The Future of Human Healthspan: Demography, Evolution, Medicine, and Bioengineering: Task Group Summaries (2008)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

The Relationship Between Demographic Mortality Rates, Aging, and Functional Human Healthspan TASK GROUP DESCRIPTION Background Does the demographic plateauing of mortality in late life show us that functional aging decelerates or alters in rate during late life? In the 1990s it was established that age-specific mortality rates plateau at late ages in both humans and well-established experimental organisms, such as Drosophila. What remains at issue is whether this constitutes a cessation or slowing of aging itself. The task group is charged with investigating methods of determining whether mortality rate plateaus constitute a slowing of aging as such, or whether some type of artifact or compositional change is responsible for this novel, important demographic phenomenon. Such methods might employ human medical data, animal model studies, or a combination of the two. Initial Challenges to Consider Some standard data on human disability and functional status do not show plateaus, apparently, but this may be in part due to small sample sizes at older ages and to the nature of the measures we use. The most commonly used measures of disability and functional status are activities of daily living (ADL) and instrumental activities of daily living (IADL), and these may saturate at older ages. This is a problem that requires methodological inves- tigation. On the other hand, some data on age-specific incidence of cancers, 41 

42 the future of human healTHspan cardiovascular diseases, and chronic infection show late-life plateaus. Why is there a disparity between this type of health measure and those based on disability? • To investigate human pathophysiology in late life, do we need func- tional measures of aggregate aging, for example, performance measures such as walking speed, grip strength, sensory function or cognitive performance, or measures of physiological function? Would other characteristics predic- tive of mortality also be useful? • What medical databases could and should be mined for extant in- formation on the time course of human pathophysiology in specific organ systems and tissues, such as heart, brain, liver, and kidney? • Which noninvasive (e.g., blood, ultrasound) technologies would be most useful in monitoring the physiological transitions that might occur as individual patients undergo the transition from the demographic aging phase to the late-life phase, and age-specific mortality rates plateau? • The ubiquity of late-life plateaus among humans and well-studied model systems, together with the difficulty of performing experiments on human cohorts, suggests the value of studying the physiology of late life among model animals. The task group is asked to suggest possible model species experimental projects that would explore the question of whether functional aging ceases in late adult life. • It has recently been established that Drosophila age-specific fecundi- ty shows a late-life plateau analogous to that of age-specific mortality rates. Of course, in humans it is difficult to study reproductive effort because of extensive parental care. However, are there useful avenues for studying reproductive pathophysiology in humans during late life. • Recent work on the evolution of aging in populations with inter- generational transfers may offer useful perspectives on late life in humans. How should evolutionary demographic theory be developed so as to best study the evolution of human late life? • Do the oldest-old pose a unique environmental design challenge, and how might this challenge be met? Initial References Lee, R. D. 2003. Rethinking the evolutionary theory of aging: Fertility, mortality, and intergenerational transfers. Proceedings of the National Academy of Sciences U.S.A. 100(16):9637-9642. 

mortality rates, aging, and functional human healthspan 43 Mueller, L. D., C. L. Rauser, and M. R. Rose. 2007. An evolutionary heterogeneity model of late-life fecundity in Drosophila. Biogerontology 8(2):147-161. Rose, M. R., C. L. Rauser, and L. D. Mueller. 2005. Late life: A new frontier for physiology. Physiological and Biochemical Zoology 78(6):869-878. Vaupel, J. W., J. R. Carey, K. Christensen, T. E. Johnson, A. I. Yashin, N. V. Holm, I. A. Iachine, V. Kannisto, A. A. Khazaeli, P. Liedo, V. D. Longo, Y. Zeng, K. G. Manton, and J. W. Curtsinger. 1998. Biodemographic trajectories of longevity. Science 280:855- 860. Task Group Members • X. Edward Guo, Columbia University • Mary Haan, University of Michigan • Scott Hofer, Oregon State University • Bruce Kristal, Brigham and Women’s University • Vikram Kumar, Brigham and Women’s Hospital • Kenneth Manton, Duke University • Laurence Mueller, University of California, Irvine • Steven Orzack, Fresh Pond Research Institute • Judith A. Salerno, National Institute on Aging • Anatoli Yashin, Duke University • Ken Wachter, University of California, Berkeley • Andrea Anderson, New York University TASK GROUP SUMMARY By Andrea Anderson, Graduate Science Journalism Student, New York University The Gompertz curve, named for British mathematician Benjamin Gompertz, predicts that after a certain age, mortality within a population begins increasing exponentially with advancing age. At first blush human mortality rates fit on this Gompertz’s law of mortality, making it a common tool for insurance companies and others. Since the 1990s, though, there has been increasing evidence from large populations that Gompertzian mortality rates may not hold in later years. On the contrary, when scientists evaluate very large cohorts of different species, including medfly,1 fruit fly,2 and—arguably—humans, mortality rate levels off after a Gompertzian period. This means mortality does not 

44 the future of human healTHspan accelerate indefinitely with age, running counter to the notion that animals senesce unstoppably after a certain point in life. In contrast, mortality may run out of steam in a sense, still claiming individuals within a population but not a higher percentage of them with each passing year. By recognizing that there are Gompertzian and non-Gompertzian phases of aging, this task group proposed that there may be something unique about late-life biology that makes it possible to relate aging to both mortality and healthspan in new ways. Namely, the group emphasized the importance of understanding what happens to health (specifically impair- ment and disability) when mortality plateaus. Within this framework they also highlighted the need to find ways of assessing impairment and disability and to better use the measures that are already available. Charge to Task Group This task group was charged with devising a plan to address and un- derstand the late-life mortality plateau within the context of human aging and healthspan. This problem raised an interesting set of questions about whether the demographic data represent genuine underlying physiological differences or some sort of statistical artifact. If it is a real effect, does the population’s disease load decrease at these late ages along with mortality or are people living with prolonged illness? In short, what can the oldest-old tell us about the process of aging itself? Strategy When the task group convened on Wednesday afternoon, it didn’t take long for the lively discussion and heated debate to begin. The group initially evaluated the general concept of mortality rate plateaus as well as several potential models for understanding mortality in general. In the process of deliberating over the proposed late-life plateau in mortality rates and what it could represent, the group also discussed a number of different angles from which to approach the problem: at an individual or population level or from a biological, societal, or financial perspective. Because the task group was charged with trying to understand what happens during the mortality rate plateau, the group spent a long time earnestly discussing whether disability and disease also plateau. This is significant, because if disability increases exponentially while mortality rates plateau, it could mean that individuals are surviving with long-term 

mortality rates, aging, and functional human healthspan 45 debilitating illnesses. On the other hand, if disability or disease plateaus, there may be a fundamental difference in the very old that could reveal secrets about aging in general. The state of those in late life is also criti- cal to understanding the question of healthspan and to recognizing the societal and financial implications of late-life mortality plateaus in human populations. In this context the group thought it was important to draw a clear distinction between disability and impairment. For the purposes of this dis- cussion, the group classified disability as a societal and functional problem, based on a person’s circumstances. To some extent disability depends on the technology and assistance available to the individual. Impairment, on the other hand, is related to the loss of a specific function or set of functions that can be measured objectively (e.g., using cognitive or perception tests). The group also bandied about ideas about just how big a part the underlying biology plays in the late-life mortality plateau. Published data showing a late plateau in age-specific fecundity in large fruit fly populations suggests there are fundamental biological differences in the latest stages of life. As well, one group member noted that the mortality rate plateau was initially interpreted as evidence for plasticity in lifespan. There was a great deal of discussion within the group about this interpretation and how we should define plasticity, if at all. For example, one option was that the mortality plateau might represent more permissive genetic heritage than previously imagined. The group also discussed whether it is possible to improve people’s quality of life at every age and, if so, how function and quality of life should be measured. Some suggestions included everything from technology that assessed voice and motion to cognitive tests to molecular markers such as cytokines. While the group did not reach a conclusion about the most ef- fective tests or how frequently sampling should be, the general consensus was that it would be advantageous to sample as often and as completely as feasible. One group member noted that in Drosophila even with the fecundity plateau, female fecundity decreases rapidly a few days before death, suggest- ing there may be markers in other species that indicate healthspan decline. If so, there may well be opportunities to manipulate the projected time of death and to decrease the amount of time before death spent with degraded health. On Thursday morning the task group presented their preliminary report to the other meeting attendees. But the feedback was mixed. Some 

46 the future of human healTHspan seemed unconvinced about the demographic evidence for a mortality pla- teau. Gerontologists pointed out that their oldest-old patients may survive longer, but they continue becoming frailer with time. This countered the task group’s assertions that aging itself—including disease load—may be curtailed in late life. Others suggested that selective survival explained the plateau in impairment and declining risk factors observed in some studies. It was back to the meeting room for more diligent debate. For instance, at least one group member suggested that it might be rash to draw analo- gies between fruit fly and human plateaus, given less convincing evidence for mortality plateaus in nonhuman mammalian systems such as mice. In the end, though, the group concluded that late-life mortality plateaus were worth exploring further—and in their final presentation the group explained their vision of how to tackle the problem. Future Challenges and Recommendations The group concluded that the demographic data suggested potential qualitative differences in late-life biology. For example, not just mortality but also disease load might plateau or even decline in later life. The question is whether these phenomena are related and, if so, how. To adequately evaluate this, the group recommended creating adequate measures for disability and impairment in late life as well as exploiting the appropriate measures that already exist. For individual and population measures already available, this includes determining the precision, bias, information, and applicability. The group also noted that several kinds of data, assessed on multiple scales, would likely be required to get to the bot- tom of the mortality plateau conundrum. Once appropriate measures are available, it should be possible to evalu- ate the age-associated changes both across populations and within individu- als in late life. This could be useful in addressing the broader questions of mortality and healthspan. For instance, better measures of impairment and disease might reveal distinct groups and vulnerabilities within these groups. A variety of noninvasive technologies were proposed—from imaging tech- niques to cognitive performance tests to the “omics” (e.g., metabolomics, proteomics, and genomics). Throughout the meeting the group struggled to reconcile the demo- graphic data with the broader concepts of healthspan and aging. It wasn’t until they drew up a conceptual model that the problem started to coalesce. The model represents a mathematical interconnection between aging, 

mortality rates, aging, and functional human healthspan 47 mortality rate, and healthspan that can theoretically accommodate various kinds of data. For example, age and healthspan can be related to each other based on the absence of disability or impairment, but also other measures of health. Since mortality is not always a function of age, the process of aging itself could theoretically be broken down into a Gompertz phase and a non-Gompertz phase. This could also help to distinguish between total life expectancy and active life expectancy. The group narrowed its focus to look at the example of a very healthy 95-year-old. They speculated that if it were possible to have perfect knowledge about the individual, it might also be possible to understand the relationship of age with healthspan and mortality. For instance, if death and health declines are completely stochastic, it might not be possible to predict health declines or death. On the other hand, one mechanism might start an inevitable decline in health ending in death, making it relatively simple to predict both decline and death. Finally, death might only be predictable after an initial health decline. In other words, after health starts deteriorating, it could allow other mechanisms that lead to death. Similarly, by understanding the oldest-old in our population using centenarian studies we might gain a better understanding of the biology, en- vironmental exposures, and/or lifestyles that are incompatible with making it to the ripe old age of 100. As this task group put it, studying centenarians might “tell us where the mines are buried.” The group also recommended using numerous animal models to try to gauge the physiological markers of healthspan. This approach may hold a great deal of promise in the short term, especially in Drosophila, where it is reportedly possible to not only observe but also manipulate the mortality plateau. As a result, the statistical methods may be easier to assess in these model systems. Animal models could also be the key to understanding what is biologically possible. Finally, the group noted that from a public health perspective, the late-life plateau may have minimal relevance if our society cannot ad- equately deal with known causes of early death, such as smoking, obesity, malnutrition, and related diseases. This is because much of the impairment and disease burden carried by the elderly results from long-term, chronic exposures that individuals experience earlier in their lives, underscoring the importance of primary prevention in early and midlife for avoiding impairment and disability. Even so, although some continue debating it, 

48 the future of human healTHspan the mortality rate plateau presents the intriguing possibility that improving health is possible at any age. And we are just starting to understand just how malleable lifespan—and healthspan—might be. NOTES 1. J. R. Carey, P. Liedo, D. Orozco, and J. W. Vaupel. Slowing of mortality rates of older ages in large medfly cohorts. Science 258(1992):457-461. 2. J. W. Curtsinger, H. H. Fukui, D. R. Townsend, and J. W. Vaupel. Demogra- phy of genotypes: Failure of the limited life span paradigm in Drosophila melanogaster. Science 258(1992):461-463.