- Email: [email protected]
Determinants of disability in older coronary patients
Determinants of disability in older coronary patients Philip A. Ades, MD, FACC, Patrick D. Savage, MS, Marc D. Tischler, MD, FACC, Eric T. Poehlman, PhD, Justine Dee, RPT, and Joelyn Niggel, RN Burlington, Vt
Background Patient-reported physical function is a major component of disability determinations and an important contributor to health-related quality of life. Prior studies of coronary disability have shown a surprisingly poor correlation between real-life activity profile and exercise capacity measured on the treadmill. The goal of the current investigation was to evaluate the relative importance of medical factors, sex, fitness-related measures, and psychologic factors as determinants of patient-reported physical function score in older persons with established coronary heart disease (CHD).
Methods Determinants of disability were studied in 51 community-dwelling patients >65 years old (71 ± 5 years, range 65-83 years) with established chronic CHD. Patient-reported physical function score (scaled 0-100) was measured by the Medical Outcomes Study Short Form physical function section. Independent variables included clinical and demographic data, treadmill testing, rest and exercise echocardiography, measures of body composition, strength, aerobic fitness, and a depression score.
Results Patients with a diagnosis of myocardial infarction had a lower physical function score than did patients with other CHD diagnoses (68 ± 19 vs 82 ± 22, P < .05). Univariate predictors of patient-reported physical function score included peak aerobic capacity (R = 0.62), treadmill test duration (R = 0.61), depression score (R = –0.60), handgrip strength (R = 0.42), and comorbidity score (R = –0.39). Peak aerobic capacity (R2 = 0.38) and depression score (cumulative R2 = 0.60) were the best independent predictors of physical function. Women had lower physical function scores than men (64 ± 22 vs 78 ± 20, P < .05) despite a similar age, diagnostic distribution, depression score, and comorbidity score. Resting left ventricular ejection fraction was not a predictor of physical function score. Conclusions Peak aerobic capacity and depression score were the best independent predictors of patient-reported physical function score in older coronary patients. These data focus on the potential for exercise training and treatment of mental depression to prevent and treat coronary disability in older coronary patients. (Am Heart J 2002;143:151-6.) Patient-reported physical functioning is a major component of disability determinations and of the need for disability-related services such as visiting nurses, household help, and Meals on Wheels.1 It is also a major component of health-related quality of life.2,3 In a study of middle-aged men with coronary heart disease (CHD) there was only a poor correlation between exercise capacity measured on the treadmill and the performance of practical activities of daily living.4 For many activities, the patients’ perception of their cardiac limitations was a better predictor of physical capacity than was the presence of cardiac symptoms. From the Division of Cardiology, Department of Medicine, University of Vermont College of Medicine, Burlington, Vt. Supported by NIA grant No. RO1AG15114-01 (P. A. A.) and General Clinical Research Center of the University of Vermont College of Medicine grant No. RR109. Submitted April 3, 2001; accepted August 3, 2001. Reprint requests: Philip A. Ades, MD, McClure 1, Cardiology, Fletcher-Allen Health Care, MCHV Campus, Burlington, VT 05401. E-mail: [email protected] Copyright © 2002 by Mosby, Inc. 0002-8703/2002/$35.00 + 0 4/1/119379 doi:10.1067/mhj.2002.119379
Compared with middle-aged patients with CHD, older patients have a diminished exercise capacity and higher rates of disability and mobility limitations.5-7 In the Framingham Disability Study, higher rates of disability and mobility limitations were noted with advancing age for patients diagnosed with coronary artery disease or congestive heart failure. Disability rates were particularly high with advancing age for patients with angina pectoris and in women. For example, disability rates in men and women aged 70 to 88 years with CHD were 49% and 79%, respectively, compared with rates of 9% and 25%, respectively, in men and women aged 55 to 69 years without CHD.6 The goals of medical care and rehabilitation of older coronary patients are to improve physical functioning and quality of life and to extend disability-free survival.8 The factors that are associated with disability in older coronary patients, however, have not been well studied. The goal of the current study was to use sophisticated methods such as rest and exercise echocardiography, dual energy x-ray absorptiometry (DEXA), measures of body composition, isokinetic dynamometer measures of muscular strength, and psychologic evaluations to com-
American Heart Journal January 2002
152 Ades et al
Table I. Characteristics of the study population by sex
No. Age (y) Peak V02 (mL/kg/m) Left ventricular ejection fraction (%) Body weight (kg) Body fat (%) Lean mass (kg) MOS physical function score Depression score (0-15) Comorbidity score (0-5) Handgrip strength (pounds)
Overall
Male
Female
P value (male vs female)
51 70.6 ± 5.3 19 ± 6 52 ± 13 76.8 ± 14 27 ± 7 53 ± 10 74 ± 21 1.9 ± 2 1.3 ± 1.1 35 ± 10
31 70.5 ± 5.0 21 ± 6 53 ± 13 79.7 ± 14 25 ± 6 57 ± 9 78 ± 20 1.9 ± 2.2 1.2 ± 1.2 38 ± 8
20 71.0 ± 6.4 14 ± 4 50 ± 16 67.4 ± 10 34 ± 7 41 ± 4 64 ± 22 1.8 ± 2.2 1.7 ± 1.0 23 ± 5
NS .005 NS .02 .002 .0001 .026 NS NS <.0001
NS, Not significant; MOS, Medical Outcomes Study, scaled 0-100.8
prehensively study medical, demographic, and psychosocial correlates of patient-reported disability in older coronary patients. We hypothesized that perceptual and psychologic factors would be as important as fitness-related factors in determining patient-reported measures of physical function. The elucidation of potentially modifiable determinants of physical functioning allows for the design and study of clinical interventions aimed at treating and preventing coronary disability in the elderly. As the size of the elderly population with CHD expands over the next several decades, interventions that successfully improve physical function in this population will have important public health implications.
Methods Study population Study volunteers were recruited from the Fletcher-Allen Health Care Cardiology Clinics, which include 15 cardiologists, and by newspaper advertisements in the Burlington, Vt, community. All subjects signed an informed consent document and the protocol was approved by the Human Subjects Committee of the University of Vermont College of Medicine. Subjects were ambulatory, community-dwelling, aged ≥65 years, not demented, and diagnosed with definite CHD; had not been hospitalized for a cardiac event within 6 months; and had never participated in cardiac rehabilitation. A total of 51 patients with chronic CHD, 31 men and 20 women, participated in this study. On average, these patients had a long-standing clinical history of CHD with a mean duration of 10.3 ± 5.3 years; therefore most patients had a history of multiple, overlapping cardiac diagnoses. Study patients described cardiac diagnoses of myocardial infarction (n = 23), coronary bypass surgery (n = 26), and percutaneous revascularization (n = 14), and 27 had been hospitalized for unstable angina. There was no difference in diagnostic distribution by sex. Mean age of the subjects was 70.6 ± 5.3 years (range 65-83 years). Mean resting left ventricular ejection fraction measured by 2-dimensional echocardiography was 52% ± 13% (range 26%-72%) (Table I).
Overall, 32 of 51 subjects were married, with women less likely than men to be currently married (6/20 vs 26/31, P < .05). Mean number of years of education was 12.8 ± 4.0 years with no difference by sex. Job type before retirement was described as white collar by 22 subjects and blue collar by 18 subjects, and 11 were housewives. All patients were taking at least one cardiovascular medication; 43 were taking aspirin or warfarin, 21 were taking a β-blocker, 15 were taking digitalis, 13 were taking long-acting nitrates, and 9 were taking calcium-blocking agents.
Study measures The primary dependent variable in this study was the patient-described physical function score (scaled 0-100) from the Medical Outcomes Study Short Form Questionnaire (MOSSF-36).2,3 This questionnaire describes a set of 10 activities ranging from heavy activities such as running and climbing several flights of stairs, to moderate activities such as grocery carrying or climbing 1 flight of stairs, to lighter activities such as dressing, bathing, stooping and kneeling. The physical function section is unchanged from the MOS long-form questionnaire because of the importance of physical function to overall quality of life.2 For each activity, patients describe whether they are limited a lot, limited a little, or not limited at all. Independent variables also included demographic data obtained by questionnaire and interview, which included a careful characterization of the cardiovascular symptoms and the presence of other comorbid conditions. In addition, patients underwent: 1. A symptom-limited stress test on the treadmill with a gradual-increment Naughton protocol with collection of expired gases for determination of peak aerobic capacity 2. Two-dimensional echocardiogram with Doppler analysis, both at rest, before the stress test and immediately after exercise. Echocardiographic measures included rest and exercise left ventricular volumes, ejection fraction, a qualitative analysis of exercise wall motion for detection of ischemia, eccentricity index (long axis divided by short axis), and indices of left ventricular diastolic inflow (a wave velocity, atrial filling fraction, E/A ratio, and deceleration time).
American Heart Journal Volume 143, Number 1
Ades et al 153
Table II. Univariate correlations with physical function score Variable Peak VO2 Treadmill duration Depression score Handgrip strength Comorbidity score Body fat (%) Exercise ejection fraction (%) Appendicular muscle mass Age (y) Lean body mass Rest ejection fraction (%) Weight (kg) Body mass index
R value 0.62 0.61 –0.60 0.42 –0.40 –0.34 0.33 0.32 –0.27 0.19 0.17 0.15 0.14
R2
value
0.38 0.38 0.27 0.17 0.16 0.11 0.11 0.10 0.07 0.04 0.03 0.02 0.02
Figure 1
P value .0001 .0001 .0001 .004 .007 .02 .05 .03 .07 .20 .26 .33 .36
Regression plot of peak aerobic capacity versus physical function score. R = 0.62, P < .001. 3. DEXA, used to determine body composition with measurement of fat mass, fat-free mass, appendicular muscle mass, and bone mineral content. Appendicular skeletal muscle mass was measured with DEXA with use of the model of Heymsfield et al.9 In this model, it is assumed that the extremities (arms and legs) are composed of skeletal mass, fat mass, and skeletal muscle mass. Thus the fat-free, bonefree tissue mass for each extremity represents skeletal muscle mass in that bone marrow, skin, and associated subcutaneous tissue contribute a negligible amount to the total mass of each extremity. Appendicular skeletal muscle mass estimated with this model has shown excellent agreement with skeletal muscle mass measured by computed tomography.10 In our laboratory, the coefficient of variation for repeat determinations in 7 women was 1% for fat-free tissue mass of the arms and 1% for fat-free tissue mass of the legs. 4. Handgrip strength measurement with a handgrip dynamometer (Jamar, Bolingbrook, Ill) 5. Isokinetic and isometric leg torque measurements taken with an isokinetic dynamometer (Lido Active Isokinetic Multi-joint System, Chattanooga Group, Hixson, Tex) 6. Characterization by a depression score by use of the Geriatric Depression Questionnaire with measurement of a depression score scaled 0 to 15 where a higher number signifies more depressive symptoms.11 7. Characterization by a comorbidity score that we developed and have used in the past.12 This score is determined by the presence of diabetes, peripheral vascular disease, cerebrovascular disease, chronic obstructive lung disease, or arthritis. If a comorbid condition is present, it is quantified by severity as follows: 0 indicates no impact on exercise capacity; 1 indicates present and had an impact on exercise response. A total comorbidity score ranging from 0 to 5 was determined for each patient.
Statistical analysis Linear regression analysis and stepwise multivariate regression analysis were used to determine relationships between study variables. Results between men and women were analyzed by nonpaired t tests. Level of significance was P < .01
for univariate regression analyses because of the large number of analyses and P < .05 for nonpaired t tests. Statistical analyses were performed using Prophet (National Center for Research Resources/National Institutes of Health) and SAS software (SAS Institute, Cary, NC).
Results The mean self-reported physical function score of the study subjects was 74 ± 21 (range 30-100) (Table I). Several aspects of the clinical history were predictive of the physical function score. Patients who described that they are “limited by chest pain or shortness of breath” or that they limited physical activities to be “safe for their heart” had lower physical function scores than patients who denied these characteristics (both P < .01). Patients who had had a myocardial infarction had a lower physical function score than patients in the other diagnostic categories (68 ± 19 vs 82 ± 22, P < .05). There was not a difference in physical function score between all revascularized patients versus nonrevascularized patients. There were no significant correlations between job type or maximal yearly salary and physical function score. Only 6 of the 51 patients demonstrated significant symptoms of mental depression as measured by a depression score of >5 of 15.11 Univariate linear regression analysis (n = 51) demonstrated a close relationship between several medical, fitness-related, and psychologic variables and the physical function score. The strongest univariate correlations between independent variables and physical function score were for aerobic capacity (peak VO2) (R = 0.62), treadmill test duration (R = 0.61), depression score (R = –0.60), handgrip strength (R = 0.42), and comorbidity score (R = –0.40) (all P < .01) (Table II, Figures 1 and 2). Echocardiography with adequate images for analysis
American Heart Journal January 2002
154 Ades et al
Figure 2
Table III. Univariate correlations with peak aerobic capacity (liters per minute) Variable
Regression plot of depression score versus physical function score. R = –0.60, P < .0001.
was accomplished in 92% of patients (47/51) at rest and in 88% of patients (45/51) immediately after exercise. Rest ejection fraction was not a strong predictor of the physical function score (Table II). No other echocardiographic variables (left ventricular mass, A-wave velocity, E/A ratio, atrial filling fraction, eccentricity index) had a significant correlation with the physical function score. Exertional echocardiographic ischemia occurred in 13 of 45 patients with adequate imaging (8/28 men and 5/17 women, P not significant [NS]). Neither a history of exertional angina (n = 27), the presence of angina at the symptom-limited stress test performed on medication (n = 3), or the demonstration of echocardiographic ischemia (n = 13) were associated with a diminished physical function score. By stepwise multivariate analysis, peak aerobic capacity (R2 = 0.38) and depression score (cumulative R2 = 0.60) were the only independent predictors of the physical function score. Sex was not an independent predictor of the physical function score. There was not a close univariate correlation between either peak VO2, exercise duration, handgrip strength, and depression score, suggesting that these fitness measures were not important determinants of the depression score (r values range from 0.01 to –0.16, P > .30 for each). Women had a lower physical function score, that is, described themselves as being more disabled, than the male patients did, 64 ± 22 versus 78 ± 20 (P = .026). The men and women in this study were similar by age, diagnostic distribution, depression score, and comorbidity score, whereas the men had higher measures of aerobic capacity (VO2 peak), strength, and lean mass than the women did (Table I). The relationship between independent variables and the more physiologic measure of peak VO2 was also analyzed (Table III). The best univariate predictors of peak VO2 (in liters per minute) were handgrip strength (R = 0.63), fat-free mass (R = 0.63), appendicular mus-
Handgrip strength Fat-free mass Appendicular muscle mass Leg isokinetic torque Weight Rest left ventricular ejection fraction Percent body fat Exercise left ventricular ejection fraction Body mass index Age Comorbidity Score
R value
R2 value
P value
0.63 0.63 0.61 0.60 0.53 0.41
0.40 0.39 0.37 0.36 0.28 0.17
.0001 .0001 .0001 .0001 .0006 .01
–0.39 0.37
0.15 0.14
.015 .026
0.27 –0.22 –0.18
0.07 0.05 0.03
.10 .17 NS
cle mass (R = 0.61), leg strength (isokinetic torque, R = 0.60), body weight (R = 0.53), resting left ventricular ejection fraction (R = 0.41), and percent body fat (R = –0.39 ) (all P < .01, Table III). The depression score was not a significant predictor of peak VO2 (R = 0.01, P NS). By multivariate analysis, 3 variables entered into the model for peak VO2. Handgrip strength alone was the best single predictor of peak aerobic capacity, with an R2 of 0.28. When resting left ventricular ejection fraction was added to the model, the cumulative R2 was 0.36. When fat-free mass (lean body mass) was added to the model, handgrip strength dropped out of the model, with the cumulative R2 measured at 0.47.
Discussion The results of this study confirm a high rate of selfreported disability in older coronary patients because 57% of patients fell into the disabled range according to the criteria of the Framingham Disability Study.6 These results also document the importance of several measures of physical fitness (peak aerobic capacity, exercise duration on a treadmill test, and handgrip strength) on measures of self-reported physical functioning in older coronary patients. More surprising was the importance of a simple, questionnaire-based depression score in predicting self-reported physical functioning in the same population. It was a more powerful predictor of physical functioning than cardiac-based measures such as left ventricular ejection fraction or the presence of exertional angina or ischemia. It should be noted that the majority of patients were not having significant depressive symptoms but that, rather, the depression score correlated with physical functioning throughout the range of scores, including patients who were not clinically depressed. The depression score was not closely correlated with other univariate predictors of physical function score such as exercise duration on
American Heart Journal Volume 143, Number 1
the treadmill, peak VO2, or handgrip strength, suggesting that it is not simply acting through an association with fitness levels but that there is a more direct relationship with physical functioning such as by mood state or level of arousal. These data are supported by a recent study from our laboratory where change in depression score was found to be the best correlate of improvement in physical function score in patients of all ages undergoing cardiac rehabilitation after a hospitalization for an acute coronary event.12 Most of the powerful predictors of physical function that were determined in the study population are modifiable. The 4 most powerful univariate predictors of physical function score, peak aerobic capacity, treadmill test duration, depression score, and handgrip strength, have each been demonstrated in younger patients after a coronary event to improve after participation in cardiac rehabilitation.12-14 Data from The Women’s Health and Aging Study and the Cardiovascular Health Study, both performed in broad samples of older subjects, confirm these findings in healthy elders because measures of strength and high body fat were found to be significant predictors of disability.15,16 It remains to be determined, however, whether specific interventions targeting these variables in older coronary patients with chronic coronary heart disease will be successful. Multiple studies have demonstrated the feasibility and effectiveness of exercise conditioning for older coronary patients in the cardiac rehabilitation setting.5,8,17-21 A recent report demonstrates that outpatient cardiac rehabilitation after a coronary event results in an improvement of physical functioning in patients spanning a wide age range.12 Patients in the current study differed from cardiac rehabilitation populations in that they were not recovering from an acute coronary event and the inactivity of a recent hospitalization. In noncoronary patients, the relationship between depression and physical functioning has received some study. In a randomized controlled trial of primary care outpatients with clinical depression, reported by Coulehan et al,22 it was demonstrated that treatment of depression resulted in an improvement of physical functioning compared with control subjects and that, similar to the current study, severity of depression and of general medical illness were not correlated. Furthermore, the presence of medical comorbidity was only a weak predictor of outcome, again suggesting an independent role of mental depression in determining physical functioning. The demonstration of lower physical function scores in women was also notable and is consistent with earlier studies.6,7 In that older women with CHD are far more likely than older men to be living alone, the maintenance of functional independence is a crucial aspect of their overall quality of life. The lower physical func-
Ades et al 155
tion scores in women were not due to differences in age, depression scores, measures of left ventricular function, or frequency of comorbidities compared with men, although there was a trend toward more frequent comorbidities in women (P = .21). More likely they were related to the combined effect of lower measures of aerobic endurance, peak aerobic capacity, and strength in the women versus men. It has also been demonstrated that as women pass through menopause muscle mass decreases, fat mass increases, and total physical activity decreases; this effect may be compounded by aging and inactivity in a coronary population.23 Perceptual factors may also be important because women were more likely than men to describe that they limited activities to be “safe for their heart.” In a study of middle-aged men with CHD, the patient’s own perceptions of the cardiac limitation varied for different activities, and for many activities it was a better determinant of physical capacity than was the presence of cardiac symptoms.4 It should be noted that we were unable to fully control for selection factors in recruiting subjects for this study because it would be expected that the most poorly functioning patients would be less likely to volunteer for participation. On the other hand, the range of physical function that was studied was broad, and thus representative intervariable relationships would most likely be revealed. Our data differ from those of earlier studies in that angina was neither highly prevalent nor a significant predictor of physical function; although more than half our patient cohort described a history of chronic angina, only a very small number of patients had angina at the symptom-limited stress test (tested on medications). Furthermore, neither a history of angina nor the presence of exercise-induced angina at the treadmill stress test was associated with a decreased physical function score. Even the presence of echocardiographically demonstrated ischemia did not predict a diminished physical function score. The finding that coronary revascularization did not predict physical function may have been due in part to the long duration of time since intervention. Limitations of this study include that we did not study a noncoronary population for comparison, although it has been conclusively demonstrated that the presence of coronary heart disease is an independent predictor of higher disability rates in older coronary patients.6,7 The low prevalence of angina, compared with previous studies, may be due in part to the more aggressive use of coronary revascularization procedures in contemporary populations along with an increased use of antianginal medications. Furthermore, our data and conclusions are limited to community-dwelling patients able to come for an afternoon of questionnaires and functional testing. The current study should serve as a foundation on
156 Ades et al
which to study the value of clinical interventions aimed at improving physical functioning in older coronary patients. In particular, the effectiveness and safety of exercise training protocols such as aerobic or resistance training in older patients with CHD, particularly women, will need to be evaluated along with interventions that target depression or mood state. Recent studies of resistance training in a noncoronary population of older women with disabilities demonstrate the value of such interventions on improving selected physical characteristics such as lower extremity strength, gait, walking endurance, and overall physical function.24,25 A preliminary study of home resistance training in elderly women with CHD was similarly favorable.26 For patients with clinical depression the value of exercise, counseling, or antidepressant medications on improving physical function will also need to be assessed. We thank Joshua Henkin, MD, and Toby Richman, RPT, for assistance with data collection. We also thank Martin LeWinter, MD, for his helpful comments and Diantha Howard, MS, for assistance with statistical analyses and data interpretation.
References 1. Disability evaluation under Social Security. Washington: US Department of Health and Human Services; 1992. SSA Publication No.: 05-10107. 2. Ware JE, Sherbourne CD. The Medical Outcomes Study: a 36 item short-form health survey (SF-36), 1: conceptual framework and item selection. Med Care 1992;30:473-83. 3. Stewart AL, Hays RD, Ware JE. The MOS short-form general health survey: reliability and validity in a patient population. Med Care 1988;26:724-35. 4. Neill WA, Branch LG, DeJong G. Cardiac disability—the impact of coronary disease on patients’ daily activities. Arch Intern Med 1981;145:1642-7. 5. Ades PA, Grunvald MH. Cardiopulmonary exercise testing before and after conditioning in older coronary patients. Am Heart J 1990;120:585-9. 6. Pinsky JL, Jette AM, Branch LG, et al. The Framingham Disability Study: relationship of various coronary heart disease manifestations to disability in older persons living in the community. Am J Public Health 1990;80:1363-8. 7. Chirikos T, Nickel J. Socioeconomic determinants of disablement from chronic disease episodes. Soc Sci Med 1986;22:1329-35. 8. Ades PA. Cardiac rehabilitation in older coronary patients. J Am Geriatr Soc 1999;47:98-105. 9. Heymsfield S, Smith R, Aulet M, et al. Appendicular skeletal muscle mass: measurement by dual-photon absorptiometry. Am J Clin Nutr 1990;52:214-8.
American Heart Journal January 2002
10. Wang Z, Visser M, Ma R, et al. Skeletal muscle mass: evaluation of neutron activation and dual-energy x-ray absorptiometry methods. J Appl Physiol 1996;80:824-31. 11. Yesavage JA, Brink TL, Rose TL. Development and validation of a geriatric depression screening scale—a preliminary report. J Psychiatr Res 1983;17:37-49. 12. Ades PA, Maloney A, Savage P, et al. Physical function in coronary patients: effect of cardiac rehabilitation. Arch Intern Med 1999; 159:2357-60. 13. Stern MJ, Gorman PA, Kaslow P. The group counseling vs exercise therapy study: a controlled intervention with subjects following myocardial infarction. Arch Intern Med 1983;143:1719-25. 14. Wenger NK, Froehlicher ES, Smith LK, et al. Cardiac rehabilitation: clinical practice guidelines. Bethesda (MD): Agency for Health Care Policy and Research and the National Heart, Lung, and Blood Institute; 1995. Report No.: AHCPR Publication No. 96-0672. 15. Rantanen T, Guralnik J, Sakari-Rantala R, et al. Disability, physical activity, and muscle strength in older women: the Women’s Health and Aging Study. Arch Phys Med Rehabil 1999;80:130-5. 16. Visser M, Langlois J, Guralnik J, et al. High body fatness, but not low fat-free mass, predicts disability in older men and women: the Cardiovascular Health Study. Am J Clin Nutr 1998;68:584-90. 17. Williams MA, Maresh CM, Esterbrooks DJ, et al. Early exercise training in patients older than age 65 years compared with that in younger patients after acute myocardial infarction of coronary bypass grafting. Am J Cardiol 1985;55:263-6. 18. Ades PA, Hanson JS, Gunther PG, et al. Exercise conditioning in the elderly coronary patient. J Am Geriatr Soc 1987;35:121-4. 19. Ades PA, Waldmann ML, Poehlman ET, et al. Exercise conditioning in older coronary patients: submaximal lactate response and endurance capacity. Circulation 1993;88:572-7. 20. Ades PA, Waldmann ML, Meyer WL, et al. Skeletal muscle and cardiovascular adaptations to exercise conditioning in older coronary patients. Circulation 1996;94:323-30. 21. Lavie CJ, Milani RV, Littman AB. Benefits of cardiac rehabilitation and exercise training in secondary coronary prevention in the elderly. J Am Coll Cardiol 1993;22:678-83. 22. Coulehan J, Schulberg H, Block M, et al. Treating depressed primary care patients improves their physical, mental, and social functioning. Arch Intern Med 1997;157:1113-20. 23. Poehlman E, Toth M, Gardner A. Changes in energy balance and body composition at menopause: a controlled longitudinal study. Ann Intern Med 1995;123:673-5. 24. Jette A, Lachman M, Giorgetti M, et al. Exercise—it’s never too late: the Strong-for-Life Program. Am J Public Health 1999;89:66-72. 25. Ades PA, Ballor DL, Ashikaga K, et al. Weight training improves walking endurance in healthy elderly persons. Ann Intern Med 1996;124:568-72. 26. King P, Savage P, Ades PA. Home resistance training in an elderly woman with coronary heart disease. J Cardiopulm Rehabil 1999; 20:126-9.
Background Patient-reported physical function is a major component of disability determinations and an important contributor to health-related quality of life. Prior studies of coronary disability have shown a surprisingly poor correlation between real-life activity profile and exercise capacity measured on the treadmill. The goal of the current investigation was to evaluate the relative importance of medical factors, sex, fitness-related measures, and psychologic factors as determinants of patient-reported physical function score in older persons with established coronary heart disease (CHD).
Methods Determinants of disability were studied in 51 community-dwelling patients >65 years old (71 ± 5 years, range 65-83 years) with established chronic CHD. Patient-reported physical function score (scaled 0-100) was measured by the Medical Outcomes Study Short Form physical function section. Independent variables included clinical and demographic data, treadmill testing, rest and exercise echocardiography, measures of body composition, strength, aerobic fitness, and a depression score.
Results Patients with a diagnosis of myocardial infarction had a lower physical function score than did patients with other CHD diagnoses (68 ± 19 vs 82 ± 22, P < .05). Univariate predictors of patient-reported physical function score included peak aerobic capacity (R = 0.62), treadmill test duration (R = 0.61), depression score (R = –0.60), handgrip strength (R = 0.42), and comorbidity score (R = –0.39). Peak aerobic capacity (R2 = 0.38) and depression score (cumulative R2 = 0.60) were the best independent predictors of physical function. Women had lower physical function scores than men (64 ± 22 vs 78 ± 20, P < .05) despite a similar age, diagnostic distribution, depression score, and comorbidity score. Resting left ventricular ejection fraction was not a predictor of physical function score. Conclusions Peak aerobic capacity and depression score were the best independent predictors of patient-reported physical function score in older coronary patients. These data focus on the potential for exercise training and treatment of mental depression to prevent and treat coronary disability in older coronary patients. (Am Heart J 2002;143:151-6.) Patient-reported physical functioning is a major component of disability determinations and of the need for disability-related services such as visiting nurses, household help, and Meals on Wheels.1 It is also a major component of health-related quality of life.2,3 In a study of middle-aged men with coronary heart disease (CHD) there was only a poor correlation between exercise capacity measured on the treadmill and the performance of practical activities of daily living.4 For many activities, the patients’ perception of their cardiac limitations was a better predictor of physical capacity than was the presence of cardiac symptoms. From the Division of Cardiology, Department of Medicine, University of Vermont College of Medicine, Burlington, Vt. Supported by NIA grant No. RO1AG15114-01 (P. A. A.) and General Clinical Research Center of the University of Vermont College of Medicine grant No. RR109. Submitted April 3, 2001; accepted August 3, 2001. Reprint requests: Philip A. Ades, MD, McClure 1, Cardiology, Fletcher-Allen Health Care, MCHV Campus, Burlington, VT 05401. E-mail: [email protected] Copyright © 2002 by Mosby, Inc. 0002-8703/2002/$35.00 + 0 4/1/119379 doi:10.1067/mhj.2002.119379
Compared with middle-aged patients with CHD, older patients have a diminished exercise capacity and higher rates of disability and mobility limitations.5-7 In the Framingham Disability Study, higher rates of disability and mobility limitations were noted with advancing age for patients diagnosed with coronary artery disease or congestive heart failure. Disability rates were particularly high with advancing age for patients with angina pectoris and in women. For example, disability rates in men and women aged 70 to 88 years with CHD were 49% and 79%, respectively, compared with rates of 9% and 25%, respectively, in men and women aged 55 to 69 years without CHD.6 The goals of medical care and rehabilitation of older coronary patients are to improve physical functioning and quality of life and to extend disability-free survival.8 The factors that are associated with disability in older coronary patients, however, have not been well studied. The goal of the current study was to use sophisticated methods such as rest and exercise echocardiography, dual energy x-ray absorptiometry (DEXA), measures of body composition, isokinetic dynamometer measures of muscular strength, and psychologic evaluations to com-
American Heart Journal January 2002
152 Ades et al
Table I. Characteristics of the study population by sex
No. Age (y) Peak V02 (mL/kg/m) Left ventricular ejection fraction (%) Body weight (kg) Body fat (%) Lean mass (kg) MOS physical function score Depression score (0-15) Comorbidity score (0-5) Handgrip strength (pounds)
Overall
Male
Female
P value (male vs female)
51 70.6 ± 5.3 19 ± 6 52 ± 13 76.8 ± 14 27 ± 7 53 ± 10 74 ± 21 1.9 ± 2 1.3 ± 1.1 35 ± 10
31 70.5 ± 5.0 21 ± 6 53 ± 13 79.7 ± 14 25 ± 6 57 ± 9 78 ± 20 1.9 ± 2.2 1.2 ± 1.2 38 ± 8
20 71.0 ± 6.4 14 ± 4 50 ± 16 67.4 ± 10 34 ± 7 41 ± 4 64 ± 22 1.8 ± 2.2 1.7 ± 1.0 23 ± 5
NS .005 NS .02 .002 .0001 .026 NS NS <.0001
NS, Not significant; MOS, Medical Outcomes Study, scaled 0-100.8
prehensively study medical, demographic, and psychosocial correlates of patient-reported disability in older coronary patients. We hypothesized that perceptual and psychologic factors would be as important as fitness-related factors in determining patient-reported measures of physical function. The elucidation of potentially modifiable determinants of physical functioning allows for the design and study of clinical interventions aimed at treating and preventing coronary disability in the elderly. As the size of the elderly population with CHD expands over the next several decades, interventions that successfully improve physical function in this population will have important public health implications.
Methods Study population Study volunteers were recruited from the Fletcher-Allen Health Care Cardiology Clinics, which include 15 cardiologists, and by newspaper advertisements in the Burlington, Vt, community. All subjects signed an informed consent document and the protocol was approved by the Human Subjects Committee of the University of Vermont College of Medicine. Subjects were ambulatory, community-dwelling, aged ≥65 years, not demented, and diagnosed with definite CHD; had not been hospitalized for a cardiac event within 6 months; and had never participated in cardiac rehabilitation. A total of 51 patients with chronic CHD, 31 men and 20 women, participated in this study. On average, these patients had a long-standing clinical history of CHD with a mean duration of 10.3 ± 5.3 years; therefore most patients had a history of multiple, overlapping cardiac diagnoses. Study patients described cardiac diagnoses of myocardial infarction (n = 23), coronary bypass surgery (n = 26), and percutaneous revascularization (n = 14), and 27 had been hospitalized for unstable angina. There was no difference in diagnostic distribution by sex. Mean age of the subjects was 70.6 ± 5.3 years (range 65-83 years). Mean resting left ventricular ejection fraction measured by 2-dimensional echocardiography was 52% ± 13% (range 26%-72%) (Table I).
Overall, 32 of 51 subjects were married, with women less likely than men to be currently married (6/20 vs 26/31, P < .05). Mean number of years of education was 12.8 ± 4.0 years with no difference by sex. Job type before retirement was described as white collar by 22 subjects and blue collar by 18 subjects, and 11 were housewives. All patients were taking at least one cardiovascular medication; 43 were taking aspirin or warfarin, 21 were taking a β-blocker, 15 were taking digitalis, 13 were taking long-acting nitrates, and 9 were taking calcium-blocking agents.
Study measures The primary dependent variable in this study was the patient-described physical function score (scaled 0-100) from the Medical Outcomes Study Short Form Questionnaire (MOSSF-36).2,3 This questionnaire describes a set of 10 activities ranging from heavy activities such as running and climbing several flights of stairs, to moderate activities such as grocery carrying or climbing 1 flight of stairs, to lighter activities such as dressing, bathing, stooping and kneeling. The physical function section is unchanged from the MOS long-form questionnaire because of the importance of physical function to overall quality of life.2 For each activity, patients describe whether they are limited a lot, limited a little, or not limited at all. Independent variables also included demographic data obtained by questionnaire and interview, which included a careful characterization of the cardiovascular symptoms and the presence of other comorbid conditions. In addition, patients underwent: 1. A symptom-limited stress test on the treadmill with a gradual-increment Naughton protocol with collection of expired gases for determination of peak aerobic capacity 2. Two-dimensional echocardiogram with Doppler analysis, both at rest, before the stress test and immediately after exercise. Echocardiographic measures included rest and exercise left ventricular volumes, ejection fraction, a qualitative analysis of exercise wall motion for detection of ischemia, eccentricity index (long axis divided by short axis), and indices of left ventricular diastolic inflow (a wave velocity, atrial filling fraction, E/A ratio, and deceleration time).
American Heart Journal Volume 143, Number 1
Ades et al 153
Table II. Univariate correlations with physical function score Variable Peak VO2 Treadmill duration Depression score Handgrip strength Comorbidity score Body fat (%) Exercise ejection fraction (%) Appendicular muscle mass Age (y) Lean body mass Rest ejection fraction (%) Weight (kg) Body mass index
R value 0.62 0.61 –0.60 0.42 –0.40 –0.34 0.33 0.32 –0.27 0.19 0.17 0.15 0.14
R2
value
0.38 0.38 0.27 0.17 0.16 0.11 0.11 0.10 0.07 0.04 0.03 0.02 0.02
Figure 1
P value .0001 .0001 .0001 .004 .007 .02 .05 .03 .07 .20 .26 .33 .36
Regression plot of peak aerobic capacity versus physical function score. R = 0.62, P < .001. 3. DEXA, used to determine body composition with measurement of fat mass, fat-free mass, appendicular muscle mass, and bone mineral content. Appendicular skeletal muscle mass was measured with DEXA with use of the model of Heymsfield et al.9 In this model, it is assumed that the extremities (arms and legs) are composed of skeletal mass, fat mass, and skeletal muscle mass. Thus the fat-free, bonefree tissue mass for each extremity represents skeletal muscle mass in that bone marrow, skin, and associated subcutaneous tissue contribute a negligible amount to the total mass of each extremity. Appendicular skeletal muscle mass estimated with this model has shown excellent agreement with skeletal muscle mass measured by computed tomography.10 In our laboratory, the coefficient of variation for repeat determinations in 7 women was 1% for fat-free tissue mass of the arms and 1% for fat-free tissue mass of the legs. 4. Handgrip strength measurement with a handgrip dynamometer (Jamar, Bolingbrook, Ill) 5. Isokinetic and isometric leg torque measurements taken with an isokinetic dynamometer (Lido Active Isokinetic Multi-joint System, Chattanooga Group, Hixson, Tex) 6. Characterization by a depression score by use of the Geriatric Depression Questionnaire with measurement of a depression score scaled 0 to 15 where a higher number signifies more depressive symptoms.11 7. Characterization by a comorbidity score that we developed and have used in the past.12 This score is determined by the presence of diabetes, peripheral vascular disease, cerebrovascular disease, chronic obstructive lung disease, or arthritis. If a comorbid condition is present, it is quantified by severity as follows: 0 indicates no impact on exercise capacity; 1 indicates present and had an impact on exercise response. A total comorbidity score ranging from 0 to 5 was determined for each patient.
Statistical analysis Linear regression analysis and stepwise multivariate regression analysis were used to determine relationships between study variables. Results between men and women were analyzed by nonpaired t tests. Level of significance was P < .01
for univariate regression analyses because of the large number of analyses and P < .05 for nonpaired t tests. Statistical analyses were performed using Prophet (National Center for Research Resources/National Institutes of Health) and SAS software (SAS Institute, Cary, NC).
Results The mean self-reported physical function score of the study subjects was 74 ± 21 (range 30-100) (Table I). Several aspects of the clinical history were predictive of the physical function score. Patients who described that they are “limited by chest pain or shortness of breath” or that they limited physical activities to be “safe for their heart” had lower physical function scores than patients who denied these characteristics (both P < .01). Patients who had had a myocardial infarction had a lower physical function score than patients in the other diagnostic categories (68 ± 19 vs 82 ± 22, P < .05). There was not a difference in physical function score between all revascularized patients versus nonrevascularized patients. There were no significant correlations between job type or maximal yearly salary and physical function score. Only 6 of the 51 patients demonstrated significant symptoms of mental depression as measured by a depression score of >5 of 15.11 Univariate linear regression analysis (n = 51) demonstrated a close relationship between several medical, fitness-related, and psychologic variables and the physical function score. The strongest univariate correlations between independent variables and physical function score were for aerobic capacity (peak VO2) (R = 0.62), treadmill test duration (R = 0.61), depression score (R = –0.60), handgrip strength (R = 0.42), and comorbidity score (R = –0.40) (all P < .01) (Table II, Figures 1 and 2). Echocardiography with adequate images for analysis
American Heart Journal January 2002
154 Ades et al
Figure 2
Table III. Univariate correlations with peak aerobic capacity (liters per minute) Variable
Regression plot of depression score versus physical function score. R = –0.60, P < .0001.
was accomplished in 92% of patients (47/51) at rest and in 88% of patients (45/51) immediately after exercise. Rest ejection fraction was not a strong predictor of the physical function score (Table II). No other echocardiographic variables (left ventricular mass, A-wave velocity, E/A ratio, atrial filling fraction, eccentricity index) had a significant correlation with the physical function score. Exertional echocardiographic ischemia occurred in 13 of 45 patients with adequate imaging (8/28 men and 5/17 women, P not significant [NS]). Neither a history of exertional angina (n = 27), the presence of angina at the symptom-limited stress test performed on medication (n = 3), or the demonstration of echocardiographic ischemia (n = 13) were associated with a diminished physical function score. By stepwise multivariate analysis, peak aerobic capacity (R2 = 0.38) and depression score (cumulative R2 = 0.60) were the only independent predictors of the physical function score. Sex was not an independent predictor of the physical function score. There was not a close univariate correlation between either peak VO2, exercise duration, handgrip strength, and depression score, suggesting that these fitness measures were not important determinants of the depression score (r values range from 0.01 to –0.16, P > .30 for each). Women had a lower physical function score, that is, described themselves as being more disabled, than the male patients did, 64 ± 22 versus 78 ± 20 (P = .026). The men and women in this study were similar by age, diagnostic distribution, depression score, and comorbidity score, whereas the men had higher measures of aerobic capacity (VO2 peak), strength, and lean mass than the women did (Table I). The relationship between independent variables and the more physiologic measure of peak VO2 was also analyzed (Table III). The best univariate predictors of peak VO2 (in liters per minute) were handgrip strength (R = 0.63), fat-free mass (R = 0.63), appendicular mus-
Handgrip strength Fat-free mass Appendicular muscle mass Leg isokinetic torque Weight Rest left ventricular ejection fraction Percent body fat Exercise left ventricular ejection fraction Body mass index Age Comorbidity Score
R value
R2 value
P value
0.63 0.63 0.61 0.60 0.53 0.41
0.40 0.39 0.37 0.36 0.28 0.17
.0001 .0001 .0001 .0001 .0006 .01
–0.39 0.37
0.15 0.14
.015 .026
0.27 –0.22 –0.18
0.07 0.05 0.03
.10 .17 NS
cle mass (R = 0.61), leg strength (isokinetic torque, R = 0.60), body weight (R = 0.53), resting left ventricular ejection fraction (R = 0.41), and percent body fat (R = –0.39 ) (all P < .01, Table III). The depression score was not a significant predictor of peak VO2 (R = 0.01, P NS). By multivariate analysis, 3 variables entered into the model for peak VO2. Handgrip strength alone was the best single predictor of peak aerobic capacity, with an R2 of 0.28. When resting left ventricular ejection fraction was added to the model, the cumulative R2 was 0.36. When fat-free mass (lean body mass) was added to the model, handgrip strength dropped out of the model, with the cumulative R2 measured at 0.47.
Discussion The results of this study confirm a high rate of selfreported disability in older coronary patients because 57% of patients fell into the disabled range according to the criteria of the Framingham Disability Study.6 These results also document the importance of several measures of physical fitness (peak aerobic capacity, exercise duration on a treadmill test, and handgrip strength) on measures of self-reported physical functioning in older coronary patients. More surprising was the importance of a simple, questionnaire-based depression score in predicting self-reported physical functioning in the same population. It was a more powerful predictor of physical functioning than cardiac-based measures such as left ventricular ejection fraction or the presence of exertional angina or ischemia. It should be noted that the majority of patients were not having significant depressive symptoms but that, rather, the depression score correlated with physical functioning throughout the range of scores, including patients who were not clinically depressed. The depression score was not closely correlated with other univariate predictors of physical function score such as exercise duration on
American Heart Journal Volume 143, Number 1
the treadmill, peak VO2, or handgrip strength, suggesting that it is not simply acting through an association with fitness levels but that there is a more direct relationship with physical functioning such as by mood state or level of arousal. These data are supported by a recent study from our laboratory where change in depression score was found to be the best correlate of improvement in physical function score in patients of all ages undergoing cardiac rehabilitation after a hospitalization for an acute coronary event.12 Most of the powerful predictors of physical function that were determined in the study population are modifiable. The 4 most powerful univariate predictors of physical function score, peak aerobic capacity, treadmill test duration, depression score, and handgrip strength, have each been demonstrated in younger patients after a coronary event to improve after participation in cardiac rehabilitation.12-14 Data from The Women’s Health and Aging Study and the Cardiovascular Health Study, both performed in broad samples of older subjects, confirm these findings in healthy elders because measures of strength and high body fat were found to be significant predictors of disability.15,16 It remains to be determined, however, whether specific interventions targeting these variables in older coronary patients with chronic coronary heart disease will be successful. Multiple studies have demonstrated the feasibility and effectiveness of exercise conditioning for older coronary patients in the cardiac rehabilitation setting.5,8,17-21 A recent report demonstrates that outpatient cardiac rehabilitation after a coronary event results in an improvement of physical functioning in patients spanning a wide age range.12 Patients in the current study differed from cardiac rehabilitation populations in that they were not recovering from an acute coronary event and the inactivity of a recent hospitalization. In noncoronary patients, the relationship between depression and physical functioning has received some study. In a randomized controlled trial of primary care outpatients with clinical depression, reported by Coulehan et al,22 it was demonstrated that treatment of depression resulted in an improvement of physical functioning compared with control subjects and that, similar to the current study, severity of depression and of general medical illness were not correlated. Furthermore, the presence of medical comorbidity was only a weak predictor of outcome, again suggesting an independent role of mental depression in determining physical functioning. The demonstration of lower physical function scores in women was also notable and is consistent with earlier studies.6,7 In that older women with CHD are far more likely than older men to be living alone, the maintenance of functional independence is a crucial aspect of their overall quality of life. The lower physical func-
Ades et al 155
tion scores in women were not due to differences in age, depression scores, measures of left ventricular function, or frequency of comorbidities compared with men, although there was a trend toward more frequent comorbidities in women (P = .21). More likely they were related to the combined effect of lower measures of aerobic endurance, peak aerobic capacity, and strength in the women versus men. It has also been demonstrated that as women pass through menopause muscle mass decreases, fat mass increases, and total physical activity decreases; this effect may be compounded by aging and inactivity in a coronary population.23 Perceptual factors may also be important because women were more likely than men to describe that they limited activities to be “safe for their heart.” In a study of middle-aged men with CHD, the patient’s own perceptions of the cardiac limitation varied for different activities, and for many activities it was a better determinant of physical capacity than was the presence of cardiac symptoms.4 It should be noted that we were unable to fully control for selection factors in recruiting subjects for this study because it would be expected that the most poorly functioning patients would be less likely to volunteer for participation. On the other hand, the range of physical function that was studied was broad, and thus representative intervariable relationships would most likely be revealed. Our data differ from those of earlier studies in that angina was neither highly prevalent nor a significant predictor of physical function; although more than half our patient cohort described a history of chronic angina, only a very small number of patients had angina at the symptom-limited stress test (tested on medications). Furthermore, neither a history of angina nor the presence of exercise-induced angina at the treadmill stress test was associated with a decreased physical function score. Even the presence of echocardiographically demonstrated ischemia did not predict a diminished physical function score. The finding that coronary revascularization did not predict physical function may have been due in part to the long duration of time since intervention. Limitations of this study include that we did not study a noncoronary population for comparison, although it has been conclusively demonstrated that the presence of coronary heart disease is an independent predictor of higher disability rates in older coronary patients.6,7 The low prevalence of angina, compared with previous studies, may be due in part to the more aggressive use of coronary revascularization procedures in contemporary populations along with an increased use of antianginal medications. Furthermore, our data and conclusions are limited to community-dwelling patients able to come for an afternoon of questionnaires and functional testing. The current study should serve as a foundation on
156 Ades et al
which to study the value of clinical interventions aimed at improving physical functioning in older coronary patients. In particular, the effectiveness and safety of exercise training protocols such as aerobic or resistance training in older patients with CHD, particularly women, will need to be evaluated along with interventions that target depression or mood state. Recent studies of resistance training in a noncoronary population of older women with disabilities demonstrate the value of such interventions on improving selected physical characteristics such as lower extremity strength, gait, walking endurance, and overall physical function.24,25 A preliminary study of home resistance training in elderly women with CHD was similarly favorable.26 For patients with clinical depression the value of exercise, counseling, or antidepressant medications on improving physical function will also need to be assessed. We thank Joshua Henkin, MD, and Toby Richman, RPT, for assistance with data collection. We also thank Martin LeWinter, MD, for his helpful comments and Diantha Howard, MS, for assistance with statistical analyses and data interpretation.
References 1. Disability evaluation under Social Security. Washington: US Department of Health and Human Services; 1992. SSA Publication No.: 05-10107. 2. Ware JE, Sherbourne CD. The Medical Outcomes Study: a 36 item short-form health survey (SF-36), 1: conceptual framework and item selection. Med Care 1992;30:473-83. 3. Stewart AL, Hays RD, Ware JE. The MOS short-form general health survey: reliability and validity in a patient population. Med Care 1988;26:724-35. 4. Neill WA, Branch LG, DeJong G. Cardiac disability—the impact of coronary disease on patients’ daily activities. Arch Intern Med 1981;145:1642-7. 5. Ades PA, Grunvald MH. Cardiopulmonary exercise testing before and after conditioning in older coronary patients. Am Heart J 1990;120:585-9. 6. Pinsky JL, Jette AM, Branch LG, et al. The Framingham Disability Study: relationship of various coronary heart disease manifestations to disability in older persons living in the community. Am J Public Health 1990;80:1363-8. 7. Chirikos T, Nickel J. Socioeconomic determinants of disablement from chronic disease episodes. Soc Sci Med 1986;22:1329-35. 8. Ades PA. Cardiac rehabilitation in older coronary patients. J Am Geriatr Soc 1999;47:98-105. 9. Heymsfield S, Smith R, Aulet M, et al. Appendicular skeletal muscle mass: measurement by dual-photon absorptiometry. Am J Clin Nutr 1990;52:214-8.
American Heart Journal January 2002
10. Wang Z, Visser M, Ma R, et al. Skeletal muscle mass: evaluation of neutron activation and dual-energy x-ray absorptiometry methods. J Appl Physiol 1996;80:824-31. 11. Yesavage JA, Brink TL, Rose TL. Development and validation of a geriatric depression screening scale—a preliminary report. J Psychiatr Res 1983;17:37-49. 12. Ades PA, Maloney A, Savage P, et al. Physical function in coronary patients: effect of cardiac rehabilitation. Arch Intern Med 1999; 159:2357-60. 13. Stern MJ, Gorman PA, Kaslow P. The group counseling vs exercise therapy study: a controlled intervention with subjects following myocardial infarction. Arch Intern Med 1983;143:1719-25. 14. Wenger NK, Froehlicher ES, Smith LK, et al. Cardiac rehabilitation: clinical practice guidelines. Bethesda (MD): Agency for Health Care Policy and Research and the National Heart, Lung, and Blood Institute; 1995. Report No.: AHCPR Publication No. 96-0672. 15. Rantanen T, Guralnik J, Sakari-Rantala R, et al. Disability, physical activity, and muscle strength in older women: the Women’s Health and Aging Study. Arch Phys Med Rehabil 1999;80:130-5. 16. Visser M, Langlois J, Guralnik J, et al. High body fatness, but not low fat-free mass, predicts disability in older men and women: the Cardiovascular Health Study. Am J Clin Nutr 1998;68:584-90. 17. Williams MA, Maresh CM, Esterbrooks DJ, et al. Early exercise training in patients older than age 65 years compared with that in younger patients after acute myocardial infarction of coronary bypass grafting. Am J Cardiol 1985;55:263-6. 18. Ades PA, Hanson JS, Gunther PG, et al. Exercise conditioning in the elderly coronary patient. J Am Geriatr Soc 1987;35:121-4. 19. Ades PA, Waldmann ML, Poehlman ET, et al. Exercise conditioning in older coronary patients: submaximal lactate response and endurance capacity. Circulation 1993;88:572-7. 20. Ades PA, Waldmann ML, Meyer WL, et al. Skeletal muscle and cardiovascular adaptations to exercise conditioning in older coronary patients. Circulation 1996;94:323-30. 21. Lavie CJ, Milani RV, Littman AB. Benefits of cardiac rehabilitation and exercise training in secondary coronary prevention in the elderly. J Am Coll Cardiol 1993;22:678-83. 22. Coulehan J, Schulberg H, Block M, et al. Treating depressed primary care patients improves their physical, mental, and social functioning. Arch Intern Med 1997;157:1113-20. 23. Poehlman E, Toth M, Gardner A. Changes in energy balance and body composition at menopause: a controlled longitudinal study. Ann Intern Med 1995;123:673-5. 24. Jette A, Lachman M, Giorgetti M, et al. Exercise—it’s never too late: the Strong-for-Life Program. Am J Public Health 1999;89:66-72. 25. Ades PA, Ballor DL, Ashikaga K, et al. Weight training improves walking endurance in healthy elderly persons. Ann Intern Med 1996;124:568-72. 26. King P, Savage P, Ades PA. Home resistance training in an elderly woman with coronary heart disease. J Cardiopulm Rehabil 1999; 20:126-9.