| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
|
|
|
|||||||
|
|||||||||
National Research Council, Institute of Neuroscience, Padova Aging Section, Italy.
Address correspondence to Nadia Minicuci, CNR-Institute of Neuroscience, Section of Padova–Aging, c/o Clinica Medica 1, Via Giustiniani, 2, 35128 Padova, Italy. E-mail: nadia.minicuci{at}unipd.it
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
An operational definition of vitality in the aged is difficult, particularly if we do not consider merely the absence of disease and disability, but rather the physical and mental functioning of a cohort of older individuals. In the past, several studies have attempted to identify predictors of healthy aging (or successful aging), but both the predictors included in the analyses and definitions of outcomes differed from one study to the other, making it very difficult to draw any conclusions (4).
Because primary prevention is a viable way of dealing with the increasing cost of medical care, there is a strong need to understand which modifiable risk factors predict healthy aging and to use this information as the scientific basis for systematic interventions designed to enhance health in elderly persons. It is worth underlining the fact that the impact of personal habits and social factors on the functional status of older individuals has often been underestimated. This study analyzes the concept of vitality and identifies the associated factors in a large cohort of older Italian men and women, monitored for 4 years.
|
METHODS |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
A first phase, administered to all participants, that included:
A second phase, administered to participants who screened positive to the first phase, consisted of clinical confirmation by a specialist (internist or neurologist) of suspected cases of cardiovascular diseases, diabetes, parkinsonism, stroke, dementia, or peripheral neuropathy through a visit and the review of medical records. Basic demographic data for nonrespondents were collected from a proxy.
Baseline data were collected in 1992 and 1993 and follow-up took place during 1996 and 1997 (length of follow-up: 3.5 ± 0.4 years). In the 1996 follow-up, data on mortality were collected by means of a copy of the death certificate (reporting date and cause of death) issued by the national registry.
Assessment of mental status.-- In the ILSA questionnaire, the MMSE modalities were: a) correct, b) wrong, c) does not answer, d) not applicable (participant was not able to read and/or write, participant had an arm in a cast or sling, or participant had a severe vision and/or hearing impairment). For analysis purposes, the "does not answer" modality was re-coded as wrong answer, and "not applicable" as missing. If more than six items were missing, the total score was set to missing.
Mental status score on the MMSE was calculated as a ratio of the number of questions answered correctly to the number of questions asked (if all questions were applicable), or as an adjusted score (number of questions answered correctly divided by the number questions answered, if some of the questions were not applicable). The total score thus became a number ranging between 0 and 1; the first quartile proved to be equal to 0.86 both at baseline and at the first follow up, whereas the third quartile was equal to 0.97 for both measurements.
Assessment of depressive symptomatology.-- In the ILSA, we found a low percentage (5.4%) of people unable to answer the Geriatric Depression Scale; this percentage decreases to 4.0 when considering participants with information on all other variables associated with vitality in the multivariate analysis phase.
The total score was calculated by dividing the number of positive answers by the number of answered questions. If six or more items were missing, the total score was set to missing. The total score thus became a number ranging between 0 and 1; the cut-off to determine depressive symptomatology (DS) was set at 0.33.
Assessment of physical functioning.-- As developed by Katz and colleagues (10), the ADLs—eating, continence, transferring in and out of bed, toileting, dressing, and bathing—have made it possible to analyze detailed observations of many basic activities in persons with chronic conditions. Each ADL item has a score between 1 and 3; a score of 1 indicates that the person is completely independent, whereas a score of 3 indicates total dependence. We used an adjusted total score varying from 0.33 to 1, obtained by dividing the total score of the answered items by the sum of scores of the same items in the hypothesis of complete dependence. A score of 0.33 means that the person is completely independent; a score between 0.33 and 0.56 indicates dependence in two ADLs at most; a score between 0.56 and 0.78 indicates dependence in three ADLs at most; and between 0.78 and 1, dependence in four or more ADLs.
Assessment of physical performance.-- The timed chair stand item was attributed a score of between 4 (best performance) and 1 (worst performance), whereas the other five items (timed step-up, tandem walk, timed one-leg stand, timed 5-meter walk, and 180° pivot) were attributed scores of between 2 (best performance) and 1 (worse performance). For all items, a score of 0 indicated that the participant was too physically impaired to perform the action. The total score of the PPT was calculated by summing the score obtained on each item and dividing by the corresponding total sum. The total score is a number between 0 and 1, where independence in PPT had a score equal to 1; mild dependence a score of between 0.66 and 1; moderate dependence a score between 0.33 and 0.66, and severe dependence scored less than 0.33.
Diagnosis of health conditions.-- The diagnostic criteria used to define the prevalence ratios of the investigated conditions included a screening phase and clinical confirmation of the disease. The criteria for the diagnosis and the prevalence ratios have already been published (13).
Definition of vitality.--
We have defined as the "high vitality group" those participants falling in the 75th percentile or higher on the MMSE (total score 
0.97) and who were independent in all ADLs (total score = 0.33). The "low vitality group" was defined as those participants falling in the 25th percentile or lower on the MMSE (total score 
0.86), and with moderate or severe disability in ADLs (total score 0.56). All of the other participants were considered to be in the "medium vitality group." Transitions were investigated both in terms of joint cognitive and physical vitality and in terms of cognitive and physical vitality separately by domain.
Statistical Analyses
The design effect was taken into account by weighting each participant according to the age distribution of the Italian reference population (1991 Census) and the sample fraction; the main reason for computing the weights was to generalize the ILSA sample to the reference Italian population from which our sample was drawn.
Associations between demographic variables (sex, marital status, educational level, and occupation), health habits (smoking and drinking habits), health status variables (diseases, comorbidity, DS, and deficiencies in the PPTs), and vitality status were investigated with the chi-square test for trend. The association with body mass index (BMI) was tested by the chi-square test.
The comparison of group mean age with vitality status was evaluated through the Generalized Linear Model (GLM) procedure. Levene's test was used to test the homoschedasticity of variances; where the assumption of homoschedasticity was not met, Welch's test was performed.
Analysis of transitions in vitality levels.-- A transition matrix of vitality levels from baseline to follow-up was constructed, and overall marginal homogeneity was evaluated by the Stuart–Maxwell test. Four categories of participants were constructed by considering vitality level at baseline and follow-up:
A multinomial logistic model was then developed with six levels of the dependent variable: positive stable (referent category), nonpositive stable, positive transition, negative transition, death, and lost to follow-up.
Analysis of transitions in cognitive and physical functioning.-- Transitions in cognitive and physical functions were also investigated separately; for both MMSE and ADLs, improvement or decline was defined as an increase or decrease, respectively, in more than one standard deviation (SD) of the difference variable (follow-up total score minus baseline total score), while all other participants were characterized as stable. Four groups of participants were defined: those who improved or remained constant in both physical functions and MMSE score; those whose physical status improved but MMSE score declined; those who improved their MMSE score but presented a decline in physical function; and those with a decline in both physical function and MMSE score. A multinomial logistic model was developed to study the predictors of transition from baseline to follow-up, in which the outcome had the same four levels as the above-mentioned four groups, plus two levels accounting for the deceased and those lost to follow-up.
Predictors of transitions.-- The variables considered in both multinomial logistic models were all sociodemographic, health habits, and conditions; analysis of confounders was implemented by evaluating variation in the estimated coefficient of the explanatory variable when the "potential confounder" was absent or present in the model, independent of the statistical significance of the confounders. All analyses were performed using the SAS statistical package (release 8.02; SAS Institute, Cary, NC).
Sample attrition.-- Of the 5632 participants originally sampled, 1010 (17.8%) did not participate in the baseline assessment and were considered nonrespondents. Among the 4620 for whom sociodemographic baseline information was available, 891 (19.3%) were lost to follow-up and 719 (15.5%) were found to be deceased at follow-up. We found a significantly higher proportion of women (65.8%), a higher mean age (72.4 years), and a lower proportion of married participants (52%) among respondents compared to those lost to follow-up (56.5%, 71.7 years, and 61.8%, respectively). Moreover, from the health status perspective, respondents had higher prevalence rates of arrhythmia (24% vs 17.6%), stroke (5.8% vs 4.0%), distal symmetric neuropathy (6.3% vs 4.1%), myocardial infarction (7.6% vs 3.8%), and hypertension (62.8% vs 53.3%). In contrast, those participants lost to follow-up seemed to present higher DS (41.7% vs 35.5%) and impairment in PPTs (54.2% vs 44.4%). All other investigated characteristics such as parkinsonism, angina, heart failure, diabetes, MMSE score, and ADLs were not statistically different between the two groups (data not shown).
A higher mean age (75.7 vs 71.7 years) and a lower proportion of women (50.6% vs 56.5%) and married people (53.8% vs 61.8%) were found among those who were deceased at follow-up than among respondents; moreover, the prevalence of conditions such arrhythmia (28% vs 24%), heart failure (15.5% vs 5.1%), myocardial infarction (11.2% vs 7.6%), diabetes (16.9% vs 12.2%), stroke (14% vs 5.8%), and distal symmetric neuropathy (10.9% vs 6.3%), were also significantly higher in those who were deceased, along with the presence of DS (53.9% vs 35.5%) and disability in both PPTs (69.8% vs 44.4%) and ADLs (55.7% vs 28.9%); mean MMSE score was instead higher among respondents (0.91 vs 0.86) (data not shown).
|
RESULTS |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Table 1 shows the distribution of sociodemographic, health, and physical characteristics by level of vitality. A statistically significant trend with vitality was found for all sociodemographic (age, sex, and marital status) and socioeconomic variables (educational level and occupation), and for one of the three health habits (smoking). BMI was also associated with vitality. Neurodegenerative diseases were associated with vitality, whereas cardiovascular diseases (with the exception of heart failure) and diabetes were not. Although the prevalence of angina showed evidence of a trend, statistical testing by the chi-square (p =.09) and chi-square for trend (p =.06) did not confirm the hypothesis. Comorbidity, presence of DS, and impairment on PPTs also showed a significant trend with vitality.
|
|
The marginal distributions show that, at baseline, 19.2% of the participants were in the high vitality group (vs 14.6% at follow-up), and 77.9% were in the medium vitality group (vs 69.7% at follow-up). The homogeneity hypothesis was statistically tested (Stuart-Maxwell test; p <.0001) for all categories simultaneously and showed that the quality of life at follow-up differed significantly from the quality of life at baseline in the same participant.
Predictors of Transitions in Vitality Levels
From the transition matrix, we constructed four categories of participants as follows:
Multinomial logistic regression analysis was performed to identify correlates associated with different trajectories of vitality, where the outcome variable had six levels: the four above-mentioned categories, the deceased, and those participants lost to follow-up. The presence of confounders was analyzed during the selection of predictors by examining the change in the estimated coefficient of the predictor when the potential confounder was added to the model.
Table 3 presents the odds ratio (OR) for the positive stable group (referent category); smoking status, arrhythmia, and comorbidity were included in the model as confounders. Age and educational level were predictors of being positive stable (independently of the choice of control group), whereas all other factors behaved differently according to the comparison group. For example, sex was a predictor of positive stable only when compared to "nonpositive stable" and "deceased." Participants with DS were always more likely to belong to the comparison groups, except for the positive transition group.
|
Table 4 presents the ORs from the multinomial logistic model for specific domain of declining versus those ORs that exhibited improvement or stability in both types of functioning. Because degree of change is a function of the initial level, the baseline values of the MMSE and ADL scores have been included in the multinomial model. This new approach, which complements the previous one, led to the identification of correlates of the different trajectories of vitality. For example, from Table 3 we can see that participants with a lower educational level were less likely to maintain a positive stable level of vitality; from Table 4 it can be inferred that transitions from the positive stable group to any other group might be due to a decline in MMSE score but not in ADL functions (OR = 3.17) or in both functioning (OR = 2.33). Age also proved to be a predictor of "improvement or stability in both ADL and the MMSE score," because older people are more likely either to lose some functioning, or to have deceased or be lost to follow-up. The presence of DS brings about an almost two-fold increase in the decline in both ADLs functions and MMSE score, and a 70% higher mortality risk.
|
|
DISCUSSION |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
When evaluating the predictors of vitality, we found some interesting results. Among the factors that predicted permanence in the negative stable group, lower educational level showed the highest impact, but older age, male sex, blue collar or housewife occupations, obesity, stroke, and DS, as well as impairment in the PPTs, were also significant risk factors. How educational level and occupation affect functional status and mortality is a moot question in both social and clinical medicine (17). Probably, they are among the strongest predictors of negative outcomes in most studies, despite the fact that the proposed pathways by which they might affect health are not completely clarified. They include psychological consequences of low income, which affect not only the social environment but also the individuals directly, with higher levels of stress. Worse lifestyle factors and less access to preventive and health services have also been described among the lower socioeconomic classes, and might be responsible for the lack of improvement in the health status of these individuals.
That obesity remains a risk factor for morbidity and disability in elderly persons is clinically well acknowledged, and its ability to predict nonpositive stability, negative transition, and loss to follow-up in the ILSA could be related to mechanical complications such as osteoarthritis and static respiratory problems, or cardiovascular and metabolic complications (18–20). Not only the increase in fat load, but also the age-associated reduction in lean body mass may play a major role in causing mobility impairment. However, because we also found an association between obesity and positive transition, we must underline two other factors. First, the importance of weight history in establishing the risks associated with obesity in old age should be kept in mind, because obese older people are not a homogeneous group. Harris and coworkers (21) evaluated the risk of late-life coronary heart disease in a group of 621 men and 960 women, with a mean age of 77 years, who were free of coronary heart disease at entry. These authors found that a BMI > 27 in middle age increased the risk for coronary heart disease later in life, but that those who had BMI > 27 for the first time in old age had no such clear pattern of increased risk. Thus older people who were overweight were found to run an increased risk only when weight history had been properly accounted for. Therefore, not only actual weight, but also weight development over preceding decades may predict outcome. Second, indices of visceral obesity may be better indicators of risk than BMI in these age groups. It is indeed known that BMI is not a reliable measure of obesity in elderly persons, due to changes in height with advancing age and to the different ratio of lean and fat mass compared to those in younger adults (22). Of extreme interest is the protective effect of overweight with decline in physical and cognitive functioning and the significant association of underweight with mortality. Older people often have poor dietary intake, which makes them particularly vulnerable. Although in most cases the effects of nutrition are small, when older people begin to develop subclinical or overt diseases, malnutrition may play a dramatic role in accelerating the physical decline leading to death (23).
Moreover, we must consider the strong association found between being dead at follow-up and older age, male sex, manual occupation, heart failure, diabetes, DS, and impairment in the PPTs. These findings stress the need to bear in mind that we are looking at transition in vitality status among a select group of survivors and, therefore, the lack of association with expected risk factors can be explained by the survival bias. Compared to males, females had a higher probability of not retaining a negative stability status during follow-up and of remaining alive. It is in fact well documented in previous studies that females have higher rates of disability, but of milder severity, than do males, and their mortality rates are also lower at advanced age.
Some very interesting findings from our analysis are related to DS. Individuals in the high vitality group at baseline have a significantly lower prevalence rate of DS compared to individuals in the low or medium vitality group. DS can be viewed as a common predictor of negative stability, negative transition, loss to follow-up, and mortality; moreover, it heightens the decline in both ADLs and MMSE scores simultaneously. These data suggest that not being depressed not only stimulates cognitive capacities, but also physical well-being. These findings are consistent with the growing literature on negative outcomes related to DS. Clinical data on depression show an increase in the risk of mortality (24), disability (25), and impaired psychosocial functioning (26). The identification of depressive symptoms as a risk factor for both physical and mental functional decline is very important. Indeed, withdrawing from physical or social activities because of DS is common in elderly persons. Moreover, the negative impact of depression on morbidity, particularly of the cardiovascular system, is becoming well known in the geriatric literature (27). Given that depressive symptoms are not a necessary component of aging, intervention to reduce their frequency should be considered as an important endpoint of geriatric care, and geriatricians need to be careful in evaluating their patients for DS. The findings on depression are particularly innovative because they refer to a population with a high prevalence of DS (28). It is a well-known fact that relative risks associated with a common risk factor are hard to estimate because they depend on the prevalence of other risk factors contributing to the same sufficient causal pathway. The fact that depression is such a common and strong factor seems to suggest that depression is closely related to vitality and there may be some conceptual overlapping of these two factors in aging.
Several limitations of our study deserve comment. First, we used a psychometric scale to assess DS, but no clinical evaluation was made of depression. However, Hautzinger and colleagues (29) have shown that there is a high correlation between the score on the scales and clinical diagnosis. Second, although we measured many traditional risk factors for physical and mental status, the association between baseline variables and pattern of vitality level may have been affected by unmeasured variables such as physical activity or chronic conditions not included in our assessment. Their inclusion would probably have increased our understanding of the pathways involved.
Our study, however, has some peculiar strengths. To our knowledge, it is the first study analyzing the association of sociodemographic and biological variables with vitality status in a large sample of older Italians, with a prospective design and a relatively long follow-up period. Moreover, we performed a clinical evaluation for the diagnosis of major chronic conditions and did not use self-reported information that might have led to a misclassification of the conditions. Finally, Butler and colleagues (30) have assessed transition in physical functioning, but we know that cognitive vitality is also an essential component of quality of life at older ages and must become a primary public health issue.
Conclusion
Analysis of the characteristics associated with vitality status provides some insight into life-course preventive measures that should be implemented so that persons reach old age with a good vitality status. In particular, we have identified some socioeconomic factors, such as low educational level and occupation status, which are responsible for keeping individuals at a medium or low vitality level, and we conclude that their functioning could improve through lifelong learning, with active mental exercise and engagement in social activities. Moreover, we have stressed the fact that proper nutrition may be an important factor in promoting physical, as well as cognitive health. Finally, DS, a frequent and modifiable condition of elderly persons, is a contributory determinant of vitality status in our population. These results deserve careful consideration from the clinical and public health perspective and should become the basis for programs to increase awareness about the real potential for achieving and maintaining vitality in old age.
|
Footnotes |
|---|
Received July 16, 2004
Accepted December 3, 2004
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| All GSA journals | The Gerontologist |
| Journals of Gerontology Series B: Psychological Sciences and Social Sciences | |
