This study prospectively included 133 consecutive outpatients with HFpEF and stable NYHA functional class II-III (figure 1). The study was conducted in a single third-level center in Spain. All patients provided informed consent, and the protocol was approved by the research ethics committee following the principles of the Declaration of Helsinki and national regulations.

Figure 1.

Central illustration. AF, atrial fibrillation; CIx, chronotropic index; CPET, cardiopulmonary exercise testing; HFpEF, heart failure with preserved ejected fraction; IRR, incidence rate ratio; SR, sinus rhythm.

(0.26MB).

Candidates were selected from the outpatient specialized HF unit. All patients met the following inclusion criteria: a) a previous history of symptomatic HF (New York Heart Association functional class ≥ II); b) normal left ventricular ejection fraction (ejection fraction> 0.50 by the Simpson method and end-diastolic diameter <60 mm); c) structural heart disease (left ventricular hypertrophy/left atrial enlargement) and/or diastolic dysfunction estimated by 2-dimensional echocardiography; and d) clinical stability, without hospital admissions in the past 3 months. Patients were excluded if they could not perform a valid baseline exercise test, had genetic or restrictive cardiomyopathies, had high suspicion of hypertrophic or amyloid cardiomyopathy, or showed any previous medical condition such as unstable angina, myocardial infarction, or cardiac surgery within the last 3 months; chronic metabolic, orthopedic, infectious disease, or pulmonary disease (including pulmonary arterial hypertension, chronic thromboembolic pulmonary disease, or moderate to severe chronic obstructive pulmonary disease); acute HF decompensation; and any other comorbidity with a life expectancy of less than 1 year.

Patients underwent maximal symptom-limited cardiopulmonary exercise testing (CPET), echocardiography, physician-perceived NYHA class, clinical examination, and laboratory tests.

The total number of WHF events was selected as the endpoint of interest. Additionally, all-cause mortality was also evaluated. WHF events inclunded hospitalizations, emergency room visits, and unplanned outpatient visits. The definition of WHF required worsening symptoms and signs of the disease and administration of parenteral diuretics. WHF events and survival status were identified from the clinical records of patients in the HF unit, hospital wards, emergency room, and electronic medical records. The endpoints were assessed by researchers blinded to the patients’ baseline characteristics, including CPET parameters. All patients included were follow-up until November 2021. The minimum duration of patient follow-up was 8.5 months.

Continuous and categorical variables are presented as mean±standard deviation, median [interquartile range (IQR)], or percentages. Bivariate negative binomial regression was used to assess the independent association between CIx as a continuous variable and the prognostic endpoints (WHF events and all-cause mortality). For the clinical endpoints, bivariate negative binomial regression evaluated the interaction between CIx and electrocardiographic rhythm (SR vs AF). Estimates are reported as incidence rate ratios (IRR). All variables listed in table 1 were evaluated for prognostic purposes. The selection of the covariates in the final multivariate models was based on biological plausibility and not only on the P value. The linearity assumption for continuous variables was simultaneously tested and transformed, if appropriate, with fractional polynomials. In addition to our exposures (CIx and the interaction AF*CIx), the covariates included in the final models for WHF events were: baseline NYHA functional class, past smoker, estimated glomerular filtration rate, N-terminal pro-brain natriuretic peptide, left ventricular ejection fraction, left ventricular end-systolic volume, left ventricular end-diastolic volume, left ventricular mass index, left atrial volume, treatment with beta-blockers, and treatment with furosemide. A 2-sided a P value <.05 was considered to be statistically significant. All analyses were performed using Stata 15.1.

Table 1 summarizes the baseline characteristics stratified by baseline electrocardiographic rhythm. Overall, patients with AF had a higher prevalence of stroke and data indicating more advanced disease (higher pulmonary artery systolic pressure, higher left atrial volume, lower systolic blood pressure at peak exercise, higher NT-proBNP levels, and lower estimated glomerular filtration rate), as shown in table 1. Likewise, patients with AF were more frequently treated with mineralocorticoid receptor antagonists and furosemide. However, there were no differences in baseline NYHA functional class, peak VO2, HR at rest, or CIx across the types of rhythm (SR vs AF).

At a median [IQR] follow-up of 2.4 [1.6-5.3] years, a total of 41 (30.8%) all-cause deaths and 146 WHF events (62 hospitalizations and 84 ambulatory episodes) in 58 patients were registered. The rates (per 10-person-years) of death and WHF events did not differ across the median of CIx (< 0.4 vs ≥ 0.4): 1.03 vs 0.75 (P=.535) and 4.44 vs 3.58 (P=.544), respectively. On multivariable analysis, CIx was not associated with WHF events (P=.573) and all-cause mortality (P=.319), as depicted in figure 2. In the same multivariate scenario, when dichotomized in the median (< 0.4 vs ≥ 0.4), CIx remained not independenlty associated with WHF events (IRR, 0.55; 95%CI, 0.23-1.32; P=.182) or death (IRR, 0.87; 95%CI, 0.50-1.56; P=.657).

Figure 2.

Association between chronotropic index and endpoints. AF, atrial fibrillation; CIx, chronotropic index; IRR, incidence rate ratio; SR, sinus rhythm.

(0.1MB).

Patients with CIx below the median (< 0.4) showed nonsignificant differences in mortality rates in SR (0.98 vs 0.45, P=.128) or AF (1.13 vs 1.28, P=.569). However, CIx <0.4 identified higher rates of WHF events in those patients with SR (4.93 vs 1.34, P=.003). In contrast, in patients with AF, CIx <0.4 showed a statistical trend to have lower rates of WHF events (2.66 vs 7.35, P=.068).

After multivariate adjustment, we confirmed a differential prognostic effect of CIx across electrocardiographic rhythms for predicting WHF events (P for interaction=.002). As depicted in figure 3, CIx was inversely and linearly associated with the risk of WHF events in patients in sinus rhythm (figure 3A). However, CIx was positively and linearly related to the risk of WHF events in those with AF (figure 3B). Similar results were found when we analyzed the differential prognostic effect of CIx only for HF hospitalizations (P for interaction=.007). Lower CIx was associated with a higher risk in those in SR (figure 4A), but we found the opposite in patients with AF (figure 4B).

Figure 3.

Differential prognostic effect of CIx across electrocardiographic rhythm for HF events. AF, atrial fibrillation; CIx, chronotropic index; IRR, incidence rate ratio; SR, sinus rhythm.

(0.17MB).

Figure 4.

Differential prognostic effect of CIx across electrocardiographic rhythm for HF hospitalizations. AF, atrial fibrillation; CIx, chronotropic index; IRR, incidence rate ratio; SR, sinus rhythm.

(0.17MB).

For all-cause mortality, the adjusted interaction between CIx and electrocardiographic rhythm was not significant (P for interaction=.529). CIx along the continuum was not associated with the risk of death in SR and AF (figure 5).

Figure 5.

Differential prognostic effect of CIx across electrocardiographic rhythm for all-cause mortality. AF, atrial fibrillation; CIx, chronotropic index; IRR, incidence rate ratio; SR, sinus rhythm.

(0.16MB).

Appropriate HR response to exercise is crucial for increasing cardiac output at maximal exercise in normal individuals9 and HF patients.10 CI is a common finding in patients with HF.1,6,11,12 In HFrEF, CI leads to exercise intolerance, impairs quality of life, and is associated with adverse events.4–6 Likewise, increased HR in HFrEF is also associated with adverse events13 and is a well-known therapeutic target.14

Regarding HFpEF, the optimal rest and exercise HR response remains elusive. Recent studies have revealed that blunted chronotropic response is a common finding associated with limited functional capacity.1,12 Despite the unclear pathophysiological mechanisms underlying CI in HFpEF, several potential mechanisms have been proposed for patients in SR: peripheral muscle dysfunction, autonomic nervous imbalance, sinus node remodeling causing a reduction in sinus node reserve, and impairment of cardiac beta-receptor responsiveness.7,11,15 Likewise, especially in patients with AF, the increased ventricular rate is a common precipitating factor for HF decompensations and a therapeutic target.14

It is well-known that the development of AF in patients with HF is associated with a worse prognosis16 and poor functional capacity.10,17 However, in HFrEF, previous studies showed that HR response to exercise in HFrEF has different patterns in patients with SR and AF.10,17 For example, in a cohort of 942 patients with HFrEF, Agostoni et al.17 reported that those with AF exhibited lower values of peak VO2 and O2 pulse but higher HR values at peak exercise than participants in SR. This finding suggested that stroke volume may be lower at peak exercise and that higher HR response acted as a compensatory mechanism for increasing stroke volume in those patients with AF.17 Thus, an increased HR response could translate into an augmented sympathetic drive triggered to maintain cardiac output.10

In HFpEF, some studies have shown that patients with AF exhibited lower values of peak VO2 and O2 pulse but no differences in peak HR compared with those in SR.18,19 Elshazly et al., 19 evaluated the CPET differences across rhythm status in 1744 young HFpEF patients (239 patients—13.7%—in AF) with a mean age of 51.2±15.4 years. They found that AF patients had lower peak VO2, O2 pulse, and systolic blood pressure at peak exercise, a higher risk of long-term total mortality, and no differences in peak HR compared with those with SR.19 Likewise, the current study did not find differences in CIx across SR vs AF. However, the present findings suggest that the type of rhythm largely influences the association between chronotropic response to exercise and WHF events. Depending on rhythm status, an exaggerated chronotropic response might identify patients with HFpEF and a higher risk of WHF. Indeed, a rapid ventricular response is a common precipitating factor of HF decompensation in patients with AF.20

Conversely, a blunted response might select those with SR at higher risk in which CI might play a crucial causative role in determining the inability to increase cardiac output during exercise. Current findings are another example of the complex and heterogeneous pathophysiology of HFpEF. This case highlights the differential role of HF response during exercise across the type of electrocardiographic rhythm.

Under the premise that our findings need further validation in future trials, we propose that the withdrawal or reduction of HR-lowering treatment, or even HR increase, in patients with HFpEF with documented CI may be a therapeutic strategy to reduce the risk of WHF events and improve functional capacity, especially in those with SR. Along this line, a recent small randomized clinical that enrolled 52 patients with stable HFpEF (80.8% on SR), previous treatment with beta-blockers (stable for at least 3 months prior to inclusion), and documented CI (CI <0.62) showed that short-term peak VO2 and the percentage of predicted peak VO2 increased by +2.1±1.29mL/kg/min (P <.001) and + 11.74±2.32% (P <.001) after beta-blocker withdrawal. Interestingly, mediation analysis showed that the main contributor to the improvement in maximal functional capacity was the magnitude of changes in HR response.2 In contrast, a tight HR control using lowering HR treatments in those patients with AF and exaggerated exercise HR response may be a valuable strategy for preventing further episodes of WHF.20–23

Due to the large number of uncertainties in the diagnosis and management of HFpEF, future studies in this field should aim to provide: a) a better understanding of the pathophysiology of chronotropic response both in patients with SR and AF; b) more precise phenotyping of HFpEF regarding HR response, evaluating the optimal range of HR in patients with HFpEF and whether it is modified by AF or SR; and c) definition of the clinical utility of beta-blockers or other HR-lowering treatment according to HFpEF phenotype,24–26 type of electrocardiogram rhythm, and HR response.

Finally, this study emphasizes the role of exercise tests in evaluating patients with HFpEF. CPET is a useful clinical tool for identifying different exercise HR response phenotypes in HFpEF.

We acknowledge that the main limitations of this study are the small sample size and the fact that this is an observational single-center study. Second, this is a selected population with high rates of beta-blocker prescription. Third, the current findings applied only to symptomatic patients with stable HFpEF. They cannot be extrapolated to other clinical scenarios, prevalent subgroups, or milder forms of the syndrome. Fourth, we did not register the longitudinal changes in the electrocardiographic type of rhythm or medical treatment during the follow-up. Finally, the low statistical power may explain some of the neutral findings.