Natriuretic peptide responses to exercise in HFpEF diagnosis: a pilot study

Abstract

Background

Diagnosing heart failure with preserved ejection fraction (HFpEF) can be challenging because natriuretic peptide (NP) plasma levels and diastolic function parameters during rest echocardiography may be normal or marginally abnormal in patients displaying exercise-induced symptoms suggestive of HF. In these patients exercise echocardiography is recommended.

We wanted to assess whether directly measured NP levels post-exercise had added diagnostic value beyond rest plasma N-terminal pro-brain NP (NT-proBNP) levels for diagnosing HFpEF.

Methods

Participants with left ventricular diastolic dysfunction (LVDD), who could not be classified as HFpEF by a panel of experts more than 4 years ago, were prospectively enrolled in the HELPFulUP observational study from August 2021 to October 2022. All participants in the HELPFulUP study underwent clinical assessment, rest and exercise-echocardiography and measurements of plasma NT-proBNP before and directly after exercise. An expert panel, blinded to exercise NT-proBNP results, but with knowledge of signs, symptoms and all other diagnostic parameters including baseline values of NT-proBNP adjudicated HFpEF status. We calculated the area under the receiver operating characteristic curve (area under the curve (AUC)) for HFpEF for rest and post-exercise NT-proBNP levels and the delta NT-proBNP.

Results

Of the 112 LVDD participants (59 women), 11 were diagnosed with HFpEF by the expert panel based on the additional information retrieved from exercise echocardiography, and 101 remained classified as LVDD. Rest (AUC=0.78 (95% CI 0.66 to 0.90) and exercise values of NT-proBNP (AUC=0.77 (95% CI 0.65 to 0.90) had similar discriminatory value. The delta NT-proBNP AUC was 0.53 (95% CI 0.31 to 0.75).

Conclusions

In this pilot study, the diagnostic performance of the change with exercise in NP levels to detect incident HFpEF, beyond measurements in rest, did not point to a strong diagnostic value.

WHAT IS ALREADY KNOWN ON THIS TOPIC.

WHAT THIS STUDY ADDS

• This is the first study exploring the diagnostic potential of exercise N-terminal pro-brain NP (NT-proBNP) for diagnosing HFpEF.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• There seems to be no incremental value of directly measured exercise-induced NT-proBNP beyond rest NT-proBNP for diagnosing HFpEF. However, due to limited sample size, no definite conclusions can be drawn. Future, larger studies are needed to also investigate the role of atrial NPs in a broader study population.

Background

Heart failure with preserved ejection fraction (HFpEF) is a clinical syndrome characterised by exercise-induced complaints, mainly shortness of breath, due to increased left ventricular (LV) filling pressures and LV diastolic dysfunction.¹ Diagnosing HFpEF in patients with exercise-induced symptoms may be challenging, because natriuretic peptide (NP) plasma levels can be normal and diastolic function during rest echocardiography may be inconclusive for HFpEF diagnosis.^{1 2} Exercise echocardiography or right heart catheterisation is recommended if rest findings with echocardiography and NP values are inconclusive.³ However, these additional diagnostic approaches have the disadvantage of being time-consuming and invasive, respectively.

NPs are released in response to increased myocardial wall stress, and levels are higher in HF with reduced ejection fraction (HFrEF) than in HFpEF. NP levels increase with higher wall stress, at equally increased LV filling pressures, but importantly, the LV wall stress is less in HFpEF than in HFrEF according to Laplace’s Law (LV wall stress = (LV pressure × LV radius)/2× LV wall thickness).⁴ Additionally, HFrEF patients constantly show elevated filling pressures, leading to a continuous NP release, while in HFpEF patients this may only occur during exercise or during a HF exacerbation.^{2 5–8} Because LV filling pressures rise with exercise, direct post-exercise NP levels and the exercise-induced rise (delta) could possibly provide added diagnostic value beyond rest plasma N-terminal pro-brain NP (NT-proBNP) levels for diagnosing HFpEF.

Methods

To test our hypothesis, we enrolled patients with LV diastolic dysfunction (LVDD) based on functional and/or morphological abnormalities with rest echocardiography and according to the opinion of an expert panel that also used other test results, as described previously.⁹ In a cross-sectional study, 4.4 years (IQR 4.2–4.7 years) after the initial assessment, the value of rest, post-exercise and delta in NT-proBNP was assessed for diagnosing HFpEF in patients that had LVDD during the initial assessment on average 4.4 years before. A power calculation considering a single predictor (delta NT-proBNP) and a binary outcome (HFpEF vs no HFpEF) with an estimated prevalence of 10% resulted in a sample size estimation of 139 patients.^{10 11} Patients provided written informed consent, and the study procedures conformed to the Declaration of Helsinki. All study measurements were approved by the Utrecht medical ethics committee (number 16–290, 21–198).

All 112 eligible patients with LVDD underwent a clinical assessment for signs and symptoms, followed by venous blood withdrawal, rest ECG and rest echocardiography (figure 1). Additionally, all underwent stepwise incremental supine bicycle exercise-echocardiography (Lode Angio, Groningen, The Netherlands; General Electric Vivid E95, Horten, Norway) targeted to 70% of predicted workload in approximately 15 min, less if limited by complaints.¹² We acquired maximal average E/e’ ratio and tricuspid regurgitation (TR) velocity at three exercise stages (low, intermediate and peak intensity level), considering E/A fusion and image quality, as described elsewhere.¹³ A second venous blood withdrawal, performed 15–20 min after peak exercise, allowed us to repeatedly measure NT-proBNP (BD Vacutainer Barricor Lithium Heparine-plasma collection tube, Becton, Dickinson and Company, USA; Atellica Immunoassay Analyzer, Siemens, USA). All examinations were carried out in the University Medical Center Utrecht by A.v.O and E.D.C.

Figure 1. Graphical abstract. We hypothesised that exercise-induced ‘overshoot’ in left ventricular filling pressures in those with HFpEF would result in a more than average increase in myocardial wall stress with exercise, leading to a substantial elevation of natriuretic peptide plasma levels. This steep rise in comparison to stage B heart failure patients would then provide discriminatory value beyond rest NT-proBNP values for diagnosing HFpEF. HFpEF, heart failure with preserved ejection fraction; NT-proBNP, N-terminal pro-brain natriuretic peptide.

A panel of at least two cardiologists and an experienced general practitioner decided on presence or absence of HFpEF with all available patient data, except the direct post-exercise NT-proBNP, and they used the HFA-PEFF score as guidance (figure 2).¹ We calculated the area under the receiver operating curve (area under the curve (AUC)) for discrimination of patients with HFpEF from those still having pre-clinical diastolic dysfunction of (i) NT-proBNP at rest and (ii) post-exercise NT-proBNP to be able to assess our primary outcome (iii) the delta in NT-proBNP. Additionally, we calculated Spearman’s correlation coefficients for exercise NT-proBNP, delta NT-proBNP and peak E/e’ ratio and TR velocity, respectively.

Figure 2. HFA-PEFF diagnostic algorithm The HFA-PEFF algorithm is a method to establish HFpEF diagnosis in patients suspected of HFpEF. The score is based on a combination of rest echocardiography functional and morphological criteria (0, 1 or 2 points per category), and a natriuretic peptide-based biomarker score (0, 1 or 2 points depending on the level). A total score ≥5 is diagnostic for HFpEF. If the findings are inconclusive (2–4 points), and then exercise testing is recommended. HFpEF, heart failure with preserved ejection fraction; LAVI, left atrial volume indexed to body surface area; LV, left ventricular; NT-proBNP, N-terminal pro-brain natriuretic peptide; TR, tricuspid regurgitation.

The data that support the findings of this study are available on reasonable request from the corresponding author. The data are not publicly available due to privacy and ethical restrictions.

Results

Patients were consecutively enrolled from August 2021 to October 2022. The mean age was 67 (±SD 8) years and 59 (52.7%) were women (table 1). Eleven patients (10.2%) of those with LVDD on average 4.4 years ago were now diagnosed with HFpEF, of which seven were women. HFpEF patients were significantly older, had higher relative wall thickness and higher rest values of NT-proBNP than the ones who remained classified as LVDD by the panel (median NT-proBNP (IQR): 156 (138, 280) vs 90 (40, 150), p value=0.001). During exercise, HFpEF patients had, compared with patients with diastolic dysfunction, a significantly higher maximal E/e’ ratio, a shorter exercise time and a higher TR velocity, while achieving a lower cardiac output and workload. Post-exercise NT-proBNP values were higher in HFpEF patients, but the delta was equal for both groups (5 (±SD 19) vs 8 (±SD 15) pg/mL, p value=0.528). Accordingly, rest NT-proBNP (AUC=0.78; 95% CI 0.66 to 0.90) and post-exercise NT-proBNP (AUC=0.77; 95% CI 0.65 to 0.90) had similar discriminatory value. The delta NT-proBNP had an AUC of 0.53 (95% CI 0.31 to 0.75) for HFpEF (figure 3). The non-significant association between exercise NT-proBNP, delta NT-proBNP and peak E/e’ ratio and TR velocity, respectively, are displayed in the scatterplots in figure 4.

Table 1. Baseline characteristics and rest and exercise echocardiographic findings stratified by HFpEF and patients with functional and/or morphological abnormalities with echocardiography who did not classify for HFpEF.

	HFpEF	Cardiac abnormalities, but no HFpEF	p Value
n	n=11	n=101
Baseline characteristics
Women (%)	7 (63.6)	52 (51.5)	0.65
Age in years (mean (SD))	72 (9)	66 (8)	0.033
Body mass index in kg/m² (mean (SD))	28 (6)	27 (5)	0.66
eGFR (mL/min/1.73 m²) (mean (SD))	73 (18)	83 (13)	0.034
Hypertension (%)	8 (72.7)	55 (54.5)	0.40
Diabetes (%)	1 (9.1)	7 (6.9)	1
Hypercholesterolaemia (%)	4 (36.4)	42 (41.6)	0.99
Atrial fibrillation (%)	0 (0.0)	6 (11.8)	0.78
Oedema (%)	3 (27.3)	12 (11.9)	0.34
NYHA class $\ge$2 (%)	7 (63.6)	24 (23.8)	0.019
Rest findings with echocardiography
Heart rate in beats per minute (mean (SD))	62 (13)	67 (10)	0.11
Systolic blood pressure in mm Hg (mean (SD))	149 (18)	143 (20)	0.32
Cardiac output in mL/min (mean (SD))	4379 (1204)	4706 (1390)	0.50
NT-proBNP in pg/mL (median (IQR))	156(138, 280)	90(40, 150)	0.001
LVEF in % (mean (SD))	57 (3)	59 (6)	0.38
E/A ratio (mean (SD))	0.79 (0.31)	0.87 (0.22)	0.27
E/e’ ratio (mean (SD))	9.68 (2.60)	8.49 (2.21)	0.10
TR velocity in cm/sec (mean (SD))	212 (28)	228 (30)	0.21
LAVI in mL/m² (mean (SD))	36 (13)	30 (8)	0.038
Relative wall thickness (mean (SD))	0.50 (0.09)	0.43 (0.09)	0.017
Exercise findings with echocardiography
Peak heart rate in beats per minute (mean (SD))	120 (23)	131 (18)	0.07
Peak heart rate as % predicted (mean (SD))	81 (17)	85 (13)	0.37
Peak systolic blood pressure in mmHg (mean (SD))	200 (31)	208 (24)	0.31
Maximal E/e’ ratio (mean (SD))	14.2 (3.1)	10.1 (2.5)	<0.001
Maximal TR velocity in cm/sec (mean (SD))	350 (19)	314 (54)	0.15
Peak cardiac output in mL/min (mean (SD))	8199 (1132)	10 879 (3152)	0.06
Peak workload in Watt (mean (SD))	107 (42)	123 (26)	0.08
Peak workload as % predicted (mean (SD))	81 (25)	88 (15)	0.16
Exercise duration in minutes (mean (SD))	12 (4)	14 (3)	0.042
Time to blood withdrawal in minutes (mean (SD))*	29 (8)	32 (5)	0.06
NT-proBNP after exercise in pg/mL (median (IQR))	173 (144, 300)	96 (44, 160)	0.002
Delta NT-proBNP in pg/mL (mean (SD))	5 (19)	8 (15)	0.53

^*Time to blood withdrawal was measured from exercise initiation onwards.

eGFR, estimated glomerular filtration rate (according to the 2021 CKD-EPI formula); LAVI, left atrial volume indexed to body surface area; LVEF, left ventricular ejection fraction; NT-proBNP, N-terminal pro-brain natriuretic peptide; NYH, New York Heart Association; TR, tricuspid regurgitation.

Figure 3. Area under the curve for rest NT-proBNP, post-exercise NT-proBNP and delta (∆) NT-proBNP Area under the curve of continuous values of Δ NT-proBNP (red curve), NT-proBNP at rest (green curve) and NT-proBNP after exercise (pg/mL) (blue curve). AUC, area under the curve; aHFpEF, heart failure with preserved ejection fraction; NT-proBNP, N-terminal pro-brain natriuretic peptide.

Figure 4. Correlation between Δ NT-proBNP and NT-proBNP after exercise with peak exercise echocardiography measurements indicative of increased filling pressures Association of (A) after exercise NT-proBNP and peak E/e’ ratio, (B) after exercise NT-proBNP and peak TR velocity, (C) delta in NT-proBNP and peak E/e’ ratio and (D) delta in NT-proBNP and peak TR velocity. No significant associations were observed. NT-proBNP, N-terminal pro-brain natriuretic peptide; TR, tricuspid regurgitation.

Discussion

Exercise echocardiography is useful for uncovering HFpEF in patients with inconclusive diastolic dysfunction with rest echocardiography. In this pilot study, measuring NT-proBNP directly after exercise echocardiography did not point to a strong diagnostic value for HFpEF diagnosis if rest NT-proBNP levels are available.

NT-proBNP release in response to exercise

Natriuretic peptides are secreted in response to wall stress, partly directly from cardiomyocyte storage granules and partly after rapid activation of the proBNP gene, which results in de novo myocyte peptide synthesis and secretion.¹⁴ Several previous studies showed that peak BNP and NT-proBNP levels were reached within 1 hour of short-term maximal exercise in healthy individuals, but also in HFrEF and HFpEF patients.^{5–8 15–17} One study also investigated associations of NT-proBNP levels at peak exercise with other parameters of HFpEF.⁸ However, the added diagnostic value of change with exercise in NT-proBNP was not studied in any of these studies, thus precluding direct comparison with our findings. However, the data provided by these other studies confirm adequate timing of exercise NT-proBNP measurements in our study.

Why direct exercise NT-proBNP measurement seems not helpful in our study

There are several possible explanations why we did not find added value of exercise NT-proBNP beyond resting NT-proBNP in our study. (i) The contrast between patients detected with HFpEF based on exercise echocardiography, but not rest echocardiography, compared with patients with insufficient abnormalities on both rest and exercise echocardiography to classify for HFpEF, was too small. (ii) Those who remained classified as diastolic dysfunction insufficient for classification as HFpEF had a lower relative wall thickness compared with HFpEF patients, which results in a relatively quicker rise in wall stress in response to elevated LV filling pressures because of Laplace’s law.⁴ (iii) Previous diagnostic HFpEF studies were performed in HFpEF patients, comparing them to healthy controls, while we assessed a more homogeneous group of patients with abnormal diastolic dysfunction more than 4 years before the assessment with exercise echocardiography. This leads to a smaller contrast between groups in our study compared with other studies. (iv) The exercise provided with stress echocardiography could be of insufficient length or resistance to adequately increase LV wall stress and thus the release of NT-proBNP to contrast those detected with HFpEF based on E/e’ and TR velocity with stress echocardiography and those who remained without HFpEF.

Future approaches

We measured NT-proBNP only once, on average 30 min after exercise initiation, which might be too early to catch the peak NT-proBNP level. Nevertheless, previous publications suggest that a peak in NT-proBNP may be reached within 1 hour of short-term maximal exercise.¹⁷ Future studies that include multiple measurements of NT-proBNP after exercise would possibly be more informative on the optimal diagnostic value. Additionally, comparison between multiple subgroups, also including healthy individuals, patients with asymptomatic hypertension and HFrEF patients, would improve our understanding of the exercise-induced change in NPs.⁸ However, we do not think that ultimately natriuretic peptide values with exercise will replace echocardiography. In keeping with this, a recent study performed in ambulatory patients with chronic dyspnoea, suspected of HFpEF, found that the single use of NT-proBNP led to unacceptably high error rates, with important interactions by obesity and AF status.¹⁸ In our pilot study, we did not measure atrial natriuretic peptides (ANP), which theoretically could discriminate HFpEF patients from those with diastolic dysfunction better, because a larger quantity than (NTpro)BNP could be released from storage granula.¹⁴ Also, because the atria are often enlarged in patients with HFpEF, a pronounced rise in atrial wall stress could induce a significant rise in ANP levels in HFpEF patients compared with LVDD patients, in whom atrial remodelling is likely less severe. Therefore, using a strategy involving ANPs could circumvent the pitfall of concentric remodelling that inhibits the rise in LV wall stress in HFpEF patients following Laplace’s law.

Limitations

The sample size of our study is rather low and should be considered a pilot study as it does not meet the pre-specified sample size. Additionally, we were unable to assess effect modifiers such as sex because of the small number of patients with HFpEF. Finally, HFpEF diagnosis was not verified using right-heart catheterisation. Also, there was no cardiopulmonary exercise testing performed synchronously to the exercise echocardiogram. Combined cardiopulmonary exercise testing and exercise echocardiography has been shown to improve diagnostic accuracy.¹⁹

Conclusions

In this pilot study, the diagnostic performance of the change with exercise in natriuretic peptide levels to detect incident HFpEF, beyond measurements in rest, did not point to a strong diagnostic value.

Data availability statement

Data are available upon reasonable request.