Full Terms & Conditions of access and use can be found athttp://www.tandfonline.com/action/journalInformation?journalCode=iann20

Download by: [Hospital Universitari Vall d'Hebron], [ Jose Barrabes] Date: 03 December 2015, At: 01:38

Annals of Medicine

ISSN: 0785-3890 (Print) 1365-2060 (Online) Journal homepage: http://www.tandfonline.com/loi/iann20

Plasma bilirubin values on admission andventricular remodeling after a first anterior ST-segment elevation acute myocardial infarction

Berta Miranda, José A. Barrabés, Jaume Figueras, Victor Pineda, JoséRodríguez-Palomares, Rosa-Maria Lidón, Antonia Sambola, Jordi Bañeras,Imanol Otaegui & David García-Dorado

To cite this article: Berta Miranda, José A. Barrabés, Jaume Figueras, Victor Pineda, JoséRodríguez-Palomares, Rosa-Maria Lidón, Antonia Sambola, Jordi Bañeras, Imanol Otaegui &David García-Dorado (2015): Plasma bilirubin values on admission and ventricular remodelingafter a first anterior ST-segment elevation acute myocardial infarction, Annals of Medicine,DOI: 10.3109/07853890.2015.1112027

To link to this article: http://dx.doi.org/10.3109/07853890.2015.1112027

Published online: 02 Dec 2015.

Submit your article to this journal

View related articles

View Crossmark data

ANNALS OF MEDICINE, 2015http://dx.doi.org/10.3109/07853890.2015.1112027

ORIGINAL ARTICLE

Plasma bilirubin values on admission and ventricular remodeling after a firstanterior ST-segment elevation acute myocardial infarction

Berta Mirandaa, Jose A. Barrabesa, Jaume Figuerasa, Victor Pinedab, Jose Rodrıguez-Palomaresa,Rosa-Maria Lidona, Antonia Sambolaa, Jordi Banerasa, Imanol Otaeguia and David Garcıa-Doradoa

aServicio de Cardiologıa, Hospital Universitari Vall d’Hebron, VHIR, Universitat Autonoma de Barcelona, Barcelona, Spain; bServicio deRadiologıa, Hospital Universitari Vall d’Hebron, VHIR, Universitat Autonoma de Barcelona, Barcelona, Spain

ABSTRACTIntroduction Bilirubin may elicit cardiovascular protection and heme oxygenase-1 overexpressionattenuated post-infarction ventricular remodeling in experimental animals, but the associationbetween bilirubin levels and post-infarction remodeling is unknown.Materials and methods In 145 patients with a first anterior ST-segment elevation acute myocardialinfarction (STEMI), we assessed whether plasma bilirubin on admission predicted adverseremodeling (left ventricular end-diastolic volume [LVEDV] increase�20% between discharge and6 months, estimated by magnetic resonance imaging).Results Patients’ baseline characteristics and management were comparable among bilirubintertiles. LVEDV increased at 6 months (P50.001) with respect to the initial exam, but themagnitude of this increase was similar across increasing bilirubin tertiles (10.8 [30.2], 10.1 [22.9], and12.7 [24.3]%, P¼ 0.500). Median (25–75 percentile) bilirubin values in patients with and withoutadverse remodeling were 0.75 (0.60–0.93) and 0.73 (0.60–0.92) mg/dL (P¼ 0.693). Absence of finalTIMI flow grade 3 (odds ratio 3.92, 95% CI 1.12–13.66) and a history of hypertension (2.04, 0.93–4.50), but not admission bilirubin, were independently associated with adverse remodeling.Bilirubin also did not predict the increase in ejection fraction at 6 months.Conclusions Admission bilirubin values are not related to LVEDV or ejection fraction progressionafter a first anterior STEMI and do not predict adverse ventricular remodeling.

KEY MESSAGES

� Bilirubin levels are inversely related to cardiovascular disease, and overexpression of hemeoxygenase-1 (the enzyme that determines bilirubin production) has prevented post-infarctionventricular remodeling in experimental animals, but the association between bilirubin levelsand the progression of ventricular volumes and function in patients with acute myocardialinfarction remained unexplored.

� In this cohort of patients with a first acute anterior ST-segment elevation myocardial infarctionreceiving contemporary management, bilirubin levels on admission were not predictive of thechanges in left ventricular volumes or ejection fraction at 6 months measured by serial cardiacmagnetic resonance imaging.

� The data are contrary to a significant protective effect of bilirubin against post-infarctionventricular remodeling.

ARTICLE HISTORY

Received 23 July 2015Revised 13 October 2015Accepted 20 October 2015Published online30 November 2015

KEYWORDSBilirubin; heme oxygenase;myocardial infarction;remodeling

Introduction

Bilirubin, the end product of heme catabolism, is

produced in a large part during red blood cell turnover,

conjugated in the liver, and secreted into the bile.

Bilirubin has antioxidant, anti-inflammatory, and anti-

proliferative effects, and there is evidence that it

may protect against cardiovascular disease (1,2).

Following the first report showing that bilirubin is a

potent antioxidant (1), its plasma levels have been

inversely associated with atherosclerotic disease or

cardiovascular events in many studies (3–11), although

such association was not present or no longer significant

after adjusting for classical cardiovascular risk factors in

other studies (12–14). In addition, heme oxygenase (HO)-

1, the enzyme that splits heme rings into biliverdin,

CONTACT Jose A. Barrabes, MD jabarrabes@vhebron.net Unidad Coronaria, Servicio de Cardiologıa, Hospital Universitari Vall d’Hebron, P. Vall d’Hebron119-129, 08035 Barcelona, Spain

� 2015 Taylor & Francis

Dow

nloa

ded

by [

Hos

pita

l Uni

vers

itari

Val

l d'H

ebro

n], [

Jose

Bar

rabe

s] a

t 01:

38 0

3 D

ecem

ber

2015

carbon monoxide, and ferrous iron, a critical and rate-

limiting step of bilirubin production, has been shown to

exert direct protective effects against vascular dysfunc-

tion and atherosclerosis (15).

The potentially beneficial effects of the HO/bilirubin

system are not limited to the vasculature. HO inhibitors

or HO-2 deletion blocked the neuroprotective effects of

a phorbol ester against hydrogen peroxide-induced

injury in neuron cultures, whereas bilirubin at nanomolar

concentrations reduced cell death in the same model

(16). In addition, brain damage following transient

middle cerebral artery occlusion or intracranial injections

of a neurotoxic agent was attenuated in HO-2-deficient

mice (17). Perhaps in relation in part with these effects,

bilirubin levels have been inversely associated with

stroke outcomes in humans (8). In the heart, myocardial

HO-1 overexpression in experimental animals has been

shown to reduce infarct size after transient global or

regional ischemia (18–22), to reduce interstitial fibrosis

after permanent coronary occlusion (23), and to attenu-

ate adverse ventricular remodeling after transient cor-

onary occlusion in vivo (20,24,25). However, to what

extent constitutive HO/bilirubin system activity may

influence post-infarction ventricular remodeling in

patients with ST-segment elevation acute myocardial

infarction (STEMI) is unknown. Accordingly, we aimed to

assess whether bilirubin values on admission predict the

change in left ventricular volumes and systolic function

in patients with a first anterior STEMI receiving contem-

porary therapy.

Materials and methods

Patients

The study sample consisted of 145 patients admitted to

our acute cardiac care unit with a first anterior STEMI

who were studied by cardiac magnetic resonance

imaging (MRI) prior or early after discharge and after 6

months. A significant part of the cases were selected

from three cohorts of patients originally aimed to study

mechanisms of myocardial damage or strategies of

myocardial protection after STEMI (26,27). Inclusion of

patients ranged from 2005 to 2014. Study protocols of

the parent studies were approved by our local ethics

committee (codes PR-AG 110/2005, ACR-AG 182/2007,

and PR-AG 282/2015), and patients gave informed

consent. No additional approval was deemed necessary

for this subanalysis.

Anterior STEMI was diagnosed by the occurrence of

anginal pain or angina-equivalent symptoms that per-

sisted after nitroglycerin administration and were

associated with ST-segment elevation�0.2 mV in�2

consecutive anterior leads (V1 to V4) and with a

significant elevation of cardiac biomarkers (creatine

kinase-MB or troponin I or T). Patients were not included

if they had a history of previous infarction, significant

myocardial or valvular disease or liver disease, or in case

of persistent clinical instability after hospital admission,

persistent atrial fibrillation or other arrhythmias, claus-

trophobia or other contraindications for MRI studies, or

unwillingness to participate.

Patients’ management was in accordance with the

guidelines at the time of inclusion (28–30). Primary

percutaneous coronary intervention (PCI) was attempted

in all cases arriving512 h after onset of symptoms and

rescue PCI in cases of failed pre-hospital thrombolysis. In

the few cases not catheterized immediately, elective

coronary angiography and PCI were performed in the

presence of high-risk clinical features or spontaneous or

induced ischemia before 2008 and systematically there-

after, according to the change in the guidelines in this

respect (29). In addition to antithrombotic therapy and

statins, all patients received angiotensin-converting

enzyme inhibitors or angiotensin receptor blockers and

beta-blocking agents unless contraindicated or not

tolerated, and aldosterone antagonists in selected cases.

Clinical, angiographic, and laboratory data

collection

Demographic data, risk factors for coronary artery

disease, major co-morbidities, clinical data during

hospitalization, and pharmacological therapy were

prospectively collected. Data on risk factors and

co-morbidities were obtained by patient report and by

assessment of previous medical history and therapies.

Severe renal failure was defined as plasma creatinine

values�2.0 mg/dL, need of renal replacement therapy,

or history of kidney transplant.

Coronary angiography was performed according to

the local protocol of our catheterization laboratory. The

culprit lesion was identified by the presence of throm-

bus, total occlusion, or delayed anterograde flow. PCI

with stenting of a significant culprit lesion was always

attempted. Anterograde flow in the culprit vessel prior

and after PCI was characterized using the Thrombolysis

in Myocardial Infarction (TIMI) scale (31). The extent of

coronary artery disease was assessed as the number of

coronary arteries with lesions�70% or left main lesions

�50%. Non-culprit PCI of distant lesions was occasion-

ally performed, in general in a staged procedure,

according to the judgement of the treating physicians.

Blood levels of creatine kinase-MB (CK-MB) were

determined upon admission, at 4–6 h intervals during

the first 24 h, and at 12-h intervals between 24 and 48 h

after admission, and peak CK-MB was recorded.

2 B. MIRANDA ET AL.

Dow

nloa

ded

by [

Hos

pita

l Uni

vers

itari

Val

l d'H

ebro

n], [

Jose

Bar

rabe

s] a

t 01:

38 0

3 D

ecem

ber

2015

The morning after admission a fasting blood sample was

obtained, and total and conjugated bilirubin levels were

measured at the central laboratory of the hospital. For

the purpose of this study, only total bilirubin was

considered.

Cardiac magnetic resonance imaging

Cardiac MRI studies were performed with 1.5 T equip-

ment (Siemens Avanto, Erlangen, Germany). Imaging

was performed in synchronization with the ECG and in

apnea. Short-axis, 4-chamber and 2-chamber cine-MRI

views (SSFP sequences, slice thickness 8 mm, space

between slices 20%, matrix 256� 256, field of view 300–

370 mm, temporal resolution550 ms) were performed to

calculate ventricular volumes, function, and mass. In the

initial exam, late enhancement sequences (matrix

256� 256, field of view 300–370 mm, optimal inversion

time to suppress the myocardium signal) were also

obtained 15 min after iv administration of 0.2 mmol/kg

dimeglumine gadopentetate (Magnevist, Bayer

HealthCare) to quantify the size of the infarction.

Studies were analyzed on a workstation (QMASS MR

7.2, Medis Medical Imaging Systems, The Netherlands)

by an experienced operator unaware of bilirubin values.

Endocardial and epicardial borders were delineated at

end-diastole and end-systole on short-axis views, and

left ventricular (LV) end-diastolic and end-systolic vol-

umes (LVEDV, LVESV) and ejection fraction (LVEF) were

calculated. Infarct size was quantified in the initial study

by delineating the enhanced areas in the late sequences

with 5 SD above average, obtained from the remote

healthy myocardium and normalized by LV mass.

Hypoenhanced zones inside enhanced areas, suggesting

microvascular obstruction, were included in the infarct

volume. Adverse left ventricular remodeling was defined

as an increase in LVEDV�20% at 6 months with respect

to baseline (32).

Statistical analysis

Continuous variables are reported as means (SD) or as

median (25–75 percentiles), and categorical variables as

counts and percentages. Associations between two

continuous variables were performed by lineal regres-

sion. Between-group differences were assessed by

Student’s t test or the Mann–Whitney U test, when

appropriate, for two independent variables, and by

paired t test or the Wilcoxon rank sum test for two

related variables. Patients were divided into three

tertiles based on total bilirubin levels on admission:

bilirubin50.64;�0.64 and50.87; and�0.87 mg/dL.

Differences across groups defined by these increasing

tertiles were assessed by the Jonckheere Terpstra test,

for continuous variables, and by the chi-square test for

trend, for categorical variables. The main analyses were

repeated in the subgroup with abnormally high bilirubin

levels (41.2 mg/dL) or in those with larger infarctions

(infarct size above the median). Multivariate logistic

regression analysis was performed to identify independ-

ent predictors of adverse remodeling. In addition to total

bilirubin in tertiles, variables previously associated with

ventricular remodeling or that were associated with this

outcome in univariate analyses (sex, history of hyper-

tension or diabetes, infarct size, and absence of final TIMI

flow grade 3 after PCI) were included in the model.

Multiple regression analysis was performed to assess the

association between bilirubin values with LVEF change

after adjusting for these potentially relevant predictors.

Statistical analysis was performed with SPSS 13.0 soft-

ware. P values50.05 were considered significant.

Results

Patient characteristics

The main demographic and baseline clinical character-

istics of the patients are summarized in Table I. Female

sex was more frequent among patients with lower

bilirubin values. There were no differences across biliru-

bin tertiles with respect to the prevalence of cardiovas-

cular risk factors, co-morbidities, previous history of

coronary artery disease, or ongoing treatments with

cardiovascular drugs. Data on clinical presentation on

admission, angiographic characteristics, in-hospital

course and therapy, and treatment at discharge were

also similar among groups (Table II). Most patients

received reperfusion therapy, mainly primary PCI, 95.9%

were catheterized, and 91.7% of the overall sample

underwent PCI during admission. Six patients were

discharged without a coronary angiography after suc-

cessful thrombolysis, and six patients that were cathe-

terized did not undergo PCI because of the finding of a

non-significant culprit lesion. Among the 139 catheter-

ized patients, 127 (91.4%) had TIMI flow grade 3 at the

culprit artery at the end of the procedure, without

between-group differences. Most patients were dis-

charged from the hospital on beta-blocking agents

and on angiotensin-converting enzyme inhibitors or

angiotensin receptor blockers, whereas aldosterone

antagonists were rarely prescribed. One patient under-

went coronary artery bypass surgery after discharge.

Table III summarizes the main laboratory data. Median

total bilirubin value in the fasting sample upon admis-

sion was 0.74 (0.60–0.92, range 0.32–4.14) mg/dL.

Twenty patients (13.8%) had bilirubin values above the

upper limit of normality, but none had a significant

ANNALS OF MEDICINE 3

Dow

nloa

ded

by [

Hos

pita

l Uni

vers

itari

Val

l d'H

ebro

n], [

Jose

Bar

rabe

s] a

t 01:

38 0

3 D

ecem

ber

2015

elevation of conjugated bilirubin values or other labora-

tory data suggesting liver disease. Laboratory data other

than bilirubin values were comparable across bilirubin

tertiles.

Cardiac magnetic resonance imaging results

All patients underwent cardiac MRI before or early after

discharge (8 [4–22] days after admission) and after 6

months (6 [6–7] months after admission). The main

Table II. Clinical presentation, angiographic characteristics, and hospital course and management of the patientsamong total bilirubin tertiles.

Tertile 1 Tertile 2 Tertile 3 P value

Time symptom onset–admission, h 2.2 (1.5–5.8) 1.8 (1.0–2.9) 1.8 (0.6–4.0) 0.116Initial systolic blood pressure, mmHg 134 (30) 142 (25) 139 (28) 0.268Initial diastolic blood pressure, mmHg 81 (18) 83 (15) 85 (19) 0.247Initial heart rate, beats/min 81 (19) 77 (17) 77 (20) 0.260Killip class I–II 48 (94.1%) 44 (97.8%) 48 (98.0%) 0.292Reperfusion therapy 47 (92.2%) 43 (95.6%) 47 (95.9%) 0.409Thrombolysis 9 (17.6%) 8 (17.8%) 8 (16.3%) 0.863Primary PCI 38 (74.5%) 35 (77.8%) 39 (79.6%) 0.545Rescue PCI 4 (7.8%) 1 (2.2%) 4 (8.2%) 0.959Coronary angiography 51 (100%) 41 (91.1%) 47 (95.9%) 0.297PCI, total 48 (94.1%) 40 (88.9%) 44 (91.7%) 0.650Initial TIMI flow grade 0–1a 40 (78.4%) 35 (85.4%) 38 (80.9%) 0.746Final TIMI flow grade 3a 47 (92.2%) 39 (95.1%) 41 (87.2%) 0.400Multivessel diseasea 22 (43.1%) 20 (48.8%) 13 (27.7%) 0.127Reinfarction 1 (2.0) 0 (0%) 0 (0%) 0.594Pulmonary edema 4 (7.8%) 1 (2.2%) 3 (6.1%) 0.698Cardiogenic shock 3 (5.9%) 2 (4.4%) 1 (2.0%) 0.337Ventricular fibrillation 2 (3.9%) 4 (8.9%) 5 (10.2%) 0.235Advanced auriculoventricular block 2 (3.9%) 0 (0%) 0 (0%) 0.092Invasive mechanical ventilation 3 (5.9%) 1 (2.2%) 1 (2.0%) 0.292Glycoprotein IIb/IIIa inhibitors 30 (63.8%) 25 (59.5%) 33 (68.8%) 0.615Dual antiplatelet therapy at discharge 49 (96.1%) 39 (86.7%) 45 (91.8%) 0.433Oral anticoagulants at discharge 2 (3.9%) 0 (0%) 3 (6.1%) 0.559Beta-blockers at discharge 41 (80.4%) 42 (93.3%) 43 (87.8%) 0.270ACE inhibitors/ARB at discharge 37 (72.5%) 33 (73.3%) 36 (73.5%) 0.917Aldosterone inhibitors at discharge 7 (13.7%) 3 (6.7%) 6 (12.2%) 0.804Statins at discharge 42 (82.4%) 40 (88.9%) 43 (87.8%) 0.431

Values are means (SD), medians (25–75 percentiles), or counts (%).aOf patients catheterized.ACE¼ angiotensin-converting enzyme; ARB¼ angiotensin receptor blockers; CABG¼ coronary artery bypass surgery;

PCI¼ percutaneous coronary intervention; TIMI¼ Thrombolysis in Myocardial Infarction.

Table I. Demographic and baseline clinical characteristics of the patients among total bilirubin tertiles.

Tertile 1 Tertile 2 Tertile 3 P value

Age, years 62 (13) 61 (13) 61 (13) 0.550Male sex 35 (68.6%) 42 (93.3%) 45 (91.8%) 0.001Body mass index, kg/m2 28.2 (3.8) 27.0 (3.3) 27.2 (3.5) 0.234Active smoking 26 (52.0%) 29 (64.4%) 20 (40.8%) 0.272Hypertension 32 (62.7%) 22 (48.9%) 27 (55.1%) 0.436Diabetes mellitus 10 (19.6%) 5 (11.1%) 7 (14.3%) 0.454Dyslipidemia 24 (47.1%) 22 (48.9%) 23 (46.9%) 0.992Previous angina 3 (5.9%) 6 (13.3%) 4 (8.2%) 0.681Previous PCI 2 (3.9%) 2 (4.4%) 2 (4.1%) 0.967Previous CABG 0 (0%) 0 (0%) 0 (0%) –Previous heart failure 1 (2.0%) 1 (2.2%) 0 (0%) 0.406Previous stroke 3 (5.9%) 2 (4.4%) 2 (4.1%) 0.674Peripheral artery disease 4 (7.8%) 2 (4.4%) 2 (4.1%) 0.409Severe chronic renal failure 1 (2.0%) 1 (2.2%) 0 (0%) 0.406Pulmonary disease 9 (18.0%) 7 (15.9%) 10 (20.4%) 0.759Previous antiplatelet agents 7 (14.0%) 3 (6.8%) 2 (4.4%) 0.096Previous beta-blockers 4 (10.3%) 4 (10.3%) 3 (7.0%) 0.602Previous ACE inhibitors/ARB 14 (35.9%) 8 (20.5%) 10 (23.3%) 0.207Previous aldosterone inhibitors 0 (0%) 0 (0%) 0 (0%) –Previous diuretics 9 (23.1%) 4 (10.0%) 6 (14.0%) 0.271Previous oral antidiabetic drugs 5 (9.8%) 5 (11.1%) 4 (8.3%) 0.811Previous insulin 4 (7.8%) 0 (0%) 0 (0%) 0.017Previous statins 16 (32.7%) 8 (18.2%) 9 (19.1%) 0.118

Values are means (SD), or counts (%).ACE¼ angiotensin-converting enzyme; ARB¼ angiotensin receptor blockers; CABG¼ coronary artery bypass surgery;

PCI¼ percutaneous coronary intervention.

4 B. MIRANDA ET AL.

Dow

nloa

ded

by [

Hos

pita

l Uni

vers

itari

Val

l d'H

ebro

n], [

Jose

Bar

rabe

s] a

t 01:

38 0

3 D

ecem

ber

2015

results of both cardiac MRI exams are summarized in

Table IV. Overall, calculated LV mass decreased signifi-

cantly between both exams, whereas LVEDV increased

by 11.3 (26.0)%, LVESV remained unchanged, and there

was a 4.6 (9.4)% increase in LVEF at 6 months with

respect to baseline. Significant LV remodeling occurred

in 42 (29.2%) patients. Infarct size in the initial exam

averaged 27.4 (19.0) g, which represented 20.6 (13.1)%

of LV mass. Infarct size was not significantly different in

patients with and without progression to adverse

remodeling (23.4 [13.1] and 19.4 [13.0]% of LV mass,

respectively, P¼ 0.099).

Association between bilirubin levels and the

progression of left ventricular volumes andfunction

Infarct size was not correlated with bilirubin levels on

admission (r¼ 0.087, P¼ 0.307, for absolute infarct mass;

r¼ 0.330, P¼ 0.699, for infarct size as percent of LV mass)

and was similar across increasing bilirubin tertiles (20.1

[12.5], 19.6 [13.1], and 22.0 [13.8]% of LV mass, respect-

ively, P¼ 0.474). Bilirubin levels were also not associated

with LVEDV change at 6 months with respect to the

initial values (r¼ 0.073, P¼ 0.38, for the absolute differ-

ence; r¼ 0.022, P¼ 0.793, for the relative change), with

LVESV progression, or with LVEF change (r¼ –0.102,

P¼ 0.224, for the absolute change; r¼ –0.107, P¼ 0.201,

for the relative change) (Figure 1). The increase in LVEDV

at 6 months with respect to the first exam was of similar

magnitude across increasing tertiles of bilirubin values

on admission (10.8 [30.2], 10.1 [22.9], and 12.7 [24.3]%,

respectively, P¼ 0.500), as was the increase in LVEF

(Figure 2). The progression of LVEDV and LVEF at 6

months with respect to the baseline exam also was not

more favorable in the 20 patients with abnormally high

bilirubin levels than in the remaining patients (7.1 [22.0]

versus 11.9 [26.6]%, respectively, P¼ 0.449, for the

relative increase in LVEDV; 0.8 [7.4] versus 5.2 [9.6]%,

respectively, P¼ 0.054, for the absolute change in LVEF).

Median bilirubin values in patients with and without

adverse remodeling were 0.75 (0.60–0.93) and 0.73

(0.60–0.92) mg/dL, respectively (P¼ 0.693). Bilirubin

levels were also not associated with LVEDV or LVEF

change or with adverse LV remodeling if the analyses

were restricted to the subgroup of patients with larger

infarctions. Absence of final TIMI 3 flow grade (odds ratio

[OR] 3.92, 95% CI 1.12–13.66, P¼ 0.032) and a history of

hypertension (OR 2.04, 95% CI 0.93–4.50, P¼ 0.076), but

not bilirubin, were independent predictors of adverse

remodeling by logistic regression analysis (Figure 3).

Neither bilirubin nor any other variable was found to be

Table III. Laboratory data among tertiles of total bilirubin.

Tertile 1 Tertile 2 Tertile 3 P value

Baseline hemoglobin, g/dL 14.0 (1.8) 14.8 (1.6) 14.7 (1.7) 0.069Baseline leukocyte count/mL 12218 (3706) 12000 (2803) 12587 (4114) 0.379Baseline glucose, mg/dL 132 (96–148) 149 (107–181) 156 (100–184) 0.840Baseline creatinine, mg/dL 0.87 (0.27) 0.95 (0.23) 0.91 (0.20) 0.077Baseline ASAT, IU/L 47 (28–101) 61 (48–179) 62 (18–141) 0.207Baseline ALAT, IU/L 31 (18–56) 44 (26–48) 27 (17–46) 0.171Fasting total bilirubin, mg/dL 0.50 (0.35–0.62) 0.75 (0.70–0.82) 1.12 (0.92–1.38) 50.001Fasting conjugated bilirubin, mg/dL 0.23 (0.20–0.25) 0.30 (0.28–0.32) 0.40 (0.36–0.48) 50.001Fasting alkaline phosphatase, UI/L 69.6 (23.3) 76.1 (27.2) 68.2 (13.2) 0.614Fasting gamma-glutaryl transpeptidase, UI/L 41 (22–94) 34 (21–104) 30 (24–44) 0.779Fasting cholesterol, mg/dL 176.9 (40.9) 185.7 (38.1) 186.2 (56.3) 0.781Fasting LDL-cholesterol, mg/dL 112.5 (31.8) 111.5 (33.5) 112.1 (40.0) 0.566Peak creatine kinase-MB, ng/mL 301.7 (232.2) 317.9 (187.1) 334.4 (258.2) 0.520

Values are means (SD) or medians (25–75 percentiles).ASAT¼ aspartate aminotransferase; ALAT¼ alanine aminotransferase; LDL¼ low-density lipoprotein.

Table IV. Main results of cardiac magnetic resonance imaging.

Initial MRI 6-months MRI P value

LV mass, g 128.0 (111.8–144.1) 119.5 (100.2–137.5) 50.001LV mass, g/m2 66.2 (60.2–76.3) 62.9 (53.8–70.5) 50.001LVEDV, mL 148.8 (38.0) 162.7 (48.2) 50.001LVEDV, mL/m2 78.2 (19.0) 85.7 (24.9) 50.001LVESV, mL 81.3 (31.6) 83.0 (41.2) 0.445LVESV, mL/m2 42.8 (16.5) 43.8 (21.7) 0.403LVEF, % 46.7 (10.8) 51.3 (12.0) 50.001Infarct size, % of LV mass 20.6 (13.1) – –

Values are means (SD), or medians (25–75 percentiles).LV¼ left ventricular; LVEDV¼ LV end-diastolic volume; LVEF¼ LV ejection fraction; LVESV¼ LV end-systolic volume;

MRI¼magnetic resonance imaging.

ANNALS OF MEDICINE 5

Dow

nloa

ded

by [

Hos

pita

l Uni

vers

itari

Val

l d'H

ebro

n], [

Jose

Bar

rabe

s] a

t 01:

38 0

3 D

ecem

ber

2015

independently associated with LVEF change by multiple

regression analysis.

Discussion

In this study in a large cohort of patients with a first

anterior STEMI receiving contemporary management

and evaluated by serial cardiac MRI, plasma bilirubin

levels on admission were not related to the progression

of LVEDV or LVEF at 6 months and did not predict

adverse LV remodeling. Absence of final TIMI flow grade

3 and a history of hypertension, but not bilirubin values,

predicted a poorer LVEDV progression in multivariable

analysis. The results are contrary to a significant

protective effect of bilirubin against post-infarction

ventricular remodeling.

Many studies have shown an inverse association

between bilirubin levels and different manifestations of

atherosclerosis or cardiovascular disease (1–11).

However, their association with post-infarction ventricu-

lar remodeling remained unexplored despite the fact

that experimental studies have reported direct cytopro-

tective effects of bilirubin against oxidant injury (16) and

a protective effect of HO-1 overexpression on myocardial

damage after ischemia and reperfusion and on subse-

quent LV remodeling. Yet et al. observed that hearts of

cardiac-specific transgenic mice overexpressing HO-1

showed improved post-ischemic functional recovery

after transient global ischemia in a Langendorff model

and a significantly reduced infarct size after transient

coronary occlusion in vivo as compared with control

mice (18). In agreement with these findings, Melo et al.

showed that adenoviral transfer of the HO-1 gene into

normal rat hearts was related to a reduced infarct size

after subsequent transient coronary occlusion (19).

These latter authors demonstrated in the same model

that HO-1 gene transfer was associated with less LV

dilatation and better recovery of LV systolic function 3

months after transient coronary occlusion as compared

with LacZ-injected control rats (25) and also with an

improved survival at 1 year (33). Recently, Hinkel et al.

confirmed that HO-1 gene transfer also reduced post-

ischemic inflammation and infarct size in pigs subjected

to transient coronary occlusion (22).

Although part of the late beneficial effects of HO-1

overexpression might be merely in consequence of

infarct size reduction, direct anti-inflammatory and anti-

fibrotic effects were also demonstrated. In this respect, a

lower expression of proinflammatory cytokines and

reduced leukocyte influx in the infarct area early after

infarction (20,22) and less fibroblast proliferation and

lower collagen and metalloproteinase-2 levels in the

subacute phase (25) were described in treated animals.

Figure 1. Association between total bilirubin plasma levels on admission and the relative change in left ventricular end-diastolicvolume (LVEDV, left panel) or the absolute change in left ventricular ejection fraction (LVEF, right panel) at 6 months with respect tothe initial cardiac magnetic resonance exam.

Figure 2. Change in left ventricular end-diastolic volume(LVEDV) and left ventricular ejection fraction (LVEF) at 6months with respect to the initial exam across increasingtertiles of plasma bilirubin on admission.

6 B. MIRANDA ET AL.

Dow

nloa

ded

by [

Hos

pita

l Uni

vers

itari

Val

l d'H

ebro

n], [

Jose

Bar

rabe

s] a

t 01:

38 0

3 D

ecem

ber

2015

In the same line, in a model of permanent coronary

occlusion in mice, injection of a HO-1 plasmid into infarct

and border tissue reduced apoptosis and interstitial

fibrosis in the infarct region and improved functional

recovery (23). HO-2 overexpression was also shown to

protect the brain against oxidative stress and/or ische-

mic damage (16,17).

The reasons for the discrepancy between those

previous studies in experimental animals and the

present observations in STEMI patients are unclear,

although several explanations are plausible. On the one

hand, those studies appropriately targeted a single

molecular pathway, but there is a multiplicity of factors

that may be involved in myocardial healing and in the

change in LV volumes and function after infarction,

particularly in complex settings (34). On the other hand,

they were performed in healthy animals subjected to a

homogeneous ischemia-reperfusion protocol, whereas

our patients had variable co-morbidities, size of the

ischemic region, duration of ischemia, treatments

applied, and response to therapy. In addition, the

variability in bilirubin values in our unselected sample

of patients was modest and possibly did not reproduce

that obtained in experimental conditions (16). Finally,

but most importantly, HO-1 overexpression may have

reduced infarct size and protected against adverse

remodeling in those studies by mechanisms independ-

ent of bilirubin formation (15). We attempted to control

in part for these factors by repeating the analyses in the

subset with larger infarctions, by dichotomizing the

sample to isolate patients with the highest bilirubin

levels or with significant LV remodeling, or by perform-

ing multivariable analyses, but the results did not

change.

Our results are in line, however, with other clinical

studies that have not confirmed a significant protective

effect of bilirubin against cardiovascular disease. In a

large cohort of overweight patients at high risk for

cardiovascular disease, baseline bilirubin levels were

inversely associated with the rates of future cardiovas-

cular events but this association was no longer signifi-

cant after adjusting for traditional cardiovascular risk

factors (14). In other studies, high bilirubin levels were

associated with either a more complex coronary artery

disease, a worse TIMI flow grade before or after PCI, or

increased in-hospital events in STEMI patients, which

suggests that these high levels may reflect in part an

acute increase in response to stress or inflammation

(12,13,35,36). Also not at variance with our results, HO-1

gene transfer by itself did not change LV fractional

shortening, volumes, or ejection fraction in mice over 1

year of follow-up (21).

In contrast to bilirubin levels, a persistently impaired

coronary flow at the culprit artery at the end of coronary

angiography/PCI procedure and a history of hyperten-

sion were independent predictors of adverse LV

remodeling, which is consistent with previous studies

(37,38). The fact that infarct size did not significantly

predict adverse remodeling is at variance with previous

observations (39). Given that quantification of infarct

tissue by MRI is dependent on the time of the exam (40),

the significant heterogeneity of the times between

admission and first cardiac MRI exam in our study may

help explain these results. However, this issue probably

did not have a major influence on the lack of association

between bilirubin levels and LV remodeling since the

results were similar in patients with earlier or a more

delayed first MRI (data not shown).

Figure 3. Forest plot illustrating the results of multivariable logistic regression analysis for the prediction of adverse left ventricularremodeling. TIMI¼ Thrombolysis in Myocardial Infarction.

ANNALS OF MEDICINE 7

Dow

nloa

ded

by [

Hos

pita

l Uni

vers

itari

Val

l d'H

ebro

n], [

Jose

Bar

rabe

s] a

t 01:

38 0

3 D

ecem

ber

2015

Ours was a relatively large and homogeneous

series of patients with a first anterior STEMI studied

by serial cardiac MRI and receiving contemporary

management with regard to the use of reperfusion

strategies and pharmacological co-therapy (41).

Although their baseline risk was somewhat lower

than that of unselected patients from the same

geographical area (41), in part due to the inclusion

criteria and the per protocol selection of 6-month

survivors with two serial MRI exams available, infarct

size was significant, as estimated by biomarker

elevation, contrast MRI, and LVEF assessment, and

the proportion developing adverse LV remodeling

was also as expected.

Several limitations need to be acknowledged. Some

important baseline variables were obtained by patient

report or by assessing the clinical records and lacked

uniform definition criteria. Because only patients with

a relatively uncomplicated course were included, the

results may not be completely representative of STEMI

patients. In this respect, the exclusion rates after the

first cardiac MRI exam could be assessed in the last

two parent studies and averaged 23%. More than one

half of these withdrawals were the patient’s decision,

and the rest were due to internal cardioverter defib-

rillator implantation or clinical deterioration, including

a few deaths. We cannot establish the causes of the

abnormally high bilirubin levels present in some

patients. Because liver disease was an exclusion

criterion and no patient had significantly abnormal

levels of conjugated bilirubin or other biochemical

tests suggesting hepatopathy, Gilbert’s syndrome may

have been the underlying cause in some cases.

However, in none of our patients had this diagnosis

been made before the index admission. An additional

limitation is that we did not measure other poten-

tial molecular pathways involved in cardiac

remodeling (34).

In conclusion, no association was found between

admission plasma bilirubin levels and 6-month progres-

sion of LV volumes or systolic function in patients with a

first anterior STEMI receiving contemporary manage-

ment. Our results suggest that bilirubin does not have a

significant influence on LV remodeling after STEMI.

Declaration of interest

The authors report no conflicts of interest. Theauthors alone are responsible for the content and writing ofthe paper.

This study was funded by a grant from the Spanish Ministryof Economy and Competitiveness through the Instituto deSalud Carlos III (PI12/01844) and by the Societat Catalana deCardiologia (Beca Ferrer 2014).

References

1. Stocker R, Yamamoto Y, McDonagh AF, Glazer AN, AmesBN. Bilirubin is an antioxidant of possible physiologicalimportance. Science. 1987;235:1043–6.

2. Kang SJ, Lee C, Kruzliak P. Effects of serum bilirubin onatherosclerotic processes. Ann Med. 2014;46:138–47.

3. Schwertner HA, Jackson WG, Tolan G. Association of lowserum concentration of bilirubin with increased risk ofcoronary artery disease. Clin Chem. 1994;40:18–23.

4. Breimer LH, Wannamethee G, Ebrahim S, Shaper AG.Serum bilirubin and risk of ischemic heart disease inmiddle-aged British men. Clin Chem. 1995;41:1504–8.

5. Djousse L, Levy D, Cupples LA, Evans JC, D’Agostino RB,Ellison RC. Total serum bilirubin and risk of cardiovasculardisease in the Framingham offspring study. Am J Cardiol.2001;87:1196–200.

6. Erdogan D, Gullu H, Yildirim E, Tok D, Kirbas I, Ciftci O,et al. Low serum bilirubin levels are independently andinversely related to impaired flow-mediated vasodilationand increased carotid intima-media thickness in both menand women. Atherosclerosis. 2006;184:431–7.

7. Perlstein TS, Pande RL, Beckman JA, Creager MA. Serumtotal bilirubin level and prevalent lower-extremity periph-eral arterial disease: National Health and NutritionExamination Survey (NHANES) 1999 to 2004. ArteriosclerThromb Vasc Biol. 2008;28:166–72.

8. Perlstein TS, Pande RL, Creager MA, Weuve J, Beckman JA.Serum total bilirubin level, prevalent stroke, andstroke outcomes: NHANES 1999-2004. Am J Med.2008;121:781–8.e1.

9. Horsfall LJ, Nazareth I, Petersen I. Cardiovascular events asa function of serum bilirubin levels in a large, statin-treated cohort. Circulation. 2012;126:2556–64.

10. Kang SJ, Kim D, Park HE, Chung GE, Choi SH, Choi SY, et al.Elevated serum bilirubin levels are inversely associatedwith coronary artery atherosclerosis. Atherosclerosis.2013;230:242–8.

11. Ryu S, Chang Y, Zhang Y, Woo HY, Kwon MJ, Park H, et al.Higher serum direct bilirubin levels were associated with alower risk of incident chronic kidney disease in middleaged Korean men. PLoS One. 2014;9:e75178.

12. Gul M, Uyarel H, Ergelen M, Akgul O, Karaca G, Turen S,et al. Prognostic value of total bilirubin in patients with ST-segment elevation acute myocardial infarction undergo-ing primary coronary intervention. Am J Cardiol.2013;111:166–71.

13. Sahin O, Akpek M, Elcik D, Karadavut S, Simsek V,Tulmac M, et al. Bilirubin levels and the burden of coronaryatherosclerosis in patients with STEMI. Angiology.2013;64:200–4.

14. Jørgensen ME, Torp-Pedersen C, Finer N, Caterson I, JamesWPT, Legler UF, et al. Association between serum bilirubinand cardiovascular disease in an overweight high riskpopulation from the SCOUT trial. Nutr Metab CardiovascDis. 2014;24:656–62.

15. Stocker R, Perrella MA. Heme oxygenase-1: a novel drugtarget for atherosclerotic diseases? Circulation.2006;114:2178–89.

16. Dore S, Takahashi M, Ferris CD, Zakhary R, Hester LD,Guastella D, et al. Bilirubin, formed by activation of hemeoxygenase-2, protects neurons against oxidative stressinjury. Proc Natl Acad Sci U S A. 1999;96:2445–50.

8 B. MIRANDA ET AL.

Dow

nloa

ded

by [

Hos

pita

l Uni

vers

itari

Val

l d'H

ebro

n], [

Jose

Bar

rabe

s] a

t 01:

38 0

3 D

ecem

ber

2015

17. Dore S, Sampei K, Goto S, Alkayed NJ, Guastella D,Blackshaw S, et al. Heme oxygenase-2 is neuroprotectivein cerebral ischemia. Mol Med. 1999;5:656–63.

18. Yet SF, Tian R, Layne MD, Wang ZY, Maemura K, SolovyevaM, et al. Cardiac-specific expression of heme oxygenase-1protects against ischemia and reperfusion injury in trans-genic mice. Circ Res. 2001;89:168–73.

19. Melo LG, Agrawal R, Zhang L, Rezvani M, Mangi AA, EhsanA, et al. Gene therapy strategy for long-term myocardialprotection using adeno-associated virus-mediated deliv-ery of heme oxygenase gene. Circulation. 2002;105:602–7.

20. Pachori AS, Melo LG, Hart ML, Noiseux N, Zhang L, MorelloF, et al. Hypoxia-regulated therapeutic gene as apreemptive treatment strategy against ischemia/reperfu-sion tissue injury. Proc Natl Acad Sci U S A.2004;101:12282–7.

21. Li Q, Guo Y, Ou Q, Wu WJ, Chen N, Zhu X, et al. Genetransfer as a strategy to achieve permanent cardioprotec-tion II: rAAV-mediated gene therapy with hemeoxygen-ase-1 limits infarct size 1 year later withoutadverse functional consequences. Basic Res Cardiol.2011;106:1367–77.

22. Hinkel R, Lange P, Petersen B, Gottlieb E, Ng JKM, Finger S,et al. Heme oxygenase-1 gene therapy provides cardio-protection via control of post-ischemic inflammation: anexperimental study in a pre-clinical pig model. J Am CollCardiol. 2015;66:154–65.

23. Tang YL, Tang Y, Zhang YC, Qian K, Shen L, Phillips MI.Protection from ischemic heart injury by a vigilant hemeoxygenase-1 plasmid system. Hypertension. 2004;43:746–51.

24. Tang YL, Qian K, Zhang YC, Shen L, Phillips MI. A vigilant,hypoxia-regulated heme oxygenase-1 gene vector in theheart limits cardiac injury after ischemia-reperfusion invivo. J Cardiovasc Pharmacol Ther. 2005;10:251–63.

25. Liu X, Pachori AS, Ward CA, Davis JP, Gnecchi M, Kong D,et al. Heme oxygenase-1 (HO-1) inhibits postmyocardialinfarct remodeling and restores ventricular function.FASEB J. 2006;20:207–16.

26. Rodrıguez-Palomares JF, Figueras-Bellot J, Descalzo M,Moral S, Otaegui I, Pineda V, et al. Relation of ST-segmentelevation before and after percutaneous transluminalcoronary angioplasty to left ventricular area at risk,myocardial infarct size, and systolic function. Am JCardiol. 2014;113:593–600.

27. Garcia-Dorado D, Garcıa-del-Blanco B, Otaegui I,Rodrıguez-Palomares J, Pineda V, Gimeno F, et al.Intracoronary injection of adenosine before reperfusionin patients with ST-segment elevation myocardial infarc-tion: a randomized controlled clinical trial. Int J Cardiol.2014;177:935–41.

28. Van de Werf F, Ardissino D, Betriu A, Cokkinos DV, Falk E,Fox KA, et al. Management of acute myocardial infarctionin patients presenting with ST-segment elevation. TheTask Force on the Management of Acute MyocardialInfarction of the European Society of Cardiology. Eur HeartJ. 2003;24:28–66.

29. Van de Werf F, Bax J, Betriu A, Blomstrom-Lundqvist C,Crea F, Falk V, et al. Management of acute myocar-dial infarction in patients presenting with persistent

ST-segment elevation: the Task Force on theManagement of ST-Segment Elevation Acute MyocardialInfarction of the European Society of Cardiology. Eur HeartJ. 2008;29:2909–45.

30. Steg PG, James SK, Atar D, Badano LP, Blomstrom-Lundqvist C, Borger MA, et al. ESC Guidelines for themanagement of acute myocardial infarction in patientspresenting with ST-segment elevation. Eur Heart J.2012;33:2569–619.

31. TIMI Study Group. The Thrombolysis in MyocardialInfarction (TIMI) trial. Phase I findings. N Engl J Med.1985;312:932–6.

32. Flachskampf FA, Schmid M, Rost C, Achenbach S, DeMariaAN, Daniel WG. Cardiac imaging after myocardial infarc-tion. Eur Heart J. 2011;32:272–83.

33. Liu X, Simpson JA, Brunt KR, Ward CA, Hall SR, Kinobe RT,et al. Preemptive heme oxygenase-1 gene delivery revealsreduced mortality and preservation of left ventricularfunction 1 yr after acute myocardial infarction. Am JPhysiol Heart Circ Physiol. 2007;293:H48–59.

34. Frangogiannis NG. The inflammatory response in myocar-dial injury, repair, and remodelling. Nat Rev Cardiol.2014;11:255–65.

35. Celik T, Kaya MG, Akpek M, Yarlioglues M, Sarli B, TopsakalR, et al. Does serum bilirubin level on admission predictTIMI flow grade and in-hospital MACE in patientswith STEMI undergoing primary PCI. Angiology.2014;65:198–204.

36. Acet H, Erts F, Akıl MA, Polat N, Aydın M, Akyuz A, et al. Anovel predictor of infarct-related artery patency beforepercutaneous intervention and in-hospital outcomes forST-segment elevation myocardial infarction patients:serum bilirubin level. Postepy Kardiol Interwencyjnej.2014;10:91–7.

37. Morishima I, Sone T, Okumura K, Tsuboi H, Kondo J,Mukawa H, et al. Angiographic no-reflow phenomenon asa predictor of adverse long-term outcome in patientstreated with percutaneous transluminal coronary angio-plasty for first acute myocardial infarction. J Am CollCardiol. 2000;36:1202–9.

38. Richards AM, Nicholls MG, Troughton RW, Lainchbury JG,Elliott J, Frampton C, et al. Antecedent hypertension andheart failure after myocardial infarction. J Am Coll Cardiol.2002;39:1182–8.

39. Tarantini G, Razzolini R, Cacciavillani L, Bilato C, Sarais C,Corbetti F, et al. Influence of transmurality, infarct size, andsevere microvascular obstruction on left ventricularremodeling and function after primary coronary angio-plasty. Am J Cardiol. 2006;98:1033–40.

40. Rodrıguez-Palomares JF, Ortiz-Perez JT, Lee DC, Bucciarelli-Ducci C, Tejedor P, Bonow RO, et al. Time elapsed aftercontrast injection is crucial to determine infarct transmur-ality and myocardial functional recovery after an acutemyocardial infarction. J Cardiovasc Magn Reson.2015;17:43.

41. Barrabes JA, Bardajı A, Jimenez-Candil J, del NogalSaez F, Bodı V, Basterra N, et al. Prognosis andmanagement of acute coronary syndrome in Spain in2012: the DIOCLES study. Rev Esp Cardiol (Engl Ed).2015;68:98–106.

ANNALS OF MEDICINE 9

Dow

nloa

ded

by [

Hos

pita

l Uni

vers

itari

Val

l d'H

ebro

n], [

Jose

Bar

rabe

s] a

t 01:

38 0

3 D

ecem

ber

2015