Successful aging, a new challenge for developing countries: The Coyoacán Cohort

Abstract

Objective

To determine the socio-demographic and health factors associated with a biomedical phenotype of successful aging (SA) among Mexican community-dwelling elderly.

Design, setting and participants

Cross-sectional study of 935 older adults aged 70 or older participating in the Mexican Study of Nutritional and Psychosocial Markers of Frailty.

Measurements

SA was operationalized in accordance with the phenotype proposed by Rowe and Kahn. Univariate and multivariate logistic regression analyses were carried out in order to identify the correlates of SA.

Results

The phenotype of SA was present in 10% of participants. Age (P < 0.001), illiteracy (P = 0.021), polypharmacy (P < 0.001), and physical pain (P < 0.001) were factors independently and inversely associated with the presence of the SA phenotype. The only variable positively associated with SA was good self-perceived health-status (P < 0.001).

Conclusion

Although age is not modifiable, several other factors associated with SA are. If we are to promote SA, efforts should be made towards improving those modifiable factors negatively associated with its presence, such as pain or polypharmacy. Also, enhancing factors positively associated to it might play a role in improving wellbeing.

Introduction

Developing countries have decisively modified their socio-demographic profile in the last few decades; the rapid population growth has slowed down, an urban profile was consolidated, and from being essentially young countries, a transition into an aging population began. This growth poses a number of challenges that give rise to the interest of determining factors that can help in increasing the number of years elderly subjects live in healthy conditions (1., 2., 3.).

Different biomedical and biopsychosocial aging phenotypes have been described in order to identify and characterize those elderly who stand out from the others because of their active participation in society due to their actual and self-perceived good physical and cognitive performance, namely successful and healthy aging phenotypes (4). However, what constitutes successful aging (SA) is still controversial, partly because of the lack of consensus for an operational definition (5).

Rowe and Kahn proposed one of the most accepted biomedical phenotypes of SA in the late 1990s. According to this phenotype, SA is present when the following three components are simultaneously identified: 1) low probability of disease or disability; 2) high physical and cognitive abilities; and 3) preserved social activity. However, as mentioned above, working definitions on SA are still controversial and may include socio-demographic as well as self-perceived constructs that may differ between developed and developing countries (5., 6.). Therefore, the social and health profile, as well as the correlates and outcomes of SA in community-dwelling elderly living in developing countries could be different from the ones described in developed ones (7, 8). Estimating the frequency at which the elderly of a society reach old age in optimal conditions of autonomy and psychosocial wellbeing as well as identifying the socio-demographic and health determinants associated with this phenotype might provide the basis to promote the implementation of tailored health agendas that promote healthy ageing in various domains (9). The main objective of the present study was to determine the socio-demographic and health factors associated with the phenotype of SA among Mexican community-dwelling elderly.

Methods

Study population

This is a croßs-sectional study of a subset of participants of the Mexican Study of Nutritional and Psychosocial Markers of Frailty (The Coyoacán Cohort), a prospective study aiming to evaluate the nutritional and psychosocial determinants of frailty among Mexican community-dwelling elderly. A detailed description of the methodology of the Coyoacán cohort has been reported elsewhere (10). Briefly, participants were identified through the database of the “Food aid, medical care and free drugs program” by the Mexican Government, which includes 95% of community-dwelling adults aged ≥ 70 years in Mexico City. The sampling procedure was randomized and stratified by age and sex and performed in Coyoacán, one of the 16 districts of Mexico City. The calculated sample size to estimate a prevalence of frailty of at least 14% was 1294 with an α= 5% and ß= 20%. Among the contacted subjects, the acceptance rate was 86.9% with a total of 1,124 participants being finally included in the study. Baseline data were collected between April 2008 and May 2009 through a questionnaire and a clinical evaluation. Data collection was conducted in two phases. In the first one, data were collected through a face-to-face interview using a standardized questionnaire administered by pollsters who were previously trained for standardized assessment. Socio-demographic factors (age, sex, social networks, education and employment status) as well as health issues (cognition, functional status, co-morbidity, smoking, alcoholism, and depressive symptoms) were investigated during this phase. In the second one, a multidisciplinary team (physician, nutritionist, and dentist) evaluated participants, and a Comprehensive Geriatric Assessment (CGA) was carried out including the evaluation of functional status, co-morbidity, pharmacological treatments, physical performance, nutritional state, oral health, blood pressure, and several measures of anthropometry. Each participant signed an informed consent, and the local Ethics Committee approved the study protocol.

Definition of successful aging

Successful aging was identified through the biomedical model proposed by Rowe and Kahn (6) among those participants meeting the following five criteria:

• •Absence of disease was identified if participants denied a previous diagnosis of any of the following four pathologies: coronary heart disease, diabetes, stroke, or cancer.
• •Absence of disability was considered if participants indicated that they did not need any assistance in order to perform all activities of daily living (ADL) from those evaluated by the Katz Index (bathing, dressing, toileting, transferring, and feeding) (11).
• •Preserved Cognitive Performance was identified through the execution of the mini-mental state examination (MMSE) (12). Due to the abnormal distribution of the MMSE score, a total score above the 20th percentile was considered as cognitively normal.
• •Appropriate Physical Performance was taken into account if participants reported being able to carry out the following five activities from the Nagi scale (13): 1) stooping, crouching or kneeling; 2) reaching or extending arms above shoulder level; 3) writing or handling small objects; 4) pushing or pulling large objects; and 5) lifting and carrying objects heavier than 5 kg (11 lbs.).
• •Active Social Participation was considered present if participants: 1) reported having worked for a salary or voluntarily during the previous week, or 2) reported having a conversation with neighbours, relatives or friends at least four times a month.

Correlates

Due to their frequent use in the clinical setting, the following variables were investigated as potential correlates:

Socio-demographic variables included age (years), sex, marital status, and illiteracy (unable to read or write a note).

Comorbidity: Participants were asked whether they had or not a physician's diagnosis of hypertension, arthritis, dysthyroidism, history of fractures, or hypercholesterolemia.

Urinary incontinence (UI) was defined by a positive response to the question: Do you have difficulty to stop or control urine?

Weight loss was identified when participants reported unintentional weight loss of 5 Kg or more in the previous year.

Participants were also asked if they experienced physical pain on a daily basis. The answer was dichotomized as “yes” versus “no”.

Falls were considered if participants reported having at least one fall during the previous year.

Depressive symptoms were assessed using the Center for Epidemiologic Studies-Depression scale (CES-D), a score of 16 or greater translated depressive symptoms (14).

Polypharmacy was considered when participants reported taking five or more different medications (15).

Spirituality was assessed asking about the importance of religion in their life. The answer was dichotomized as important versus not important.

Self-reported health status and financial situation were rated as excellent, very good, good, fair or poor. Excellent, very good, and good responses were operationalized as “good”.

Sample

From an initial study population of 1,124 participants, 189 were excluded because the phenotype of SA could not be constructed and information from 935 participants was finally included in the statistical analyses. Those participants excluded were no different from the included ones regarding age (P = 0.247), ADL disability (P = 0.784), cognitive status (P = 0.385), physical functioning (P =0.755), or social interaction (P = 0.441).

Statistical analysis

The variables are described using frequencies and proportions or means and standard deviations where appropriate. For the comparison between successful and nonsuccessful agers, χ2 test, Student t test or nonparametric tests were used according to the variable type and its distribution. In order to develop an explanatory model for the SA phenotype, unadjusted logistic regression models were created to identify the socio-demographic and health variables associated with the dependent variable. Finally, a forward multiple logistic regression method was used to find the best-fit model. The α value was set at 0.10 but the variables significantly associated with SA at the 0.05 level were retained. For all statistical tests 95% confidence intervals (CI) were given. The analyses were conducted using Stata® statistical package (Stata/SE 10.0 for Macintosh 2007, StataCorp, College Station, TX).

Results

Table 1 shows the socio-demographic characteristics and health status of participants. Mean age was 79.6 (SD ± 7.1) years and 59.4% of subjects were women. Mean education level was 6.69 (SD ± 5.4) years and 41.7% of participants were married. Current smoking was reported in 8.9% and depressive symptoms in 30.8%. The phenotype of SA was present in 10% (n = 94) of subjects.

Table 1.

Socio-demographic characteristics and health status

	All n = 935	Successful Agers n = 94	Non Successful Agers n = 841	P
Age, mean ±SD	79.6 (±7.1)	76.4 (±4.9)	80.0 (±7.2)	<0.001
Women (%)	59.4	50.0	60.5	0.049
Illiteracy (%)	86.5	85.1	98.9	< 0.001
Married (%)	41.7	51.0	40.6	0.207
Spirituality (%)	71.6	67.0	72.1	0.293
Visual Impairment (%)	50.3	52.1	50.1	0.720
Hypertension (%)	53.8	42.5	55.0	0.021
Hypercholesterolemia (%)	28.0	27.6	28.0	0.934
Dysthyrodism (%)	3.3	1.0	3.5	0.199
Arthritis (%)	15.7	10.6	16.9	0.153
Urinary Incontinence (%)	16.4	8.5	17.3	0.028
History of any Fracture (%)	34.6	28.7	35.3	0.203
History of Hip Fracture (%)	8.6	0	9.4	< 0.001
Former or Current Smoker (%)	8.9	6.3	9.2	0.352
Weight Loss (%)	34.5	21.2	36.0	0.004
Falls (%)	37.9	35.6	38.0	0.711
Physical Pain (%)	40.0	10.6	43.2	< 0.001
Depressive Symptoms (%)	30.8	36.1	30.2	0.235
Good self-reported health status (%)	46.6	81.9	42.6	< 0.001
Good financial self-perception (%)	31.1	41.5	29.9	0.022
Polypharmacy (%)	49.3	24.4	52.0	< 0.001

In comparison with non-successful agers, those who met the phenotype of SA were younger (P < 0.001), more likely to be male (P = 0.049), had higher educational level (P < 0.001), were more likely to live with a partner (P = 0.037), more frequently reported good self-perceived health status (P < 0.001), and had good financial perception (P = 0.022). Also, successful agers reported to have less hypertension (P = 0.021), UI (P = 0.028), and physical pain (P < 0.001).

The univariate logistic regression analysis showed that the following variables were positively associated with the phenotype of SA: being male, having higher educational level, living with a partner; as well as reporting good self-perceived health and financial status (Table 2). Conversely, the variables negatively associated with SA were: age, hypertension, UI, polypharmacy, and physical pain. However, the multivariate logistic regression analysis showed that age (OR 0.92, CI 95% 0.88 to 0.96; P < 0.001), illiteracy (OR 0.70, CI 0.01 to 0.94; P = 0.021), physical pain (OR 0.26, 95% CI 0.01 to 0.52; P < 0.001), and polypharmacy (OR 0.33, 95% CI 0.19 to 0.55; P < 0.001) were independently, but inversely, associated with the probability to show the phenotype of SA whereas only good self-reported health status had a positive and independent association to SA (OR 3.90, IC 95% 2.22 to 6.87; P < 0.001) (Table 3).

Table 2.

Univariate logistic regression analyses of successful aging

Variable	OR (CI 95%)	P
Age (years)	0.91 (0.87 to 0.95)	<0.001
Male	1.53 (1.00 to 2.35)	0.050
Illiteracy	0.06 (0.01 to 0.44)	0.006
Married	1.57 (1.02 to 2.42)	0.036
Good financial self- perception	1.65 (1.07 to 2.56)	0.023
Good health status self-perception	6.11 (3.55 to 10.51)	<0.001
Hypertension	0.60 (0.39 to 0.93)	0.022
Urinary Incontinence	0.44 (0.20 to 0.93)	0.032
Physical Pain	0.15 (0.07 to 0.30)	<0.001
Polypharmacy	0.24 (0.12 to 0.52)	<0.001

Abbreviations: OR, Odds ratio; CI, Confidence intervals.

Table 3.

Correlates of successful aging

Variable	OR (CI 95%)	P
Age (years)	0.92 (0.88 to 0.96)	< 0.001
Illiteracy	0.70 (0.01 to 0.94)	0.021
Good health-status self-perception	3.90 (2.22 to 6.87)	<0.001
Physical Pain	0.26 (0.01 to 0.52)	<0.001
Polypharmacy	0.33 (0.19 to 0.55)	<0.001

Abbreviations: OR, Odds ratio; CI, Confidence intervals.

Discussion

As expected, in this study of Mexican community-dwelling elderly, age, illiteracy, physical pain, and polypharmacy were negatively and independently associated with a biomedical model of successful aging whereas good self-perceived health status was the only positive correlate. To our knowledge, no previous work has been published describing SA among Latin- American elderly populations.

Worldwide, the frequency of SA varies significantly ranging from 0.4% to 95%. This inconsistency could be explained by the diversity of operational definitions (5). It has been observed that the larger the number of components included on a biomedical model of SA, the lower the probability of a population to meet such criteria. However, in the present study the prevalence of the phenotype was 10%, which is similar to that reported in previous work using the components described by Rowe and Kahn (5, 9, 16). This finding alludes to assume that when a biomedical model of SA has been used, the frequency of successful aging may not differ substantially in populations from developing countries when compared with reports from other latitudes.

It is not surprising that age is inversely associated with SA since the core elements of the biomedical phenotype are co-morbidity and disability. A study carried out in over 6,500 subjects aged 65 or older, with a follow-up of 4 years, showed that increasing age was the most important risk factor for functional decline and therefore for disability rending SA, as per this phenotype, less attainable for the older old (9, 17). On the other hand, younger age has been repeatedly described as one of the factors associated with the phenotype of SA (5). Interestingly, age was not the strongest correlate of SA in the present study, other factors such as self-perceived health-status, the presence of physical pain, or polypharmacy were more strongly associated. This finding coincides with the view that more comprehensive phenotypes should be taken into account when defining SA.

In addition, illiteracy was negatively associated with SA. Educational level has been repeatedly described as a factor positively associated with this phenotype; nevertheless, higher educational levels and not illiteracy have been described elsewhere (18, 19). Previous work has shown an association between disability and illiteracy, which suggests that this cognitive resource, no mater how extensive, is extremely important when maintaining an appropriate functional status in older subjects (20, 21). It is noteworthy that apart from illiteracy, other demographic variables were not associated with the phenotype of SA in our sample; nevertheless, this finding has been already described elsewhere and seems to persist in different cultural settings (5), which further underscores the importance of physical health and social interaction.

In contrast to age, educational level or self-perceived health status, polypharmacy has not been described before as a correlate of SA. The presence of this association may reflect more thoroughly the effect of health status compared with co-morbidity indexes often used in biomedical models of SA. Polymedicated elderly patients have greater risks of suffering adverse drug reactions and drug interactions as a consequence of physiological changes associated with the aging process itself and with the influence of disease (22). Nowadays, inadequate drug prescription in elderly patients is considered to be a public health problem leading to higher morbidity, mortality, and greater use of health services. Furthermore, polypharmacy is associated with several geriatric syndromes including UI, cognitive impairment, falls, and frailty. Improving polypharmacy has been proposed as a measure to counteract the effects of frailty thus it is not nonsensical to think that its resolution may have beneficial effects on the way we age (23, 24).

Another factor inversely associated with SA was pain. Pain is one of the main reasons for consultation in geriatric clinics, it leads to great social, functional, financial, and even cognitive impairment (25). Tate et al. reported that one of the essential factors required for older adults to report good self-perceived health-status is the absence of pain (16). On that line of thought, good self-reported health status was the only factor positively associated with the probability of expressing the phenotype of SA in the present study. The subjective component of health has shown to be influenced by social, cultural, and psychological factors. However, it has also been reported to be a very close measure to the objective health status in elderly populations (26, 27). Moreover, self-reported health status has also been associated with several health-related outcomes including mortality and frailty (28, 29). Therefore, our results support the usefulness of this subjective indicator as an important measure in clinical and epidemiologic settings.

Several limits must be acknowledged. The cross-sectional nature of the study prevents from establishing the direction of the associations previously described. In addition, the operational definition of SA may represent another limit since it is based on a biomedical model, which misses the subjective components of well-being, and not on a biopsychosocial one; however, the results are similar to those reported elsewhere which suggests that the operational aging model is valid (5, 8, 9). Although the associations found in the present study may seem intuitive or redundant, their presence underscores the availability of clinical markers, used on a daily basis in real clinical practice that might be modifiable in order to promote successful aging if a biomedical point of view is held. Moreover, other clinical issues such as dietary characteristics should be explored (30). In spite of its limits, the present study was conducted in a population-based sample of a scarcely studied population with regards to SA. The description of the correlates of SA in different populations with different cultural, socio-demographic, and economic characteristics may further contribute to the understanding of this phenomenon.

In conclusion age, illiteracy, physical pain, polypharmacy, and self-reported health status were independently associated with the phenotype of successful aging in this population. Even though age is not a modifiable factor, improving pain, polypharmacy, and self-reported health-status is possible and may promote successful aging. The identification of different aging phenotypes, including SA, and their correlates may enable public health initiatives to better promote healthy aging. However, further investigation is still needed.

Conflict of interest

None.

Disclosure statement

All authors state no financial interest, stock, or derived direct financial benefit.

Sponsor's Role

None