Effects of Inspiratory Muscle Training and Yoga Breathing Exercises on Respiratory Muscle Function in Institutionalized Frail Older Adults

A Randomized Controlled Trial

Abstract

Background: 

In older adults, respiratory function may be seriously compromised when a marked decrease of respiratory muscle (RM) strength coexists with comorbidity and activity limitation. Respiratory muscle training has been widely studied and recommended as a treatment option for people who are unable to participate in whole-body exercise training (WBET); however, the effects of inspiratory muscle training and yoga breathing exercises on RM function remain unknown, specifically in impaired older adults.

Purpose: 

To evaluate the effects of inspiratory threshold training (ITT) and yoga respiratory training (YRT) on RM function in institutionalized frail older adults.

Methods: 

Eighty-one residents (90% women; mean age, 85 years), who were unable to perform WBET (inability to independently walk more than 10 m), were randomly assigned to a control group or one of the 2 experimental groups (ITT or YRT). Experimental groups performed a supervised interval-based training protocol, either through threshold inspiratory muscle training device or yoga breathing exercises, which lasted 6 weeks (5 days per week). Outcome measures were collected at 4 time points (pretraining, intermediate, posttraining, and follow-up) and included the maximum respiratory pressures (maximum inspiratory pressure [MIP] and maximum expiratory pressure [MEP]) and the maximum voluntary ventilation (MVV).

Results: 

Seventy-one residents completed the study: control (n = 24); ITT (n = 23); YRT (n = 24). The treatment on had a significant effect on MIP YRT (F_6,204 = 6.755, P < .001, η² = 0.166), MEP (F_6,204 = 4.257, P < .001, η² = 0.111), and MVV (F_6,204 = 5.322, P < .001, η² = 0.135). Analyses showed that the YRT group had a greater increase of RM strength (MIP and MEP) and endurance (MVV) than control and/or ITT groups.

Conclusion: 

Yoga respiratory training appears to be an effective and well-tolerated exercise regimen in frail older adults and may therefore be a useful alternative to ITT or no training, to improve RM function in older population, when WBET is not possible.

INTRODUCTION

With aging, respiratory muscle (RM) strength decreases over time.¹ The loss of RM strength is due in large part to geometric changes in the thoracic cage, the reduction of costovertebral joint mobility, the degradation of neuromuscular recruitment patterns, as well as the loss of muscle fiber-type (sarcopenia).^1,2 Hence, maximum respiratory pressures decline with aging and become critically low in the eighth and ninth decades.³ When this decline in RM strength is complicated by the onset of comorbidities such as stroke, myocardial infarction, severe musculoskeletal problems, or dementia, an older adult will often require professional long-term care and, in most cases, institutionalization.⁴ The loss of RM strength (declining maximum respiratory pressures) in addition to all of the changes due to the presence of different diseases create a scenario of a downward spiral, as inferred by declining mobility and the loss of tolerance for physical activity. Respiratory muscle can be strengthened with regular physical activity,⁵ and consequently whole-body exercise training as well as activities of daily living are directly linked with the improvement or maintenance of RM function.^3,4 However, in frail older adults who are lacking in general strength, whole-body physical conditioning is rarely possible.⁶ Therefore, if older adults are unable to undergo whole-body exercise training (eg, walking) because of comorbid conditions, specific RM strength training is a useful alternative training method to prevent the clinical deterioration in this population with disabilities who are susceptible to disease.^7–10

One of the most common techniques used for RM strengthening is respiratory threshold loading.^11–12 Inspiratory threshold loading devices (eg, Threshold IMT; Respironics HealthScan Inc, Cedar Grove, New Jersey) have emerged as a simple, relatively and an effective method to increase inspiratory muscle strength and endurance, independent of breathing pattern.¹⁰ Gosselink¹³ noted an additional advantage—the threshold loading device shortens inspiratory time and increases time for exhalation and relaxation, which could delay the onset of inspiratory muscle fatigue. Several studies have investigated the effect of threshold inspiratory muscle training (IMT) devices in healthy participants^14,15 and in participants with respiratory, neuromuscular, and cardiovascular diseases.^16–19 A recent meta-analysis²⁰ concluded that people with more advanced weakness benefit more than those with less severe disease. However, none of these studies incorporated institutionalized frail older adults in their cohorts.

Yoga breathing exercises (Pranayama) are less commonly used for RM strength training. Pranayama combines the inspiration and expiration through one or both nostrils, controls the time of the breathing cycle, and requires an activation of chest and abdomen RMs.²¹ Most studies^22–25 evaluate comprehensive yoga programs, including (a) physical postures or Asanas; (b) different methods of timed breathing or Pranayama; and (c) meditation sessions. Only 2 previous studies assessed the effect of yoga breathing exercises on RM strength, showing a significant increase in maximum respiratory pressures, one in a healthy young population²⁶ and one in a healthy older population.²⁷

However, there is a complete paucity of literature on the training effects of yoga breathing exercises on RM strength in institutionalized frail older adults. Therefore, the purpose of the study was to evaluate the effects of inspiratory threshold training (ITT) and yoga respiratory training (YRT) on RM function in institutionalized frail older adults. The hypothesis was that both RM training programs (ITT and YRT) would improve RM strength and endurance when compared with the control group in institutionalized frail older population.

METHODS

Design

This was an open-label randomized controlled trial in which 81 residents with activity limitation were allocated into 3 balanced groups: 1 control group and 2 training groups (ITT and YRT). The trained participants performed a 6-week supervised RM training program and did not perform other forms of exercise training. The control group did not participate in any form of training during the study period. This study was approved by the University of Valencia ethics committee for human research, was performed within 2008 and 2009, and has been registered in the ClinicalTrials.gov (NCT ID: NCT01624272).

Setting and Participants

The sample was made up of all eligible older adults from 4 nursing homes located in Valencia, Spain. The included nursing homes met similar criteria to classify the residents and similar professional health services. A flow diagram (Figure 1) describes the participant eligibility and randomization of the population who met the following inclusion criteria: (1) older than 65 years; (2) medically stable; and (3) a Mini-Mental State Examination score of 21 points or greater^28,29 to assess the level of cognitive function (participants without moderate or severe cognitive deterioration to assure their proper collaboration in assessment and training processes). Exclusion criteria for the participants were (1) ability to independently walk more than 10 m; (2) chronic pulmonary disease; (3) any acute cardiorespiratory episode within 2 months before the study; (4) muscular, neurological, or neuromuscular disease that could interfere in the proper performance of assessment and/or training protocols; (5) smoking for short term or within the last 5 years before the study; and (6) a terminal disease diagnosis. All participants were informed of the risks and benefits of the study and agreed to participate by signing a consent form.

Figure 1: Enrollment, intervention allocation, follow-up, and data analysis, according with updated guidelines for reporting parallel group randomized trials. IMT indicates inspiratory muscle training; ITT, inspiratory threshold training; max, maximum; min, minimum; YRT, yoga respiratory training.

In sum, the sample was characterized by individuals spending a great deal of their time sitting and lying down, though some walked short distances between rooms (bed room, dining room, bath room, etc). None of the participants left the nursing home for walks (38% used a wheelchair for displacements inside the nursing home, and 62% were unable to independently walk >10 m). Physical functioning and participation in daily activities were not therefore considered as outcomes.

Randomization and Interventions

The recruitment strategies were as follows (Figure 1): first, the research group members (coordinators) selected the nursing homes according to their location (metropolitan area of Valencia) and similar standards (eg, residents' admission and classification criteria, offered services, professional health qualifications) to guarantee the same quality of care; second, health care professionals specialized in geriatrics (a physician and a psychologist per nursing home) selected the older people who met the inclusion criteria mentioned earlier; and third, the same health care professionals identified diagnosed diseases and other circumstances that would exclude the potential participants from the study. Nursing home health care professionals exclusively conducted the recruitment process following the manual of procedures developed by the research group, but they were not involved in any other tasks related to the study.

Then, participants were randomly allocated to 1 of 3 groups: control (n = 27), ITT (n = 27), and YRT (n = 27). A statistician who had no contact with the participants and/or the health care professionals who undertook the measurements and the intervention performed the random allocation sequence (random number generator in SPSS; SPSS, Inc, Chicago, Illinois).

Outcomes and Follow-up

All participants were measured at 4 intervals: at baseline (T₁) and at the end of week 3 (T₂), week 6 (T₃), and week 9 (T₄). Baseline measurements covered demographic characteristics, lung function and RM function, functional and cognitive capacity, and diagnosed diseases. In addition, primary outcomes (maximum inspiratory and expiratory pressures) and secondary outcome (maximum voluntary ventilation [MVV]) were assessed at the 4 time points by a physical therapist not involved in any of the training protocols (ITT or YRT).

Pulmonary volumes and capacities were measured according to the standards published by the American Thoracic Society³⁰: Slow vital capacity, forced vital capacity, forced expiratory volume in the first second, peak inspiratory flow, peak expiratory flow, and MVV were collected by using a Jaeger spirometer (Flow Screen; VIASYS Healthcare GmbBH, Hoechberg, Germany). For RM function, a pressure gauge (Series 2000 Magnehelic Pressure Gauge; Dwyer Instruments, Michigan City, Indiana) was used to measure maximum inspiratory pressure (MIP) and maximum expiratory pressure (MEP), in accordance with the standards required by the American Thoracic Society.³¹ The reference values used to obtain the percentage of the predicted values were those reported by Enright et al³² and Neder et al³³ for MIP-MEP and MVV, respectively.

Testing Sessions

The control group underwent no training but did undergo pulmonary testing at all scheduled time periods (T₁ through T₄). After the initial baseline testing (T₁), the ITT and YRT groups began training for two 3-week blocks of time for 5 days per week, followed by a 3-week follow-up period (T₄) in which no training took place. T₄ was a testing period to determine whether there was carry-over in the training effects. At the end of each 3-week block (T₂, T₃, T₄), all participants underwent repeat pulmonary testing.

Training Protocols

Inspiratory threshold training and YRT shared common features, among them were (1) supervised on a daily basis; (2) performed in groups of 8 to 10 residents; (3) interval-based programs consisting of 7 cycles of 2-minute work and 1-minute rest (this protocol is published as a practical guide for clinicians by Hill et al.¹⁰); (4) sessions took place 5 times per week over a 6-week period, for a total of 30 morning sessions; (5) 2-day familiarization period at the beginning of the protocol, to reduce naivete and increase performance compliance; and (6) safety of training protocols was monitored, using a pulseoximeter (SmartOx WM18000; Weinmann Medical Technology, Hamburg, Germany), before and immediately after the training session. Participants who participated in less than 80% of the sessions in both programs were dropped from the analyses. In the same way as for the recruitment process, the intervention counted with a manual of procedures that guided the physical therapists during the training sessions in the different nursing homes.

The specific features of the ITT were as follows: (1) The Threshold IMT device is an adjustable spring-loaded negative pressure breathing device in which the person must exert a sufficient inspiratory effort to open the valve and allow air to pass through the device to the lungs. (2) The working range of pressures is from 7 to 41 cm H₂O. (3) Participants were permitted to select their own breathing pattern, and expiration was unloaded.¹⁰ The interval-based program consisted of 7 cycles of 2-minute work and 1-minute rest (intervention details are available on request). As a starting point for the first period of training (weeks 1-3), the participants breathed against a load 30% to 50% of their own baseline MIP value (this load is recommended by Gosselink¹³ and Hill et al¹⁰). During this period, the load was progressed every 2 days, according to the participant's tolerance. The T₂ MIP value was used to readjust the inspiratory load of each participant to ensure that the participants were training at optimal loads for the second period of training (weeks 4-6). Similarly, the load was increased according to the participant's tolerance. Two trained physical therapists supervised each session and kept activity journals, recording the participant's increases in inspiratory load (cm H₂O) and the relative difficulty of breathing for each session, using the 0- to 10-point Borg Scale.³⁴

The specific features of the YRT were as follows: (1) The weekly program of breathing exercises was developed by a master yoga instructor (Yoga Siromani) and was easily reproducible by trained physical therapists (intervention details are available on request). (2) Increasing the complexity of the exercises over time to encourage a training effect incorporated principles of increasing repetitions of the exercise, increasing resistance by alternating the exercises (breathing through both nostrils or alternative-nostril breathing) and the timing of the breathing cycle (inspiration: apnea postinspiration: expiration: apnea postexpiration), and performing an active expiration in all cases. Similarly, 2 physical therapists kept activity journals, recording the participants' breathing difficulty by using the 0- to 10-point Borg scale.³⁴ The yoga respiratory protocol was centered on rhythmic slow, deep, and nostril breathing, not including Yogasanas and meditation, as described in most yoga studies.^22–25

Statistical Analyses

Data analyses were conducted by using SPSS 19. Descriptive statistics were calculated for all variables and are reported as mean (1 SD) for quantitative variables and percentages for qualitative variables. Normality as well as other statistical assumptions was tested, and outliers identified before using data-modeling techniques. Variables were all screened for normality and outliers by graphical and statistical means, q-q plots, and Kolmogorov-Smirnov tests for normality, box and whiskers graphs, and z scores (> ±3) for outliers, the methods recommended in the statistical literature.³⁵

Inferential statistics were used to test for statistically significant effects among the variables in the study. Baseline levels on the variables of interest were compared across the 3 groups by means of analysis of variance (ANOVA) (for quantitative dependent variables) or chi-square tests (for qualitative dependent variables), depending on the nature of the variables. ANOVA 3 (group) × 4 (time) were used to analyze the dependent variables in the randomized clinical trial. Effect size was measured with partial eta-squared measures (η²). Results were considered significant if P < .05. Follow-up simple effects for the interactions were performed with Bonferroni adjustments. In addition, nonparametric (Kruskal-Wallis) tests were used to compare the number of comorbidities across groups.

No reliable estimates of expected effect sizes in the context of frail older adults for both treatments existed. Therefore, a priori calculation of sample size was not possible, and sample size was determined, as such groups were always larger than those used in the previous literature on both treatments.

RESULTS

Participants, Compliance, and Dropout

A pool of 355 residents' files was screened from the database of the 4 nursing homes (Figure 1); 81 residents were randomly placed into 1 of the 3 groups, resulting in the 3 balanced groups of 27 participants. Ten participants were lost to follow-up: (a) 3 residents died during the study because of exacerbations of their chronic diseases, and 1 resident died during hospitalization stay after a fall; (b) 2 residents were excluded from analysis because of discontinued intervention; (c) 1 resident dropped voluntarily out of the intervention; and (d) 3 residents needed rest because of exacerbations of their chronic diseases.

No significant difference (P = .351) was found in the total number of sessions attended by participants from both experimental groups: ITT group, 24.8 (3.0) sessions (95% confidence interval [CI], 23.5-26.1); YRT group, 24.9 (3.6) sessions (95% CI, 23.3-25.5).

Baseline Characteristics

No significant differences for variables between the 3 groups were found, except for the variable “height” and “functional and cognitive capacity” (Table 1). Regarding RM function, the values were less than predicted in the complete sample (MIP, 69%; MEP, 59%; MVV, 36%). In the case of the predicted values of MVV (% predicted), however, it must be borne in mind that only 15 of 71 cases (those younger than 80 years) took part in this analysis, and as a consequence, there may be a large bias in this particular statistics (reference values by Neder et al³³). For this reason, this variable (MVV, % predicted) was excluded from analyses.

Table 1: Baseline Characteristics for Each Group: Mean (SD) and 95% Confidence Intervals Shown in Parentheses

Moreover, 17% of the participants were former smokers (control, n = 3; ITT, n = 7; YRT, n = 2; P = .101). None of the participants (nonsmokers and former smokers) were treated with oxygen to maintain their oxygen saturation (SpO₂ ≥ 94%) during the time the study progressed.

The high prevalence of morbidity was due to cardiovascular and degenerative skeletal diseases (Table 1). The other frequent-diagnosed diseases were blindness, deafness, overweight, stroke, and diabetes mellitus sequelae (spasticity, wounds, and amputations).

Effects of Training Protocols on Primary Outcomes

The mixed-factorial ANOVA on MIP, in absolute values (Figure 2), found both a significant time effect (F_3,204 = 40.449, P < .001, η² = 0.373) and a group effect (F_2,68 = 3.465, P = .037, η² = 0.092). The treatment effect showed up in a significant interaction effect (F_6,204 = 6.755, P < .001, η² = 0.166). A second ANOVA on MIP, in percentage values, was performed. The main effects results show a significant time effect (F_3,204 = 38.107, P < .001, η² = 0.359). Group effect was also statistically significant (F_2,68 = 5.512, P = .006, η² = 0.140). The statistically significant interaction effect shows again the effectiveness of treatment (F_6,204 = 6.774, P < .001, η² = 0.166), in line with the results already commented for the absolute values of MIP (Table 2).

Figure 2: Maximum inspiratory pressure absolute values (cm H₂O) for each group and time point: means and standard errors bars (SD). ITT indicates inspiratory threshold training; YRT, yoga respiratory training.

Table 2-a: Effects of Control, Inspiratory Threshold Training, and Yoga Respiratory Training Groups

Table 2-b: Effects of Control, Inspiratory Threshold Training, and Yoga Respiratory Training Groups

Follow-up analyses (simple interaction effects, with Bonferroni adjustment of error I and its CIs) showed an increase of MIP means from baseline testing in the 3 groups (Table 2). However, the YRT group showed a significantly greater increase of inspiratory muscle strength than control and ITT groups. After 6 weeks of training and 3 weeks of nonintervention, the MIP in the YRT group was significantly larger than in the other 2 groups, while ITT and control did not significantly differ.

For the MEP, in absolute values (Figure 3), the results of main effects showed a significant time effect (F_{3, 204} = 53.903, P < .001, η² = 0.442), as well as a group effect (F_2,68 = 5.112, P = .009, η² = 0.131). The statistically significant interaction effect demonstrated the effectiveness of treatment (F_6,204 = 4.257, P < .001, η² = 0.111). Another ANOVA on MEP, in percentage values, was estimated. The main effects were both significant: a time effect (F_3,204 = 54.228, P < .001, η² = 0.444) and a group effect (F_2,68 = 6.701, P = .002, η² = 0.165). The interaction effect was also statistically significant (F_6,204 = 5.048, P < .001, η² = 0.129), in line with the results already commented for the MEP absolute values. Interactions' simple effects (Table 2) showed equal means (P > .05) of MEP in the first assessment and significant differences from the intermediate assessment. However, as the MEP absolute values only showed significant differences between control and YRT groups, the percentage values showed significant differences between control and ITT groups versus the YRT group.

Figure 3: Maximum expiratory pressure absolute values (cm H₂O) for each group and time point: means and standard errors bars (SD). ITT indicates inspiratory threshold training; YRT, yoga respiratory training.

In general, the analyses of both primary outcomes demonstrated that YRT was superior to the other 2 groups. In addition to all the evidence already presented, the training load in cm H₂O and percentage of baseline MIP were measured in the ITT group, each training session during the 6 weeks of the study. Training loads in cm H₂O (Figure 4) significantly improved during the training sessions (F_5,100 = 72.022, P < .001, η² = 0.783). Similarly, the percentage of baseline MIP also significantly increased over the 6 weeks (F_5,100 = 67.193, P < .001, η² = 0.771).

Figure 4: Training data for the inspiratory threshold-training group over the 6 weeks of training. Note that the increases in the training load are expressed both as percentages of baseline maximum inspiratory pressure and as the absolute week pressure (range, 7-41 cm H₂O).

Effect of Training Protocols on Secondary Outcome

Further analyses were calculated on MVV, in absolute values (Figure 5). A mixed-factorial ANOVA on MVV found a significant main effect of time (F_3,204 = 23.374, P < .001, η² = 0.256) and effect of group (F_2,68 = 2.325, P = .106, η² = 0.064). Again, the result of the interaction was statistically significant for the absolute values of the dependent variable MVV (F_6,204 = 5.322, P < .001, η² = 0.135), thus giving support to a treatment effect. Regarding the simple effects to assess the statistical significance of differences within each time point, a greater increase in the YRT group appeared, but unlike the trend of the MIP and MEP, this is not significant for T2 and T3. In the follow-up, there was only a significant difference between the YRT group and the control group (P = .012).

Figure 5: Maximum voluntary ventilation absolute values (L/min) for each group and time point: means and standard errors bars (SD). ITT indicates inspiratory threshold training; YRT, yoga respiratory training.

Finally, a mixed-factorial ANOVA on perceived breathing difficulty by Borg Scale CR10 was performed. No significant time effect (F_5,205 = 1.210, P = .306, η² = 0.029) and group effect (F_1,41 = 1.203, P = .279, η² = 0.029) were found. The result of the interaction was not statistically significant (F_5,205 = 1.015, P = .387, η² = 0.024). The Borg Scale's mean for total sessions for ITT group was 3.2 (0.8) (95% CI, 2.9-3.5) and for YRT group was 3.5 (0.6) (95% CI, 3.2-3.8). Mean difference was not statistically significant (t₂₅ = −1.53, P = .13) (Table 3).

Table 3: Perceived Breathing Difficulty (Borg Scale [0–10]) by Participants of the 2 Trained Groups^a

DISCUSSION

This is the first randomized controlled trial that evaluates the effects of both ITT and YRT performed in institutionalized older adults with significant activity limitation and RM weakness. The general hypothesis was that both RM training programs (ITT and YRT) would improve RM strength and endurance when compared with the control group in the population under scrutiny. This hypothesis was only partially supported by the data, as YRT was effective when compared with gains in the control group, but ITT improvements were not statistically different from those in the control group.

Most randomized controlled trials to evaluate IMT effects on RM functioning^7,8,20 have found significant effects on the primary outcomes in this study. Contrary to these established effects, this study has not found these effects. The majority of studies tested the effects of IMT on other populations, especially in people with chronic obstructive pulmonary disease. Our results are counterintuitive, since the meta-analyses by Lötters et al⁷ concluded that IMT increases the MIP and MVV in people with chronic obstructive pulmonary disease, especially those with muscle weakness. Therefore, in the current sample of the frail older adults with muscular weakness, it was expected to have a significant benefit. With respect to an older population, there were only 3 RCT studies available for comparison: Belman and Gaesser,³⁶ Watsford and Murphy,⁹ and Fonseca et al.³⁷ Unfortunately, 2 studies^9,36 did not use the Threshold IMT device, used a lower age range, the exercise capacity was normal, and the participants were community-dwelling. The study of Fonseca et al³⁷ also had several differences with the present study: lower mean age, the older adults had an adequate exercise capacity, and the dependent variables were not related to measures of RM strength, but measures of functional autonomy. In addition, the absence of experimental effect of the ITT could also be due to the characteristics of training protocol applied in this study. Hill et al¹⁰ recommended an ITT protocol of 3 sessions per week during 8 weeks, whereas our study's protocol had 5 sessions per week over 6 weeks. However, the total number of sessions in our study exceeds the protocol of Hill et al.¹⁰ The control group featured greater MIP improvements than the ITT group over the first 3 checkpoints, and this could be due to the training load being too low during the first 4 weeks of study. Nevertheless, the lowest training load in our study was more than the 30% suggested by Hill et al.¹⁰ However, the effect of the training load on effectiveness of ITT protocol should be further investigated. No previous study with the same population and training protocol has been performed that permits strict comparison with this study. Therefore, the extreme debilitation of the older adults and the institutionalization⁴ seems a potential explanation for the statistically nonsignificant effects of the ITT in this study. Furthermore, the trend of increasing MIP and MEP at T₃ and T₄ seems to point out that a longer training duration could confirm or rule out the effect of the ITT in the frail older adults. Finally, the improvement found in MVV from weeks 6 to 9, after training was stopped, could be due to the training physiological mechanism of adaptation (eg, supercompensation principle) and could indicate a possible training effect.³⁸

Regarding the YRT and its effectiveness, current results presented strong evidence for its positive effects on RM strength (MIP and MEP) and endurance (MVV). Despite there being a recent study,²⁷ which confirms the improvement of respiratory function (MIP and MEP) in healthy older adults, this is the only randomized controlled trial available in the literature with frail older adults. This supports the effectiveness of YRT and could suggest that a new study is needed to evaluate the effectiveness of a longer ITT protocol in the frail older adults with comorbidity and activity limitation. Finally, this difference between the effects of the ITT and YRT did not result in significant differences related to the level of compliance between the 2 training protocols in the institutionalized older population with disabilities. We could speculate about the gain in the YRT group. First, the results of the YRT group could be explained by the greater expansion of the chest to a larger total lung capacity,¹³ giving the older adults the opportunity to have a more optimal muscle length-tension relationship, such that their muscles could generate a greater power after training than they could generate before training. An improved length-tension relationship would provide them with greater power to do the MEP maneuver.^13,39,40 Second, the yoga breathing exercises could train the brain to have better motor unit recruitment, thus giving the participants a more optimal recruitment pattern to generate greater power output.

Strengths and Limitations

To our knowledge, the singular novelty of this study is the comparison of 2 breathing techniques (the more conventional and recommended IMT technique and the unusual and less-known yoga breathing exercises) in institutionalized older adults with significant activity limitation and muscle weakness that has been insufficiently studied to date. Another novelty is to study the improvement of RM strength and endurance with yoga breathing exercises in a randomized controlled trial. Also, the importance of this evidence lies in the following aspects: (a) the mean age of the sample (85 (7) years), since previous research has examined older adults in an age range between 60 and 75 years; (b) the measuring at 4 time points of the main dependent variables (MIP, MEP, and MVV), unlike previous research in which only pre- and postintervention were tested; and (c) finally, our sample size is greater than the average sample size, between 4 and 15 participants per group in previous randomized controlled trial studies.^{9,14,15,22,36,37}

A number of limitations of this study are as follows: (a) our protocol lasted for 6 weeks, while Hill et al¹⁰ has recently recommended a protocol lasting 8 weeks or more; (b) the possibility that the results could depend on the functional and cognitive differential capacities; (c) a learning effect could be present across groups, including controls. The potential learning effect in the first few weeks of training, despite the familiarization sessions at the beginning of the protocol, could be attributable to an improved neuromuscular recruitment pattern rather than RM conditioning.^41,42 Future research should measure RM performance and other outcomes related to quality of life after allowing time for learning and train to different intensities, at least 8 weeks.

Clinical Implications

The significant increase in the proportion of old people in the population, the increase of life expectancy, the progressive functional decline, the demand for long-term specialized care, and the prevalence of decreased RM strength lead us to recommend yoga breathing exercises as a measure to improve and maintain RM function in the population with activity limitation. Apart from the direct benefits of yoga on the RM strength, this technique may be recommended on secondary reasons, such as ease of implementation, it is extremely inexpensive, and may be used in nonambulatory population.

CONCLUSIONS

In this randomized controlled trial, YRT appears as an effective and well-tolerated exercise regimen in frail older adults and may therefore be a useful alternative to ITT or no training, to improve RM function in the older population with general muscle weakness, when whole-body exercise training is not possible.

ACKNOWLEDGMENTS

The authors thank all the residents, the yoga instructor (Lesia Kowalyk), the physical therapists who participated in the training protocols (José Moret Vilar, Juan Francisco Donoso Hurtado, Pablo Martín Sánchez, Clara Guzmán Sospedra, and Estefanía Jiménez Picazo), and the nursing homes' health care professionals responsible for the recruitment.