Disease Undergoing Cardiopulmonary

Predictors of Postoperative Atrial Fibrillation in Patients with Coronary Artery

Bypass: A Possible Role for MyocardialIschemia and Atrial Inflammation

Maria Lucia Narducci, MD, PhD, Gemma Pelargonio, MD, PhD, Teresa Rio, MD, Milena Leo, MD,

Antonio Di Monaco, MD, Francesco Musaico, MD, Vincenzo Pazzano, MD, Francesco Trotta, MD,

Giovanna Liuzzo, MD, PhD, Anna Severino, PhD, Luigi Marzio Biasucci, MD, Andrea Scapigliati, MD,

Franco Glieca, MD, Franco Cavaliere, MD, Antonio Giuseppe Rebuzzi, MD, Massimo Massetti, MD,

and Filippo Crea, MD

Objective: To evaluate the preoperative presence of C-

reactive protein (CRP) and troponin T(hs-TnT) in patients

with coronary artery disease (CAD) undergoing cardiopul-

monary bypass (CPB) in order to better clarify the role of

atrial inflammation and/or myocardial ischemia in the devel-

opment of postoperative atrial fibrillation (POAF).

Design: Prospective, nonrandomized study.

Setting: University hospital.

Participants: Thirty-eight consecutive ischemic patients

admitted to the authors’ hospital for CAD undergoing

elective on-pump coronary artery bypass grafting (CABG).

Intervention: Elective on-pump CABG.

Measurements and Main Results: Peripheral blood sam-

ples were collected from all patients before and 24 hours

after CABG to assess high sensitive (hs)-CRP and troponin

T (hs-TnT) levels. The patients’ heart rhythm was monitored

by continuous ECG telemetry. Biopsies from the right atrial

appendage were obtained at the beginning of the CABG

procedure in order to perform immunohistochemistry for

Journal of Cardiothoracic and Vascular Anesthesia, Vol ], No ] (Month), 2

CRP and reverse transcription polymerase chain reaction for

CRP mRNA expression.

Fourteen patients out of 38 (36%) developed POAF. Atrial

CRP was found in 31 patients (82%), 10 with POAF and 21 with

sinus rhythm (71% v 87% respectively, p ¼ ns). None of the

atrial samples was positive for CRP mRNA. Atrial CRP did not

correlate with serum hs-CRP levels and with occurrence of

POAF, but with the incidence of diabetes (p ¼ 0.010). Post-

operative hs-TnT levels, but not hs-CRP levels, were identified

as the only predictor of POAF occurrence (p ¼ 0.016).

Conclusions: In patients undergoing CABG, neither periph-

eral nor tissue preoperative CRP levels, but only

postoperative hs-TnT levels, correlated with POAF, suggest-

ing the primary role of an ischemic trigger of atrial fibrillation.

& 2013 Elsevier Inc. All rights reserved.

KEY WORDS: atrial fibrillation, inflammatory mediators,coronary artery bypass grafts, myocardial ischemia,immunochemistry

123456789101112

From the Cardiovascular Sciences Department, Catholic Universityof Sacred Heart, Rome, Italy.

Address reprint requests to Maria Lucia Narducci, Catholic Uni-versity of Sacred Heart, Cardiovascular Sciences Department, Largo AGemelli 8, 00135, Rome, Italy. E-mail: lianarducci@yahoo.it© 2013 Elsevier Inc. All rights reserved.1053-0770/2601-0001$36.00/0http://dx.doi.org/10.1053/j.jvca.2013.06.002

DESPITE MEDICAL AND SURGICAL developments,the incidence of atrial fibrillation (AF) after coronary

artery bypass graft surgery (CABG) is 16% to 50%,1–7 withconsequent prolonged hospitalization and increased risk ofstroke and mortality (5% at 1 year and 16% at 10 years).8–11

The mechanisms responsible for postoperative atrial fibrilla-tion (POAF) still are debated.12,13 In particular, a number ofstudies suggest a link between oxidative stress and inflam-mation associated with cardiopulmonary bypass (CPB) andPOAF.14–21

It is well known that CPB causes an acute systemicinflammatory response22,23 characterized by the increase ofserum C-reactive protein (CRP) levels, peaking on the secondpostoperative day with a time course similar to that of POAF.17

In addition to systemic inflammation caused by cardiopul-monary bypass, atrial inflammation might contribute to theoccurrence of POAF. Chen at al found a higher number ofCD45-positive cells in the right atrial appendage with AF ascompared to that in patients in sinus rhythm undergoing valvesurgery.24 Accordingly, the authors found CRP in atrialcardiomyocytes more frequently in patients with AF than inthose with Wolff-Parkinson-White syndrome undergoingcatheter-based ablation procedures.25

Interestingly, other studies have highlighted a potentialrole for perioperative myocardial injury in the developmentof POAF.26–30 In order to better elucidate the potential roleof preoperative atrial inflammation and/or myocardial ische-mic injury in the development of POAF, the authorsevaluated right atrial appendage specimens taken at theearly phases of the surgical procedure for the presence of

CRP and serum hs-TnT levels in the perioperative period ina selected population of patients undergoing CABG. Theauthors also investigated the time course of systemicmarkers of inflammation, such as hs-CRP levels, whiteblood cell (WBC) count, and fibrinogen levels in this settingof cardiac surgery.

METHODS

Thirty-eight consecutive patients with coronary artery disease(CAD) (without previous history of AF admitted to the authors’department for elective CABG from January 2009 to February 2010)were enrolled in this prospective study. Exclusion criteria included liveror kidney failure, recent infections or inflammatory disease in the last 6months, autoimmune diseases, neoplasm, recent trauma, or surgery inthe last 6 months. Demographic data, cardiovascular risk factors suchas hypertension, diabetes, dyslipidemia, smoking, familial history ofCAD, history of chronic kidney disease, CAD clinical features, medicaltherapy, ECG findings, and echocardiographic parameters, such as leftventricular ejection fraction (LVEF) and left atrial size, were collected.In particular, chronic kidney disease was defined as estimatedglomerular filtration rate, determined with Modification of Diet for

013: pp ]]]–]]] 1

NARDUCCI ET AL2

Renal Disease formula, less than 60 mL/min/1.73 m2 in the last 3months as described elsewhere.31 The study was approved by the ethicscommittee of the authors’ university, and all patients gave writteninformed consent.

Following median sternotomy and administration of an IV bolus ofheparin (300 IU/kg) to obtain an activated coagulation times (ACT)4480s, normothermic cardiopulmonary bypass was instituted bycannulation of the right atrium and ascending aorta. Right atrialappendage biopsies were performed at the early phase of theintervention at the cannulation site.

Then, myocardial protection was accomplished by anterogradeintermittent blood cardioplegia isothermic to the systemic perfusiontemperature. Additional boluses of heparin (100 IU) were administeredif needed to maintain ACT in the range. Whenever possible, the leftinternal mammary artery was used to graft the left anterior descendingcoronary artery and saphenous vein grafts for the remaining coronaryarteries. At the end of CPB and achievement of stable hemodynamicconditions, anticoagulation was reversed by administration of prot-amine sulphate (1-1.5 mg per 100 IU of heparin administered in theprevious hour). Surgical parameters, such as CPB time and temper-ature, aortic clamping time, and body temperature during the inter-vention, were collected for each patient.

The formalin-fixed, paraffin-embedded specimens, obtained by rightatrial appendage biopsy during CPB, were sectioned at 5μm and stainedwith hematoxylin and eosin and with hematoxylin and Van Gieson toconfirm the presence of atrial cardiomyocytes. Myocardial infiltration ofinflammatory cells was examined histologically. Sections from eachspecimen displaying atrial cardiomyocytes were cut at 5mm, mounted onglass, and dried overnight at 371C. All sections then were deparaffinizedin xylene, rehydrated through a graded alcohol series, and washed inPBS, which also was used for all subsequent washes and for antibodydilution. Endogenous peroxidase activity was blocked by 5% hydrogenperoxide. For immunohistochemistry, tissue sections were heated twicein a microwave oven for 5 minutes each at 700 W in citrate buffer (pH ¼6). The murine monoclonal antibody (clone CRP-8, IgG1, Sigma, St.Louis, MO) directed against human CRP was used at 1:500 dilution andincubated overnight at 41C. Then, the sections were processed with thestandard streptavidin-biotin-immunoperoxidase method (DAKO Univer-sal Kit, DAKO Corp, Carpinteria, CA). Diaminobenzidine was used asthe final chromogen, and hematoxylin as the nuclear counterstain. Themonoclonal antibody CRP-8 displayed its reactivity against native anddenatured CRP without cross-reaction with human serum, amyloid Pcomponent, human haptoglobin, human alpha-1-acid glycoprotein andhuman IgG. All specimens were analyzed to assess cell number and theimmunoreactivity to CRP by a single investigator who was blinded to thepatients’ characteristics. Atrial specimens were scored as positive ornegative for the presence or absence of CRP.

The total RNA was extracted using Trizol (Invitrogen, GrandIsland, NY) as described previously, blinded to histologic results. Allatrial specimens were evaluated for CRP mRNA expression. cDNAwas synthesized from 1 lg total RNA in 20 lL reaction using oligo (dT)18 Primer and RevertAidTM M-MuLV Reverse Transcriptase (Fermen-tas, St. Leon-Rot, Germany). mRNA expressions of CRP, AT1 andglyceraldehyde-3-phosphate dehydrogenase (GAPDH) were examinedby RT-PCR. Sequence-specific PCR primers used were CRP, forward50-TCGTATGCCACCAAGAGACAAGACA-30, reverse 50-AAC-ACTTCGCCTTGCACTTCATACT-30 [11], giving 440 bp PCRproduct; AT1, forward 50-ACTAGGCATCAT ACGTGACTGTAG-30, reverse 50-TGTTGAAAGGTTTG AGTGGG-30, giving 196 bpPCR product; GAPDH, forward 50-AACATCATCCCTGCCTC-TACTGG-30, reverse 50-CTCCGACGCCTGCTTCACC-30, giving189 bp PCR product. PCR products were separated by electrophoresison 2% agarose gel. Expression of mRNA was quantified as relative tointernal control GAPDH.

Immediately before surgery and 24 hours after, blood samples weredrawn from the radial artery. White blood cell count, serum fibrinogen,and serum high sensitivity (hs)-CRP levels were assessed as inflamma-tory markers. Serum hs-TnT levels were assessed as myocardial necrosismarker. Serum hs-CRP was assayed using a latex-enhanced immunone-phelometric assay (BN II, Siemens Diagnostic Glasgow, DE). Cardiachs-TnT was quantitatively determined using a 1-step electroimmuno-assay based on electrochemilumiscence technology (Elecsys 2010,Roche), with a lower detection limit of 0.01 ng/mL.

All patients were monitored by ECG telemetry until discharge.Patients who developed POAF were subjected to electrical or pharma-cologic cardioversion.

Continuous variables are expressed as median accompanied byinterquartile range, as Shapiro-Wilk test showed that the data were notnormally distributed, whereas categoric variables are expressed asabsolute frequencies and percentages. Comparisons between groupswere performed with Mann-Whitney U test, Wilcoxon test for pairedsamples, or chi-square test, as appropriate. Univariate logistic regres-sion analysis was performed, including all clinical features, surgicalvariables, and laboratory measurements (both inflammatory and myo-cardial damage markers) possibly related to POAF or atrial CRPlocalization. Then, variables showing a significant association(p o 0.10) were included in multivariate logistic regression. All testswere 2-sided, and a value of p o 0.05 was settled for statisticalsignificance. Statistical analysis was performed using SPSS 19 (SPSS,Inc Chicago, IL).

RESULTS

Thirty-eight patients were enrolled in the study (mean age61 years, male gender 73%). Clinical presentation was an acutecoronary syndrome in 20 patients (54%), without patients withST-elevation myocardial infarction. Eighteen patients (46%)presented with stable angina. Thirty patients (79%) had severe3-vessel disease. All 38 (100%) patients were in sinus rhythmon admission. Postoperative monitoring by ECG telemetry inthe authors’ department was obtained for mean time intervalsof 9 � 3 days. Fourteen patients out of 38 (36%) developed AFduring the postoperative period (POAF group) POAF occurredearly (within 72 hours from CABG) in 12 (86%) and late (after72 hours from CABG) in only 2 (14%) patients. The remaining24 patients (64%) exhibited constant sinus rhythm in thepostoperative period (SR group).

Clinical, surgical features and laboratory findings of POAFand SR patients are shown in Table 1. No significant differencewas observed between the 2 groups with regard to age, gender,prevalence of cardiovascular risk factors, clinical presentationof CAD, medical therapy, or surgical parameters. Of note, allpatients with severe left ventricular dysfunction (LVEF o35%)developed POAF (p ¼ 0.018); furthermore, POAF patients hada higher rate of 3-vessels coronary disease (100% v 67%,respectively; p ¼ 0.015) and of chronic kidney disease (p ¼0.012) than the SR group.

Preoperative hs-CRP levels, WBC count, fibrinogen, andhs-TnT levels did not differ significantly between the 2 groups.After CABG, serum levels of hs-CRP and hs-TnT increasedsignificantly in both RS and POAF groups as compared tobaseline levels (hs-CRP: p ¼ 0.013 for POAF group andp o 0.001 for SR group, [Fig 1]; hs-TnT: p o 0.001 for bothgroups, [Fig 2]). Of note, serum postoperative hs-TnT levelswere significantly higher in the POAF group than in the SR

PREDICTORS OF POSTOPERATIVE ATRIAL FIBRILLATION 3

group (0.52 [0.32-0.86] ng/mL versus 0.30 [0.23-0.42] ng/mL,respectively, p ¼ 0.016) (Fig 2), while postoperative levels ofhs-CRP were similar in both groups (4.02 [3.57-4.93] mg/dLversus 4.54 [3.55-5.71] mg/dL, respectively, p ¼ 0.34) (Fig 1).

Fibrinogen and WBC count significantly increased afterCABG in all patients as compared to baseline levels (preop-erative fibrinogen [median 308 mg/dL; range 241-388 mg/dL])versus postoperative fibrinogen (median 405 mg/dL; range368-446 mg/dL), p o 0.001; preoperative WBC (median 7245� 109/L; range 6302-9375 x 109/L) versus postoperative WBC(median 11,525 � 109/L; range 9382-15087 � 109/L);p o 0.001). As reported in Table 1, both preoperative andpostoperative fibrinogen levels and WBC count did not differsignificantly between the POAF and SR groups.

No correlation was found between preprocedural hs-CRPlevels and preprocedural hs-TnT levels (p ¼ 0.804) and

Table 1. Clinical and Surgical Vari

POAF Group (n ¼ 14; 36

Age 71 (68-77)

Gender (male) 9 (64%)

Risk factors and comorbidities

Diabetes 7 (50%)

Hypertension 14 (100%)

Smoking 2 (14%)

Dyslipidemia 7 (50%)

Chronic kidney disease 6 (43%)

Clinical features

Acute Coronary Syndrome 5 (36%)

3-vessel coronary disease 14 (100%)

LVEF o35% 3 (21%)

Left atrium diameter 440 mm 8 (57%)

Therapy

β-blockers 13 (93%)

Ace-inhibitors or ARBs 7 (50%)

Ca-antagonists 3 (21%)

Warfarin 5 (36%)

Statins 8 (57%)

Aspirin 6 (43%)

Diuretics 8 (57%)

Surgical parameters

CPB time (min) 63

Aortic clamp time (min) 55

CPB temperature (1C) 35

Body temperature (1C) 37,0

CABG number ≥3 8 (57%)

Preprocedural laboratory findings

hs-CRP (mg/dL) 0.94 (0.30-1.98)

WBCs count (n � 109/L) 7220 (5797-8750)

Fibrinogen (mg/dL) 353 (268-474)

hs-TnT (ng/mL) 0.000 (0.00-0.01)

Postoperative laboratory findings

hs-CRP (mg/dL) 4.02 (3.6-4.9)

WBCs count (n � 109/L) 9450 (8965-15310)

Fibrinogen (mg/dL) 402 (345-416)

hs-TnT (ng/mL) 0.52 (0.32-0.86)

NOTE: All values are presented as absolute number (%), except for age

range).

Abbreviations: ARBs, angiotensin-receptor antagonists; CABG, corona

sensitivity C-Reactive Protein; hs-TnT, high-sensitivity troponin T; LVEF, le

SR, sinus rhythm; WBC, white blood cells.

between postprocedural hs-CRP levels and postprocedural hs-TnT levels (p ¼ 0.788).

At univariate analysis, postoperative hs-TnT levels emergedas predictors of POAF (OR 17.6 [1.2-263.7]; p ¼ 0.038)(Table 2). At multivariate analysis, including age and hs-TnT,postoperative serum levels of hs-TnT were the only predictivefactor of AF development after CABG (OR ¼ 15.0 [1.0-232.4];p ¼ 0.05) (Table 2).

No inflammatory infiltrates suggesting myocarditis werefound in atrial specimens at histopathologic analysis.C-reactive protein was found in atrial cardiomyocytes of 10POAF patients and of 21 SR patients (71% v 87% respectively,p ¼ ns). Of note, immunohistochemical analysis revealed aspecific localization of CRP within the cytoplasm of atrialcardiomyocytes (Fig 3). No trace of vascular, capillary, orintercellular CRP was found in atrial biopsies.

ables in POAF and SR Groups

%) SR Group (n ¼ 24; 64%) p Value

69 (59-75) 0.35

18 (75%) 0.28

15 (63%) 0.18

21 (88%) 0.24

10 (42%) 0.39

19 (79%) 0.27

2 (8%) 0.012

10 (42%) 0.38

16 (67%) 0.015

0 (0%) 0.018

7 (29%) 0.84

19 (79%) 0.71

15 (63%) 0.38

7 (29%) 0.28

13 (54%) 0.16

17 (71%) 0.23

10 (42%) 0.22

9 (38%) 0.81

57 0.98

49 0.84

36 0.88

36,8 0.97

13 (54%) 0.54

0.38 (0.21-1.30) 0.28

8090 (6387-10537) 0.61

328 (241-410) 0.53

0.00 (0.00-0.00) 0.83

4.54 (3.6-5.7) 0.34

12150 (10005-15755) 0.74

409 (363.5-488) 0.95

0.30 (0.23-0.42) 0.016

and laboratory findings, which are expressed as median (interquartile

ry artery bypass graft; CBP, cardiopulmonary bypass; hs-CRP, high-

ft ventricular ejection fraction; POAF, postoperative atrial fibrillation;

Fig 1. Preoperative and postoperative systemic high-sensitivity

C-reactive protein (hs-CRP) levels in patients who developed post-

operative atrial fibrillation (POAF) and in patients maintaining persis-

tent sinus rhythm after CABG (SR). High-sensitivity C-reactive protein

(hs-CRP) was assayed at baseline (preoperative), 24 h after CABG

(postoperative) in all patients. Asterisk* indicates p ¼ 0.013 baseline

hs-CRP v 24 h after CABG hs-CRP in postoperative atrial fibrillation

(POAF) group. Double asterisk** indicates p o 0.001 baseline hs-CRP

v 24 h after CABG hs-CRP in SR group. Data are presented as box-plot,

with median, 25-75 percentiles (box) and 5-95 percentiles (whiskers).

Fig 2. Preprocedural and postoperative peripheral high-sensitivity

troponin-T (hs-TnT) levels in patients with postoperative atrial fibrillation

(POAF group) and in patients maintaining sinus rhythm after CABG

surgery (SR group). High-sensitivity troponin-T (hs-TnT) was assayed at

baseline (preoperative), 24 h after CABG (postoperative) in all patients.

Asterisk* indicates p o 0.001 baseline hs-TnT v 24 h after CABG hs-TnT

in postoperative atrial fibrillation (POAF) group. Double asterisk** indi-

cates p ¼ 0.016 postoperative hs-TnT between POAF group and sinus

rhythm (SR) group. Triple asterisk*** indicates p o 0.001 baseline hs-TnT

v 24 h after CABG hs-TnT in SR group. Data are presented as box-plot,

with median, 25-75 percentiles (box) and 5-95 percentiles (whiskers).

NARDUCCI ET AL4

Polymerase chain reaction for CRP failed to identify thepresence of CRP mRNA in all right atrial specimens. Thecharacteristics of patients according to the presence or absenceof CRP in atrial specimens are depicted in Table 3. Atrial CRP-positive patients showed a higher prevalence of diabetes andsevere left ventricular dysfunction as compared to CRP-negative patients (67% versus 14%, p ¼ 0.01; 9% versus0%, p ¼ 0.012, respectively). All other clinical and laboratoryvariables, including preoperative and postoperative serum hs-CRP levels, were similar in the 2 groups. At univariate andmultivariate analysis, diabetes emerged as the only predictor ofCRP presence in atrial specimens (OR 12.6 [1.3-119.2]; p ¼0.027 at univariate and OR 12.1 [1.3-116.3]; p ¼ 0.03 atmultivariate analysis respectively) (Table 4).

DISCUSSION

In the authors’ selected population of patients undergoingCABG, postoperative serum hs-TnT levels were the onlyindependent predictors of POAF. Preoperative serum hs-CRPlevels were not associated with POAF. The novelty of theirstudy is represented by the examination of atrial preoperativeinflammatory status performing immunohistochemical analysisfor CRP in right atrial tissue of ischemic patients undergoingCABG. The authors found the presence of CRP in thecytoplasm of atrial cardiomyocytes without any evidence ofatrial myocarditis. Notably, 82% of the patients exhibitedintracytoplasmatic presence of CRP, without any differencebetween POAF and SR patients. Interestingly, diabetes was theonly predictor of CRP in atrial cardiomyocytes.

Several studies showed that troponin levels alwaysincrease after cardiac surgical interventions,32,33 and highlevels of troponin are associated with increased operative

death and adverse events after CABG.26,27,34 The elevationof this biomarker reflects a condition of myocardial necrosisbut does not indicate its mechanism.35 Both ischemic andnon-ischemic myocardial injuries (myocarditis, aortic dis-section, pulmonary embolism, congestive heart failure, renalfailure) may determine a rise of cardiac troponin.36 In theparticular context of CABG surgery, many mechanisms havebeen hypothesized to explain this phenomenon: Myocardialischemia or injury resulting from coronary occlusion, prolongedbypass time, myocardial cell injury related to surgical stressbecause of incomplete myocardial protection, reperfusioninjury, and intramyocardial vessel manipulation. It is difficultto affirm that myocardial ischemia is certainly the cause ofpostoperative troponin rise in POAF patients. However, manyobservations are in favor of an ischemic trigger of postoperativetroponin rise and of consequent POAF occurrence. First, thestudy population was a selected group of patients undergoingisolated CABG surgery without any cardiac disease except forcoronary artery disease. Moreover, postoperative clinical con-ditions and laboratory parameters were stable in all patients,thus excluding new pathologic settings as myocarditis, pulmo-nary embolism, acute congestive heart failure and/or renalfailure. Thirdly, surgical parameters (cardiopulmonary bypasstime and temperature, body temperature, aortic clamp time,number of coronary artery bypass grafts) were similar in allpatients. Finally, the prevalence of 3-vessel coronary arterydisease in POAF patients was significantly higher than in SRpatients (100% v 67%, Table 1). In contrast, neither pre- orpostoperative systemic inflammatory markers (serum hs-CRPlevels, WBC count, fibrinogen levels) nor the atrial presence ofintramyocardial CRP predicted POAF.

Table 2. Logistic Regression Analysis for Predictors of Postoperative Atrial Fibrillation (POAF) Occurrence

Variables Univariate OR (95% CI) p Value Multivariate OR (95% CI) p Value

Age 1.1 (0.9-1.2) 0.10 1.07 (0.98-1.17) 0.15

Gender (male) 0.6 (0.1-2.5) 0.48

Diabetes 0.6 (0.2-2.3) 0.45

3-vessel coronary disease 1.4 (0-5.2) 0.99

LVEF o35% 3.5 (0-7.4) 0.99

Left atrium diameter 440 mm 3.2 (0.8-12.8) 0.09

CPB time (min) 1.0 (0.98-1.1) 0.23

Aortic clamp time (min) 1.0 (0.98-1.1) 0.23

CBP temperature (1C) 1.1 (0.1-16.9) 0.92

Body temperature (1C) 0.1 (0.001-2.7) 0.14

CABG number ≥3 1.1 (0.3-4.3) 0.86

Preprocedural WBCs count (n � 10 ̂9/L) 1.0 (1.0-1.03) 0.28

Preprocedural neutrophils (%) 0.95 (0.90-1.01) 0.11

Preprocedural fibrinogen (mg/dL) 1.0 (0.99-1.0) 0.12

Preprocedural hs-Tn T (ng/mL) 0.7 (0.1-3.8) 0.65

Postoperative hs-Tn T (ng/mL) 17.6 (1.2-263.7) 0.038 15.0 (1.0-232.4) 0.05

Preprocedural hs-CRP (mg/dL) 1.0 (0.98-1.1) 0.25

Postoperative hs-CRP (mg/dL) 0.99 (0.97-1.01) 0.34

Atrial tissue CRP 0.4 (0.06-1.9) 0.23

Abbreviations: CABG, coronary artery bypass graft; CBP, cardiopulmonary bypass; hs-CRP, high-sensitivity C-reactive protein; hs-TnT, high-

sensitivity troponin T; LVEF, left ventricular ejection fraction; POAF, postoperative atrial fibrillation; WBC, white blood cells.

Table 3. Clinical Variables and Laboratory Findings According to the Presence or Absence of CRP in Atrial Tissue

Atrial Tissue CRP Positivity (n ¼ 31; 82%) Atrial Tissue CRP Absence (n ¼ 7; 18%) p Value

Age 70 (61-78) 67 (60-69) 0.84

Gender (male) 21 (68%) 6 (86%) 0.12

Risk factors and comorbidities

Diabetes 21 (67%) 1 (14%) 0.010

Hypertension 29 (93%) 6 (85%) 0.19

Smoking 10 (32%) 2 (28%) 0.26

Hypercolesterolemia 20 (65%) 6 (85%) 0.43

Chronic kidney disease 8 (26%) 0 (0%) 0.65

Clinical features

Acute coronary syndrome 12 (38%) 3 (43%) 0.55

3-vessel coronary disease 25 (80%) 5 (71%) 0.32

LVEF o35% 3 (9%) 0 (0%) 0.012

Left atrium diameter 440 mm 12 (39%) 3 (43%) 0.27

Therapy

β-blockers 25 (80%) 7 (100%) 0.27

Ace-inhibitors or ARBs 17 (55%) 5 (71%) 0.38

Ca-antagonists 10 (32%) 0 (0%) 0.59

Warfarin 15 (48%) 3 (43%) 0.85

Statins 19 (61%) 6 (85%) 0.42

Aspirin 12 (38%) 4 (57%) 0.66

Diuretics 16 (52%) 1 (14%) 0.37

Preprocedural laboratory findings

Hs-CRP (mg/dL) 0.40 (0.23-1.31) 0.99 (0.36-3.05) 0.21

WBCs count (n � 109/L) 7250 (6350-9290) 6530 (4360-14880) 0.87

Fibrinogen (mg/dL) 342 (243-387) 274 (236-453) 0.66

Hs-TnT (ng/mL) 0 (0,00-0,01) 0 (0,00-0,00) 0.81

Postoperative laboratory findings

Hs-CRP (mg/dL) 4.05 (3.5-4.9) 4.93 (3.1-17.6) 0.64

WBCs count (n � 109/L) 11550 (9460-15755) 8610 (8543-8700) 0.55

Fibrinogen (mg/dL) 409 (379-458) 355 (342-388) 0.78

Hs-TnT (ng/mL) 0.38 (0.24-0.62) 0.45 (0.14-0.45) 0.89

NOTE. All values are presented as absolute number (%), except for age and laboratory findings that are expressed as median (interquartile range).

Abbreviations: ARBs, angiotensin-receptor blockers; hs-CRP, high-sensitivity C-reactive protein; hs-TnT, high-sensitivity troponin T; LVEF, left ventricular

ejection fraction; WBCs, white blood cells.

PREDICTORS OF POSTOPERATIVE ATRIAL FIBRILLATION 5

Fig 3. Immunostaining of right atrial appendages specimens for C-reactive protein (CRP). Right atrial appendage cardiomyocytes were

immunohistochemically stained with C-reactive protein (CRP) antibodies. (A) Example of positive staining for CRP, noted as brown in the

cytoplasm with diffuse pattern of localization (original magnification 20�), in a diabetic patient. (B) Example of negative staining for C-reactive

protein (original magnification 40�) in a patient without diabetes.

NARDUCCI ET AL6

The relationship between preoperative systemic hs-CRP andPOAF has been addressed in several studies with conflictingresults. Lo et al showed that high baseline CRP levels were anindependent predictor of POAF in patients undergoing on-pump and off-pump surgery.37 Accordingly, a recent study38 in656 patients undergoing off-pump CABG found that higherpreoperative CRP levels were independently associated withPOAF. Another recent study found that both systemic andintracardiac CRP levels predicted the onset of AF afterCABG.39 In contrast, several other studies failed to find asignificant association between CRP levels and POAF.40–45 Assuggested by Anselmi et al,46 the association between CRP andPOAF is not direct but probably mediated by CPB-associatedoxidative stress and necrosis, which, on the one hand, triggersinflammation and, on the other, may trigger POAF in suscep-tible patients.

Table 4. Logistic Regression Analysis for Predictors of C

Variables Univariate OR (95% CI)

Age 1.0 (0.95-1.13)

Gender (male) 0.4 (0.03-3.3)

Diabetes 12.6 (1.3-119.2)

Hypertension 2.4 (0.19-31.1)

Smoking 1.2 (0.2-7.2)

Dyslipidemia 0.3 (0.03-2.9)

Chronic kidney disease 4.9 (0.003-9.9)

Acute Coronary Syndrome 0.8 (0.2-4.4)

3-vessel coronary disease 1.7 (0.3-10.8)

LVEF o35% 0.08 (0.006-1.10)

Left atrium diameter 440 mm 0.2 (0.03-1.2)

Preprocedural WBCs count (n � 10̂9/L) 1.0 (0.98-1.33)

Preprocedural neutrophils (%) 0.99 (0.9-1.1)

Preprocedural fibrinogen (mg/dL) 1.0 (0.99-1.01)

Preprocedural hs-Tn T (ng/mL) 3.6 (0.02-3.2)

Postoperative hs-Tn T (ng/mL) 8.9 (0.2-442.9)

Preprocedural peripheral hs-CRP (mg/dL) 0.99 (0.98-1.1)

Postoperative peripheral hs-CRP (mg/dL) 0.99 (0.97-1.0)

Abbreviations: ARBs, angiotensin-receptor antagonists; CBP, cardiopu

high-sensitivity troponin T; LVEF, left ventricular ejection fraction; WBC, w

Thirty-one patients (82% of the authors’ population) exhib-ited intracytoplasmatic presence of CRP without any differencebetween POAF and SR patients. Preoperative atrial tissue CRPdid not correlate with serum preoperative and postoperative hs-CRP levels and was not produced by atrial cardiomyocytes asdemonstrated by the lack of mRNA expression for CRP. Theintracytoplasmatic localization of CRP in right atrial specimensmight suggest a long-term accumulation of CRP, possiblyassociated with a long-lasting stimulus. Interestingly, diabeteswas the only predictor of CRP in atrial cardiomyocytes. Themechanisms of this association cannot be derived from theresults of the authors’ study.

Of note, myocardial biopsies were performed at the begin-ning of CPB before the surgical stress, thus reflecting theintramyocardial atrial inflammatory environment existingbefore the intervention. There are no biopsy data assessing

-reactive Protein (CRP) Localization in Atrial Tissue

p Value Multivariate OR (95% CI) p Value

0.33 1.03 (0.94-1.13) 0.45

0.36

0.027 12.1 (1.3-116.3) 0.03

0.50

0.85

0.30

0.99

0.84

0.59

0.06

0.07

0.90

0.67

0.59

0.53

0.28

0.25

0.14

lmonary bypass; hs-CRP, high-sensitivity C-reactive protein; hs-TnT,

hite blood cells.

PREDICTORS OF POSTOPERATIVE ATRIAL FIBRILLATION 7

the potential local effect of the surgical stress on atrialcardiomyocites, so what the authors’ study demonstrates ismerely that pre-existing atrial inflammation does not predict thedevelopment of POAF. It would be of interest to evaluate theinflammatory response in additional atrial tissue specimenscollected just after the surgical procedure.

In the authors’ study, they aimed to evaluate the possiblerole of myocardial ischemia and atrial inflammation in thepathogenesis of POAF. In fact, the possible mechanismsresponsible for the onset of POAF still are debated.12,13 Evenif several studies suggest a prominent role for systemicinflammation in the development of POAF,17,47–51 their dataare consistent with those of other studies that highlighted apossible role for myocardial injury/ischemia in the patho-genesis of POAF.26,40 Particularly, their data are in agree-ment with those of Ahlsson et al, who observed that serumCK-MB levels but not CRP levels predicted POAF in 524patients undergoing CABG.40 Moreover, a number of studieshave shown that oxidative stress occurring during ischemia/reperfusion is likely to be an important link between CPBand POAF.14–16 Accordingly, minimized perfusion circuitshave been found to be associated with a lower prevalence ofPOAF.52 Of note, in an experimental model, anotherintriguing biomolecular link between ischemia and AF wasfound in ischemia-induced alterations of connexins, proteins

expressed by atrial cells, known for being implicated inarrhythmias.53

However, it is important to underline that the differentpathogenetic hypotheses linked to POAF are not mutuallyexclusive. Particularly, Abramov et al54 demonstrated that therise of serum-TnI levels was steeper after cardiac surgery thanafter myocardial infarction, suggesting an increased postoperativemyocardial permeability, typical of an inflammatory process.

The most important limitation of the authors’ study is thesmall sample size, although it represents the largest series ofpatients undergoing immunohistochemical analysis of atrialspecimens for CRP.

A second limitation is represented by isolated right atrialappendage analysis without left atrial analysis. Finally, post-operative systemic hs-CRP levels have been assessed 24 hourspostoperatively while CRP peaks at 48 hours, as demonstratedby Bruins.17

In conclusion, available data on the mechanisms of POAF inpatients undergoing CABG are conflicting. The authors’findings support the notion that the severity of ischemia duringCPB contributes to POAF rather than systemic inflammation asassessed by measuring hs-CRP levels, a sensitive marker ofinflammation. Furthermore, preoperative atrial tissue CRP wasfound in the majority of patients with a significant correlationwith diabetes but not with POAF.

REFERENCES

1. Almassi GH, Schowalter T, Nicolosi AC, et al: Atrial fibrillationafter cardiac surgery: A major morbid event? Ann Surg 226:501-511; discussion 11-3, 19972. Hogue CW Jr, Hyder ML: Atrial fibrillation after cardiac operation:

Risks, mechanisms, and treatment. Ann Thorac Surg 69:300-306, 20003. Mariscalco G, Engstrom KG: Postoperative atrial fibrillation is

associated with late mortality after coronary surgery but not aftervalvular surgery. Ann Thorac Surg 88 :1871-1876, 20094. Aranki SF, Shaw DP, Adams DH, et al: Predictors of atrial

fibrillation after coronary artery surgery. Current trends and impact onhospital resources. Circulation 94:390-397, 19965. Ahlsson A, Fengsrud E, Bodin L, et al: Postoperative atrial

fibrillation in patients undergoing aortocoronary bypass surgery carriesan eightfold risk of future atrial fibrillation and a doubled cardiovas-cular mortality. Eur J Cardiothorac Surg 37:1353-1359, 20106. Villareal RP, Hariharan R, Liu BC, et al: Postoperative atrial

fibrillation and mortality after coronary artery bypass surgery. J AmColl Cardiol 43:742-748, 20047. Siebert J, Anisimowicz L, Lango R, et al: Atrial fibrillation after

coronary artery bypass grafting: Does the type of procedure influencethe early postoperative incidence? Eur J Cardiothorac Surg 19:455-459, 20018. El-Chami MF, Kilgo P, Thourani V, et al: New-onset atrial

fibrillation predicts long-term mortality after coronary artery bypassgraft. J Am Coll Cardiol 55:1370-1376, 20109. Mathew JP, Fontes ML, Tudor IC, et al: A multicenter risk

index for atrial fibrillation after cardiac surgery. JAMA 291:1720-1729, 200410. Auer J, Weber T, Berent R, et al: Postoperative atrial fibrillation

independently predicts prolongation of hospital stay after cardiacsurgery. J Cardiovasc Surg 46:583-588, 200511. Lahtinen J, Biancari F, Salmela E, et al: Postoperative atrial

fibrillation is a major cause of stroke after on-pump coronary arterybypass surgery. Ann Thorac Surg 77:1241-1244, 2004

12. Attaran S, Shaw M, Bond L, et al: Atrial fibrillation postcardiacsurgery: A common but a morbid complication. Interact CardiovascThorac Surg 12:772-777, 201113. Maesen B, Nijs J, Maessen J, et al: Postoperative atrial

fibrillation: A maze of mechanisms. Europace 14:159-174, 201214. Carnes CA, Chung MK, Nakayama T, et al: Ascorbate attenuates

atrial pacing-induced peroxynitrite formation and electrical remodelingand decreases the incidence of postoperative atrial fibrillation. Circ Res89:E32-E38, 200115. Shiroshita-Takeshita A, Schram G, Lavoie J, et al: Effect of

simvastatin and antioxidant vitamins on atrial fibrillation promotion byatrial-tachycardia remodeling in dogs. Circulation 110:2313-2319, 200416. Kim YM, Kattach H, Ratnatunga C, et al: Association of atrial

nicotinamide adenine dinucleotide phosphate oxidase activity with thedevelopment of atrial fibrillation after cardiac surgery. J Am CollCardiol 51:68-74, 200817. Bruins P, te Velthuis H, Yazdanbakhsh AP, et al: Activation of the

complement system during and after cardiopulmonary bypass surgery:Postsurgery activation involves C-reactive protein and is associated withpostoperative arrhythmia. Circulation 96:3542-3548, 199718. Hak L, Mysliwska J, Wieckiewicz J, et al: Interleukin-2 as a

predictor of early postoperative atrial fibrillation after cardiopulmonarybypass graft (CABG). J Interferon Cytokine Res 29 :327-332, 200919. Abdelhadi RH, Gurm HS, Van Wagoner DR, et al: Relation of

an exaggerated rise in white blood cells after coronary bypass orcardiac valve surgery to development of atrial fibrillation postoper-atively. Am J Cardiol 93:1176-1178, 200420. Lamm G, Auer J, Weber T, et al: Postoperative white blood cell

count predicts atrial fibrillation after cardiac surgery. J CardiothoracVasc Anesth 20:51-56, 200621. Gibson PH, Cuthbertson BH, Croal BL, et al: Usefulness of

neutrophil/lymphocyte ratio as predictor of new-onset atrial fibrillationafter coronary artery bypass grafting. Am J Cardiol 105:186-191,2010

NARDUCCI ET AL8

22. Cremer J, Martin M, Redl H, et al: Systemic inflammatoryresponse after cardiac operations. Ann Thorac Surg 61:1714-1720,199623. Taylor KM: SIRS—the systemic inflammatory response syn-

drome after cardiac operations. Ann Thorac Surg 61:1607-1608, 199624. Chen MC, Chang JP, Liu WH, et al: Increased inflammatory cell

infiltration in the atrial myocardium of patients with atrial fibrillation.Am J Cardiol 102:861-865, 200825. Narducci ML, Pelargonio G, Dello Russo A, et al: Role of tissue

C-reactive protein in atrial cardiomyocytes of patients undergoingcatheter ablation of atrial fibrillation: Pathogenetic implications. Euro-pace 13:1133-1140, 201126. Leal JC, Petrucci O, Godoy MF, et al: Perioperative serum

troponin I levels are associated with higher risk for atrial fibrillation inpatients undergoing coronary artery bypass graft surgery. InteractCardiovasc Thorac Surg 14:22-25, 201227. Leal JC, de Paula Neto A, Avanci LE, et al: Risk stratification

with troponin I in patients undergoing myocardial revascularizationsurgery. Arq Bras Cardiol 80:279-288, 200328. Sezai A, Hata M, Niino T, et al: Study of the factors related to

atrial fibrillation after coronary artery bypass grafting: A search for amarker to predict the occurrence of atrial fibrillation before surgicalintervention. J Thorac Cardiovasc Surg 137:895-900, 200929. Kaplan S, Ozisik K, Morgan JA, et al: Measurement of troponin

T and I to detect cardioprotective effect of aminophylline duringcoronary artery bypass grafting. Interact Cardiovasc Thorac Surg 2:310-315, 200330. Knayzer B, Abramov D, Natalia B, et al: Atrial fibrillation and

plasma troponin I elevation after cardiac surgery: Relation toinflammation-associated parameters. J Card Surg 22:117-123, 200731. Levey AS, Coresh J, Balk E, et al: National Kidney Foundation

practice guidelines for chronic kidney disease: Evaluation, classifica-tion, and stratification. Ann Intern Med 139:137-147, 200332. Nesher N, Alghamdi AA, Singh SK, et al: Troponin after cardiac

surgery: A predictor or a phenomenon? Ann Thorac Surg 85:1348-1354, 200833. Bonnefoy E, Filley S, Kirkorian G: Troponin I, troponin T, or

creatine kinase-MB to detect perioperative myocardial damage aftercoronary artery bypass surgery. Chest 114:482-486, 199834. Thielmann M, Massoudy P, Neuhäuser M, et al: Prognostic

value of preoperative cardiac troponin I in patients undergoingemergency coronary artery bypass surgery with non-ST-elevation orST-elevation acute coronary syndromes. Circulation 114(1 Suppl):I448-I453, 200635. Jaffe AS, Babuin L, Apple FS: Biomarkers in acute cardiac

disease: The present and the future. J Am Coll Cardiol 48:1-11, 200636. Thygesen K, Alpert JS, Jaffe AS, et al: Third universal definition

of myocardial infarction. Eur Heart J 33:2551-2567, 201237. Lo B, Fijnheer R, Nierich AP, et al: C-reactive protein is a risk

indicator for atrial fibrillation after myocardial revascularization. AnnThorac Surg 79:1530-1535, 2005

38. Kinoshita T, Asai T, Suzuki T, et al: Off-pump bilateral versussingle skeletonized internal thoracic artery grafting in high-risk patients.Circulation 124(11 Suppl):S130-S134, 201139. Kaireviciute D, Blann AD, Balakrishnan B, et al: Character-

isation and validity of inflammatory biomarkers in the prediction ofpostoperative atrial fibrillation in coronary artery disease patients.Thromb Haemost 104:122-127, 201040. Ahlsson AJ, Bodin L, Lundblad OH, et al: Postoperative atrial

fibrillation is not correlated to C-reactive protein. Ann Thorac Surg 83:1332-1337, 200741. Ishida K, Kimura F, Imamaki M, et al: Relation of inflammatory

cytokines to atrial fibrillation after off-pump coronary artery bypassgrafting. Eur J Cardiothorac Surg 294:501-505, 200642. Fontes ML, Mathew JP, Rinder HM, et al: Atrial fibrillation after

cardiac surgery/cardiopulmonary bypass is associated with monocyteactivation. Anesth Analg 101:17-23, 200543. Fontes ML, Amar D, Kulak A, et al: Increased preoperative white

blood cell count predicts postoperative atrial fibrillation after coronaryartery bypass surgery. J Cardiothorac Vasc Anesth 23:484-487, 200944. Hogue CW Jr, Palin CA, Kailasam R, et al: C-reactive protein

levels and atrial fibrillation after cardiac surgery in women. Ann ThoracSurg 82:97-102, 200645. Arribas-Leal JM, Pascual-Figal DA, Tornel-Osorio PL, et al:

[Epidemiology and new predictors of atrial fibrillation after coronarysurgery]. Rev Esp Cardiol 60:841-847, 200746. Anselmi A, Possati G, Gaudino M: Postoperative inflammatory

reaction and atrial fibrillation: Simple correlation or causation? AnnThorac Surg 88:326-333, 200947. Ommen SR, Odell JA, Stanton MS: Atrial arrhythmias after

cardiothoracic surgery. N Engl J Med 336:1429-1434, 199748. Chung MK, Martin DO, Sprecher D, et al: C-reactive protein

elevation in patients with atrial arrhythmias: Inflammatory mechanismsand persistence of atrial fibrillation. Circulation 104:2886-2891, 200149. Dernellis J, Panaretou M: Relationship between C-reactive

protein concentrations during glucocorticoid therapy and recurrentatrial fibrillation. Eur Heart J 25:1100-1107, 200450. Sata N, Hamada N, Horinouchi T, et al: C-reactive protein and

atrial fibrillation. Is inflammation a consequence or a cause of atrialfibrillation? Jpn Heart J 45:441-445, 200451. Aviles RJ, Martin DO, Apperson-Hansen C: Inflammation as a

risk factor for atrial fibrillation. Circulation 108:3006-3010, 200352. El-Essawi A, Hajek T, Skorpil J, et al: Are minimized perfusion

circuits the better heart lung machines? Final results of a prospectiverandomized multicentre study. Perfusion 26:470-478, 201153. Severino A, Narducci ML, Pedicino D, et al: Reversible atrial

gap junction remodeling during hypoxia/reoxygenation and ischemia:A possible arrhythmogenic substrate for atrial fibrillation. Gen PhysiolBiophys 31:439-448, 201254. Abramov D, Abu-Tailakh, Freiger M, et al: Plasma troponin

levels after cardiac surgery vs after myocardial infarction. AsianCardiovasc Thorac Ann 14:530-535, 2006