Decoding the Pathophysiology of HFpEF With High-Resolution Phenotyping

Corresponding Author

The barriers to creating new treatments for heart failure with preserved ejection (HFpEF) have so far proven impenetrable. To prevail, the field may require a more rigorous understanding of HFpEF pathophysiology. This knowledge would serve interlocking objectives, from refining disease nosology to guiding discovery efforts. Knowledge of the cause(s) of symptoms and outcomes, coupled with the ability to measure them at scale, would enable the subtyping of patients by matching pathophysiology. Knowledge of these causal mechanisms would also enable rational therapy design, and, in a well selected population, improve the odds of a successful clinical trial.

A chief complaint among patients with HFpEF is exercise intolerance, a phenotype whose pathophysiology has been the subject of intense scrutiny. To this end, many investigators have applied a reductionist framework premised on defining exercise capacity as the rate of oxygen (O₂) consumption at peak exercise (Vo₂). Mechanisms of exercise intolerance are then taken to be defects in the individual steps of O₂ transport and consumption that move O₂ from the mouth to skeletal muscle mitochondria (the O₂ pathway). This breakdown of Vo₂ into its constituent O₂ pathway steps has the attractive feature that it can be performed quantitatively. Another attractive feature is it can be performed hierarchically, although this requires care. For example, the most common approach is to decompose Vo₂ in to 2 terms: ${\dot{V}}_{O_2}=Q\cdot \Delta A{V}_{O_2}$. The first term, cardiac output, reflects convective O₂ transport mediated by the heart. The second term, peripheral O₂ extraction, lumps all noncardiac O₂ pathway steps together, including alveolar ventilation, diffusive O₂ transport in the lungs, O₂ carried by hemoglobin, vascular redistribution of blood flow, diffusive O₂ transport to skeletal muscle, and mitochondrial respiration. The measurements needed for this 2-term analysis can be performed with familiar tools of clinical cardiology, including cardiopulmonary exercise testing, echocardiography, and arterial and venous catheters. Unfortunately, the appeal of this approach to decoupling the cardiac and noncardiac components of VO₂ is deceptive because the value of $\Delta A{V}_{O_2}$ depends on cardiac output as well, a consequence of the competition between convective and diffusive transport of O₂. To truly decouple the O₂ pathway into independent steps requires finer measurements and a more exacting analysis. Such an approach was adopted by Zamani et al. (1), to analyze Vo₂ in HFpEF in this issue of JACC: Basic to Translational Science.

Zamani et al. (1) sought to improve our understanding of the noncardiac causes of exercise intolerance in patients with HFpEF. Study participants performed 2 types of exercise: supine cycle ergometry with echocardiographic monitoring; and forearm exercise (isometric handgrip) with invasive monitoring. Cycle ergometry confirmed the widely recognized phenotype of reduced peak Vo₂ in HFpEF. Moreover, using echocardiographic estimates of peak cardiac output, the investigators calculated peak $\Delta A{V}_{O_2}$ and found that it too was reduced in HFpEF, as others have also demonstrated. The second exercise modality, isometric handgrip, was monitored with a catheter in the antecubital vein and a brachial artery flow probe. These tools permitted the measurement and estimation of O₂ transport properties in the local muscle bed, including blood flow, arterial and venous blood gases, and Vo₂. The investigators then used this high-resolution phenotyping to calculate the forearm muscle diffusion conductance for O₂ (D_M). Unlike prior efforts to characterize D_M in HFpEF, the investigators’ estimate did not rely on assumptions regarding local blood flow and local venous O₂ tension, because they measured these quantities directly. This forearm technique had the added advantage that it could be scalable to a broader population than is currently reachable with alternative techniques (e.g., the use of pulmonary artery catheters).

A key finding from Zamani et al. (1) was that forearm $\Delta A{V}_{O_2}$was reduced in patients with HFpEF relative to control subjects, but forearm D_M was not. This finding prompts at least 2 important questions: 1) if forearm D_M is similar between patients and control subjects, can it be extrapolated that locomotor muscle D_M might also be similar between groups, contrary to previous estimates? (2); and 2) what explains the drop in forearm $\Delta A{V}_{O_2}$in lieu of a defect in D_M?

Several observations likely preclude the extrapolation of the forearm findings to other muscle beds. First, at least 2 other studies used similar methodology to Zamani et al. (1) to estimate D_M in distinct muscle beds (3,4). Among control participants from all 3 studies, the values of D_M (normalized to muscle mass) in the forearm (1), whole arm (3), and single-knee extensors (4) differed widely. Similarly, muscle mass-normalized Vo₂ in these 3 muscle beds was also distinct. Though it could be argued that the control groups from each of these studies were not identical, data from within the Zamani et al. (1) study itself casts doubt on the likelihood that D_M is an invariant property across muscle beds. In particular, it was notable that in addition to D_M forearm Vo₂ (normalized to muscle mass) was also quite comparable between patients with HFpEF and control subjects. In other words, the drop in peak Vo₂ observed in patients during cycle ergometry was not recapitulated by the forearm muscles. This strongly suggests that the O₂ pathway determinants of Vo₂ in locomotor muscle differ from those of forearm muscle in patients with HFpEF. Finally, previous work has shown that modulation of arm Vo₂ and its determinants can be dissociated from whole-body Vo₂ (cycle ergometry). Boushel et al. (3) performed an exercise training study in which they found peak Vo₂ and D_M in the arm (arm exercise) increased after training, whereas peak whole-body Vo₂ remained unchanged.

If D_M differs between patients with HFpEF and control subjects in some muscle beds but not others, this fact may have important implications. It would argue against the existence of a circulating factor that uniformly compromises microcirculatory structure or function—and thus D_M—across all muscle beds. Rather, it is not difficult to imagine that differences in locomotor activity between patients and control subjects could explain differences in locomotor O₂ transport and consumption. Forearm activity might simply be more similar between groups, thereby explaining the similarities in Vo₂ and D_M. Another possibility is that patients with HFpEF are not less mobile per se, but their locomotor muscles adapt less to a given amount of mobility—they are less trainable.

To explain the drop in forearm $\Delta A{V}_{O_2}$in patients with HFpEF in the absence of a defect in forearm D_M, the investigators considered alternative O₂ pathway steps. First, they noted that forearm blood flow trended higher in patients with HFpEF. An isolated rise in blood flow, ceteris paribus, would be expected to cause a fall in $\Delta A{V}_{O_2}$, together with a sublinear rise in Vo₂. That Vo₂ was unchanged suggested the existence of additional O₂ pathway defects. The investigators entertained the possibility that anemia and or impaired mitochondrial function in HFpEF could contribute to the drop in forearm $\Delta A{V}_{O_2}$. To this end, they noted that patients with HFpEF were more obese than control subjects and the degree of adiposity was correlated with the drop in forearm $\Delta A{V}_{O_2}$. Furthermore, a strong epidemiological association between HFpEF and obesity has been previously recognized (5). Although it can be difficult to tease apart correlation from causation, the latter is made plausible by biological links between adiposity and anemia, muscle metabolism, and mitochondrial function.

The repeated failure of clinical trials in HFpEF strongly suggests that its pathophysiology remains insufficiently understood. The prototypical chronic symptom of this syndrome, exercise intolerance, is governed by mechanisms that are distributed over multiple organs, cell types, and subcellular systems. The resulting system properties of Vo₂ in turn give rise to tremendous mechanistic heterogeneity among patients with HFpEF (2). In the face of this complexity, progress will likely require comprehensive high-resolution phenotyping combined with quantitative causal analysis. Careful studies such as the work by Zamani et al. (1) will ultimately pave the way to improving disease nosology and discovering new therapies.