Compression of morbidity in the elderly James F. Fries* Stanford University School of Medicine, 1000 Welch Road, Suite 203, Palo Alto, CA 94304-5755, USA Abstract The Compression of morbidity paradigm envisions reduction in cumulative lifetime morbidity through primary prevention by postponing the age of onset of morbidity to a greater amount than life expectancy is increased, largely by reducing the lifestyle health risks which cause morbidity and disability. Recent data document slowly improving age-speci®c health status for seniors, indicate that postponement of the onset of disability by at least 10 years is feasible, and prove eectiveness of some lifestyle interventions by randomized controlled trials. Human aging is increasingly represented by frailty, with declining reserve function of many organ systems, including the immune system. Enhancement of immune function in this setting raises medical, ethical, and social issues which are sometimes in con¯ict. # 2000 Elsevier Science Ltd. All rights reserved. Keywords: Compression of morbidity; Aging; Primary prevention The health of persons over 65 years of age is a medi- cal problem in the developed nations. Over 80% of all illness, morbidity, mortality, and medical costs are concentrated in the years after age 65. The illness bur- den is mostly made up of chronic illness, complicated by the frailty of increasing age. In this century we have seen a transfer of the illness burden from the acute infectious diseases prevalen t in 1900 to chronic disease by mid-century, followed by a decline in major chronic illnesses, and now with a slow transition from chronic disease per se to the associated problems of aging [1]. 1. An overview of the aging process 1.1. Mortality changes over time Changes in survival curves in developed nations over this century are instructive, since they lead us beyond a simplistic, if popular, notion of ever-increas- ing life expectancy. Life expectancy from birth is aected most strongly by changes in infant mortality rates and in deaths in early life. Successive survival curves (Fig. 1) have become more rectangular, as marked reduction in deaths at early ages ¯atten the in- itial part of the curve. There are now few deaths prior to age 50 or 60. At the same time, the downslope of the curves has become steeper, although the insertion point has changed little [2]. A natural barrier to bio- logic longevity may be visualized through these succes- sive curves. In the United States, life expectancy from birth has increased from 47 to 75 years. However, life expectancy from advanced ages has changed relatively little. Since 1980, life expectancy for females from age 65 has increased only by 0.5 years. Almost unnoticed has been a striking occurrence; life expectancy for bot h sexes from age 85, 6.1 years, has not changed signi®- cantly during the past 20 years [3]. With the increasing size of su ccessive birth cohorts and with increased sur- vival to age 65 or age 85 the absolute numbers of senior citizens will increase markedly in coming years, while life expectancy for the average senior will increase little. Vaccine 18 (2000) 1584±1589 0264-410X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved. PII: S0 2 6 4 - 4 1 0 X ( 9 9 ) 0 0 4 9 0 - 9 www.elsevier.com/locate/vaccine * Tel.: +1-650-723-6003; fax: +1-650-723-9656. E-mail address: j@leland.stanford.edu (J.F. Fries). 1.2. Declines in organ reserve function with age Data from longitudinal studies of aging show a con- sistent decline in the maximum function of the various organs with age, the decline being essentially linear at a rate of 1.5% per year after age 30 (Fig. 2). Data on maximal performance, such as world record marathon times, similarly show a nearly linear decline with age at the same rate from age 30 to age 80 [2]. Physiologic normal values, however, as represented by the central tube in Fig. 3, remai n constant with increasing age. Normal pH, blood chemistry, white cell count, and other homeostatic values do not vary over the lifespan, representing the internal physiologic en- vironment essential for cellular function. However, with the linear decrease in organ reserve in multiple organs the ability to respond physiologically to a per- turbation decreases exponentially, represented by the decreasing area of the polygon in the ®gure. As a result, mortality rates increase exponentially, with a doubling of mortality rates each 8 years after age 30 [2]. With these realities of human aging come some paradoxes, particularly in the context of emerging scienti®c advances, some of which are more fully ela- borated in this volume. Decline in organ reserve is inevitable, yet we can increase organ reserve quite readily, at almost any age. For example, an increase in exercise can increase cardiopulmonary reserve very substantially, even at advanced ages [4,5]. 1.3. Enhanced personal autonomy and modi®able determinism In philosophic writings, a classical con¯ict has been between the advocates of free will and those of deter- minism, and this con¯ict now exists in modern dress. Determinism is represented by molecular genetics, with the notion that your health over a lifespan is ulti- mately determined only by your genes. Free will is rep- resented by the advocates of health promotion, seeking voluntary changes in behavioral risk factors, such as lack of exercise, cigarette smoking, obesit y, and dietary fat, which can enhance organ reserve, preserve func- tion, and extend life. In this view, health requires that you take care of yourself. The tension between these two paradigms is complicated by the new science, where the role of telomerase, apoptosis, the growth of pluri-potential stem cell lines, and other developments remains to be determined. Fig. 1. The rectangularization of survival curves. Successive curves, at 20 year intervals, illustrate (a) the ¯attening of the early part of the curves from decreases in premature deaths, (b) the steepening of the downslope as natural life limits are approached, and (c) the nearly constant insertion site of the curves. Fig. 2. Declining reserve organ function with age in several organ systems; decline is essentially linear and similar for dierent organs. Fig. 3. Declining ability to return to the homeostatic state (rep- resented by the tube) after a perturbation. The polygon area rep- resents the multiple organ system reserve of the organism and decreases exponentially with age. J.F. Fries / Vaccine 18 (2000) 1584±1589 1585 2. The Compression of Morbidity 2.1. The hypothesis and the paradigm The Compression of Morbidity hypothesis was introduced in 1980 [1] and has become the dominant paradigm underlying health improvement programs and policies directed at more successful aging [6,7]. Most morbidity results from chronic processes and is concentrated in the years prior to death. The goal is the compression of mo rbidity between its time of onset, currently averaging 55 years, and the age of death, currently averaging 75 years, as diagrammed in Fig. 4. A crescendo of increasing morbidity occurs, on average, over this period. With social and scienti®c progress it is likely that both the age of onset of in®r- mity and the age at death will increase over time. At issue is the relative movement of the two points. If mortality decreases predominate there may be more morbidity in a typical life, the so-called ``failure of suc- cess'' [8,9], as shown in the second line of the ®gure. If postponement of the age of onset of morbidity pre- dominates, then morbidity is compressed and the aver- age illness burden is less, the period of adult vigor is prolonged, life quality is improved, and the need for medical care and associated costs may be reduced. For many, the ideal life is one vigorous and vital over the life span, with a terminal collapse of physical and mental function limited to a few months or years immediately preceding death [2]. A central issue, when considering the eventual impact of new scienti®c dis- coveries, is whether they may add health or add illness to the average life. If morbidity is compres sed by a new advance then the advantage is clear; if the process of dying is prolonged then major ethical issues arise [10]. 2.2. Trends in morbidity compression Over the past 20 years there has been some com- pression of morbidity in the elderly, even without health policies directed at morbidity compression [11,12]. The major chronic illnesses, including heart disease, cancer, and stroke, are about 70% preventa- ble. The true causes of death are not the chronic dis- eases but the underlying causes of these diseases, which are led by cigarette smoking, sedentary living, obesity, and diets high in saturated fat and low in complex carbohydrates [13]. The potential for post- ponement of morbidity is now much better under- stood. Freedman and Martin [14], among others, have shown signi®cant age-speci®c functional improvement in seniors over a recent seven-year period. Cross-sec- tional studies have suggested compression of morbidity in favored groups, such as those with higher levels of education [15], higher socio-economic class [16], and those with fewer lifestyle health risks. Daviglus et al. [17] showed substantial decreases in Medicare costs for those with few health risk factors in mid-life. Reed et al. [18] related healthy aging to prospectively deter- mined health risks. Clearly, behavioral health risks makes a very major contribution to both morbidity and mortality. Selected medical advances such as total joint replacement and cataract extraction, or any treat- ment which makes a major contribution to life quality, also contributes to the compression of morbidity. The central measure of success of the compression of mor- bidity is reduction in the average cumulative lifetime morbidity [19]. 3. How long may the onset of morbidity be postponed? 3.1. Estimating postp onement of morbidity Our research group has been prospectively following substantial numbers of aging seniors in two cohorts longitudinally over the past 16 years to identify the factors which postpone onset of morbidity, the magni- Fig. 4. Compression of Morbidity. The top line provides a graphic estimate of current lifetime morbidity, disability, or cost, expressed as a shaded area between onset of disability at an average age of 55 years and an average age of death of 75 years. Possible future scen- arios are shown below. The second line represents the extension of morbidity if life expectancy is increased but the onset age of morbid- ity is not. The third shows compression of morbidity if the onset age of morbidity rises more rapidly than does the life expectancy. J.F. Fries / Vaccine 18 (2000) 1584±15891586 tude of the postponement, and the eects of lifestyle health risks upon cumulative lifetime disability. We use disability as a de®nable and measurable surrogate for morbidity and cumulative disabi lity over the elder years as a surrogate for cumula tive lifeti me disability [5,19]. 3.2. The University of Pennsylvania Study In the University of Pennsylvania Study we follow 1741 university attendees studied in 1939 and 1940, surveyed again in 1962, and followed annually since 1986. Health risk strata were developed for persons at high, moderate, or low risk, based upon cigarette smoking, body mass index, and lack of exercise, and assigned by risk status in 1962 (average age 43 years). Cumulative disability from 1986 to 1994 (average age 75 years) served as the surrogate for lifetime disability. Persons with high health risks in 1962 or in 1986 had twice the cumulative disability of these in the low risk strata. Deceased low risk subjects had only one-half the cumulative lifetime disability of high risk subjects and also had only one-half the amount of disability in the last one or two years of life. The same results obtained in males and in females, and in those without disability in 1962 and 1986. This 100% decrease in dis- ability rates was oset only partially by a 50% decrease in mortality rates in the high risk strata, demonstrating compres sion of morbidity. Onset of dis- ability (Fig. 5) was postponed by 7.75 years in the low risk stratum as compared with the high risk stratum [19]. 3.3. The Precursors of Arthritis Study In the Precursors of Arthritis Study we have fol- lowed 537 senior runners and vigorous exercisers with 423 age-matched community controls. We controlled for self-selection bias by a longitudinal ``intention to treat'' study since 1984, by separate analyses for gen- der, by accounting for pre-study dropouts by including all who had ever begun an exercise program at any point in life, by separately analyzing those without in- itial disabi lity, and by controlling for joint X-ray sta- tus, history of arthritis, and other factors [5, 20]. Present data show the exercising group to have less Fig. 5. Progression of disability over time in low, medium, and high risk groups from the University of Pennsylvania study. The average post- ponement of disability in the low risk group as compared with the high risk group is 7.75 years. J.F. Fries / Vaccine 18 (2000) 1584±1589 1587 than one-half the cumulative disability of the sedentary controls and this major dierence between groups actually increased over the 13 years of observation. Results held for men and women, those without initial disability, and for all of the other subgroups. The pro- portion of those disabled was also reduced by more than on-half in the exercise groups. The postponement of disability for the exercising group was 8.7 years for minimum disability and approx. 12 years for higher levels of disability. 4. Reduction in need and demand for medical service Epidemiologic studies of associations, no matter how strong and consistent, ultimately require random- ized study of interventions for proof of causality, and these are now available. Health promotion programs in senior populations directed at risk prevention, health improvement, and medical cost reduction have been studied in a number of large and well performed randomized trials, with positive results. The Bank of America randomized study of 4500 retired subjects reduced risks by 12% in the intervention groups in the ®rst year compared with controls and reduced medical care and costs, as con®rmed by a study of medical claims [22], by even a greater percentage. The Califor- nia Public Employment Retirement System study [23], involved 57,000 subjects in a 1-year randomized trial with similarly dramatic results con®rmed by claims data endpoints. Randomized studies of self-manage- ment programs for chronic diseases such as arthritis [24] and Parkinson's Disease [25] have had similarly positive results. Carefully designed high quality health improvement programs have now been proven eective and cost-eective both preventing illness and reducing medical need [26]. 5. Issues for contemplation The human aging process, when not prematurely stopped by trauma or disease, moves towards multiple organ system frailty [6,7,9]. The immediate cause of death shifts from external towards intrinsic factors. The formally assigned ``cause of death'' becomes increasingly irrelevant compared with the underlying frailty, the inability of the aging organism to withstand even a minor perturbation. Frailty is like an old cur- tain rotted by the sun, where an attempt to repair a tear in one place is followed by a tear in another. In a similar context, William Osler considered pneu- monia to be the ``old mans' friend'', terminating the in®rmities of frailty. In¯uenza epidemics result in an increase in death rates, but they are followed by a 6- month period of death rates which are actually below expected baselines. Pneumovax and in¯uenza vaccines can prevent a number of speci®c perturbations and as- sociated attributable deaths, but only in the context of multi-organ frailty. The ultimat e health bene®ts of suc- cessful interventions in the terminally frail may prove in substantial part illusory when so many competing risks remain. Age-associated declines in the immune and in¯am- matory responses have been well de®ned. Similar declines occur in other organ systems. What would be the full range of eects if declines in the immune and in¯ammatory systems were prevented or reversed in the setting of declines in other organ systems with multi-organ failure and increased frailty? Is there an important signal in the app arently accelerated aging of Dolly, the telomerase-impaired cloned sheep? The cen- tral question: would enhanced immune responsiveness increase health or increase morbidity? Answers to such questions are clearly speculative, but it seems that one answer might carry a bit of a paradox with it; that improved immune responsiveness might be construc- tive in terms of decreasing morbidity if achieved prior to terminal multiple organ frailty, but might increase lifetime morbidity, albeit only slightly, if achieved only shortly before the time of natural death. Acknowledgements This paper was supported by a grant from the National Institutes of Health (AR43584) to ARAMIS (Arthritis, Rheumatism, and Agin g Medical Infor- mation System). References [1] Fries JF. Aging, natural death, and the compression of morbid- ity. N Engl J Med 1980;303:130±5. [2] Fries JF, Crapo LM. Vitality and aging. San Francisco: Freeman, 1981. [3] Kranczer S. US life expectancy. Statistical Bulletin 1997;Oct(Dec):8±14. [4] Stewart AL, King AC, Haskell WL. Endurance exercise and health-related quality of life in 50±65 year-old adults. Gerontologist 1993;33:782±9. [5] Fries JF, Singh G, Morfeld D, et al. Running and musculoske- letal pain with age: a six-year longitudinal study. Arthritis Rheum 1996;39:64±72. [6] Fries JF. Compression of morbidity: Near or far? Milbank Mem Fund Q 1990;67(2):208±32. [7] Fries JF. Compression of morbidity 1993: Life span, disability, and health care costs. Facts Research Gerontol 1993;7:183±90. [8] Gruenberg EM. The failure of success. Milbank Q 1997;55:3± 34. [9] Olshansky SJ, Rudberg MA, Carnes BA, et al. Trading o longer life for worsening health: The expansion of morbidity hy- pothesis. J Aging Health 1991;3(2):194±216. J.F. Fries / Vaccine 18 (2000) 1584±15891588 [10] Verbrugge LM. Longer life but worsening health? Trends in health and mortality of middle-aged and older persons. Milbank Q 1984;62:475±519. [11] Nusselder WJ, Mackenbach JP. Rectangularization of the survi- val curve in the Netherlands, 1950-1992, Gerontologist 1996;36773±82. [12] Campion EW. Aging better. N Engl J Med 1998;338:1064±6. [13] Mcginnis JF, Foege W. Actual causes of death in the United States. JAMA 1993;270(19):2207±12. [14] Freedman VA, Martin LG. Understanding trends in functional limitations among older Americans. Am J Pub Health 1998;88(10):1457±62. [15] Leigh JP, Fries JF. Education, gender and the compression of morbidity. Int J Aging Hum Dev 1994;39:233±46. [16] House JS, Kessler RC, Herzog AR, et al. Age, socioeconomic status, and health. Milbank Mem Fund Q 1990;68:383± 411. [17] Daviglus ML, Liu K, Greenland P, et al. Bene®t of a favorable cardiovascular risk-factor pro®le in middle age with respect to medicare costs. N Engl J Med 1998;339(16):1122±9. [18] Reed DM, Foley DJ, White LR, et al. Predictors of healthy aging in men with high life expectancies. Am J Pub Health 1998;88(10):1463±8. [19] Vita AJ, Terry RB, Hubert HB, et al. Aging, health risks, and cumulative disability. N Engl J Med 1998;338:1035±41. [20] Fries JF, Singh S, Morfeld D, et al. Running and the develop- ment of disability with age. Ann Intern Med 1994;121(7):502±9. [22] Fries JF, Bloch DA, Harrington H, Richardson N, Beck R. Two-year results of a randomized controlled trial of a health promotion program in a retiree population: the Bank of America study. Am J Med 1993;94:455±62. [23] Fries JF, Harrington H, Edward R, Kent LA, Richardson N. Randomized controlled trial of cost reductions from a health education program: the California Public Employees' Retirement System (PERS) study. Am J Health Promotion 1994;8:216±23. [24] Fries JF, Carey C, McShane DJ. Patient education in arthritis: Randomized controlled trial of a mail-delivered program. J Rheumatol 1997;24(7):1378±83. [25] Montgomery Jr EB, Leiberman A, Singh G, Fries JF. Patient education and health promotion can be eective in Parkinson's Disease: A randomized controlled trial. Am J Med 1994;97:429± 35. [26] Fries JF, Koop CE, Sokolov J, Beadle CE, Wright D. Beyond health promotion: reducing need and demand for medical care. Health Aairs 1998;17(2):70±84. J.F. Fries / Vaccine 18 (2000) 1584±1589 1589 . Vaccine 18 (2000) 1584±1589 1585 2. The Compression of Morbidity 2.1. The hypothesis and the paradigm The Compression of Morbidity hypothesis was introduced. contributes to the compression of morbidity. The central measure of success of the compression of mor- bidity is reduction in the average cumulative lifetime morbidity