Abstract
Aims:
Cardiorespiratory fitness (CRF) is closely linked to health status and clinical outcomes in heart failure (HF) patients. We aimed to test whether biomarkers can reflect CRF and its change over time.
Methods and Results:
This post-hoc analysis utilized data from ambulatory cohorts of heart failure with reduced ejection fraction (HFrEF) (IRONOUT) and heart failure with preserved ejection fraction (HFpEF) (RELAX). Cardiopulmonary exercise testing, 6 minute walk distance (6MWD), and serum biomarkers were measured at baseline and 16 or 24 week follow-up (for IRONOUT and RELAX respectively). Biomarkers included N-terminal pro–B-type natriuretic peptide (NT-proBNP), soluble ST2, growth differentiation factor-15 (GDF-15), and galectin-3. Analysis included 225 patients with HFrEF and 216 with HFpEF. Baseline peak VO2, VE/VCO2 slope and 6MWD showed a mild correlation with the doubling of all four tested biomarkers in HFrEF and HFpEF. Following multivariable adjustment (including all biomarkers), the only significant association between change in biomarker and functional parameter in HFrEF was change in NT-proBNP and change in VE/VCO2 slope (3.596% increase per doubling, 95%CI [0.779, 6.492]; p=0.012). In HFpEF a decrease in peak VO2 was associated with an increase in NT-proBNP (−0.726 ml/min/kg per doubling, 95%CI [−1.100, −0.353]; p<0.001) and a decrease in 6MWD was associated with an increase in GDF-15 (−31.606 m per doubling, 95%CI [−61.404, −1.809]; p=0.038).
Conclusions:
In these ambulatory trial cohorts, NT-proBNP was associated with baseline and change in CRF in HFrEF and HFpEF. In contrast, novel biomarkers do not appear suitable as a reliable surrogate for serial assessment of exercise capacity in HF patients given lack of consistent independent association with CRF beyond traditional risk factors and NT-proBNP.
Clinical Trial Registration:
and
Keywords: Heart failure, preserved ejection fraction, reduced ejection fraction, biomarkers, exercise function, cardiorespiratory fitness
INTRODUCTION
Among patients with heart failure (HF), cardiorespiratory fitness as measured by cardiopulmonary exercise testing (CPET) or the 6 minute walk test are powerful predictors of morbidity and mortality [1-4]. Cardiorespiratory fitness is a key objective measure recommended for risk stratification during evaluation for advanced HF therapies [5]. Further, cardiorespiratory fitness is a variable that is responsive to various therapies in patients with chronic HF (exercise training, medications, mechanical circulatory devices) [5-8], thus supporting the value of serial measurements of cardiorespiratory fitness. A decrease in cardiorespiratory fitness between examinations is strongly associated with a higher risk of adverse patient-reported and clinical outcomes, when compared to patients experiencing an increase in cardiorespiratory fitness [9]. However, performing standardized functional testing in patients as part of routine clinical care, especially CPET, is complicated by multiple barriers such as need for equipment, trained personnel, time, and patient effort.
A series of clinically established and novel biomarkers of cardiovascular stress, inflammation, and fibrosis in patients with HF have been associated with adverse clinical outcomes and are either used or proposed for use as risk predictors [10]. Serial testing of biomarkers to help guide HF management, whether for medical/decongestive therapies or repeat risk stratification, has been investigated retrospectively and prospectively with mixed results [11-14]. Nonetheless, whether biomarkers can serve as reliable and practical alternative to serial CPET for longitudinal monitoring of patient status remains to be determined [15, 16]. Moreover, data linking traditional cardiorespiratory fitness measures to specific biomarkers may invoke biologic processes most responsible for impaired functional status, such as filling pressures and wall stress (N-terminal pro-B-type natriuretic peptide [NT-proBNP]), hypertrophy and fibrosis (soluble ST2, [ST2] and galectin 3), and inflammation (growth differentiation factor [GDF-15]). Identifying a potential link between cardiopulmonary performance and novel biomarkers may also inform our understanding of potential (differential) pathophysiological relationships between key HF subgroups such as heart failure with preserved ejection fraction (HFpEF) and heart failure with reduced ejection fraction (HFrEF).
In the present analysis, we sought to test the association between biomarkers and cardiorespiratory fitness at baseline and during follow up in two well characterized chronic HF clinical trial cohorts from the Heart Failure Network (HFN) trials: the IRONOUT (Iron Repletion effects ON Oxygen UpTake in Heart Failure) trial of patients with HFrEF, and the RELAX (Phosphodiesterase-5 Inhibition to Improve Clinical Status and Exercise Capacity in Heart Failure with Preserved Ejection Fraction) trial of patients with HFpEF.
METHODS
Statistical Analysis
Overview
This post-hoc analysis was performed using data from the National Heart, Lung, and Blood Institute (NHLBI)-sponsored HFN IRONOUT and RELAX trials (Figure 1). Each protocol was approved by the Institutional Review Boards at each site, and written informed consent was obtained from all patients before randomization. IRONOUT was a double-blind, placebo-controlled, randomized trial designed to test the efficacy and safety of oral iron polysaccharide compared to placebo among patients with HFrEF (defined as an EF ≤40% and New York Heart Association [NYHA] functional class II-IV symptoms) and iron deficiency (i.e. defined as a ferritin 15-100 ng/mL or ferritin 100-299 ng/mL with a transferrin saturation <20%)[17]. RELAX was a double-blind, randomized trial designed to test the effect of the phosphodiesterase-5 inhibitor sildenafil compared with placebo on exercise capacity and clinical status in HFpEF. RELAX enrolled ambulatory patients who had an EF ≥50% and objective evidence of HF [18, 19]. Subjects were required to have elevated N-terminal pro–B-type natriuretic peptide (NT-proBNP, ≥400 pg/mL) or elevated invasively measured filling pressures and reduced exercise capacity (≤60% age-, sex- and body size-specific predicted VO2). In both trials, patients with advanced chronic kidney disease were excluded (estimated glomerular filtration rate < 20 mL/min/1.73m2).
Figure 1.
CONSORT diagram
Biomarkers and Exercise Testing
Participants underwent baseline studies including a history and physical examination, echocardiography, CPET, six-minute walk test, and serum biomarkers. CPET and biomarker measurements were performed at baseline and follow-up at 16 weeks in IRONOUT and 24 weeks in RELAX. In both studies, CPET was performed using an identical protocol and interpreted by the HFN CPET core laboratory (Massachusetts General Hospital, Boston, MA) as previously reported [19, 20]. Patients unable to achieve a respiratory exchange ratio greater than or equal to 1.0 on baseline screening CPET were excluded. Plasma biomarker measurements were performed by the HFN biomarker core laboratory (University of Vermont, Burlington, VT) as previously described [19, 20] and included NT-proBNP (i.e., marker of filling pressures and wall stress), ST2 (i.e., cardiovascular stress and fibrosis), GDF-15 (i.e., inflammation) and Galectin 3 (i.e., mediator of inflammation and fibrosis).
Statistics
Baseline characteristics were summarized with medians and quartiles for continuous variables and counts and percentages for categorical variables. Multiple imputation by chained equations was used to impute missing biomarkers and covariates for the regression models, which were aggregated across 100 imputed datasets. Missing outcomes were not imputed; patients without a baseline or follow-up outcome were omitted from the corresponding model. Imputation and regression were stratified by clinical trial/HF subgroup, with separate models for IRONOUT (HFrEF) and RELAX (HFpEF) patients.
This analysis investigated the associations between each of three exercise capacity outcomes (peak VO2, VE/VCO2 slope, and 6-minute walk distance [6MWD]) and each of the four plasma biomarkers mentioned previously (NT-proBNP, ST2, GDF-15, and Galectin 3). The baseline values of the outcomes and biomarkers were compared cross-sectionally, and the changes in the outcomes from baseline to 16 (IRONOUT, HFrEF) or 24 (RELAX, HFpEF) weeks were compared to the corresponding changes in the biomarkers. Each outcome-biomarker association was evaluated in a univariable model, an adjusted model controlling for six pre-specified potential confounders (age, sex, body mass index, prior HF hospitalization in past year, systolic blood pressure, and blood urea nitrogen), and a multivariable model further controlling for the other biomarkers. All models also included a random intercept for study site. Variables were transformed as appropriate to address linearity and influential observations. All biomarkers were log2 transformed. Generalized linear regression models were fit using the gamma, inverse Gaussian, and normal distributions, and the distribution with the smallest Akaike information criterion was chosen; the normal distribution was selected in each case. Sensitivity analyses controlling for treatment and treatment-biomarker interactions were also conducted for the change-from-baseline models to assess the impact of omitting treatment, which had null effects in both original trials, from the primary models.
Two-sided p values <0.05 were considered statistically significant. Statistical analyses were completed using SAS software, version 9.4 (SAS Institute Inc., Cary, North Carolina).
RESULTS
Baseline characteristics
This analysis included all 225 patients with HFrEF (IRONOUT and 216 with HFpEF (RELAX). Patients in both trial cohorts were predominantly white and male with a mean age >60 years (Table 1). The HFpEF ambulatory cohort had more evidence of congestion and a greater comorbidity burden than the HFrEF cohort. As expected, the HFrEF cohort had a high rate of guideline-directed medication use but diuretic use was >75% in both cohorts. Both patient groups had markedly impaired baseline exercise capacity as measured by peak VO2, VE/VCO2 slope, and 6MWD (Table 2). In both study groups, median levels of NT-proBNP and GDF-15 were above the upper limit of normal, whereas median levels for Galectin-3 and ST2 were within the reference range (Table 2/Supplement Table 1).
Table 1.
Baseline characteristics in the two trial cohorts
| Characteristic | HFrEF (IRONOUT Trial) (N=225) |
HFpEF (RELAX Trial) (N=216) |
|---|---|---|
| Demographics | ||
| Age, years: median (Q1, Q3) [N] | 63 (55, 70) [225] | 69 (62, 77) [216] |
| Female sex: n/N (%) | 80 / 225 (35.6%) | 104 / 216 (48.1%) |
| Self-reported white race: n/N (%) | 164 / 225 (72.9%) | 197 / 216 (91.2%) |
| Clinical Assessment | ||
| Body mass index, kg/m2: median (Q1, Q3) [N] | 29.2 (25.7, 33.8) [224] | 32.9 (28.3, 39.1) [216] |
| Elevated jugular venous pressure[1]: n/N (%) | 26 / 222 (11.7%) | 95 / 209 (45.5%) |
| Heart rate, beats/min: median (Q1, Q3) [N] | 71 (64, 79) [225] | 69 (61, 78) [216] |
| Left ventricular ejection fraction, %: median (Q1, Q3) [N] | 25 (20, 34) [225] | 60 (56, 65) [213] |
| New York Heart Association functional classification[2]: n/N (%) | ||
| II | 149 / 225 (66.2%) | 101 / 216 (46.8%) |
| III | 75 / 225 (33.3%) | 115 / 216 (53.2%) |
| Peripheral edema >= trace: n/N (%) | 63 / 225 (28.0%) | 125 / 216 (57.9%) |
| Systolic blood pressure, mmHg: median (Q1, Q3) [N] | 112 (98, 125) [225] | 126 (113, 138) [216] |
| Medical History | ||
| Atrial fibrillation/flutter[3]: n/N (%) | 86 / 222 (38.7%) | 111 / 216 (51.4%) |
| Chronic kidney disease >= stage 3: n/N (%) | 124 / 225 (55.1%) | 119 / 216 (55.1%) |
| Diabetes mellitus: n/N (%) | 88 / 225 (39.1%) | 93 / 216 (43.1%) |
| HF hospitalization in past year: n/N (%) | 97 / 225 (43.1%) | 79 / 216 (36.6%) |
| Hypertension: n/N (%) | 162 / 224 (72.3%) | 183 / 216 (84.7%) |
| Ischemic heart disease: n/N (%) | 175 / 225 (77.8%) | 84 / 216 (38.9%) |
| Medications at Enrollment | ||
| Mineralocorticoid receptor antagonist: n/N (%) | 136 / 225 (60.4%) | 23 / 216 (10.6%) |
| Angiotensin converting enzyme inhibitor or angiotensin II receptor blocker: n/N (%) | 189 / 225 (84.0%) | 152 / 216 (70.4%) |
| Beta-blocker: n/N (%) | 216 / 225 (96.0%) | 164 / 216 (75.9%) |
| Loop diuretic: n/N (%) | 185 / 224 (82.6%) | 166 / 216 (76.9%) |
| Laboratory Results | ||
| Blood urea nitrogen, mg/dL: median (Q1, Q3) [N] | 22 (16, 30) [223] | 25 (17, 33) [176] |
| Creatinine, mg/dL: median (Q1, Q3) [N] | 1.2 (1.0, 1.5) [225] | 1.2 (0.9, 1.5) [216] |
| Cystatin C, mg/L: median (Q1, Q3) [N] | 1.1 (0.8, 1.3) [223] | 1.3 (1.1, 1.7) [214] |
| Hemoglobin, g/dL: median (Q1, Q3) [N] | 12.6 (11.8, 13.4) [224] | 12.9 (11.9, 13.8) [215] |
Recorded as elevated/distended in IRONOUT, and as >= 8 cm H2O in RELAX.
One patient was classified as having class I symptoms in IRONOUT.
Recorded as atrial fibrillation in IRONOUT, and as atrial fibrillation/flutter in RELAX.
Table 2.
Biomarker Profile and Functional Parameters in HFrEF and HFpEF Ambulatory Patients
| Characteristic | HFrEF (IRONOUT Trial) (N=225) |
HFpEF (RELAX Trial) (N=216) |
|---|---|---|
| Laboratory Results * | ||
| Galectin-3, ng/mL: median (Q1, Q3) [N] | 14 (11, 18) [196] | 14 (11, 18) [208] |
| GDF15, pg/mL: median (Q1, Q3) [N] | 1331 (875, 2046) [201] | 2324 (1537, 3713) [161] |
| NT-proBNP, pg/mL: median (Q1, Q3) [N] | 1111 (453, 2412) [222] | 700 (283, 1553) [213] |
| ST2, ng/mL: median (Q1, Q3) [N] | 28 (22, 38) [204] | 34 (27, 47) [174] |
| CPET / Functional Capacity | ||
| Peak VO2, mL/min/kg: median (Q1, Q3) [N] | 13.2 (11.1, 15.7) [224] | 11.7 (10.2, 14.4) [215] |
| Peak respiratory exchange ratio: median (Q1, Q3) [N] | 1.13 (1.06, 1.20) [224] | 1.09 (1.02, 1.15) [215] |
| VE/VCO2 Slope: median (Q1, Q3) [N] | 34 (30, 40) [224] | 33 (28, 38) [210] |
| Six-minute walk distance, m: median (Q1, Q3) [N] | 363 (292, 428) [225] | 308 (229, 383) [216] |
Reference values: Galectin-3 < 22.1 ng/mL; GDF-15 </= 750 pg/mL; NT-proBNP < 300pg/mL; ST-2 </= 35ng/mL
Relationship between functional capacity and biomarkers at baseline
Baseline peak VO2 showed a weak correlation with the doubling of all four tested biomarkers across both trials (Figure 2). Similar, but less prominent, was a correlation between VE/VCO2 slope and biomarkers at baseline (Supplement Figure 1). Finally, 6MWD was correlated with a doubling of biomarkers at baseline (Supplement Figure 2). In unadjusted and adjusted analysis, most biomarkers were significantly associated with functional capacity, whether measured by CPET or 6MWD (Supplement Table 2). Following multivariable adjustment (accounting for all four biomarkers), in HFrEF, NT-proBNP and GDF-15 at baseline were significantly associated with baseline peak VO2 (NT-proBNP: −0.344 mL/min/kg per doubling, 95% CI [−0.585, −0.103], p=0.005; and GDF-15: −1.528 mL/min/kg per doubling, 95% CI [−2.179, −0.876], p<0.001) and VE/VCO2 slope (NT-proBNP: 3.398% increase per doubling, 95% CI [1.501, 5.330], p<0.001; and GDF-15: 5.370% increase per doubling, 95% CI [0.244, 10.759], p=0.040) (Table 3, Figure 3). Only GDF-15 was associated with a lower 6MWD (−33.579 m per doubling, 95% CI [−57.645, −9.513], p=0.006).
Figure 2.
Correlation Between Baseline Peak VO2 and Baseline Biomarkers by HF Type in HFrEF and HFpEF
Table 3.
Multivariable Association of Baseline Biomarkers with Baseline Functional Capacity by HF Type
| Adjusted[1] | |||
|---|---|---|---|
| Outcome | Biomarker | Mean Difference (95% CI)[2] |
P-value |
| HFrEF (IRONOUT Trial) | |||
| Peak VO2, mL/kg/min | Log2(Galectin-3), per doubling | 0.291 (−0.545, 1.126) | 0.495 |
| Log2(GDF15), per doubling | −1.528 (−2.179, −0.876) | <.001 | |
| Log2(NT-proBNP), per doubling | −0.344 (−0.585, −0.103) | 0.005 | |
| Log2(ST2), per doubling | −0.794 (−1.619, 0.030) | 0.059 | |
| VE/VCO2 slope, % relative difference | Log2(Galectin-3), per doubling | −2.959 (−8.897, 3.366) | 0.351 |
| Log2(GDF15), per doubling | 5.370 (0.244, 10.759) | 0.040 | |
| Log2(NT-proBNP), per doubling | 3.398 (1.501, 5.330) | < .001 | |
| Log2(ST2), per doubling | 3.347 (−2.998, 10.107) | 0.308 | |
| Six-minute walk distance, m | Log2(Galectin-3), per doubling | −7.292 (−38.253, 23.669) | 0.644 |
| Log2(GDF15), per doubling | −33.579 (−57.645, −9.513) | 0.006 | |
| Log2(NT-proBNP), per doubling | −2.322 (−11.099, 6.454) | 0.604 | |
| Log2(ST2), per doubling | −10.228 (−40.492, 20.036) | 0.508 | |
| HFpEF (RELAX Trial) | |||
| Peak VO2, mL/kg/min | Log2(Galectin-3), per doubling | −0.033 (−0.695, 0.629) | 0.922 |
| Log2(GDF15), per doubling | −1.010 (−1.487, −0.532) | <.001 | |
| Log2(NT-proBNP), per doubling | −0.464 (−0.676, −0.251) | <.001 | |
| Log2(ST2), per doubling | 0.341 (−0.343, 1.024) | 0.329 | |
| VE/VCO2 slope, % relative difference | Log2(Galectin-3), per doubling | −1.265 (−6.846, 4.650) | 0.668 |
| Log2(GDF15), per doubling | 4.438 (−0.078, 9.158) | 0.054 | |
| Log2(NT-proBNP), per doubling | 2.523 (0.653, 4.427) | 0.008 | |
| Log2(ST2), per doubling | 1.557 (−3.969, 7.400) | 0.588 | |
| Six-minute walk distance, m | Log2(Galectin-3), per doubling | 2.626 (−23.743, 28.995) | 0.845 |
| Log2(GDF15), per doubling | −4.407 (−23.284, 14.470) | 0.647 | |
| Log2(NT-proBNP), per doubling | −4.555 (−12.804, 3.693) | 0.279 | |
| Log2(ST2), per doubling | −24.963 (−49.844, −0.083) | 0.049 | |
Adjusted for age, sex, body mass index, prior HF hospitalization in past year, systolic blood pressure, blood urea nitrogen, site, and all other biomarkers.
Estimated difference in given units of outcome per given unit difference in biomarker.
Figure 3:
Functional Capacity Change vs. Biomarker Change by Heart Failure Type. (A) HFrEF (IRONOUT Trial), Change from Baseline to 16 Weeks Post-Baseline; (B) HFpEF (RELAX Trial, Change from Baseline to 24 Weeks Post-Baseline
In HFpEF, NT-proBNP and GDF-15 were associated with peak VO2 (NT-proBNP: −0.464 mL/min/kg per doubling, 95% CI [−0.676, −0.251], p<0.001; and GDF-15: −1.010 mL/min/kg per doubling, 95% CI [−1.487, −0.532], p<0.001). For VE/VCO2 slope only NT-proBNP was found to have a significant association. For 6MWD there was only an association with ST2 in the multivariable analysis (Table 3, Figure 3).
Relationship between change in functional capacity and change in biomarkers
Associations between changes in exercise functional parameters and parallel changes in biomarkers from baseline to follow up were evaluated in 196 patients with HFrEF (IRONOUT and 189 with HFpEF (RELAX with measured functional capacity at follow-up (Supplement Table 3). Univariable (unadjusted and adjusted) results are presented in Figure 3 and Supplement Table 4. Following a multivariable analysis including all biomarkers (Table 4, Figure 4), in patients with HFrEF, the only significant association between a biomarker and functional parameter was a change in NT-proBNP and a change in VE/VCO2 slope (3.596% increase per doubling, 95% CI [0.779, 6.492]; p=0.012). In HFpEF, an increase in NT-proBNP was associated with a decrease in peak VO2 (−0.726 ml/min/kg per doubling, 95% CI [−1.100, −0.353]; p<0.001) and an increase in GDF-15 was associated with decrease in 6MWD (−31.606 m per doubling, 95% CI [−61.404, −1.809]; p=0.038).
Table 4.
Multivariable Association of Biomarker Change with Functional Capacity Change by HF Type
| Adjusted[1] | |||
|---|---|---|---|
| Outcome | Biomarker | Mean Difference (95% CI)[2] |
P-value |
| HFrEF (IRONOUT Trial), Change from Baseline to 16 Weeks Post-Baseline | |||
| Peak VO2, mL/kg/min | Log2(Galectin-3), per doubling | 0.501 (−0.572, 1.575) | 0.360 |
| Log2(GDF15), per doubling | −0.195 (−1.176, 0.785) | 0.696 | |
| Log2(NT-proBNP), per doubling | −0.206 (−0.573, 0.161) | 0.271 | |
| Log2(ST2), per doubling | −0.954 (−1.938, 0.030) | 0.057 | |
| VE/VCO2 slope, % relative difference | Log2(Galectin-3), per doubling | 6.461 (−2.257, 15.956) | 0.151 |
| Log2(GDF15), per doubling | 0.175 (−7.119, 8.041) | 0.964 | |
| Log2(NT-proBNP), per doubling | 3.596 (0.779, 6.492) | 0.012 | |
| Log2(ST2), per doubling | 4.578 (−2.929, 12.667) | 0.239 | |
| Six-minute walk distance, m | Log2(Galectin-3), per doubling | −8.112 (−43.901, 27.676) | 0.657 |
| Log2(GDF15), per doubling | −27.096 (−58.850, 4.658) | 0.094 | |
| Log2(NT-proBNP), per doubling | −2.187 (−14.490, 10.117) | 0.728 | |
| Log2(ST2), per doubling | −3.443 (−36.437, 29.552) | 0.838 | |
| HFpEF (RELAX Trial), Change from Baseline to 24 Weeks Post-Baseline | |||
| Peak VO2, mL/kg/min | Log2(Galectin-3), per doubling | 0.193 (−0.516, 0.903) | 0.593 |
| Log2(GDF15), per doubling | −0.599 (−1.369, 0.171) | 0.127 | |
| Log2(NT-proBNP), per doubling | −0.726 (−1.100, −0.353) | < .001 | |
| Log2(ST2), per doubling | −0.123 (−1.100, 0.854) | 0.805 | |
| VE/VCO2 slope, % relative difference | Log2(Galectin-3), per doubling | −3.942 (−9.097, 1.505) | 0.153 |
| Log2(GDF15), per doubling | 2.756 (−2.991, 8.843) | 0.355 | |
| Log2(NT-proBNP), per doubling | 2.922 (−0.048, 5.980) | 0.054 | |
| Log2(ST2), per doubling | −0.294 (−7.394, 7.351) | 0.938 | |
| Six-minute walk distance, m | Log2(Galectin-3), per doubling | −17.050 (−45.540, 11.440) | 0.241 |
| Log2(GDF15), per doubling | −31.606 (−61.404, −1.809) | 0.038 | |
| Log2(NT-proBNP), per doubling | −10.783 (−25.518, 3.953) | 0.152 | |
| Log2(ST2), per doubling | 1.492 (−34.373, 37.357) | 0.935 | |
Adjusted for age, sex, body mass index, prior HF hospitalization in past year, systolic blood pressure, blood urea nitrogen, site, baseline biomarker levels, baseline outcome, and all other biomarkers.
Estimated change in given units of outcome from baseline to 16 (24) weeks per given unit change in biomarker from baseline to 16 (24) weeks.
Figure 4.
Multivariable Association of Biomarker Change with Functional Capacity Change by HF Type with Placebo Group Sensitivity Analysis
The decision to omit treatment from the follow-up analyses was assessed in two ways. A sensitivity analysis in only placebo control patients was performed to check for consistency with the primary analyses in treated and placebo patients (Supplement Table 5, Figure 4). Further analyses adjusting for treatment and treatment-biomarker interactions found a statistically significant interaction only for Peak VO2 vs. Galectin-3 in HFpEF (not shown). One significant result (p < 0.05) among 24 tests was not considered strong evidence for including treatment.
DISCUSSION
In two well-characterized ambulatory trial cohorts, a biomarker panel testing various domains of cardiovascular stress generally showed mild correlations with baseline cardiorespiratory fitness in patients with chronic HFrEF and HFpEF. Most notably, there was only a weak association between change in cardiorespiratory fitness and a change in biomarkers over 4-6 months both in HFrEF and HFpEF. In contrast to NT-proBNP, novel biomarkers demonstrated inconsistent independent associations with cardiorespiratory fitness, suggesting limited incremental utility in the serial monitoring of exercise capacity in HF patients.
The primary purpose of this study was to assess if a select panel of HF biomarkers were associated with CPET performance in patients with HFpEF and HFrEF potentially allowing substitution for cardiorespiratory fitness testing. Developing new tools/biomarkers to predict and track functional capacity is of high importance to patients and providers. The biomarkers we chose to test in this post-hoc analysis of two clinical cohorts represent a wide array of surrogates of pathophysiological processes believed to underlie the development and progression of HF. Pathophysiological processes including left ventricular wall stretch from increased volume or pressure (NT-proBNP), inflammation (GDF-15), and cardiovascular stress/fibrosis (ST2 and Galectin 3) represent underlying mechanism and/or disease severity. The three novel biomarkers (GDF-15, ST2 and Galectin 3) in particular were put forward as surrogates of myocardial remodeling given uncovered biological links [21-24]. Irrespective of the mechanistic role of these surrogate markers, all the novel biomarkers have been shown to incrementally improve risk prediction for adverse clinical outcomes such as acute HF and sudden cardiac death in chronic HF often exceeding that of the natriuretic peptides [21, 25, 26]. Some evidence suggests that the same associations may extend to NT-proBNP [15, 16, 27, 28], GDF-15 [29], ST2 [30], Galectin 3 [31, 32] and to measures of cardiorespiratory fitness, which could present a particularly helpful objective tool in the long-term management of chronic HF patients. Our data support prior reports and extend them through the use of a well-phenotyped trial patient population, use of standardized CPET protocols and clinically practical 6MWD and inclusion of HFrEF and HFpEF patients. At baseline peak VO2, VE/VCO2 slope and 6MWD were only mildly correlated with the four biomarkers in HFrEF and HFpEF. When adjusted for all biomarkers, only NT-proBNP and GDF-15 remained independently associated with peak VO2 and VE/VCO2 slope, irrespective of HFrEF or HFpEF status.
More importantly, we evaluated whether a change in cardiorespiratory fitness correlated with a change in any of the four tested biomarkers and thus would allow detection of trends in cardiorespiratory fitness. We observed a wide range in functional parameter changes in both trials under investigation. For example, a change in peak VO2 by 6% (~0.7-0.8 mg/ml/min2) from the baseline is considered clinically significant and was observed in the majority of the patients across both trials [9]. There was only a weak to modest correlation between a change in functional parameters and biomarkers across study follow up. In a multivariable analysis, we found no independent association between GDF-15, ST2 or Galectin 3 with any of the tested CPET parameters, despite prior evidence to suggest potential benefit for serial testing of GDF-15 [33], ST2 [34, 35] and Galectin 3 [26] in chronic HF, given their independent prognostic information for left ventricular remodeling, adverse clinical outcomes and correlation with medical treatment. The only significant association between novel biomarkers was between GDF-15 with 6MWD, the least well validated measure of cardiorespiratory fitness in our analysis [36]. Interestingly, only NT-proBNP was independently associated with a change in both peak VO2 (in HFpEF) and VE/VCO2 slope (in HFpEF). Our findings support the potential utility for NT-proBNP to assess improvement or deterioration of cardiorespiratory fitness but do not support the serial use of novel biomarkers for the longitudinal monitoring of exercise capacity in HF patients.
The pathophysiological determinants of cardiorespiratory fitness such as cardiovascular congestion, left ventricular remodeling driven by macrovascular/microvascular ischemia, fibrosis and inflammation [5] can be captured by a multimodal biomarker profile. However, there appears to be only a limited independent value for most of the novel biomarkers when adjusted for traditional clinical, biochemical parameters and other biomarkers. NT-proBNP appears to stand out when compared to the novel biomarkers, which could suggest that mid-term changes in cardiorespiratory fitness are predominately the result of hemodynamic or congestion-related changes that might be best reflected by NT-proBNP. Various dimensions of ventricular remodeling represented by GDF-15, ST2 and Galectin 3, while good surrogates of overall disease severity, might not be equally well suited to track cardiorespiratory fitness. Further, unlike NT-proBNP, GDF-15, ST2 and Galectin 3 are not cardiac-specific, but merely (over-)expressed in the heart as the result of the HF disease process [37-39]. Thus, the potential contribution of other non-cardiac processes to levels of novel biomarkers may make them less sensitive to changes in cardiac status.
Limitations
This is a post-hoc analysis of two randomized, controlled, double-blinded trials, was limited to 16/24 weeks follow-up time, and the studies were not powered to detect changes in the cardiorespiratory fitness endpoints. The two trial populations are not directly comparable. Further, follow-up time and degree of baseline cardiorespiratory impairment varied between cohorts, and so this aspect of the analysis is also not comparable between the two cohorts. These results may not be generalizable to all ambulatory HFrEF and HFpEF phenotypes, especially given that the sample was skewed towards white participants. Lastly, this study must be interpreted in the context of patients with complete biomarkers and cardiorespiratory fitness data at baseline and follow-up, excluding patients with interval death or lost-to-follow-up who may have had greatest disease severity.
Conclusions
In these ambulatory trial cohorts, NT-proBNP was associated with baseline cardiorespiratory fitness in HFrEF and HFpEF and change in cardiorespiratory fitness in HFrEF and HFpEF. In contrast, the novel biomarkers were only modestly associated with the cardiorespiratory fitness but after adjustment for traditional risk factors and NT-proBNP there was no consistent independent association with the outcomes of interest. This suggests that there was only limited utility in using these biomarkers in the serial monitoring of exercise capacity in HF patients beyond what’s already available.
Supplementary Material
Central Figure:
Review of Results
Acknowledgments
Funding: The research reported in this publication was supported by the National Heart, Lung, and Blood Institute (NHLBI) of the National Institutes of Health (NIH) under award number U10 HL084904 (for the Coordinating Center) and award numbers U10 HL110297, U10 HL110342, U10 HL110309, U10 HL110262, U10 HL110338, U10 HL110312, U10 HL110302, U10 HL110336, and U10 HL110337 (for the Regional Clinical Centers). Database management and statistical analyses were performed by the Duke Clinical Research Institute (Durham, NC). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH/NHLBI.
Conflict of Interest: M.F. is supported by an American Heart Association Grant, 17MCPRP33460225 and NIH T32 grant 5T32HL007101. S.J.G. is supported by the NIH T32 postdoctoral training grant (5T32HL069749) and a Heart Failure Society of America/ Emergency Medicine Foundation Acute Heart Failure Young Investigator Award funded by Novartis. M.V. is supported by the NHLBI T32 postdoctoral training grant (T32HL007604). All other authors declare no relevant financial disclosures.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
REFERENCES
- 1.Kodama S, et al. , Cardiorespiratory fitness as a quantitative predictor of all-cause mortality and cardiovascular events in healthy men and women: a meta-analysis. Jama, 2009. 301(19): p. 2024–35. [DOI] [PubMed] [Google Scholar]
- 2.Mancini DM, et al. , Value of peak exercise oxygen consumption for optimal timing of cardiac transplantation in ambulatory patients with heart failure. Circulation, 1991. 83(3): p. 778–86. [DOI] [PubMed] [Google Scholar]
- 3.Arena R and Sietsema KE, Cardiopulmonary exercise testing in the clinical evaluation of patients with heart and lung disease. Circulation, 2011. 123(6): p. 668–80. [DOI] [PubMed] [Google Scholar]
- 4.Bittner V, et al. , Prediction of mortality and morbidity with a 6-minute walk test in patients with left ventricular dysfunction. SOLVD Investigators. JAMA, 1993. 270(14): p. 1702–7. [PubMed] [Google Scholar]
- 5.Ross R, et al. , Importance of Assessing Cardiorespiratory Fitness in Clinical Practice: A Case for Fitness as a Clinical Vital Sign: A Scientific Statement From the American Heart Association. Circulation, 2016. 134(24): p. e653–e699. [DOI] [PubMed] [Google Scholar]
- 6.O'Connor CM, et al. , Efficacy and safety of exercise training in patients with chronic heart failure: HF-ACTION randomized controlled trial. JAMA, 2009. 301(14): p. 1439–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Ahmad T, et al. , Effects of left ventricular assist device support on biomarkers of cardiovascular stress, fibrosis, fluid homeostasis, inflammation, and renal injury. JACC Heart Fail, 2015. 3(1): p. 30–9. [DOI] [PubMed] [Google Scholar]
- 8.Ahmad T, et al. , Prognostic Implications of Long-Chain Acylcarnitines in Heart Failure and Reversibility With Mechanical Circulatory Support. J Am Coll Cardiol, 2016. 67(3): p. 291–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Swank AM, et al. , Modest increase in peak VO2 is related to better clinical outcomes in chronic heart failure patients: results from heart failure and a controlled trial to investigate outcomes of exercise training. Circ Heart Fail, 2012. 5(5): p. 579–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Ibrahim NE and Januzzi JL Jr., Established and Emerging Roles of Biomarkers in Heart Failure. Circ Res, 2018. 123(5): p. 614–629. [DOI] [PubMed] [Google Scholar]
- 11.Felker GM, et al. , Effect of Natriuretic Peptide-Guided Therapy on Hospitalization or Cardiovascular Mortality in High-Risk Patients With Heart Failure and Reduced Ejection Fraction A Randomized Clinical Trial. JAMA, 2017. 318(8): p. 713–720. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Daubert MA, et al. , NT-proBNP Goal Achievement Is Associated With Significant Reverse Remodeling and Improved Clinical Outcomes in HFrEF. JACC Heart Fail, 2019. 7(2): p. 158–168. [DOI] [PubMed] [Google Scholar]
- 13.Bajaj NS, et al. , Effect of NT-proBNP-Guided Therapy on All-Cause Mortality in Chronic Heart Failure With Reduced Ejection Fraction. J Am Coll Cardiol, 2018. 71(8): p. 951–952. [DOI] [PubMed] [Google Scholar]
- 14.Pfisterer M, et al. , BNP-guided vs symptom-guided heart failure therapy: the Trial of Intensified vs Standard Medical Therapy in Elderly Patients With Congestive Heart Failure (TIME-CHF) randomized trial. JAMA, 2009. 301(4): p. 383–92. [DOI] [PubMed] [Google Scholar]
- 15.Kruger S, et al. , brain natriuretic peptide levels predict functional capacity in patients with chronic heart failure. J Am Coll Cardiol, 2002. 40(4): p. 718–22. [DOI] [PubMed] [Google Scholar]
- 16.Maeder M, et al. , B-type natriuretic peptide kinetics and cardiopulmonary exercise testing in heart failure. Int J Cardiol, 2007. 120(3): p. 391–8. [DOI] [PubMed] [Google Scholar]
- 17.Lewis GD, et al. , Effect of Oral Iron Repletion on Exercise Capacity in Patients With Heart Failure With Reduced Ejection Fraction and Iron Deficiency: The IRONOUT HF Randomized Clinical Trial. JAMA, 2017. 317(19): p. 1958–1966. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Redfield MM, et al. , Effect of phosphodiesterase-5 inhibition on exercise capacity and clinical status in heart failure with preserved ejection fraction: a randomized clinical trial. JAMA, 2013. 309(12): p. 1268–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Redfield MM, et al. , PhosphdiesteRasE-5 Inhibition to Improve CLinical Status and EXercise Capacity in Diastolic Heart Failure (RELAX) trial: rationale and design. Circ Heart Fail, 2012. 5(5): p. 653–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Lewis GD, et al. , Oral Iron Therapy for Heart Failure With Reduced Ejection Fraction: Design and Rationale for Oral Iron Repletion Effects on Oxygen Uptake in Heart Failure. Circ Heart Fail, 2016. 9(5). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Bayes-Genis A, et al. , Soluble ST2 monitoring provides additional risk stratification for outpatients with decompensated heart failure. Rev Esp Cardiol, 2010. 63(10): p. 1171–8. [DOI] [PubMed] [Google Scholar]
- 22.Xu J, et al. , GDF15/MIC-1 functions as a protective and antihypertrophic factor released from the myocardium in association with SMAD protein activation. Circ Res, 2006. 98(3): p. 342–50. [DOI] [PubMed] [Google Scholar]
- 23.Kempf T, et al. , The transforming growth factor-beta superfamily member growth-differentiation factor-15 protects the heart from ischemia/reperfusion injury. Circ Res, 2006. 98(3): p. 351–60. [DOI] [PubMed] [Google Scholar]
- 24.Sharma UC, et al. , Galectin-3 marks activated macrophages in failure-prone hypertrophied hearts and contributes to cardiac dysfunction. Circulation, 2004. 110(19): p. 3121–8. [DOI] [PubMed] [Google Scholar]
- 25.Ahmad T, et al. , Biomarkers of myocardial stress and fibrosis as predictors of mode of death in patients with chronic heart failure. JACC Heart Fail, 2014. 2(3): p. 260–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Anand IS, et al. , Serial measurement of growth-differentiation factor-15 in heart failure: relation to disease severity and prognosis in the Valsartan Heart Failure Trial. Circulation, 2010. 122(14): p. 1387–95. [DOI] [PubMed] [Google Scholar]
- 27.Kruger S, et al. , The value of cardiopulmonary exercise testing and brain natriuretic peptide plasma levels in predicting the prognosis of patients with chronic heart failure. Eur J Intern Med, 2006. 17(2): p. 96–101. [DOI] [PubMed] [Google Scholar]
- 28.Koglin J, et al. , Role of brain natriuretic peptide in risk stratification of patients with congestive heart failure. J Am Coll Cardiol, 2001. 38(7): p. 1934–41. [DOI] [PubMed] [Google Scholar]
- 29.Sharma A, et al. , Utility of Growth Differentiation Factor-15, A Marker of Oxidative Stress and Inflammation, in Chronic Heart Failure: Insights From the HF-ACTION Study. JACC Heart Fail, 2017. 5(10): p. 724–734. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Felker GM, et al. , Soluble ST2 in ambulatory patients with heart failure: Association with functional capacity and long-term outcomes. Circ Heart Fail, 2013. 6(6): p. 1172–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.de Boer RA, et al. , The fibrosis marker galectin-3 and outcome in the general population. J Intern Med, 2012. 272(1): p. 55–64. [DOI] [PubMed] [Google Scholar]
- 32.Tang WH, et al. , Usefulness of plasma galectin-3 levels in systolic heart failure to predict renal insufficiency and survival. Am J Cardiol, 2011. 108(3): p. 385–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Stahrenberg R, et al. , The novel biomarker growth differentiation factor 15 in heart failure with normal ejection fraction. Eur J Heart Fail, 2010. 12(12): p. 1309–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Gaggin HK, et al. , Head-to-head comparison of serial soluble ST2, growth differentiation factor-15, and highly-sensitive troponin T measurements in patients with chronic heart failure. JACC Heart Fail, 2014. 2(1): p. 65–72. [DOI] [PubMed] [Google Scholar]
- 35.Weir RA, et al. , Serum soluble ST2: a potential novel mediator in left ventricular and infarct remodeling after acute myocardial infarction. J Am Coll Cardiol, 2010. 55(3): p. 243–50. [DOI] [PubMed] [Google Scholar]
- 36.Opasich C, et al. , Six-minute walking performance in patients with moderate-to-severe heart failure; is it a useful indicator in clinical practice? Eur Heart J, 2001. 22(6): p. 488–96. [DOI] [PubMed] [Google Scholar]
- 37.Sanada S, et al. , IL-33 and ST2 comprise a critical biomechanically induced and cardioprotective signaling system. J Clin Invest, 2007. 117(6): p. 1538–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Moore AG, et al. , The transforming growth factor-ss superfamily cytokine macrophage inhibitory cytokine-1 is present in high concentrations in the serum of pregnant women. J Clin Endocrinol Metab, 2000. 85(12): p. 4781–8. [DOI] [PubMed] [Google Scholar]
- 39.AbouEzzeddine OF, et al. , Galectin-3 in heart failure with preserved ejection fraction. A RELAX trial substudy (Phosphodiesterase-5 Inhibition to Improve Clinical Status and Exercise Capacity in Diastolic Heart Failure). JACC Heart Fail, 2015. 3(3): p. 245–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.