Prognostic impact of chronotropic incompetence in transthyretin cardiac amyloidosis: a multicentre study

Abstract

Introduction

Chronotropic incompetence (CI) is a frequent but underappreciated feature of cardiac amyloidosis and may contribute to exercise intolerance. However, its prognostic significance remains incompletely defined. We investigated the prevalence, functional correlates, and prognostic value of chronotropic incompetence in patients with transthyretin amyloid cardiomyopathy.

Methods

In this multicentre retrospective study, 212 stable outpatients with transthyretin amyloid cardiomyopathy, in sinus rhythm and naïve to disease-specific therapy, underwent maximal cardiopulmonary exercise testing. Chronotropic response was assessed as peak heart rate expressed as a percentage of the age-predicted value (pHR%). Chronotropic incompetence prevalence was evaluated using clinically relevant thresholds. Prognostic performance for 1-year cardiovascular mortality was assessed using Cox regression, receiver operating characteristic analysis, and Kaplan–Meier estimates.

Results

Chronotropic incompetence defined by a pHR% ≤75% was present in 35% of patients and was associated with markedly impaired functional capacity, including lower peak oxygen uptake and reduced ventilatory efficiency. During 1-year follow-up, 10 cardiovascular deaths occurred. Among exercise-derived variables, pHR% demonstrated strong prognostic value, with each 1% increase associated with a 5.5% relative reduction in cardiovascular mortality risk (hazard ratio 0.945; P = .011). A pHR% threshold of 75% provided optimal discrimination (area under the curve 0.71) and identified a subgroup with significantly lower survival (log-rank P < .05).

Conclusion

A blunted chronotropic response is common in transthyretin amyloid cardiomyopathy and conveys adverse short-term prognosis. A pHR% ≤75% represents a clinically meaningful and easily obtainable threshold for functional and prognostic stratification, offering a pragmatic alternative when comprehensive cardiopulmonary exercise testing assessment is not available.

Graphical Abstract

Graphical Abstract.

In 212 consecutive transthyretin cardiac amyloidosis (ATTR-CM) outpatients undergoing CPET, chronotropic incompetence assessed by percentage of predicted peak heart rate was associated with lower exercise capacity and identified patients at higher risk of cardiovascular mortality. A pHR threshold ≤75% showed prognostic relevance and good discriminatory performance.

Introduction

Amyloidoses are a complex spectrum of diseases in which the accumulation of misfolded proteins within multiple tissues produces highly variable morphological, functional, and clinical profiles.¹ Cardiac involvement typically begins with impaired diastolic function, leading to elevated left ventricular filling pressures² and clinical manifestations consistent with heart failure with preserved ejection fraction (HFpEF). As the disease progresses, the restrictive physiology is accompanied by a global reduction in systolic function and cardiac output, both of which are critical predictors of worsening heart failure and adverse prognosis.^3,4 Nevertheless, increasing disease awareness and earlier diagnostic pathways have shifted the clinical presentation⁵, and for much of the disease trajectory, exertional dyspnoea, and reduced functional capacity remain the most prominent manifestations of amyloid cardiomyopathy. Consequently, growing evidence supports cardiopulmonary exercise testing (CPET) as the most comprehensive non-invasive method for functional assessment⁶ and to target pharmacological disease-modifying interventions.⁷ When rigorously quantified by CPET, patients with cardiac amyloidosis frequently exhibit moderate-to-severe functional limitation, reflected by reduced peak oxygen uptake (pVO₂).⁸ Exercise impairment is partly driven by postcapillary pulmonary hypertension arising from left-sided diastolic dysfunction, which affects ventilatory efficiency.⁹ This abnormal response is further amplified by an exaggerated skeletal-muscle metaboreflex.¹⁰ Both mechanisms manifest as an elevated ventilation to carbon dioxide production relationship (VE/VCO₂) slope, a variable that holds independent prognostic value in amyloid cardiomyopathy, in addition to pVO₂^11,12 and circulatory power.¹³ However, beyond and beneath these metabolic metrics, chronotropic incompetence (CI) has emerged as an additional contributor to functional limitation.¹⁴ Indeed, in restrictive cardiomyopathies such as amyloid cardiomyopathy, heart rate (HR) plays a crucial role in maintaining cardiac output during exercise, given the structural constraints that limit preload reserve and stroke-volume augmentation.^4,6 The reported prevalence of CI in cardiac amyloidosis varies according to the definition and cut-off adopted,¹⁵ showing a prevalence higher than 40% in a multicentre amyloid cardiomyopathy cohort¹⁶ as well as a significant relationship with functional status.

Therefore, in a relatively large multicentre cohort of stable outpatients, we sought to investigate: (i) the prevalence of CI in a homogeneous cohort of patients with transthyretin amyloidosis (ATTR) and cardiac involvement; (ii) the impact of CI presence on their functional profile; (iii) a possible prognostic relevance of CI in terms of cardiovascular death at 1-year follow-up.

Methods

Study sample

We analysed data of 277 consecutive stable outpatients from three expert centres (Centro Cardiologico Monzino, Milan, Italy; Fondazione Toscana Gabriele Monasterio, Pisa, Italy; Medical University of Vienna, Austria) diagnosed between 2019 and 2023 with wild-type or variant ATTR, in accordance with the current guidelines.¹⁷ Primary inclusion criteria were stable clinical conditions (no admissions for heart failure or invasive procedure in the last 6 months and unchanged medications for at least 3 months), availability of a maximal symptom-limited CPET performed on a cycle ergometer, absence of comorbidities, or non-cardiac amyloid-related organ involvement which directly interfere with exercise performance such as moderate-to-severe anaemia (haemoglobin levels <100 g/L), severe obstructive/restrictive lung disease, significant peripheral vascular disease, exercise-induced angina, and/or ST changes as well as moderate-to-severe polyneuropathy or orthopaedic disabilities. Furthermore, to avoid confounders regarding the HR kinetics,¹⁸ patients with atrial fibrillation (AF) at the time of CPET, those with second- or higher-degree atrioventricular block as well as with a pacemaker-dependent rhythm were excluded. When present at study run-in, we converted β-blocker dosage to equivalent doses of bisoprolol (i.e. the daily dosage in patients taking carvedilol was divided by 5, the dose of metoprolol was divided by 20 while the nebivolol dose was left unchanged) to avoid confounding.¹⁹ Noteworthy, all patients enrolled were still naïve to disease-specific therapy and were not enrolled in clinical trials on disease-modifying agents. Indeed, in our centres, consistent with their clinical condition, all the referred patients with suspected amyloid cardiomyopathy received a baseline comprehensive clinical and instrumental evaluation, including a functional one by means of a maximal CPET.

Ethical approval has been obtained from the appropriate local ethics committee or Institutional Review Board and informed consent has been obtained for each patient. The study has been performed in accordance with the principles stated in the Declaration of Helsinki. Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.

Cardiopulmonary exercise testing

A maximal, symptom-limited CPET was performed on an electronically braked cycle ergometer connected to a metabolic chart. A personalized ramp protocol was chosen, aiming at a test duration of 10 ± 2 min. The exercise was preceded by a 2-min resting phase with breath-by-breath gas exchange monitoring followed by a 3-min unloaded warm-up. CPET was self-terminated by the patient when referring to maximal effort and as confirmed by a peak respiratory exchange ratio (RER) ≥1.05. A breath-by-breath analysis of O₂, carbon dioxide (CO₂), and ventilation (VE) was performed, and peak values were computed as the highest observed measurements (20 s average). The predicted pVO₂ was determined by using the sex, age, and weight-adjusted Hansen–Wasserman equations. The anaerobic threshold (AT) was identified through a V-slope analysis of VO₂ and CO₂ production (VCO₂) and was confirmed through the specific behaviour of the ventilatory equivalents of O₂ (VE/VO₂) and CO₂ (VE/VCO₂), as well as through the end-tidal pressure of O₂ and CO₂. A 12-lead electrocardiogram, blood pressure, and HR were recorded. Specifically, peak HR (pHR) was collected during CPETs, whereas rest HR was measured after at least 2 min of rest in a seated position on the cycle ergometer. Peak HR was also analysed as a percentage of the maximum predicted value according to the standard formula (pHR% = (pHR/220—age) × 100).¹⁶

Statistical analysis

Continuous variables were summarized as mean ± standard deviation (SD) when normally distributed and as median with interquartile range (IQR, 25th–75th percentile) when skewed. Categorical variables are presented as counts and percentages. Group comparisons were performed using the χ² test (or Fisher’s exact test when required) for categorical variables, the independent two-sample t-test for normally distributed continuous variables, and the Wilcoxon rank-sum test for non-normally distributed variables. Because patients were enrolled across three centres, the dataset exhibited a multilevel structure. To account for centre-related heterogeneity, descriptive analyses were complemented by random-effects regression models incorporating centre-specific Gaussian intercepts. The prespecified primary end-point was cardiovascular (CV) mortality at 1 year. Survival times were explored in the overall cohort, univariate and multivariate Cox proportional hazards models were fitted to examine the association between variables and the end-point. Results are expressed as hazard ratios (HRs) with 95% confidence intervals (CIs). The discriminatory performance of percentage of predicted pHR% for identifying patients at risk for 1-year CV death was assessed using receiver operating characteristic (ROC) curves. The area under the curve (AUC) with 95% CI was calculated, and optimal thresholds were determined using Youden’s index. Diagnostic performance at clinically relevant pHR% cut-off values was expressed as sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). Kaplan–Meier curves were constructed to estimate 1-year CV survival according to the optimal pHR% threshold, with comparisons based on the log-rank test. All analyses were performed using R software (R Foundation for Statistical Computing, Vienna, Austria). All tests were two-sided, and a P value ≤ .05 was considered statistically significant.

Results

General characteristics of the study population

Starting from the initial pool of 277 outpatients with amyloid cardiomyopathy, a total of 212 patients (71%) met the inclusion/exclusion criteria (204 wild-type ATTR and 8 variant ATTR) and were therefore considered for the present analysis (Medical University of Vienna: 108; Fondazione Toscana Gabriele Monasterio: 60; Centro Cardiologico Monzino: 44). Indeed, data from 65 patients were ruled out due to one or more of the previously mentioned exclusion criteria, atrial fibrillation being the most prevalent hindering condition (Figure 1). Within the study population, in accordance with each centre’s usual practice (see Methods), the time interval between baseline CPET and the definitive diagnosis of amyloid cardiomyopathy was generally short (4.6, IQR 3.2–5.8 months). Regarding medical therapy, we reported only β-blocker use, as this was the sole treatment in our cohort with the potential to interfere with HR kinetics. β-blockers were prescribed in 87 patients (Supplementary Table S1), primarily for one or more of the following indications: arterial hypertension (45 patients, 52%), a history of paroxysmal AF (23 patients, 26%), ischaemic heart disease (7 patients, 8%), and palpitations (12 patients, 14%). Overall, β-blocker therapy was used in 41% of the cohort, typically at relatively low doses at the time of CPET assessment.

Figure 1.

Study screening procedures and chronotropic incompetence prevalence in the effective cardiac amyloidosis sample according to different pHR% cut-off values

Table 1 reports in detail the main clinical and exercise test data collected at the study run-in in the overall sample and categorized according to the prespecified end-point. The study population consisted mostly of elderly male patients with a moderate exercise limitation. In comparison with event-free survivors, patients who experienced cardiovascular death at 1 year showed a markedly more compromised clinical and functional profile along with lower left ventricular ejection fraction, indicating more advanced cardiac dysfunction. As expected, despite their markedly skewed distribution, NT-proBNP levels were significantly higher in this group compared with event-free survivors at 1 year (Table 1).

Table 1.

Main clinical variables of the overall study sample (n = 212 patients) and categorized according to the prespecified end-point (cardiovascular death at 1 year)

General data	Overall population N (212)	Survivors at 1 year (n = 202)	CV death at 1 year (n = 10)	P values
Age, years	77.1 ± 6.8	76.9 ± 6.7	78.7 ± 8.9	NS
Male, n %	192 (90)	183 (90)	8 (80)	NS
Body mass index, kg/m2	26.5 ± 3.9	26.6 ± 3.9	24.3 ± 3.7	NS
NT-proBNP, pg/ml	1825 [2163]	1758 [2078]	4383 [6125]	<.001
LVEF	52.1 ± 10.4	52.46 ± 10.47	45.20 ± 8.16	.030
β-Blockers, n (%)	87 (41)	83 (41)	4 (40)	NS
β-Blockers dosagea, mg	2.5 [2.5]	2.50 [2.50]	1.88 [1.25]	NS
Exercise testing variables
Rest HR, b.p.m.	72.0 ± 13.9	72.3 ± 14.9	67.0 ± 9.4	NS
Rest SBP, mm Hg	126.5 ± 18.9	127.2 ± 18.9	112.5 ± 13.5	.016
pHR, b.p.m.	119.8 ± 26.3	121.0 ± 26.1	99.9 ± 19.7	.011
pHR%, % of predicted	83.8 ± 18.0	84.6 ± 17.9	70.8 ± 13.9	.017
ΔHR, b.p.m.	47.8 ± 22.8	48.7 ± 22.7	33.0 ± 16.7	.031
pVO2, ml/kg/min	14.9 ± 4.2	15.1 ± 4.2	11.2 ± 3.0	.004
pVO2, % of predicted	71.9 ± 18.5	73.0 ± 18.1	52.4 ± 14.1	.001
O2 pulse peak, mL/beat	9.8 ± 2.15	10.0 ± 2.89	7.9 ± 1.8	.022
VE/VCO2 slope	39.2 ± 6.3	37.8 ± 7.5	46.6 ± 10.5	.001
Peak WL, W	80.3 ± 31.4	82.3 ± 30.4	46.0 ± 25.7	<.001
Peak SBP, mm Hg	160.8 ± 26.9	162.2 ± 26.8	136.50 ± 18.86	.003

Data are expressed as mean ± SD, as absolute number of patients (% on total sample) or as median [75th–25th percentile]. NT-proBNP, N-terminal pro-B-type natriuretic peptide; LVEF, left ventricular ejection fraction; SBP, systolic blood pressure; HR, heart rate; pHR, peak HR; ΔHR, (pHR—resting HR); WL, workload; pVO₂, peak oxygen consumption.

^abisoprolol dose equivalent (see Methods for details).

All key exercise-derived parameters were substantially impaired among patients with subsequent CV death. Specifically, they exhibited lower pHR, a blunted increase in HR (ΔHR: 33.0 ± 16.7 vs. 48.7 ± 22.7, P = .031) and reduced pHR% (70.8 ± 13.9 vs. 84.6 ± 17.9, P = .013). Functional capacity was markedly diminished, as reflected by lower pVO₂ as per cent predicted (52.4 ± 14.1 vs. 73.0 ± 18.1, P = .001).

Chronotropic incompetence prevalence and association with functional capacity

Figure 1 shows the difference in CI prevalence according to three different pHR% cut-off values. Particularly, adopting the pHR% ≤75% cut-off, CI was identified in 76 patients (35%). Baseline clinical and exercise variables stratified according to the latter pHR% cut-off value are provided in Supplementary Table S2. Patients with CI presented impaired pVO₂, either in terms of absolute values corrected for body weight (mL/kg/min) (12.9 ± 3.3 vs. 16.0 ± 4.3, P < .001) or expressed as a percentage of the maximum predicted (%) (63.1 ± 15.4 vs. 76.7 ± 18.3, P < .001) and a higher VE/VCO₂ slope (39.7 ± 8.7 vs. 37.5 ± 7.3, P = .05) (Figure 2).

Figure 2.

Bar graphs comparing key prognostically relevant cardiopulmonary exercise variables across the pHR% cut-off of 75%. Patients with chronotropic incompetence show reduced functional capacity (pVO₂) and impaired ventilatory efficiency (VE/VCO₂ slope)

Prognostic assessment and risk stratification according to chronotropic incompetence

At 1-year follow-up, 10 cardiovascular deaths (5%) were recorded. Table 2 presents the univariate Cox proportional hazard analysis for the main exercise-testing variables. Treatment peak VO₂ confirmed its well-established prognostic role, with lower pVO₂ values strongly associated with a higher risk of cardiovascular death. Notably, all exercise-induced HR–derived parameters were significant predictors of the pre-specified end-point. Among them, pHR% emerged as a particularly informative marker, with each 1% increase associated with a 5.5% relative risk reduction (HR 0.945; P = .011). Importantly, pHR% remained significantly associated with the primary outcome after adjusting for age, beta-blocker therapy, and left ventricular ejection fraction (LVEF) (HR 0.96, 0.94–0.99, P = .036). This association was lost when pVO₂% was included the model (Supplementary Table S3).

Table 2.

Significant univariate Cox proportional survival analysis according to the main exercise testing and prognostic variables for the prespecified end-point of cardiovascular death at 1 year

Variables	Cardiovascular death at 1 year
	Hazard ratio (95% CI)	P values	C-index
pHR, b.p.m.	0.959 (0.9829–0.990)	.010	0.736
pHR%, % of predicted	0.945 (0.903–0.989)	.011	0.729
ΔHR, b.p.m.	0.954 (0.917–0.994)	.025	0.716
pVO2, ml/kg/min	0.712 (0.567–0.891)	.003	0.788
pVO2, % of predicted	0.922 (0.881–0.965)	<.001	0.830
VE/VCO2 slope	1.111 (1.046–1.180)	<.001	0.747
Peak WL, W	0.942 (0.912–0.975)	<.001	0.824
Peak SBP, mm Hg	0.960 (0.934–0.987)	.004	0.779
NT-proBNPa	1.526 (1.246–1.870)	<.001	0.833

HR, hazard ratio; CI, confidence interval. Only variables with a statistical significance at 15% were included in the table. See Table 1 for other abbreviations.

^aValues reported per 1000 pg/ml increase of NT-proBNP.

To explore the discriminatory ability of CI, ROC curve analysis identified a pHR% threshold of 75% as the optimal cut-off for predicting cardiovascular mortality (sensitivity 70%; specificity 67%; AUC 71%) (Figure 3). The diagnostic accuracy of alternative clinically relevant thresholds is summarized in Table 3.

Figure 3.

Left panel: receiving operator curve showing the point with the best sensitivity and specificity of the pHR% in the entire study sample (n = 212). Right panel: Kaplan–Meier estimator of CV death events for the pHR% according to a 75% cut-off value. See Table 2 for the original univariate Cox analysis

Table 3.

Accuracy of the pHR% variable according to different clinical cut-off value for the prespecified end-point of cardiovascular death at 1 year

HR variables	Sensitivity, %	Specificity, %	PPV, %	NPV, %	AUC
pHR% ≤80%	70	55	7	97	71
pyre% ≤75%	70	67	9	98	71
pHR% ≤70%	40	88	14	96	71

PPV, positive predictive value; NPV, negative predictive value; AUC, area under the curve

When stratifying the cohort according to the abovementioned cut-off, Kaplan–Meier survival curves demonstrated a clear separation of event-free survival throughout follow-up, with patients exhibiting pHR% ≤75% showing a significantly higher incidence of cardiovascular death (log-rank P < .05) (Figure 3).

As previously mentioned, patients who experienced CV death at 1 year had significantly higher N-terminal pro-B-type natriuretic peptide (NT-proBNP) concentrations (4383 [6125] vs. 1758 [2078], P < .001) and, accordingly, the NT-proBNP was associated with the primary outcome at the univariate analysis (Table 2). At ROC curve analysis, this biomarker confirmed the well-known prognostic accuracy, being 2485 pg/ml the most accurate cut-off value (sensitivity 64%; specificity 100%; AUC 83%). Noteworthy, when patients were stratified according to both NT-proBNP and pHR% thresholds, a higher proportion of events was observed in the subgroup with both elevated NT-proBNP and reduced pHR% (Figure 4).

Figure 4.

Scatter plot showing the distribution of patients according to pHR% and NT-proBNP levels. Dashed lines indicate the predefined cut-offs for pHR% and NT-proBNP, dividing the population into four quadrants

Discussion

The present multicentre analysis, conducted on a relatively large cohort of stable consecutive outpatients with ATTR cardiomyopathy, showed that a blunted exercise-induced HR response is not only common and closely linked to reduced functional capacity, but also associated with an increased risk of 1-year cardiovascular mortality. Specifically, our data suggest the use of a pHR% ≤75% as a clinically meaningful threshold in ATTR cardiomyopathy.

In recent years, CI has been increasingly recognized as a pathophysiologic dimension of HFpEF, both contributing to dampened cardiac output adaptation to increasing exercise demands and leading to increased left ventricular filling pressure.²⁰ Cardiac amyloidosis, a typical underlying aetiology of HFpEF, in perfect alignment is characterized by a significant prevalence of CI, ranging from 16% up to 59%, according to parameter and cut-off value adopted, the stage of disease, and the background pharmacological therapy.^14–16,21 Compared with the typical ‘garden variety’ of HFpEF patients, where a cardiac β–receptor desensitization have been identified,^22,23 the derangement in HR response during exercise in the cardiac amyloidosis endotype is upstream and possibly related to the autonomic nervous system more than cardiac infiltration.^14,15,24 Specifically, in a recently published paper by Patel and colleagues conducted on a remarkable cohort of patients with ATTR, CI was found to be independent of the cardiac infiltration degree evaluated at cardiac magnetic resonance imaging.¹⁵ However, in ATTR amyloid cardiomyopathy both the autonomic dysfunction and the direct disruption of the heart electrical system concur to a detrimental effect on chronotropism²⁵, leading first to atrial myopathy and sinus node isolation²⁶ and subsequently affecting the specialized conduction cells.²⁷ The occurrence of AF and/or of atrioventricular or intraventricular conduction defects, leading to pacemaker implantation alter definitively the physiological HR response to exercise²⁸, wasting a compensative mechanism in restrictive physiology. In this context, our findings align closely with those of our previous multicentre analysis on the prevalence and functional correlates of CI in amyloid cardiomyopathy. In that earlier study, a substantial proportion of patients exhibited a blunted chronotropic response, with prevalence estimates ranging from 16% to 59% depending on the adopted pHR% threshold. Importantly, lower pHR% values were consistently associated with a parallel deterioration in functional capacity, reflected by a stepwise reduction in pVO₂ and a worsening of ventilatory efficiency across decreasing pHR% strata.¹⁶ Notably, the same pHR% ≤75% threshold emerged as the most accurate discriminator of significant exercise impairment, defined as pVO₂ <60% of predicted, showing the best accuracy values. The present study extends those observations by demonstrating that the prognostic relevance of this cut-off goes beyond functional stratification. Indeed, a pHR% ≤75% not only identifies patients with a markedly impaired exercise profile but also delineates a subgroup at increased short-term cardiovascular mortality.

While the link between CI and impaired exercise capacity has a solid pathophysiological foundation and has been meticulously characterized, our results add the novel evidence that a blunted increase in heart rate is associated with a worse prognosis, namely with an increase in 1-year CV mortality. Indeed, cardiac involvement in amyloidosis often dictates the prognosis of the syndrome²⁹, with two-thirds of deaths resulting from cardiovascular causes, mainly worsening HF (67%) and sudden death (23%).³⁰ The reduction in cardiac output coupled with increased filling pressure could be a driver of neurohormonal activation, thus fostering a vicious cycle of worsening HFpEF. Alternatively, CI could mirror the extent of cardiac electrical involvement³¹, anticipating the occurrence of life-threatening bradyarrhythmias or pulseless electrical activity.³² The present research paper, conducted within a cohort of ATTR patients all on sinus rhythm, after accounting for possible cardiac and extra-cardiac secondary factors affecting HR kinetics, gave us the ideal context where to test and confirm this prognostic relationship between CI and cardiac amyloidosis. CPET remains a cornerstone for functional assessment and prognostic stratification in cardiac amyloidosis, with pVO₂¹¹, circulatory power¹³ and VE/VCO₂ slope³³ being particularly informative on adverse clinical outcomes. Yet, CPET availability is often limited to tertiary or university centres. In this context, the chronotropic response, and specifically the pHR%, emerges as a non-metabolic and widely accessible parameter that not only provides meaningful pathophysiological insight into a patient’s clinical phenotype but may also represent a pragmatic alternative tool for risk assessment. Importantly, our data strengthen the reliability of the commonly used pHR% ≤75% threshold for CI definition: beyond its physiological rationale, this cut-off is now corroborated by prognostic evidence, as a blunted HR response was strongly associated with 1-year cardiovascular mortality in our cohort.

Eventually biomarkers, and NT-proBNP in particular, have consistently demonstrated a strong and independent prognostic value in cardiac amyloidosis, largely irrespective of (and rather complementary to) functional capacity parameters, such as peak VO₂.¹³ In our cohort, we confirm the prognostic relevance of NT-proBNP, with a threshold comparable with that reported in previous studies.³⁴ In this context, pHR%, while not directly reflecting aerobic capacity, may still contribute to a more nuanced risk stratification by identifying a clinically meaningful subgroup of patients at higher risk of adverse outcomes (Figure 4).

Limitations

The retrospective design of the present analysis, together with the limited number of cardiovascular events, represents an important limitation that must be acknowledged. Although a parsimonious Cox regression model was performed, including age, β-blocker therapy, and LVEF selected a priori for their established influence on chronotropic response, the small number of events necessarily constrained the number of covariates that could be reliably included. As a consequence, comprehensive multivariable adjustment was not feasible, and residual confounding cannot be excluded. Specifically, the aim and ambition of our study was not to demonstrate that CI assessment (and specifically pHR%) could compete with metabolic metrics in risk stratification of patients with cardiac amyloidosis, but to provide a scalable alternative. The exclusion of patients with AF or pacemaker-dependent rhythm, while resulting in the selection of a lower-risk subgroup, allowed us to isolate and examine the intrinsic impairment of heart rate kinetics without the confounding influence of rhythm abnormalities or pacing.

Regarding background medication, a significant percentage of patients were receiving β-blocking agents at the time of the study run-in, these agents known to be deleterious with respect to the haemodynamic phenotype of cardiac amyloidosis (HFpEF).³⁵ However, it should be noted that the relatively low bisoprolol-equivalent dosage as well as the lack of association with the primary outcome (Table 1, Supplementary Table S2). Additionally, CI remained significantly associated with 1-year CV death even after adjustment for β-blocker therapy.

Our cohort included only patients who were naïve to disease-specific therapy. Therefore, we cannot speculate on whether specific treatments might restore a more physiological heart rate response²¹ or on the impact of subsequent changes in medical therapy. After CPET, 110 patients (52%) initiated treatment with a TTR stabilizer (tafamidis) and 4 out of the 10 patients who experienced the primary outcome were receiving this medication at the time of the event. However, as recently demonstrated, mortality curves across ATTR-CM trials begin to diverge approximately 12–18 months after therapy initiation.³⁶ Accordingly, since we focused on CV death at 1 year, it might be hypothesized that tafamidis would not impact disease trajectory within this timeframe.

Finally, from a technical standpoint, although CPET assessments were performed and interpreted by highly experienced physicians, the analysis was not centralized; each participating centre used its own equipment, both for cycle ergometry and metabolic gas analysis, which may have introduced inter-site variability in CPET-derived measurements. Additionally, compared with treadmill exercise, the use of a cycle ergometer elicits a slightly less physiological exercise response, typically resulting in lower pVO₂ and pHR. However, we believe that the use of a cycle ergometer may enhance the generalizability of our findings, as it reflects routine clinical practice. In this context, our study provides a simple cut-off that may allow community centres to assess the presence and clinical relevance of CI in ATTR patients using a standard bicycle test.

Conclusions

In conclusion, beyond its pathophysiological significance, our data suggest that a pHR% ≤75% threshold represents a clinically meaningful and prognostically validated cut-off value to be adopted in stable ATTR outpatients with cardiac involvement, as it identifies patients with a worse functional status and at increased risk of short-term CV mortality. Thus, this non-invasive and easy-to-obtain exercise parameter might be an acceptable alternative prognostic tool when CPET is not accessible and, contextually, it could ultimately assist in guiding therapeutic decisions and resource allocation.

Supplementary data

Supplementary data are available at ESC Heart Failure online.

Declarations

Disclosure of Interest

The authors report there are no competing interests to declare.

Data Availability

The data that support the findings of this study are available on request from the corresponding author (E.F.).

Funding

DigiCardiopaTh PhD program, Sapienza University of Rome to E.F. Pfizer research grant ID (87737205) to R.B.E.

Ethical Approval

Ethical Approval was not required.

Pre-registered Clinical Trial Number

None supplied.