Heart failure with preserved ejection fraction (HFpEF) has been previously defined by various criteria,¹ with a common theme being the presence of preserved left ventricular ejection fraction. Defining a heart disease on the basis of a normal cardiac finding should beg the question: What is distinctly abnormal about the heart in HFpEF? Answering this question may aid in overcoming a track record of limited success to date with pharmacological approaches to HFpEF.

Article, see p 1664

This week’s study by Burrage et al² is timely in this context, because it investigated the association of cardiac energetics assessed noninvasively via phosphorus magnetic resonance spectroscopy (³¹P MRS) across a spectrum of HFpEF disease severity. A total of 43 participants underwent magnetic resonance spectroscopy at rest followed by supine exercise at a constant low workload of 20 W for 6 min with repeat whole-heart cine imaging at the end of exercise, including assessment of lung water using MR proton density mapping.

Principal findings demonstrate a gradient of myocardial energetic deficit across 4 clinical groups of participants: 11 age-matched controls, 9 with diabetes mellitus, 14 with clinical HFpEF, and 9 with amyloid cardiomyopathy with restrictive filling or high estimated filling pressures. Of note, the “pre-HFpEF” group with diabetes had phosphocreatine (PCr)/ATP ratios (1.71 [95% CI, 1.61–1.01]) similar to those of the HFpEF group (1.66 [95% CI, 1.44–1.89]), which raises the question of whether the energetic deficits arise early in the HFpEF disease spectrum or whether they may be related to nonhemodynamic metabolic factors. Interestingly, PCr/ATP was correlated with body mass index, but not glycated hemoglobin, across all individuals, raising the possibility that elevated body mass index may play a particularly important role in impairment of myocardial energetics. A particular strength of this work is the consistent finding that PCr/ATP at rest is associated with multiple complementary measures of cardiac performance, including NT-proBNP (N-terminal pro-B-type natriuretic peptide) levels, left ventricular E/e’, right ventricular function, and left ventricular filling rate during exercise. In addition, PCr/ATP levels in HFpEF are highly consistent with those reported in a previous study by Phan et al in which the PCr/ATP was 1.57±0.52 in HFpEF versus 2.14±0.63 in controls with related impaired time to peak filling in HFpEF.³ Taken together, these findings substantiate an important role of resting cardiac energetics in signaling abnormal ability of the heart to perform high-energy consuming processes of early diastolic relaxation and contractility augmentation.

Another important aspect of this work is the comprehensive magnetic resonance imaging–based assessment of functional reserve capacity during exercise, complementing previous multimodality studies that demonstrate the importance of exercise in unmasking abnormal cardiac performance that may not be evident at rest (Table). For example, although there was augmentation of diastolic filling rates and biventricular ejection fractions and stable atrial volumes with exercise in normal controls, HFpEF patients demonstrated blunted left ventricular ejection fraction and right ventricular ejection fraction augmentation, failure to augment diastolic filling rate, and biatrial dilation, illustrating the power of exercise to add phenotypic resolution by unmasking cardiac dysfunction in patients with HFpEF versus controls.

Table. Characteristic Cardiovascular Abnormalities in HFpEF Unmasked During Exercise and Their Clinical Relevance

E/e’ indicates the ratio between early mitral inflow velocity (E) and mitral annular early diastolic velocity (e’); HFA-PEFF, Heart Failure Association-PEFF with P representing pretest assessment, E representing echocardiography, F representing functional testing, and F representing final aetiology; HFpEF, heart failure with preserved ejection fraction; PCWP pulmonary capillary wedge pressure; V_E/Vco₂, ratio of minute ventilation to carbon dioxide elimination characterized either as a slope or a nadir of the ratio during exercise; and Vo₂, oxygen uptake.

The authors emphasize quantification of lung water by proton density imaging that correlated with severity of reduction in cardiac energetics (PCr/ATP). Pulmonary edema is a well-recognized finding in acute decompensated heart failure with flooding of alveoli that results from an imbalance between fluid filtration and removal in the setting of high left-sided hydrostatic pressure. Initially, fluid accumulates in the more compliant interstitial compartment surrounding bronchioles and arterioles before invading the interstitial region of the alveolar–capillary septum that can lead to impairment in gas exchange with resultant hypoxemia. However, exercise is quite distinct from acute decompensated heart failure. Blood transfers from the splanchnic to the pulmonary vasculature, promoting increased capillary volume that may contribute to “greater lung water.” Yet an important distinction is that exercise in left-sided heart failure alone does not lead to hypoxemia characteristic of frank pulmonary edema.⁴ Thus, in this study, the physiologic significance of a 4.4% increase in proton density/lung water observed in HFpEF and a 6.2% increase in patients with cardiac amyloidosis during exercise is unclear because systemic oxygen saturations were not reported and exact localization of the increased lung water is not known. Furthermore, we do not know whether maximum exercise capacity was at all limited by the development of lung water. However, the study was conducted in the supine position, and orthopnea with increased venous return is known to be experienced by HFpEF patients when they are lying down. Whether lung water would have been evident during exercise in the more physiologically relevant upright position remains an open question. Nonetheless, the overall constellation of findings indicative of maladaptive cardiac responses to exercise do track with reduced PCr/ATP ratios and increased exercise lung water, which should prompt further investigation of the functional significance of exercise-induced increased pulmonary proton density.

The findings by Barrage et al should be put in the context of other efforts to characterize myocardial and extracardiac energetics in HFpEF. To date, noninvasive assessment of in vivo myocardial energetics in heart failure have largely been characterized by either ¹¹C acetate positron emission tomography⁵ or ³¹P MRS.³¹¹C acetate positron emission tomography has been validated in preclinical models to reveal the myocardial oxidative energetics by injecting a ¹¹C acetate tracer, which is taken up by the heart and then rapidly converted to acetyl coenzyme A and metabolized to carbon dioxide through the tricarboxylic acid cycle with oxidative phosphorylation. The largest HFpEF patient cohort study characterizing myocardial energetics with ¹¹C acetate positron emission tomography (N=19 HFpEF patients) revealed mechanical, energetic, and flow-reserve dysfunction unmasked during dobutamine stress in HFpEF compared with controls.⁶ In contrast, the current study used ³¹P MRS to study the endogenous presence of the PCr/ATP ratio at rest as a measure of myocardial energetics. ³¹P MRS uses the unique signature of the phosphorous nuclear magnetic resonance spectra to identify the off-resonance shift of ³¹P nuclei in PCr and ATP molecules. The key advantages of ³¹P MRS in comparison with ¹¹C acetate positron emission tomography include no ionizing radiation, ease of repeated measurements, and direct measurement of high energy phosphate energetics. However, because of the inherent insensitivity of magnetic resonance in typically polarizing only a few nuclei per million molecules and the low endogenous concentrations of ³¹P nuclei (mmol/L), myocardial ³¹P MRS results in a low signal-to-noise ratio, which commonly limits measurements to a single 1-cm³ voxel on the septum. Furthermore, the requirement for special magnetic resonance imaging hardware (both coil and scanner modifications), sensitivity to B₀ inhomogeneity, limited nuclear magnetic resonance spectral resolution at clinical field strengths, and long scan times that confine ³¹P MRS to a state of rest may limit its widespread clinical use. Thus, we applaud the authors in executing such a technically demanding study, while also providing a unique insight into the HFpEF energetic phenotype.

In the future, assessment of myocardial energetics may overcome current technical limitations of magnetic resonance spectroscopy with newer technologies including a horizon imaging technology called hyperpolarized magnetic resonance spectroscopy that can boost the signal-to-noise ratio of magnetic resonance spectroscopy by 10 000 with the added advantage of characterizing metabolic flux through in vivo pyruvate metabolism.⁷ In looking toward the future of ³¹P MRS imaging studies in HFpEF it should be noted that ³¹P MRS application to skeletal muscle has shown promising characterization of the dysfunctional energetics in the HFpEF cohort under both rest and exercise stress conditions.⁸ An intriguing question may be to what extent more accessible characterization of skeletal muscle ³¹P MRS may mirror cardiac energetics in HFpEF, particularly given the systemic nature of the disease. This is particularly interesting when considering that exercise training, a rare example of an effective intervention that improves functional capacity in HFpEF, confers salutary effects primarily through peripheral skeletal muscle adaptations.

None.

Disclosures Dr Ho has received research grants from the National Institutes of Health and Bayer AG, and research supplies from EcoNugenics. Dr Lewis has received research funding from the National Institutes of Health (R01-HL 151841, R01-HL131029, and U01-HL 160278) and American Heart Association (15GPSGC-24800006), as well as from Amgen, Cytokinetics, Applied Therapeutics, AstraZeneca, and Sonivie, in relation to projects that are distinct from this work; honoraria from Pfizer, Merck, Boehringer-Ingelheim, Novartis, American Regent, Cyclerion, Cytokinetics, and Amgen for advisory boards outside of the current study; and receives royalties from UpToDate for scientific content authorship related to exercise physiology. Dr Nguyen reports no conflicts.

The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association.

For Sources of Funding and Disclosures, see page 1681.

Circulation is available at www.ahajournals.org/journal/circ

射血分数保留的心力衰竭 (HFpEF) 先前已由各种标准定义，¹一个共同的主题是存在保留的左心室射血分数。根据正常的心脏发现定义心脏病应该引出一个问题：HFpEF 中的心脏有什么明显异常？回答这个问题可能有助于克服迄今为止对 HFpEF 的药理学方法取得有限成功的记录。

文章，见第 1664 页

^{Burrage 等人2}本周的研究在此背景下是及时的，因为它调查了通过磷磁共振波谱 ( ³¹ P MRS) 无创评估的心脏能量学在 HFpEF 疾病严重程度范围内的关联。共有 43 名参与者在休息时接受了磁共振波谱分析，然后在 20 W 的恒定低工作量下进行 6 分钟的仰卧运动，并在运动结束时重复全心电影成像，包括使用 MR 质子密度映射评估肺水。

主要研究结果表明 4 组临床参与者的心肌能量缺乏梯度：11 名年龄匹配的对照组、9 名糖尿病患者、14 名临床 HFpEF 和 9 名淀粉样心肌病伴限制性充盈或高估计充盈压。值得注意的是，患有糖尿病的“pre-HFpEF”组的磷酸肌酸 (PCr)/ATP 比率 (1.71 [95% CI, 1.61–1.01]) 与 HFpEF 组 (1.66 [95% CI, 1.44–1.89]) 相似)，这提出了能量不足是否在 HFpEF 疾病谱早期出现或者它们是否可能与非血流动力学代谢因素有关的问题。有趣的是，在所有个体中，PCr/ATP 与体重指数相关，但与糖化血红蛋白无关，增加了体重指数升高可能在心肌能量损害中发挥特别重要作用的可能性。这项工作的一个特别优势是一致的发现，即静息时的 PCr/ATP 与心脏功能的多种互补测量相关，包括 NT-proBNP（N 末端前 B 型利钠肽）水平、左心室 E/e' 、右心室功能和运动期间的左心室充盈率。此外，HFpEF 中的 PCr/ATP 水平与 Phan 等人先前研究中报告的水平高度一致，其中 HFpEF 中的 PCr/ATP 为 1.57±0.52，而对照组为 2.14±0.63，与 HFpEF 的峰值填充时间相关受损. 这项工作的一个特别优势是一致的发现，即静息时的 PCr/ATP 与心脏功能的多种互补测量相关，包括 NT-proBNP（N 末端前 B 型利钠肽）水平、左心室 E/e' 、右心室功能和运动期间的左心室充盈率。此外，HFpEF 中的 PCr/ATP 水平与 Phan 等人先前研究中报告的水平高度一致，其中 HFpEF 中的 PCr/ATP 为 1.57±0.52，而对照组为 2.14±0.63，与 HFpEF 的峰值填充时间相关受损. 这项工作的一个特别优势是一致的发现，即静息时的 PCr/ATP 与心脏功能的多种互补测量相关，包括 NT-proBNP（N 末端前 B 型利钠肽）水平、左心室 E/e' 、右心室功能和运动期间的左心室充盈率。此外，HFpEF 中的 PCr/ATP 水平与 Phan 等人先前研究中报告的水平高度一致，其中 HFpEF 中的 PCr/ATP 为 1.57±0.52，而对照组为 2.14±0.63，与 HFpEF 的峰值填充时间相关受损.³总而言之，这些发现证实了静息心脏能量学在发出心脏异常能力以执行早期舒张期舒张和收缩力增强的高能量消耗过程的信号中的重要作用。

这项工作的另一个重要方面是对运动期间功能储备能力的综合磁共振成像评估，补充了先前的多模态研究，这些研究证明了运动在揭示在休息时可能不明显的异常心脏表现的重要性（表）。例如，虽然正常对照组的舒张期充盈率和双心室射血分数增加，心房容积稳定，但 HFpEF 患者左心室射血分数和右心室射血分数增加迟钝，舒张期充盈率未能增加，双心房扩张，通过揭示 HFpEF 患者与对照组的心功能不全，说明运动的力量以增加表型分辨率。

桌子。运动过程中发现的 HFpEF 的特征性心血管异常及其临床相关性

E/e'表示二尖瓣早期流入速度（E）与二尖瓣环舒张早期速度（e'）的比值；HFA-PEFF，心力衰竭协会-PEFF，P 代表测试前评估，E 代表超声心动图，F 代表功能测试，F 代表最终病因；HFpEF，射血分数保留的心力衰竭；PCWP 肺毛细血管楔压；V _E /Vco ₂，每分钟通气量与二氧化碳消除的比率，以运动期间比率的斜率或最低点为特征；Vo ₂，摄氧量。

作者强调通过质子密度成像对肺水进行量化，这与心脏能量减少 (PCr/ATP) 的严重程度相关。肺水肿是急性失代偿性心力衰竭中公认的发现，伴有肺泡充盈，这是由于在高左侧静水压力的情况下液体过滤和清除之间的不平衡造成的。最初，液体在细支气管和小动脉周围更顺应的间质隔室积聚，然后侵入肺泡-毛细血管间隔的间质区域，这可能导致气体交换受损并导致低氧血症。然而，运动与急性失代偿性心力衰竭截然不同。血液从内脏转移到肺血管，促进毛细血管体积的增加，这可能有助于“更多的肺水”。然而，一个重要的区别是，仅在左侧心力衰竭中进行运动不会导致明显肺水肿的低氧血症特征。⁴因此，在这项研究中，在 HFpEF 中观察到的 4.4% 的质子密度/肺水增加以及在运动期间心脏淀粉样变性患者增加 6.2% 的生理意义尚不清楚，因为没有报告全身氧饱和度以及增加的确切定位。肺水不详。此外，我们不知道最大运动能力是否完全受到肺水发展的限制。然而，该研究是在仰卧位进行的，已知 HFpEF 患者在躺下时会出现静脉回流增加的端坐呼吸。在生理上更相关的直立位置运动期间肺水是否明显仍然是一个悬而未决的问题。尽管如此，

Barrage 等人的研究结果应放在其他努力描述 HFpEF 中心肌和心外能量学特征的背景下。迄今为止，心力衰竭体内心肌能量学的无创评估主要通过¹¹ C 醋酸盐正电子发射断层扫描⁵或³¹ P MRS 来表征。^{3 11} C 醋酸盐正电子发射断层扫描已在临床前模型中得到验证，可通过注射^{11来揭示心肌氧化能量学。}C醋酸示踪剂，被心脏吸收，然后迅速转化为乙酰辅酶A，并通过三羧酸循环氧化磷酸化代谢为二氧化碳。^用11 C醋酸正电子发射断层扫描（N=19 HFpEF 患者）表征心肌能量学的最大 HFpEF 患者队列研究显示，与对照组相比，HFpEF 在多巴酚丁胺应激期间暴露出机械、能量和流量储备功能障碍。⁶相比之下，目前的研究使用³¹ P MRS 来研究静息时 PCr/ATP 比率的内源性存在，以作为心肌能量学的量度。³¹P MRS 使用磷核磁共振光谱的独特特征来识别PCr 和 ATP 分子中^{31 P 核的非共振偏移。与11 C 醋酸盐正电子发射断层扫描相比， 31} P MRS的主要优势包括无电离辐射、易于重复测量以及直接测量高能磷酸盐能量学。然而，由于磁共振固有的不敏感性，通常每百万分子仅极化几个核，并且³¹ P 核 (mmol/L) 的低内源性浓度，心肌³¹ P MRS 导致低信噪比，这通常将测量限制在单个 1-cm ³隔膜上的体素。此外，对特殊磁共振成像硬件（线圈和扫描仪修改）的要求、对 B ₀不均匀性的敏感性、临床场强下的有限核磁共振光谱分辨率以及将³¹ P MRS 限制在静止状态的长扫描时间可能限制了其广泛的临床应用。因此，我们赞赏作者执行这样一项技术要求很高的研究，同时也提供了对 HFpEF 能量表型的独特见解。

未来，心肌能量学评估可能会通过更新的技术克服当前磁共振波谱的技术限制，包括称为超极化磁共振波谱的水平成像技术，该技术可以将磁共振波谱的信噪比提高 10000通过体内丙酮酸代谢表征代谢通量的优势。⁷在展望 HFpEF 中³¹ P MRS 成像研究的未来时，应该注意的是，³¹ P MRS 应用于骨骼肌已显示出在休息和运动压力条件下 HFpEF 队列中功能失调的能量学的有希望的特征。⁸一个有趣的问题可能是骨骼肌³¹ P MRS 的更容易获得的表征在多大程度上可以反映 HFpEF 中的心脏能量学，特别是考虑到该疾病的系统性。考虑到运动训练是一种罕见的有效干预措施，可提高 HFpEF 的功能能力，这一点尤其有趣，它主要通过外周骨骼肌适应来产生有益的影响。

没有任何。

披露何博士已获得美国国立卫生研究院和拜耳公司的研究资助，以及 EcoNugenics 的研究用品。Lewis 博士获得了美国国立卫生研究院 (R01-HL 151841、R01-HL131029 和 U01-HL 160278) 和美国心脏协会 (15GPSGC-24800006) 以及 Amgen、Cytokinetics、Applied Therapeutics、AstraZeneca 的研究资助和Sonivie，与与本作品不同的项目有关；辉瑞（Pfizer）、默克（Merck）、勃林格殷格翰（Boehringer-Ingelheim）、诺华（Novartis）、美国丽晶（American Regent）、Cyclerion、Cytokinetics 和安进（Amgen）为当前研究之外的顾问委员会提供的酬金；并从 UpToDate 获得与运动生理学相关的科学内容作者的版税。Nguyen 博士报告没有冲突。

本文所表达的观点不一定是编辑或美国心脏协会的观点。

有关资金来源和披露信息，请参见第 1681 页。

流通可在 www.ahajournals.org/journal/circ