The Effect of Exercise Training, IMT and FES on Diastolic Function, Exercise Capacity and Quality of Life in Patients With Heart Failure With Preserved Ejection Fraction: a Systematic Review and Meta-analysis

Background: We aimed to better understand the effects of two exercise training interventions [endurance training and a combination of endurance and resistance training (combined exercise)] and two physiotherapies [functional electrical stimulation (FES) and inspiratory muscle training (IMT)] on diastolic function, exercise performance and quality of life (QoL) in heart failure with preserved ejection fraction (HFpEF) patients. Methods: Cochrane Library, EMBASE and MEDLINE via PubMed for randomized controlled trials were searched from their inception to May 2020. The methodological quality was assessed using the Physiotherapy Evidence Database sacle. All analyses were used by STATA. Results: A total of 14 articles on 13 trials were included in this meta-analysis with 673 HFpEF patients. The pooling revealed that peak oxygen uptake was improved by endurance training [MD (95% CI): 1.89 ml.kg -1 .min -1 (1.32, 2.46), P < 0.001], FES [MD (95% CI): 2.28 ml.kg -1 .min -1 (0.92, 3.65), P = 0.001] and IMT [MD (95% CI): 2.72 ml.kg -1 .min -1 (1.44, 3.99), P < 0.001]. Similar results were observed for 6-minute walk test and QoL. Besides, endurance training increased arterial venous oxygen difference [MD (95% CI): 1.41 (0.09, 2.74), P = 0.036]. Combined exercise was benecial to the ratio of peak early to late diastolic mitral inow velocities [MD (95% CI): -2.90 (-4.97, -0.83), P = 0.006] and the early diastolic mitral annual velocity [MD (95% CI): 1.40 (0.68, 2.12), P = 0.006]. IMT improved ventilation/carbon dioxide ratio slope [MD (95% CI): -3.36 ml.kg -1 .min -1 (-6.17, -0.54), P = 0.019]. on central diastolic function. Moreover, IMT improve pathophysiologic manifestations of HF via breathing against an external inspiratory load. Prior literature has reported improvement in exercise capacity, functional autonomy, diaphragmatic thickness and mobility, and cardiac autonomic control. Our result showed that IMT signicantly improved VE/VCO 2 slope, but no positive effect was found in other types of physical therapy. Various studies of HF patients have shown that selective respiratory muscle training improved submaximal and maximal exercise capacity and/or perceived dyspnea during daily living activities in HFrEF patients [17, 49-54]. Interestingly, the magnitude of outcome in HFpEF patients following IMT intervention was similar [17, 54] or even stronger [49, 51, 52] compared to other important trials aiming to evaluate the effect of IMT in HFrEF patients. The development of diaphragmatic fatigue is a delay, leading to a reduction in the recruitment of accessory respiratory muscles, and an improved ventilatory eciency 55].

Studies on the effect of physical therapy in HFpEF patients published until May 31, 2020 were searched by Cochrane Library, EMBASE and MEDLINE via PubMed. We used a mix of medical subject heading (MeSH) and keywords including exercise training, aerobic exercise, endurance training, combined training, inspiratory muscle training, functional electrical stimulation, and heart failure with preserved ejection fraction. A manual search of the list of references of all identi ed studies and review articles was performed for additional relevant studies.

Study selection
Inclusion criteria included the following: (1) published as randomized controlled trials (RCTs). (2) Patients (aged ≥ 18 years) with HFpEF were considered. (3) Patients (physical therapy group) undergoing endurance exercise, combined exercise, FES and IMT were considered. (4) Control group (non-intervention group) included those with a sedentary life style and those having some life-style education.
Articles that failed to meet the inclusion criteria were removed, including reviews, animal studies, non-RCTs, non-English language, those investigations with only healthy participants, and intervention durations with less than 4 weeks. Exercise was de ned to allow for inclusion of a broad range of structured physical activities and included endurance, resistance, combined training (endurance and resistance). Additionally, physiotherapies of FES and IMT were included in the de nition of exercise for the purpose of this meta-analysis.

Main outcomes
Primary outcomes of this study were exercise capacity measured by [peak VO 2 and 6-minute walk test (6MWT)], QoL measured by [Minnesota Living with Heart Failure Questionnaire (MLHFQ) total score]. Secondary outcomes were evaluated by diastolic function [measured by the ratio of peak early to late diastolic mitral in ow velocities (E/A), the ratio of early diastolic mitral in ow to annular velocities (E/e'), and the early diastolic mitral annular velocity (e')], ventilation/carbon dioxide ratio slope (VE/VCO 2 ) slope and arterial-venous oxygen difference (A-VO 2 Diff).

Data extraction
The following information was summarized: study characteristics (e.g. authors, year of publication and country), participant characteristics (e.g. age and sample size of different group), study methods/design, period of exercise intervention and outcomes. Full-text articles were retrieved for all title-abstracts that met the eligibility by two reviewers (Chenchen Zhuang and Xufei Luo), and a consensus reached by consulting a third reviewer (Qiongying Wang) if required.

Risk of Bias
We evaluated the risk of bias for inclusion in this meta-analysis by the Physiotherapy Evidence Database (PEDro) scale [22]. The methodological quality of each article was independently assessed by the two reviewers using a total scale, including eligibility criteria, point estimates and variability, between-group comparisons, intention-to-treat analysis, adequate follow-up, blinded assessors, blinded subjects, blinded therapists, baseline comparability, concealed allocation and random allocation. When a disagreement occurred, a third reviewer was consulted.

Data synthesis and analysis
For each outcome, the effect size in our study was assessed by the baseline to follow-up change. If the outcome was measured on the same scale, we used weighted mean difference (MD) and 95% con dence intervals (CI). When the I 2 statistic was lower than 30% and P < 0.10, a xed-effect model was used; otherwise, a random effects model was used. We assessed publication bias through Egger's regression test. Subgroup analysis was conducted to investigate the effect of different modalities of physical therapies on diastolic function, QoL and exercise performance. Furthermore, to account for differences in the functional capacity and QoL between intervention and control groups, we conducted meta-regression analysis for these outcomes (change in peak VO 2 and change in MLWHF total score). All analyses were based on intention-to-treat data. All analyses were used by STATA version 14.0 (Stata Corporation, College Station, TX).

Characteristics of studies
The basic characteristics of each study were summarized in Table 1. A total of 14 articles on 13 trials including 673 HFpEF patients were included in the present meta-analysis. One of all RCTs included not only FES intervention, but also IMT intervention [33]. Our original intention was to evaluate the effects of physiotherapy (FES and IMT) and exercise training (endurance training and combined exercise), respectively. The mean age of participants ranged from 60.5 to 75 years, and the proportion of women ranged from 35.6 to 100%. Primary outcomes of the study were exercise capacity (measured by peak VO 2 and 6MWT), QoL (measured by MLHFQ total score). Secondary outcomes that were also evaluated in the study included diastolic function (measured by E/A, E/e' and e'), VE/VCO 2 slope and A-VO 2 Diff.
The quality of included RCTs is presented in Table 2. In none of the studies was there objective evidence of imbalance in baseline characteristics between the intervention and control groups. The moderate risk of bias was due to inadequate blinding of participants and therapists, allocation concealment and intention-to-treat methodologies. Some may be contributed to insu cient reporting rather than inherent design weakness.

Diastolic function
Eight trials with 416 patients reported on E/A ( Fig. 2. A). The heterogeneity was small (I 2 = 0%, P = 0.606), using a xed-effect model. E/A was no change by

Quality of life
Thirteen trials with 560 patients reported on the MLHFQ total score (Fig. 3). The heterogeneity was small (I 2 = 28.5%, P = 0.158), using a xed-effect model.

Exercise performance
Our meta-analysis could be performed for two functional capacity indicators: 6MWT and peak VO 2 ; one exercise physiology parameter: VE/VCO 2 slope.

Discussion
The present meta-analysis summarized data from 14 on 13 RCTs that examined the effects of two exercise training interventions and two physiotherapies on diastolic function, exercise performance and QoL in HFpEF patients. FES and IMT improved exercise performance and QoL, and the outcomes were similar to endurance training, but the change of diastolic function on FES, IMT and endurance training was not found. Notably, combined exercise has a potential to improve diastolic function on HFpEF patients.
The previous studies have focused primarily on overall exercise training rather than the different types of exercise training. The meta-analysis of Gomes-Neto et al. [35] only conducted endurance training and failed to represent other exercise modality. They showed that exercise training improved peak VO 2 and QoL, but exercise training failed to improve diastolic function. The meta-analysis of Fukuta et al. [10] showed that exercise training improved exercise capacity in HFpEF patients. But a methodological limitation was the combination of results from Fukuta et al. that investigated endurance training and combined exercise. Besides, The previous meta-analyses held inconsistent opinions on diastolic function in HFpEF patients during exercise training [8][9][10]. Thus, the present meta-analysis compensated for the weakness of the previous studies.
The present meta-analysis observed that functional capacity indicators (peak VO 2 and 6MWT) and QoL were improved by FES and IMT, as well as exercise training. The positive result of QoL in endurance training or combined exercise is in accordance with the previous meta-analyses [8][9][10]. In addition, the evaluation of A-VO 2 Diff is only reported by endurance training, showing a positive result. The peak VO 2 may affect oxygen delivery and/or utilization via cardiac, vascular and skeletal muscle function [36][37][38][39]. Besides, the active skeletal muscle is the major reason to induce the augment of peak VO 2 in HFpEF patients, including oxidative enzyme activity, capillary density and mitochondrial density [29].
Our ndings showed that combined exercise improved E/e' and e', failed to improve E/A, whereas, endurance training, FES and IMT failed to improve diastolic function. Smart et al. [26] suggested that 16-week endurance training may not be su cient to elicit alteration in myocardial properties. Meanwhile, Fujimoto et al. [40] reported that 1-year endurance training had little effect on LV compliance and cardiovascular stiffness in HFpEF patients. Conversely, Edelmann et al. [23] reported that combined exercise improved diastolic function, as evidenced by a signi cant reduction in E/e' and e'. Diastolic function may be impacted by peripheral determinants (e.g. oxygen use by skeletal muscle, body mass index, etc), not by central determinants (e.g. E/A, E/E', etc) [41]. In response to exercise training, cardiac relaxation may be compounded by abnormalities in skeletal muscle oxygen use, which augments cardiac output, and flow into a small, stiff and slowly relaxing heart [42,43]. Thus, the different results of diastolic function between combined exercise and endurance training may be induced by resistance training, which adds muscle bulk to peripheral muscles and increases bone mass.
FES improves selected muscle in a coordinated manner and the aerobic oxidative capacity, which results in a functional pattern such as walking, cycling, or rowing [44][45][46][47][48]. FES seems to act primarily on peripheral vascular system, whereas its effects may be indirect on central diastolic function. Moreover, IMT improve pathophysiologic manifestations of HF via breathing against an external inspiratory load. Prior literature has reported improvement in exercise capacity, functional autonomy, diaphragmatic thickness and mobility, and cardiac autonomic control.
Our result showed that IMT signi cantly improved VE/VCO 2 slope, but no positive effect was found in other types of physical therapy. Various studies of HF patients have shown that selective respiratory muscle training improved submaximal and maximal exercise capacity and/or perceived dyspnea during daily living activities in HFrEF patients [17,[49][50][51][52][53][54]. Interestingly, the magnitude of outcome in HFpEF patients following IMT intervention was similar [17,54] or even stronger [49,51,52] compared to other important trials aiming to evaluate the effect of IMT in HFrEF patients. The development of diaphragmatic fatigue is a delay, leading to a reduction in the recruitment of accessory respiratory muscles, and an improved ventilatory e ciency [53,55].

Strengths and limitation
Although this study shows that FES and IMT may be effective for HFpEF, which may be an compensate rehabilitation to patients unable to experience exercise training. Moreover, combined exercise exhibits a positive effect on diastolic function All included studies were RCTs, and none were rated as poor in methodological quality according to the PEDro score, which suggests that the data presented are reliable. However, some included RCTs have small sample sizes, and according to the subgroup analysis of the different modalities of physical therapy on cardiac function, QoL and exercise performance. Second, there are only two studies reported combined exercise, however, they share one set of date using different results [23,24].
Further research is needed to pay more attention to the mechanism of hemodynamics and muscle function on HFpEF patients experiencing physical therapy. Besides, larger scale, good quality RCTs are needed for further investigating the effect of combined exercise, FES and IMT on patients with HFpEF.

Conclusions
In conclusion, our meta-analysis suggests that FES and IMT tend to a positive effect to HFpEF patients, the positive outcomes may be superior to exercise training. Besides, IMT tends to improve VE/VCO 2 slope. FES and IMT are surrogate rehabilitation for patients unable of exercise training.

Consent for publication
The present study was approved by the Ethics Committee Board of Lanzhou University Second Hospital (D2019-098).

Competing interests
No potential con ict of interest was reported by the authors.