Effects of a 12-Week Exercise Training Program on Physical Function in Institutionalized Frail Elderly

With the increase in life expectancy, the Brazilian elderly population has risen considerably. However, longevity is usually accompanied by problems such as the loss of functional capacity, cognitive decline, frailty syndrome, and deterioration in anthropometric parameters, particularly among those living in long-term care facilities. This randomized controlled trial aimed to verify the effects of exercise training on biochemical, inflammatory, and anthropometric indices and functional performance in institutionalized frail elderly. The sample consisted of 37 elderly people of both genders, aged 76.1 ± 7.7 years, who were randomly allocated into 2 groups: 13 individuals in the exercise group (EG) and 24 in the control group (CG). Anthropometrics, clinical history, functional tests, and biochemical evaluation were measured before and after the completion of a physical exercise program, which lasted for 12 weeks. The 12-week exercise program for frail elderly residents in a long-term care facility was efficient in improving muscle strength, speed, agility, and biochemical variables, with reversal of the frailty condition in a considerable number. However, no effects in anthropometric and inflammatory parameters were noted.


Introduction
Aging is a continuous process in which a progressive decline takes place in most physiological systems. is may expose the elderly to frailty, a syndrome characterized by a decrease in the homeostatic reserve and reduction of the body's ability to endure and perform, leading to a cumulative, vicious cycle of decline in multiple physiological systems and to vulnerability to adverse outcomes [1,2].
In addition to physiological problems, aging is usually related to socioeconomic disadvantages. It is usually accompanied by an increase in expenditure, particularly with medications, in addition to decreasing income, since there is loss of production capacity and dependence almost exclusively of retirement income. On the other hand, the Brazilian family has changed signi cantly over the years, especially after the women's entry into the labor market, making home care di cult and increasing the demand for long-stay institutions. However, institutionalization may result in damage to functional decline and physical dependence, since older people living in the current models of long-term care facilities (LTCFs) do not seem to be properly stimulated [3].
Frailty decreases functional capacity by a ecting the strength necessary to perform everyday activities [4] such as walking, raising from a chair, and reducing balance [5]. It is also insidiously related to a higher incidence of falls and disability as well as to hospitalization and mortality [1,2,6]. Consequently, frailty has healthcare cost implications, and its prevention could lead to a reduction in public costs and individual burden, which grows proportionally as the world population ages [7,8].
Moreover, there is evidence that a decrease in physical capacity of older adults is related to an abnormal biochemical milieu such as insulin resistance, dyslipidemia, and systemic inammation [9][10][11][12], but some controversy remains in this regard. Exercise training may represent a nonpharmacological strategy to prevent or treat frailty [13].
In terms of physical tness, muscle strength seems to play a pivotal role in functional independence, with low muscle strength associated with a variety of comorbidities [14,15] as well as with disability to perform everyday activities [16]. In addition to strength training, bene ts for older adults have been observed in studies that evaluated multicomponent training programs [17][18][19][20]. However, an intervention capable of providing greater bene ts has not been de ned yet [20]. Also, physical exercise programs speci c to the elderly population do not seem to be accessible, especially for institutionalized older people [18].
us, the aim of this study was to verify the e ect of physical training on biochemical, in ammatory, and anthropometric traits as well as on functional performance of institutionalized frail elderly.

Study Characterization and Sample.
is is a randomized experimental study conducted with institutionalized elderly volunteers. e sample consisted of 37 individuals of both genders, aged 60 years or older, who were residents of an LTCF in Brasília, Brazil. Subjects were randomly allocated to 2 groups after a series of recruitment steps as follows: 13 individuals in the exercise group (EG) and 24 in the control group (CG), according to a 2 : 1 ratio rationale between the CG and EG ( Figure 1). Randomization was carried out by means of simple lottery applied to each subject in a process based on blindingly choosing cards which exhibited two options of letter (A or B), each representing one branch (CG or EG), with the amount of one letter in twofold excess compared to the other. e ratio of 2 : 1 between the CG and EG relies in the ethical premise of avoiding exposure of a vulnerable individual to a treatment arm of the study when e cacy is under consideration. is study was approved by the institutional research ethics committee, and the volunteers signed a consent form. Subjects aged over 60 years were diagnosed as prefrail or frail and who did not present limitations that precluded cognitive tests were recruited. Participants were declared t by a medical doctor allowing them to undergo the exercise training program. e elderly diagnosed with dementia, Alzheimer's disease, or Parkinson's disease were excluded from the present study.  evaluation assessed before and after the completion of a physical exercise program that lasted for 12 weeks. e elderly in the EG participated in a training program performed 3 times a week, with sessions lasting 40 minutes, while subjects in the CG were instructed to maintain their usual activities of everyday life and to not change their physical activity habits during the study period.

Anthropometric Variables.
To measure body mass, individuals stood on a Britannia ® digital scale with a resolution of 0.1 kg wearing as few clothes as possible, immobile until the value was stable on the display. Height was assessed with a measuring tape with the subjects erected and arms relaxed, after deep inhalation. e body mass index (BMI) was determined as kg/m 2 . e conicity index was determined by measures of weight, height, and waist circumference using the following mathematical equation: . (1)

Handgrip Strength.
After participants became familiar with the equipment, they remained seated, with their shoulders in neutral position, elbows exed at 90°, and st in neutral position. Subjects were instructed to perform a maximum isometric contraction. ree attempts were made with alternate limbs, with a 60-second interval between attempts. e highest achieved reading was recorded for subsequent analyses. No verbal encouragement was o ered during the test [21].

Timed Up and Go Test (TUG).
A stopwatch was used to measure the time spent to get up from a chair, walk a 3-meter distance on habitual speed, go around an obstacle, and sit again. We used a chair with a 45-inch tall seat, a 65-inch tall armrest, and a full, straight backrest [22].

Sitting and Lift Test.
e sitting and lift test was conducted in a chair with a 45 cm tall seat but with no lateral, armrest support and with a full, straight backrest leaning on a wall. Each time a participant stood up, it was counted out loud, and ve repetitions were timed [23]. ere was a period of familiarization to ensure that the participant's sitting position occupied the largest part of the seat.

Mini-Mental State Examination (MMSE).
Answers were obtained from questions presented in the interview format, and a score cut was set based on the scores of 13 points for illiterate people, 18 points for people with 1 to 8 years of schooling, and 26 points for those with more than 8 years of schooling [24,25].

Depression and Functionality Levels.
e Katz scale was used to assess the autonomy of the elderly to perform the basic activities of daily living, with conditions classi ed according to scores achieved as total (0 point), severe (1 or 2 points), moderate (3 or 4 points), or light dependency (5 points) as well as independent (6 points). e scale of Yesavage was used for the screening of depression, assessing cognitive and behavioral aspects typically a ected among older adults. In the present study, 11 to 15 points accounted for severe depression, whereas 6 to 10 points for mild depression and 0 to 5 as indicative of no depression. e Katz and Yesavage scales were implemented as described elsewhere [26].

Blood Drawing and
Testing. Blood samples were obtained through a venous puncture. Serum triglycerides (TGs), high-density lipoprotein (HDL-c), total cholesterol (TC), and glucose were analyzed by enzyme-based colorimetric methods using commercially available kits (Advia 2400, SIEMENS Healthcare Diagnostics Inc., Tarrytown, USA).

Identi cation of Frailty.
Frailty was identi ed based on criteria described by Fried et al., which include unintentional weight loss, reports of exhaustion, reduction of walking speed, muscle weakness, and low level of physical activity. Elderly who did not display any of the criteria mentioned above were classi ed as nonfrail, whereas patients tting in 1 or 2 criteria were considered prefrail. Only elderly displaying 3 or more criteria were classi ed as frail.
Loss of more than 4.5 kg or 10% of the body weight over the last year was considered as weight reduction. Exhaustion was identi ed when there was self-reported fatigue. Low walking speed was identi ed by measuring the time required to walk a distance   (Table 1). Physical tness and exercises for the training program were chosen after an initial assessment of the volunteers to gather variables to improve the independence of elderly people.
To implement the program, the EG was divided into 3 subgroups of approximately 5 elderly people in each group so that exercises were conducted in a personalized and individualized way. e subgroup was followed at all times by 2 physical educators and 3 monitors who did not know which elders were in each of the groups. e intensity and complexity of exercises were de ned based on the initial assessment and adjusted weekly considering the subjective perception of e ort (SPE) reported by each participant.
Owing to the clinical condition of participants, characterized by a high level of functional impairment, medical clearance was granted to perform up to moderate intensities of exercise, what resulted in exercises between 5 and 7 of SPE being selected according to an adapted Omni scale [28]. e intensity was collected during the exercise sessions during the 12 weeks.

Statistical Analysis.
e sample size was estimated based on an "a priori" calculation, using the dependent variables "TUG" and "handgrip strength." After the analysis of preliminary data, with an alpha of 0.05 and a power of 0.80 for a two-tailed test, and assuming uneven groups, a sample of at least 12 participants for the EG and of 18 for the CG would be required to detect a signi cant di erence in the order of 50% between treatment arms.
Clinical, anthropometric, biochemical, and in ammatory parameters on the postintervention moment were compared between the groups (intervention and control) using analysis of covariance (ANCOVA) models for variables with Gaussian distribution. Nevertheless, the nonparametric ANCOVA was employed for the variables that did not display Gaussian distribution in at least one of the groups. For ANCOVA models, measures obtained during postintervention were deemed dependent variables, whereas the groups (exercise and control) were deemed independent variables, with baseline measures as covariates. Intragroup comparisons of measurements were performed using the Student's t-test for paired samples or using a nonparametric Wilcoxon test whether measures displayed a Gaussian distribution or not, respectively. Comparisons of mean measurements between groups were conducted by the independent t-test. Comparisons of frequencies between groups used the chi-square test. A p value of less than 0.05 was considered signi cant.

Results
EG participants had a 61.5% attendance to exercise sessions and reported an average SPE of 5.2 during intervention. Tables 2 and 3 display descriptive statistics of clinical, anthropometric, biochemical, functional, and in ammatory variables observed in the sample, with their respective intragroup and intergroup comparisons.
Functional performance variables signi cantly improved in the EG when baseline and postintervention values were compared. e EG showed statistically superior left and right handgrip strength measures by 33 and 26%, respectively, in comparison to the CG endpoint levels. Regarding TUG and stand up/sit down tests, the EG had statistically lower values than the CG (38% for TUG and 29% for stand up/sit down). Also in the EG, anthropometric traits and MMSE exhibited no signi cant di erences when baseline and postintervention were compared. None of these variables in the CG presented di erences between the baseline and postintervention moments.
Regarding the biochemical and in ammatory traits, values checked for glucose, insulin, total cholesterol, triglycerides, vitamin D 3 , and CRP showed signi cant di erences when comparing baseline and postintervention moments in the EG, while only glucose values di ered in the CG. No other changes were observed concerning biochemical and in ammatory variables among groups or moments. Figure 2 shows the absolute prevalence of criteria for frailty as well as frequencies of elderly classi ed as frail and prefrail in pre-and postintervention moments. e EG showed a ≈ 34% reduction in the prevalence of criteria for frailty, while the CG showed a reduction of ≈6% between evaluations conducted before and after the exercise program. A decrease of more than ≈73% in the number of elderly classi ed as frail was noted in the EG, with most participants migrating to a prefrail status. On the other hand, the CG remained similar between pre-and postintervention periods.

Discussion
e main ndings of this study showed an improvement in the biochemical and functional capacity (strength, speed, and agility) of residents of long-term care institutions after undertaking a 12-week exercise training program. Improvements  20′  20′  15′  15′  --------Flexibility  --10′  10′  10′  10′  10′  10′  ----Aerobic resistance  20′  20′  15′  15′  15′  15′  10′  10′  15′  15′  10′  10′  Strength training  ----15′  15′  20′  20′  25′  25′  30′  30′  Total time  40′  40′  40′  40′  40′  40′  40′  40′  40′  40′  40′  40′ in biochemical, anthropometric, or in ammatory variables were not apparent when crude endpoints values of the exercise and control groups were compared. However, traits such as serum glucose, insulin, total cholesterol, triglycerides, vitamin D 3 , and CRP showed signi cant improvements when baseline and postintervention stages of the EG are compared. Furthermore, a reduction in the absolute count of criteria elements for frailty was observed as well as in the proportion of the elderly considered frail in the EG after intervention. e baseline physical status of the elderly participants presented extremely low performance scores when compared to studies conducted elsewhere, rending a sample with a major dysfunctional status [29][30][31][32]. Exercise-induced improvements in physical performance of institutionalized older adults are corroborated by previous reports [19], resulting in decreased risk of falls and attenuation of frailty. Nevertheless, Faber et al. reported the e ects of 2 exercise programs with moderate intensity conducted in 15 longterm care institutions and found that both programs had positive e ects on physical performance in prefrail elderly people but not in frail individuals [33]. On the other hand, the intervention described in the present study was successful in reversing the frailty phenotype in more than 70% of the frail elderly [34].  Also, a systematic review of randomized clinical trials with frail elderly patients showed contradictory data regarding improvements in risk for falls, mobility, balance, strength, functional capacity, and body composition [20]. Even among those who have shown improvements in functional capacity, it is not appropriate to attribute the success of these results to a speci c element of the intervention protocol. Although most studies usually describe the intervention program in detail, some reports fail to describe appropriately load-related parameters, such as heart rate, lactate concentration, or SPE during interventions, and that might at least in part contribute to the contradictory results in the literature. e mean PSE of 5.2 obtained herein renders the intensity of our exercise as moderate and supports a bene cial e ect of this exercise intensity on functional parameters [35][36][37], risk of falling [33], balance [38], and self-reported health status [39]. Other studies show that exercise programs with higher intensity seem to generate better results [40,41]. In this context, we believe that, besides intensity, other variables such as the duration, method, and type of the exercise may in uence results in studies involving exercise programs in frail elderly, since responsiveness to training, time, and history of exposure cannot be neglected. e literature shows that participation of the elderly population in regular physical activity is considered protective against various components of frailty [42]. Moreover, physical tness among aged individuals is associated with better functioning of glucose and lipid metabolisms, with greater lean mass as well as with a lower, systemic in ammation [9,11,12,43]. However, the physical training program conducted herein was not e cient in promoting improvements in the investigated interleukins and anthropometric variables. is suggests that the intervention was e ective but limited to promoting adaptations on systems directly related to functionality (mainly neuromuscular e ects) and on few metabolic traits (glucose, insulin, total cholesterol, triglycerides, vitamin D 3 , and CRP). One possible explanation for such a narrowed repercussion can be based on the low adherence to our exercise program (61.5%). For instance, the Brazilian Guidelines for Cardiac Rehabilitation [44] a rm that patients who rmly adhere to programs display more pronounced adaptations in hemodynamic, metabolic, myocardial, vascular, and psychological parameters, not observed in the conditions herein. Another study analyzed the association between adherence and physiological changes in middle-aged adults, concluding that high adherence promoted improvements in physical parameters and quality of life [45]. In general, the literature presents consistent evidence that adherence to the exercise program associates with improved clinical outcomes, functional performance, and quality of life [46], and the levels of attendance described herein can be rendered as a limitation of the study.
us, more studies testing models of physical exercise with institutionalized elderly people are necessary to yield more bene cial protocols, having in mind protocols already in literature that were successful in reducing systemic inammation and regulating the metabolic pro le to boost anabolism and muscle protein synthesis [13]. But any intervention should take into consideration individual aspects. e type, intensity, and duration of the exercise program must be carefully chosen, and the frailty status, medical recommendations, history of diseases, and physical activity should be considered. In this case, programs based on multiple components, which address physical qualities such as mobility, balance, exibility, aerobic capacity, and strength, similar to the model used in this study may be used. Part of literature supports this type of intervention, stating that these programs are more e ective at improving general health of frail elderly people [19,30,47]. e present study is not without limitations. For example, a low participation rate in the exercise sessions was noted. Despite constant stimulation and customized work, adherence to the program was below expectations. Another limitation was the extremely low physical conditioning of the participants, as well as the elevated prevalence of depression, which explains greater social isolation and lower adherence to the exercise program, even though, at the end of the exercise program, there was a signi cant reduction in depression rates.
Nevertheless, our results corroborate that, regardless of physical limitations and frailty severity, elderly people are capable of participating in personalized training programs when their limits are considered. Additionally, strength training must be taken into account and should be carried out in an adapted, assisted way [32]. Training should o er bene ts even when it is carried out in a nonintensive way. In addition, physical exercises can be a very e ective strategy to implement bene cial routines to residents of long-term care institutions, particularly those with greater impairment as well as those diagnosed with frailty.

Conclusion
is study showed that a 12-week exercise training program in elderly patients with the syndrome of frailty living in longterm care institutions was e cient in improving aspects of their functional capacity, including muscle strength, speed, and agility, as well as glucose, insulin, total cholesterol, triglycerides, vitamin D 3 , and CRP. Moreover, the intervention helped to reduce the amount of criteria for frailty syndrome, and there was a reversal of the frailty condition in a considerable number of elderly who attended the program. However, there was no repercussion regarding anthropometric, biochemical, and in ammatory features.

Conflicts of Interest
e authors declare that there are no con icts of interest.