Panel Members
Richard M. Cawthon, M.D., Ph.D.
University of Utah
Steven R. Cummings, M.D., FACP
University of California, San Francisco
J. David Curb, M.D., Co-chairperson
John A. Burns School of Medicine
Douglas C. Ewbank, Ph.D.
University of Pennsylvania
Jeffrey Kaye, Ph.D., Co-chairperson
Oregon Health Sciences University
Richard A. Kerber, Ph.D.
University of Utah
Thomas T. Perls M.D., MPH
Boston University
Eugenia Wang, Ph.D.
University of Louisville
Elad Ziv, M.D.
University of California, San Francisco
NIA Staff
Evan Hadley, M.D.
Geriatrics and Clinical Gerontology Program
Jennifer Harris, Ph.D.
Behavioral and Social Research Program
Anna McCormick, Ph.D.
Biology of Aging Program
Marilyn Miller, Ph.D.
Neuroscience and Neuropsychology of Aging Program
Marcelle Morrison-Bogorad, Ph.D.
Neuroscience and Neuropsychology of Aging Program
Winifred K. Rossi, M.A.
Geriatrics and Clinical Gerontology Program
Richard Suzman, Ph.D.
Behavioral and Social Research Program
Huber Warner, Ph.D.
Biology of Aging Program
Introduction
Exceptional individuals whose productive life span greatly exceeds the average have long been subjects of fascination. Understanding the factors that contribute to exceptional longevity and/or exceptional “health span” (survival without disease or disability) could lead to better means to maintain health and prevent disease throughout life. As knowledge of biology, medicine, and public health grows, opportunities to identify such factors will continue to increase.
While the very old share the characteristic of unusually long survival, they differ within and across groups in sociodemographic and socioeconomic characteristics, lifestyle, and health profiles. The ability to interpret results of studies on the exceptionally old has been limited by a lack of descriptive information on demographic, health, and functional status of the subjects studied. The development of a set of standard measures that characterize various aspects of individuals in study populations would clarify analyses of risk and protective factors by specifying more thoroughly the phenotypes to which these factors are related, and allow better synthesis and comparison of results of studies in this field.
Such measures would be especially useful because much scientific interest in exceptional survival is not on longevity per se, but survival with preserved health and/or function. A variety of such phenotypes exist, e.g., survival with intact cognition, survival without physical disability, or survival without severe chronic disease. Another set of exceptional survival phenotypes of interest, in persons with chronic diseases or risk factors, includes survival to a much older age than expected, given their conditions or risk factors. There is no single “correct” exceptional survival phenotype: many are of public health and biologic interest, and each may be affected by different risk factors. However, whichever is chosen, interpretation of results would be enhanced by a precise specification of the phenotype, including measures selected from a core set that was used in all studies of exceptional survival. This would be especially true if the primary data on these core measures collected in individual studies were available for subsequent analyses for comparisons and pooling with other studies.
Widespread use of such standard measures could enhance research on exceptional survival in the following ways:
- For studies of factors predicting exceptional survival or associated with exceptional survival, it would provide more precise information about the “survival” phenotype to which these factors are related.
- It would facilitate analyses of the relationship of factors to various exceptional survival phenotypes, by allowing delineation of subgroups within studies who meet specific sociodemographic, health or functional criteria (e.g., survival to a specified age without disease). The ability to examine more homogeneous subgroups may aid in the detection of predictive factors.
- In considering results of different studies on factors predisposing to or against exceptional survival, it would enhance the ability to determine whether apparently discrepant results between studies are explicable by phenotypic differences among the study populations.
- The use of common measures enhances the potential for pooling data on subjects from more than one study who meet specific phenotypic criteria, thereby providing greater statistical power.
To address these issues, the National Institute on Aging convened a panel in November 2001with clinical, demographic, epidemiologic, and genetics research expertise, to identify a standard set of measures that could be used to characterize individuals or populations with exceptional survival (ES). The term “exceptional survival” was selected because, in contrast to “exceptional longevity” (which conventionally refers only to attained age per se), it may be applied to other important survival outcomes such as survival without disease and or disability, as well as to longevity per se.
The panel recommended measures in the following domains: age, sociodemographic characteristics, health, functional status, and psychological characteristics. The panel made recommendations for two levels of measures: a minimum set to be used in all ES studies, and more intensive or detailed measures that are important for studies with emphasis on particular domains. The recommendations for the minimum set of measures are summarized in the Table, and described below, along with recommendations for more detailed or intensive measures.
The recommended measures are intended to complement additional measures reflecting the foci of specific studies. Their use could enhance the knowledge gained in various studies that examine differing exceptional survival phenotypes, and factors that influence, by enhancing opportunities to pool and compare data across studies. The panel recognizes that persons with exceptional survival in different countries and social groups will be studied. Different social and cultural environments in themselves may affect the significance of specific diagnostic and function outcome measures. Hence in many cases, it will be important to use additional measures to accommodate these differences. However, use of the recommended core measures as a point of reference should nonetheless improve comparability across studies.
Age
Age determination. At a minimum, stated age should be validated with a birth certificate or other documentation. In studies in which age validation is especially crucial (e.g., on supercentenarians), additional forms of proof of age at different times in individuals’ lives (e.g. census records, marriage license, school report card with age noted, employment record, parental age on child’s birth certificate) and familial reconstitution may be necessary. In family studies, analyzing as many generations as possible to ascertain agreement of multiple family members’ birth dates with census records is important. These approaches can detect, among other things, the use of a false identity, for example that of a relative from a previous generation.
Survivorship of study subjects relative to source population. Since life spans vary in different groups, subjects of a particular age range may constitute a more or a less exceptional degree of survivorship, depending on the source population for the study. To take these differences into account, life span information on the source population should be obtained and reported when possible. When possible, the age range of study subjects should be presented in terms of percentile of survivorship, as well as chronologic age.
In many cases, the only data on source populations will be for a broader population (e.g., national censuses) than the source of study participants, but in other cases survival data on more specific source groups will be available. Where possible, percentile survivorship data should be adjusted for major demographic factors such as gender, educational level, and income. In making comparisons with other groups it is helpful not only to understand where individuals fall within the age distribution of their own group, but also where they stand within survival of those individuals in comparison groups.
Sociodemographic Characteristics
Birth Cohort. Many studies of exceptional longevity include multiple generations of families, or longevous individuals whose birth dates span more than a decade. There have been marked changes over the last century in the proportion of individuals surviving to very advanced age. Diseases and trauma that once prevented some individuals from surviving to extreme old age are now being prevented, delayed or at least palliated. In addition, exposure to different diseases at various ages differs among different birth cohorts. These factors not only influence the degree of “exceptionality” reflected by a given age, but also the interpretation of factors influencing survival. The presence or absence of a major infectious disease epidemic or war during a birth cohort’s life span may influence the relative strength of factors affecting survival to old age. To increase the interpretability results of studies on exceptional survival, the panel recommends that, at a minimum, the distribution of ages of birth be reported by decade of birth.
Other Sociodemographic Characteristics. Source populations from which exceptionally long-lived individuals are drawn can vary in other sociodemographic characteristics besides age structure. Such characteristics have been associated with differences in survival time (1-3), and are thus important factors to be included in analyses of exceptional survival. Therefore, all studies of exceptional survival should provide standard collection of a minimum set of sociodemographic covariates that may significantly affect survival to extreme old age. At a minimum, information should be reported on sex, race, ethnicity, education, smoking, and place of residence. Where possible, data on parental ethnicity, socioeconomic status, access to health care, physical activity and, diet should also be provided.
Health
Studies on exceptional survival, at a minimum, should document self-reported history and age of onset of the following major chronic diseases, using ICD definitions and codes:
- Diseases of the heart
- Malignant neoplasms
- Cerebrovascular diseases
- Chronic lower respiratory tract diseases
- Diabetes mellitus
- Alzheimer’s disease
(For cognitively impaired subjects proxy reports should be used where appropriate. (4))
Comorbidities should also be assessed using a standardized scale of illness severity, the Cumulative Illness Rating Scale (CIRS). (5)
Studies with a strong focus on health and comorbidity should collect and report additional data via medical chart review, physical examination, or autopsy, and, when possible, noninvasive clinical measures such as blood pressure, heart rate, body mass index, pulmonary function measures such as FEV1, bone density, presence or absence of cataracts, auditory acuity, and circulating levels of creatinine and inflammatory markers such as C-reactive protein. Storage of serum for future additional phenotypic measures is recommended if feasible.
Functional Status
Exceptionally old persons vary widely in functional status. Hence information on the prevalence of disabilities of subjects in specific studies is crucial for understanding the nature of the population under consideration. Use of widely accepted assessment instruments would facilitate comparisons among studies.
Activities of daily living. Disabilities in these activities are extremely common in the very old. Studies should assess subjects’ functional status using the basic Activities of Daily Living (ADLs) scale (6) or the Barthel scale (7). While each of these scales has been validated, there may be unique clinical and research situations that make one preferable to the other. Studies should also assess higher-level function Instrumental Activities of Daily Living (IADLs). (8)
Where possible, performance-based measures of physical function should be included in studies of exceptional longevity. These provide a check on self-reported data on physical function, and have predictive validity for subsequent disability. A minimal assessment of physical performance should include gait speed, timed chair stands, and grip strength. (9-10).
Cognitive function. The minimum goal of cognitive screening in studies of exceptional survival population must be to determine if there is some level of cognitive impairment that suggests dementia. However, it is likely that screening tests alone in this population will underestimate the degree of cognitive impairment. Ideally cognitive screens should minimize cultural and educational biases and be less sensitive to sensory and motor changes frequently associated with exceptional survival.
Studies should include an assessment of general cognitive functioning. In general the Mini-Mental State Examination (MMSE) or a similar instrument should be used. If MMSE is not used, data on the alternate tool’s comparability to MMSE scores should be reported. (11)
In studies examining dementia-free survival, diagnosis of dementia in individuals with low cognitive function scores should be conducted utilizing a standardized interview and neurological examination and the neuropsychological test battery from the Consortium to Establish a Registry for Alzheimer's disease (CERAD). (12) In general self or proxy reports of cognitive performance should not be used. Remote (telephone) and abbreviated cognitive assessment tools may be of limited validity, especially in the oldest individuals.
Psychological and Social Characteristics
Coping abilities and personality may be important in exceptional survival. Personality traits and coping strategies may be assessed using the Sixteen Personality Factor Questionnaire (13) or the Neuroticism, Extraversion and Openness (NEO) Inventory. (14)
A screen for depressive symptoms should also be part of a minimum set of measurements. Examples using self-rating (implying that some level of adequate cognitive capacity is present) include the Beck Depression Inventory, the Zung Self-Rating for Depression Scale, the Geriatric Depression Scale, the Hamilton Depression Rating, and the Center for Epidemiological Studies – Depression Scale. (15-19)
The degree of participation in social networks is an important characteristic of older persons that is related to risk for adverse events. It should be characterized at least briefly in studies on exceptional survival. An abbreviated version of the Duke Social Support Index with 11 items has been described (20). Of these, the first four items alone make a simple, compact module of social network participation that can be easily compared across studies.
Group Survival Characteristics
In some studies of exceptional survival, the focus, or the sampling frame, is based on the survival properties of a group of individuals (e.g., pedigrees with a high proportion of members reaching exceptionally old age, relatives of centenarians, or individuals with a particular genotype that is hypothesized to contribute to exceptional survival). Survivorship to exceptional old age is affected by survival rates during all segments of the life span. The pattern of survival rates in different age segments may vary among groups, e.g., some having especially high survivorship only in early life, others only in late life. Such differences are of biological and health significance. Hence, in studies involving long-lived families or other groups, data on survival curves and age-specific
mortality rates of over the life span should be collected and reported, rather than merely average life spans, or the proportion of individuals reaching exceptional age. In the reporting of such data, individuals whose survival data are censored because they are still alive, should be distinguished from those who have died.
Data Reporting and Availability
At a minimum, data on the recommended measures in the foregoing domains should be reported in publications as means (or prevalences) and standard deviations.
Where appropriate and where publication page limits permit, more detailed characterization of distributions of values of these measures, and differences among study subgroups (e.g., genders) in publications is useful. Cohort life tables (if available) in which at least some of these factors are taken into account can be helpful in determining how the factors affect probability of survival to extreme old age.
If this information is too lengthy for publication, it is recommended that it be available in a publicly accessible format (e.g., website) if possible, or furnished by the authors upon request. Maintenance of the original data in a form that would allow grouping into different categories that may be desirable for subsequent analyses of individual studies or pooled data from several studies is encouraged.
Conclusion
The use of such a core set of common measures is consistent with investigations of differing exceptional survival phenotypes of interest by different investigators, while maintaining the ability to pool and compare data on a particular phenotype across studies. Adopting these common measures in studies of exceptional survival should aid in clarifying the characteristics of exceptional survivors, as well as the role of contributory factors to a variety of exceptional survival phenotypes. In addition, as data from these and other measures accumulate, they may aid in clarifying commonalities and differences among groups with different exceptional survival phenotypes, and aid in the further refinement of tools for their characterization.
References
- Sorlie, P. D. , Backlund, E. & Keller, J. B. (1995) US mortality by economic, demographic, and social characteristics: the National Longitudinal Mortality Study
Am. J. Pub. Health 85, 949-956
- Elo IT, Preston SH. Educational differentials in mortality: United States, 1979-85.
Soc Sci Med 1996 Jan;42(1):47-57
- Vaupel JW, Jeune B. The emergence and proliferation of centenarians. In: Jeune B, Vaupel JW, eds. Exceptional longevity: From prehistory to the present. New York: Springer, 1995, pp 109-116.
- International Classification of Diseases, 9th Revision Modification Clinical Sixth Edition National Center for Health Statistics, Health Care Financing Administration, 2001.
- Miller MD, Paradis CF, Houck PR, Mazumdar S, Stack JA, Rifai AH, et al. Rating chronic medical illness burden in geropsychiatric practice and research: application of the cumulative illness rating scale. Psychiatry Res 1992; 41: 237-248.
- Katz S, Dowens TD, Cash HR. Progress in development of index of ADL. Int J Health Serv 1976;6:493-508.
- Mahoney FI, Barthel D. Functional evaluation: the Barthel Index. Maryland State Med J. 1965;14:56-61.
- Lawton MP; Brody EM. Instrumental Activities of Daily Living Scale (IADL) IN: Gallo JJ; Reichel W & Andersen L. Handbook of geriatric assessment. Rockville, Md. : Aspen Publishers. Pg.71-73, 1988
- Guralnik JM, Simonsick EM, Ferrucci L, et al. A short physical performance battery assessing lower extremity function: association with self-reported disability and prediction of mortality and nursing home admission. J Gerontol 1994;49:M85-M94
- Rantanen T, Masaki K, Foley D, Izmirlian G, White L, Guralnik JM. Grip strength changes over 27 yr in Japanese-American men. J Appl Physiol 1998 85(6):2047-53
- Folstein MF, Folstein SE, McHugh PR, 1975. Mini-mental state: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 12:189-198
- Welsh K, Butters N, Hughes J, Mohs R, Heyman A. Detection of abnormal memory decline in mild cases of Alzheimer's disease using CERAD neuropsychological measures. Arch Neurol 1991 Mar;48(3):278-81
- Cattell RB, Eber HW, Tatsuoka MM. Handbook for the Sixteen Personality Factor Questionnaire (16pF) 1907. Champaign, IL.
- Costa PT Jr, McCrae RR. Domains and facets: Hierarchical personality assessment using the Revised NEO Personality Inventory. J Personality Assessment, 1995; 64:21-50.
- Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. An inventory for measuring depression. Archives of General Psychiatry, 1961; 4, 561-571.
- Zung WW, Richards CB, Short MJ. Self-rating depression scale in an outpatient clinic. Further validation of the SDS. Arch Gen Psychiatry;13(6):508-15
- Yesavage JA. Geriatric Depression Scale. Psychopharmacol Bull 1988;24(4):709-11
- Hamilton M. A rating scale for depression. J Neurology Neurosurgery Psychiatry. 1960;23:56-62.
- Radloff LS, Teri L. Use of the Center for Epidemiological Studies-Depression Scale with older adults. Clinical Gerontologist. 1986;5 (1-2)(Jun):119
- Koenig HG, Westlund RE, George LK, Hughes DC, Blazer DG, Hybels C. Abbreviating the Duke Social Support Index for use in chronically ill elderly individuals. Psychosomatics 1993 Jan-Feb;34(1):61-9
Acknowledgments
The panel is grateful to Ken R. Smith, Ph.D., University of Utah, for advice on items relating to sociodemographic status and social support.
