The Relationship Between Frailty, Obesity and Social Deprivation in Non-Institutionalized Elderly People

Abstract

Background

With the increasing prevalence of obesity and the risk of increased dependency among the elderly, it becomes important to characterize the link between obesity and frailty. The relationship between obesity and social deprivation would be bidirectional, with each influencing the other.

Objectives

Main objective was to study the relationship between frailty as defined by Fried and obesity (Body Mass Index (BMI) and abdominal obesity). Secondary objective was to assess the relationship between frailty and social deprivation.

Materials and Methods

This was a cross-sectional study, with data collected between January 2014 and December 2015 using a senior periodic health prevention examination form used in the 4 sites of the health examination center, in Rhone, among non-institutionalized seniors (≥ 65 years). Frailty was defined according to Fried’s criteria. Obesity was defined by a BMI ≥ 30 kg / m² and a waist circumference > 88 cm for women and >102 cm for men. We studied the association between obesity according to BMI ≥ 30kg / m² on the one hand and abdominal obesity on the other hand with frailty according to Fried. The analyzes were adjusted for gender, age, education level, not being in a relationship and social deprivation quantified by the assessment score of deprivation and health inequalities (EPICES score).

Results

1593 senior health prevention examination forms were studied. According to BMI, senior women were almost twice as likely to be frail when obese (RR = 1.92, 95% CI [1.06 − 3.45], p = 0.018). The results were similar for abdominal obesity in women aged 65–74 years (RR = 2.12, 95% CI [1.03−4.35], p = 0.029). There was no relationship in men for both types of obesity. Seniors who were socially deprived were 2.7 times more likely to be frail than non-deprived seniors (adjusted RR = 2.76, 95% CI [1.808 − 4.203], p <0.001).

Conclusions

Obesity (BMI ≥ 30kg / m² and high waist circumference) was associated with increased frailty among older, non-institutionalized women who came for a periodic health prevention examination. Screening and prevention of obesity in the elderly appears to be a major public health issue, and remains a priority target for action.

Introduction

In recent years, several definitions of frailty have been proposed (1) two of which are the most frequently used in the literature: the first is based on a «physical» phenotype, called the Fried phenotype (Fried 2001) (2), which is the most widely known and best characterized, and the second takes into account criteria based on the concept of «multi-domain frailty» - integrating cognition, mood, motivation, motor skills, balance, capacities for the activities of daily life, nutrition, social condition, and comorbidities (Rockwood 2005) (3). These two approaches of frailty are complementary. Frailty reflects a decrease in physiological reserve capacities that alter the mechanisms of adaptation to stress (4). Its clinical expression is modulated by comorbidities and psychological, social, economic, and behavioral factors (3), and is a marker of risk of mortality and adverse events, including disabilities, dependency in activities of daily living, falls, hospitalizations, and institutional admissions (2, 5, 6). Early identification of frailty is therefore an important public health issue.

Worldwide, the number of obese people has doubled since 1980 (7). Obesity is therefore becoming a concern in the elderly because the prevalence of obesity among older people continuously increases (8).

Furthermore, some studies show that obesity is associated with decreased physical performance (9) that is closely related to frailty (10). This is confirmed in institutionalized elderly patients in whom an association between frailty and obesity is reported (11, 12, 13, 14). However, there are few published studies examining this association among non-institutionalized seniors (14, 15, 16, 17) and none conducted on seniors, who came to a periodic health prevention examination. We therefore conducted a study to investigate this; the main focus was the relationship between frailty and obesity defined by BMI and high waist circumference. The secondary focus was to study the relationship between frailty and deprivation.

Materials and methods

Design of the study and participants

This was a cross-sectional study with data collected between January 2014 and December 2015 using the standardized periodic health prevention examination form used in the HEC (Centre d'Examen de Santé du Rhône de la Caisse Primaire d'Assurance Maladie). The population included non-institutionalized seniors (≥ 65 years) who had voluntarily completed a periodic health prevention examination in one of the 4 sites of HEC. The seniors had either received an invitation letter because they were a priori deprived or they had made the request. Seniors were given the usual health check, plus an assessment of the risks inherent to this population. Seniors with language difficulties and those with severe neurocognitive disorders were excluded (inability to understand the exercises).

Variables studied

Frailty

Frailty was defined according to Fried phenotype (2), when ≥ 3 of the 5 following criteria were present: involuntary recent weight loss ≥ 4.5 kg (or ≥ 5% of weight) for 1 year; tiredness felt (positive response to question «Do you feel tired ?»); slow gait speed, defined by walking speed < 4 m / 5 sec (0.8 m / sec); handgrip strength used as a proxy of overall muscle strength. Measurement was performed with the dominant hand using a hydraulic hand-held BASELINE dynamometer, with seniors in the seated position with elbow flexed at 110°. Seniors were instructed to squeeze the handle as hard as possible for 3 to 5 sec. The measurement was repeated twice after a recovery period of 30 seconds. The highest value of the stronger hand was used in the analyses, expressed in kilograms. Decreased muscle strength was defined by the recommended cut-offs score < 20 kg for women and < 30 kg for men; sedentarity, defined by achieving less than 1.5 h of physical activity per week (such as walking for pleasure, jogging, swimming…).

Obesity

BMI was calculated as weight in kilograms divided by height in meters squared. Underweight was defined as BMI <18.5 kg / m², normal body weight between 18.5 and 24.99 kg / m², overweight 25 to 29.99 kg / m² and obesity when BMI ≥ 30 kg / m² (moderately obese between 30 to 34.99 kg / m², severely obese between 35 to 39.99 kg / m² and morbid obese when BMI ≥ 40 kg / m²). Abdominal obesity was defined by waist circumference > 88 cm in women and > 102 cm in men (according to the criteria of the third report of National Cholesterol Education Program Adult Treatment Panel (NCEP ATP III (2001)) (18). Waist circumference was measured only for seniors aged 65 to 74 years.

Social deprivation

The assessment score of deprivation and health inequalities in the HECs (Score d’Evaluation de la Précarité et des Inégalités de santé dans les Centres d’Examen de Santé, Score EPICES; Appendix 1) is composed of 11 binary questions, each answer is assigned a coefficient corresponding to its weight (calculated by multiple regression). The score thus obtained is quantitative, it varies from 0 (lack of deprivation) to 100 (highest deprivation). The threshold > 30 is considered as deprivation threshold according to EPICES. This score is an individual indicator of social deprivation that takes into account the multidimensional nature of deprivation. It was developed in 1998 by a working group made up of members of the HECs, the technical center for support and training of health examination centers (Centre Technique d’Appui et de Formation des centres d’examen de santé) and the Nancy school of public health from a survey taking into account several elements of deprivation: employment, socio-economic classification, housing, family composition, social ties, financial difficulties, life events, and perceived health. This score is useful for detecting but also for quantifying deprivation. It helps to support and target deprived populations (19). The education level and the relationship status were also studied.

Statistical analysis

The data collected during the periodic health prevention examination were captured by the HEC physicians in their professional software: automated health examination management system (SAGES, Sages Software SA, Mérignac, France). They were then exported to Statistical Package for the Social Sciences (SPSS) software version 20.0 (SPSS Inc, Chicago, IL, USA) for Windows and then analyzed. The association between obesity defined by a BMI ≥ 30kg / m² and frailty according to Fried phenotype, and the association between frailty and abdominal obesity were analyzed. Comparisons between groups were made using the Chi-squared test for qualitative variables and the Student test for quantitative variables. The threshold of significance was been set at 5%. When the normality of distribution was not met, Fisher's exact test was used. Binary logistic regression was used to determine the association between obesity and frailty, adjusting for potential confounders, such as demographic factors (age and sex) and social factors (not being in a relationship, the EPICES score and education level).

Results

A total 1 593 senior health prevention examination forms were completed from January 1, 2014 to December 31, 2015; there were 718 women and 875 men, and the mean age was 70.7 years (Table 1).

Table 1.

Characteristics of the population

	Senior population, n=1593 n (%)
Sex
Women	718 (45.1)
Men	875 (54.9)
		Women n = 718	Men n = 875	p
Age (years)
Mean ± SD	70.7 ± 5.3	71.0 ±5.6	70.4 ±5.1
Range	65–93
65-69	834 (52.4)	357 (49.0)	477 (54.5)
70–74	418 (26.2)	186 (25.9)	232 (26.5)
75–79	217 (13.6)	110 (15.3)	107 (12.2)
≥80	124 (7.8)	65 (9.1)	59 (6.7)
In a relationship				< 0.001
Yes	973 (61.1)	316 (44.0)	657 (75.1)
No	620 (38.9)	402 (56.0)	218 (24.9)
EPICES score
Mean ± SD	28.1 (21.3)	29.0 (20.7)	27.4 (21.8)
≤30	892 (56.2)
>30	696 (43.8)
Education level				0.019
No schooling / primary education	528 (33.1)	216 (30.1)	312 (35.7)
Secondary/higher education	1065 (66.8)	502 (69.9)	563 (64.3)
BMI (kg/m2)
< 18.5	28 (1.8)	22 (3.1)	6 (0.7)
18.5 – 24.99	586 (36.8)	296 (41.2)	290 (33.1)
25 – 29.99	638 (39.4)	248 (34.5)	380 (43.4)
30 – 34.99	270 (16.9)	103 (14.3)	167 (19.1)
35 – 39.99	67 (4.2)	39 (5.4)	28 (3.2)
≥40	14 (0.9)	10 (0.9)	4 (0.9)
		Women n = 542	Men n = 712
Waist circumference * (cm)
≤ 88 cm (W) / 102 cm (M)		331 (61.8)	476 (67.0)
> 88 cm (W) / > 102 cm (M)		211 (38.2)	232 (33.0)

*Waist circumference was measured only for those aged 65 to 74 years.

The overall frequency of frailty was 7.6% (Table 2). Among women, 9.4% (63/674) were considered as frail, and among men 6.1% (50/822) were so. In the total population, among those aged ≥ 80 years 19.3% were frail, for those between 75 and 79 years 9.7% were so, for those between 70 and 74 years 8.5% were so, and among those < 70 years 4.7% were so. In the total population, 22% (n=351) was obese, including 23.7% (n=199) of men and 20.6% (n=152) of women (16.9% was moderately obese, i.e. 19.1% men and 14.3% women and 5.1% was severely or morbid obese i.e. 4.6% men and 6.3% women) (Table 1).

Table 2.

Fried phenotype criteria

	Senior population n (%)
Involuntary weight loss, n=1562
Yes	57 (3.6)
Handgrip strength (kg), n=1540
Women, n= 694
Mean + SD	21.1 ±5.0
Range	5.3−45.3
Handgrip strength < 20 kg	265 (38.2)
Men, n=846
Mean + SD	37.0 ± 8.0
Range	4.3−62.0
Handgrip strength < 30 kg	136 (16.1)
Walking speed (4m/5s), n=1532
Mean ± SD	4.2 ± 1.3
Range	1.1−14.6
Tiredness felt, n= 1560
Yes	365 (23.4)
Sedentarity (< 1.5 h), n= 1557
Yes	190 (12.2)
Frail*, n= 1496
Yes	113 (7.6)

After adjusting on age, education level and EPICES score, obese seniors were one and a half times more likely to be frail than non-obese seniors (RR = 1.55, 95%CI [1.01-2.38], p = 0.036). Women were almost twice as likely to be frail if they were obese (RR = 1.92, 95%CI [1.06-3.45], p = 0.018). For men, there was no significant association between obesity and frailty (RR = 1.12, 95%CI [0.57-2.19], p = 0.744) even after adjusting on age (for ≥ 75 years, RR = 0.99, 95%IC [0.31-3.22], p = 0.991, for ≥ 80 years, RR = 1.26, 95%IC [1.09-1.45], p = 0.187; Table 3).

Table 3.

Association between obesity (BMI ≥ 30kg / m²) and frailty

	RR	[95%CI]	p	RR adjusted*	[95%CI]	p
Total	1.93	[1.28−2.91]	0.002	1.55	[1.01−2.38]	0.036
Women	2.94	[1.71−5.04]	< 0.001	1.92	[1.06−3.45]	0.018
Men	1.12	[0.57−2.19]	0.744
Men ≥ 75 years	0.99	[0.31−3.22]	0.991
Men ≥ 80 years	1.26	[1.09−1.45]	0.187

The threshold of significance was set at 0.05; * Adjustment on age class, education level and EPICES score.

The association between BMI and frailty exhibited a J-curve. No underweight senior was frail. Among normal weight women, 6.1% were frail. The proportion of frail elderly per class of BMI increased with increasing BMI: 8.5% of overweight seniors considered frail, and 16.5% for moderately or severe obese patients, reaching 40% for morbid obese. Regarding men, the association between BMI and frailty also showed a J-shaped curve. As for women, no consultant with a small BMI was frail. With equal BMI, the proportion of frail people was higher in women (relative to the total population for each age group) than that in men (Figure 1).

Figure 1.

Proportion participants considered frail according to BMI category stratified on sex. No underweight senior was frail. Higher BMIs had a larger percentage of frail senior than smaller

There was a significant association between high waist circumference and frailty in women (RR = 2.12, 95%CI [1.03−4.35], p=0.029); there was no significant association in men (RR = 1.42, 95%CI [0.67−2.97], p=0.358).

Independently of the sex, socially deprived seniors, i.e. with an EPICES score >30, were 2.7 times more likely to be frail than non-deprived seniors, after adjusting for confounding factors (age, sex, study level, obesity according to BMI) (adjusted RR = 2.77, 95%CI [1.81−4.20], p<0.001).

Discussion

The present study found that non-institutionalized obese senior women were about twice more likely to be frail than those who were not obese. Surprisingly, in men there was no significant association between frailty and either definition of obesity used herein. Furthermore, socially deprived seniors were just over two-and-half times more likely to be frail than those who were not.

Many studies have been conducted in institutionalized patients (11, 12, 13, 14), but very few in non-institutionalized seniors (14, 15, 16, 17). Woods et al. were the first to investigate the potential link between BMI and frailty among non-institutionalized seniors (17); the authors conducted a prospective study of 40 657 women aged 65 to 79 in the Women's Health Initiative Observational Study (WHI-OS). At baseline, BMI was strongly associated with frailty in obese women, which was what we found herein. Other authors have, however, investigated this association among both women and men (14, 15) and among these García-Esquinas et al. presented results for each sex (14); this longitudinal study, conducted in Spain and based on two cohorts, included 1801 non-institutionalized patients aged ≥60 years and 1289 institutionalized and non-institutionalized aged ≥65 years. The authors found that both obesity (BMI and abdominal obesity) were associated with frailty, and, in particular, in the non-institutionalized population, the association between abdominal obesity and frailty was only found in women (14) as we found herein. To explain the difference according to sex regarding the relationship between obesity and frailty, it is interesting to note that Rockwood et al, who defined frailty (3) and subsequently developed a frailty index, found that, for a given age, women were more frail than men (20).

In the present study we used two definitions of obesity: BMI and waist circumference. For BMI, the measurement of the height is questionable in the geriatric population owing to the effect of osteo-articular diseases; at constant weight, BMI is overestimated in the elderly, and the error is greater overestimated in women than in men (21). BMI gives a poor indication of the degree of adiposity, and a recent study has focused on redefining obesity and overweight thresholds by age and sex; the obesity threshold in men aged 60 to 69 was a BMI ≥ 29 kg / m², and in those aged ≥ 70 years this was a BMI ≥ 30 kg / m², while in women these were, respectively, a BMI ≥ 30kg / m² and ≥ 31kg / m² (22). However, using these revised thresholds, we found no difference in the association between obesity and frailty in this study population both for women and for men. Body composition changes over time; lean mass decreases in favor of fat mass and lipid infiltration in skeletal muscles is observed, leading to sarcopenic obesity (23−26). Thus adiposity is certainly under-evaluated by BMI, and this is more pronounced with ageing (27, 28). One might think that the difference in fat distribution between men and women (29) perhaps favor the appearance of frailty in women compared to men. Further studies would be needed to investigate whether hormonal factors in the adipose tissue (30) of the obese intervene in the onset of frailty and confer protection to men compared to women.

Previous studies have reported that patients who are underweight and as well those who are obese are at higher risk of frailty, representing a «U» shaped curve for the relationship between BMI and frailty in non-institutionalized seniors (17, 31) as well as those who are institutionalized (12). In the present study, the distribution of frail seniors in function of BMI category was a «J» shaped; among those who were underweight none were frail both among women and men. Since the study population is composed entirely of seniors coming for a prevention program exam, underweight seniors were likely to be in good health, paying attention to their health and to their fitness and that they had a constitutional leanness.

The frailty phenotype according to Fried describes alterations of functional reserves based on medical determinants (2) but does not explore the social and environmental side that may play a role in the development of obesity. In the present study, we found that deprived seniors were significantly more likely to be frail than those not in such a situation. Many authors have suggested the value of integrating psychosocial determinants into prevention strategies (32). The simultaneous consideration of the social dimension by using the EPICES score and the frail phenotype according to Fried is essential in order to better explore the global environment of a frail elderly person. An EPICES score > 30 seems to be an indicator of poor compliance with the recommendations made to prevent loss of autonomy (33).

Among the strengths of this study, is that it included a large number of seniors and that we used Fried's criteria because the physical tests are easily performed in consultation. There are however certain limitations that should also be noted. First, there were missing data for ≥ 1 Fried criterion (51 for at ≥ 1 criterion). Second, there was an evaluation bias for two Fried's criteria, which are only declarative: tiredness felt and secondly the estimate of the weight loss that could overestimate or underestimate the number of frail seniors. Third, although we had information about many confounding factors we cannot rule out some residual confounding from factors such as ethnicity (17), cardiovascular diseases (34), smoking and alcohol consumption (16). Finally, one of the controversial points of Fried phenotype is that the definition addresses frailty only in terms of physical functions while the frailty syndrome incorporates other disorders, including cognitive, mood and deprivation (3). By adjusting to the EPICES score, we were able to integrate the socio-economic dimension into the relationship between obesity and frailty.

Conclusion

We found that obesity is clearly related to frailty among non-institutionalized senior women, but there was no association between obesity and frailty among men. Given the increasing proportion of obese people among the elderly, this should be a major public health concern.

Ethical standards

All procedures performed in the study were in accordance with the ethical standards of the institutional research committee.

Conflict of interest

None.

Electronic Supplementary Material

Supplementary material is available for this article at https://doi.org/10.1007/s12603-020-1465-6 and is accessible for authorized users.

Supplementary material, approximately 74.5 KB.