smoked per day, and the presenceor absenceof glucose intolerance, left ventricular hypertrophy, cancer (except nonmelanomaskin cancer), and chronic obstructive pulmonary disease.An additional analysislooked for an interaction between ageand physical activity and found it not to be statistically significant (p 0.68), allowing us to analyze the younger and older subjectstogether. RESULTS The baseline characteristics of these women are shown in Table II. There were some differences among the activity quartiles. Women who were more active were somewhat younger, less likely to have glucose intolerance, and were less likely to have cancer at baseline. The survival curves for each activity quartile are

Volume 128, Number 5 American Heart Journal

Sherman

et al.

881

,.

60

50 ?r 1

2

3

4

5

6

7

8

9

10

11

12

13

14

Years of Followup Fig. 1. Kaplan-Meier curve of survival by level of physical activity.

shown in Fig. 1. After 16 years of follow-up, 319 (23%) women had died. The two more active quartiles had lower mortality rates throughout essentially the entire follow-up period. Table III shows the mortality rates by activity quartile, adjusted only for age. The two more active groups had an overall mortality rate about 30 % lower than that of the two more sedentary ones. We also adjusted the results for baseline levels of cardiac risk factors and the presence of cancer or chronic obstructive pulmonary disease (Table IV). The findings were very similar to the unadjusted results. We also analyzed the data excluding all deaths during the first six years, in an attempt to control for early deaths due to subclinical disease at baseline (Table V). The results were quite similar to those presented above. The causes of death for each of the activity quartiles are shown in Table VI. There was not a large difference in the rates of cardiovascular death among the activity quartiles. The largest trends were seen for deaths due to cancer, other causes, and unknown causes.

DISCUSSION

In our study, the 16-year mortality rate of active women was two thirds that of more sedentary women. The results were unchanged by controlling for major confounding variables. Several previous studies have examined the relation between physical activity and mortality in women. Salonen et a1.13monitored 3688 women aged 35 to 59 for approximately 6 years. After adjustments were made for age and baseline levels of cholesterol, blood pressure, body mass index, and cigarettes per day, women who were more active at work had a lower risk of having an acute myocardial infarction or a cerebrovascular accident. Unfortunately, that study used one-sided t tests and 90 % confidence intervals, making it more difficult to compare with other studies. It also included women both with and without cardiovascular disease, making it a much different study population. Brunner et all4 studied physical activity at work in relation to cardiovascular disease among men and women on Israeli kibbutzim. They found a higher rate of ischemic heart disease in sedentary female

002

Sherman et al.

American

Table IV. Sixteen-year multivariate

relative risk of mortality

and cardiovascular

Death Quartile Least 2nd 3rd Most

disease by level of physical activity

CVD death

RR

95% CI

RR

November 1994 Heart Journal

CVD

95% CI

RR

95% CI

active

1.0

-

1.0

-

1.0

-

active

0.93 0.65 0.68

0.70-1.23 0.47-0.90 0.49-0.94

0.90 0.72 1.29

0.53-1.53 0.40-1.28 0.77-2.16

0.96 1.11 0.98

0.71-1.29 0.82-1.50 0.72-1.34

Risks for each quartile are relative to least-active quartile after adjustments for age, systolic blood pressure, serum cholesterol, cigarettes smoked per day, Metropolitan life insurance chart weight, and presence or absence of glucose intolerance, left ventricular hypertrophy, chronic obstructive pulmonary disease, and cancer (excluding nonmelanoma skin cancer). CVD, Cardiovascular disease; RR, relative risk; CI, cotidence interval.

Table V. Sixteen-year multivariate relative risk of death and cardiovascular disease,excluding people who died within first 6 years, by level of physical activity Death Quartile Least 2nd 3rd Most

CVD death

RR

95% CI

RR

CVD

95% CI

active

1.0

-

1.0

-

active

0.91 0.63 0.66

0.66-1.25 0.44-0.90 0.46-0.95

0.93 0.78 1.50

0.51-1.70 0.41-1.48 0.85-2.66

RR 1.0 1.00 1.20 1.04

95% CI 0.73-1.37 0.88-1.63 0.76-1.43

Risks for each quartile are relative to least active quartile after adjustments for age, systolic blood pressure, serum cholesterol, cigarettes smoked per day, Metropolitan life insurance chart weight, and presence or absence of each of glucose intolerance, left ventricular hypertrophy, chronic obstructive pulmonary disease, and cancer (excluding nonmelanoma skin cancer). CVD, Cardiovascular disease; RR, relative risk; Cl, confidence interval.

Table Vi. Causesof death by level of physical activity Quartile Least active

2nd

3rd

Most active

352 107

351 90

350 62

351 60

6 5 10

12 4 5

6 4 6

9 8 7

10 35 23 18

5 35 15 14

3 21 15 7

6 17 7 6

No. of subjects

Total deaths CHD Nonsudden

Sudden Cerebrovascular accident Other

CVD

Cancer Other Unknown

These figures represent actual number of deaths in each quartile ity. CHD, Coronary heart disease; CVD, cardiovascular disease.

workers than in their

of activ-

more active colleagues. They did not, however, correct for other cardiovascular risk factors. In Kannel and Sorlie’d5 earlier analysis from Framingham, physical activity levels in women were not associated with overall mortality or carcliovascular morbidity and mortality rates. Both their study and ours ‘used Framingham Heart Study subjects, but they used the physical activity assessment from

1955 through 1956 with 14 years of follow-up. We had roughly the same length of follow-up, but the physical activity was assessed between 1969 and 1973. The earlier paper split the women into three activity groups and, unfortunately, the middle one had 71% of the subjects. The failure of the previous study to find an association between activity levels and morbidity and mortality in women is most likely the result of the study design, which yielded too few events in the top and bottom groups to be able to reach any statistical conclusion. Furthermore, the women were much younger then, so there were fewer events to analyze. In the Gothenburg study, Lapidus and Bengtssod2 monitored 1462 middle-aged Swedish women for 12 years. Women who were more active during leisure hours had a lower risk of stroke, and those who were more active at work had a lower overall mortality rate. In the most convincing study to date, Blair et al.ll monitored 3120 women for an average of 8 years and assessed physical fitness rather than physical activity. Women who were more physically fit lived longer, even after adjustments were made for baseline differences in levels of cardiac risk factors. The more fit women had significantly fewer cancer deaths, and there was a trend toward fewer cardiovascular deaths. We found fewer cancer deaths in the more active

Volume 128, Number 5 American Heart Journal

women, but we did not find an association between activity levels and cardiovascular deaths. The causes of death for our subjects merit further discussion. We found no association between the level of physical activity and cardiovascular morbidity and mortality rates. Although studies have consistently shown that more active men are less likely to develop heart disease, the more rigorous studies in women have not shown an effect on the incidence of heart disease. It is unclear whether there is really no effect on cardiovascular morbidity or whether there is an effect but the morbidity rate is not high enough for the effect to be measured. We did find fewer deaths from cancer, even after adjusting for baseline differences among the quartiles. Although the pathophysiologic mechanism for such an association is not clear, other studies have also shown a similar relation.‘i, lgw21There appeared to be fewer deaths classified as “other” or “unknown” in the more active quartiles. Unfortunately, no further information is available on deaths in these two categories. Our study has three basic limitations. One is that it is a cohort study, not a randomized trial. We showed that people who said they were more active tended to live longer. It is quite possible that the sedentary group was less active because they were sicker, with the higher mortality rate the result of greater comorbidity rather than lack of physical activity. We attempted to control for this in two different ways. First, we adjusted the data for baseline factors that might affect mortality: age, weight, blood pressure, cholesterol level, smoking history, glucose intolerance, left ventricular hypertrophy, cancer, and chronic obstructive pulmonary disease. Controlling for all these factors did not affect the benefit seen in the two most active quartiles. The second attempt to remove the effect of greater baseline illness in the sedentary group was a multivariate analysis, with all deaths in the first 6 years excluded. This should have helped to remove baseline differences in physical activity resulting from subclinical cardiac disease, cancer, or other problems, This analysis also did not attenuate the beneficial effect of exercise. These two analyses suggest that the lower mortality rates seen in the more active group was due to exercise rather than an unmeasured confounding factor. The second limitation of the study lies in the assessment of physical activity, and there are two potential problems. First, the questions we used have not been formally validated, although the clearly delineated levels of activity (ranging in five steps from sleep to heavy activity) constitute a broad continuum

Sherman

et al.

883

and provide face validity. Furthermore, the measure has also been used in prior studies to show an association between physical activity and overall morbidity and mortality rates.15q 22,23 Second, surveys of physical activity levels are known to be only moderately accurate,24 especially shorter ones such as this. Inaccuracies in assessments of physical activity, however, tend to minimize any true association, thereby strengthening our findings. Finally, it is possible that the apparent protective effect of being more active is the result of some factor we did not measure, such as socioeconomic status or education level. Some of the effect could be mediated through these confounders. In conclusion, we found that women who were more active had a lower overall mortality rate. This finding supports previous studies with similar findings and suggests that the studies that found no benefit might not have had sufficient sample size or follow-up time. REFERENCES 1. Morris JN, Everitt MG, Pollard R, Chave SPW, Semmence AM. Vigorous exercise in leisure-time: Protection against coronary heart disease. Lancet 1980;2:1207-10. 2. Paffenbarger RS Jr, Hyde RT, Wing AL, Hsieh C-C. Physical activity, all-cause mortality, and longevity of college alumni. N Engl J Med 1986;314:605-13. 3. Leon AS, Connett J, Jacobs DR Jr, Rauramaa R. Leisure-time physical activity levels and risk of coronary heart disease and death: the Multiple Risk Factor Intervention Trial. JAMA 1987;258:2388-95. 4. Harris SS, Caspersen CJ, DeFriese GH, E&es EH Jr. Physical activity counseling for healthy adults as a primary preventive intervention in the clinical setting: report for the U.S. Preventive Services Task Force. JAMA 1989;261:3590-8. 5. Hagberg JM, Goldring D, Heath GW. Effect of exercise training on the hlood pressure and hemodynamic features of hypertensive adolescents. Am J Cardiol 1983;52:763-8. 6. Roman 0, Camuzsi AL, Villalon E, Klenner C. Physical training program in arterial hypertension: a long-term prospective follow-up. Cardiology 1981;67:230-43. I. Smith EL, Gilligan C, McAdam M, Ensign CP. Smith PE. Deterring hone loss by exercise intervention in premenopausal and postmenopausal women. Calcif Tissue Int X%39,44:312-21. 8. Dalsky GP, Stocke KS, Ehsani AA, Slatopolaky E, Lee WC, Birge SJ Jr. Weight-bearing exercise training and lumbar hone mineral content in postmenopausal women. Ann Intern Med 1988:108:824-8. 9. Doyne EJ, Ossip-Klein DJ, Bowman ED, Osborn KM, McDougall-Wilson IB, Neimeyer RA. Running versus weight lifting in the treatment of depression. J Consult Clin Psycho1 1987;55:748-54. 10. Carney RM, Templeton B, Hong BA, Harter HR, Hagberg JM, Schechtman KB, Goldberg AP. Exercise training reduces depression and increases the performance of pleasant activities in hemodialysis patients. Nephron 198’7;47:194-8. 11. Blair SN, Kohl HW III, Paffenbarger RS Jr, Clark DG, Cooper KH, Gibbons LW. Physical fitness and all-cause mortality: a prospective study of healthy men and women. JAMA 1989;262:2395-461. 12. Lapidus L, Bengtsson C. Socioeconomic factors and physical activity in relation to cardiovascular disease and death: a 12-year follow-up of participants in a population study of women in Gothenberg, Sweden. Br Heart J 1986;55:295-301. 13. Salonen JT, Puska P, Tuomilehto J. Physical activity and risk of myocardial infarction, cerebral stroke and death. Am J Epidemiol 1982:115:526-37.

Antimisiaris

et al.

14. Brunner D, Manelis G, Modan M, Levin S. Physical activity at work and the incidence of myocardial infarction, angina pectoris and death due to ischemic heart disease: an epidemiologic study in Israeli collective settlements (kibbutzim). J Chron Dis 1974;27:217-33. 15. Kannel WB, Sorlie P. Some health benefits of physical activity: the Framingham Study. Arch Intern Med 1979;139:857-61. 16. Dawber TR. The Framingham study: the epidemiology of atherosclerotic disease. Cambridge, Mass.: Harvard University Press, 1980. 17. Stokes J III, Kannel WB, Wolf PA, Cupples LA, D’Agostino RB. The relative importance of selected risk factors for various manifestations of cardiovascular disease among men and women from 35 to 64 years old: 30 years of follow-up in the Framingham Study. Circulation 1987;75(suppl V):V-65-73. 18. Lawless JF. Statistical models and methods for lifetime data. New York: J. Wiley and Sons, 1982. 19. Whittemore AS, Wu-Williams AH, Lee M, Zheng S, Gallagher RP, Jiao DA, Zbou L, Wang XH, Chen K, Jung D. Diet, physical activity, and

American

20.

21. 22.

23.

24.

November 1994 Heart Journal

colorectal cancer among Chinese in North America and China. J Nat1 Cancer Inst 1990;82:915-26. Ballard-Barbash R, Schatzin A, Albanes D, Schiffman MH, Kreger BE, Kannel WB, Anderson KM, Helsel WE. Physical activity and risk of large bowel cancer in the Framingham study. Cant Res 1990;50:3610-3. Albanes D, Blair A, Taylor PR. Physical activity and risk of cancer in the NHANES I population. Am J Public Health 1989;79:744-50. Garcia-Palmieri MR, Costas R Jr, Cruz-Vidal M, Sorlie PD, Havlik RJ. Increased physical activity: a protective factor against heart attacks in Puerto Rico. Am J Cardiol 1982;50:749-55. Kannel WB, Belanger A, D’Agostino RB, Israel I. Physical activity and physical demand on the job and risk of cardiovascular disease and death: the Framingham Study. AM HEART J 1986;112:820-5. Haskell WL, Leon AS, Caspersen CJ, Froelicher VF, Hagberg JM, Harlan W, Holloszy JO, Regensteiner JG, Thompson PD, Washburn RA. Cardiovascular benefits and assessment of physical activity and physical fitness in adults. Med Sci Sports Exert 1992;24(6 suppl):S201-20.

Effects of amiodarone on the circadian rhythm and power spectral changes of heart rate and QT interval: Significance for the control of sudden cardiac death Effects of chronic amiodarone therapy on the circadian rhythmic&y and power spectral changes of heart rate and QT intervals from Holter recordings were evaluated in three groups of patients: group 1 baseline (n = 10); group 2, treated for 3 to 6 months (n = 11); and group 3, treated for >l year (n = 13). Amiodarone reduced heart fate, which reached steady state at 3 to 6 months; bradycardia was evident during the entire 24 hours. The corrected QT (QT,) interval increased as a function of treatment duration. It was 467 _t 39, 530 -+ 26 (p < O.OOl), and 561 k 36 (p < 0.0002) msec for groups 1, 2, and 3, after 6 months, respectively. The circadian rhythmicity of QT, was abolished in group 3. Power spectral analysis showed a tendency for amiodarone to reduce both R-R and QT interval variabilities, suggesting inhibition of autonomic control on the heart by the drug. The effectiveness of amiodarone against ventricular arrhythmias may result in part from the sustained bradycardia in concert with continuous uniform prolongation of myocardial repolarization. (AM HEART J 1994;128:884-91.)

Michael Antimisiaris, MD,a Jonnalagedda S. M. Sarma, PhD,” Mary P. Schoenbaum, RN, MS,” Param P. Sharma, MD,a K. Venkataraman, and Bramah N. Singh, MD, DPhil, with the statistical assistance of Peter Christenson, PhDb Los Angeles and Duarte, Calif. From the BDivision of Cardiology, Veterans Affairs Medical Center of West Los Angeles and BCity of Hope National Medical Center, Duarte, and the aDepartments of Medicine bBiomathematics, University of California, Los Angeles, School of Medicine. Suunorted in Dart bv mants from the Veterans Administration Research Se;&e (Washington, b. C.) and the American Heart Association, the Greater Los Angeles Affiliate.

884

Received

for publication

MD,a

Dec. 20, 1993; accepted Feb. 1,1994.

Reprint requests: Bramah N. Singh, MD, Division of Cardiology 691/111E, Veterans Administration Medical Center, Wilshire and Sawtelle Boulevards, Los Angeles, CA 90073. Copyright

Q 1994 by Mosby-Year

0002-8703/94/$3.00

+ 0

4/l/68133

Book, Inc.