carbohydrate antigen-125 guided therapy in acute heart failure · mini-focus issue: novel...

J A C C : H E A R T F A I L U R E V O L . 4 , N O . 1 1 , 2 0 1 6

ª 2 0 1 6 B Y T H E AM 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 UN DA T I O N I S S N 2 2 1 3 - 1 7 7 9 / $ 3 6 . 0 0

P U B L I S H E D B Y E L S E V I E R h t t p : / / d x . d o i . o r g / 1 0 . 1 0 1 6 / j . j c h f . 2 0 1 6 . 0 6 . 0 0 7

MINI-FOCUS ISSUE: NOVEL BIOMARKERS

Carbohydrate Antigen-125–GuidedTherapy in Acute Heart FailureCHANCE-HF: A Randomized Study

Julio Núñez, MD,a Pau Llàcer, MD,b Vicente Bertomeu-González, MD,c Maria José Bosch, MD,d Pilar Merlos, MD,e

Sergio García-Blas, MD,a Vicente Montagud, MD,f Vicent Bodí, MD,a Vicente Bertomeu-Martínez, MD,c

Valle Pedrosa, MD,e Andrea Mendizábal, MD,b Alberto Cordero, MD,c Jorge Gallego, MD,d Patricia Palau, MD,d

Gema Miñana, MD,a Enrique Santas, MD,a Salvador Morell, MD,f Angel Llàcer, MD,a Francisco J. Chorro, MD,a

Juan Sanchis, MD,a Lorenzo Fácila, MD,f for the CHANCE-HF Investigators

ABSTRACT

Fro

Me

defSe

ob

au

the

Re

OBJECTIVES This study sought to evaluate the prognostic effect of carbohydrate antigen-125 (CA125)–guided therapy

(CA125 strategy) versus standard of care (SOC) after a hospitalization for acute heart failure (AHF).

BACKGROUND CA125 has emerged as a surrogate of fluid overload and inflammatory status in AHF. After an episode of

AHF admission, elevated values of this marker at baseline as well as its longitudinal profile relate to adverse outcomes,

making it a potential tool for treatment guiding.

METHODS In a prospective multicenter randomized trial, 380 patients discharged for AHF and high CA125 were

randomly assigned to the CA125 strategy (n ¼ 187) or SOC (n ¼ 193). The aim in the CA125 strategy was to reduce CA125

to #35 U/ml by up or down diuretic dose, enforcing the use of statins, and tightening patient monitoring. The primary

endpoint was 1-year composite of death or AHF readmission. Treatment strategies were compared as a time to first event

and longitudinally.

RESULTS Patients allocated to the CA125 strategy were more frequently visited, and treated with ambulatory intra-

venous loop diuretics and statins. Likewise, doses of oral loop diuretics and aldosterone receptor blockers were more

frequently modified. The CA125 strategy resulted in a significant reduction of the primary endpoint, whether evaluated as

time to first event (66 events vs. 84 events; p ¼ 0.017) or as recurrent events (85 events vs. 165 events; incidence rate

ratio: 0.49; 95% confidence interval: 0.28 to 0.82; p ¼ 0.008). The effect was driven by significantly reducing reho-

spitalizations but not mortality.

CONCLUSIONS The CA125 strategy was superior to the SOC in terms of reducing the risk of the composite of

1-year death or AHF readmission. This effect was mainly driven by significantly reducing the rate of rehospitalizations.

(Carbohydrate Antigen-125-guided Therapy in Heart Failure [CHANCE-HF]; NCT02008110) (J Am Coll Cardiol HF

2016;4:833–43) © 2016 by the American College of Cardiology Foundation.

m the aServicio de Cardiología, Hospital Clínico Universitario, INCLIVA, Universitat de Valencia, Valencia, Spain; bServicio de

dicina Interna, Hospital de Manises, Valencia, Spain; cServicio de Cardiología, Hospital de San Juan, Alicante, Spain; dServicio

Medicina Interna, Hospital de la Plana, Castellón, Spain; eServicio de Cardiología, Hospital de Manises, Valencia, Spain; and the

rvicio de Cardiología, Hospital General Universitario de Valencia, Valencia, Spain. This study was funded with public funds

tained in competitive calls: grant EC10-108 of the Ministry of Health Call for Independent Clinical Research in year 2010. The

thors also received the support of CAIBER (CAI11/01/0039), SCReN-Spanish Clinical Research Network (PT13/0002/0031) from

National RþDþI Plan of the Institute of Health Carlos III (Ministry of Economy and Competitiveness: Co-financed by European

gional Development Fund “A way to make Europe”), Red de Investigación Cardiovascular, Programa 7 (RD12/0042/0010 and

https://clinicaltrials.gov/ct2/show/NCT02008110

http://crossmark.crossref.org/dialog/?doi=10.1016/j.jchf.2016.06.007&domain=pdf

http://dx.doi.org/10.1016/j.jchf.2016.06.007

ABBR EV I A T I ON S

AND ACRONYMS

ACEI = angiotensin-converting

enzyme inhibitors

AHF = acute heart failure

ARB = angiotensin II receptor

antagonists

BNP = B-type natriuretic

peptide

CA125 = carbohydrate antigen-

125

CI = confidence interval

FED = furosemide equivalent

dose

HR = hazard ratio

HF = heart failure

IRR = incidence rate ratio

NT-proBNP = N-terminal pro–

B-type natriuretic peptide

SOC = standard of care

RD/12/0042

2013 by Ser

disclose.

Manuscript

Núñez et al. J A C C : H E A R T F A I L U R E V O L . 4 , N O . 1 1 , 2 0 1 6

CA125-Guided Management Therapy in Heart Failure N O V E M B E R 2 0 1 6 : 8 3 3 – 4 3

834

M orbidity and mortality ratesremain very high after dischargefrom acute heart failure (AHF).

Thus, development of new strategies aimedto decrease the risk during this vulnerablephase constitutes a research priority (1–3).This has led to a renewed interest in the useof biomarkers aimed at guiding the intensityof therapy. Indeed, some studies have shownthat the trajectory of some biomarkers corre-lated with the course of the disease, whichmade them theoretically attractive for thispurpose (4). However, it has been chal-lenging to isolate their capacity to predictoutcomes from their potential as atreatment-guiding tool. For instance, the effi-cacy of natriuretic peptide–guided strategieshas revealed heterogeneous and conflictingresults, particularly in highly comorbid andelderly subjects (5,6). There is growing evi-

dence indicating that AHF may be a case mix of phe-notypes contributing to the lack of well-definedevidenced-based therapies in this setting (7,8).

SEE PAGE 844

A great amount of evidence supports the role ofsodium and fluid overload in AHF (9); nevertheless,its severity and organ distribution are largely het-erogeneous (10,11). Fluid overload has been tradi-tionally assessed through symptoms and signs,despite their limited accuracy (10,11). In recent years,carbohydrate antigen-125 (CA125), a widely availablebiomarker used for ovarian cancer monitoring (12),has emerged as a potential surrogate of fluid reten-tion and inflammation activity in AHF (13). Publisheddata has shown that high levels of this glycoprotein,which is present in up to two-thirds of patients hos-pitalized for AHF (13,14), correlate with the severity ofAHF (13–15) and relate to morbidity and mortality (12).The potential role of this glycoprotein for monitoringand guiding post-discharge therapy has been sug-gested on the basis of the significant correlation be-tween CA125 serum fluctuations with clinicaloutcomes (16–18), and response to loop diuretics andstatin therapy (19,20).

With this rationale in mind, we designed this ran-domized multicenter clinical trial with the aim to

/0068) FEDER, and PIE15/00013. Dr. Núñez received support to or

vier and Ferrer. All other authors have reported that they have no

received March 1, 2016; revised manuscript received June 23, 20

evaluate the role of CA125 as a guiding tool strategyfor patients recently discharged for AHF.

METHODS

STUDY DESIGN AND OVERSIGHT. This is aninvestigator-initiated, multicenter, open-label, ran-domized controlled trial conducted from December 7,2011, to July 17, 2014, in 5 academic centers in Spain.Informed consent was obtained from all patients ortheir nominated representative. The InstitutionalReview Board of each center and Agencia Española deMedicamentos y Productos Sanitarios approved thestudy. The study design was published previously(21). The trial was overseen by an independent dataand safety monitoring board. The funders of the studyhad no role in study design, data collection, dataanalysis, data interpretation, or writing of the report.All analyses were performed by an independentcompany (CLINSTATS LLC, Reading, Pennsylvania).

PATIENTS. Candidate patients were selected from acohort of patients recently discharged for AHF. In-clusion and exclusion criteria for the trial are detailedin Online File 1. Briefly, the trial included patientswith at least 1 episode of AHF in the last 180 days, NewYork Heart Association functional class $II at themoment of enrollment, CA125 >35 U/ml, and eitherechocardiographic evidence of a structural or func-tional abnormality of the heart congruent with heartfailure (HF) diagnosis or elevation of natriuretic pep-tides (N-terminal pro–B-type natriuretic peptide [NT-proBNP] >1,000 pg/ml or B-type natriuretic peptide[BNP] >100 pg/ml), or both.

RANDOMIZATION. Trained physicians enrolled par-ticipants at each site. After providing the informedconsent, patients were randomly assigned, with aremote, web-based computer-generated blockrandomization procedure in a 1:1 ratio, to either theCA125 strategy or standard of care (SOC). Patients,outcomes evaluators, and personnel involved in datamanagement were masked to group assignment.

STUDY TREATMENTS. All patients were followed upin the outpatient HF clinics of each center with pre-specified visits at 1, 6, and 12 months after randomi-zation. Additional outpatient visits were permitted inboth arms at the discretion of the physician in charge ofthe patient. In all pre-specified visits, a sample of blood

ganize a CHANCE-HF researcher meeting at 2012 and

relationships relevant to the contents of this paper to

16, accepted June 23, 2016.


J A C C : H E A R T F A I L U R E V O L . 4 , N O . 1 1 , 2 0 1 6 Núñez et al.N O V E M B E R 2 0 1 6 : 8 3 3 – 4 3 CA125-Guided Management Therapy in Heart Failure

835

for standard laboratory testing and CA125 measure-ment (Elecsys CA125 II assay, Roche Diagnostics, Basel,Switzerland) was obtained. CA125 results were blindedto investigators and patients belonging to the SOC arm.

Patients allocated to SOC were treated followingstandard guidelines regarding the use of angiotensin-converting enzyme inhibitors (ACEI), angiotensin IIreceptor antagonists (ARB), beta-blockers, aldoste-rone antagonists, ivabradine, statins, diuretics, andother treatments such as anticoagulants, antiar-rhythmics, digoxin, nitrates and vasoactive drugs,devices, revascularization, and surgical procedures(22,23). No specific algorithm for drug therapy intro-duction or intensification was used. The need foradditional outpatient visits and intravenous diureticadministration were left to the discretion of theclinician in charge of the patient.

The CA125 strategy followed a pre-specified algo-rithm (Online File 2) aimed at keeping CA125 levels at35 U/ml or less by means of diuretic dose optimiza-tion, enforcing the use of statins, and increasing thefrequency of monitoring visits.OUTCOMES. The primary endpoint was the compos-ite of 1-year all-cause mortality or readmission forAHF. Readmission for AHF was defined as any un-planned hospitalization requiring a stay >24 h, andcaused by substantive worsening of HF requiringintravenous administration of diuretics, inotropes, orvasodilators. Secondary endpoints included: 1) thecomposite of 1-year death or readmission for anycause; 2) all-cause mortality and days alive out ofhospital; 3) recurrent hospitalizations; 4) episodes ofworsening HF not requiring hospitalization; and5) depiction of CA125 and natriuretic peptides’ tra-jectories over follow-up. Specific definitions of end-points are described elsewhere (21).

Safety endpoints included the evaluation andcomparison of incidence rates of renal dysfunc-tion (estimated glomerular filtration rate [eGFR]<60 min/ml/m2), hypertransaminasemia, hyper-kalemia, hypokalemia, and elevated total creatinekinase.STATISTICAL ANALYSIS. Sample s ize . On the basisof prior observations, we expected that the CA125strategy would translate into 35% relative riskreduction of the primary endpoint. Thus, using an80% power to detect at least such difference at 1 year,and using a 1:1 allocation ratio, 180 patients wereestimated on each group (n ¼ 360) (21). The totalsample was increased to 380 patients to account forineligibility criteria and loss to follow-up, which wasassumed to be around 5% (21).

All statistical comparisons were made under theintention-to-treat principle. Results are presented as

frequencies, mean � SD, or median (interquartilerange [IQR]), as appropriate. Between-group com-parisons were performed using t test, Mann-Whitneytest, or Fisher exact test. Differences in count amongstrategies are presented as incidence rate ratio (IRR).Time-to-event analys i s . Patient follow-up wascensored at the time of death, new diagnosis of can-cer, cardiac valve surgery, or transcatheter interven-tion. Differences between treatment groups weredepicted with the Kaplan-Meier method and testedby the Peto-Peto test. The proportional hazardsassumption was checked for the type of strategyagainst each endpoint using Schoenfeld residuals.When the proportionality assumption was untenable,a restricted mean survival time (RMST) difference wasused as a global measure of treatment efficacy (24).A Cox proportional hazards model stratified by centerwas used to estimate hazard ratio (HR). Pre-specifiedsubgroup analyses included >75 years of age, gender,ischemic etiology, left ventricular ejection fraction(LVEF) >50%, CA125 above median, natriuretic pep-tides above median, and eGFR <60 ml/min/1.73 m2.Long i tud ina l ana lys i s . The cumulative rate ofreadmissions (AHF and all cause) were plotted againsttime and compared by the Ghosh and Lin estimator(25), which accounts for death as a terminal event. Thelongitudinal counterpart of the primary endpointincluded all episodes of AHF rehospitalization asrecurrent events. One simple strategy for incorpo-rating death into analyses of this endpoint is toconsider this outcome as an additional event in therecurrent event process (26). That is, one considers acomposite of recurrent AHF hospitalizations anddeath. This updated recurrent event process wasanalyzed using multilevel negative binomial regres-sion for longitudinal data. A secondary longitudinalendpoint included all episodes of all-cause rehospi-talization, with death accounted in a similar way. Thedifference between strategies from both longitudinalendpoints is expressed as IRR.CA125 and natriuretic peptides’ longitudinal trajectories.Measured values of CA125 and natriuretic peptides atpre-planned visits were plotted against time and theirtrajectory difference among strategies tested bymeansof jointmodeling of longitudinal and survival data (27).A short description about this method is presented inOnline File 3. Missing values of NT-proBNP (n ¼ 61)were imputed using BNP and other covariates within amultiple imputation framework (chained equationsimputation).Sensitivity analyses. This analysis aimed to evaluate therole of NT-proBNP changes on the primary endpoint,and how this marker influences the endpoint–treat-ment strategy association. Only those values of NT-





836

proBNP that precede the occurrence of the endpointwere included. Two joint models were built: Model 1included logNT-proBNP and treatment strategy asmain effects and Model 2 included the interaction be-tween logNT-proBNP and the treatment strategy.

A 2-sided p value of <0.05 was considered to bestatistically significant for all analyses. All analyseswere performed using Stata 14.1 (StataCorp LP, Col-lege Station, Texas) and R package (R DevelopmentCore Team, Vienna, Austria).

RESULTS

PATIENTS. Of 380 patients included in this study, 187were randomly assigned to the CA125 strategy and 193 towere assigned to SOC (Figure 1). Randomization tookplace at discharge from the index admission in 97.9%(n ¼ 372), with the rest within the first week afterdischarge.

FIGURE 1 Patient Flow Chart

CA125 ¼ carbohydrate antigen-125; COPD ¼ chronic obstructive pulmon

The mean age of the study population was 73.7 �11.1 years, 44.2% were women, and 40.3% exhibitedLVEF >50%. Baseline characteristics across treatmentarms are shown in Table 1. Overall, there were nodifferences between both treatment strategies withthe exception of higher proportion of hypertensionand diabetes mellitus in the CA125-guided arm.

PRE-SPECIFIED VISITS. According to the protocol,patients were evaluated at randomization (visit 0,n ¼ 380), 30 days (visit 1, n ¼ 363), 6 months (visit 2,n ¼ 334), and 1 year (visit 3, n ¼ 298). The median timesince randomization was 26 days (IQR: 23 to 33 days),183 days (IQR: 175 to 196 days), and 363 days (IQR:352 to 371 days) to visit 1, visit 2, and visit 3,respectively.

INTERVENTION. Ambulatory v i s i t s . During thecourse of the trial, the group allocated to the CA125strategy had more visits (pre-specified and optional)

ary disease.

TABLE 1 Baseline Characteristics

CA125-GuidedTherapy(n ¼ 187)

SOC(n ¼ 193) p Value

Demographics

Age, yrs 74 � 11 73 � 11 0.41

Male 107 (57.2) 105 (54.4) 0.61

Weight, kg 75.2 � 17.3 75.7 � 18.1 0.80

Body mass index, kg/m2 28.5 � 5.1 28.7 � 6.0 0.77

Medical history

Hypertension 167 (89.3) 158 (81.9) 0.04

Diabetes mellitus 100 (53.5) 82 (42.5) 0.04

Insulin-dependent diabetes mellitus 44 (23.5) 36 (18.7) 0.26

Atrial fibrillation 110 (58.8) 114 (59.1) 0.99

Coronary artery disease 61 (32.6) 61 (31.6) 0.91

Myocardial infarction 45 (24.1) 44 (22.8) 0.81

Stroke 16 (8.6) 16 (8.3) 0.99

Peripheral artery disease 22 (11.8) 17 (8.8) 0.40

Prior history of valvular heart disease 63 (33.7) 73 (37.8) 0.45

Chronic renal disease 71 (38.0) 59 (30.6) 0.13

COPD 32 (17.1) 24 (12.4) 0.25

Tobacco and alcohol use

Active smoker 24 (12.8) 20 (10.4) 0.52

Previous smoker 55 (29.6) 55 (28.5) 0.82

Alcohol abuse 12 (6.4) 15 (7.8) 0.69

Medical devices

Pacemaker 20 (10.7) 15 (7.8) 0.38

ICD 19 (10.2) 17 (8.8) 0.73

Biventricular pacemaker 8 (4.3) 8 (4.1) 0.99

Physical examination

Heart rate, beats/min 79 � 19 80 � 19 0.52

Systolic blood pressure, mm Hg 127 � 25 122 � 22 0.06

Diastolic blood pressure, mm Hg 69 � 14 69 � 14 0.80

Peripheral edema 103 (55.1) 108 (55.9) 0.86

Electrocardiogram and echocardiography

QRS duration, ms 100 (80–130) 100 (80–120) 0.59

LVEF, % 45.8 � 16.8 44.7 � 17.2 0.56

LVEF <35% 65 (35) 68 (36) 0.91

LVEF <50% 109 (58.3) 103 (53.4) 0.33

LVDD, mm 55.2 � 10.5 54.5 � 10.9 0.57

LVSD, mm 40.2 � 12.3 40.7 � 13 0.73

LA, mm 46.1 � 8.9 46.2 � 8.7 0.93

TAPSE, mm 17.8 � 4.5 17.4 � 6.3 0.48

PASP, mm Hg* 48.4 � 15.6 45.8 � 14.6 0.16

Laboratory results

Hemoglobin, g/dl 12.2 � 1.9 12.4 � 2.0 0.52

Iron deficiency 82 (43.9) 97 (50.3) 0.22

Sodium, mEq/l 140 (137–142) 140 (138–143) 0.02

Potassium, mg/dl 4.2 � 0.6 4.2 � 0.5 0.31

Blood urea nitrogen, mg/dl 60 (42–83) 55 (42–80) 0.20

Creatinine, mg/dl 1.20 (0.93–1.51) 1.13 (0.94–1.50) 0.56

CA125, U/ml 103 (64–174) 86 (55–160) 0.12

BNP, pg/ml† 582 (321–1,010) 662 (448–1,154) 0.33

NT-proBNP, pg/ml‡ 4,570 (2,251–9,849) 3,773 (1,947–8,192) 0.18

Values are mean � SD, n (%), or median (interquartile range). Iron deficiency was defined as serumferritin <100 mg/l or as ferritin 100 to 299 mg/l with a transferrin saturation <20%. *n ¼ 277. †n ¼ 61. ‡n ¼ 319.

BNP ¼ B-type natriuretic peptide; CA125 ¼ carbohydrate antigen-125; COPD ¼ chronic obstructive pulmonarydisease; ICD ¼ implantable cardioverter-defibrillator; LA ¼ left atrial diameter; LVDD ¼ left ventricular diastolicdiameter; LVEF ¼ left ventricular ejection fraction; LVSD ¼ left ventricular systolic diameter; NT-proBNP ¼ N-terminal pro–B-type natriuretic peptide; PASP ¼ pulmonary artery systolic pressure; SOC ¼ standard of care;TAPSE ¼ tricuspid annulus plane systolic excursion.


837

compared with those allocated to SOC (5.97 visits/person-year vs. 5.23 visits/person-year, respectively;IRR: 1.14; p ¼ 0.003).

Beta-b lockers , ACEI or ARB, and aldosteronereceptor b lockers . There was no difference in pre-scription frequencies for beta-blockers, ACEI or ARB,and aldosterone receptor blockers either at randomi-zation or at the end of the study. For aldosteronereceptor blockers, doses were more frequently up-titrated among the CA125-guided patients (Table 2).Diuret ics . Oral loop diuretics. There were no differ-ences in the frequency or dose of loop diuretics, eitherat randomization or at the end of the study (Table 2).However, the CA125-guided strategy was associatedwith higher frequency of furosemide equivalentdose (FED) adjustments (up- or down-titration) ascompared with SOC (Table 2). Figure 2 depicts the tra-jectory of FED stratified by treatment strategy and thesimultaneous status of CA125. FED values wereconsistently higher when CA125 was >35 U/ml in allpatients (110 � 71 vs. 80 � 55; p < 0.001). Indeed, FEDlongitudinal profile was close to 20 mg/day higher inpatients with CA125>35 U/ml (p¼0.038). In thosewithCA125 #35 U/ml, FED was not different between the 2strategies (Figure 2).Ambulatory intravenous furosemide. The percentage ofpatients and rates at which intravenous furosemidewas administered in HF ambulatory units was higherfor the CA125-guided strategy (21.0% vs. 11.0%; p ¼0.008; 0.58 administrations/person-year vs. 0.28administrations/person-year; IRR: 2.11; p < 0.001).Specifically, 19 (10.2%) patients in the CA125-guidedarm had 2 or more sessions of intravenousfurosemide compared with 11 (5.7%) of thoseassigned to SOC (p ¼ 0.019) (Figure 2).Thiazides. Frequency of thiazide prescriptions atrandomization and at the end of the study was similaramong strategies; nevertheless, there was a trendtoward a higher number of dose adjustments (anddose increase) in the active arm (Table 2).

Stat ins . Statin therapy was more frequently pre-scribed in the CA125-guided group at randomization(82.4% vs. 53.4%; p < 0.001) and at the end of the trial(78.1% vs. 41.5%; p < 0.001). The rate at which statindoses changed did not differ between the 2 arms.However, a decrease in statin doses was morefrequently observed in the SOC arm (Table 2).

Others . No significant differences were found for theproportion of patients treated with ivabradine anddigoxin either at randomization or at the end of trial(Table 2). Likewise, no differences regarding revas-cularization procedures (3.9% vs. 4.2%; p¼1.00), car-diac valve replacement (0.6% vs. 0.5%; p¼1.00),

TABLE 2 Treatment Across Strategies


SOC(n ¼ 193) IRR p Value

FED

Randomization 186 (99.5) 187 (96.9) 0.12

End of the trial 180 (96.3) 184 (95.3) 0.80

FED at discharge, mg/24 h 94.0 � 49.980 (60–120)

96.3 � 56.280 (40–120)

0.670.73

FED at the end of the trial, mg/24 h 108.0 � 89.980 (40–120)

97.3 � 71.780 (40–80)

0.200.31

Dose adjustments* 2.91 1.47 1.98 <0.001

Dose increases* 1.30 0.83 1.57 <0.001

Dose decreases* 1.57 1.15 1.37 <0.001

IV furosemide

Outpatient IV administration 40 (21.4) 21 (10.9) 0.01

Outpatient IV administration† 0.58 0.28 2.11 <0.001

Aldosterone receptor blockers

Randomization 100 (53.5) 107 (55.4) 0.76

End of the trial 102 (54.6) 102 (52.9) 0.76

$50% goal doses at baseline 93 (94.9) 98 (91.6) 0.41

$50% goal doses at the end of trial 92 (92.0) 90 (89.1) 0.63

Dose adjustments* 0.32 0.22 1.50 0.05

Dose increases* 0.16 0.08 2.01 0.03

Dose decreases* 0.17 0.14 1.21 0.48

ACEI/ARB

Randomization 134 (71.7) 140 (72.5) 0.91

End of the trial 141 (75.4) 143 (74.1) 0.81




Dose increases* 0.30 0.27 1.11 0.60

Dose decreases* 0.19 0.19 1.01 0.95

Beta-blockers

Randomization 160 (85.6) 160 (82.9) 0.49

End of the trial 157 (84.0) 154 (79.8) 0.35




Dose increases* 0.41 0.37 1.11 0.55

Dose decreases* 0.24 0.20 1.16 0.52

Statins

Randomization 154 (82.4) 103 (53.4) <0.001

End of the trial 146 (78.1) 80 (41.5) <0.001




Dose increases* 0.02 0.01 2.09 0.42

Dose decreases* 0.06 0.13 0.48 0.04

Continued on the next page



838

implantable cardioverter-defibrillator implant (2.8%vs. 6.3%; p ¼ 0.137), cardiac resynchronization ther-apy (1.7% vs. 1.1%; p ¼ 0.678), or ultrafiltration tech-niques (4.4% vs. 2.2%; p ¼ 0.405) were found acrossboth strategies. The percentage of patients at whichintravenous iron was administered was higher for theCA125-guided strategy (Table 2).

OUTCOMES EVALUATED AS TIME TO FIRST EVENT.

Pr imary endpoint . The CA125 strategy resulted in a

reduction of the proportion of events (66 [35.3%] vs.84 [43.5%]; p ¼ 0.101). Kaplan-Meier plot showed the2 curves widely separated over the follow-up, butwith differences that decreased over time (Figure 3A).For this endpoint, the variable type of strategydid not follow the proportionality assumption(Schoenfeld residuals p ¼ 0.018), not allowing us tomeasure its effect with a single HR. Instead, we esti-mated HRs using Cox regression at 3 months (HR:0.48; 95% confidence interval [CI]: 0.30 to 0.75;p ¼ 0.002), 6 months (HR: 0.64; 95% CI: 0.43 to 0.95;p ¼ 0.028), 9 months (HR: 0.71; 95% CI: 0.50 to 1.01;p ¼ 0.058), and 12 months (HR: 0.72; 95% CI: 0.52 to0.99; p ¼ 0.049). More importantly, however, wasthat the difference in RMST between the 2 strategieswas significant (0.08 years; 95% CI: 0.02 to 0.15 years;p ¼ 0.017) in favor of the CA125 strategy (0.74 yearsvs. 0.66 years). Expressed in days, this means thatpatients in the CA125 strategy had, on average, 30days more time free of event in a 1-year framework.

We found no interactions between treatmentstrategies and any of the pre-specified subgroups(Online File 4).Secondary endpo ints . Composite of death or anyrehospitalization. The number of events in the CA125strategy versus SOC was 84 (44.9%) versus 97 (50.3%),respectively (p ¼ 0.297). The difference in RMST be-tween the 2 strategies was used as the main approachfor efficacy measure, on the basis that the propor-tionality assumption was untenable (Schoenfeld re-siduals p ¼ 0.050). The RMST for the CA125 strategyand SOCwas 0.67 years versus 0.61 years, respectively,with a difference of 0.054 years (95% CI: –0.018 to0.125 years; p ¼ 0.142) (Figure 3B).All-cause mortality and days alive and out of thehospital. There were no differences in the proportionof events between the 2 strategies (31 [16.6%] vs. 35[18.1%]; p ¼ 0.689). The HR as a measure of globaleffect was not significant (HR: 0.92; 95% CI: 0.57 to1.49; p ¼ 0.735) (Online File 5A). Likewise, nodifferences were found for mortality rates usingdays alive and out of the hospital (Peto-Peto test:p ¼ 0.732) (Online File 5B).LONGITUDINAL ANALYSIS. Recurrent hospitalizationsincluding outside hospital death as additionalreadmission. The CA125 strategy showed a signifi-cant reduction in the rate of recurrent AHF (as a partof the primary endpoint) and all-cause rehospitaliza-tion: 85 events versus 165 events (IRR: 0.49; 95%CI: 0.28 to 0.82; p ¼ 0.008) and 132 events versus 202events (IRR: 0.61; 95% CI: 0.39 to 0.96; p ¼ 0.033),respectively. Figures 3C and 3D show the mean numberof recurrences for AHF and all-cause hospitalization,respectively.




TABLE 2 Continued


SOC(n ¼ 193) IRR p Value

Thiazide diuretic agents

Baseline 24 (12.8) 24 (12.4) 0.99

End of the trial 47 (25.1) 50 (25.9) 0.91

HCTZ equivalent dose atdischarge, mg/24 h

28.9 � 13.9 26.0 � 11.8 0.44

HCTZ equivalent dose at theend of the trial, mg/24 h

26.7 � 17.3 24.6 � 12.2 0.49


Dose increases* 0.14 0.04 3.83 0.09

Dose decreases* 0.19 0.11 1.70 0.33

Digoxin

Randomization 19 (10.2) 16 (8.3) 0.60

End of the trial 22 (11.8) 19 (9.8) 0.62

Ivabradine

Randomization 7 (3.7) 7 (3.6) 0.99

End of the trial 13 (7.0) 17 (8.8) 0.57

IV iron

Iron IV administration 40 (21.4) 23 (11.9) 0.02

Values are n (%), mean � SD, or median (interquartile range). *Rates (episodes of changes per person-year).

ACEI ¼ angiotensin-converting enzyme inhibitors; ARB ¼ angiotensin II receptor blockers; CA125 ¼ carbohy-drate antigen-125; FED ¼ furosemide equivalent dose; HCTZ ¼ hydrochlorothiazide; ICD ¼ implantable car-dioverter defibrillator; IRR ¼ incidence rate ratio; IV ¼ intravenous; SOC ¼ standard of care.


839

Recurrent hosp i ta l i zat ions (no ad justment fordeath as termina l event) . There was a superiorityof the CA125 strategy in reducing rates for recurrentAHF hospitalizations (0.40 events/person-year vs.0.85 events/person-year; IRR: 0.47; 95% CI: 0.35 to0.63; p < 0.001) and all-cause hospitalizations (0.67events/person-year vs. 1.06 events/person-year; IRR:0.63; 95% CI: 0.50 to 0.80; p < 0.001).Emergency room vis i t s for AHF not requ i r inghosp i ta l i zat ion . Patients in the CA125-guided armhad fewer visits to the emergency department (withduration<24h) as comparedwith the SOCgroup (10.7%vs. 17.1%; p ¼ 0.077). This difference becomes signifi-cant when expressed as events per person-year(0.16 events/person-year vs. 0.28 events/person-year;p ¼ 0.008).CA125 and NT-proBNP tra jector ies . CA125 valuesdecreased over time in both treatment arms, espe-cially between randomization and visit 1. Lower meanvalues were found in CA125 strategy, particularlyafter the third month (p ¼ 0.027) (Online File 6).LogNT-proBNP decreased over time in both armswithout significant differences at randomization and30 days. At 6 months and 1 year logNT-proBNP waslower in SOC (Online File 7).

SENSITIVITY ANALYSES. NT-proBNP changes andthe pr imary endpoint . The strategy guided byCA125 reduced the incidence of the compositeendpoint by 38% on average, despite adjustment bylongitudinal values of logNT-proBNP (HR: 0.62; 95%CI: 0.44 to 0.86; p ¼ 0.005). In addition, the predic-tive effect of logNT-proBNP was uniformly presentbetween the 2 arms with no difference among them(p value for interaction ¼ 0.924).

SAFETY. Rates of eGFR <60 ml/min/1.73 m2 werehigher in the CA125-guided arm (IRR: 1.25; 95% CI: 1.07to 1.47; p ¼ 0.004); however, this difference did nottranslate into an increase in the rate for acute renalfailure hospitalizations (3.1% vs. 1.5%; p ¼ 0.335). Inaddition, we found no differences in the frequency ofhospitalizations due to hyperkalemia (0.53% vs. 0%;p ¼ 0.493), hypokalemia (1.06% vs. 0.52%; p ¼ 0.619),or any other safety endpoint for the SOC and CA125-guided arms, respectively.

DISCUSSION

Despite the associated high risk of adverse outcomes,the evidence about the optimal therapeutic approachafter discharge from AHF is largely unknown (1–3).To explore new avenues about the optimal manage-ment of these patients, we designed a randomizedcontrolled trial to compare the effect of a CA125-guided strategy versus SOC on the composite of

1-year death or readmission for AHF. At its face value,this strategy was characterized by an increase in thefrequency of patients’ monitoring, optimization ofthe use of diuretics, and increase in the use of statins.Under this strategy, there was a significant reductionin the primary endpoint, the cumulative risk ofreadmissions (HF-related [w50%] as well as all-causereadmission [w40%]), and visits to emergency room(w40%). Interestingly, this effect was homogeneousamong all pre-specified subgroups and independentof changes in natriuretic peptides. Participantswere selected by the levels of CA125. By using thisscheme, a significant proportion of high-risk patients(elderly subjects, women, and patients with pre-served ejection fraction) were well represented in thistrial.

In contrast to other biomarker-guided trials inwhich management strategies are on the basis ofevidence-based therapies (5,6), our CA125-guidedapproach focused on empirical, and sometimescontroversial, interventions.

By exploiting the assumption about the surrogacyof CA125 for fluid overload, the dose titration of loop-diuretic treatment was an important component ofthe active strategy. Indeed, in the CA125 strategy, theFED titration (either up or down) was about 50% morefrequent as compared with the control arm. Addi-tionally, up to 21% of these patients received ambu-latory administration of intravenous furosemide on



FIGURE 2 Furosemide Equivalent Dose Administration Across Treatment Strategies and CA125 Status

*Values obtained during follow-up (excluding values at randomization). CA125 ¼ carbohydrate antigen-125; FEDpo ¼ furosemide equivalent

dose, oral administration; FEDiv ¼ furosemide equivalent dose, intravenous administration; SOC ¼ standard of care.



840

the basis of clinical profile and CA125 response.Furthermore, ARB were more frequently up-titrated.Perhaps the limited sensitivity and specificity ofsymptoms or signs for the clinical assessment of fluidoverload (9,10) have contributed to the inertia in theuse of loop diuretics in this setting.

Even though contemporary guidelines do notendorse a routine use of statins in HF (23,24), someobservational data have suggested a beneficial effectby reducing readmission rates in patients with HFand elevated inflammatory activity (28). Fueled by theobserved correlation between CA125 and inflammationmarkers in HF (13), we decided to add statins as acomponent of the CA125 strategy. In the end, 80% ofpatients in the active arm received statins withoutrelevant safety issues. Probably as a consequenceof patient’s close monitoring, the frequency of intra-venous iron administration was higher in the CA125strategy.

On the contrary, there were no differences onthe use of beta-blockers, ACEI or ARB, and otherpharmacological and nonpharmacological treatments

between the 2 strategies. Worth noticing is that theuse of these evidence-based therapies met the stan-dard recommended by the guidelines.

Logistic aspects about this biomarker, such as itswide availability, standardized measurement, and lowcost, offer additional advantages for its fast imple-mentation in the routine management of HF. In addi-tion, plasmaCA125 appears not significantly influencedby gender, LVEF, and renal function. Interestingly, thehalf-life reported for this biomarker (around 5 to 12days) (13) could provide useful clinical and patho-physiological information of the prior weeks, in asimilar fashion as glycated hemoglobin in diabetes.STUDY LIMITATIONS. This trial is not exempt oflimitations. First, not including patients with a low-normal CA125 limits the external validity of our find-ings. We believe that including these patients wouldhave made it easier to detect a significant differencebetween the 2management strategies, andwould havereinforced the role of this biomarker for guiding ther-apy. Nonetheless, the inclusion of these patients in alarger trial deserves further consideration in the

FIGURE 3 Time to First Event and Recurrent Events Endpoints

(A) Kaplan-Meier curve for the composite of death or acute heart failure (AHF) rehospitalization (primary endpoint). (B) Kaplan-Meier curve for the composite of death or

all-cause rehospitalization (secondary endpoint). (C) Recurrent events plot for AHF rehospitalizations (primary endpoint). (D) Recurrent events plot for all-cause

rehospitalizations (secondary endpoint). Recurrent event estimates (C, D) are adjusted for informative censoring due to mortality and the curves tested among strategies

using the Ghosh and Lin test (25). CA125 ¼ carbohydrate antigen-125; SOC ¼ standard of care.


841

future. Second, not having the randomly assignedstrategy being blinded from the physician and nursesin charge of the patient is indeed a potential threat tothe internal validity of the trial. However, the infor-mation about patients’ levels of CA125 ismandatory fortreatment monitoring. Third, this is a strategy com-parison (holistic approach), and thus it is not possibleto isolate the effect of individual components thatmake up the active strategy. Fourth, in some cases, thetrajectory of CA125 could have been influenced by

other nondiagnosed subclinical conditions differentfrom HF. Fifth, the fact that most of the effect of theguided strategy involved the readmission componentof the endpoints suggests that we may need a biggersample size and longer trial duration to detect a sig-nificant effect on mortality. Sixth, due to limitedsample size, we cannot rule out false negative resultsdue to type II error when testing the interactions in thesubgroup analyses. Last, the lack of effect on days aliveand out of the hospital, despite the significant

PERSPECTIVES

COMPETENCY IN MEDICAL KNOWLEDGE: Mor-

bimortality following an episode of AHF remains high.

Unfortunately, the optimal therapeutic approach in this

setting is not well defined, especially regarding the in-

tensity of depletive therapies, use of statins, and fre-

quency of monitoring. CA125 has shown to be a reliable

prognostic marker, and a proxy of fluid overload and

inflammatory activity. The fact that plasma levels of this

marker respond to the intensity of diuretic use opens the

potential use of this biomarker as a guiding therapy tool.

The results of this trial supported the superiority of a

CA125-guidedstrategy (as comparedwithSOC), in terms

of reducing the risk of the composite of 1-year death or

AHF readmission. Building on the hope that our results

will be reproduced in further trials, a CA125-guided

strategy will pave the road toward a more accurate

management of patients with AHF.

TRANSLATIONAL OUTLOOK: Larger studies are

needed to confirm our results and define the real

clinical impact of the CA125-guided therapy after an

episode of AHF.



842

reduction on readmissions, suggests that the avoidedhospitalizations in the active arm were in those withlower severity and hence shorter stays.

CONCLUSIONS

After being discharged for AHF, a CA125-guidedtherapy aiming to keep CA125 levels below 35 U/mlwas associated with a reduced risk of 1-year adverseoutcomes. This effect was mainly driven by signifi-cantly reducing the rate of rehospitalizations.

ACKNOWLEDGMENTS The authors thank the studycoordinators, nurses, and staff at the investigativesites and especially all of the patients involved inthis trial. Special acknowledgment to Marta Peiró(INCLIVA), Gemma Romero (INCLIVA), Anna Mollar(INCLIVA), Estefania Montalvo (INCLIVA), AmparoVillaescusa (INCLIVA), and INCLIVA workers forthe logistic and institutional support. Theseindividuals received no compensation for theircontributions.

REPRINT REQUESTS AND CORRESPONDENCE: Dr.Julio Núñez, Servicio de Cardiología, Hospital ClínicoUniversitario, Avenida Blasco Ibáñez 17, 46010Valencia, Spain. E-mail: [email protected].

RE F E RENCE S

1. Dharmarajan K, Hsieh AF, Kulkarni VT, et al.Trajectories of risk after hospitalization for heartfailure, acute myocardial infarction, or pneumonia:retrospective cohort study. BMJ 2015;350:h411.

2. Ambrosy AP, Fonarow GC, Butler J, et al. Theglobal health and economic burden of hospitali-zations for heart failure: lessons learned fromhospitalized heart failure registries. J Am CollCardiol 2014;63:1123–33.

3. Krumholz HM, Lin Z, Keenan PS, et al. Rela-tionship between hospital readmission and mor-tality rates for patients hospitalized with acutemyocardial infarction, heart failure, or pneumonia.JAMA 2013;309:587–93.

4. Januzzi JL, Troughton R. Are serial BNP mea-surements useful in heart failuremanagement? Serialnatriuretic peptide measurements are useful in heartfailure management. Circulation 2013;127:500–7.

5. Troughton R, Michael Felker G, Januzzi JL Jr.Natriuretic peptide-guided heart failure manage-ment. Eur Heart J 2014;35:16–24.

6. PfistererM, Buser P, Rickli H, et al. BNP-guided vssymptom-guided heart failure therapy: BNP-guidedvs symptom-guided heart failure therapy: the Trialof Intensifiedvs StandardMedical Therapy in ElderlyPatients With Congestive Heart Failure (TIME-CHF)randomized trial. JAMA 2009;301:383–92.

7. Felker GM, Lee KL, Bull DA, et al. Diureticstrategies in patients with acute decompensatedheart failure. N Engl J Med 2011;364:797–805.

8. Di Somma S, Magrini L. Drug therapy for acuteheart failure. Rev Esp Cardiol (Engl Ed) 2015;68:706–13.

9. Parrinello G, Torres D, Paterna S, di Pasquale P,Licata G. The pathophysiology of acute heartfailure: the key role of fluid accumulation. AmHeart J 2008;156:e19.

10. Gheorghiade M, Follath F, Ponikowski P, et al.,European Society of Cardiology, European Societyof Intensive Care Medicine. Assessing and gradingcongestion in acute heart failure: a scientificstatement from the acute heart failure committeeof the heart failure association of the EuropeanSociety of Cardiology and endorsed by the Euro-pean Society of Intensive Care Medicine. Eur JHeart Fail 2010;12:423–33.

11. Peacock WF, Soto KM. Current techniques offluid status assessment. Contrib Nephrol 2010;164:128–42.

12. Marcus CS, Maxwell GL, Darcy KM,Hamilton CA, McGuire WP. Current approachesand challenges in managing and monitoringtreatment response in ovarian cancer. J Cancer2014;5:25–30.

13. Núñez J, Miñana G, Núñez E, Chorro FJ, Bodí V,Sanchis J. Clinical utility of antigen carbohydrate125 in heart failure. Heart Fail Rev 2014;19:575–84.

14. Núñez J, Sanchis J, Bodí V, et al. Improvementin risk stratification with the combination of the

tumour marker antigen carbohydrate 125 and brainnatriuretic peptide in patients with acute heartfailure. Eur Heart J 2010;31:1752–63.

15. D’Aloia A, Faggiano P, Aurigemma G, et al.Serum levels of carbohydrate antigen 125 in pa-tients with chronic heart failure: relation to clinicalseverity, hemodynamic and Doppler echocardio-graphic abnormalities, and short-term prognosis.J Am Coll Cardiol 2003;41:1805–11.

16. Núñez J, Núñez E, Sanchis J, et al. Antigencarbohydrate 125 and brain natriuretic peptideserial measurements for risk stratificationfollowing an episode of acute heart failure. Int JCardiol 2012;159:21–8.

17. Miñana Escrivá G, Núñez J, Sanchis J, et al.Carbohydrate antigen 125 serial measurementsafter an admission for acute heart failure and riskof early readmission. Med Clin (Barc) 2012;139:479–86.

18. Núñez J, Núñez E, Bayés-Genís A, et al. Long-term serial kinetics of N-terminal pro B-typenatriuretic peptide and carbohydrate antigen 125for mortality risk prediction following acute heartfailure. Eur Heart J Acute Cardiovasc Care 2016May 19 [Epub ahead of print].

19. Núñez J, Núñez E, Miñana G, et al. Differentialmortality association of loop diuretic dosage ac-cording to blood urea nitrogen and carbohydrateantigen 125 following a hospitalization for acuteheart failure. Eur J Heart Fail 2012;14:974–84.

mailto:[email protected]

http://refhub.elsevier.com/S2213-1779(16)30325-0/sref1



























































































843

20. Núñez J, Sanchis J, Núñez E, et al. Benefitsof statin therapy according to plasma antigencarbohydrate 125 following an admission foracute heart failure. Rev Esp Cardiol 2011;64:1100–8.

21. Núñez J, Merlos P, Fácila L, et al., CHANCE-HF Investigators. Prognostic effect of carbohy-drate antigen 125-guided therapy in patientsrecently discharged for acute heart failure(CHANCE-HF). Study design. Rev Esp Cardiol2015;68:121–8.

22. Dickstein K, Cohen-Solal A, Filippatos G,et al. ESC guidelines for the diagnosis andtreatment of acute and chronic heart failure2008: the Task Force for the Diagnosis andTreatment of Acute and Chronic Heart Failure2008 of the European Society of Cardiology.Developed in collaboration with the HeartFailure Association of the ESC (HFA) andendorsed by the European Society of IntensiveCare Medicine (ESICM). Eur Heart J 2008;29:2388–442.

23. McMurray JJ, Adamopoulos S, Anker SD, et al.ESC Guidelines for the diagnosis and treatment ofacute and chronic heart failure 2012: the TaskForce for the Diagnosis and Treatment of Acuteand Chronic Heart Failure 2012 of the EuropeanSociety of Cardiology. Developed in collaborationwith the Heart Failure Association (HFA) of theESC. Eur Heart J 2012;33:1787–847.

24. Royston P, Parmar MK. The use of restrictedmean survival time to estimate the treatment ef-fect in randomized clinical trials when the pro-portional hazards assumption is in doubt. Stat Med2011;30:2409–21.

25. Ghosh D, Lin DY. Nonparametric analysis ofrecurrent events and death. Biometrics 2000;56:554–62.

26. Rogers JK, Pocock SJ, McMurray JJ, et al.Analysing recurrent hospitalizations in heart fail-ure: a review of statistical methodology, withapplication to CHARM-Preserved. Eur J Heart Fail2014;16:33–40.

27. Rizopoulos D. JointModels for Longitudinal andTime-to-Event Data: With Applications in R. BocaRaton, FL: Chapman and Hall/CRC Press; 2012.

28. McMurray JJ, Kjekshus J, Gullestad L, et al.Effects of statin therapy according to plasma high-sensitivity C-reactive protein concentration in theControlled Rosuvastatin Multinational Trial inHeart Failure (CORONA): a retrospective analysis.Circulation 2009;120:2188–96.

KEY WORDS biomarker guided-therapy,carbohydrate antigen-125, clinical outcomes,clinical trial, heart failure

APPENDIX For a complete list of theCHANCE-HF investigators and an expandedMethods section as well as supplementalfigures and a table, please see the onlineversion of this article.





















































carbohydrate antigen-125 guided therapy in acute heart failure · mini-focus issue: novel...

Documents