Effectiveness of Respiratory Exercises on Perceived Symptoms of Fatigue among Multiple Sclerosis Patients: A Systematic Review

: Multiple sclerosis (MS) is an autoimmune disease in which fatigue is one of the most frequent and disabling symptoms, and it is believed to be associated with respiratory involvement. Individuals who are physically inactive for long lengths of time display greater symptomatic fatigue. The objective of this systematic review was to analyze the effectiveness of breathing exercises within rehabilitation programs in improving fatigue in patients with MS. A systematic search of electronic databases, including PubMed, Web of Science, Scopus, and PEDro, was conducted up until November 2022. Nine articles, with a total of 290 participants, were selected. The studies combined breathing exercises with other treatment techniques, such as Ai-Chi, Pilates, and upper and lower limb exercises. Four studies used the Modiﬁed Fatigue Severity scale, observing a reduction in the perception of fatigue, in favor of the experimental group. Incorporating respiratory exercises into physical exercise programs, such as Ai-Chi and Pilates, may help to reduce the perception of fatigue. However, the heterogeneity in the protocols and outcome measures makes the generalization of the results difﬁcult. Nonetheless, further studies that include speciﬁc respiratory variables are needed to analyze whether this perceived improvement is associated with enhanced pulmonary capacity.


Introduction
Multiple sclerosis (MS) is an autoimmune disease characterized by an inflammatory demyelinating affectation of the central nervous system [1]. It is the most common neurological non-traumatic disabling disease in young adults, affecting almost 2.5 million people worldwide, and women to a greater extent [2,3]. The prevalence of MS varies depends on geographical area; in North America and Europe, it is high, with more than 100 cases per 100,000 inhabitants, while Eastern Asia and sub-Saharan Africa have an average of 2 cases per 100,000 inhabitants [4]. In Spain, there is a prevalence of 26.6 per 100,000 inhabitants, and an incidence of 1.34 per 100,000 inhabitants, with the higher-latitude areas presenting the highest figures [5].
A predisposition to MS is complex, and involves both environmental and genetic factors [6]. It can trigger important physical and mental symptoms, such as reduced mobility, balance disorders, muscle weakness, cognitive impairment, decreased gait function, double vision, spasticity, and fatigue [7]. Respiratory infections and complications pose a significant risk of mortality among this population, with a staggering 11.7-fold higher

Study Design
A systematic review was carried out on 18 November 2022, following the standards of the PRISMA declaration [29] (see PRISMA checklist in the Supplementary Material). It was registered in the database of the International Prospective Registry of Systematic Reviews, PROSPERO (identifier: CRD42023390576).

Search Strategy
The descriptors used for the search were the following: "Multiple Sclerosis" [Mesh], "Multiple Sclerosis", "Sclerosis, Multiple", "Breathing Exercises" [Mesh], "Breathing Exercises", "Exercise, Breathing", "Respiratory Muscle Training", and "Muscle Training, Respiratory", in the following databases: PUBMED, Web of Science, Scopus, and PEDro. As for the Boolean operators, "AND" and "OR" were used, according to Table 1. The search strategy was conducted in all fields. Certain databases that did not provide any results were omitted.  The search strategy was developed by two researchers independently (LTA and RLL). A third researcher (MAVM) was consulted if there was any disagreement. After an initial review, 111 articles were considered potentially relevant, and an exhaustive reading of their full text was carried out, with special attention paid to the intervention, relief of symptoms, number of sessions, and duration of treatment. Reference lists in the selected studies were manually searched for eligible studies, and papers found to cite the selected studies were assessed.

Selection Criteria
The following PICOS eligibility criteria were used in the selection of articles: Articles were excluded if they used a sample population with other pathologies, such as Parkinson's or others related to nervous system involvement, and those articles whose main treatment component was not breathing exercises. Also excluded were articles not available in the English or Spanish language, observational studies, case study reports, and systematic reviews.

Study Selection Process and Data Extraction
Two authors (LTA and RLL) independently retrieved relevant articles and extracted data from the included studies. If needed, any disagreements were discussed with, and resolved by, a third author (MAVM). The reference lists of included articles were also screened for more relevant articles, and the corresponding authors of some articles were contacted for nonreported information. The following information was codified: author, year, type of study, sample size, age, duration, measured variables, (scales and measures), and aim of results (intervention, comparison, follow up).

Risk of Bias Assessment
To assess the methodological quality of clinical trials, the Physiotherapy Evidence Database (PEDro) scale-one of the most widely used in the field of physical therapist interventions [30]-was applied by the two leading researchers independently (LTA and RLL). The PEDro Scale consists of 10 criteria, each receiving either a yes or no score. The PEDro score allocates three points according to the methods of blinding used (blinding of subject, therapist, and assessor), two points for randomization procedures (random allocation, concealment of allocation), two points for the reporting of appropriate data (baseline characteristics and point-estimates and measures of variability), and 1 point each for the analysis of the data (intent-to-treat analysis) and the adequacy of the follow-up.

Study Selection
Finally, nine articles were selected that met the inclusion criteria and aligned with the objective of the study. The search strategy used to identify the clinical trials for this review in the different databases is described in Table 1. Figure 1 shows the process for the selection of articles.

Study Characteristics
The nine articles contain a total of 290 participants. A summary of the main characteristics of each study is presented in Tables 2 and 3.
In addition, an analysis of the content of the studies was carried out, based on the following variables:    Table 3. Characteristics of the intervention in the selected articles.

Author, Year Intervention (EG) Comparison (CG) Follow Up Scales and Measures
Klefbeck et al., 2003 [31] IMT training, 10 min, 2 times/day, every 2 days, with 4 h between each session, for 10 weeks, participated from home, weekly telephone contact with help and feedback.
Breathing exercises were part of their physiotherapy session. No telephone contact to resolve doubts.
The measurements were taken at the beginning, at the end of the 10 weeks of treatment, and one month after the end of the treatment.
IMT with instructions on use and progression. Optional telephone advice.
Measurements were taken at the beginning and end of the 10 weeks of treatment.
Conventional treatment in a therapy room.
Measurements were taken before treatment, and at 4 and 10 weeks.  Upper limb and respiratory exercises 60 min/2 days a week at a center, and at least 20 min/3 days a week performed independently. In total, 4 weeks.
No treatment, but they attended the center twice a week for 60 min to socialize.
Measurements were taken before and after treatment.
No treatment, but they attended the center twice a week for 60 min to socialize for 8 weeks.
Measurements were taken before and after treatment.

Study Characteristics
The nine articles contain a total of 290 participants. A summary of the main characteristics of each study is presented in Tables 2 and 3.
Various tools were used to measure pain in the study by Castro-Sanchez et al. (2012) [33], including the Visual Analogue Scale (VAS), Pain Rating Index (PRI), Present Pain Intensity Scale (PPI), and the McGill Pain Questionnaire (MPQ) [33]. They also assessed disability using the Roland-Morris Disability Questionnaire (RMDQ) [33], and spasticity using the Visual Analogue Scale of Spasticity (spasm VAS) [33].
The CORE was measured with several tests, such as the side bridge test, the modified Biering-Sorensen test, the trunk flexion test, and the prone bridge test [36].
Sleep quality and insomnia were measured using the Pittsburgh Sleep Quality Questionnaire (PSQI) and the Insomnia Severity Index (ISI) [38].
Finally, psychological stress was measured in one of the studies by means of the Scale of Clinical Results in the Routine Assessment (CORE-OM) [38]; and depression, using the Beck Depression Inventory [33].

Intervention and/or Treatments Applied
Four of the studies used devices to perform respiratory muscle training [31,32,35,39], as follows.
The Inspiratory Muscle Training (IMT) device blocks airflow until the patient generates enough inspiratory pressure to overcome the resistance provided by the valve [32]. In Klefbeck et al. (2003) [34], patients had one week to become familiar with the Threshold device, and training consisted of three sets of 10 resisted breaths, performed at home, with follow-up. The training resistance began at between 40% and 60% of the previously calculated PImax, if they did not report a value greater than 17 on the Borg scale. In contrast, the control group (CG) performed breathing exercises as part of their physiotherapy session, but did not receive any follow-up, or contact to resolve any questions [31].
The Respiratory Muscle Training (RMT) device is another customized instrument for resistance training in the respiratory muscles, developed by the Center for Research and Education in Special Environments at the University at Buffalo [35]. The patients performed three sessions, one in the laboratory and the other two at home, using a nose clip and a T-shaped mouthpiece with inlet and outlet valves. The nozzle was connected to a pressure transducer and a laptop computer. Following a 10 s cycle, the patients inhaled and exhaled with pre-set resistances based on the percentage of their PImax and PEmax. The CG also received a device with instructions on its use and the progression of its application, with the option of telephone counselling [35].
Ghannadi et al. (2022) used a POWERbreathe device, with an initial resistance of 30% of the initial MIP. A follow-up was conducted with patients via telephone. The CG did not undergo any treatment; they only received brochures on lifestyle education and the importance of regular physical activity [39].
Castro-Sanchez et al. (2012) [37] and Bayraktar et al. (2013) [37] applied an Ai-Chi treatment in a swimming pool. During the 60 min session, relaxing music was played, and the Ai-Chi exercises were supported using abdominal breathing for 10 min, which involved exercises with deep breathing and wide and slow movements in the arms, legs and trunk. They focused on balance, strength, relaxation, flexibility, and breathing with 16 movements: "contemplating", "floating", "uplifting", "folding", "soothing", "gathering", "freeing", "transferring", "accepting", "accepting with grace", "rounding", "flowing", "relaxing" and "sustaining" [33]. The CG performed the same exercise program as the experimental group (EG) during the relaxation periods, but in a supine position on a mat [33]. Bayraktar et al. (2013) [34] applied Ai-Chi therapy, with 60 min sessions with a 15 min warm-up period, based on free movements in the extremities or activities with different pool materials. They combined deep breathing and worked on balance, strength, relaxation, flexibility, and breathing. Ai-Chi's own program lasted 30 min and consisted of the 16 movements mentioned [34]. The CG performed active leg and abdominal exercises at home [34]. Bulguroglu et al. (2017) [36] applied Pilates therapy [36], wherein the subjects were divided into three groups: Mat Pilates, Reformer Pilates, and the CG. In both Pilates groups, the participants were taught the key elements of this therapy: breathing, focus, the location of the rib cage, and the location of the shoulders, head, and neck. Movements were controlled by the physiotherapist, and corrections were made through tactile and verbal prompts and images. Stretching and postural exercises were used to cool down [36]. In the Mat Pilates group, the difficulty of the exercises was increased using different positions and elastic bands (TheraBand). In the Reformer Pilates group, the level of difficulty of the exercises was maintained through different positions and the resistance of springs. Finally, the CG did a program at home with relaxation and breathing exercises [36].
As for other interventions, Grubic et al. (2019) and (2021) [40,41] used upper and lower limb exercises accompanied by breathing exercises in their treatment. The patients performed exercises both at the Multiple Sclerosis Society Center (MSSC) under the guidance of a physical therapist, and at home. The program included a 15 min warm-up phase, followed by diaphragmatic and thoracic breathing exercises, as well as movements in the upper limbs [37] and lower limbs [38]: a range of motion, coordination, and strengthening with minimal resistance (0.5 kg dumbbells or TheraBands). At the end of each session, 10 min were devoted to stretching the muscles used. The CG did not receive any treatment, but attended the center for social activities [37].

Methodogical Quality Assessment
The methodological quality of the clinical trials included was assessed using the PEDro scale [30], (see Table 4): Table 4. Evaluation of the methodological quality of the articles included, using the PEDro Scale. Klefbeck et al.

Item (PEDRro
Ray et al. [35] 2013 Bayraktar et al. [34] 2013 Grubic et al. [37] 2019 Grubic et al. [38] Ghannadi et al. [39] N: the criteria is not satisfied; Y: the criteria is satisfied. Criterion 1: this item is not used to calculate the PEDro score. Of the nine articles included in this review, five of them scored greater than or equal to 6 on the PEDro scale [32,33,[37][38][39], indicating a good methodological quality. All the included studies presented appropriate data in point-estimates and measures of variability. Eight studies showed adequate baseline comparability or between-group comparisons. These articles did not meet some of the scale items: neither the subjects nor the therapists were blinded to the study (however, this issue could not be omitted, because of the peculiarity of the interventions), or allocations were not concealed (seven studies), among others risk of bias. However, most of studies had a low or moderate risk of bias for the characteristics analyzed. Bayraktar et al. (2013) [34] featured the lowest score, not satisfying the random and concealed allocation, baseline comparability, methods of blinding, follow-up, or between-group comparisons.

Discussion
The objective of this systematic review was to describe the effectiveness of breathing exercises on fatigue in patients with MS. Despite us finding a low number of published articles in this field, when considered together, the studies included in this review show different alternatives for carrying out respiratory muscle training. Whether they opt for a direct treatment of these muscles [31,32,35,39], or combine such training with active exercises [33,34,[36][37][38], the studies indicate that treatment with breathing exercises leads to a reduction in fatigue in patients with MS [33][34][35][36][37][38][39]. In addition, some evaluate the positive impact on improving the quality of life among these patients, as they are able to carry out different tasks more easily during the day. Five articles included in this review [32,33,[37][38][39] display a good methodological quality.
The implementation of respiratory exercises primarily involved utilizing a device specifically designed for the direct strengthening of the respiratory muscles, as well as incorporating other types of therapies that combine diaphragmatic and respiratory muscle work. Most of the authors opt for a device to work resistance during inspiration [31,32,39], while others apply resistance during both inspiration and expiration [35].
According to Klefbeck et al. (2003) [31], IMT aided by a Threshold device benefits the inspiratory muscle strength, while also maintaining said improvement for up to one month after the end of treatment. These results are supported by Fry et al. (2007) [32], who also obtain improvements in the inspiratory muscle strength and even a reduction in the inspiratory muscle deterioration after treatment. However, neither of the two studies achieved significant improvements in the FSS scale values, so these patients did not show changes in fatigue after specific respiratory muscle strength training with devices. On the other hand, Ghannadi et al. (2022) [39] achieved an improvement in inspiratory function, and reduced fatigue, according to the MFIS scale. Participants reported feeling better and being able to do more tasks throughout the day with less fatigue. These results are very similar to those reported by Ray et al. (2013) [35], who obtained improvements in the values of the MFIS scale and a reduction in fatigue in patients, who also received RMT.
Several studies [31,32,35,39] agree that RMT and IMT substantially improve inspiratory muscle strength and reduce muscle deterioration, an observation already made by authors such as Huang et al. (2020) [40], in which patients with severe MS improved their PImax and PEmax values after a ten-week IMT treatment. However, Ray et al. (2013) and Ghannadi et al. (2022) [35,39] show a reduction in fatigue, yet they do also describe improvements in the ability to perform everyday tasks.
One aspect to consider regarding this systematic review is that, when comparing the studies, differences in the level of severity of the MS were detected. In three studies, patients had a mild-to-moderate disease state, with scores of up to a maximum of 6.5 points according to the EDSS scale [32,35,39]; Klefbeck et al. (2003) [31] included patients with a severe-to-major condition, with a score between 6.5 and 9.5 points [31]. However, it is observed that the results of these studies are not directly related to the EDSS scale score, or to a reduction or not in fatigue after respiratory treatment. In contrast, the studies by Gosselink et al. (2000) [41] and Ray et al. (2015) [26] identify weakness in the respiratory muscles as a possible factor that increases fatigue in patients with MS. Their hypothesis associates reduced fatigue with an increase in the strength of the respiratory muscles [39]. Therefore, it would have been interesting to describe whether the perception of reduced fatigue shown by people with MS in this systematic review could be due to an improvement in the respiratory muscles, but there was not enough information in the articles on this variable to allow us to test this hypothesis.
Most of the studies included in this review choose to combine breathing exercises with other treatment techniques, such as Ai-Chi [33,34], Pilates [36], and the exercising of the upper and lower limbs [37,38]. Castro-Sanchez et al. (2012) [33] and Bayraktar et al. (2013) [34], after the application of Ai-Chi [33,34], achieve a significant reduction in the levels of fatigue in the treatment group. These results can be compared with a recently published meta-analysis by Amedoro et al. (2020) [42] on the effect of aquatic therapy in MS, where they highlight not only the perception of reduced physical fatigue, but also improvements in the psychological, motivational, and emotional aspects [42].
Bulguruglu et al. (2017) [36] use Pilates as a treatment in two groups of patients with a mild level of disability. After eight weeks of treatment, the authors conclude that there is a reduction in fatigue in both EGs, reporting significantly positive results according to the FSS, and an improvement in the quality of life, thanks to there being less difficulty in carrying out daily tasks. The study suggests that fatigue decreased as a result of the postural control and balance achieved following the Pilates treatment, something that was also described in the cohort study conducted by Soysal-Tomruk et al. (2016) [43]. This study and that by Gandolfi et al. (2015) [44] show that sensory integration is related to fatigue, although this is questioned by Brichetto et al. (2015) [45], in whose study no direct relationship was found between these variables.
Two of the studies included in this review propose a combined treatment of diaphragmatic exercises with upper and lower limb training. In 2019, Grubić et al. [37] choose to work the upper members, while, in 2021, Grubić et al. [38] include both the upper and lower limbs. Both studies report a significant reduction in fatigue in the EG. The hypothesis defended by these authors is that fatigue is not influenced by a single aspect, but that there are multiple factors through which it can modified [37]; and they indicate improvements in the quality of life of the patients, who had initially indicated a lack of motivation to carry out work at home due to their feeling of fatigue and lack of energy [37,38]. The authors argue that the improvement in both aspects may also be related to the possibility of carrying out the exercises in a place of social support, with other patients with MS, as opposed to carrying out the treatment in laboratories or hospital centers, an aspect that is also revealed in a study by Tacchino et al. (2017) [46]. This improvement could be the result of group exercise creating more motivation to exercise, due to social interaction. Therefore, according to this hypothesis, the impact of a socially supportive and less stressful family environment with other people with MS on exercise motivation cannot be ruled out [38,46].
To elucidate the correlation between fatigue and respiratory muscle weakness, it has been noted that increased neural stimulation is required, to support alveolar ventilation and ensure proper gas exchange in the presence of weakened respiratory muscles, leading to fatigue, as well as the perception of fatigue [39,47]. Interventions such as IMT can lead to various outcomes, including enhanced diaphragm thickness, decreased blood lactate concentration resulting from the improved lactate uptake by trained respiratory muscles, improved ventilatory efficiency due to changes in motor recruitment patterns, and a reduced perception of respiratory effort and reduced perceived exertion [48]. The increased strength and endurance in the inspiratory muscles can minimize the gap between the energy demand and effort required during exercise [49].

Study Limitations
The limitations to this review include the differences in the degree of disease (according to the EDSS scale) in the patients included, as well as the diversity in the action protocols and duration of the studies selected. In addition, because fatigue is a subjective symptom, it is difficult for researchers and patients themselves to measure. None of the studies maintained the treatment for longer than 20 weeks, so it is unknown whether any improvement would be maintained in the long term. Future investigations should be carried out considering longer periods of time. During the search and selection process of the articles, Health Sciences databases, such as CINAHL, EMBASE, or Cochrane Library, were not used, so there may be some potentially relevant articles that were not selected for the study. We could not rule out some degree of selection, publication, and reference bias from the literature searches included. The majority of the papers were published in English; hence, there was a likelihood of missing data available in other languages. In addition, this systematic review included studies that were not randomized controlled trials, so the risk of bias is greater. Due to the low number of suitable studies in this field, and the heterogeneity regarding their various exercise intervention protocols and the outcome variables, a methodically sound meta-analysis might not be appropriate to conduct.

Clinical Practice and Policy Implications of This Study
Among the practical implications of this review, one of its aims was to describe alternatives within conservative treatment, through protocols that include breathing exercises to improve fatigue, which is one of the most common symptoms in patients with MS. It has been highlighted that the effectiveness of conservative treatment leads to an improvement in quality of life and in the performance of everyday tasks. This approach could lead to a reduction in the number of drugs in these polymedicated patients and, therefore, a reduction in the economic cost that pharmacology entails, both for health management centers and for families.

Conclusions
This systematic review examines the effectiveness of respiratory exercises on the perceived symptoms of fatigue among MS patients. The majority of the included studies conclude that incorporating these respiratory exercises resulted in a reduction in fatigue perception after engaging in physical exercise programs, versus the control groups.
We have found a low number of published articles in this field, and a meta-analysis was impossible to conduct. Further studies are needed in order to analyze whether this perceived improvement in fatigue is associated with an enhanced pulmonary capacity. Nevertheless, more studies, a larger number of participants, and short-, medium-, and long-term follow-ups are required, to confirm the current results.