James F. Fries

Aging, Natural Death, and the Compression of Morbidity

The average length of life has risen from 47 to 73 years in this century, but the maximum life span has not increased. Therefore, survival curves have assumed an ever more rectangular form. Eighty per cent of the years of life lost to nontraumatic, premature death have been eliminated, and most premature deaths are now due to the chronic diseases of the later years. Present data allow calculation of the ideal average life span, approximately 85 years. Chronic illness may presumably be postponed by changes in life style, and it has been shown that the physiologic and psychologic markers of aging may be modified. Thus, the average age at first infirmity can be raised, thereby making the morbidity curve more rectangular. Extension of adult vigor far into a fixed life span compresses the period of senescence near the end of life. Health-research strategies to improve the quality of life require careful study of the variability of the phenomena of aging and how they may be modified.

The Patient-Reported Outcomes Measurement Information System (PROMIS): progress of an NIH Roadmap cooperative group during its first two years.

Background:The National Institutes of Health (NIH) Patient-Reported Outcomes Measurement Information System (PROMIS) Roadmap initiative (www.nihpromis.org) is a 5-year cooperative group program of research designed to develop, validate, and standardize item banks to measure patient-reported outcomes (PROs) relevant across common medical conditions. In this article, we will summarize the organization and scientific activity of the PROMIS network during its first 2 years. Design:The network consists of 6 primary research sites (PRSs), a statistical coordinating center (SCC), and NIH research scientists. Governed by a steering committee, the network is organized into functional subcommittees and working groups. In the first year, we created an item library and activated 3 interacting protocols: Domain Mapping, Archival Data Analysis, and Qualitative Item Review (QIR). In the second year, we developed and initiated testing of item banks covering 5 broad domains of self-reported health. Results:The domain mapping process is built on the World Health Organization (WHO) framework of physical, mental, and social health. From this framework, pain, fatigue, emotional distress, physical functioning, social role participation, and global health perceptions were selected for the first wave of testing. Item response theory (IRT)-based analysis of 11 large datasets supplemented and informed item-level qualitative review of nearly 7000 items from available PRO measures in the item library. Items were selected for rewriting or creation with further detailed review before the first round of testing in the general population and target patient populations. Conclusions:The NIH PROMIS network derived a consensus-based framework for self-reported health, systematically reviewed available instruments and datasets that address the initial PROMIS domains. Qualitative item research led to the first wave of network testing which began in the second year.

Pathologic observations in systemic sclerosis (scleroderma): A study of fifty-eight autopsy cases and fifty-eight matched controls

Abstract Fifty-eight autopsy cases of systemic sclerosis were compared with fifty-eight matched controls to determine the extent, frequency and type of involvement in various organs. The organs found to be frequently and significantly involved by this disease were the skin, gastrointestinal tract, lungs, kidneys, skeletal muscle and pericardium. Myocardial and pleural involvement was also present, but was less easy to separate from similar lesions seen in matched controls. Specific involvement of the brain, anterior pituitary, cardiac valves, liver, spleen, gallbladder, urinary bladder, pancreas and peritoneum could not be demonstrated. In organs containing smooth muscle, such as the gastrointestinal tract, and in skeletal muscle, atrophy was usually a more prominent finding than was fibrosis. Vascular lesions suggestive of severe hypertension were seen in the kidneys of patients with scleroderma in the absence of hypertension. Moreover, vascular lesions suggestive of pulmonary hypertension were found in the lungs of patients without either clinical or pathologic evidence of pulmonary hypertension. Similar vascular lesions were also noted in other organs.

AGING, HEALTH RISKS, AND CUMULATIVE DISABILITY

BACKGROUND Persons with lower health risks tend to live longer than those with higher health risks, but there has been concern that greater longevity may bring with it greater disability. We performed a longitudinal study to determine whether persons with lower potentially modifiable health risks have more or less cumulative disability. METHODS We studied 1741 university alumni who were surveyed first in 1962 (average age, 43 years) and then annually starting in 1986. Strata of high, moderate, and low risk were defined on the basis of smoking, body-mass index, and exercise patterns. Cumulative disability was determined with a health-assessment questionnaire and scored on a scale of 0 to 3. Cumulative disability from 1986 to 1994 (average age in 1994, 75 years) or death was the measure of lifetime disability. RESULTS Persons with high health risks in 1962 or 1986 had twice the cumulative disability of those with low health risks (disability index, 1.02 vs. 0.49; P<0.001). The results were consistent among survivors, subjects who died, men, and women and for both the last year and the last two years of observation. The onset of disability was postponed by more than five years in the low-risk group as compared with the high-risk group. The disability index for the low-risk subjects who died was half that for the high-risk subjects in the last one or two years of observation. CONCLUSIONS Smoking, body-mass index, and exercise patterns in midlife and late adulthood are predictors of subsequent disability. Not only do persons with better health habits survive longer, but in such persons, disability is postponed and compressed into fewer years at the end of life.

Osteoarthritis: New Insights. Part 2: Treatment Approaches

There is no known cure for osteoarthritis, and the goal of contemporary management of the patient with osteoarthritis remains control of pain and improvement in function and health-related quality of life with avoidance, if possible, of therapeutic toxicity. Recent studies have demonstrated the potential of treatments ranging from newly approved oral medications to nutriceuticals, patient education interventions, and surgery. Increasingly, appropriate treatment of osteoarthritis combines one or more oral agents with exercise and other biomechanical techniques. This article is part 2 of a two-part summary of a National Institutes of Health (NIH) conference, Stepping Away from OA: Prevention of Onset, Progression, and Disability of Osteoarthritis. The conference brought together experts from diverse backgrounds and provided a multidisciplinary and comprehensive summary of recent advances in the prevention of osteoarthritis onset, progression, and disability. For research questions and opportunities identified at the conference, see www.nih.gov/niams/reports/oa/oareport.htm(accessed on 25 May 2000). Systemic and Topical Treatments Dr. Marc C. Hochberg (University of Maryland School of Medicine, Baltimore, Maryland), Dr. Timothy McAlindon (Boston University School of Medicine, Boston, Massachusetts), and Dr. David T. Felson (Boston University School of Medicine): Drug therapy for pain management is most effective when combined with nonpharmacologic strategies (1, 2). In 1995, the American College of Rheumatology issued guidelines for the medical management of osteoarthritis of the hip and knee (2, 3). Since then, several systematic reviews of drug therapy for osteoarthritis have been published (4-6). The following recommendations for systemic and topical treatments (except for glucosamine and chondroitin, which were not evaluated) are derived from updated recommendations of the American College of Rheumatology for the treatment of osteoarthritis. Systemic Treatments Nonopioid Analgesics For many patients with osteoarthritis, the relief of mild to moderate joint pain afforded by the simple analgesic acetaminophen is comparable to that achieved with a nonsteroidal anti-inflammatory drug (NSAID) (7, 8). Accordingly, although acetaminophen fails to adequately relieve pain in many patients, it merits a trial as initial therapy on the basis of its overall cost, efficacy, and toxicity profile (9, 10). The daily dose of acetaminophen should not exceed 4 g. Although it is one of the safest analgesics, acetaminophen can be associated with clinically important adverse events, such as prolongation of the half-life of warfarin (11). At therapeutic doses acetaminophen rarely causes hepatic toxicity, but it should be used cautiously in patients with existing liver disease and avoided in patients with chronic alcohol abuse because of known increased risk in these patients (12-14). Even though acetaminophen was reported to be weakly associated with end-stage renal disease, the Scientific Advisory Committee of the National Kidney Foundation recommended it as the drug of choice for analgesia in patients with impaired renal function (15). Tramadol, a centrally acting oral analgesic, is a synthetic opioid agonist that inhibits reuptake of norepinephrine and serotonin. It has been approved by the U.S. Food and Drug Administration for treatment of moderate to severe pain and can be considered for use in patients in whom acetaminophen therapy has failed and who have contraindications to NSAIDs, including the cyclooxygenase-2 (COX-2)specific inhibitors. Although numerous studies have examined use of tramadol to treat general pain, few controlled studies have examined its use in osteoarthritis. The efficacy of tramadol has been found to be comparable to that of ibuprofen in patients with hip and knee osteoarthritis (16), and it is useful as adjunctive therapy in patients with osteoarthritis whose symptoms were inadequately controlled with NSAIDs (17). Daily doses of tramadol have generally been in the range of 200 to 300 mg given in four divided doses. Side effects are common and include nausea, constipation, and drowsiness. Despite the opioid pharmacology of tramadol, a comprehensive surveillance program has failed to demonstrate significant abuse, and tramadol remains an unscheduled agent. Seizures have been reported as a rare side effect, either at doses above the recommended range or at doses within the recommended range in patients with a history of epilepsy and those taking concomitant medications that lower the seizure threshold. NSAIDs For patients who do not obtain adequate symptom relief with nonopioid analgesics, use of NSAIDs should be considered. The choice between a nonselective NSAID and a COX-2specific inhibitor should be made after evaluation of risk factors, particularly for upper gastrointestinal and renal toxicity. Data from epidemiologic studies show that among persons 65 years of age or older, 20% to 30% of all hospitalizations and deaths due to peptic ulcer disease were attributable to therapy with NSAIDs (18-20). Furthermore, the risk for a catastrophic gastrointestinal event in elderly patients taking NSAIDs is dose dependent (18). Risk factors for upper gastrointestinal bleeding in patients treated with NSAIDs include age 65 years or older, history of peptic ulcer disease or previous upper gastrointestinal bleeding, concomitant use of oral corticosteroids or anticoagulants, and possibly smoking and alcohol consumption (21-23). Risk factors for reversible renal failure in patients with intrinsic renal disease who are treated with NSAIDs include age 65 years or older, hypertension or congestive heart failure, and concomitant use of diuretics and angiotensin-converting enzyme inhibitors (24). Additional considerations involved in a practitioners decision to treat an individual patient with osteoarthritis include existing comorbid conditions and concomitant therapy, as well as the side effects and costs of specific treatments. The options for medical management of the patient with osteoarthritis who is at increased risk for a serious adverse upper gastrointestinal event, such as bleeding, perforation, or obstruction, are use of a COX-2specific inhibitor or a nonselective NSAID with gastroprotective therapy. Two COX-2specific inhibitors, celecoxib and rofecoxib, have been approved by the U.S. Food and Drug Administration for use in patients with osteoarthritis (25, 26). Celecoxib has been found to be more effective than placebo and as effective as naproxen for symptoms in patients with hip or knee osteoarthritis (27-29). Rofecoxib has also been found to be more effective than placebo and is comparable in efficacy to both ibuprofen and diclofenac in patients with hip or knee osteoarthritis (30, 31). Endoscopic studies have shown that celecoxib and rofecoxib are associated with an incidence of gastroduodenal ulcers lower than that of comparator NSAIDs and similar to that of placebo (25). These data suggest an advantageous safety profile compared with nonselective NSAIDs, especially for treatment of high-risk patients (21-23). However, no large long-term studies have been published that were designed to demonstrate differences between COX-2specific inhibitors and nonselective NSAIDs with respect to major gastrointestinal clinical outcomes; such studies are in progress. A further advantage of COX-2specific inhibitors with respect to upper gastrointestinal bleeding is that celecoxib and rofecoxib do not have a clinically significant effect on platelet aggregation or bleeding time. In addition, at doses recommended for treatment of osteoarthritis, these drugs appear to be better tolerated than comparator nonselective NSAIDs, with a lower incidence of dyspepsia and other gastrointestinal side effects. As with nonselective NSAIDs, however, COX-2specific inhibitors can cause renal toxicity. Caution must be exercised, therefore, if these drugs are used in patients with mild to moderate renal insufficiency, and they should not be used in patients with severe renal insufficiency. In addition, celecoxib is contraindicated in patients with a history of allergic reaction to a sulfonamide. The alternative to use of a COX-2specific inhibitor is use of a nonselective NSAID with a gastroprotective agent, an approach endorsed by the American College of Gastroenterology (23). As noted earlier, serious adverse upper gastrointestinal events attributed to NSAIDs in the elderly are dose dependent. Therefore, if nonselective NSAIDs are used, therapy should be begun at low, analgesic doses and increased to full anti-inflammatory doses only if lower doses do not provide adequate relief of symptoms. In a study of 8843 patients with rheumatoid arthritis, misoprostol at a dosage of 200 g four times daily reduced the incidence of serious ulcer complications, including perforation, bleeding, and obstruction, by 51% (32). In a 12-week randomized, double-blind, placebo-controlled endoscopy study, misoprostol at a dosage of 200 g three times per day had comparable efficacy in prevention of both gastric and duodenal ulcers; however, 200 g twice daily conferred significantly less protection against gastric ulcers (33). Side effects, particularly diarrhea and flatulence, may occur with this agent in a dose-dependent manner (33). Alternative approaches to prophylaxis with misoprostol include use of omeprazole or high-dose famotidine, both of which have been shown in carefully conducted endoscopy studies to be effective in treating and preventing NSAID-induced gastropathy (34-37). Histamine-2 blockers in usual doses, however, have not been found to be as effective as misoprostol (36), whereas omeprazole (20 mg/d or 40 mg/d) was as effective as misoprostol (200 g twice daily) in treatment of existing ulcers and was better tolerated and associated with a lower rate of relapse (37). Of note, proton-pump inhibitors have not been approved by the U.S. Food and Dr

The Stanford Health Assessment Questionnaire: Dimensions and Practical Applications

The ability to effectively measure health-related quality-of-life longitudinally is central to describing the impacts of disease, treatment, or other insults, including normal aging, upon the patient. Over the last two decades, assessment of patient health status has undergone a dramatic paradigm shift, evolving from a predominant reliance on biochemical and physical measurements, such as erythrocyte sedimentation rate, lipid profiles, or radiographs, to an emphasis upon health outcomes based on the patients personal appreciation of their illness. The Health Assessment Questionnaire (HAQ), published in 1980, was among the first instruments based on generic, patient-centered dimensions. The HAQ was designed to represent a model of patient-oriented outcome assessment and has played a major role in many diverse areas such as prediction of successful aging, inversion of the therapeutic pyramid in rheumatoid arthritis (RA), quantification of NSAID gastropathy, development of risk factor models for osteoarthrosis, and examination of mortality risks in RA.Evidenced by its use over the past two decades in diverse settings, the HAQ has established itself as a valuable, effective, and sensitive tool for measurement of health status. It is available in more than 60 languages and is supported by a bibliography of more than 500 references. It has increased the credibility and use of validated self-report measurement techniques as a quantifiable set of hard data endpoints and has contributed to a new appreciation of outcome assessment. In this article, information regarding the HAQs development, content, dissemination and reference sources for its uses, translations, and validations are provided.

Nonsteroidal anti-inflammatory drug-associated gastropathy: Incidence and risk factor models

PURPOSE The most prevalent serious drug toxicity in the United States is increasingly recognized as gastrointestinal (GI) pathology associated with the use of nonsteroidal anti-inflammatory drugs (NSAIDs). The incidence of serious GI events (hospitalization or death) associated with NSAID use was therefore prospectively analyzed in patients with rheumatoid arthritis (RA) and patients with osteoarthritis. PATIENTS, METHODS, AND RESULTS The study consisted of 2,747 patients with RA and 1,091 patients with osteoarthritis. The yearly hospitalization incidence during NSAID treatment was 1.58% in RA patients and was similar in all five populations studied. The hazard ratio of patients taking NSAIDs to those not taking NSAIDs was 5.2. The incidence in osteoarthritis may be less. The risk of GI-related death in RA patients was 0.19% per year with NSAIDs. Multivariate analyses assessing risk factors for serious GI events were performed in the 1,694 (98 with an event) RA patients taking NSAIDs at the predictive visit. The main risk factors were higher age, use of prednisone, previous NSAID GI side effects, prior GI hospitalization, level of disability, and NSAID dose. A rule is presented that allows estimation of the risk for the individual patient with RA. CONCLUSION Knowledge of the risk factors for NSAID-associated gastropathy and their inter-relationships provides a tool for identification of the patient at high risk and for initiation of appropriate therapeutic action.

Toward an Epidemiology of Gastropathy Associated With Nonsteroidal Antiinflammatory Drug Use

The thesis of this paper is that gastropathy associated with nonsteroidal antiinflammatory drugs (NSAIDs) is the most frequent and, in aggregate, the most severe drug side effect in the United States. This work is based on a consecutive series of 2400 patients with rheumatoid arthritis followed prospectively for an average of 3.5 yr by ARAMIS, the American Rheumatism Association Medical Information System. We present a preliminary exploration of the magnitude of the problem, the population at risk, and the patients within that population who are at particularly high risk. Patients on NSAIDs had a hazard ratio for gastrointestinal (GI) hospitalization that was 6.45 times that of patients not on NSAIDs. Characteristically, high-risk patients for GI hospitalization and GI death are older, have had previous upper abdominal pain, have previously stopped NSAIDs for GI side effects, and have previously used antacids or H2-receptor antagonists for GI side effects. They also are frequently on corticosteroids. In contrast, patients attributing relatively minor symptoms to the drug tend to be younger and more frequently female. Our preliminary analysis is univariate and, as these variables are interdependent, firm conclusions regarding the relative importance of these risk factors will require reevaluating our data base as it is expanded using multivariate analysis. The syndrome of NSAID-associated gastropathy can be estimated to account for at least 2600 deaths and 20,000 hospitalizations each year in patients with rheumatoid arthritis alone.

Measuring and monitoring success in compressing morbidity.

The Compression of Morbidity paradigm, which was presented as an hypothesis in 1980 (1), noted that most illness was chronic and occurred in later life and postulated that the lifetime burden of illness could be reduced if the onset of chronic illness could be postponed and if this postponement could be greater than increases in life expectancy. Figure 1 illustrates this concept. Estimated present lifetime morbidity is portrayed with three possible future scenarios: life extension, shift-to-the-right, and compression of morbidity. The lines represent the length of life, and the shaded triangles depict lifetime morbidity. Two arrows are shown for each scenario: The left arrow represents the median age at onset of chronic morbidity and therepresents the median age at death. Alternative health futures are determined by the relative movement of these arrows over time. If the arrows separate, lifetime morbidity increases, and if they come closer, morbidity is compressed. In each scenario, some extension of life expectancy is envisioned. The illustrative use of age 55 years as the present age of onset of chronic morbidity is drawn from our data showing this to be the median age of detectable chronic disability (2). Figure 1. Possible scenarios for future morbidity and longevity. In 1980, most demographers and health policy workers believed that the life extension scenario was occurring and was the most likely, and most unfortunate, future for health. As medical progress increasingly prolonged life, those extra months and years would be spent in poorer health. The process was termed the failure of success (3). The Compression of Morbidity represented a positive concept, with the ideal of a long life with a relatively short period of terminal decline. Initially, the concept was sometimes portrayed as naively optimistic and, perhaps, a threat to the preparation required to care for ever-larger elderly populations (4, 5). A problem in 1980 was the lack of data on trends in morbidity; such data would have required serial population surveys with similar methods, and these were not available. There was not even agreement on the definition of morbidity, nor is there now. It seemed reasonable in 1980 to postulate that lifetime morbidity had been rising with the emergence of chronic illnesses, but it could be argued that this period was ending with declining incidences of major chronic illnesses, such as cardiovascular disease (1). The Compression paradigm focuses attention on the quality of life over its quantity and considers morbidity as a lifetime cumulative area-under-the-curve concept rather than just a cross-sectional particular point of time, such as a specific age. It suggests that the national burden of illness may be reduced by postponing the onset of infirmity. Thus, the national illness burden that is increasing because of the growing number of elderly persons in the population may be offset, at least in part, by a lower average illness burden for the individual, with positive consequences for the stability of the health care system. This paper draws on recent data from national surveys, observational longitudinal studies, and randomized, controlled trials to bring together the current evidence for morbidity compression in the United States and to outline the research agenda for the continued monitoring of trends in morbidity and disability. Given appropriate data, the Compression hypothesis can be tested by using age-specific disability rates or cumulative lifetime disability as surrogates for the less easily quantitated morbidity rates. Trends in age-specific disability can then be compared with trends in age-specific mortality, and trends toward a postponed point of incident disability can be compared with trends toward a postponed age at death. If age-specific disability declines faster than age-specific mortality, then compression is established. Even better, if given serial cohorts of longitudinal data, trends in cumulative lifetime disability could be assessed. In addition, if disability could be postponed by specific interventions (such as exercise, weight control, total joint replacement, influenza vaccination, or smoking cessation) by more than projected increases in the length of life from these interventions, then strategies for reduction in the national burden of illness by dissemination of these interventions would become possible. Trends in Mortality Since morbidity compression depends on the relative trends in morbidity and mortality, trends in life expectancy must be examined. From 1980 to 1998, life expectancy (U.S. data; both sexes combined) rose at a rate of 0.150 year per year, for an increase of 2.7 years, to 76.6 years, somewhat slower than in the two previous decades (6). The more relevant number, however, is life expectancy from age 65 because it excludes early-life mortality. Life expectancy from age 65 rose at a rate of only 0.066 year per year, for an increase of 1.2 years over 18 years, to a life expectancy of 82.7 years; this is a slightly lower rate of increase than in previous decades. From age 85, life expectancy rose by 0.017 year per year for an increase of 0.3 years or only about 4 months over 18 years. If present trends continue, gains in life expectancy after 65 years of age will remain modest. Because of larger birth cohorts and an increasing proportion of persons surviving to age 65, however, the number of elderly persons will increase markedly. Present trends, however, will not continue indefinitely. Barring discovery and dissemination of a breakthrough in understanding of the basic science of aging, the rates of increases will continue to slow. Constant decreases in mortality rates, about 1% per year, yield diminishing increases in terms of life expectancy over time because of the mathematical relationship between the two statistics; the absolute value of the mortality rate change decreases over time (7). This can be seen with the point of paradox, in which life expectancy from birth, rising more rapidly, exceeds life expectancy from age 65 years, rising more slowly, after an intersection point at a life expectancy of 87.8 years in the year 2076. Because this is not possible, present trends cannot continue indefinitely. Trends in Morbidity Over the past two decades, it is possible that a great natural experiment occurred in which preventive measures, which are most relevant to postponement of morbidity, had been broadly implemented. On a national basis, there could have been massive reductions in health risk factors, such as cigarette smoking, obesity, and sedentary lifestyle. Then, with emerging data on trends in disability, particularly from the National Long-Term Care Survey (NLTCS) (8) and the National Health Interview Survey (NHIS) (9), we could assess the benefits from these changes against changes in mortality. Unfortunately, although smoking did decrease significantly, the prevalence and amount of obesity increased, and a trend toward more sedentary lifestyles continued. Over the last two decades, therefore, lifestyle changes probably did not play a major role in compression of morbidity. Nevertheless, compression of morbidity is occurring, and at a relatively rapid rate. In the NLTCS data (Table), disability in persons older than 65 years of age decreased from 26.2% to 19.7% from 1982 to 1999, a decrease of approximately 2% per year and substantially greater than the decreases in mortality rates over this period, which declined at about 1% per year (10). The decline was seen both in activities of daily living disability and instrumental, more complex activities of daily living. It has been argued that a decline in age-specific disability of 1.5% per year would be sufficient to maintain the solvency of Medicare for at least 70 years (11). Moreover, the declines in disability in the NLTCS accelerated in the past 5 years of observation and began to be seen in minority as well as white subpopulations (8). Data from the NHIS from 1982 to 1996 show strikingly similar declines in disability, again with an accelerated rate of decline in more recent years (9). Table. Long-Term Trends in Disability Recently, Freedman and colleagues (12) performed a systematic review of 16 studies of trends in disability and found them consistent in the finding of declining disability trends. Thus, the literature, although of uneven quality, shows consensus. Only the NLTCS data had a sample of all Medicare-eligible persons, both institutionalized and not; covered a sufficient time period (1982 to 1999); and had well-defined disability measures. This is the single best study (Figure 2) (13). Figure 2. Recent trends in disability among older Americans. NLTCS NHIS ADL IADL The reasons for the decline in disability seem to be multifactorial, with no single identifiable cause. There may have been contributions from reductions in cigarette smoking; medical advances, such as better treatment of hypertension, diabetes, coronary artery disease, and rheumatoid arthritis; total joint replacements; and medical preventive measures, such as colon cancer screening, influenza and pneumococcal vaccines, and cardiac-dose aspirin (10). It might be that rising expectations for healthier aging became self-fulfilling, perhaps through a mechanism of perceived self-efficacy (14). The association between level of education and health is strong, and self-efficacy might at least partly explain this association; education levels in the elderly rose substantially over the past two decades (15). Social factors do not seem to have played a role; access to care has not improved, and access to prescription drugs may have become more difficult (10). Whatever the reasons, compression of morbidity is currently occurring at the population level in the United States. Health Risks and Lifetime Disability Longitudinal studies of the relationships of lifestyle risk factors to subsequent disability levels in advantaged elderly pop

Reduced Disability and Mortality Among Aging Runners: A 21-Year Longitudinal Study

BACKGROUND Exercise has been shown to improve many health outcomes and well-being of people of all ages. Long-term studies in older adults are needed to confirm disability and survival benefits of exercise. METHODS Annual self-administered questionnaires were sent to 538 members of a nationwide running club and 423 healthy controls from northern California who were 50 years and older beginning in 1984. Data included running and exercise frequency, body mass index, and disability assessed by the Health Assessment Questionnaire Disability Index (HAQ-DI; scored from 0 [no difficulty] to 3 [unable to perform]) through 2005. A total of 284 runners and 156 controls completed the 21-year follow-up. Causes of death through 2003 were ascertained using the National Death Index. Multivariate regression techniques compared groups on disability and mortality. RESULTS At baseline, runners were younger, leaner, and less likely to smoke compared with controls. The mean (SD) HAQ-DI score was higher for controls than for runners at all time points and increased with age in both groups, but to a lesser degree in runners (0.17 [0.34]) than in controls (0.36 [0.55]) (P < .001). Multivariate analyses showed that runners had a significantly lower risk of an HAQ-DI score of 0.5 (hazard ratio, 0.62; 95% confidence interval, 0.46-0.84). At 19 years, 15% of runners had died compared with 34% of controls. After adjustment for covariates, runners demonstrated a survival benefit (hazard ratio, 0.61; 95% confidence interval, 0.45-0.82). Disability and survival curves continued to diverge between groups after the 21-year follow-up as participants approached their ninth decade of life. CONCLUSION Vigorous exercise (running) at middle and older ages is associated with reduced disability in later life and a notable survival advantage.