Quantifying skeletal muscle energy metabolism during exercise in heart failure with preserved ejection fraction using in vivo  ³¹P magnetic resonance spectroscopy

Abstract

Background and Aims

Patients with heart failure and preserved ejection fraction (HFpEF) experience significant exercise intolerance, yet its underlying mechanisms remain poorly defined and are multifactorial. Iron deficiency (ID) occurs frequently in HFpEF and may contribute to exercise impairment. This study evaluated mitochondrial oxidative muscle metabolism in HFpEF using in vivo  ³¹phosphorus magnetic resonance spectroscopy (³¹P-MRS) employing two exercise protocols, and assessed whether ID influences exercise energetics.

Methods

In this parallel analysis of prospective studies at two sites, patients with HFpEF and control individuals performed either isometric exercise (isolated leg protocol) or dynamic exercise (cardiopulmonary protocol) with concomitant phosphocreatine recovery assessment using in vivo  ³¹P-MRS. Associations between clinical factors and oxidative metabolism were evaluated. ID was defined as ferritin <100 µg/L, or ferritin 100–299 µg/L with transferrin saturation <20%.

Results

Fifty-eight patients with HFpEF and 16 controls performed isometric exercise (n = 46 HFpEF; n = 16 control) or cardiopulmonary exercise (n = 12 HFpEF). Phosphocreatine recovery halftime after isometric exercise was prolonged in patients versus controls [27 (23–32) vs 24 (19–28) seconds, respectively; P = .03]. Phosphocreatine recovery halftime after dynamic exercise in patients was 39 (27–57) seconds. Both cohorts consisted of patients with and without ID (n = 19 and 27, and n = 6 and 6, respectively), who had comparable exercise and oxidative muscle capacity (all P > .42). High-sensitive C-reactive protein was associated with prolonged phosphocreatine recovery halftime (P =. 01).

Conclusions

Patients with HFpEF exhibit impaired whole-muscle oxidative capacity of skeletal muscle, as shown by two different ³¹P-MRS protocols with upper leg measurements, independent of ID status.

Study registration

NTR6605, NTR7297 (https://onderzoekmetmensen.nl/nl/trial/55673); NCT05750940 (https://clinicaltrials.gov/study/NCT05750940).

Introduction

Limitations in exercise capacity are common in heart failure with preserved ejection fraction (HFpEF), resulting in functional limitation and reduced quality of life.¹ Multiple pathophysiological changes may contribute to exercise intolerance in HFpEF, including impaired cardiovascular performance, structural and functional alterations in skeletal muscle and the (micro)vasculature.^2,3 In addition, multiple other comorbidities, such as pulmonary disease, sarcopenia, and iron deficiency (ID), could also contribute to exercise intolerance in HFpEF.^1,4

Emerging evidence suggests impediments in mitochondrial energy metabolism may also contribute to impaired oxygen utilization and, eventually, exercise intolerance in HFpEF. Prior studies have largely employed cardiopulmonary exercise test or 6-min walking test distance (6MWD), yielding a limited capacity to quantify the contribution of mitochondrial impediments. In vivo ³¹P-magnetic resonance spectroscopy (³¹P-MRS) permits non-invasive quantification of phosphocreatine (PCr) recovery kinetics as a direct probe of mitochondrial oxidative capacity,⁵ potentially identifying impaired muscle metabolism as underlying mechanism and target for exercise intolerance in HFpEF. However, different modes of exercise can be employed, including isometric muscle testing and dynamic cardiopulmonary exercise testing, capturing both muscle-specific local mitochondrial oxidative capacity and the integrated physiological response to whole-body exertion.⁵

Therefore, this study employed both isometric and dynamic exercise protocols using in vivo  ³¹phosphorus magnetic resonance spectroscopy (³¹P-MRS) to assess skeletal muscle oxidative function in HFpEF comprehensively. Moreover, we evaluated if ID contributed to impaired oxidative metabolism in HFpEF, and explored other clinical factors associated with oxidative metabolism impairments.

Methods

Detailed methods can be provided on request by corresponding author.

Study design

This parallel analysis of two prospective studies was derived from separate pre-registered prospective trials (NTR6605; NTR7297; NCT05750940), which were approved by the Medical Ethical Review Committee of the Maastricht University Medical Center (MUMC+) and the University Medical Center Groningen (UMCG). All patients provided written informed consent before enrolment. The study was conducted in accordance with the Declaration of Helsinki and the Medical Research Involving Human Subjects Act.

Study population

Patients diagnosed with HFpEF and controls without HF were included based on predefined in- and exclusion criteria (see study registrations). The main inclusion criteria entailed preserved left ventricular ejection fraction in combination with echocardiographic HFpEF characteristics and elevated circulating natriuretic peptide levels (NT-proBNP). The main exclusion criteria were significant peripheral artery disease or contraindications for MRI according to the safety protocols of the local hospitals, or iron supplements within 6 months before inclusion. Patient inclusion was stratified by iron status based on serum ferritin and transferrin saturation (TSAT) levels according to the ESC-HF guidelines into (i) no ID, or (ii) ID; including both absolute ID (ferritin <100 µg/L) and relative ID (ferritin 100–299 µg/L and TSAT <20%).

Study procedures

Two distinct exercise protocols with concomitant ³¹P-MRS analysis were employed for isometric and dynamic exercise, respectively. The patients included in the isometric exercise cohort underwent an isolated leg exercise protocol. During this protocol patients performed isometric contractions until 50% of PCr depletion was reached compared with resting levels. Subsequently, recovery of PCr was measured during post-exercise rest. In the dynamic exercise cohort patients underwent a supine cycling cardiopulmonary exercise protocol. In this protocol, patients performed an incremental workload protocol up to exhaustion. In each case, ³¹Phosphorus spectra were recorded continuously from the vastus lateralis muscle in the resting state, during exercise, and post-exercise recovery.

Outcome parameters

The primary endpoint of this analysis was the rate of PCr resynthesis following exercise (in seconds), assessed non-invasively using ³¹P-MRS and quantified as PCr halftime (Figure 1A), informing on the oxidative capacity of the vastus lateralis muscle.⁵

Figure 1.

(A) ³¹Phosphorus magnetic resonance spectroscopy parameters during and post-exercise using two different exercise set-ups. In the isolated leg muscle protocol (isometric exercise cohort), phosphocreatine recovery was measured after each patient reached a phosphocreatine depletion of 50%. In the cardiopulmonary exercise protocol (dynamic exercise cohort), phosphocreatine recovery was measured after patients cycled up to exhaustion. Pi, inorganic phosphate; PCr, phosphocreatine; ATP, adenosine triphosphate. (B and C) Impaired exercise capacity and mitochondrial energy metabolism parameters in heart failure with preserved ejection fraction and controls. Box-plots showing (B) impaired exercise capacity [397 (355–469) vs 514 (453–601) metres, P < .001] and (C) phosphocreatine recovery halftime in heart failure with preserved ejection fraction patients and controls within the isometric exercise cohort. Comparable results from the dynamic exercise cohort revealed impaired exercise capacity and phosphocreatine recovery halftime (data not shown). (D and E) Correlates of oxidative skeletal muscle metabolism in heart failure with preserved ejection fraction patients. Scatter plots of data from patients with heart failure with preserved ejection fraction to visualize the correlation between phosphocreatine recovery halftime and high-sensitive C-reactive protein (D) and exercise performance (E). Sex had a significant interaction with the correlation between phosphocreatine recovery halftime and high-sensitive C-reactive protein. Therefore, sex-stratified results are displayed (D). All data in these plots represents patients from the isometric exercise cohort. The authors thank Pablo I. Sanchez (UMCG) for supporting the design of the scanner figures. *P < .05

Results

Clinical characteristics

The study population consisted of the isometric exercise cohort (HFpEF n = 46, hypertensive controls n = 16) and the dynamic exercise cohort (HFpEF n = 12) (Table 1). Heart failure and preserved ejection fraction patients were older than controls (75 [69–79] vs 65 [58–74] years, P < .001), had more frequently atrial fibrillation (27 (59%) vs 2 (13%), P = .001), had higher NT-proBNP levels (400 [192–721] vs 61 [38–96] pg/ml, P < .001) and signs of diastolic dysfunction or elevated cardiac filling pressures on echocardiography (LA volume index 44 [33–60] vs 29 [23–33] ml/m², P < 0.001; E/e′ average 10.5 [7.9–13.5] vs 7.8 [6.7–11.0], P = .099), and more impaired exercise tolerance (Figure 1B). During rest, high-energy phosphate levels were normal in all subjects with HFpEF and controls, and values were in accordance with previous human ³¹P-MRS studies.⁵

Table 1.

Clinical and ³¹phosphorus magnetic resonance spectroscopy characteristics of heart failure with preserved ejection fraction patients and controls

	Isometric exercise cohort			Dynamic exercise cohort
	HfpEF (n = 46)	Controls (n = 16)	P-value	HfpEF (n = 12)
Female sex, n(%)	28 (61%)	10 (63%)	0.908	7 (58%)
Age (years)	75 [69–79]	65 [58–74]	<0.001	70 [62–76]
Body mass index (kg/m2)	28.1 [25.1–32.9]	28.4 [26.8–31.4]	0.618	28.2 [25.4–29.4]
Medical history, n (%)
Hypertension	40 (87%)	16 (100%)	0.325	10 (83%)
Atrial fibrillation	27 (59%)	2 (13%)	0.001	5 (42%)
Diabetes mellitus	5 (11%)	2 (13%)	0.999	3 (25%)
Laboratory values
eGFR CKD-EPI (ml/min/1.73 m)	62 [47–77]	69 [63–81]	0.044	72 [47–91]
NT-proBNP (pg/ml)	400 [192–721]	61 [38–96]	<0.001	818 [377–2065]
hsCRP (mg/L) (n = 30/16/11/2)	2.2 [1.0–3.4]	1.6 [0.5–3.2]	0.258	2.2 [1.0–3.5]
Echocardiography
LV ejection fraction (%)	61 [56–64]	59 [55–63]	0.600	53 [48–55]
LV mass index (gr/m2)	77 [66–87]	67 [57–86]	0.164	90 [79–99]
LA volume index (ml/m2)	44 [33–60]	29 [23–33]	<0.001	35 [30–42]
E/e′ average	10.5 [7.9–13.5]	7.8 [6.7–11.0]	0.099	9.8 [8.7–11.9]
31P-MRS
PCr recovery halftime (s)	27.4 [22.6–32.2]	24.2 [19.0–27.8]	0.030	39 [27–57]
PCr tau time constant (s)	19 [16–22]	17 [13–19]	0.030	26 [18–38]
PCr depletion (%)	30 [26–37]	27 [24–40]	0.711	—
End-exercise pH	7.1 [7.0–7.1]	7.1 [7.1–7.1]	0.359	6.8 [6.7–6.9]

eGFR CKD-EPI, estimated glomerular filtration rate according to chronic kidney disease epidemiology collaboration formula; HFpEF, heart failure with preserved ejection fraction; hsCRP, high-sensitive C-reactive protein; LA, left atrium; LV, left ventricle; NT-proBNP, N-terminal prohormone brain natriuretic peptide.

Skeletal muscle oxidative metabolism by isometric exercise

Acidosis at end-exercise was prevented in all subjects (pH >6.9). Heart failure and preserved ejection fraction patients vs controls had prolonged post-exercise PCr recovery halftime [27.4 (22.6–32.2) vs 24.2 (19.0–27.8) seconds, P = .030] (Figure 1C). Six (13%) HFpEF patients and 0 (0%) controls expressed a PCr recovery halftime >35 s (equal to τ_PCr of 50 s), reflecting impaired whole-muscle oxidative recovery.⁵

Skeletal muscle oxidative metabolism by dynamic exercise

Patients performed dynamic exercise up to exhaustion, and PCr recovery was measured using the tau time constant post-exercise and converted to PCr recovery halftime. Phosphocreatine recovery halftime was 39 (27–57) seconds in HFpEF patients.

Impact of iron deficiency and clinical factors on muscle oxidative metabolism in heart failure and preserved ejection fraction

After stratified inclusion for ID status, HFpEF patients had ID in 19 (41%) and 6 (50%) in the isometric and dynamic exercise cohorts. Heart failure and preserved ejection fraction patients with and without ID had comparable clinical characteristics, including exercise capacity, echocardiography, and also comparable PCr recovery halftime [isometric exercise cohort 28.6 (23.2-32.8) vs 27.3 (21.0-23.1) seconds, P = .496; dynamic exercise cohort 38 (27-95) vs 41 (24-50) P = .810]. Patients with only absolute ID compared with those without ID showed similar results in both cohorts (P = .734 and P = .454, respectively), so was ID based on TSAT <20%.

Linear regression (HFpEF patients n = 46 of isometric exercise cohort) showed no association between PCr recovery halftime with ID parameters. Phosphocreatine halftime was associated with higher hsCRP (R² = 0.202)—which appeared stronger in females-, and shorter 6MWD had an association trend (Figure 1D and E). Heart failure and preserved ejection fraction patients with impaired (n = 6) vs normal (n = 40) whole-muscle oxidative recovery had non-significant trends of worse exercise capacity (6MWD 345 ± 109 vs 420 ± 76 metres, P = .065) and higher hsCRP [3.40 (2.43, 4.55) vs 1.59 (1.00, 3.06) mg/L, P = .062].

Discussion

In this parallel analysis of in vivo  ³¹P-MRS exercise studies, we found significant deviations in oxidative skeletal muscle metabolism (prolonged PCr recovery halftime) after exercise in HFpEF compared with controls. Iron deficiency had no direct association with skeletal muscle oxidative metabolism in HFpEF. Rather, inflammation was associated with deviations in oxidative skeletal muscle function, and appeared primarily in patients with female sex, suggesting a possible sex-specific role for inflammation on skeletal muscle metabolism in HFpEF.

The combination of two different ³¹P-MRS exercise platforms facilitated the assessment of mitochondrial energy metabolism in HFpEF in one of the largest cohort so far on in vivo MRS-derived exercise data.^6,7 The prolonged PCr resynthesis time after isometric single-leg knee extension in this study aligns with previous studies evaluating ex- and in vivo mitochondrial oxidative capacity in HFpEF.⁶ Recently, Lewsey et al. reported worse muscle metabolism and exercise capacity in older compared with younger obese HFpEF patients, using an age cut-off of 65 years.⁷ During plantar flexion until exhaustion, they found with MRS mainly impaired muscle metabolism during exercise, while PCr recovery time was similar, possibly related to the type and termination of exercise. It may be argued that elevated cardiac filling pressures are the cornerstone of exercise intolerance in HFpEF patients, but acutely decreasing these filling pressures with nitroglycerine did not improve peak VO2.⁸ Taking all available data together, patients with HFpEF may have a more metabolic-oriented limitation in exercise tolerance and could benefit from targeting mitochondrial function in muscle cells.

Our finding that ID did not correlate with impaired oxidative muscle capacity was somewhat unexpected compared with previous studies.⁹ A prior mechanistic HFrEF trial with iron supplementation reported improved functional class and PCr recovery halftime primarily in anaemic patients with absolute ID, with less clear improvements in non-anaemic ID patients.¹⁰ Anaemic vs non-anaemic patients had higher CRP levels, which may have additionally contributed to the different effects.

Furthermore, markers of inflammation were correlated to oxidative muscle capacity during isolated leg exercise in HFpEF on a metabolic level in this study, and possibly more so in females. This observation aligns with the hypothesis of systemic inflammation and organ impairment in HFpEF.¹¹

Phosphocreatine recovery halftime appeared even more prolonged in patients after cardiopulmonary than isometric exercise, suggesting that dynamic rather than isometric exercise imposes greater systemic oxidative demands extending from peripheral adaptations.¹² This could hypothetically explain cardiovascular or pulmonary exercise limitations and aggravate (subtle) muscular mitochondrial impairments. Moreover, the two study protocols tested skeletal muscle metabolism in two different HFpEF populations with in general similar clinical baseline characteristics. Differences included a small average age difference and higher NT-proBNP values in the cardiopulmonary exercise cohort, possibly reflecting different disease stages and resultant mitochondrial functions. Because functional exercise is dependent on adequate upregulation and integration of multiple physiological and mitochondrial-related processes, combining dynamic and isometric exercise may be useful to pinpoint specific metabolic impairments per individual patient.¹³

Limitations

The observational and cross-sectional character of the study design should be taken into account, precluding causal conclusions. Moreover, the effect of adipose tissue, muscle mass, and collagen quantities and characteristics on mitochondrial function which could be relevant to exercise performance could not be assessed with the available acquisitions of the present study. Furthermore, the patient cohorts included in this study are heterogeneous and multimorbid; however, this is concordant with clinical HFpEF populations.¹⁴ Lastly, a mild age difference between HFpEF and control patients may have confounded results, as mitochondrial function declines with increasing age, but this decline is expected to be much lower than the difference we observed between groups.¹⁵

The results of this study encourage future research on targeting specific inflammation pathways that could restore mitochondrial function in HFpEF to benefit both peripheral and cardiac muscle function.

Conclusion

Patients with HFpEF exhibit impaired whole-muscle oxidative of skeletal muscle, as shown by two different ³¹P-MRS protocols with upper leg measurements, independent of ID status.

Author contributions

J.W.: data curation, formal analysis, investigation, visualization, writing—original draft, S.N.V.: data curation, formal analysis, investigation, writing—original draft, J.A.L.J.: data curation, investigation, methodology, resources, supervision, writing—review & editing, A.B.A.: data curation, investigation, writing—review & editing, J.M.: data curation, investigation, writing—review & editing, A.J.S.-K.: writing—review & editing, R.J.R.: writing—review & editing, B.L.M.S.: writing—review & editing, C.K.: writing—review & editing, A.J.H.M.H.: writing—review & editing, P.v.d.M.: writing—review & editing, V.B.S.-H.: methodology, resources, supervision, writing—review & editing, B.D.W.: funding acquisition, supervision, writing—review & editing, and V.P.M.v.E.: funding acquisition, supervision, writing—review & editing

Declarations

Declaration of Interest

J.W. reports grants from Corvia Medical, CSL Vifor, and Boehringer Ingelheim, outside the submitted work. C.K reports grant support from Pfizer, AstraZeneca, Philips, and TomTec imaging systems, consulting or speaker fees from Alnylam, Bayer, Pfizer, Alnylam, Philips/TomTec Imaging Systems, BMS, and Novartis, outside the submitted work. P.V.d.M. is supported by a grant from the European Research Council (ERC CoG 101045236, DISSECT-HF); received consultancy fees and/or grants from Novartis, Pharmacosmos, Vifor Pharma, Astra Zeneca, Pfizer, Pharma Nord, BridgeBio, Novo Nordisk, Daiichi Sankyo, Boehringer Ingelheim, and Ionis, all paid to the institution (UMCG). B.D.W. has received consultancy and speaker fees from Bayer, Boehringer Ingelheim, Novartis outside of the submitted work, all paid to the institution (UMCG). V.P.M.v.E. reports personal fees from Bayer, personal fees from Janssen, personal fees from Merck, grants and personal fees from Vifor, grants from Dutch Heart Foundation, outside the submitted work, all paid to the institution (UM). The other authors declare no other potential conflict of interest.

Data Availability

The data underlying this article will be shared on reasonable request to the corresponding author.

Funding

This study was supported by a research grant from Vifor Pharma Nederland BV. This work was supported by the Hartstichting (Dutch Heart Foundation) [grant numbers CVON2017-21-SHE PREDICTS HF and CVON2015-10-Early HFpEF to V.P.M.v.E. and Senior Dekker Clinical Scientist Grant to BDW (Grant Number 2019T064)]; the Partnership of UMCG–Siemens for building the future of Health (IPA 37 and IPA 39) to S.N.V. and B.D.W.; the Health Foundation Limburg to V.P.M.v.E.; and J.A.L.J. was supported in part by National Institutes of Health grant R01 HL173346. The funders had no influence on the study design, data collection, analyses, interpretation of the data, and writing of the report.