Effectiveness of Pulmonary Rehabilitation among COVID-19 Patients: A Systematic Review and Meta-Analysis

Background: Many COVID-19 patients presented with detrimental features, such as impaired respiratory function, physical capacity, and overall poor quality of life. The present study evaluates the effectiveness of pulmonary rehabilitation on COVID-19 patients. Methods: We searched PubMed, Scopus, ScienceDirect, and Google Scholar from 2019 to 2021. The protocol was registered in PROSPERO with the registration number CRD42021273618. We performed statistical analyses via random effects and expressed the outcomes as standardized mean difference (SMD) for continuous variables, with 95% confidence intervals (CI). Results: We included six trials involving 432 patients. The primary outcome showed a significant improvement in physical function (SMD 0.83, 95% CI −0.58 to 1.09; p < 0.001; four trials, 266 participants; high-quality evidence). There was significant difference in anxiety (SMD −0.80, 95% CI −1.23 to −0.37; p = 0.003), physical activity intensity levels (SMD −1.27, 95% CI −2.23 to −0.32; p = 0.009), sleep quality (MD −0.05, 95% CI −0.83 to −0.16; p = 0.004), peripheral muscle performance of lower limbs (SMD 0.90, 95% CI −0.60 to 1.20; p < 0.001), and dyspnoea outcomes (SMD −0.55, 95% CI −0.87 to −0.23; p = 0.007). Conclusions: Pulmonary rehabilitation is an effective adjuvant therapy that minimizes COVID-19 severity in the intervention group compared to the conventional treatment. The findings of this study will need to be considered in the framework of the clinical outcome as observed in the intervention outcome. Additionally, safer data on guideline rehabilitation would be needed to examine whether pulmonary rehabilitation would be a fruitful intervention to reduce COVID-19 severity.


Introduction
In China's Hubei region, an abrupt occurrence of a contagious respiratory disease called COVID-19 was announced by the WHO in 2019 and spread universally [1]. By 15 October 2019, 38,394,169 confirmed cases and 1,089,047 mortalities had been reported by the WHO worldwide [2]. It was declared a pandemic as it spread over 170 nations and over 30 million people worldwide, causing the deaths of over 1 million people as of

Protocol and Registration
The protocol was registered in PROSPERO with the registration number CRD42021273618.

Research Question
Studies about the effectiveness of respiratory rehabilitation among COVID-19 patients were selected based on the "PICOS" (PRISMA-P 2016) technique: Respiratory function.
Peripheral muscle performance of lower limbs.

Data Sources
Two independent authors (S.B.A. and H.A.) conducted an electronic literature search up to September 1 by combining MeSH terminology and keywords with the Boolean operators "OR" and "AND" to find relevant literature. The keywords are ("physical activit*" OR "exercise" OR "pulmonary rehabilitation" OR "telerehabilitation" OR "Respiratory rehabilitation" OR "training" OR "fitness") AND ("Covid-19" OR "SARS-CoV-2" OR "2019-nCoV") (Supplementary File S1). Unvaccinated COVID-19 patients with no age limit. 10. Publications with no language limitation and with full text available. 11. Pulmonary rehabilitation. 12. Randomized controlled trials, and controlled clinical studies.

Exclusion Criteria
All review articles, case reports, commentary, letters, and short communication.

Study Selection
Two authors, S.B.A. and H.A. scanned the papers based on a linear evaluation of names, abstracts, and complete texts (in cases of doubt). The remaining articles were evaluated entirely based on the qualifying criteria before making a final selection. This method was used independently, with the assistance of a third researcher (A.A.I.) in the case of any conflicts or doubts.

Data Extraction
After reading the full article, two authors (S.B.A. and A.A.I.) conducted independent sampling and data extraction from qualifying studies. The studies that were included produced substantial data, which was extracted and published. This includes the first author, journal name, population, year of publication, gender, method (exercise name, duration, intensity, sets, reps, intervention timing, study duration, and outcome measures.

Assessment of Risk of Bias
We checked the risk of bias based on random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessors, completeness of outcome data, selectivity of outcome reporting, and other bias, as discussed in the Cochrane Handbook for Systematic Reviews of Interventions [20].

Measurement of Treatment Effect
We used risk ratios (RR) and 95% confidence interval (CI) to draw forest plots for trials with categorical variables, and estimation of risk differences (RD) and 95% CI were reported as well. If the outcomes were continuous variables, we planned to analyse the data using mean differences (MD) or standardized mean difference (SMD) and 95% CI. We determined the presence of heterogeneity in two phases. First, we compared demographics, contexts, treatments, and outcomes to see any noticeable variability. Second, we used the I 2 statistic [21] to analyse statistical heterogeneity. We conducted a subgroup evaluation on the duration of intervention when it was feasible.

Sensitivity Analysis
We ran a sensitivity analysis to see how the risk of bias affected sequence generation and allocation concealment in the papers that were included.

Summary of Findings Table
We used the Cochrane Collaboration's Grades of Recommendation, Assessment, Development, and Evaluation (GRADE) approach to evaluate the quality of evidence in systematic reviews. The GRADE system defines four levels of quality, with randomized trial evidence being the highest. Depending on the existence of four elements, it might be degraded as moderate, low, or even extremely poor-quality evidence: (i) constraints in study design and implementation; (ii) indirectness of evidence; (iii) unexplained heterogeneity or inconsistency of results; and (iv) imprecision of outcomes. The GRADEpro software was used to show the quality of evidence for each particular outcome, and the evaluation is being phased in alongside the 'Summary of findings' (SoF) table [21].
The SoF table is made up of the following elements: • Key findings that were summarized (participants, comparative and baseline data, and results) [22].

•
Statistical results that have been condensed. • A summary of the evidence's quality, the degree of the effect, and the source of information utilized in the assumed risk.

Results
A total of 9249 studies were retrieved from PubMed, Scopus, Science Direct, and Google scholar using all MeSH keywords in Figure 1. After identifying duplicate articles, a total of 8687 studies were screened for further selection. After reading the articles' title and abstract, a total of 8605 were excluded according to our inclusion and exclusion criteria. Thus, the remaining 82 articles were proceeded for further selection by reading the full texts, out of which 76 were excluded. The remaining 6 articles that met the eligibility criteria were used for data extraction.

Participants
Two out of the six trials were from high-income countries [23,24], and four trials were from middle-income countries [17,[25][26][27]. Four out of the six trials recruited their respondents from hospital settings [17,[25][26][27], while two trials reported enrolling their participants through an informative text or email distributed on social media platforms (WhatsApp and Facebook), television stations, and newspapers, all of which included interviews with members of the research team [23,24]. Three of the six trials performed their exercise at the hospital [13,22,23], while three conducted the exercise at their home [23][24][25]. Table 1 describes the characteristics of the included trials.

Participants
Two out of the six trials were from high-income countries [23,24], and four trials were from middle-income countries [17,[25][26][27]. Four out of the six trials recruited their respondents from hospital settings [17,[25][26][27], while two trials reported enrolling their participants through an informative text or email distributed on social media platforms (WhatsApp and Facebook), television stations, and newspapers, all of which included interviews with members of the research team [23,24]. Three of the six trials performed their exercise at the hospital [13,22,23], while three conducted the exercise at their home [23][24][25]. Table 1 describes the characteristics of the included trials.  Two trials excluded patients if the patient had any of the following diseases: chronic kidney disease, respiratory conditions in the last 12 months, chronic neurological disease, chronic lung disease, and hypertension [23,24]. One trial reported excluding patients if they did not have a smartphone and had cognitive dysfunction [26]. One trial included only those with mild COVID-19 infection without detailing the exclusion criteria [27]. One trial excluded patients with moderate or severe heart disease with haemorrhagic stroke or neurodegenerative diseases [17]. One trial excluded patients from the study if they had dyspnoea of 4-5 episodes, a resting heart rate of more than 100 bpm, uncontrolled chronic illness (e.g., diabetes mellitus with random blood glucose > 16.7 mmol/L, haemoglobin hbA1C > 7.0%), cerebrovascular disease, mental disorder, and participated in other rehabilitation programmes [25].

Intervention
Patients in our included studies were randomly assigned to intervention and control groups. In two trials, the intervention was a pulmonary telerehabilitation programme [23,25]. One trial combined a telerehabilitation programme with additional exercises [24]. In two trials, the intervention was a pulmonary rehabilitation programme [17,27]. In one trial, the intervention was a breathing exercise [26].
There was a difference in the duration of the intervention between our included studies. In three trials, the intervention was one week [23,24,27]. In another trial, the intervention was for 10 days [26]. In two trials, the intervention was for six weeks [17,25]. In three trials, the intervention was done at home [23][24][25], while another three trials were conducted in the hospital [17,26,27].

Excluded Studies
Out of 29 full-text articles, 23 were exempted because they did not satisfy our inclusion criteria, including 1 trial that did not report on the effectiveness of pulmonary/respiratory rehabilitation among COVID-19 patients [28], 2 studies with unclear data, 15 articles with no intervention, and 6 with no control.

Random Sequence and Allocation Concealment
The method of randomization was described in four trials and the random sequence generation was judged as a low risk of bias [17,[24][25][26].
In one trial, the allocation was concealed by central randomization [26]. Liu, Zhang [17] applied computer-generated randomization, and two trials used balanced

Random Sequence and Allocation Concealment
The method of randomization was described in four trials and the random sequence generation was judged as a low risk of bias [17,[24][25][26].
In one trial, the allocation was concealed by central randomization [26]. Liu, Zhang [17] applied computer-generated randomization, and two trials used balanced

Random Sequence and Allocation Concealment
The method of randomization was described in four trials and the random sequence generation was judged as a low risk of bias [17,[23][24][25].
In one trial, the allocation was concealed by central randomization [25]. Liu, Zhang [17] applied computer-generated randomization, and two trials used balanced randomiza-tion [23,24]. In the remaining two trials, the method of randomization was not described; thus, we judged random sequence generation as an unclear risk of bias [26,27] (Supplementary File S1).

Blinding of Participants, Personnel, and Outcome Assessment
In two trials, the patients were blinded during the entire study process and based on the blinding of participants, and the trials were deemed to have a low risk of bias [23,24]. In one trial, one patient in the control group was randomized mistakenly, and the blinding of participants was, thus, judged as a high risk of bias [25]. In one trial, the blinding of the patients was not feasible; thus, the blinding of participants was evaluated as a high risk of bias [26]. In one trial, the patients were aware of all rehabilitation procedures; thus, the blinding of participants was judged as a high risk of bias [17]. The information regarding the blinding of the participant was not provided in one trial and was therefore judged an unclear risk of bias [27]. The assessors were blinded in four trials, and blinding of the outcome assessment was judged a low risk of bias [23][24][25][26]. Meanwhile, two trials did not report if the assessors were blinded and blinding of the outcome were judged an unclear risk of bias [17,27]; however, Liu, Zhang [17] stated that efforts had been made to blind assessors and participants to group allocation, but this cannot be guaranteed.

Incomplete Outcome Data
Six trials measured the primary outcomes and were included in the meta-analysis. In two trials, the intervention was a respiratory telerehabilitation programme [23,25]. One trial combined a telerehabilitation programme with additional exercises [24]. In two trials, the intervention was a pulmonary rehabilitation programme [17,27]. In one trial, the intervention was a breathing exercise [26]. Five trials reported that all participants completed the study, and the bias due to incomplete outcome data was judged as a low risk [17,23,24,26,27]. Of the five trials, three trials measured the primary outcome at one week [23,24,27], one trial at 10 days [26], and one trial at six weeks [17]. The sixth trial also measured the primary outcome at six weeks [25]; however, six patients of the intervention group (10%) did not complete the post-treatment assessment. Two patients who discontinued the intervention, one because of chest pain and one for unspecified reasons, missed the post-treatment evaluation but returned for the follow-up assessment. Contact was lost with four additional patients in the telerehabilitation programme group and five patients from the control group at the final follow-up. However, intention to treat analysis was applied, and the incomplete outcome data was judged to have a low risk of bias [25].

Selective Reporting
All six trials reported the outcomes as specified in their methods section [17,[23][24][25][26][27] and were regarded as low risks of bias.

Other Potential Sources of Bias
We detected no other potential sources of bias.

Outcomes
The primary outcomes in this review were physical function and quality of life. The physical function of participants in the four trials was measured using a six-minute walking test before and after the intervention [17,[23][24][25]. Four trials reported on the quality of life using sleep quality score [27], social support scale [17,26], the physical function of short Form Health Survey-12 and Short Form-36 [17,25], and the mental function of Short Form Health Survey-12 and Short Form-36 [17,25].
The secondary outcomes were dyspnoea, pulmonary function, physical activity intensity, anxiety, depression, and peripheral muscle performance of lower limp. Two trials reported on dyspnoea using the Multidimensional Dyspnoea-12 questionnaire [23], and the modified Medical Research Council questionnaire [25]. Two trials reported on pulmonary function post-intervention using a spirometer [17,25]. Two trials reported on physical activity intensity levels post-intervention using the Perceived Exertion Scale [23,24]. Three trials reported on anxiety post-intervention using the anxiety scale [17,26,27]. Two trials reported on depression post-intervention using the self-depression scale [17,26]. Three trials reported on peripheral muscle performance of lower limp post-intervention using a 30 s sit-to-stand test [23,24], while one trial used squat time [25].

Primary Outcomes
The primary outcomes in this review were physical function and quality of life. A "six-minute walk test" was used in four trials to assess physical function [17,[23][24][25], while "physical health score" and "mental health score" were used in two trials to assess the quality of life [17,25].
Healthcare 2022, 10, x FOR PEER REVIEW 13 of 20 the modified Medical Research Council questionnaire [26]. Two trials reported on pulmonary function post-intervention using a spirometer [17,26]. Two trials reported on physical activity intensity levels post-intervention using the Perceived Exertion Scale [24,25]. Three trials reported on anxiety post-intervention using the anxiety scale [17,27,28]. Two trials reported on depression post-intervention using the self-depression scale [17,27]. Three trials reported on peripheral muscle performance of lower limp post-intervention using a 30 s sit-to-stand test [24,25], while one trial used squat time [26].

Primary Outcomes
The primary outcomes in this review were physical function and quality of life. A "six-minute walk test" was used in four trials to assess physical function [17,[24][25][26], while "physical health score" and "mental health score" were used in two trials to assess the quality of life [17,26].

Physical Function
Pulmonary rehabilitation improved physical function (SMD 0.83, 95% CI −0.58 to 1.09; I² statistic = 0%; p < 0.001; four trials, 266 participants; high-quality evidence) [17,[24][25][26] (Figure 4, Table 2) compared to the standard treatment group.  Table 2) between the pulmonary rehabilitation group and standard treatment group.     ⊕ Very low CI: confidence interval; MD: mean difference; SMD: standardized mean difference; RCT: randomized control trials; a: the included studies recorded a small sample size for both the control and intervention groups; b: participants were aware of all rehabilitation procedures; c: there is moderate heterogeneity in the involved studies; d: there is substantial heterogeneity in the study's outcome; e: there is considerable heterogeneity in the studies; f: blinding was not feasible for participants and researchers in the study; only the evaluator who gave the link of questionnaires and data analyst were blinded for the treatment; g: information regarding the blinding of the participant and the assessor was not provided.
formance of lower limp.

Summary of Main Results
The present review was designed to incorporate all randomized controlled trials evaluating the effectiveness of pulmonary rehabilitation among COVID-19 patients. There was a significant difference in physical function, anxiety, dyspnoea, physical activity intensity levels, sleep quality, and peripheral muscle performance of the lower limbs in the intervention group following pulmonary rehabilitation compared to the standard treatment group. There was no difference in the quality of life, pulmonary function, depression, and social support outcomes between the intervention and standard treatment groups for the limited number of trials included.

Overall Completeness and Applicability of Evidence
We conducted a detailed and elaborated literature review to evaluate the effectiveness of pulmonary rehabilitation among COVID-19 patients. The RCT included in this review comprehensively illustrate pulmonary rehabilitation outcome among COVID-19 patients. Six trials were included in the meta-analysis. We detected a significant improvement in the intervention groups for the various parameters: physical function, anxiety, dyspnoea, peripheral muscle performance of the lower limbs, physical activity intensity level, and sleep quality.

Quality of the Evidence
The quality of trial evidence varies from moderate to very low certainty. For most trials in most domains, there was a low or unclear risk of bias. There was no evidence of selective reporting bias. In the original research and subsequent review, a lack of proper random sequence generation may contribute to treatment effect bias in the original trial and the subsequent review. The risk of performance bias was presented in three trials. Performance bias was unclear in one trial due to the lack of information regarding the blinding of the participant. Only one trial reported a loss to follow-up of 10% among the intervention groups as they did not complete the post-treatment assessment, and intention-to-treat analysis was carried out. The random-effect meta-analysis of the study showed low to moderate heterogeneity, where we have done random-effects meta-analysis, there was no shift in the effect estimate, and although the 95% (CI) was wider in all cases, the overall level of evidence contributing to this review as assessed using the GRADE approach is moderate to very low quality.

Potential Biases in the Review Process
We intended to limit publication bias by exploring several databases without language restrictions and analysing the reference lists of all relevant studies for additional information. All included studies met all the inclusion, and we did not introduce any bias during the review process, all the studies were vividly reviewed and checked for secondary citation.

Limitation of the Study
All the studies included in this meta-analysis illustrate a similar direction of effect; however, we discovered substantial heterogeneity for quality-of-life outcome. In our analysis, we were unable to explain this due to limited trials. We cannot say with absolute certainty that we have identified all the trials in this field. Considering the fact that there were six trials included, we were unable to create a funnel plot for publication bias relating to each outcome.

Agreements and Disagreements with Other Studies or Reviews
To the best of our knowledge, this represents the first systematic review and metaanalysis been carried out to determine the effectiveness of pulmonary rehabilitation among COVID-19 patients. Three different reviews examined the rehabilitation programme for COVID-19 patients [29][30][31]. Fila, Rocco [29] evaluated the effects of a rehabilitation programme on COVID-19 patients and showed significant improvement in dyspnoea, respiratory function, quality of life, and anxiety among the patients who participated in the rehabilitation programme. The study included 32 articles. Goodwin, Allan [30] included six cohort studies examining longitudinal changes in physical function, three cross-sectional studies investigating the difference in physical function and fitness compared with healthy controls, and one randomized controlled trial investigating the effects of an exercise intervention following SARS-CoV infection. Bernal-Utrera, Anarte-Lazo [31] conducted a scoping review to evaluate the effects of rehabilitation in 29 studies on COVID-19 patients and found a reduction in the severity and progress of COVID-19-related diseases, improved quality of life, and pulmonary function.

Implications for Practice
Pulmonary rehabilitation has a significant effect on improving physical function, pulmonary function, dyspnoea, anxiety, depression, physical activity intensity level, and sleep quality. Hence, encouraging pulmonary rehabilitation only to minimize COVID-19 infection without adhering to other established standard medical treatment does not seem to be justified, but it might be valuable to patients' adjuvant therapy. Nevertheless, pulmonary rehabilitation will ensure an effective adjuvant therapy to the conventional treatment, thus improving pulmonary function and quality of life. Though, in communities with a low occurrence of COVID-19 transmission, where a conservative approach is practiced as a means of reducing the disease burden, the impact of this finding would be less appreciated practical-wise. The findings of this review would need to be considered in the context of the clinical outcome as observed in the intervention outcome. Additionally, safer data on guideline rehabilitation would be needed to examine adequately whether pulmonary rehabilitation would be a fruitful intervention to reduce the COVID-19 severity.

Implications for Research
If further studies were conducted to examine the use of pulmonary rehabilitation in COVID-19 patients, they should include a comprehensive physical function test as an outcome and outline safety information. Data on aerobic exercise and resistance exercise for COVID-19 and other respiratory infections should also be collated. Suppose studies are done in remote and less privileged regions or settings with poor access to standard medical care, the adjuvant treatment should include a structured and tolerable pulmonary rehabilitation programme of sufficient duration to ensure the progression of COVID-19 infection and severity are controlled.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/healthcare10112130/s1, Supplementary File S1: Search strategy and risk of bias assessment.