Real-Life Outcomes of a Multicomponent Exercise Intervention in Community-Dwelling Frail Older Adults and Its Association with Nutritional-Related Factors

Most of the studies on physical exercise in older adults have been conducted through randomized clinical trials performed under tight experimental conditions. Data regarding Real-Life physical exercise intervention programs in older adults with different conditions and in different settings, are lacking. This is an interventional, prospective and pragmatic Real-Life study in which fifty sedentary and frail individuals were enrolled. We aimed at determining if a Real-Life exercise intervention outweighs previously reported improvements in a Clinical Trial (NCT02331459). We found higher improvements in the Real-Life exercise intervention vs. the Clinical Trial in functional parameters, such as Fried’s frailty criteria, Tinetti, Barthel and Lawton & Brody scales. Similar results were found in the dietary habits, emotional and social networking outcomes determined through the Short-MNA, Yesavage, EuroQol and Duke scales. The Real-Life intervention led to a significant reduction in the number of falls, visits to the primary care centers and emergency visits when compared to the results of our previously published Clinical Trial. The implementation of a Real-Life exercise intervention is feasible and should be a major priority to improve health-span in the older population.


Introduction
Application of clinical trial results to clinical practice is not straightforward [1]. Issues such as reproducibility and replicability due to restrictive enrolment criteria, experimental design limitations, financial issues and biological variability can all underlie the disparity between the outcomes achieved in clinical trials compared to those achieved in clinical practice.
Replicability and reproducibility are fundamental assumptions in science. Clinicians are generally interested in the replicability of a trial that refers to the ability to obtain consistent results across studies aimed at answering the same scientific question, each of which has obtained its own data [2]. According to Popper, "Non-reproducible single occurrences are of no significance to science" [3]. From the perspective of a practicing physician, reproducing the precise experimental conditions of a trial in ordinary clinical setting is, in most cases, very difficult. Clinical trials are characterized by strict control of the variables under study. This is far from being achieved in Real-Life clinical practice, where the heterogeneity of patients and experimental settings makes this control challenging.

Multicomponent Exercise Program
In our previous Clinical Trial [10], the multicomponent exercise interventio 65 min 5 days per week for 24 weeks. Briefly, the sessions were delivered in group supervised and involved a combination of the following activities: propriocept neuromotor exercises (10-15 min), cardiorespiratory training (initially at 40% o mum heart rate increasing progressively to 65%), strength training (initially at 2 repetition maximum up to 75%) and stretching. Patient exercise compliance wa (95% confidence interval [CI] 38.7%-55.7%). The neuromotor exercises included p sway and dynamic balance, coordination and flexibility of the lumbo-pelvic area. diorespiratory training included walking around a circuit and climbing stai strength training was performed with resistance bands and included isometric, co and eccentric exercises with arms, hands and legs. The stretching exercises include legs and neck. The details of time, intensities and progression of the exercise train be seen in [10]. The ratio of trainers to participants was 1: 15. In the Real-Life study intervention, the sessions were conducted and superv a sport scientist, in group and included five minutes of warming-up, 20 min of s exercises from week 1 to week 8 at 45%-55% of one-repetition maximum, from w week 16 at 65% one-repetition maximum, from week 17 to week 24 at 70%-75% o etition maximum; 20 min of cardiorespiratory exercises from week 1 to week 8

Clinical Trial
Real-Life

Multicomponent Exercise Program
In our previous Clinical Trial [10], the multicomponent exercise intervention lasted 65 min 5 days per week for 24 weeks. Briefly, the sessions were delivered in groups, were supervised and involved a combination of the following activities: proprioception and neuromotor exercises (10-15 min), cardiorespiratory training (initially at 40% of maximum heart rate increasing progressively to 65%), strength training (initially at 25% of 1 repetition maximum up to 75%) and stretching. Patient exercise compliance was 47.3% (95% confidence interval [CI] 38.7%-55.7%). The neuromotor exercises included postural sway and dynamic balance, coordination and flexibility of the lumbo-pelvic area. The cardiorespiratory training included walking around a circuit and climbing stairs. The strength training was performed with resistance bands and included isometric, concentric and eccentric exercises with arms, hands and legs. The stretching exercises included arms, legs and neck. The details of time, intensities and progression of the exercise training can be seen in [10]. The ratio of trainers to participants was 1:15.
In the Real-Life study intervention, the sessions were conducted and supervised by a sport scientist, in group and included five minutes of warming-up, 20 min of strength exercises from week 1 to week 8 at 45%-55% of one-repetition maximum, from week 9 to week 16 at 65% one-repetition maximum, from week 17 to week 24 at 70%-75% onerepetition maximum; 20 min of cardiorespiratory exercises from week 1 to week 8 at 55% HRmax, from week 9 to week 16 at 65%-70% HRmax, from week 17 to week 24 at 70%-75% HRmax; 10 min of neuromotor training and 5 min of stretching (Tables A3 and A4).
It has been shown that exercise programs that seem to result in better outcomes are those performed 3 days per week [18]. Thus, the Real-Life intervention was performed 3 days per week, for 60 min, for 24 weeks.
Heart rate was monitored and supervised in every participant by the sport scientist during all the training programs. A heart rate higher than that designed for the aerobic exercise, dizziness symptoms and muscle or joint pain were criteria to stop the intervention.
Briefly, the main changes made in the Real-Life intervention lie in the number of weekly sessions and the fact that the participants train all the physical capacities in each session. The type of exercise, the materials used and the trainer to participant ratio were the same as in our previous RCT (1:15).

Measurements
The following parameters were recorded: age, gender, social situation, marital status. Anthropometric data: abdominal, brachial and leg girths with a SECA anthropometric belt; lean mass; and fat mass percentages were determined by bioelectrical impedance analysis (Tanita. Inner Scan V BC-601). Functional assessment included: Barthel Index (basic activities of daily living), Lawton & Brody (instrumental activities of daily living), Tinetti (fall risk) and hand grip strength with a Jamar (c) Hydraulic Hand Dynamometer. Cognitive, emotional and social determinations were assessed using the MMSE, Duke social support, EuroQol quality-of-life scale (EQ-5D) and geriatric depression scale from Yesavage. We also determined frailty [19,20] and the nutritional status of the individuals with the Short-MNA scale [21].
Prevalence of other geriatric syndromes, number and risk of falls, number of voluntary hospital admissions, visits to the emergency service and visits to the primary care center in the previous 6 months were also recorded.
All the data were registered using an Apple iPAD, stored at Microsoft's Azure cloud and protected with a VPN.

Statistical Analysis
Statistical analysis was performed using the GraphPad Prism 8 software. Categorical variables were described as the frequency and percentage, and quantitative variables as the mean and standard deviation (SD). Descriptive analyses were carried out for each of the two groups. The between-group differences in the frequency distribution across categories were analyzed using the χ2-test, whereas the mean differences between groups were analyzed using the independent-samples t-test with quantitative variables that showed a normal distribution with the Shapiro-Wilk test, while the variables with non-normal distribution were treated with the non-parametric Wilcoxon test for paired data. An ANCOVA analysis was also performed for the main outcome variables using exercise time and adherence to the intervention program as confounding variables.
The threshold for statistical significance was established at a bilateral α value of 0.05. Figure 2a shows that frailty decreases significantly in the Real-Life exercise intervention. These results are even more pronounced than those achieved in the Clinical Trial intervention group [10]. No significant differences in the basal state of frailty were found between groups before the intervention. However, after the intervention, the Real-Life group shows higher improvements when comparted to the Clinical Trial and the Control Group (see Table 1). The clinical characteristics of the different groups of patients are shown in Table A1. This shows that most of the baseline characteristics of the patients were not altered in the different groups studied. Even if we randomly selected patients with the same characteristics, we observed that the number of falls previous to the Real-Life intervention happened to be higher than in the Clinical Trial, as was the hyperlipidemia. On the other hand, the control group seemed to have a statistically significant lower number of smokers. This is obviously a limitation of the study, but we believe that its biological significance does not hinder the validity of the results and the conclusions achieved in our study.

Effect of a Supervised, Personalized and Social Exercise Program on Age-Associated Frailty
Group (see Table 1). The clinical characteristics of the different groups of patients are shown in Table A1. This shows that most of the baseline characteristics of the patients were not altered in the different groups studied. Even if we randomly selected patients with the same characteristics, we observed that the number of falls previous to the Real-Life intervention happened to be higher than in the Clinical Trial, as was the hyperlipidemia. On the other hand, the control group seemed to have a statistically significant lower number of smokers. This is obviously a limitation of the study, but we believe that its biological significance does not hinder the validity of the results and the conclusions achieved in our study.

A Real-Life Exercise Intervention Improves Basic as Well as Instrumental Activities of Daily Living
Maintaining both basic and instrumental activities of daily living in their patients is a major undertaking for geriatricians. Basic and instrumental activities of daily living were determined using the Barthel and Lawton & Brody Scales, respectively. Both scales were improved after the Real-Life or the Clinical Trial exercise interventions when compared with the participants' basal values (see Figure 2b,c). The improvements were more pronounced in individuals who performed the Real-Life intervention than in those involved in the Clinical Trial. Individuals who did not perform exercise got worse on these scales. Table 1 also indicates that, after the 6-month intervention, the Real-Life and Clinical Trial groups show significant improvements when compared to the Control Group in terms of basic and instrumental activities of daily living.

Enrolling in an Exercise Program Improves Nutritional Habits in Older Adults
The Short-MNA scale determines the risk of malnutrition. This is a useful tool to understand the dietary habits of a population whose nutrition may be inadequate for a variety of reasons. Figure 3 shows that the participants in the Real-Life and Clinical Trial studies scored higher on the Short-MNA scale after the interventions. Subjects in the control group showed lower values in the mini nutritional assessment scale.
derstand the dietary habits of a population whose nutrition may be inadequate for a variety of reasons. Figure 3 shows that the participants in the Real-Life and Clinical Trial studies scored higher on the Short-MNA scale after the interventions. Subjects in the control group showed lower values in the mini nutritional assessment scale.  Figure 4a shows that an exercise program lowers the risk of falls when the gait and balance (Tinetti scale) data is compared between the basal values and those obtained after a 6-month intervention (pre vs. post). Similarly, the participants who did not perform exercise had an increased risk of falling. Table 1 also indicates that, after the 6-month intervention, the Real-Life and Clinical Trial groups show significant improvements when compared to the Control Group.

Exercise Lowers the Number of Falls and Mitigates the Fear of Falling in a Real-Life Intervention
We determined the number of falls in our populations before and after the interventions. Individuals in the control group showed an increase in the number of falls in the 6 months period studied. On the contrary, the participants involved in the Clinical Trial and those who were engaged in the Real-Life intervention showed a very significant decrease in the number of falls (see Figure 4b). The effectiveness of the intervention in reducing the number of falls was again more pronounced in the case of the Real-Life than in the Clinical Trial intervention.  Figure 4a shows that an exercise program lowers the risk of falls when the gait and balance (Tinetti scale) data is compared between the basal values and those obtained after a 6-month intervention (pre vs. post). Similarly, the participants who did not perform exercise had an increased risk of falling. An important psychological factor to be considered is the fear of falling. Falls are such an intense threat to the health and wellbeing of the old population that the fear of falling is a significant cause of concern. Persons who did not exercise (control) did not experience a lowering in their fear of falling (see Figure 4c). In contrast, those who exercised did lower it both in the Clinical Trial and the Real-Life intervention. The lowering of the fear of falling was significantly more pronounced in the Real-Life than in the Clinical Trial intervention groups.

Real-Life Exercise Intervention Improves the Quality of Life in Old Adults
Exercise resulted in a clear improvement in quality of life (as determined by the EQ-5D scale). Figure 5a shows that patients who did not perform exercise significantly lost perceived quality of life in the six months trial duration. On the contrary those who performed the exercise significantly increased their quality of life.
The Duke scale is a questionnaire for the social support perceived by the patient. Again, those who did not follow the exercise intervention significantly lost social support. No changes in social support perception were found in the participants from the Clinical Trial exercise group, whereas those who exercised in the Real-Life intervention improved We determined the number of falls in our populations before and after the interventions. Individuals in the control group showed an increase in the number of falls in the 6 months period studied. On the contrary, the participants involved in the Clinical Trial and those who were engaged in the Real-Life intervention showed a very significant decrease in the number of falls (see Figure 4b). The effectiveness of the intervention in reducing the number of falls was again more pronounced in the case of the Real-Life than in the Clinical Trial intervention.
An important psychological factor to be considered is the fear of falling. Falls are such an intense threat to the health and wellbeing of the old population that the fear of falling is a significant cause of concern. Persons who did not exercise (control) did not experience a lowering in their fear of falling (see Figure 4c). In contrast, those who exercised did lower it both in the Clinical Trial and the Real-Life intervention. The lowering of the fear of falling was significantly more pronounced in the Real-Life than in the Clinical Trial intervention groups.

Real-Life Exercise Intervention Improves the Quality of Life in Old Adults
Exercise resulted in a clear improvement in quality of life (as determined by the EQ-5D scale). Figure 5a shows that patients who did not perform exercise significantly lost perceived quality of life in the six months trial duration. On the contrary those who performed the exercise significantly increased their quality of life.

Real-Life Exercise Intervention Can Result in Substantial Savings in Healthcare Cost Expenses
Reductions in the number of emergency visits or in visits to primary care centres does not only reflects improvements in the health status of frail individuals but also results in a significant reduction in public health expenditure. As seen in Figure 6a, the number of visits to primary care centres decreased in the Clinical Trial intervention group. Moreover, this number was further reduced in Real-Life intervention participants. The reduction in the visits to primary care centres of the Real-Life group resulted in a significant difference when compare with the Control Group after the intervention (see Table 1).
Equally important is the number of visits to emergency care centres. These were not decreased in the clinical trial exercise intervention group but were significantly decreased in the Real-Life exercise intervention participants (Figure 6b). The Real-Life group also shows a significant decrease in the number of visits to emergency care centres after the intervention when compared with the Control Group ( Table 1).
The data shown in Figure 6 point towards the critical importance of exercise in lowering public health expenditure for the ever-growing numbers of frail old persons in our society.
The exercise intervention also improved grip strength and anthropometric parameters such as lean mass and fat mass percentages and abdominal and brachial girths (See Figures A1-A3). The Duke scale is a questionnaire for the social support perceived by the patient. Again, those who did not follow the exercise intervention significantly lost social support. No changes in social support perception were found in the participants from the Clinical Trial exercise group, whereas those who exercised in the Real-Life intervention improved their social support (see Figure 5b). Figure 5c shows that participants from the Clinical Trial control group significantly increased their depressive state as determined by the Yesavage scale. The exercise intervention both in the Clinical Trial and in the Real-Life studies very significantly improved the participant's depressive state. After the intervention, the comparison between groups shows how the reduction of depression criteria, as well as the increase in the perception of quality of life, were more evident in the Real-Life group. However, in perceived social support no significant differences were found (see Table 1).

Real-Life Exercise Intervention Can Result in Substantial Savings in Healthcare Cost Expenses
Reductions in the number of emergency visits or in visits to primary care centres does not only reflects improvements in the health status of frail individuals but also results in a significant reduction in public health expenditure. As seen in Figure 6a, the number of visits to primary care centres decreased in the Clinical Trial intervention group. Moreover, this number was further reduced in Real-Life intervention participants. The reduction in the visits to primary care centres of the Real-Life group resulted in a significant difference when compare with the Control Group after the intervention (see Table 1).
in the Real-Life exercise intervention participants (Figure 6b). The Real-Life group also shows a significant decrease in the number of visits to emergency care centres after the intervention when compared with the Control Group ( Table 1).
The data shown in Figure 6 point towards the critical importance of exercise in lowering public health expenditure for the ever-growing numbers of frail old persons in our society.
The exercise intervention also improved grip strength and anthropometric parameters such as lean mass and fat mass percentages and abdominal and brachial girths (See Figures A1-A3).  Equally important is the number of visits to emergency care centres. These were not decreased in the clinical trial exercise intervention group but were significantly decreased in the Real-Life exercise intervention participants (Figure 6b). The Real-Life group also shows a significant decrease in the number of visits to emergency care centres after the intervention when compared with the Control Group ( Table 1).
The data shown in Figure 6 point towards the critical importance of exercise in lowering public health expenditure for the ever-growing numbers of frail old persons in our society.
The exercise intervention also improved grip strength and anthropometric parameters such as lean mass and fat mass percentages and abdominal and brachial girths (See Figures A1-A3).

Effect of Time and Adherence to the Intervention in the Main Outcomes of the Real-Life Intervention
Using time and adherence as confounding variables and making a global analysis, we can say that the outcomes obtained with the Real-Life intervention outweigh those obtained in the Clinical Trial (see Table 1).
The Real-Life intervention reduced Fried's frailty criteria to a greater extent than the clinical trial groups. We also found a more pronounced increase in both dominant and non-dominant hand grip strength.
Due to the intra-group variability inherent in the Real-Life interventions, we found smaller differences in the post-intervention comparison between groups for the Barthel and Lawton & Brody functional scales. However, the improvements due to the Real-Life intervention in the instrumental activities of daily living are even greater than those achieved in the Clinical Trial group.

Short-Term Effects of a Multicomponent, Social, Personalized and Supervised Exercise Program
Life expectancy has been increasing at approximately 2 years per decade for the last 150 years [22]. Much of this increase, especially from the second half of the 20th century, has been due to increased survival of middle-aged people into old age, but not all of this life extension is spent in good health. An average woman or man with a life expectancy of around 82 years can expect to live 19 of those years (~20%) in poor health [23]. Frailty is a good valuable target in attempts to improve the health-span [19].
There is growing evidence that frailty is not an inevitable and unalterable process; on the contrary, it is amenable to intervention [24][25][26][27].
Any intervention to delay the onset of frailty and, most importantly, the transition from frailty to disability, should of course be effective, but it is important that the beneficial effects are seen in the relatively short term.
It is very well established that life-long practice of salutary habits results in a prolongation of life-and health-span [25,26], but in the clinic, we frequently find persons who are above the age of 70 and who have not carried out a life of salutary habits, especially physical activity. A major question is whether a short-term exercise program could significantly improve the health=span of these persons. In our previous reports on a Clinical Trial [10], we showed that relatively short-term (six months) exercise training results in a significant improvement in health-span in old frailty individuals.
Our results on a Real-Life study show that, with a personalized, multicomponent, supervised and social intervention, we can outweigh the improvements achieved in a Clinical Trial in terms of quality of life of individuals at risk of becoming disabled. Moreover, we have found that enrolment in an exercise program improves nutritional habits in older adults. The participants in the Real-Life and Clinical Trial exercise studies scored higher on the Short-MNA scale after both interventions. It is well known that physical exercise effects on fitness are influenced by nutritional status. The cross-talk between these two lifestyle factors, exercise and nutrition, during aging deserve further research and attention [28].
The age-associated loss of function is intrinsic to all the cellular systems. However, the decline in muscle mass and function, preferentially of our lower body, probably represents the most dramatic and significant of all changes during the aging process [29][30][31]. Muscle power begins to decline after the age of 30 and continues to decline linearly with advancing age [32]. From the age of 50, there is a progressive loss of muscle mass (1-2% per year) and of muscle strength (2-5% per year) [33,34] with clinical consequences. Frailty and other age-related diseases increase muscle catabolism which has important implications in metabolic diseases [35]. The results from our Real-Life intervention showed that there was an improvement in body composition, an increase in muscle mass and a decrease in fat mass (See Table 1 and Figures A2 and A3). More importantly, it was accompanied with improvements in grip strength (See Table 1 and Figure A1). These changes have relevant clinical implications, muscle strength is a strong predictor of slow gait speed, severe mobility limitation, risk of falls and hospitalization and high mortality rate [36]. Older adults with low muscle strength have~2-fold greater risk of mortality compared to stronger old individuals [37].

A Real-Life Exercise Intervention Improves Adherence to the Exercise Program
The optimal frequency for a multicomponent exercise program aimed at the frail older adult is 2-3 days a week [9,38]. It is beneficial for all outcomes, but the physical and psychosocial determinants show the bigger improvements [18]. We have found that a Clinical Trial exercise intervention significantly improves functional parameters in old frail individuals. Interestingly, an adaptation of this strictly controlled randomized clinical trial to a Real-Life setting results in even better outcomes in terms of functionality for the participants.
One of the main adaptations of the Real-Life exercise program was the reduction from five to three exercise sessions a week, with an adherence of 79% (95% confidence interval [CI] 72%-86%), while the adherence to the Clinical Trial program was 47% (95% confidence interval [CI] 39%-56%).
This modification means that, on average, a participant in the Real-Life exercise intervention attended a total of 57 sessions out of 72, while the patients in the Clinical Trial attended an average of 56 sessions out of a total of 120.
The main characteristic of a Real-Life intervention in physical exercise is that the development of the sessions is performed in an everyday environment and close to the participant's home. This accessibility of the program leads to a greater adherence. In agreement with the results found in our study, other pragmatic Real-Life interventions have also reported higher adherence when compared to clinical trials carried out in groups with similar characteristics [4,39].
We also think that the superior benefits achieved with the Real-Life intervention could be explained by the fact that all the physical capacities were trained in every session by the participants in this program, while in the Clinical Trial a different one was trained each day. This allowed a better adaptation to training in the Real-Life intervention [40].
These results, together with those of the recently mentioned pragmatic interventions [4,39], confirm that the implementation of studies in everyday locations is favorable for participants. Moreover, they increase adherence to the interventions, establishing them as an important focus for further research in the Real-Life context.

Economic Impact of the Real-Life Exercise Intervention
The cost of caring for a disabled person is at least 16 times more than the cost of caring for a vigorous one [41].
It is obvious that a major aim for health sciences, including medicine, nursing, physical activity, etc., is to prevent the transition from vigorousness to frailty and from frailty to disability. The figures become especially impressive when one thinks in terms of one given country. For instance, in Spain, in August 2020, there were over 1,346,000 individuals who were disabled out of a total population of approximately 47 million citizens. If we consider that the cost or caring for each one of these disabled persons is, as previously discussed, 14,000€ per year, then the cost of disability in a country like Spain amounts to 18 billion euros per year. In accordance with the health cost rates in the Valencian region the Real-Life exercise intervention resulted in a reduction in spending of 16,628€. This is more than the total reduction in health costs of the clinical trial intervention (11,163€). Table A2 shows that this saving is driven by both reduction in the number of visits to primary care centres and to emergency wards. The reduction in health expenses due to the 6-month Real-Life or the Clinical Trial intervention were −16,629€ (−56.2%) and −11,163€ (−38.7%), respectively [42]. However, those patients that did not follow the exercise program ended with an increase in the average cost of the primary care and emergency visits from 21,485€ to 25,676€ in just half a year [42].
Any measure that we may take, like the ones we describe in this paper, to lower the cost of disability will mean not only a tremendous improvement in human well-being, but also a very serious saving in social costs.

Conclusions
Adherence to a Real-life, social, personalized and supervised multicomponent exercise program results in remarkable improvements in terms of well-being, nutritional habits and in the reduction in health costs. The practical application of the multicomponent physical exercise program resulted in better results than those previously obtained in a randomized Clinical Trial. The implementation of this type of intervention should be a major priority for social security and health services.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the patient(s) to publish this paper.

Acknowledgments:
We thank Marilyn Noyes for reviewing the English aspects of this work and to all the patients that have participated in the study.

Conflicts of Interest:
The authors declare no conflict of interest.
Appendix A Table A1. Baseline characteristics of the participants.