Limits to Human Life Expectancy: Evidence, Prospects and Implications

Citation: Manton, et al. (1991) Limits to Human Life Expectancy: Evidence, Prospects and Implications.
Internet Archive Scholar (search for fulltext): Limits to Human Life Expectancy: Evidence, Prospects and Implications
Tagged: uw-madison (RSS), wisconsin (RSS), sociology (RSS), demography (RSS), prelim (RSS), qual (RSS), WisconsinDemographyPrelimAugust2009 (RSS)

Summary

There are three major schools of thought on the debate over limits to human life expectancy. "Traditionalists"argue that the limit is not much greater that current life expectancy in developed nations-that is, around 85 years. "The limit is viewed as due to biological senescence [i.e., natural causes associated with aging], which is not affected by changing the mortality of specific causes" (p. 603). "Visionaries" believe that biomedical research will raise the limits imposed by senescence. Humans with modified senescence could then arguably live an average of 125 years or more. Finally, an "empiricist" view "contends that we are not currently near a life expectancy limit, because mortality is declining and progress is being made in the treatment and management of the chronic diseases and disabilities that dominate mortality at later ages" (p. 603). Although no limits are imposed, empiricists observe that if recent mortality declines of about 2 percent per year continue, life expectancies will be around 95-100 by 2080 in developed nations. To evaluate these perspectives, Manton et al.(1) review the traditionalist and visionary estimates of limits to life expectancy, and their methods and data, (2) examine long-lived populations to determine a lower bound to a theoretical limit, and (3) build a model to estimate life expectancy limits using information on multiple, time-varying risk factors. A common approach used by traditionalists and visionaries is to eliminate "exogenous" causes of death (e.g., car accidents and infectious diseases) and then estimate life expectancy only by the remaining "endogenous" causes. However, increases in life expectancy need not be due to cause elimination-delayed disease onset alone can result in longer life. For this and other reasons (e.g., failure of such models to account for risk factor reduction), cause elimination models can estimate "the effect of eliminating individual causes of death . . . but not for determining a life expectancy limit, because, as more causes are eliminated, life expectancy increases nonlinearly without limit" (p. 610). That said, Olshansky et al. (traditionalists) use this technique to estimate a theoretical maximum life expectancy of 82.0 for males and 88.0 for females (1990). Similarly, Strehler (visionary) uses cause elimination to estimate a theoretical maximum life expectancy of 85-90 (1975). When an added assumption of "Gompertzian kinetics" (i.e., altered senescence) is introduced, the maximum increases to 100-125. Traditionalists also utilize methods of "rectangularization" to estimate theoretical limits. As exogenous causes of death are eliminated, only senescence-the universal decline in the physiological state of organisms with advancing age-will remain, thus identifying the limits to human life expectancy. Using such a technique, Fries (1980) estimates that the limit is 85.6 for both sexes. However, "There is a problem in assessing 'rectangularization' regardless of the procedure used: namely, if genetic heterogeneity in all ages at death exists, the 'ultimate' survival curve is not 'square'" (p. 612). Since mortality data do not contain enough information to produce "meaningful estimates of life expectancy," Manton et al. review populations long-lived populations (p. 613). All of these populations live under conditions that are conducive to healthy lifestyles-namely (a) avoidance of risk factors (e.g., smoking), (b) nutritional control and (c) exercise. The populations reviewed are generally religious groups (e.g., Mormons) who strongly encourage healthy personal behavior-although nutritional habits were not necessarily optimal (e.g., no restrictions on animal fats). Using standardized mortality ratios, Manton et al. translate the SMRs into life expectancy estimates of anywhere from 85 to 100 years. Given the logic of statistical inference, these populations must represent a lower bound to a theoretical limit. It is important to note that these life expectancies exceed "traditionalist" estimates. Finally, Manton et al. construct a multivariate model of chronic disease risk factors to estimate a limit to life expectancy. The sample of 5,029 males and females came from the Framingham Heart Study. Using 11 measures of risk factors (e.g., blood pressure) that were set to "optimal levels" and a Gompertz function to represent biological variability, Manton et al. estimate a theoretical maximum of 99.9 for males and 97.0 for females. The traditionalist view is refuted both by some observed human populations and Manton's model. "With reference to the United States, a life expectancy of 100 has implications for the Social Security and Medicare Trust Funds, private pension systems, health insurance, and the health care system" (p. 631).