Biological age--a concept whose time has come: a preliminary study.

Objective: Chronology poorly predicts biological age (BA) or physiologic reserve (PR). An objective approach to the heterogeneity of aging would greatly help clinical decision making in the elderly.

Materials And Methods: The first pilot study evaluated 130 "healthy" volunteers, ages 70 to 95 years. A summary BA/PR index was developed, using measures of endurance, strength, flexibility, balance, cognition, depression, comorbidity, and exercise. The second study applied the BA/PR concept to prediction of death after a first elective coronary artery bypass graft, using a Veterans Administration database.

Results: The BA/PR index was a better predictor of 3-year functional outcomes and death than was chronological age. In the coronary artery bypass graft study, the inclusion of BA/PR variables significantly improved prediction of 6-month and long-term death for Veterans Administration patients.

Conclusions: The usefulness of a biological age (BA/PR) approach in predicting outcomes in the elderly was supported. Needed research should develop tools for routine "tracking" of the aging process.

**********

Elderly people constitute the most rapidly increasing segment of the US population and present the highest health care costs, per capita and as a group. Treatments and procedures have demonstrated effectiveness in this population, albeit with more complications and higher mortality rates than in younger patients. (1-9) As individual differences increase with age, levels of physiologic and functional ability range from the very fit to the very frail beyond age 70 years. This marked heterogeneity of old age must be characterized objectively if optimal selection of patients for treatments and interventions is to be achieved.

The Venn diagram (Fig. 1) illustrates the concept of physiologic reserve as the outcome of the interaction of aging, lifestyle, and disease. The assessment of aging per se is particularly complicated because age-related functional decline varies widely in reference to specific organ systems (10) and because subclinical disease is not easily differentiated from the effects of aging or lifestyle changes. (11) Physiologic reserve (PR) assessment is the key to estimating biological age (BA), since a loss of reserve is the hallmark of aging.

The concept of BA has been studied for decades. (12) Biomarkers of aging are appealing conceptually, but relating biomarkers to functional capacity or biological aging has been elusive, (13,14) and attributing variations in aging to genetics is not supported by the data. (15-17) Most investigators have concluded that no single or simple combination of biological or genetic markers is likely to provide a useful clinical estimate of aging. (18-22) Geriatricians have documented the clinical utility of using a variety of measures and markers of functional status in multiple studies of elderly populations or large samples, often limited by frailty and/or disease and dementia. (24-40) Regulatory pressures now require nursing homes to estimate functional status. (22,41) Biological age is also getting attention in the lay press, and several web sites are now providing questionnaires to "Calculate Your BA." (23)

[FIGURE 1 OMITTED]

We have explored the usefulness of the BA/PR concept in assessing aging heterogeneity. To establish proof of principle, we conducted two pilot studies. The first study demonstrated that functional heterogeneity is present and can be objectively assessed, even in a restricted narrow sample of "healthy" high socioeconomic status (SES) elderly persons living independently. The second study applied the PR/BA concept to coronary artery bypass graft (CABG) surgery outcomes, using a Veterans Administration (VA) database. (42-44)

Both pilot studies test the hypothesis that BA/PR measures can be used to characterize aging heterogeneity and will provide important information about outcomes, even with narrow samples and admittedly imperfect variables. Simply stated, the host, not just the disease, is a major determinant of outcomes in the elderly.

Study I: Assessing Heterogeneity and Developing a Biological Age Index

Design

Recruitment and procedures. After institutional review board approval was given, 130 volunteers over the age of 70 years were recruited from a wellness center at Vanderbilt University, a residential retirement center, and the senior club of a local church. We selected "healthy" elderly individuals with relatively high SES. People with acute illnesses or overt functional limitations were excluded. A medical history and a physical/functional evaluation were conducted, using validated tools selected for their perceived ability to characterize fitness as well as frailty.

Glucose, albumin, hematocrit, and creatinine were determined from a fasting blood sample. Physical performance measures included the 6-minute walk test for overall fitness and cardiovascular endurance, (45,46) two measures of balance, the Tinetti Balance and Gait scale, (47,48) and the timed "one foot balance," (49) the standing forward reach flexibility test, (50) the hand dynamometer grip strength test, (51,52) and a strength and velocity measure (30-seconds/number of hand weight lifts: 4 pounds for females, 8 pounds for males). Cognitive performance testing included the Mini-Mental State Examination (MMSE), (53,54) the modified Mini-Mental State (3 MS), (55) the Controlled Oral Word Association test (COWA), (56) and the Trail Making test forms A and B. (57) Depression (Geriatric Depression Scale), (58) SES, exercise participation, basic activities of daily living (ADL) (59) and instrumental activities of daily living (IADL), (60) nutritional status (Mini Nutritional Assessment), (61) and comorbidity (Charlson index) (62,63) were evaluated by self-report and interview. The need for help with basic and instrumental activities of daily living skills (ADL and IADL) was assessed at baseline and at 1-year, and 3-year follow-up. ADL activities included bathing, dressing, grooming, transfer in and out of bed, getting to the bathroom on time, eating, and walking. IADL activities were telephone use, getting to places, shopping, preparing meals, housework, taking medicines, handling money, and recent changes in driving and housing.

Statistical analyses. Descriptive statistics were calculated for all variables. A BA/PR index was developed by psychometric criteria from classical test theory, (64) namely by rejecting items with inadequate part-whole correlations less than 0.10. The part-whole correlation is the correlation between one item and the sum of all the other items. Items that do not correlate with other items reduce a test's internal consistency reliability because they measure something extraneous. Rejecting extraneous items maximizes the Cronbach [alpha] internal consistency reliability for the resulting index, and this reliability forms an upper limit for the predictive validity of the index.

Scores on the selected measures were standardized (mean = 0; SD = 1) and the BA/PR summary score was computed, weighing each item equally by adding individual z-averaged scores. The ability of the index to predict 3-year categorical outcomes (death, transfer to assisted living, or ability to drive) was investigated by using [chi square] tests. For continuous outcomes (number of ADL and IADL limitations), a longitudinal multilevel regression (66) with random coefficients was run with SAS PROC MIXED. (67) This was applied to repeated measurements at baseline and at 1-year and 3-year follow-up. In this hierarchical linear (68) or random regression (69) slope-as-outcome model, the most important result is the slope of change in the number of ADL or IADL limitations over time. (69)

Results

Mean sample age was 78.8 [+ or -] 6.25 (range, 69 to 95), with 22.5% of the participants 85 years or older and 60% women. SES was high, with 15.7 ([+ or -] 2.6) years of education and a yearly income in excess of $50,000. Abnormal laboratory values were rare [eg, albumin (1.5%), hematocrit (2.4%), and creatinine (5.3%)]. Self-perceived health was rated as "as good" (34.8%) or "better" (56.8%) than same-age individuals. One participant on dialysis and another one with much lower education (5 years) and income were excluded from the analysis.

A BA/PR index score was obtained by adding z-averaged scores on nine selected measures with part-whole correlation greater than 0.10. The components of the BA/PR index are listed in Table 1. Significant sex differences were only found on the 6-minute walk; z-scores were therefore calculated within each sex for this test. The distribution of BA/PR scores (mean = 0, SD = 5.00, range = -14.78-10.51) showed significant non-normality (Shapiro-Wilk W = 0.97, P < 0.01), with a moderate skew (skew = 0.59 toward disability). The internal consistency of the nine-item index, or its ability to measure a single dimension, showed adequate reliability (Cronbach [alpha] = 0.77) and indicates that scores on the nine measures share considerable communality. A standard part of standard internal consistency analysis is reviewing the impact of each item's deletion on the overall Cronbach [alpha], since deleting items usually lowers reliability, as indicated by the well-known 1910 Spearman-Brown relation. (70,71) For example, deleting the Charlson index would have lowered the [alpha] from 0.774 to 0.771. With adequate reliability often considered as [alpha] = 0.80 or higher, (65,72) all nine items were retained in the index.

The negative correlations with age, ranging between r = -0.19 for the Tinetti balance scale and r = -0.56 for the 6-minute walk test, indicate a modest decline in performance with advancing age (Table 1). Timed tests such as "time standing on one foot," Trails B, and the 6-minute walk showed the widest performance ranges (eg, at age 85 years, distance walked ranged from 500 to 2,000 feet.

Table 2 presents 3-year mortality rates and functional outcomes for four categories of participants classified by median-split values on chronological age ("old" = age 79+; "young" = age < 79) and physiologic reserve ("frail" less than median and "fit" greater than or equal to median BA/PR index score). Table 2 presents data for 82% of cases with completed follow-ups (n = 106). At the 3-year follow-up, 11 of the original 128 volunteers had missing data (7 refused to answer, 2 had moved, and 2 were lost to follow-up), and 11 had died. Nine of the 11 deaths at 3 years had occurred in the old/frail category (P < 0.01) and none in the old/fit group. The remaining two deaths in the fit/young category were due to a dissecting ascending aortic thoracic aneurysm and metastatic prostate cancer. All health-related moves to assisted living housing facilities (n = 4) were found in the old/frail group (P < 0.01). Being both biologically frail and chronologically old also predicted loss of ability to drive a car (P < 0.001).

Changes in basic and instrumental activities of daily living (ADL and IADL) were investigated longitudinally from baseline to 3-year follow-up as a function of time, chronological age, and BA/PR scores. The ADL outcomes did not show statistically significant group differences in basic activities of daily living changes. All four groups showed an increase in instrumental (IADL) limitations over time, starting with an average of 0.39 limitations at baseline and gaining 0.29 limitations per year (P < 0.05). Figure 2 presents the slopes of IADL changes. Those who started with more limitations deteriorated faster (r = 0.57, P < 0.0001). Chronological age did not predict IADL baseline values, but the BA/PR scores did. Chronologically older individuals did not deteriorate more quickly (P = 0.30) if they were biologically fit. The two biologically old, low BA/PR groups gained an additional 0.48 IADL limitations per year (P < 0.0001), as compared with the two biologically young (or "fit") groups. The old/old or "frail" participants (mean age = 86, n = 44) deteriorated even faster (+0.14 limitations per year) than the frail/young group (age = 75 years, n = 21).

[FIGURE 2 OMITTED]

Study II. BAI/PR and CABG Surgical Outcomes

Design

Study data base

The primary goal of the Department of VA Processes, Structures and Outcomes of Care in Cardiac Surgery (PSOCS) study was to link processes to outcomes of care. (42,73-74) The PSOCS data involved 4,969 patients with CABG and were collected by trained research nurses within 72 hours of surgery in 14 VA hospitals (September 1992 to December 1996); late death was recorded. Our secondary analysis of the PSOCS database was restricted to male patients undergoing first elective CABG surgery (n = 3,077), without prior cardiac surgery, and who had completed the Short-Form Health Status (SF-36) quality-of-life measure. (33) Of those 2,205 patients with CABG, 606 were 70 years of age or older (mean = 73.6, SD = 3.1) and 1,599 age 70 years or younger (mean = 59 years, SD = 7.5)

Selection of variables

Since several of our first BA/PR study measures were not part of the 315 variables/PSOCS database, we selected 45 of the closest BA and disease indicators available. The 21 aging and lifestyle predictors (BA) were mostly self-report measures of functional status (ADL, exercise, mobility), quality of life (SF-36 mental and physical), and lifestyle (smoking, education, social support, employment status, weight loss, and physical activity). Also included were peak expiratory volume (FE[V.sub.1]) and cognitive tests (Trails A, neurotap test). The 24 disease indicators (cardiovascular disease and comorbidity) were classified into high or low technology. Hi-tech disease markers (n = 11) required specialized equipment or highly trained personnel (eg, number of diseased coronary vessels, electrocardiographic interpretation, or cardiomegaly on radiography). Low-tech disease indicators could be reliably based on the patient's self report (eg, diuretic use, angina class, prior heart surgery, or myocardial infarction). A table with descriptive statistics on the 45 variables for the under-70 and over-70 age categories can be obtained from the corresponding author. Outcome measures were 6-month death and long-term survival. The number of years of survival was censored by length of follow-up with patients counted as alive or dead at the time of data collection (from 3.2 to 7.5 years after surgery).

Statistical analyses

Regression analyses were used to study the relative contribution to outcomes (Ys) of the three sets of predictors (Xs = high tech disease, low-tech disease, low-tech biological age). The model testing the incremental value of aging and lifestyle [Y = F (high tech + low tech disease), then BA] favors disease indicators by forcing them first into the model. Logistic regression was used to test 6-month outcomes, whereas long-term survival analyses used Cox regression. (77) Each analysis produced an overall fit index: c-area for logistic regression and pseudo [R.sup.2] for survival analysis. (57) Although the pseudo-[R.sup.2] is only an approximate measure of effect size, it provides a way to compare results from different statistical models, such as linear regression and Cox regression. (77,78) A disease and a BA/PR summary index scores were computed by adding z-averaged individual scores on selected measures (see study 1). Missing values were handled by established methods. (75,76)

Results

Of the 606 patients with CABG 70 years of age or older, there were 41 deaths (6.8%) at 6-month follow-up and 159 deaths (26.2%) at long-term follow-up (3.2 to 7.5 years). The average survival (censored) was 4.06 years ([+ or -]1.72). Regressions predicting 6-month death showed similar contribution for high-tech disease, low-tech disease, and low-tech BA predictors tested separately (c-areas = 0.73, 0.72, 0.73). Combining high and low tech disease predictors did not add much to model fitness (c-area = 0.75). When BA variables were added, there was a significant improvement in predictive validity: c-areas increased from 0.75 to 0.87 [(high + lowtech) + BA predictors]. In predicting long-term survival after CABG, the three single-predictor models show similar results: pseudo-[R.sup.2] = 0.08, 0.12, 0.08, respectively, for high-tech, low-tech, and BA predictors. Adding low-tech to hightech disease predictors was especially useful: pseudo-[R.sup.2] increased to 0.17; it was 0.20 with the addition of BA variables.

[FIGURE 3 OMITTED]

Long-term survival analyses as a function of BA and disease (cardiovascular disease and comorbidity) in younger and older patients appear in Figure 3 (A and B). The two index scores were based on the disease (18/24) and the 18 of 21 BA indicators with part-whole correlation ([greater than or equal to] 0.10). Patients were classified into four categories, using median splits to define high and low scores. In both younger and older VA patients with CABG, being biologically fit with more severe disease or biologically frail with less disease predicted a similar and intermediate length of survival.

For patients older than 70 years of age (Fig. 3A), chronological age did not affect prediction [mean age was 73.6 in all four groups (F (3, 602) = 0.04, P = 0.99, NS)], but survival time was statistically different in the four disease/BA strata, as indicated by the life-table test for equality [[chi square] (3, n = 606) = 46.89, P < 0.0001]. In younger patients (Fig. 3B), survival was predicted by both BA/disease scores [[chi square] (3, n = 1,599) = 65.11, P < 0.001] and chronological age [F (3,1595) = 13.09, P < 0.0001], with mean age ranging between 57.8 and 61 years.

Discussion

Outcomes of treatment and procedures in the elderly can be gratifying but with higher rates of complications and death than in the young. This is often attributed to greater comorbidity and severity of the disease being treated. In patients who are very old, especially if they are also frail, competing risks lessen the prognostic implications of developing a new disease or condition. (29,79,80) Individual variations in the decreased physiologic reserve associated with the aging process per se may be a major contributor to outcomes variability. (81) Frailty and similar descriptive terms (26-32) have often been used by experienced clinicians to characterize the heterogeneity of aging. Since the older population is increasing and is also often healthier, finding ways to characterize fitness in elderly individuals without frailty or gross clinical deficits should be extremely useful in clinical decision-making.

Our first pilot study demonstrated that heterogeneity of functioning exists, even in a small highly selected sample of 128 healthy male and female volunteers (mean age = 78.8) living independently and without major limitations due to chronic disease. Timed tests (6-minute walk, one-foot balance, 30-second biceps curls, and the Trails B cognitive test) were especially useful in demonstrating a wide range of individual performance. By contrast, the MMSE, a widely used cognitive screening test, did not show variability in this highly educated sample with very few low MMSE scores.

The study also indicated that a useful BA/PR index could be developed by simply adding z-averaged scores on easily available and simple measures of endurance, balance, strength, flexibility, cognition, depression, exercise participation, and disease status (Charlson index of comorbidity). Scores on the nine-item BA/PR index were more useful than chronological age in predicting 3-year outcomes such as increased number of limitations in instrumental activities of daily living (IADL), moving to assisted living, giving up driving, and death, Being older (above the sample's mean age of 79 years) was an important risk factor for "biologically" frail individuals but not for those who were biologically "young or fit" (Table 2 and Fig. 2). The BA/PR did not predict basic ADL deterioration--not surprising in this high-functioning sample with a low rate of ADL limitations at baseline and minimal deterioration over time. This was also found in a longitudinal study of 282 older people living in the community as IADL but not ADL impairment occurred. (25)

Deficit-based indices including dementia, sensory loss, ADL limitations, medical conditions, and mobility have often been used to describe aging heterogeneity and define biological age (BA). In a large study (n = 2,914, mean age = 82.2), BA defined by 20 deficits was a strong predictor of a 52% mortality rate at 5 years. (29) The frailty index showed an exponential 3% increase with age, not only in people with dementia and/or cognitive deficit (68% of the sample) but also in the noncognitively impaired. (30) Frailty, defined as the presence of at least three of five deficits (exhaustion, unintentional weight loss, low grip strength, slow walk, and low physical activity), predicted falls, ADL deterioration, hospitalization, and death over a 3-year follow-up in the Cardiovascular Health Study. (26) Other reports confirm these results. (25-31) In the East Boston Senior Health Project, (19) perceived health rating (from poor to excellent) by 3,800 community dwellers was a strong predictor of 5-year death.

As effective as these predictors are in large samples of the elderly, such deficit-based and frailty indexes have limited usefulness with the elderly samples used in our studies. In the BA sample, 90% rated their health as "good" or "better than expected for age," and anemia, low albumin, poor cognition, and so forth, were rare. Timed functional tests of strength, walking speed, balance, and cognition were the key determinants of BA in these high-functioning elderly. Thus, even in this narrow subset of the elderly, important heterogeneity predictors were found.

In the PSOCS study, low-tech information (low-tech disease or low-tech BA) predicted CABG outcomes as well as high-tech disease information (requiring sophisticated equipment and/or highly trained personnel). The addition of BA information improved prediction in younger as well as older VA patients with CABG. Low disease severity and high BA/PR scores (biologically "fit or young") were associated with the longest survival time. Chronological age was a weak but statistically significant predictor of survival in younger patients (Fig. 3B) but not in the older sample (Fig. 3A). The predictive ability of the BA/PR information and the relatively modest age effect were unexpected with our younger (under 70) patients. Possible explanations are the restricted age range in younger veterans (18% only age 55 or younger) and the underrepresentation of "biologically fit" individuals in the PSOCS sample. VA patients tend to have more severe cardiovascular disease and comorbidity and are possibly biologically older than community patients, as indicated by comparisons between Medicare patients (82) and the BARI (Bypass Angioplasty Revascularization Investigation) sample. (83)

Purpose and limitations of this study

The primary objectives of our pilot studies were to demonstrate aging heterogeneity in a narrow sample of high-functioning elderly and to establish proof of principle regarding the usefulness of developing BA/PR estimates, while extending the data available from large study samples that include many frail and cognitively impaired elderly. It was not the intent and it is beyond the scope of our data to develop a specific instrument for general clinical application.

Sample selection bias was present in both studies. Larger study samples, including low SES and lower functioning elderly, would increase heterogeneity and probably strengthen the effect sizes beyond those obtained with our small high SES healthy community sample. This is supported by large studies of older adults reporting a significant association between 5-year mortality outcomes, difficulty in any IADL and relatively lower income, (85) or the proportion of 20 deficits. (29) Similarly, selection bias excluded the more severely diseased patients and reduced heterogeneity and effect size in the PSOCS study. We found 6-month mortality rates of 4.9% in our sample of elective CABG patients with completed SF-36 (n = 2,205) versus 10.6% for urgent and emergent procedures (n = 1,892), confirming the well-documented fact that surgical priority is a powerful determinant of outcome (42,43) and the need to assess CABG in the prereferral phase of medical decision.

In both pilot studies, the BA/PR predictors should only be considered illustrative of the kinds of data (eg, usefulness of "timed" tests in our healthy sample) and domains to be included. Several of the nine BA/PR items, weighted equally, address mobility (6-minute walk, balance tests, exercise participation). Restricted mobility has been found to be an important predictor of clinical functional disability, morbidity, and mortality in the elderly, (47-49,85,86) but other predictors may be equally or more predictive. (28,29) The PSOCS is a rich database but was not designed to assess BA/PR, and missing information was more frequent for BA variables than for disease measures. How to best select and weigh the components of the BA/PR index could be determined by large and well-designed prospective studies extending the information developed over the last several decades, primarily by geriatricians studying broad general populations of elderly persons. (84)

Implications for practice

The principles and approaches of geriatric assessment developed by geriatricians should be applied in a routine fashion by prospective serial evaluation of elderly patients. The informal, intuitive, and ad hoc BA evaluation used by most physicians should be extended to systematically collect BA information in all elderly individuals, using simple, easily obtained performance tests. This "tracking of the aging process" could serve as a prelude to identifying those patients who could best benefit not only from treatment but also from the sophisticated, often expensive, and sometimes risky procedures used to assess diseases such as cardiovascular disease.

Older patients use different value systems and individual priorities to weigh treatment benefits and negative outcomes that include death but also disability, discomfort, drug side effects, and financial risks. (87) Understanding and incorporating patient outcome priorities in the decision process should be helped by more realistic/specific outcome predictions, taking into account the biological age/physiologic reserve of each older person. In preparing patients for elective procedures, BA/PR results could also be used as an incentive for training since, even in the very old, multiple functions (eg, strength and endurance; and therefore BA/PR scores) can improve with training. (88,89)

Using the BA/PR assessment prior to referral for major treatments and procedures is likely to enhance patient and physician participation and limit the number of BA missing data. Time required to counsel patients is a major deterrent to a useful BA/PR approach. This emphasizes the need for simple and easy BA assessment tools, the usefulness of routine assessment "tracking" of the aging process, and the critical importance of a close interaction between patients, referring physicians, and the specialists involved in testing, procedures, and treatment.

References

1. American Heart Association. 2000 Heart and Stroke Statistical Update. Available at: http://www.amhrt.org/statistics/09medicl.htlm. Accessed June 2000.

2. Aziz S, Grover FL. Cardiovascular surgery in the elderly in Cardiovascular Disease. Cardiovasc Clin 1999;17:213-232.

3. Mullany CJ, Mock MB, Brooks MM, et al. Effect of age in the Bypass Angioplasty Revascularization Investigation (BARI randomized trial). Ann Thorac Surg 1999;67:316-403.

4. Grover FK, Shroyer LA, Hammermeister KE. Calculating risk and outcome; The Veterans Affairs Database. Ann Thorac Surg 1996;62:S6-S11.

5. Hochman JS, Sleeper LA, Webb JG, et al. Early revascularization in acute myocardial infarction complicated by cardiogenic shock. N Engl J Med 1999;341:625-634.

6. Berger AK, Schulman KA, Gersch BJ, et al. Primary coronary angioplasty vs thrombolysis for the management of acute myocardial infarction in elderly patients. JAMA 1999;282:341-348.

7. Thiemann DR, Coresh J, Schulman SP, et al. Lack of benefit for intravenous thrombolysis in patients with myocardial infarction who are older than 75 years. Circulation 2000;101:2239-2246.

8. Simoons ML, Maggioni AP, Knatterud G, et al. Individual risk assessment for intracranial hemorrhage during thrombolytic therapy. Lancet 1993;342:1523-1528.

9. Panotopoulos PT, Axtell K, Anderson AJ, et al. Efficacy of the implantable cardioverter defibrillatory in the elderly. J Am Coll Cardiol 1997;29:556-560.

10. Friesinger GC, Gravanis ME. Aging and the cardiovascular System. In: Gravanis ME, ed. Health Disease in Cardiovascular Disorders. St. Louis: Mosby; 1995.

11. Kuller LH, Sutton-Tyrell K. Aging and cardiovascular disease: use of subclinical measurements in cardiovascular disease in the elderly. Cardiovasc Clin 1999;17:51-66.

12. Hayflick L. How and Why We Age. New York, Ballantine Books, 1994, 11-17, 137-149.

13. Lamberts SW, van den Beld AW, van der Lely AJ. The endocrinology of aging. Science 1997;278:419-424.

14. Miller RA. When will the biology of aging become useful? Future landmarks in biomedical gerontology. J Am Geriatr Soc 1997;45:1258-1267.

15. Finch CE, Tanzi RE. Genetics of aging. Science 1997;278:407-411.

16. Perls T, Kunkel L, Puca A. The genetics of aging. Curr Opin Genet Dev 2002;3:362-369.

17. Puca AA, Daly MJ, Brewster SJ, et al. A genome-wide scan for linkage to human exceptional longevity identifies a locus on chromosome 4. Proc Natl Acad Sci U S A 2001;98:10505-10508.

18. Inouye SK, Peduzzi PN, Robison JT, et al. Importance of functional measures in predicting mortality among older hospitalized patients. JAMA 1998;279:1187-1193.

19. Walston J, McBurnie MA, Newman A, et al. Frailty and activation of the inflammation and coagulation systems with and without clinical comorbidities: results from the Cardiovascular Health Study. Arch Intern Med 2002;162:2333-2341.

20. Nakamura E, Miyao K. Further evaluation of the basic nature of the human biological aging process based on a factor analysis of age-related physiological variables. J Gerontol 2003;58A:196-204.

21. Bulpitt CJ. Assessing biological age; practicality? Gerontology 1995;41:315-321.

22. Farquharson SW, Gupta R, Heald RJ, et al. Surgical decisions in the elderly: the importance of biological age. J R Soc Med 2001;94:370-371.

23. Bright B. Body language: How old are you really? Biological age tests purports to give you the answer. Wall Street Journal R6. October 11, 2004.

24. Applegate WB, Blass JB, Williams TF. Instruments for the functional assessment of older patients. N Engl J Med 1990;322:1207-1214.

25. Reuben DB, Rubenstein LV, Hirsch SH, et al. Value of quality of life and functional status as a predictor of mortality: Results of a prospective study. Am J Med 1992;93:663-669.

26. Fried LP, Tangen CM, Walston J, et al. Cardiovascular Health Study Collaborative Research Group. Frailty in older adults: Evidence for a phenotype. J Gerontol A Biol Sci Med Sci 2001;56A:M146-M156.

27. Idler EL, Kasl S. Health perceptions and survival: Do global evaluations of health status help predict mortality? J Gerontol 1991;46:555-565.

28. MacDonald SWS, Dixon RA, Cohen A, Hazlit JE. Biological age and 12-year cognitive change in older adults: findings from the Victoria Longitudinal Study. Gerontology 2004;50:64.

29. Mitniski AB, Graham JE, Mogilner AJ, et al. Frailty, fitness and late-life mortality in relation to chronological and biological age. BMC Geriatr 2002;2:1.

30. Mitniski AB, Song XS, Rockwood RK. The estimation of relative fitness and frailty in community-dwelling older adults using self-report data. J Gerontol Med Sci 2004;39A:627-632.

31. Pressley JC, Patrick CH. Frailty bias in co-morbidity risk adjustments of community-dwelling elderly populations. J Clin Epidemiol 1999;52:753-760.

32. Fried LP, Ferrucci L, Darer J, et al. Untangling the concepts of disability, frailty, and comorbidity: Implications for improved targeting and care. J Gerontol 2004;59:255-263.

33. Rumsfeld JS, MaWhinney SM, McCarthy M Jr., et al. Health-related quality of life as a predictor of mortality following coronary artery bypass graft surgery. JAMA 1999;281:1298-1303.

34. Sarkisian CA, Lachs MS. 'Failure to thrive' in older adults. Ann Intern Med 1996;1072-1081.

35. Bierman AS. Functional status. J Gen Intern Med 2001;16:785.

36. Gill TM, Robison JT, Tinetti ME. Two components of the disability continuum among community-living older persons. Ann Intern Med 1998;128:96-101.

37. Reuben DB, Siu AL. An objective measure of physical function of elderly outpatients: The Physical Performance Test. J Am Geriatr Soc 1990;38:1105-1112.

38. Katz S, Ford AB, Moskowitz RW, et al. Studies of illness in the aged: The index of ADL: A standardized measure of biological and psychosocial function. JAMA 1993;185:914-919.

39. Branch LG, Katz S, Kniepmann K, et al. A prospective study of functional status among community elders. Am J Public Health 1984;74:266-268.

40. Fillenbaum GG, Smyer MA. The development, validity, and reliability of the OARS multidimensional functional assessment questionnaire. J Gerontol 1981;36:428-434.

41. Bogardus ST, Towle V, Williamns CS, et al. What does the medical record reveal about functional status? J Gen Intern Med 2001;16:728.

42. Process, Structures, and Outcomes of Care in Cardiac Surgery. Department of Veterans Affairs Cooperative Studies. Med Care 1995;10(suppl):OS1-OS107.

43. Jones RH, Hannan EL, Hammermeister KE, et al. Identification of preoperative variables needed for risk adjustment of short-term preoperative mortality after coronary artery bypass graft surgery. J Am Coll Cardiol 1996;28:1478-1487.

44. Clark RE. Calculating risk and outcome: The Society of Thoracic Surgeons database. Ann Thorac Surg 1996;62:S2-S5.

45. Guyall GH, Sullivan MJ, Thompson PJ, et al. The 6 minute walk: A new measure of exercise capacity in chronic heart failure. Can Med Assn J 1985;182:919-923.

46. Bittner V, Weiner DH, Yusuf S, et al. Prediction of mortality and morbidity with a 6-minute walk test in patients with left ventricular dysfunction. JAMA 1993;270:1702-1707.

47. Tinetti ME. Performance-oriented assessment of mobility problems in elderly patients. J Am Geriatr Soc 1986;34:119-126.

48. Gill TM, Williams CS, Tinetti ME. Assessment risk for the onset of functional dependence among older adults: The role of physical performance. J Am Geriatr Soc 1995;43:603-609.

49. Vellas BJ, Wayne SJ, Romero L, et al. One leg balance as an important predictor of injurious falls in older persons. J Am Geriatr Soc 1997;45:734-738.

50. Duncan PW, Weiner DK, Chandler J, et al. Functional reach: A new clinical measure of balance. J Gerontol 1990;45:M192-M197.

51. Gill D, Reddon J, Renney C, et al. Hand dynamometer: Effects of trial and sessions. Percept Mot Skills 1985;195-198.

52. Mathiowetz V, Kashman N, Volland G, et al. Grip and pinch strength: Normative data for adults. Arch Phys Med Rehabil 1985;66:69-74.

53. Folstein MF, Folstein PR, McHugh SE. Mini-mental state: A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189-198.

54. Tombaugh TN, McIntyre NJ. The Mini-Mental State examination: A comprehensive review. J Am Geriatr Soc 1992;40:922-935.

55. Teng EL, Chui HC. The modified Mini-Mental State (3MS) Examination. J Clin Psychiatr 1987;48:314-341.

56. Lezak MD. Neuropsychological Assessment. New York, Oxford University Press, 3rd ed, 1995, pp 545-546.

57. Reitan RM. Validity of the Trail Making Test as an indication of organic brain damage. Percept Mot Skills 1958;8:271-276.

58. Yesavage JA, Brink TL, Rose TL, et al. Development and validation of a geriatric depression screening scale: A preliminary report. J Psychiatr Res 1982;17:37-49.

59. Katz S, Ford AB, Moskowitz RW, et al. Studies of illness in the ages: The index of ADL: A standardized measure of biological and psychosocial function. JAMA 1963;185:914-919.

60. Lawton MP, Brody EM. Assessment of older people: self-maintaining and instrumental activities of daily living. Gerontologist 1969;9:179-186.

61. Guigoz Y, Vellas B, Garry PJ. Mini-Nutritional Assessment: A practical assessment tool for grading the nutritional assessment of elderly patients. Nutr Rev 1992;2 (suppl):15-59.

62. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic co-morbidity in longitudinal studies: development and validation. J Chron Dis 1987;5:373-383.

63. Katz JN, Chang LC, Sangha O, et al: Can co-morbidity be measured by questionnaire rather than medical record review? Med Care 1996;1:73-84.

64. Lord FM, Novick MR. Statistical theories of mental test scores. Oxford, England, Addison-Wesley, 1968.

65. Nunnally JC, Bernstein IH. Psychometric theory. New York, McGraw-Hill, 1994.

66. Diggle P, Liang KY, Zeger SL. Analysis of Longitudinal Data. Oxford/New York, Clarendon Press/Oxford University Press, 1994, pp 60.

67. Littell RC, Milliken GA, Stroup, WW, et al. SAS System for Mixed Models. Cary, SAS Institute, 1996.

68. Raudenbush SA Bryk AS. Hierarchical Linear Models: Applications and Data Analysis Methods. Newbury Park, Sage Publications, 2nd ed, 2002.

69. Gibbons RD, Hedeker D, Elkin I, et al. Some conceptual and statistical issues in analysis of longitudinal psychiatric data. Application to the NIMH Treatment of Depression Collaborative Research Program data set. Arch Gen Psychiatry 1993;50:739.

70. Brown W. Some experimental results in the correlation of mental abilities. Br J Psychol 1910;3:3296-3322.

71. Spearman C. Correlation calculated from faulty data. British J Psychol 1910;3:3171-3195.

72. Clark LA, Watson D. Constructing Validity: Basic Issues in Objective Scale Development. Psychol Assess 1995;73:309-319.

73. Grover FL, Johnson RR, Shroyer AL, et al. The Veterans Affairs Continuous Improvement in Cardiac Surgery Study. Ann Thorac Surg 1994;58:1845-1851.

74. Hammermeister KE. Participatory continuous improvement. Ann Thorac Surg 1994;58:1815-1821.

75. Little RJA, Rubin DB. Statistical Analysis with Missing Data. New York, Wiley, John & Sons, 1987.

76. Hill MA. SPSS Missing Value Analysis 7.5. Chicago, SPSS Incorporated, 1997.

77. Cox, DR, Oaks D. Analysis of Survival Data. London, Chapman and Hall, 1984.

78. Allison PD. Survival Analysis Using the SAS System. Cary, SAS Institute, 1995.

79. Welch W, Albertsen P, Nease RF, et al. Estimating treatment benefits for the elderly: The effect of competing risks. Ann Intern Med 1996;124:577-584.

80. Katz S, Branch LG, Granson MD, et al. Active life expectancy. N Engl J Med 1983;309:1218-1224.

81. Kohn, RR. Cause of death in very old people. JAMA 1982;247:2793-2797.

82. Petersen LA, Norman SL, Daley J, et al. Outcome of myocardial infarction in Veterans Health Administration patients as compared with Medicare patients. N Engl J Med 2000;343:1934-1934.

83. Brooks MM, Detre KM. The design, patient population and outcomes from the Bypass Angioplasty Revascularization Investigation (BARI) randomized trial and registries: Semin Interv Cardiol 1999;4:191-199. Lombardo C and Detre KM. Personal Communication. December 2000.

84. Gill TM. Geriatric medicine: It's more than caring for old people. Am J Med 2002;113:85-90.

85. Fried LP, Kronwal RA, Newman AB, et al. Risk factors for 5-year mortality in older adults: The Cardiovascular Health Study. JAMA 1998;279585-592.

86. Fried LP, Bandeen-Roche KB, Chaves PH, et al. Preclinical mobility predicts incident mobility in older women. J Gerontol A Biol Sci Med Sci 2000;55A:M43-M52.

87. Fries, JF. The hierarchy of quality of life assessment, the Health Assessment Questionnaire (HAQ), and issues mandating development of a toxicity index. Control Clin Trial 1991;12:1068-1175.

88. Ades PA, Ballor DL, Ashikaga T. Weight training improves walking endurance in healthy elderly persons. Ann Intern Med 1996;124:568-572.

89. Vincent KR, Braith RW, Feldman MS, et al. Improved cardiorespiratory endurance following 6 months of resistance exercise in elderly men and women. Arch Intern Med 2002;162:673-678.

I have found that if you love life, life will love you back.
--Arthur Rubinstein

Jacqueline Goffaux, PHD, Gottlieb C. Friesinger, MD, Warren Lambert, PHD, Laurie W. Shroyer, PHD, Thomas E. Moritz, MS, Martin McCarthy, Jr, PHD, William G. Henderson, PHD, and Karl E. Hammermeister, MD

From Vanderbilt University, Department of Medicine, Division of Cardiovascular Medicine, Nashville, Tennessee; Vanderbilt University, Institute for Public Policy Studies, Nashville, Tennessee; the Cardiology Section (111B), Denver VA Medical Center, Denver, Colorado; The Edward Hines Jr VA Hospital, Cooperative Studies Program Coordinating Center, Chicago, Illinois; Northwestern University Medical School, Chicago, Illinois; and the University of Colorado, Colorado Health Outcomes Program F-443, Aurora, Colorado.

Reprint requests to Gottlieb C. Friesinger, MD, Vanderbilt University Medical Center, 383 PRB, 2220 Pierce Avenue, Nashville, TN 37232-6300. E-mail: [e-mail].

Accepted June 23, 2005.

This study was supported by philanthropic funds from the Cardiovascular Medicine Division, Vanderbilt University and Health Services Research and Development Service of the Department of Veterans Affairs, Washington, DC. In addition, the support of the Vanderbilt University Kim Dayani Human Performance Center is acknowledged.

RELATED ARTICLE: Key points

* Major heterogeneity in biological age/physiological reserve (BA/PR) was present in high functioning "healthy" individuals, age 70 years or older.

* Simple measures, and especially timed tests (eg, Trails B, 6-minute walk, standing on one foot), can be used to characterize this heterogeneity.

* BA/PR measures were more predictive of adverse outcomes than chronological age in pilot studies involving older samples of (a) 128 independent-living "healthy" individuals and (b) 606 veterans undergoing coronary artery bypass graft.

* Further research is needed to determine the most predictive BA/PR measures and to extend the concept to broader groups.

Table 1. BA/PR Index measures

                                       Baseline
                            Pearson r  sample            N = 128
Measures                    with age   Mean (SD)         range

Six-minute walk (feet) (a)  -0.56      1531.29 (388.70)  480-2,237
Weight lifts in 30 s (b)    -0.41        19.96 (5.71)     10-35
Tinetti balance (1/16)      -0.19        15.43 (1.50)      5-16
One-foot stand (s) (c)      -0.46        24.64 (32.05)     0-181
Standing reach (in)         -0.44        13.58 (3.27)      2-21
Physical activity (d)       -0.30         5.54 (3.21)      0-12
Trails B (s) (e)             0.33       134.95 (76.32)    46-561
Depression (GDS30) (f)       0.25         4.78 (4.20)      0-19
Charlson comorbidity (g)     0.21         1.41 (1.16)      0-5

Table presents raw scores for each of the nine tests used to compute the
index. BA/PR index = sum of the nine test scores in z-scores (mean = 0;
SD = 1).
(a) z-scores based on sex-specific means/SD; (b) hand weight: 4 lb,
females; 8 lb, males; (c) best three trials average time right/left foot
stand; (d) total number of times per week (vigorous + long walk + short
walk + physical work); (efg) low scores = good for: Trails B, Geriatric
Depression Scale (0-30), Charlson number comorbid conditions.

Table 2. Three-year outcomes as a function of BA/PR and chronological
age

                              Frail/old     Fit/old       Frail/young
                              Mean (SD)     Mean (SD)     Mean (SD)

No.                           28            20            18
Deaths at 3 yr (a)             9             0             0
Moved to assisted living (b)   4             0             0
Not driving (b)               16             2             7

Age                           85.79 (3.08)  81.15 (1.95)  74.94 (2.81)
BA Index                      -4.41 (3.5)   +3.17 (1.73)  -3.09 (4.09)

IADL baseline                  0.68 (1.2)    0.05 (0.22)   0.5 (0.9)
IADL 1 yr                      1.4 (1.6)     0.3 (0.07)    1.3 (1.3)
IADL 3 yr                      2.3 (2)       0.75 (0.85)   1.6 (1.3)

                              Fit/young
                              Mean (SD)     N = 106 range

No.                           40             20-40
Deaths at 3 yr (a)             2              0-9
Moved to assisted living (b)   0              0-4
Not driving (b)                3              2-16

Age                           72.87 (2.64)   69-91
BA Index                      +4.54 (2.05)  -14.78-10.51

IADL baseline                  0.02 (0.27)    0-4
IADL 1 yr                      0.13 (0.34)    0-5
IADL 3 yr                      0.44 (1.23)    0-7

(a) P < 0.01, (b) P < 0.001, [chi square] test; median split
classification into high/low on index of physiologic reserve and
chronologic age categories. This table includes complete cases with all
three repeated measurements (n = 106).