J A C C : H E A R T F A I L U R E V O L . 7 , N O . 2 , 2 0 1 9

ª 2 0 1 9 T H E A M E R I C A N C O L L E G E O F C A R D I O L O G Y F O U N D A T I O N .

P U B L I S H E D B Y E L S E V I E R . A L L R I G H T S R E S E R V E D .

NT-proBNP Goal AchievementIs Associated With SignificantReverse Remodeling and ImprovedClinical Outcomes in HFrEF

Melissa A. Daubert, MD,a,b Kirkwood Adams, MD,c Eric Yow, MS,b Huiman X. Barnhart, PHD,a,b

Pamela S. Douglas, MD,a,b Susan Rimmer, RDCS,b Casey Norris, MS,b Lawton Cooper, MD, MPH,d Eric Leifer, PHD,d

Patrice Desvigne-Nickens, MD,d Kevin Anstrom, PHD,a,b Mona Fiuzat, PHARMD,a Justin Ezekowitz, MBBCH, MSC,e

Daniel B. Mark, MD, MPH,a,b Christopher M. O’Connor, MD,a,f James Januzzi, MD,g G. Michael Felker, MDa,b

ABSTRACT

ISS

FrocUeU

ch

Ro

wa

su

an

Co

Ins

co

Ma

OBJECTIVES This study aims to assess the association between biomarker-guided therapy and left ventricular (LV)

remodeling.

BACKGROUND In patients with heart failure with reduced ejection fraction (HFrEF), it is unclear if lowering natriuretic

peptides reflects structural and functional changes in the heart. This study aims to assess the association between

biomarker-guided therapy and left ventricular (LV) remodeling.

METHODS The GUIDE-IT (Guiding Evidence Based Therapy Using Biomarker Intensified Treatment in Heart Failure) Echo

Substudy was a multicenter study that randomized 268 patients with HFrEF (EF #40%) to either pro-B-type natriuretic

peptide (NT-proBNP)-guided treatment or usual care. Echocardiograms were performed at baseline and 12 months in 124

patients. Remodeling indices and clinical outcomes were compared between treatment arms and by achievement of the

NT-proBNP goal of <1,000 pg/ml at 12 months.

RESULTS At 12 months, the changes in EF and LV volumes were similar between the biomarker-guided and usual care

arms with no difference in clinical outcomes; however, lowering NT-proBNP to <1,000 pg/ml, regardless of treatment

strategy, was associated with a significantly greater increase in EF compared with those not reaching goal (9.9 � 8.8%

vs. 2.9 � 7.9%; p < 0.001) and lower LV volumes. The extent of reverse remodeling correlated with the change in NT-

proBNP: a decrease of 1,000 pg/ml was associated with an increase in EF of 6.7% and a reduction in systolic and diastolic

volumes of 17.3 ml/m2 and 15.7 ml/m2, respectively. Adverse events were significantly lower among patients achieving

the NT-proBNP goal (p < 0.001).

CONCLUSIONS Among patients with HFrEF, lowering NT-proBNP to <1,000 pg/ml by 12 months was associated with

significant reverse remodeling and improved outcomes. A greater reduction in NT-proBNP was associated with more

extensive reverse remodeling. (Guiding Evidence Based Therapy Using Biomarker Intensified Treatment [GUIDE-IT];

NCT01685840) (J Am Coll Cardiol HF 2019;7:158–68) © 2019 the American College of Cardiology Foundation.

Published by Elsevier. All rights reserved.

N 2213-1779/$36.00 https://doi.org/10.1016/j.jchf.2018.10.014

m the aDuke University Medical Center, Durham, North Carolina; bDuke Clinical Research Institute, Durham, North Carolina;

niversity of North Carolina, Chapel Hill, North Carolina; dNational Heart, Lung, and Blood Institute, Bethesda, Maryland;

niversity of Alberta, Edmonton, Alberta, Canada; fInova Heart and Vascular Institute, Falls Church, Virginia; and the gMassa-

usetts General Hospital, Boston, Massachusetts. Duke Clinical Research Institute was the recipient of research grants from

che Diagnostics. The GUIDE-IT Trial was funded by the National Heart, Lung, and Blood Institute. The GUIDE-IT Echo Substudy

s funded by Roche Diagnostics. Dr. Daubert has received grant support from Roche Diagnostics. Dr. Douglas has received grant

pport from the National Institutes of Health. Drs. Cooper, Leifer, and Desvigne-Nickens are faculty of the National Heart, Lung,

d Blood Institute; and were members of the Guiding Evidence Based Therapy Using Biomarker Intensified Treatment Steering

mmittee. Drs. Adams, Anstrom, Fiuzat, O’Connor, Mark, Januzzi, and Felker have received grant support from the National

titutes of Health and Roche Diagnostics. All other authors have reported that they have no relationships relevant to the

ntents of this paper to disclose. Barry Greenberg, MD, served as Guest Editor for this paper.

nuscript received July 27, 2018; revised manuscript received October 5, 2018, accepted October 12, 2018.

J A C C : H E A R T F A I L U R E V O L . 7 , N O . 2 , 2 0 1 9 Daubert et al.F E B R U A R Y 2 0 1 9 : 1 5 8 – 6 8 NT-proBNP Goal and Reverse Remodeling

159

AB BR E V I A T I O N S

AND ACRONYM S

CI = confidence interval

EDVi = end-diastolic volume

index

ESVi = end-systolic volume

index

HF = heart failure

HFrEF = heart failure with

reduced ejection fraction

IQR = interquartile range

KM = Kaplan-Meier

LV = left ventricle

NT-proBNP = amino-terminal

pro-B-type natriuretic peptide

T he prevalence of heart failure (HF) isincreasing with the aging population andconstitutes an alarming public health prob-

lem (1). Although evidence-based HF therapy hasimproved the prognosis among patients with HFwith reduced ejection fraction (HFrEF), the rate ofadverse events remains high. The natriuretic pep-tides, specifically B-type natriuretic peptide andamino-terminal pro-B-type natriuretic peptide (NT-proBNP), are released from the myocardium inresponse to hemodynamic stress and are objectivebiomarkers that provide important prognostic infor-mation in patients with HFrEF. Elevated concentra-tions of these biomarkers are powerfully predictiveof adverse events (2–5). Conversely, a decrease in natri-uretic peptide concentration over time in patientswith HFrEF has been associated with improved clin-ical outcomes (6–8). Despite these associations, it isunclear if changes in natriuretic peptides reflect struc-tural and functional changes in the heart.

A potential link between lowering NT-proBNP andreverse left ventricular (LV) remodeling was firstdemonstrated in the PROTECT (Pro-BNP OutpatientTailored Chronic Heart Failure Therapy) study (2,9).In PROTECT, patients with HFrEF randomized toreceive biomarker-guided HF care had significantlylower NT-proBNP concentrations, a greater increasein EF, and smaller LV volumes compared with usualHF care (2,9). This single-center experience with arelatively short duration of follow-up was insuffi-cient, however, to delineate the complex relationshipbetween change in natriuretic peptide concentration,cardiac remodeling, and clinical events; therefore,the large, multicenter GUIDE-IT (Guiding EvidenceBased Therapy Using Biomarker Intensified Treat-ment in Heart Failure) trial and EchocardiographicSubstudy were designed to prospectively study theinterrelationship of these 3 key elements. We hy-pothesized that patients with HFrEF treated withbiomarker-guided therapy would demonstrate morereverse remodeling and have fewer adverse eventsthan patients treated with usual care. The GUIDE-ITEchocardiographic Substudy provided a unique op-portunity to evaluate the effect of biomarker-guidedtherapy on cardiac structure and function and assessthe association between NT-proBNP reduction,reverse remodeling, and clinical outcomes.

METHODS

The study design for the main GUIDE-IT trial has beenpreviously published (10). Briefly, GUIDE-IT was amulticenter trial that randomized patients withHFrEF (EF #40%) to either NT-proBNP–guided HF

treatment strategy or usual guideline-directed HF care. The objective was tocompare a strategy of medical therapy titra-tion aimed at achieving and maintaining anNT-proBNP target of <1,000 pg/ml to usualcare in patients with HFrEF. The study wasdiscontinued early after 894 patients wereenrolled (81% of planned) because biomarker-guided treatment strategy was not moreeffective than usual care in improving out-comes (11).

The GUIDE-IT Echo Substudy was a pre-specified substudy designed to understandthe influence of biomarker-guided therapy oncardiac structure and function compared withusual care among chronic patients with

HFrEF with at least moderate systolic dysfunction(EF #40%) and an NT-proBNP concentration >2,000pg/ml within 30 days before randomization. The EchoSubstudy was conducted in 18 sites in the UnitedStates and Canada. Patients were eligible if they metthe same inclusion and exclusion criteria as the mainGUIDE-IT trial and agreed to echocardiography atbaseline and 12 months (10). All participants providedwritten informed consent for participation in theEcho Substudy, which was obtained concurrentlywith the main trial informed consent. All patientswere consented and enrolled before main trialrandomization to preserve the randomization schemain the Echo Substudy. Patients were followed for upto 24 months after randomization. The Echo Substudywas approved by the institutional review board ateach study site. The GUIDE-IT study was funded bythe National Heart, Lung, and Blood Institute(NCT01685840); the Echo Substudy was funded byRoche Diagnostics Inc. (Rotkreutz, Switzerland).

STUDY ENDPOINTS. The primary endpoint of theGUIDE-IT Echo Substudy was change in LV end-systolic volume index (ESVi) from baseline to 12months among patients receiving NT-proBNP–guidedtreatment compared with patients treated with usualcare. Secondary endpoints were change in: end-diastolic volume index (EDVi), EF, diastolic parame-ters, right heart function, and valvular regurgitation.Additionally, remodeling indices were assessedamong patients achieving NT-proBNP goal of <1,000pg/ml at 12 months compared with those notachieving the NT-proBNP goal ($1,000 pg/ml),regardless of treatment strategy. NT-proBNP con-centration was assessed using Roche Elecsys assay.Patients with missing or unmeasurable baseline or 12-month echocardiograms were excluded from quanti-tative analyses. The composite outcome of death or

Daubert et al. J A C C : H E A R T F A I L U R E V O L . 7 , N O . 2 , 2 0 1 9

NT-proBNP Goal and Reverse Remodeling F E B R U A R Y 2 0 1 9 : 1 5 8 – 6 8

160

HF hospitalization after 12 months was comparedbetween treatment groups and by achievement of NT-proBNP goal.

ECHOCARDIOGRAPHIC ANALYSIS. The GUIDE-ITEcho Substudy protocol required 2-dimensionalechocardiographic imaging with spectral and colorDoppler acquisitions from standard parasternal andapical views. The echocardiographic parametersanalyzed were: EF, LV volumes, LV mass, left atrialvolume, ratio of early transmitral peak velocity toearly diastolic peak annular velocity (E/e0), valveregurgitation, right ventricular systolic function asassessed by tricuspid annular plane systolic excursion,right ventricular systolic pressure, stroke volume,cardiac output, and global longitudinal strain. Mea-surements were indexed for body surface area whenappropriate. Cardiac volumes and EF were quantifiedby the biplane Simpson’s method using DigisonicsDigiview software (version 3.8.4, Houston, Texas). LVlongitudinal strain was obtained in the apical viewsand averaged to determine a global longitudinal value.Strain calculations were performed using TOMTECsoftware version 4.6.2 (Unterschleissheim, Germany).

Echocardiographic analysis was performed in anindependent core laboratory at Duke ClinicalResearch Institute (Durham, North Carolina). Allechocardiographic interpretations were performed byblinded readers in accordance with best practices andcurrent guidelines (12–15). There was high inter-reader reproducibility across readers (intraclass cor-relation coefficient: 0.95 for EDV, 0.98 for ESV, and0.99 for EF).

STATISTICAL ANALYSIS. The sample size of the EchoSubstudy was selected to detect a clinically mean-ingful change in ESVi of 10 ml/m2 between the treat-ment groups at 12 months. With 64 patients in eachtreatment arm with quantitative echocardiography atbaseline and 12 months, there would be at least 80%power at the 0.05 level of statistical significance todetect a difference of 10 ml/m2 between treatmentgroups assuming a SD of 20 ml/m2. Accounting for a20% loss to follow-up and an unmeasurable rate forLV volumes of 45%, a total of 290 patients (145 pertreatment strategy) was planned.

Continuous and categorical variables werecompared between groups using Wilcoxon rank sumtest and Fisher exact test, respectively. Continuousvariables are presented as mean � SD or median(interquartile range [IQR]), as appropriate; categoricalvariables are shown as counts and percentages.Treatment group analyses were made on the principleof intention-to-treat. Univariable and multivariablelogistic regression modeling were used to determine

factors independently associated with attainment ofNT-proBNP goal. The multivariable model includedthe following variables: age, sex, coronary arterydisease, ischemic cardiomyopathy, natural log ofbaseline NT-proBNP, and EF at baseline. The associ-ation between the degree of LV remodeling and NT-proBNP change was initially evaluated with locallyestimated scatterplot smoothing plots. Visual in-spection of this exploratory analysis suggested theuse of a 2-piece linear spline model. Transformationanalysis was performed to determine the optimal NT-proBNP cutpoint for EF, ESVi, and EDVi, defined asthe 12-month NT-proBNP concentration below whichthere was a significant increase in EF and decrease inLV volumes, respectively. Clinical outcomes after 12months were evaluated with a landmark analysis thatincluded time to death or HF hospitalization from 12months to study end. For the time-to-event analysis,Kaplan-Meier (KM) cumulative risk curves were usedto estimate the event rate; log-rank statistics wereused for comparisons between groups. The thresholdfor statistical significance was 2-sided with a type Ierror rate of 0.05. There was no adjustment per-formed for multiple comparisons. Analyses wereperformed using SAS version 9.4 (SAS Institute Inc.,Cary, North Carolina).

RESULTS

Of the 894 patients in the main GUIDE-IT trial, 268(92% of planned) were enrolled in the Echo Substudybetween August 2013 and July 2016. Of these, 135patients completed the study before early trialcessation. Mean follow-up was 18.7 months. Echo-cardiographic image quality was excellent; as a result,the unmeasurable rate for quantitative LV volumesand EF was only 12%, which was markedly lower thanprojected for this chronic HF population.

TREATMENT GROUP ANALYSIS. Among the 268 pa-tients in the Echo Substudy, 141 were randomizedinto the NT-proBNP-guided arm and 127 into theusual care arm. After excluding patients who diedbefore 12 months, were lost to follow-up, or lacked a12-month echocardiogram because of early studytermination (Online Table 1), there were 67 patients inthe biomarker-guided arm and 57 patients in theusual care arm that had paired baseline and 12-monthechocardiograms with quantitative measurements(Figure 1A). Baseline clinical characteristics were notsignificantly different among patients with HFrEF ineither of the treatment groups, except for a greaterproportion of patients with prior myocardial infarc-tion in the usual care arm (Table 1). The NT-proBNPvalues were similar at baseline. The treatment

FIGURE 1 Consolidated Standards of Reporting Trials Diagrams

(A) Randomized treatment group analysis. (B) NT-proBNP goal analysis. Echo ¼ echocardiogram; f/u ¼ follow-up; GUIDE-IT ¼ Guiding

Evidence Based Therapy Using Biomarker Intensified Treatment; NT-proBNP ¼ pro-B-type natriuretic peptide.

J A C C : H E A R T F A I L U R E V O L . 7 , N O . 2 , 2 0 1 9 Daubert et al.F E B R U A R Y 2 0 1 9 : 1 5 8 – 6 8 NT-proBNP Goal and Reverse Remodeling

161

groups were also well balanced with respect to thebaseline echocardiography results (Table 2), whichdemonstrated markedly dilated LV volumes andseverely reduced systolic function: EF ¼ 29.0 � 9.7%in the NT-proBNP–guided arm and 27.7 � 10.7% in theusual care arm (p ¼ 0.37).

At 12 months, median NT-proBNP concentrationsin the biomarker-guided arm were not significantlydifferent from the usual care arm: 1,072 (IQR: 440 to3,067) pg/ml versus 1,286 (IQR: 504 to 2,923) pg/ml;p ¼ 0.73. The primary outcome of ESVi decreased by15.2 � 20.4 ml/m2 in the biomarker-guided armcompared with a decrease of 17.4 � 28.4 ml/m2 in theusual care arm (p ¼ 0.82). The decreases in EDVi weresimilar (Table 2). Improvement in EF was not signifi-cantly different between the treatment groups: þ6.0� 8.0% in the NT-proBNP–guided arm versus þ6.6 �10.5% in the usual care arm (p ¼ 0.75).

NT-proBNP GOAL ANALYSIS. In the NT-proBNP goalanalysis, we pooled the 2 treatment groups, consid-ering them as a function of achieved NT-proBNPconcentration at 12 months. There were 116 patientswith paired echocardiograms and a 12-month NT-proBNP value. Of these, 52 patients achieved theNT-proBNP goal of <1,000 pg/ml at 12 months,whereas 64 had an NT-proBNP value of $1,000 pg/mlat 12 months (Figure 1B). Age, sex, race, and baselineNT-proBNP values were not significantly differentbetween groups (Table 3); however, patients notachieving NT-proBNP goal at 12 months were signifi-cantly more likely to have ischemic heart disease.After multivariable adjustment, patients withischemic cardiomyopathy remained significantlymore likely to have an NT-proBNP $1,000 pg/ml at 12months (odds ratio: 4.2; 95% confidence interval [CI]:1.1 to 16.0; p ¼ 0.04). A higher baseline NT-proBNP

TABLE 1 Baseline Characteristics of the GUIDE-IT Echo Substudy by

Treatment Group

NT-proBNP Guided(n ¼ 67)

Usual Care(n ¼ 57) p Value

Age, yrs 58.1 � 13.4 62.7 � 10.4 0.08

Male 48 (71.6) 36 (63.2) 0.31

Race, black 32 (47.8) 19 (34.5) 0.14

Body mass index, kg/m2 29.4 � 7.1 31.0 � 7.6 0.13

Hypertension 49 (73.1) 45 (78.9) 0.45

Hyperlipidemia 35 (52.2) 35 (61.4) 0.31

Diabetes mellitus 35 (52.2) 23 (40.4) 0.19

Smoking 22 (32.8) 21 (36.8) 0.64

Coronary artery disease 30 (44.8) 32 (56.1) 0.21

Myocardial infarction 17 (25.4) 26 (45.6) <0.05

Ischemic cardiomyopathy 27 (40.3) 28 (49.1) 0.32

NYHA functional class III/IV 24 (35.8) 18 (31.6) 0.86

NT-proBNP, pg/ml 2,182 (1,320–4,252) 2,597 (1,405–4,362) 0.81

Atrial fibrillation 11 (16.4) 11 (19.3) 0.68

Renal disease 24 (35.8) 17 (29.8) 0.48

Peripheral artery disease 7 (10.4) 9 (15.8) 0.38

Values are mean � SD, n (%), or median (interquartile range).GUIDE-IT ¼ Guiding Evidence Based Therapy Using Biomarker Intensified Treatment; NYHA ¼

New York Heart Association; NT-proBNP ¼ pro-B-type natriuretic peptide.

Daubert et al. J A C C : H E A R T F A I L U R E V O L . 7 , N O . 2 , 2 0 1 9

NT-proBNP Goal and Reverse Remodeling F E B R U A R Y 2 0 1 9 : 1 5 8 – 6 8

162

was also independently associated with the inabilityto achieve the NT-proBNP goal at 12 months (oddsratio: 2.5; 95% CI: 1.4 to 4.3; p ¼ 0.002). There was nosignificant difference in baseline EF or LV volumesbetween achieved NT-proBNP groups (Table 4).

At 12 months, the median NT-proBNP in the goal-achieving group was 6-fold lower than the nongoalgroup: 424 (IQR: 180 to 650) pg/ml versus 2,669 (IQR:1503 to 5184) pg/ml. Compared with those notachieving goal, patients achieving the NT-proBNPgoal of <1,000 pg/ml at 12 months had significantlygreater reverse LV remodeling (Figure 2). LV ESVi andEDVi were significantly more reduced among patientsachieving the NT-proBNP goal than those who did not(Table 4). Correspondingly, patients achieving theNT-proBNP goal had an absolute increase in EF of 9.9� 8.8% compared with 2.9 � 7.9% among those notreaching goal (p < 0.001). Global longitudinal strainwas also significantly improved among patientsachieving goal (�2.8 � 3.3) compared with those whodid not (�0.6 � 3.0) (p ¼ 0.003). There was signifi-cantly less moderate or severe mitral regurgitationamong patients achieving NT-proBNP goal at 12months: 5.8% versus 18.8%; p ¼ 0.04. There was alsoa trend toward improved LV diastolic function andright heart function in patients achieving goalcompared to those who did not.

NT-proBNP AND REVERSE LV REMODELING. At 12months, NT-proBNP concentrations correlated with

LV volumes and EF (Figure 3). Notably, locally esti-mated scatterplot smoothing curves showed adistinct inflection point at approximately 1,000 pg/mlat which the decreases in LV volumes and the in-crease in EF were greater below this threshold thanabove it (Online Figure 1). Transformation analysesdetermined that the optimal 12-month cutpoint forNT-proBNP was 1,028 pg/ml for EF, 941 pg/ml for ESVi,and 1,286 pg/ml for EDVi, thus approximating theprespecified NT-proBNP target goal of 1,000 pg/ml.These results were confirmed with 2-piece linearspline analyses that revealed significantly steeperslopes below the NT-proBNP cutpoints for LV volumesand EF than above (Figure 3).

Of importance, the extent of reverse LV remodelingcorrelated with changes in NT-proBNP. Linear splineregression modeling demonstrated that the greaterthe reduction in NT-proBNP, the more extensive thereverse LV remodeling. Specifically, an NT-proBNPdecrease of 1,000 pg/ml corresponded to an absoluteincrease in EF of 6.7% and a reduction in ESVi andEDVi of 17.3 and 15.7 ml/m2, respectively (Figure 4).

CLINICAL OUTCOMES. The composite outcome ofdeath or HF hospitalization after 12 months was notsignificantly different between treatment arms andoccurred in 9 patients (13%) in the biomarker-guidedgroup and 7 patients (12%) in the usual care groupwith KM rates of 20% (95% CI: 0.10 to 0.37) and 18%(95% CI: 0.09 to 0.36), respectively (p ¼ 0.77). Incontrast, patients achieving an NT-proBNP goalof <1,000 pg/ml at 12 months had substantially betterclinical outcomes after 12 months. The compositeendpoint of death or HF hospitalization after 12months was significantly lower among patientsachieving NT-proBNP goal (KM rate of 0%) comparedwith patients who did not achieve goal at 12 months(KM rate of 30%, 95% CI: 0.19 to 0.46); p < 0.001(Table 5). Furthermore, median NT-proBNP concen-trations after 12 months continued to decrease in thegoal-achieving group compared to an increase in NT-proBNP among those patients who did not achievegoal: 310 (IQR: 134 to 654) pg/ml versus 3,038 (IQR:1,382 to 4,711) pg/ml, respectively.

DISCUSSION

The GUIDE-IT Echo Substudy evaluated the impact ofbiomarker-guided therapy on cardiac structure andfunction compared with usual care among HFrEFpatients. Reduction in LV volumes and improvementin EF were not significantly different between treat-ment strategies and, like the main GUIDE-IT trial,there was no difference in clinical outcomes betweenstudy arms. When evaluated by achievement of

TABLE 2 Baseline Echocardiography and Change at 12 Months by Treatment Group

Baseline Change at 12 Months

NT-proBNP Guided(n ¼ 67)

Usual Care(n ¼ 57) p Value

NT-proBNP Guided(n ¼ 67)

Usual Care(n ¼ 57) p Value

ESVi, ml/m2 87.5 � 38.5 90.7 � 34.9 0.39 �15.2 � 20.4 �17.4 � 28.4 0.82

EDVi, ml/m2 119.6 � 38.6 122.7 � 36.8 0.46 �13.3 � 20.6 �15.9 � 29.8 0.95

Ejection fraction, % 29.0 � 9.7 27.7 � 10.7 0.37 þ6.0 � 8.0 þ6.6 � 10.5 0.75

Cardiac index, l/min/m2 2.0 � 0.7 1.9 � 0.6 0.97 þ0.1 � 0.6 0.0 � 0.6 0.56

E/e0 21.0 � 10.2 21.0 � 9.6 0.74 �0.4 � 10.9 �1.4 � 8.1 0.88

LA volume index, ml/m2 46.7 � 15.2 48.8 � 16.4 0.46 �4.9 � 12.1 �3.1 � 13.1 0.73

RA area, ml 21.5 � 7.8 21.2 � 6.8 0.79 �1.6 � 4.2 �1.4 � 6.2 0.95

TAPSE, cm 1.7 � 0.5 1.7 � 0.6 0.70 0.0 � 0.6 þ0.1 � 0.6 0.39

RV systolic pressure, mm Hg 37.4 � 14.3 36.4 � 10.8 0.97 �1.7 � 13.0 �3.2 � 11.3 0.30

Global LV strain, % �8.7 � 3.4 �9.7 � 3.8 0.20 �2.4 � 3.6 �1.1 � 3.6 0.19

Values are mean � SD.EDVi ¼ end-diastolic volume index; E/e0 ¼ ratio of early transmitral peak velocity to early diastolic peak annular velocity; ESVi ¼ end-systolic volume index; LA ¼left atrial;

LV ¼ left ventricular; RA ¼ right atrial; RV ¼ right ventricular; TAPSE ¼ tricuspid annular plane systolic excursion; other abbreviations as in Table 1.

TABLE 3 Baseline Characteristics of Patients by Achievement of NT-proBNP Goal at

12 Months

NT-proBNP <1,000 pg/mlat 12 Months (n ¼ 52)

NT-proBNP $1,000 pg/mlat 12 Months (n ¼ 64) p Value

Age, yrs 57.7 � 12.7 63.1 � 10.6 0.22

Male 34 (64.2) 46 (71.9) 0.45

Race, black 18 (34.6) 27 (42.9) 0.41

Body mass index, kg/m2 30.5 � 7.0 30.2 � 7.9 0.15

Hypertension 35 (67.3) 53 (82.8) 0.05

Hyperlipidemia 23 (44.2) 44 (68.8) <0.05

Diabetes mellitus 19 (36.5) 34 (53.1) 0.08

Smoking 15 (28.3) 25 (39.1) 0.25

Coronary artery disease 18 (34.6) 42 (65.6) <0.05

Myocardial infarction 12 (23.1) 29 (45.3) <0.05

Ischemic cardiomyopathy 12 (23.1) 40 (62.5) <0.05

NYHA functional class III/IV 16 (30.9) 24 (37.5) 0.81

NT-proBNP, pg/ml 1,760 (807–2,603) 3,579 (1,973–5,777) 0.98

Atrial fibrillation 4 (7.7) 18 (28.1) <0.05

Renal disease 9 (17.3) 31 (48.4) <0.05

Peripheral artery disease 2 (3.8) 13 (20.3) <0.05

Biomarker-guided arm 30 (57.7) 33 (51.6) 0.51

Values are mean � SD, n (%), or median (interquartile range).Abbreviations as in Tables 1 and 2.

J A C C : H E A R T F A I L U R E V O L . 7 , N O . 2 , 2 0 1 9 Daubert et al.F E B R U A R Y 2 0 1 9 : 1 5 8 – 6 8 NT-proBNP Goal and Reverse Remodeling

163

NT-proBNP goal <1,000 pg/ml at 12 months, however,irrespective of treatment strategy, those patients whoattained the NT-proBNP goal had significantly greaterreverse remodeling and better clinical outcomes thanthose not achieving goal. This study also demon-strated an important link between lowering NT-proBNP and improvements in cardiac structure andfunction such that a greater reduction in NT-proBNPover time was associated with more extensivereverse remodeling.

Previous studies have shown that biomarker-guided HF treatment is associated with better out-comes than usual care (2,16,17). In particular, thepreviously mentioned PROTECT study randomized151 patients to receive NT-proBNP–guided therapy orusual HF care and found that, after a mean follow-upof 10 months, biomarker-guided therapy resulted insignificantly lower NT-proBNP concentrations andreduced adverse events compared with usual care (2).Similarly, a meta-analysis of 11 trials showed areduction in all-cause mortality with natriureticpeptide-guided therapy compared with usual care(hazard ratio: 0.62 [95% CI: 0.45 to 0.86]; p ¼ 0.004)(17). One the basis of this foundational work, theGUIDE-IT trial was designed to be the largest study tocompare biomarker-guided HF management withusual guideline-directed HF care in patients withHFrEF; however, unlike prior studies, the GUIDE-ITtrial showed no difference in clinical outcomes be-tween these treatment strategies (11).

The GUIDE-IT Echo Substudy may provide uniqueinsight into the lack of treatment effect in the mainGUIDE-IT trial and help explain the discrepant resultswhen compared with prior studies. First, a similardegree of reverse LV remodeling, or more specifically,a similar reduction in LV volumes and increase in EF,

was seen in the biomarker-guided and usual carearms. The extent to which clinical outcomes areinfluenced by LV recovery would be expectedlyequivalent in both groups. Second, both treatmentarms were on guideline-directed HF care, hadfrequent follow-up visits, and nearly equivalent re-ductions in NT-proBNP levels at 12 months; thus, it isnot surprising that both treatment groups showedimprovement in cardiac structure and function at 12months. This could explain why clinical outcomeswere also similar in the Echo Substudy and the mainGUIDE-IT trial. Finally, the findings from the Echo

TABLE 4 Baseline Echocardiography and Change at 12 Months by Achievement of NT-proBNP Goal

Baseline Change at 12 Months

NT-proBNP <1,000 pg/mlat 12 Months (n ¼ 52)

NT-proBNP $1,000 pg/mlat 12 Months (n ¼ 64) p Value

NT-proBNP <1,000 pg/mlat 12 Months (n ¼ 52)

NT-proBNP $1,000 pg/mlat 12 Months (n ¼ 64) p Value

ESVi, ml/m2 83.9 � 33.1 90.6 � 36.8 0.58 �24.6 � 28.8 �8.9 � 17.3 <0.001

EDVi, ml/m2 116.5 � 35.3 122.7 � 37.8 0.57 �22.0 � 31.8 �8.0 � 15.9 0.006

Ejection fraction, % 29.8 � 11.0 27.8 � 8.9 0.44 þ9.9 � 8.8 þ2.9 � 7.9 <0.001

Cardiac index, l/min/m2 2.1 � 0.7 1.9 � 0.7 0.19 þ0.1 � 0.6 �0.1 � 0.6 0.17

E/e’ 18.3 � 7.5 22.3 � 10.4 0.05 �2.8 � 5.7 þ0.8 � 12.0 0.07

LA volume index, ml/m2 40.8 � 13.2 53.1 � 16.4 <0.001 �6.6 � 13.2 �2.1 � 12.4 0.14

Right atrial area, ml 17.9 � 5.2 24.0 � 8.1 <0.001 �2.2 � 3.8 �0.6 � 6.1 0.05

TAPSE, cm 1.8 � 0.5 1.6 � 0.6 0.04 þ0.2 � 0.6 �0.1 � 0.5 0.07

RV systolic pressure, mm Hg 34.4 � 8.6 40.4 � 14.8 0.18 �4.0 � 10.4 �1.9 � 13.2 0.41

Global LV strain, % �10.5 � 4.0 �8.7 � 3.2 0.11 �2.8 � 3.3 �0.6 � 3.0 0.003

Values are mean � SD.Abbreviations as in Table 2.

Daubert et al. J A C C : H E A R T F A I L U R E V O L . 7 , N O . 2 , 2 0 1 9

NT-proBNP Goal and Reverse Remodeling F E B R U A R Y 2 0 1 9 : 1 5 8 – 6 8

164

Substudy suggest that perhaps it is not the treatmentstrategy, but the response to treatment as assessed bychange in NT-proBNP that is important.

Patients with greater reductions in NT-proBNP at12 months had more extensive improvement in LVstructure and function and significantly feweradverse outcomes regardless of treatment strategy. Asimilar association between lowering NT-proBNP and

FIGURE 2 Changes in Left Ventricular Structure and Function by Ac

Patients achieving the NT-proBNP goal of <1,000 pg/ml at 12 months ha

improvement in GLS compared with patients not attaining goal. EDVi ¼ in

end-systolic volume; GLS ¼ global longitudinal strain; other abbreviatio

indices of reverse remodeling was demonstrated inthe echocardiographic subgroup analysis of thePROTECT study (9). The GUIDE-IT Echo Substudyconfirms these findings and is the first study to havesufficient power to quantitate the magnitude ofchange in NT-proBNP and correlate this with indicesof reverse remodeling (i.e., for every decrease of1,000 pg/ml in NT-proBNP there was an incremental

hievement of NT-proBNP Goal at 12 Months

d a greater increase in EF, reduction in left ventricular volumes, and

dexed end-diastolic volume; EF ¼ ejection fraction; ESVi ¼ indexed

ns as in Figure 1.

FIGURE 3 Relationship Among NT-proBNP, Left Ventricular Volumes, and EF at 12 Months

Dot plots of ESVi (top), EDVi (middle), and EF (bottom) compared with NT-proBNP at 12 months. The piece-wise linear regression line

represents the mean. Slope indicates 100 pg/ml change in NT-proBNP concentration. The significance of the slope was tested above and

below 1,000 pg/ml. Abbreviations as in Figures 1 and 2.

J A C C : H E A R T F A I L U R E V O L . 7 , N O . 2 , 2 0 1 9 Daubert et al.F E B R U A R Y 2 0 1 9 : 1 5 8 – 6 8 NT-proBNP Goal and Reverse Remodeling

165

FIGURE 4 NT-proBNP Change and Absolute Change in Ejection Fraction and Left Ventricular Volumes at 12 Months

Regression modeling demonstrates that greater decreases in NT-proBNP correspond to greater increases in ejection fraction (EF) and greater

decreases in left ventricular volumes. EDVi ¼ indexed end-diastolic volume; ESVi ¼ indexed end-systolic volume.

TABLE 5 Clinical Ou

Death/HF hospitalizatio

Death

HF hospitalization

*Log-rank test.

CI ¼ confidence interval

Daubert et al. J A C C : H E A R T F A I L U R E V O L . 7 , N O . 2 , 2 0 1 9

NT-proBNP Goal and Reverse Remodeling F E B R U A R Y 2 0 1 9 : 1 5 8 – 6 8

166

increase in EF and corresponding reduction inindexed LV volumes). These results suggest thatchanges in NT-proBNP not only reflect the responseto HF therapy, but can also serve as a noninvasiveindicator of the dynamic state of cardiac structureand function in patients with HFrEF. Such knowledgemay be leveraged to decide on the timing of echo-cardiography in patients with chronic HFrEF (18). Ithas been suggested that reassessment of LV size andfunction might not be necessary in the context of alow and stable NT-proBNP (4,9,19). The GUIDE-ITEcho Substudy supports such an approach, which islikely to be cost-effective and safe compared withroutine monitoring.

tcomes After 12 Months

NT-proBNP<1,000 pg/ml at

12 Months (n ¼ 52)

NT-proBNP$1,000 pg/ml at

12 Months (n ¼ 64)

p Value*Number ofEvents

KM Rate(95% CI)

Number ofEvents

KM Rate(95% CI)

n 0 0.0 15 0.30 (0.19–0.46) <0.001

0 0.0 7 0.16 (0.08–0.31) 0.04

0 0.0 13 0.27 (0.17–0.43) 0.002

; HF ¼ heart failure; KM ¼ Kaplan-Meier.

In the GUIDE-IT Echo Substudy, patients with anonischemic HF etiology were significantly morelikely to achieve the NT-proBNP goal at 12 monthsand patients who achieved this goal had a greaterpropensity for reverse LV remodeling. In contrast,patients who failed to achieve the NT-proBNP goalwere more likely to have coronary artery disease,atrial fibrillation, and renal dysfunction, and there-fore constituted a sicker population. The failure toattain the NT-proBNP goal may be an epiphenomenonin which the primary issue is actually the greaterburden of myocardial scarring and comorbidities thatare limiting the extent of reverse remodeling andresulting in more adverse events. Further study isneeded to assess the mechanisms underlying thisdifferential response to therapy.

This study also provided mechanistic insight onthe complex interplay among NT-proBNP change,reverse remodeling, and clinical outcomes. In theGUIDE-IT Echo Substudy, there was a mean decreaseof 42 and 48 ml for EDV and ESV, respectively, amongpatients achieving the NT-proBNP goal comparedwith a mean decrease in EDV of 16 ml and ESV of 18 mlin those patients not achieving goal at 12 months. In ameta-analysis of drug- or device-related changes inLV volumes among patients with HFrEF (20), thera-pies that decreased EDV or ESV by a mean of 11 ml

PERSPECTIVES

COMPETENCY IN MEDICAL KNOWLEDGE: The greater the

reduction in NT-proBNP with HF treatment, the more extensive

the improvement in LV structure and function in patients with

HFrEF. Among patients with HFrEF, HF treatment that results in

lowering NT-proBNP to <1,000 pg/ml at 1 year is associated

with reverse remodeling and improved clinical outcomes.

TRANSLATIONAL OUTLOOK: Patients with a nonischemic

cardiomyopathy are more likely to achieve a NT-proBNP

goal <1,000 pg/ml at 1 year than patients with an ischemic

cardiomyopathy. Further study is needed to assess the mecha-

nisms underlying this differential response to therapy.

J A C C : H E A R T F A I L U R E V O L . 7 , N O . 2 , 2 0 1 9 Daubert et al.F E B R U A R Y 2 0 1 9 : 1 5 8 – 6 8 NT-proBNP Goal and Reverse Remodeling

167

were associated with a 65% to 75% likelihood offavorable effects on mortality, with greater decreasesportending better outcomes, as was seen in thisstudy. Furthermore, achieving the NT-proBNP goal inthe GUIDE-IT Echo Substudy was associated with alack of adverse events in the follow-up period. Such asalubrious response was also reported in the PRIDE(NT-proBNP Investigation of Dyspnea in the Emer-gency Department) study, which also had a 0%death rate among patients with a NT-proBNPconcentration <986 pg/ml, and thus was 1 of theoriginal studies establishing 1,000 pg/ml as the NT-proBNP target (21). The GUIDE-IT Echo Substudy ex-tends this work and supports the hypothesis thatpatients with HFrEF on optimized therapy who ach-ieve the NT-proBNP goal of <1,000 pg/ml, can reversethe maladaptive remodeling associated with HF pro-gression and reduce the incidence of adverse clinicalevents.

To our knowledge, the GUIDE-IT Echo Substudy isthe largest study to definitively show that a reductionin NT-proBNP is associated with salutary changes incardiac structure and function in patients withHFrEF. Specifically, these results suggest that the NT-proBNP concentration of 1,000 pg/ml is an optimaltarget for HF therapy because attainment of this goalis associated with significant reverse remodeling andimproved clinical outcomes. Additionally, this is thefirst study to demonstrate the proportionality be-tween NT-proBNP and reverse remodeling, such thata greater reduction in NT-proBNP corresponds tomore extensive reverse remodeling.

STUDY LIMITATIONS. This was a pre-specified, sub-group analyses of the GUIDE-IT Echo Substudy. Therelatively small number of patients was a result ofthe GUIDE-IT trial early termination and therequirement for paired echocardiograms to assess forcardiac remodeling over time. For these reasons,quantitative analysis at baseline and 12 months wasonly available in approximately one-half of the sub-study patients, yet nearly 80% power was main-tained because of the low rate of unmeasurableechocardiograms in this study. The duration of NT-proBNP <1,000 pg/ml was not assessed in thisstudy; however, the evaluation of whether reverseLV remodeling and improved outcomes is associatedwith duration below goal, specific pharmacologictherapies, and/or dose titrations is ongoing. Furtherstudy is needed to evaluate the exact mechanismsunderlying both a decrease in NT-proBNP andreverse remodeling. It is speculated that a common

mechanism may be decreased intracardiac pressures,decreased wall stress, and more efficient energyexpenditure; however, additional studies with inva-sive monitoring and a pathologic correlate would beneeded for confirmation, which may not be possiblein this high-risk HFrEF population. Finally, there wasonly a small number of adverse events, likelybecause of the truncated trial duration; however, theprofound difference in outcomes between patientsachieving the NT-proBNP goal and those who did notsupports that therapy aimed at attaining this targetmay significantly improve clinical outcomes amongpatients with HFrEF.

CONCLUSIONS

Whether patients with HFrEF in the GUIDE-IT EchoSubstudy were treated with biomarker-guided HFtherapy or usual care, changes in cardiac structureand function were similar and there was no signifi-cant difference in clinical outcomes. Lowering NT-proBNP to a goal <1,000 pg/ml, regardless of treat-ment strategy, was associated with a significantlygreater increase in EF, more extensive reductions inLV volumes, and markedly reduced adverse eventscompared with those not achieving goal at 12 months.The improved outcomes associated with achievingthe NT-proBNP goal may be explained, in part, byreverse LV remodeling.

ADDRESS FOR CORRESPONDENCE: Dr. Melissa A.Daubert, Duke University and Duke Clinical ResearchInstitute, 2400 Pratt St, Room 0311 Terrace Level,Durham, North Carolina 27705. E-mail: melissa.daubert@duke.edu.

Daubert et al. J A C C : H E A R T F A I L U R E V O L . 7 , N O . 2 , 2 0 1 9

NT-proBNP Goal and Reverse Remodeling F E B R U A R Y 2 0 1 9 : 1 5 8 – 6 8

168

RE F E RENCE S

1. Yancy CW, Jessup M, Bozkurt B, et al. 2013ACCF/AHA guideline for the management of heartfailure: a report of the American College of Car-diology Foundation/American Heart AssociationTask Force on Practice Guidelines. J AM Coll Car-diol 2013;62:e147–239.

2. Januzzi JL Jr., Rehman SU, Mohammed AA,et al. Use of amino-terminal pro-B-type natriureticpeptide to guide outpatient therapy of patientswith chronic left ventricular systolic dysfunction.J Am Coll Cardiol 2011;58:1881–9.

3. Fonarow GC, Peacock WF, Phillips CO, et al.Admission B-type natriuretic peptide levels and in-hospital mortality in acute decompensated heartfailure. J Am Coll Cardiol 2007;49:1943–50.

4. Cleland JG, McMurray JJ, Kjekshus J, et al.Plasma concentration of amino-terminal pro-brainnatriuretic peptide in chronic heart failure: predic-tion of cardiovascular events and interaction withthe effects of rosuvastatin: a report from CORONA(Controlled Rosuvastatin Multinational Trial inHeart Failure). J Am Coll Cardiol 2009;54:1850–9.

5. Sachdeva A, Horwich TB, Fonarow GC. Comparisonof usefulness of each of five predictors of mortalityand urgent transplantation in patients with advancedheart failure. Am J Cardiol 2010;106:830–5.

6. Anand IS, Fisher LD, Chiang YT, et al. Changesin brain natriuretic peptide and norepinephrineover time and mortality and morbidity in the Val-sartan Heart Failure Trial (Val-HeFT). Circulation2003;107:1278–83.

7. Bettencourt P, Azevedo A, Pimenta J, Frioes F,Ferreira S, Ferreira A. N-terminal-pro-brain natri-uretic peptide predicts outcome after hospitaldischarge in heart failure patients. Circulation2004;110:2168–74.

8. Masson S, Latini R, Anand IS, et al. Prognosticvalue of changes in N-terminal pro-brain natri-uretic peptide in Val-HeFT (Valsartan Heart FailureTrial). J Am Coll Cardiol 2008;52:997–1003.

9. Weiner RB, Baggish AL, Chen-Tournoux A, et al.Improvement in structural and functional

echocardiographic parameters during chronicheart failure therapy guided by natriuretic pep-tides: mechanistic insights from the ProBNPOutpatient Tailored Chronic Heart Failure (PRO-TECT) study. Eur J Heart Fail 2013;15:342–51.

10. Felker GM, Ahmad T, Anstrom KJ, et al.Rationale and design of the GUIDE-IT study:Guiding Evidence Based Therapy Using BiomarkerIntensified Treatment in Heart Failure. J Am CollCardiol HF 2014;2:457–65.

11. Felker GM, Anstrom KJ, Adams KF, et al. Effectof Natriuretic Peptide-Guided Therapy on Hospi-talization or Cardiovascular Mortality in High-RiskPatients With Heart Failure and Reduced EjectionFraction: a randomized clinical trial. JAMA 2017;318:713–20.

12. Douglas PS, DeCara JM, Devereux RB, et al.Echocardiographic imaging in clinical trials:American Society of Echocardiography Standardsfor echocardiography core laboratories. J Am SocEchocardiogr 2009;22:755–65.

13. Lang RM, Badano LP, Mor-Avi V, et al. Rec-ommendations for cardiac chamber quantificationby echocardiography in adults: an update from theAmerican Society of Echocardiography and theEuropean Association of Cardiovascular Imaging.J Am Soc Echocardiogr 2015;28:1–39.

14. Nagueh SF, Smiseth OA, Appleton CP, et al.Recommendations for the evaluation of left ven-tricular diastolic function by echocardiography: anupdate from the American Society of Echocardi-ography and the European Association of Cardio-vascular Imaging. J Am Soc Echocardiogr 2016;29:277–314.

15. Zoghbi WA, Enriquez-Sarano M, Foster E, et al.Recommendations for evaluation of the severityof native valvular regurgitation with two-dimensional and Doppler echocardiography. J AmSoc Echocardiogr 2003;16:777–802.

16. Berger R, Moertl D, Peter S, et al. N-terminalpro-B-type natriuretic peptide-guided, intensivepatient management in addition tomultidisciplinary

care in chronic heart failure a 3-arm, prospective,randomized pilot study. J Am Coll Cardiol 2010;55:645–53.

17. Troughton RW, Frampton CM, Brunner-LaRocca HP, et al. Effect of B-type natriureticpeptide-guided treatment of chronic heart failureon total mortality and hospitalization: an individ-ual patient meta-analysis. Eur Heart J 2014;35:1559–67.

18. Yancy CW, Januzzi JL Jr., Allen LA, et al. 2017ACC Expert Consensus Decision Pathway forOptimization of Heart Failure Treatment: answersto 10 pivotal issues about heart failure withreduced ejection fraction. J Am Coll Cardiol 2018;71:201–30.

19. Kubanek M, Goode KM, Lanska V, Clark AL,Cleland JG. The prognostic value of repeatedmeasurement of N-terminal pro-B-type natriureticpeptide in patients with chronic heart failure dueto left ventricular systolic dysfunction. Eur J HeartFail 2009;11:367–77.

20. Kramer DG, Trikalinos TA, Kent DM,Antonopoulos GV, Konstam MA, Udelson JE.Quantitative evaluation of drug or device effectson ventricular remodeling as predictors of thera-peutic effects on mortality in patients with heartfailure and reduced ejection fraction: a meta-analytic approach. J Am Coll Cardiol 2010;56:392–406.

21. Januzzi JL Jr., Camargo CA, Anwaruddin S,et al. The N-terminal Pro-BNP investigation ofdyspnea in the emergency department (PRIDE)study. Am J Cardiol 2005;95:948–54.

KEY WORDS biomarker-guided therapy,heart failure with reduced ejection fraction,reverse remodeling, NT-proBNP

APPENDIX For a supplemental table andfigure, please see the online version of thispaper.