Effects of Exercise Training on Cardiopulmonary Function and Quality of Life in Elderly Patients with Pulmonary Fibrosis: A Meta-Analysis

(1) Objective: Our objective was to conduct a meta-analysis of randomized controlled trials that have evaluated the benefits of exercise training for elderly pulmonary fibrosis (PF) patients. (2) Methods: Studies in either English or Chinese were retrieved from the China National Knowledge Infrastructure (CNKI) and the Wanfang, PubMed, Web of Science and SPORTDiscus databases from inception until the first week of April 2021. Age, body mass index (BMI), and exercise frequency, intensity, type, and duration were considered for each participant. The specific data recorded were the six-minute walk distance (6MWD), maximal rate of oxygen consumption (peak VO2), predicted forced vital capacity (FVC% pred), predicted diffusing capacity of the lung for carbon monoxide (DLCO% pred), predicted total lung capacity (TLC% pred), St. George’s respiratory questionnaire (SGRQ) total score and a modified medical research council score (mMRC). (3) Results: Thirteen studies comprised this meta-analysis (eleven randomized controlled trials and two prospective studies design), wherein 335 patients were exercised and 334 were controls. The results showed that exercise training increased the 6MWD (Cohen’s d = 0.77, MD = 34.04 (95% CI, 26.50–41.58), p < 0.01), peak VO2 (Cohen’s d = 0.45, MD = 1.13 (95% CI, 0.45–1.82), p = 0.0001) and FVC% pred (Cohen’s d = 0.42, MD = 3.94 (95% CI, 0.91–6.96), p = 0.01). However, exercise training reduced scores for the SGRQ (Cohen’s d = 0.89, MD = −8.79 (95% CI, −10.37 to −7.21), p < 0.01) and the mMRC (Cohen’s d = 0.64, MD = −0.58 (95% CI, −0.79 to −0.36), p < 0.01). In contrast, exercise training could not increase DLCO% pred (Cohen’s d = 0.16, MD = 1.86 (95% CI, −0.37–4.09), p = 0.10) and TLC% pred (Cohen’s d = 0.02, MD = 0.07 (95% CI, −6.53–6.67), p = 0.98). Subgroup analysis showed significant differences in frequency, intensity, type, and age in the 6MWD results (p < 0.05), which were higher with low frequency, moderate intensity, aerobic–resistance–flexibility–breathing exercises and age ≤ 70. Meanwhile, the subgroup analysis showed significant differences in exercise intensity and types in the mMRC results (p < 0.05), which were lower with moderate intensity and aerobic–resistance exercises. (4) Conclusions: Exercise training during pulmonary rehabilitation can improved cardiopulmonary endurance and quality of life in elderly patients with PF. The 6MWDs were more noticeable with moderate exercise intensity, combined aerobic–resistance–flexibility–breathing exercises and in younger patients, which all were not affected by BMI levels or exercise durations. As to pulmonary function, exercise training can improve FVC% pred, but has no effect on DLCO% pred and TLC% pred.


Literature Search Strategy
Using guidelines provided by the Cochrane Collaboration, a comprehensive search strategy was devised and applied to the following electronic databases in the first week of April 2021 with no date restrictions: (1) China National Knowledge Infrastructure (CNKI) and the Wanfang, PubMed, Web of Science, SPORTDiscus databases. Articles published in English and Chinese were included, and all terms were searched as free text and keywords where applicable. Scientific databases were searched according to three criteria: participants ("pulmonary fibrosis patients"), medical interventions ("training or exercise", "exercise training", "pulmonary rehabilitation", "physical exercise", "exercise program" or "physical training") and outcomes ("cardiopulmonary fitness or function", "pulmonary function or lung function", "quality of life", "health-related quality of life", "HRQL", or "QOL"). All search strategies were performed in English and Chinese in the relevant databases. All literature was imported into Endnote X9 (Thomson Reuters, Carlsbad, CA, USA), which also removed duplications. Two reviewers screened all titles and abstracts. Once abstracts suggested that studies were potentially suitable, the full-text versions were screened and then included in the review if they fulfilled the selection criteria. A third reviewer was consulted in cases of disagreements. Additional searches included reference list screening and citation tracking (Google Scholar) of all studies.

Selection Criteria 2.3.1. Inclusion Criteria
(I) Studies included PF patients referring to the elderly, defined by the World Health Organization as those aged over 60 years old [22].
(II) Medical interventions related to the operating group consisting of aerobic, resistance, flexibility, and breathing exercises. For the control group, physical therapy and medication under the supervision of a therapist, playing Wii (a video game), and educational lectures.
(III) Studies that included any of the following criteria: six-minute walk distance (6MWD) [23], maximal rate of oxygen consumption (peak VO 2 ), predicted forced vital capacity (FVC% pred), predicted diffusion capacity of the lung for carbon monoxide (DLCO% pred), predicted total lung capacity (TLC% pred), St. George's respiratory questionnaire (SGRQ) [24], and a modified medical research council (mMRC) score [14]. An Egger test based on regression was used to analyze publication bias.
(IV) The study design was either a randomized controlled trial or a prospective study design.

Exclusion Criteria
(I) Case reports; (II) Non-English/Chinese study; (III) Participants with an inventory of interpersonal problems (IIP), connective tissue disorders or extra parenchymal causes of restriction; (IV) Cross sectional, retrospective, systematic reviews, editorial letters or conference abstracts without the full text available.

Data Management
Data (means and standard deviations (SD)) pertaining to participant and study characteristics were extracted and entered into an Excel spreadsheet.

Outcomes
(I) Cardiopulmonary function: Within this review, cardiopulmonary function indexes included peak VO 2 , FVC% pred, DLCO% pred, and TLC% pred. Peak VO 2 was used to predict cardiovascular disease in adults [34,35] and overall mortality [36,37]. The peak VO 2 was measured by spiroergometry; specifically, exercising on a bicycle ergometer or treadmill until the subject reached their maximum. The single-breath diffusing capacity for carbon monoxide (DLCO) was also included. All values were expressed as a percentage of the predicted values reported. The FVC% pred was used to evaluate lung function, determining to what degree it had decreased. It is also useful for assessing the progression of lung disease and to evaluate the effectiveness of treatment [38]. The DLCO% pred was widely used in the diagnosis, classification, treatment, monitoring, and prognosis of PF patients [39].
(II) Quality of life: The six-minute walk test (6MWT) assessed functional limitations and determined functional capacity. As a self-paced and submaximal test, the 6MWD also reflects the ability to conduct daily activities [40,41]. The SGRQ is a disease-specific quality of life assessment tool validated for PF [42][43][44], where a high score implies a poor quality of life [45]. The mMRC scale is a self-rating tool to detect the degree to which breathlessness limits daily activity [46,47].

Statistical Analysis
The effect size was calculated according to Cohen's d [48]. Cohen suggested that d values of 0.2, 0.5, 0.8 represent small, medium, and large effect sizes, respectively [49]. This was calculated using Equations (1) and (2): The mean difference (MD) and SD were calculated using Equations (3) and (4): The number 1 represents the baseline, and number 2 represents the follow-ups. We assumed an R value of 0.40 to impute the missing SD of the mean within-group change according to Follman et al. [50]. In this study, the effect size was represented by d, and the result size by MD. If a study reported results for different durations, each of them was treated as a separate trial [51]. The Cochrane systematic review software Review Manager (version 5.3.5) was used to map the forest. In addition, the 95% confidence intervals were calculated by this software. Meta-analysis was conducted to evaluate the effects of training interventions on PF. Heterogeneity in the studies was analyzed by a forest plot, and the heterogeneity was quantitatively determined by I 2 . This study had a low heterogeneity; therefore, a fixed-effect model was adopted for meta-analysis. If there was statistical heterogeneity among the results, its source was analyzed further, and significant heterogeneity was treated by subgroup analysis.

Publication Bias
For all studies, the potential publication bias was evaluated by Egger' regression test [57]. Egger' regression tests were performed for 6MWD, Peak VO2, FVC% pred, DLCO% Figure 8. mMRC. Notes: Symbols: for single studies, the squares indicate the mean difference, and the relative size of the square is an indication of the weighting of this study towards the overall effect. The endpoints of the horizontal lines are the upper and lower 95% confidence intervals. The large diamonds represent the summed data for the subgroups and all studies included in the meta-analysis; the midpoint of the diamond indicates the mean difference, whereas the endpoints are the upper and lower 95% confidence intervals. Abbreviations: 95% CI, 95 percent confidence interval; IV, inverse variance; SD, standard deviation. If an included study reported results for different durations, each different duration was treated as a separate trial [51]. Vainshelboim 2016 (1) represents a study with a 3-month exercise duration. Vainshelboim 2016 (2) represents a study with an 11-month exercise duration. Holland 2008 (1) represents a study with a 2.25-month exercise duration. Holland 2008 (2) represents a study with a 6.5-month exercise duration. Perez-Bogerd 2018 (1) represents a study with 3-month exercise duration. Perez-Bogerd 2018 (2) represents a study with a 6-month exercise duration. Perez-Bogerd 2018 (3) represents a study with a 12-month exercise duration.

Discussion
The results in this study indicated that exercise training could improve cardiopulmonary endurance and the quality of life.

Cardiopulmonary Endurance
In this study, cardiopulmonary endurance was evaluated with peak VO 2 , which is widely used to assess cardiopulmonary endurance by researchers [58]. In our study, a medium effect size was found, which indicated that exercise training improve peak VO 2 performance in PF patients. Due to the high heterogeneity, subgroup analysis on exercise duration was performed. The confidence intervals of the exercise duration subgroups overlapped, and the effect size was very small when exercise duration was less than or equal to 3 months (Table 3); therefore, the peak VO 2 might not be affected by exercise duration. Exercise types in older adults should include aerobic, resistance, flexibility, balance training, etc. [59]. Our results further indicate that combined aerobic-resistanceflexibility-breathing training can improve the cardiopulmonary endurance of elderly patients with PF.
The findings in the current study support the previous hypothesis that exercise training improves cardiopulmonary endurance in patients with PF [39]. One GQ study showed that exercise training significantly improved peak VO 2 in the elderly in both healthy and disease contexts [60]. Exercise training (2-3 times per week) can effectively improve joints' range of motion and muscle endurance [52]. Especially in the elderly, exercise training preserves bone mass and reduces the risk of falling [61]. The increase in peak VO 2 in the operate group [62] is presumably because long-term exercise training increases cardiopulmonary endurance through improving blood circulation, lowers blood pressure, and improves cardiovascular function [62]. Therefore, the current synthesized evidence supports the opinion that exercise training can improve the cardiopulmonary endurance of PF patients.

Pulmonary Function
Pulmonary function was evaluated with the FVC% pred, DLCO% pred and TLC% pred in this study. One study showed that the loss of pulmonary function may lead to ventilatory limitation in exercise training in the active elderly, which decreases the accumulation of health benefits during physical activity [63]. FVC% pred can be used as an indicator of disease progression, which can be combined with other variables to predict disease progression more accurately [64]. PF patients expand their lungs with more difficulty due to a narrower airway [65]; one study revealed that exercise training can expand airways to increase FVC% pred in healthy subjects [66]. This may provide insights for PF patients.
The results in the current study indicate that exercise training improves FVC% pred performance in patients with PF. Disease progression in PF is monitored by a decline in forced vital capacity (FVC). An absolute or relative decline in FVC% pred of ≥10% is associated with mortality [67][68][69]. Two of the GQ studies on FVC% pred supported exercise training [16,26], whereas two other GQ studies did not support it [19,27]. The combined evidence supported the positive effect of exercise training on FVC% pred. However, the limitations were that data were extracted from one author and all studies were from the same group. Therefore, the effect of exercise training on FVC% pred is still inconclusive and needs to be further studied.
DLCO% pred provided an objective index of disease severity and prognosis [70], which is related to the rate of oxygen uptake by hemoglobin [71]. This study showed that exercise frequency, intensity, type, and duration did not affect the DLCO% pred. The number of studies on the DLCO% pred was abundant and the pool of subjects was large (n = 8, n = 98, respectively); therefore, the lack of benefit from exercise training on DLCO% pred of PF patients was validated.
Only two GQ studies focusing on the effect of aerobic-resistance-flexibility-breathing exercise reported TLC% pred; the results showed that aerobic-resistance-flexibility-breathing exercise had no benefits on the TLC% pred. No study with an adequate sample size (n ≥ 30) was found to evaluate the effects of exercise training on TLC% pred. The effect of exercise training on TLC% pred in patients with PF needs further study. Compared with other physical therapy methods, exercise training has merely no side-effects on patients with PF; thus, patients will have a higher tolerance to exercise training [72]. However, prospective evidence is still needed.

Quality of Life
Quality of life was evaluated with 6MWD, SGRQ and mMRC in our study. Our results showed that exercise training improved 6MWD performance in PF patients. We further explored the effects of exercise frequency, intensity, duration, age and BMI on 6MWD ( Table 3). The subgroups analysis showed that there were no differences in 6WMD between two BMI levels and two exercise duration groups, whereas there were differences among different exercise frequency, intensity, type and age groups. The differences between age groups can be supported indirectly by a recent study [4]. Our findings indicated that the effects of 6MWD are more obvious in moderate exercise intensity, combined exercise of four types and with younger patients; meanwhile, the effects were not affected by BMI level or exercise duration. In this study, the combined effect size of 6MWD was medium (Cohen's d = 0.77). Therefore, elderly patients with PF can derive benefits from exercise training on 6WMD.
In this study, exercise training reduced SGRQ performance in PF patients (Figure 7). The SGRQ is a disease-specific quality of life assessment tool validated for both chronic obstructive pulmonary disease (COPD) and PF [42][43][44]. There were 76 items in the questionnaire, including three parts to measure symptoms, activity restriction and the social and emotional impact of the disease. A higher score implies a poorer quality of life [45]. Compared with the control group, the operational group scored lower on the SGRQ. Due to high heterogeneity, based on the included information (exercise frequency, intensity, type, duration, age and BMI), we conducted subgroup analysis. The results showed that no SGRQ differences could be found in frequency, intensity, type, duration, age or BMI groups. Thus, the SGRQ was probably not affected by frequency, intensity, type, duration, age or BMI. The effect of BMI on the SGRQ is still not clear, because there was one study which did not include BMI, and there were a few studies on BMI greater than 25. Figure 8 shows a decreased mMRC score in the operational group compared with the control group. The mMRC scale is a self-rating tool to measure the degree of disability that breathlessness poses on daily activities [46,47]. The higher the score, the more severe the disability. Subgroup analysis showed that mMRC was not affected by exercise duration. Limited by the small pool of subjects, the findings were still inconclusive. In the future, the effects of exercise frequency, age and BMI on mMRC in PF patients need to be focused.
Overall, a large effect size suggests that exercise training reduces SGRQ performance, and medium effect size indicates that exercise training increase 6MWD performance. Improvement in the 6MWD equates with an improved quality of life in patients [73]. Additionally, a large effect size from this review indicated that exercise training had a positively impact on the mMRC. The PF patients were more breathless and tended to be less physically active [74,75]; consequently, their functional capacity and quality of life became worse [74][75][76]. Through exercise training during pulmonary rehabilitation, PF patients achieved an improvement in exercise ability and ventilation function, which alleviated dyspnea during sub-maximal exercises such as activities of daily living [52], the fact of which was demonstrated by a decrease in the mMRC after the exercise intervention. An active exercise training lifestyle can improve the quality of life by increasing feelings of vitality [77], well-being [78,79], and reduce the risk of cognitive decline and dementia [80][81][82][83]. Therefore, the comprehensive evidence in current study reveals that exercise training can improve the quality of life of patients with PF.

Advantages and Future Directions
In summary, this review has evaluated the effects of aerobic, resistance, flexible, and breathing exercise on cardiopulmonary endurance, pulmonary function, and quality of life in PF patients. However, compelling studies are still lacking in evaluating FVC% pred, TLC% pred and mMRC; therefore, more studies are needed in the future, especially on single interventions.
To the best of our knowledge, this meta-analysis has two advantages. Firstly, the data extraction was more reasonable and standard than other reviews due to comprehensive literature retrieval strategies. We searched for studies from five countries, three continents, and in two languages (English/Chinese), which further reduced regional bias and language bias. Secondly, we analyzed two methods of exercise training effects (Cohen's d, and mean difference) which evaluated clinical and statistical effects.
The main limitations of this study were that the disease severity, variability and progression of the PF patients included were varied, which may have affected the results. Another limitation was that the included exercise training regimens were combined; thus, a single exercise type could not be evaluated.

Conclusions
Exercise training during pulmonary rehabilitation can improve cardiopulmonary endurance and quality of life in elderly patients with PF. The 6MWD were more noticeable with moderate exercise intensity, combined aerobic-resistance-flexibility-breathing exercises and in younger patients, all which were not affected by BMI levels or exercise durations. Regarding pulmonary function, exercise training can improve FVC% pred, but has no effect on DLCO% pred and TLC% pred.  Institutional Review Board Statement: The study did not involve humans or animals.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.