Exercise in the Oldest Old: Some New Insights and Unanswered Questions

In this issue of the Journal of the American Geriatrics Society, three articles speak to the feasibility and utility of an exercise prescription for the “oldest old”: adults with preexisting functional dependency or multiple chronic diseases. These new studies extend our current knowledge about the safety and efficacy of exercise programs initiated in the eighth and ninth decades of life, where the focus of healthcare practitioners is shifted from disease prevention toward disease treatment and the looming issue of disability. Together, they provide further evidence for the inclusion of a geriatric exercise prescription within the mainstream of the medical practice of geriatricians, rather than as an optional adjunct to standard care of the oldest old.1 At the same time, though, they raise important unanswered questions in this cohort about pre-exercise screening, optimal exercise modalities for specific outcomes, dose-response relationships, and mechanisms of exercise-derived health benefits. In a well-designed randomized, controlled trial, Binder et al.2 describe the effects of a moderately intensive, multimodal exercise program on physiological impairments and self-reported disability in 115 adults aged 78 and older selected for preexisting physical frailty. Fifteen percent of individuals responding to advertisements were excluded because they were considered too frail or medically unsuitable for vigorous exercise. Twenty-seven percent of randomized participants dropped out of the study before the 9-month reassessment, mostly because of medical problems unrelated to the exercise training. Only two exercise-related adverse events were reported over the course of the trial, both shoulder injuries during resistance training. Several questions come to mind when reviewing these outcomes. The requirement of completing a maximal exercise tolerance test, in addition to full medical screening by a physician, chest x-ray, and blood tests, may have eliminated individuals at high risk for exercise-related adverse events, although the authors do not present data on the percentage of abnormal screening tests. The impressive safety record of this study once again counters the persistent myth that moderate frailty or advanced age is a contraindication to appropriately robust exercise.3 Are these safety data generalizable to frail older people in clinical exercise programs who are not so rigorously (and expensively) screened and monitored? Although important prognostic information is obtained from cardiac stress testing in the clinical setting, the issue of stress testing before exercise prescription in asymptomatic older people continues to be a controversial one.4 Many physicians and exercise physiologists do not recommend stress testing before moderate- or low-intensity exercise programs in older people because of the cost implications, potential barriers to participation, false-positive results, and lack of evidence that such screening is necessary. It would be helpful to know what medical conditions were considered exclusionary, what constituted “too frail,” whether current smokers could have been included, and whether telemetry provided information not obtainable via simpler methods such as pulse monitoring or ratings of perceived exertion. There are several examples of moderate- and high-intensity resistance and balance training programs in older adults with chronic disease that have been safely conducted without prior stress testing.5–7 It remains to be seen whether moderate-to-high–intensity aerobic training can be carried out in this way as well. Those assigned to 9 months of supervised high-intensity flexibility and balance training, resistance exercise, and aerobic training significantly improved their muscle strength, balance, aerobic capacity, perceived health status, and activities of daily living function, compared with those randomized to a low-intensity, home-based flexibility program 2 days per week. One of the strengths of this study is the use of a theoretically grounded exercise prescription targeted to the remediable physiological impairments most often linked to functional dependency in older people. The program was successfully phased in over 6 months to allow physical and psychological adaptation to this new behavior in an attempt to maximize adherence, safety, and efficacy. The lack of appreciable objective benefits from low-intensity exercise (as commonly prescribed to older or frailer adults) should dissuade healthcare professionals from using doses and modalities of exercise that are below the threshold required for physiological adaptation or therapeutic efficacy, even if they seem harmless. The lack of perceived efficacy from such low-intensity regimes can only serve to undermine the promotion of exercise as medicine and lead to dropout and discouragement on the part of both patient and practitioner. As is the case with most multicomponent interventions, it is not possible to determine the relative potency or necessity of any of the individual elements of the prescription. This is not unimportant, because the program as described is somewhat complex, costly, and dependent on supervision and cardiac monitoring at baseline and throughout. Several questions come to mind. Because flexibility training in the control group resulted in no appreciable benefits, was it a necessary part of the active intervention? Which of the exercise components was most important for the improvement in disability? For example, other studies in disabled adults with osteoarthritis of the knee have shown similar efficacy of aerobic and resistance training8 on functional independence. Without a direct comparison of various exercise modalities in the same cohort, it is not clear that this more-complicated regimen is necessary or more effective than single modalities. The authors do not present the relationships between changes in proximate outcomes (aerobic capacity, strength, balance, flexibility) and changes in activities of daily living; such analyses might have shed light on the question of relative potency of individual prescriptive elements. It seems logical that identification of specific physiological impairments followed by a modified exercise prescription (resistance training for weakness, etc.) would be superior to a “blanket” exercise prescription for the treatment of frailty, but this individualized approach has not yet been reported in clinical trials of exercise and disability. No information is provided as to the etiology of the functional impairment in this cohort. The high prevalence of arthritis (>75%) suggests that this was the most likely etiology of disability, given the absence of other common causes such as depression, stroke, cognitive decline, neuromuscular disease, or sensory impairment. It would be important to know whether the exercise intervention was also applicable to individuals with frailty unrelated to arthritis. Finally, the proportion of women in this trial (52%) was far lower than would be predicted in frail older people aged 80 and older. It is possible that gender plays a role in the treatment of disability with exercise, and it would be helpful to know whether frail women adhere and respond as well as men do, given their higher prevalence of disability. The current trial was not powered for such subgroup analyses; future studies will need to address these issues. The second study, by Vaitkevicius et al.,9 provides further evidence that moderate/high intensity endurance training (21 minutes per session at approximately 75% of maximal heart rate 2 days per week over 6 months) results in a modest improvement in peak oxygen consumption in individuals aged 80 and older with chronic disease. The uncontrolled design and small sample size limit this study, but it does suggest that even this small volume of training can reverse 5 to 10 years worth of decline in maximal oxygen consumption. The improvement of 6.5% in aerobic capacity was less than the 14% to 15% seen in Binder et al.'s study2 or in studies of the young old,10 perhaps due to the absence of resistance training contributions to aerobic capacity,11 lower training volume and frequency,12 and lack of high-intensity interval training. The increase would presumably have been even lower if those who did not meet criteria for peak oxygen consumption (VO2peak) per unit time at follow-up were included in the analysis. The importance of this dose-response effect is supported by the multiple regression models, indicating that time spent training was predictive of the increase in aerobic capacity observed. Future studies will need to define the clinical implications of an improvement in aerobic capacity of this magnitude and define thresholds of efficacy.12 The other finding of interest in this study is that those with the lowest endurance capacity at baseline had the greatest response to training. This is consistent with the resistance training literature in regard to strength gains, and raises an important point about exclusionary criteria. In this study, as well as in Binder et al.'s study,2 those with the lowest functional capacity (too frail, VO2peak <10 mL/kg/min, or unable to meet criteria for VO2peak) were ineligible. If the model is correct, then these individuals would have benefited the most from the training programs and perhaps should have been included. Future studies should address this issue, because it has important implications for the clinical dissemination of such programs to individuals with more-advanced disability. Although the participants in this study are described as “frail,” the evidence for this is limited, and they appear to be healthier than the cohort studied by Binder et al.2 and had a baseline aerobic capacity in the normal range for age.13 Other limitations on the generalizability of this study to frail older people arise from the fact that participants were disproportionately male and had a relatively high dropout rate (37%). Long-term effective treatment of disability with exercise will require interventions that can be sustained over time in a majority of individuals and resumed after periods of intercurrent illness, even in the face of increasing comorbidity. A small number of such trials have reported7,8 to improve disability relative to sedentary controls over periods of 9 to 18 months, but further work is needed in this area. The third study, by Morey et al.,14 describes the effects of adherence to a clinical multimodal exercise program on mortality in a cohort of veterans followed for an average of 3.7 years. Approximately 50% of the participants (aged 71, 80% male) attended at least 47 Gerofit sessions (60% compliance) over the first 6 months of enrollment. Regardless of their adherence in the subsequent period, this early behavior change provided a survival benefit of 25% relative to nonadherers after adjustment for some potential confounders. The divergence in mortality curves is most apparent between 4 and 7 years after exercise prescription, after which time both groups begin to die at about the same rate. The mortality benefit is similar in magnitude to effects observed in much larger cohort studies in somewhat younger and healthier populations.15–17 As the authors suggest, the mechanism of benefit is not clear, and it is possible that unmeasured confounders related to adherence and mortality or cointerventions due to closer observation of adherers were responsible for the observed mortality differences. This study, although intriguing, raises more questions than it answers. Which of the modalities of exercise prescribed (aerobic, resistance, flexibility) were related to the outcome? What was the contribution of the extensive screening protocol (geriatrician review, exercise tolerance test, cardiologist referral and treatment as needed, multidisciplinary team assessment, and individualized prescription) to mortality rates or intervention safety? Why were there so few participants (135) over 10 years of the program? Were these participants different from nonparticipant veterans in important ways? What characteristics define adherence in the oldest old, compared with younger adults? For example, are previously described factors such as obesity, educational level, social support networks, self-efficacy, depression, need for transportation, caregiver roles, or functional independence predictive of adherence? What was the pattern of exercise behavior after the first 6 months? Linkage of changes in behavior to mortality would provide stronger evidence of a causal link than simply examining initial adherence status. What changes in health status or fitness were observed between groups as possible proximate mediators of mortality? Much of this information may be available to the authors, and such analyses would increase the insights this cohort provides, enabling others to design behavioral programs targeted at overcoming barriers to participation. Together, these trials suggest that the oldest old can benefit from exercise in terms of physiological capacity, functional independence, and mortality. Although the mechanisms by which exercise reduces mortality18 or improves disability in the oldest old are not clear, these benefits do not appear to be restricted to young or middle-aged individuals, before the development of disability.19 Answers to the questions raised by these new studies are critical for the effective clinical application of exercise for the oldest old, because the requirements for medical screening, mode of delivery, cost, optimal adherence, and availability of resources and trained personnel will dictate feasibility, access, and, ultimately, success in this cohort.