Abstract
Background:
Although clinical studies have demonstrated the association between a single N-terminal pro-B type natriuretic peptide (NT-proBNP) measurement and clinical outcomes in chronic heart failure, the biomarker is frequently measured serially in clinical practice.
Objectives:
To determine the added prognostic value of repeated NT-proBNP measurements compared to single measurements alone for chronic heart failure patients.
Methods:
In the GUIDE-IT (Guiding Evidence Based Therapy Using Biomarker Intensified Treatment in Heart Failure) study, 894 study participants with chronic heart failure with reduced ejection fraction (HFrEF) were enrolled at 45 outpatient sites in the United States and Canada. Repeated NT-proBNP levels were measured over a two-year study period. Associations between repeated NT-proBNP measurements and trial endpoints were assessed using a joint longitudinal and survival model.
Results:
After adjustment for baseline covariates, each doubling of baseline NT-proBNP level was associated with a 1.17 (95% CI: 1.08–1.28, p=0.0003) hazard ratio for the primary trial endpoint of cardiovascular death or HF hospitalization. Serial measurements increased the adjusted hazard ratio for the primary trial endpoint to 1.66 (95% CI 1.50–1.84; p<0.0001), and similar increased risk was observed across secondary trial endpoints. In joint modeling, increase in NT-proBNP occurred weeks prior to onset of adjudicated events.
Conclusions:
Repeated NT-proBNP measurements are a strong predictor of outcomes in HFrEF with increase in concentration occurring well before event onset. These results may support routine NT-proBNP monitoring to assist in clinical decision making.
Keywords: GUIDE-IT, guideline-directed medical therapy, heart failure and reduced ejection fraction, natriuretic peptide, NT-proBNP, serial measurement
Central Illustration:
Improved Reclassification of Patient Risk for HF Event using Serial NT-proBNP Measurements
Addition of a NT-proBNP level at 3 months to a baseline measurement enhanced risk stratification of patients. For patients with a HF event, 64 patients were correctly reclassified to a higher risk category. For patients without a HF event, 192 were correctly reclassified to a lower risk category. Risk categories were defined as low (<15%), intermediate (15–25%) and high (>25%) predicted risk of HF event at one year.
INTRODUCTION
Measurement of N-terminal pro B-type natriuretic peptide (NT-proBNP) concentrations is an accepted strategy to predict worse clinical outcomes in heart failure (HF) with reduced ejection fraction (1–2). Treatment with medical therapy generally leads to a decrease in NT-proBNP levels in parallel with benefits from such treatments, while persistent elevations identify a particularly high-risk profile associated with more congestion, worse myocardial remodeling and heightened likelihood for hospitalization (3–5). Although established as a prognostic measure in chronic HF, most studies of NT-proBNP testing have examined the value of single measurements of NT-proBNP that do not consider proximity to events. Given the dynamics of NT-proBNP concentrations relative to disease stability, the most clinically informative approach may be to use repeated measurements that more accurately reflect the dynamic and progressive course of HF as well as response to interventions (6). Absence of available data using truly serial measurements of NT-proBNP as done in clinical practice has led to ambiguity about the utility of such testing (1–2). In addition, the extent to which NT-proBNP changes precede the onset of clinical HF events, and therefore offer early warning of such events, remains unclear.
Prior studies have demonstrated that serial BNP monitoring during outpatient follow-up after acute coronary syndrome predicted mortality and HF risk (7). More recently, long-term NT-proBNP changes over time were associated with risk for developing HF in patients without HF (8). A meta-analysis of natriuretic peptide monitoring in heart failure patients (acute or chronic; preserved or reduced ejection fraction subtypes) found that natriuretic peptide guided therapy did not confer additional benefit over regular care (9). This study did acknowledge that the available evidence was of moderate quality due to variable methods between the included trials.
The GUIDE-IT (Guiding Evidence Based Therapy Using Biomarker Intensified Treatment in Heart Failure) trial, included in the aforementioned meta-analysis, tested the impact of titrating guideline-directed medical therapy (GDMT) using NT-proBNP versus usual care alone (10). The availability of multiple NT-proBNP measurements in both study arms allows for an appraisal of the additional risk information afforded by serial NT-proBNP measurements. Accordingly, in this study, we used data from the GUIDE-IT trial to examine the prognostic impact of repeated measurements of NT-proBNP in stable patients with chronic HF. We hypothesized that repeated NT-proBNP measurements enhance prediction of risk for adverse events compared to baseline measurements alone and further sought to understand timing of adjudicated HF events relative to changes in NT-proBNP.
METHODS
The design and results of the GUIDE-IT trial have been previously published (10, 11). The trial was a randomized, multicenter clinical trial conducted at 45 sites in the United States and Canada. Patients with HF with reduced ejection fraction (HFrEF; left ventricular ejection fraction ≤ 40%), NT-proBNP concentration greater than 2,000 pg/mL, or B-type natriuretic peptide concentration greater than 400 pg/mL within the prior 30 days or a history of a HF event (HF hospitalization or equivalent) were randomized to either an NT-proBNP guided strategy or usual care. Patients randomized to the guided strategy had GDMT titrated to achieve NT-proBNP less than 1,000 pg/mL. Patients in the usual-care arm were managed as recommended in HF clinical practice guidelines targeting optimal GDMT use. The primary endpoint was time to first HF hospitalization or cardiovascular mortality, and the trial was stopped early due to futility. Both arms had comparable outcomes, treatment, and similar reduction of NT-proBNP. The trial data are publicly available through the Biologic Specimen and Data Repository Information Coordinating Center of the National Heart, Lung, and Blood Institute. The Yale University Institutional Review Board approved requisition of the data, and a waiver of the requirement to obtain informed consent was provided to conduct this analysis. All trial participants provided informed consent.
Statistical Analysis
Baseline clinical characteristics were compared between study participants who experienced the primary trial composite endpoint of HF hospitalization or cardiovascular mortality and those who did not; categorical variables were compared using the Pearson chi-square test and continuous variables were compared using the Wilcoxon-rank-sum test. Baseline characteristics are expressed as median (interquartile range [IQR]) for continuous variables and frequency (percentage) for categorical variables.
Joint longitudinal and survival models were developed to demonstrate the added prognostic value for the trial’s clinical endpoints of NT-proBNP testing over baseline testing alone. The joint models combined mixed-effects linear regression to model serial NT-proBNP measurements with Cox proportional hazards models for the study endpoints as previously described (12, 13). The proportional hazards assumption was checked and was appropriate (Supplemental Figure 1). The mixed effects NT-proBNP model had linear and quadratic fixed effects terms and a linear random effects term where NT-proBNP was base 2 logarithmically transformed. Both unadjusted and adjusted Cox models were fit. The adjusted Cox models accounted for baselines variables previously identified in the GUIDE-IT predictive models for the primary endpoint and all-cause mortality, respectively (14). The primary endpoint model adjusted for ICD or pacemaker, baseline sodium, NYHA class (I/II versus III/IV), sleep apnea, Log of baseline creatinine, Hispanic or Latino ethnicity, Black race, and baseline heart rate. The all-cause mortality model adjusted for baseline sodium, history of ischemic heart disease, heart failure duration, baseline diastolic blood pressure, and baseline heart rate. Average estimated NT-proBNP concentrations were plotted for patients experiencing the primary trial endpoint of CV death or HF hospitalization compared to censored patients. The prognostic improvement for a 12-month primary outcome of repeat NT-proBNP testing was also assessed through risk reclassification tables and c-indices of participants who were primary endpoint-free to 3 months. The trial’s 33 non-cardiovascular mortalities were treated as independent censorings if they occurred before the primary event.
Each participant was assigned to a 12-month risk category of low (< 15%), medium (15–25%), and high (>25%) first using a Cox model for baseline NT-proBNP alone, and then using a Cox model for baseline and 3-month NT-proBNP. If the 3-month measurement improved the prognostication, we would expect that participants who had an event by 12 months would more likely be assigned to an equivalent risk or higher risk category using both NT-proBNP measurements than using the baseline NT-proBNP measurement alone. Similarly, we would expect that participants who did not have an event by 12 months would more likely be assigned to an equivalent risk or lower risk category using both NT-proBNP measurements than using the baseline NT-proBNP measurement alone. Participants censored event-free before 12 months had their 12-month outcome multiply imputed using the Kaplan-Meier event probability estimate.
To understand how NT-proBNP values changed over time, transition frequencies and probabilities were calculated for patients split into three groups (NT-proBNP less than 1,000 pg/mL, between 1000–2999 pg/mL, and greater than 3000 pg/mL). In addition, a visual heat map of the absolute risk of cardiovascular death or HF hospitalization at six months was constructed based on the risk conferred by baseline and 3-month NT-proBNP (12). Stata (StataCorp, College Station, Texas) was used for descriptive data analysis. R statistical software (version 4.1.1, R Foundation, Vienna, Austria) and the JM package (version 1.5–2) was used for the joint survival and longitudinal models (12).
RESULTS
There were 328 (36.7%) patients who had a HF hospitalization or died of cardiovascular causes during the two-year study period (Table 1). Baseline NT-proBNP concentrations were higher among patients with an event compared to those patients without an event and conferred an increased adjusted hazard of HF hospitalization or cardiovascular death (HR 1.17 for doubling of NT-proBNP value; 95% CI 1.08–1.28; p=0.0003) (Table 2). We subsequently considered the importance of follow-up NT-proBNP concentrations compared to baseline alone. For the joint longitudinal and survival model, the adjusted hazard ratio using the repeated NT-proBNP measurements for HF hospitalization or cardiovascular death was larger indicating greater prognostic information from repeated measurements than the baseline measurement alone (HR 1.66; 95% CI 1.50–1.84; p<0.001) (Table 2).
Table 1:
Baseline Characteristics of Patients Stratified by Primary Trial Endpoint (Heart Failure Hospitalization or Cardiovascular Death)
| Overall Sample (N=894) | No Event (N=566) | Event* (N=328) | p-value | |
|---|---|---|---|---|
| Age, yrs | 63 (53–71) | 62 (53–71) | 64 (53–72) | 0.49 |
| Baseline BMI, kg/m2 | 28.8 (24.6–33.8) | 28.2 (24.4–33.2) | 29.3 (25.0–34.9) | 0.05 |
| Female | 286 (32.4) | 188 (33.2) | 98 (29.9) | 0.33 |
| Race | ||||
| White | 490 (55.4) | 337 (59.5) | 153 (46.7) | <0.001 |
| Black | 324 (36.7) | 179 (31.6) | 145 (44.2) | |
| Asian | 27 (3.05) | 17 (3.0) | 10 (3.05) | |
| Hispanic/Other | 53 (6.0) | 33 (5.83) | 20 (6.10) | |
| Comorbidities | ||||
| Ischemic heart disease | 447 (50.6) | 251 (44.4) | 196 (59.8) | <0.001 |
| Diabetes mellitus | 410 (46.4) | 231 (40.8) | 179 (54.6) | <0.001 |
| Atrial fibrillation | 145 (16.4) | 98 (17.3) | 47 (14.3) | 0.31 |
| Sleep apnea | 201 (22.7) | 103 (18.3) | 98 (30.2) | <0.001 |
| ICD | 391 (44.2) | 193 (34.1) | 198 (60.4) | <0.001 |
| Depression treated with medications | 140 (15.8) | 83 (14.7) | 57 (17.4) | 0.31 |
| GWTG heart failure risk prediction score | 39 (34–43) | 38 (34–42) | 40 (36–45) | <0.001 |
| Site description | ||||
| Academic | 501 (56.7) | 308 (54.4) | 193 (58.8) | 0.21 |
| Community | 393 (44.5) | 258 (45.6) | 135 (41.2) | |
| Country of Enrollment | ||||
| United States | 756 (85.5) | 462 (81.6) | 294 (89.6) | 0.001 |
| Canada | 138 (15.6) | 104 (18.4) | 34 (10.4) | |
| Baseline Medications | ||||
| Beta-blocker | 845 (95.6) | 545 (96.3) | 300 (91.5) | <0.001 |
| ACE Inhibitor/ARB | 714 (80.8) | 469 (82.9) | 245 (74.7) | 0.003 |
| MRA | 444 (50.2) | 285 (50.4) | 159 (48.5) | 0.84 |
| Loop diuretic | 834 (94.3) | 522 (92.2) | 312 (95.1) | 0.10 |
| Ejection Fraction, % | 24 (20–30) | 25 (30–30) | 22.5 (20–30) | 0.33 |
| NHYA functional class | ||||
| I | 59 (6.67) | 47 (8.30) | 12 (3.66) | <0.001 |
| II | 446 (50.5) | 319 (56.4) | 127 (38.7) | |
| III | 357 (40.4) | 186 (32.9) | 171 (52.1) | |
| IV | 17 (1.92) | 6 (1.06) | 11 (3.35) | |
| Unknown | 15 (1.70) | 8 (1.41) | 7 (2.13) | |
| Baseline laboratory values | ||||
| NT-proBNP, pg/mL | 2607 (1467–5264) | 2311 (1264–4705) | 3377 (1876–6524) | <0.001 |
| Serum sodium, mg/dL | 139 (136–141) | 139 (137–141) | 138 (136–141) | 0.05 |
| Serum potassium, mg/dL | 4.3 (4.0–4.7) | 4.3 (4.0–4.7) | 4.3 (4.0–4.7) | 0.56 |
| GFR, mL/min/1.73m2 | 51.7 (37.9–67.1) | 55.3 (40.6–70.1) | 46.5 (33.9–60.2) | <0.001 |
Event defined as any CV death or HF hospitalization (primary trial endpoint)
Table 2:
Hazard Ratios of Trial Endpoints with Repeated NT-proBNP Measurements Using a Joint Longitudinal and Survival Model
| Baseline NT-proBNP Level | Serial NT-proBNP+ | |||
|---|---|---|---|---|
| N=Number of Events | HR* (95% CI) | p-value | HR* (95% CI) | p-value |
| HF Hospitalization or Cardiovascular Mortality (N=328) | ||||
| Model 1 | 1.25 (1.17–1.35) | <0.0001 | 1.62 (1.49–1.77) | <0.0001 |
| Model 2A | 1.17 (1.08–1.28) | 0.0003 | 1.66 (1.50–1.84) | <0.0001 |
| All-Cause Mortality (N=143) | ||||
| Model 1 | 1.56 (1.39–1.74) | < 0.0001 | 1.92 (1.68–2.18) | < 0.0001 |
| Model 2B | 1.59 (1.38–1.83) | < 0.0001 | 1.75 (1.53–2.01) | < 0.0001 |
| HF Hospitalization (N=288) | ||||
| Model 1 | 1.23 (1.14, 1.33) | < 0.0001 | 1.58 (1.44, 1.74) | < 0.0001 |
| Model 2A | 1.16 (1.06, 1.27) | 0.002 | 1.64 (1.47, 1.84) | < 0.0001 |
Hazard ratio reported as per doubling of NT-proBNP
Repeated measures joint longitudinal and survival model
Model 1 only includes NTproBNP as a covariate, either baseline or repeatedly measured
Model 2A: adjusted for ICD or pacemaker, sodium, NYHA class (I/II vs. III/lV), sleep apnea, Log of creatinine, Hispanic or Latino, Black, Heart rate
Model 2B: adjusted for sodium, history of ischemic heart disease, HF duration, diastolic blood pressure, heart rate
The next consideration was the added value from serial measurements following early medication intensification and early NT-proBNP response to such therapy. For the 621 patients free of primary events to 3-months with baseline and 3-month NT-proBNP measurements, Table 3 reports the risk reclassification of the three-month NT-proBNP measurements with the baseline measurement to predict a 12-month primary outcome. The prespecified categories of predicted risk were >15%, 15–25%, and >25%. Among the 190 participants who had an event by 12 months, including imputed events, 64 (34%) participants were correctly reclassified to a higher risk category while 30 (16%) were incorrectly reclassified to a lower risk category (Central Illustration). Among the 431 participants who did not have an event by 12 months, including imputed non-events, 192 (45%) were correctly reclassified to a lower risk category while 66 (15%) were incorrectly reclassified to a higher risk category. Using both baseline and 3-month measurements had a c-index of 0.73 (95% CI: 0.69–0.76) while using only the baseline measurement had a c-index of 0.54 (95% CI: 0.50–0.59).
Table 3:
Reclassification of participants into predefined risk categories for 12-month primary outcome using the 3-month NTproBNP measurement versus of the baseline NTproBNP measurement.
| Predicted 12-month event rate using the baseline NT-proBNP level | Predicted 12-month event rate using the baseline and 3-month NT-proBNP levels | |||
|---|---|---|---|---|
| <15% | 15–25% | >25% | total | |
| Participants who had an event by 12 months | ||||
| <15% | 1 | 0 | 0 | 1 |
| 15–25% | 24 | 52 | 64 | 140 |
| >25% | 1 | 5 | 43 | 49 |
| Total | 26 | 57 | 67 | 190 |
| Participants who were event-free at 12 months | ||||
| <15% | 14 | 0 | 3 | 17 |
| 15–25% | 150 | 108 | 63 | 321 |
| >25% | 20 | 22 | 51 | 93 |
| 184 | 130 | 117 | 431 | |
Of the 621 patients free of primary events at 3 months, 109 (18%) patients had a baseline NT-proBNP less than 1000 pg/mL (Table 4A). Among these 109 patients, 86 (79%) maintained a measurement less than 1000 pg/ml at 3 months. Among these 86 patients, only 5 (6%) had a subsequent event through the remainder of the study period. In contrast, the post-3-month event rate was 47% for the 17 patients who increased from less than 1000 pg/mL at baseline to between 1000 and 2999 pg/mL at 3 months, and 33% for the 6 patients who increased to greater than 3000 pg/mL at 3 months.
Table 4A:
Baseline and 3-month NT-proBNP categories (N = 621): # of first primary events/#patients (percentage)
| 3-month NT-proBNP | |||||
|---|---|---|---|---|---|
| Baseline NT-proBNP | < 1000 | 1000–2999 | ≥ 3000 | total | |
| < 1000 | 5/86 (6%) | 8/17 (47%) | 2/6 (33%) | 15/109 | |
| 1000–2999 | 8/84 (10%) | 48/130 (37%) | 19/38 (50%) | 75/252 | |
| ≥ 3000 | 4/28 (14%) | 14/86 (16%) | 62/146 (42%) | 80/260 | |
| total | 17/198 | 70/233 | 83/190 | 170/621 | |
Of the 621 patients without events by 3 months, 260 (42%) had a baseline NT-proBNP greater than 3000 pg/mL. Among these 260 patients, 146 (56%) maintained a measurement greater than 3000 pg/mL at 3 months. Among these 146 patients, 62 (42%) had a subsequent event during the study period. In contrast, the post-3-month event rate was 16% for the 86 patients who decreased from greater than 3000 pg/mL at baseline to between 1000 and 2999 pg/mL at 3 months, and 14% for the 28 patients who decreased to less than 1000 pg/mL at 3 months. In the 440 patients free of primary events to 6 months, similar associations between risk and a higher follow-up versus previous NT-proBNP were seen relative to changes from 3 to 6 months and subsequent events (Table 4B). Summary statistics for baseline, 3-month, and 6-month NT-proBNP are given in Supplemental Table 1. Heat maps showing the prediction of HF hospitalization or CV death at 6 months as a function of baseline and 3-month NT-proBNP (Figure 1A) as well as at 12 months as a function of baseline and 3-month NT-proBNP levels (Figure 1B) and 3-month and 6-month NT-proBNP levels (Figure 1C). NT-proBNP levels at 6 months (P<0.0001) more strongly predicted 12-month outcomes compared to 3-month NT-proBNP (P=0.42) so Figure 1D plots the 12-month rate as function of the 6-month level alone.
Table 4B:
3-month and 6-month NT-proBNP categories (N = 440): # of first primary events/#patients (percentage)
| 6-month NT-proBNP | total | ||||
|---|---|---|---|---|---|
| 3-month NT-proBNP | < 1000 | 1000–2999 | ≥ 3000 | ||
| < 1000 | 3/102 (3%) | 7/36 (19%) | 3/25 (12%) | 13/163 | |
| 1000–2999 | 7/42 (17%) | 20/86 (23%) | 8/27 (30%) | 35/155 | |
| ≥ 3000 | 7/25 (28%) | 12/35 (34%) | 28/62 (45%) | 47/122 | |
| total | 17/169 | 39/157 | 39/114 | 95/440 | |
Figure 1A: 6-month HF Hospitalization + CV Death Rates as a Function of the Baseline and 3-month NT-proBNP.
Dots represent the 621 patients who were event-free to 90 days and had both NT-proBNP measurements. In a model with both biomarker levels, levels at 3-months more strongly predicted the primary outcome (p< 0.00001) than the baseline level (p=0.00017).
Figure 1B: 12-month HF Hospitalization + CV Death Rates as a Function of the Baseline and 3-month NT-proBNP Values.
Dots represent individual patients with risk percentiles according to baseline and 3-month NT-proBNP levels. In a model with both biomarker levels, levels at 3-months more strongly predicted the primary outcome (p< 0.00001) than the baseline level (p=0.00017).
Figure 1C: 12-month HF Hospitalization + CV Death Rates as a Function of the 3-month and 6-month NT-proBNP.
Dots represent individual patients with risk percentiles according to 3-month and 6-month NT-proBNP levels. In a model with both biomarker levels, levels at 6-months strongly predicted the primary outcome (p< 0.0001) whereas the 3-month level did not (p=0.42).
Figure 1D:
12-month HF hospitalization + CV death rate as a function of 6-month NT-proBNP with 95% confidence bars
Supplemental Figures 2 and 3 are similar plots for the all-cause mortality endpoint.
Lastly, the timing of NT-proBNP change relative to event onset was considered. On average, in joint modeling, NT-proBNP concentrations showed a positive slope approximately 200 days prior to an event for those patients with HF hospitalization or cardiovascular death rising more steeply in proximity to the event, while those patients without an event had a stable NT-proBNP trajectory (Figure 2).
Figure 2: Association of NT-proBNP with Heart Failure Hospitalization or Cardiovascular Mortality in GUIDE-IT.
Plots represent the average estimated NT-proBNP pattern for patients with (red) and without (blue) a heart failure event. Patients with an event had uptrend in NT-proBNP level occurring around several weeks prior to the event; patients without an event did not demonstrate a similar trend.
DISCUSSION
In this secondary analysis from GUIDE-IT, we examined the association between repeated NT-proBNP measurements and risk for adjudicated adverse HF events. We demonstrated that repeated NT-proBNP measurements over time are a stronger predictor of outcomes in HFrEF than single measurements, with increase in concentration occurring well before event onset. While NT-proBNP is established as a risk predictor in the clinical management of HF, frequency of measurement is often at the discretion of the managing clinician. The results of this study provide important clinical insights, demonstrating that repeated NT-proBNP measurements for patients with chronic active HFrEF, receiving medical therapy in the outpatient setting adds substantial prognostic value and remains an important strategy for risk stratification and management.
Most studies of natriuretic peptide-based risk prediction have examined the prognostic value of single measurements of NT-proBNP. The GUIDE-IT trial selected an NT-proBNP level of 1000 pg/mL as the therapeutic target for titration of GDMT in the natriuretic peptide guided arm, and subsequent analyses affirmed the prognostic benefit of achieving NT-proBNP less than 1000 pg/mL (16). Conversely, there has been evidence from GUIDE-IT and other trials on the adverse outcomes associated with NT-proBNP non-responders (i.e. those patients unable to achieve NT-proBNP measurement less than 1000 pg/mL) (3, 17–18). These associations are important, but single natriuretic peptide measurements may not reflect a patient’s complete clinical trajectory and data are sparse on the predictive value of serial versus single measurements. We demonstrated that the association between repeated NT-proBNP measurements and endpoints was stronger than the association of the baseline NT-proBNP alone. Despite clinical policy documents recommending serial measurement of NT-proBNP at approximately 3–6 month intervals in stable chronic HFrEF, there exists a relative paucity of data supporting such a recommendation (2). The results from this study affirm this recommendation, suggesting that 3-monthly measurement of NT-proBNP during HF therapy increases discrimination for events compared to less frequent measurement. More recent measurements add significantly to prior values, even if the prior concentrations were either low or elevated. Another important finding of this analysis is that the onset of NT-proBNP rise may occur well ahead of clinical events providing a window of opportunity to intervene.
The utility of NT-proBNP monitoring has not yet been demonstrated in clinical trials. A meta-analysis of available data showed no benefit in outcomes for natriuretic peptide monitoring across the HF spectrum, but this conclusion was made with only a moderate level of evidence due to variable methods between the trials (9). The neutral result observed in GUIDE-IT may be attributed to several features of the trial design and treatment patterns observed. There was a similar reduction in NT-proBNP levels between the guided and usual care arms, and medical therapy was also similar between the groups. In addition, the trial was unblinded and employed a 1:1 patient randomization strategy which enabled the same physician to carry out both treatment strategies and potentially bias the overall results. The usual care arm also had required follow-up visits which may not have been scheduled in the absence of NT-proNP data. It is also unknown how serial NT-proBNP monitoring could impact outcomes across clinicians of different backgrounds. GUIDE-IT enrolled patients under the care of HF specialists, who may be able to detect early signs of worsening HF based on history and exam, whereas other clinicians may benefit from NT-proBNP trajectories. Lastly, GUIDE-IT enrolled patients with elevated baseline NT-proBNP indicative of advanced HF, and incremental benefit may have been difficult to achieve between the treatment groups. Given these factors, it will be important to further study how NT-proBNP monitoring can impact outcomes for HF patients in blinded trials and with clinicians of different backgrounds.
Considered in aggregate, the findings of this study suggest that progression of NT-proBNP levels during HF treatment emphasize importance of biomarker monitoring—serial measurement of NT-proBNP during HF treatment—for early indications of clinical decompensation. In clinical practice, the NT-proBNP measurement can be repeated over time to not only risk stratify patients cross-sectionally, but also intervene should an NT-proBNP change occur. Given associations between secular trends in NT-proBNP with cardiac remodeling, such interventions might include reconsideration of adequacy of guideline directed medical therapy, assessment for therapeutic adherence, and examining for congestion (17, 19). Beyond this, evaluating for other factors associated with decompensated HF such as coronary ischemia or paroxysmal atrial arrhythmia might also be advisable. Conversely, routine, low concentrations of NT-proBNP provide reassurance that progressive remodeling is likely absent and stability is much more likely. Monitoring therefore provides utility to focus healthcare efforts more accurately.
Study Limitations
These data are observational and cannot definitively establish cause-effect relationships. In addition, these data may be subject to selection bias, as enrolled patients were required to meet pre-specified clinical characteristics and it is unclear if the findings from HF trials are generalizable to the broader HF population. Further, patients were treated by HF clinicians, and it is possible that treatment patterns may have been different with providers that have less experience in such care. Combined sacubitril-valsartan was newly introduced during the GUIDE-IT trial, with very few subjects receiving the drug, and sodium glucose cotransporter-2 inhibitors were not yet available as part of comprehensive guideline-directed medical therapy.
Conclusions
This analysis from the GUIDE-IT study demonstrates importance of repeated measurements of NT-proBNP as a strong predictor of adverse outcomes for patients with chronic, stable HFrEF compared with baseline measurements alone. They show how secular trends in NT-proBNP are predictive of future events well before such outcomes occur. Taken together these results suggest that NT-proBNP monitoring at periodic intervals (such as every 3 months) may be helpful in clinical practice to identify patients at risk for decompensation.
Supplementary Material
Clinical Perspectives.
Competency in Medical Knowledge
In GUIDE-IT, repeated NT-proBNP measurements for patients with HFrEF conferred a greater risk for HF event than the baseline measurement alone signaling better prognostic information. Among patients with HF events, there was an uptrend in NT-proBNP levels months in advance of the event.
Translational Outlook
Repeated NT-proBNP measurements confer additional prognostic information with uptrend well in advance of HF events. These data may support routine, serial NT-proBNP monitoring for ambulatory HF patients to assist in risk assessment and early intervention.
Disclosures:
The GUIDE-IT trial was supported by the National Heart, Lung, and Blood Institute of the National Institutes of Health. Additional substudies were supported by Roche Diagnostics. The sponsors had no role in the design and conduct of the present study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.
Drs. Fuery, Leifer, Samsky, and Sen have no disclosures to report. Dr. O’Connor has received grant or research support from Merck; and consulting fees from Merck, Bayer, and Abiomed. Dr. Fiuzat has received grant support from the National Institutes of Health and Roche Diagnostics. Dr Ezekowitz has received research grant support and consulting fees from Bayer, Merck, Servier, Amgen, Sanofi, Novartis, Cytokinetics, American Regent, Otsuka, and Applied Therapeutics. Dr Piña has participated on Advisory Boards for Vifor and AstraZeneca and is a Steering Committee member for Novartis. Dr. Whellan has received grants from NHLBI, Merck, Amgen, Cytokinetcis, and Norvo Nordisk; has acted as a consultant to Cytokinetics and Novo Nordisk; and has served on clinical endpoint committees/data safety monitoring boards for CVRx and NIH. Dr Mark reports grant support to institution from HeartFlow and Merck. Dr Felker has received research grants from NHLBI, American Heart Association, Amgen, Bayer, BMS, Merck, Cytokinetics, and CSL-Behring; he has acted as a consultant to Novartis, Amgen, BMS, Cytokinetics, Innolife, Medtronic, Cardionomic, Boehringer-Ingelheim, American Regent, Abbott, Astra-Zeneca, Regeneron, Reprieve, Myovant, Sequana, Windtree Therapuetics, Rocket Pharma, and Whiteswell, and has served on clinical endpoint committees/data safety monitoring boards for Amgen, Merck, Medtronic, EBR Systems, V-Wave, LivaNova. Dr. Desai works under contract with the Centers for Medicare and Medicaid Services to develop and maintain performance measures used for public reporting and pay for performance programs. He reports research grants and consulting for Amarin, Amgen, Astra Zeneca, Bayer, Boehringer Ingelheim, Bristol Myers Squibb, Cytokinetics, Novartis, SCPharmaceuticals, and Vifor. Dr. Januzzi is a Trustee of the American College of Cardiology, a Director at Jana Care and Imbria, has received grant support from Abbott, Applied Therapeutics, HeartFlow Inc, Innolife and Roche Diagnostics, consulting income from Abbott, AstraZeneca, Beckman-Coulter, Boehringer-Ingelheim, Janssen, Novartis, Merck, Quidel, Roche Diagnostics and Siemens, and participates in clinical endpoint committees/data safety monitoring boards for Abbott, AbbVie, Bayer, CVRx, Pfizer and Takeda. Dr. Ahmad has received consulting fees from Sanofi, Amgen, and Cytokinetics; and has received research funding from Boehringer Ingelheim, AstraZeneca, Cytokinetics, and Relypsa. All authors report that no relationships are relevant to the contents of this manuscript.
Abbreviations:
- NT-proBNP
N-terminal pro B-type natriuretic peptide
- HF
Heart Failure
- GUIDE-IT
Guiding Evidence Based Therapy Using Biomarker Intensified Treatment in Heart Failure
- IQR
Interquartile range
- NYHA
New York Heart Association
- ICD
Implantable Cardioverter-Defibrillator
- GDMT
Guideline-directed Medical Therapy
- HFrEF
Heart Failure with Reduced Ejection Fraction
Footnotes
Study Registration: NCT01685840
REFERENCES
- 1.Heidenreich PA, Bozkurt B, Aguilar D, et al. 2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure: Executive Summary: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. J Am Coll Cardiol. 2022. May 3;79(17):1757–1780. doi: 10.1016/j.jacc.2021.12.011. [DOI] [PubMed] [Google Scholar]
- 2.Writing Committee, Maddox TM, Januzzi JL Jr, Allen LA, et al. 2021 Update to the 2017 ACC Expert Consensus Decision Pathway for Optimization of Heart Failure Treatment: Answers to 10 Pivotal Issues About Heart Failure With Reduced Ejection Fraction: A Report of the American College of Cardiology Solution Set Oversight Committee. J Am Coll Cardiol. 2021. 16;77(6):772–810. doi: 10.1016/j.jacc.2020.11.022. [DOI] [PubMed] [Google Scholar]
- 3.Gaggin HK, Truong QA, Rehman SU, et al. Characterization and prediction of natriuretic peptide “nonresponse” during heart failure management: results from the ProBNP Outpatient Tailored Chronic Heart Failure (PROTECT) and the NT-proBNP-Assisted Treatment to Lessen Serial Cardiac Readmissions and Death (BATTLESCARRED) study. Congest Heart Fail. 2013;19(3):135–42. doi: 10.1111/chf.12016. [DOI] [PubMed] [Google Scholar]
- 4.Weiner RB, Baggish AL, Chen-Tournoux A, et al. Improvement in structural and functional echocardiographic parameters during chronic heart failure therapy guided by natriuretic peptides: mechanistic insights from the ProBNP Outpatient Tailored Chronic Heart Failure (PROTECT) study. Eur J Heart Fail. 2013;15(3):342–351. doi: 10.1093/eurjhf/hfs180. [DOI] [PubMed] [Google Scholar]
- 5.Daubert MA, Adams K, Yow E, et al. NT-proBNP goal achievement is associated with significant reverse remodeling and improved clinical outcomes in HFrEF. J Am Coll Cardiol HF 2019;7(2):158–168. doi: 10.1016/j.jchf.2018.10.014. [DOI] [PubMed] [Google Scholar]
- 6.Masson S, Latini R, Anand IS, et al. Prognostic value of changes in N-terminal pro-brain natriuretic peptide in Val-HeFT (Valsartan Heart Failure Trial). J Am Coll Cardiol 2008; 52: 997–1003. doi: 10.1016/j.jacc.2008.04.069. [DOI] [PubMed] [Google Scholar]
- 7.Morrow DA, de Lemos JA, Blazing MA, Sabatine MS, Murphy SA, Jarolim P, White HD, Fox KA, Califf RM, Braunwald E; Investigators. Prognostic value of serial B-type natriuretic peptide testing during follow-up of patients with unstable coronary artery disease. JAMA. 2005. Dec 14;294(22):2866–71. doi: 10.1001/jama.294.22.2866. [DOI] [PubMed] [Google Scholar]
- 8.Jia X, Al Rifai M, Hoogeveen R, Echouffo-Tcheugui JB, Shah AM, Ndumele CE, Virani SS, Bozkurt B, Selvin E, Ballantyne CM, Nambi V. Association of Long-term Change in N-Terminal Pro-B-Type Natriuretic Peptide With Incident Heart Failure and Death. JAMA Cardiol. 2023. Mar 1;8(3):222–230. doi: 10.1001/jamacardio.2022.5309. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Khan MS, Siddiqi TJ, Usman MS, Sreenivasan J, Fugar S, Riaz H, Murad MH, Mookadam F, Figueredo VM. Does natriuretic peptide monitoring improve outcomes in heart failure patients? A systematic review and meta-analysis. Int J Cardiol. 2018. Jul 15;263:80–87. doi: 10.1016/j.ijcard.2018.04.049. [DOI] [PubMed] [Google Scholar]
- 10.Felker GM, Anstrom KJ, Adams KF, 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):713–720. doi: 10.1001/jama.2017.10565. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Felker GM, Ahmad T, Anstrom KJ, et al. Rationale and design of the GUIDE-IT study: Guiding Evidence Based Therapy Using Biomarker Intensified Treatment in Heart Failure. J Am Coll Cardiol HF. 2014; 2(5):457–65. doi: 10.1016/j.jchf.2014.05.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Rizopoulos D. JM: An R Package for the Joint Modelling of Longitudinal and Time-to-Event Data. Journal of Statistical Software 2010; 35(9), 1–33. 10.18637/jss.v035.i0921603108 [DOI] [Google Scholar]
- 13.van Vark LC, Lesman-Leegte I, Baart SJ, et al. Prognostic Value of Serial ST2 Measurements in Patients With Acute Heart Failure. J Am Coll Cardiol. 2017; 70(19): 2378–2388. doi: 10.1016/j.jacc.2017.09.026. [DOI] [PubMed] [Google Scholar]
- 14.O’Connor C, Fiuzat M, Mulder H, et al. Clinical factors related to morbidity and mortality in high-risk heart failure patients: the GUIDE-IT predictive model and risk score. Eur J Heart Fail. 2019; 21(6):770–778. doi: 10.1002/ejhf.1450. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Wolsk E, Claggett B, Diaz R, et al. Risk Estimates of Imminent Cardiovascular Death and Heart Failure Hospitalization Are Improved Using Serial Natriuretic Peptide Measurements in Patients With Coronary Artery Disease and Type 2 Diabetes. J Am Heart Assoc. 2022; 11(8):e021327. doi: 10.1161/JAHA.121.021327. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Januzzi JL Jr, Ahmad T, Mulder H, et al. Natriuretic Peptide Response and Outcomes in Chronic Heart Failure With Reduced Ejection Fraction. J Am Coll Cardiol. 2019; 74(9):1205–1217. doi: 10.1016/j.jacc.2019.06.055. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Januzzi JL Jr, Prescott MF, Butler J, et al. Association of Change in N-Terminal Pro-B-Type Natriuretic Peptide Following Initiation of Sacubitril-Valsartan Treatment With Cardiac Structure and Function in Patients With Heart Failure With Reduced Ejection Fraction. JAMA. 2019; 322(11):1085–1095. doi: 10.1001/jama.2019.12821. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Ibrahim NE, Januzzi JL Jr. The Future of Biomarker-Guided Therapy for Heart Failure After the Guiding Evidence-Based Therapy Using Biomarker Intensified Treatment in Heart Failure (GUIDE-IT) Study. Curr Heart Fail Rep. 2018; 15(2):37–43. doi: 10.1007/s11897-018-0381-0. [DOI] [PubMed] [Google Scholar]
- 19.Murphy SP, Prescott MF, Maisel AS, et al. Association Between Angiotensin Receptor-Neprilysin Inhibition, Cardiovascular Biomarkers, and Cardiac Remodeling in Heart Failure With Reduced Ejection Fraction. Circ Heart Fail. 2021; 14(6):e008410. doi: 10.1161/CIRCHEARTFAILURE.120.008410. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.