Psychological factors and cardiac repolarization instability during anger in implantable cardioverter defibrillator patients

Abstract Background Evidence indicates that emotions such as anger are associated with increased incidence of sudden cardiac death, but the biological mechanisms remain unclear. We tested the hypothesis that, in patients with sudden death vulnerability, anger would be associated with arrhythmic vulnerability, indexed by cardiac repolarization instability. Methods Patients with coronary artery disease (CAD) and an implantable cardioverter defibrillator (ICD; n = 41) and healthy controls (n = 26) gave an anger‐inducing speech (anger recall), rated their current (state) anger, and completed measures of trait (chronic) levels of Anger and Hostility. Repolarization instability was measured using QT Variability Index (QTVI) at resting baseline and during anger recall using continuous ECG. Results ICD patients had significantly higher QTVI at baseline and during anger recall compared with controls, indicating greater arrhythmic vulnerability overall. QTVI increased from baseline to anger recall to a similar extent in both groups. In ICD patients but not controls, during anger recall, self‐rated anger was related to QTVI (r = .44, p = .007). Trait (chronic) Anger Expression (r = .26, p = .04), Anger Control (r = −.26, p = .04), and Hostility (r = .25, p = .05) were each associated with the change in QTVI from baseline to anger recall (ΔQTVI). Moderation analyses evaluated whether psychological trait associations with ΔQTVI were specific to the ICD group. Results indicated that Hostility scores predicted ΔQTVI from baseline to anger recall in ICD patients (β = 0.07, p = .01), but not in controls. Conclusions Anger increases repolarization lability, but in patients with CAD and arrhythmic vulnerability, chronic and acute anger interact to trigger cardiac repolarization lability associated with susceptibility to malignant arrhythmias.

Increased beat-to-beat variability in the QT interval (QT Variability Index or QTVI), a measure of cardiac repolarization lability, is also predictive of increased risk of cardiac arrhythmias and sudden death (Atiga et al., 1998;Berger et al., 1997;Dobson etal.,2011;Haigneyetal.,2004;Piccirilloetal.,2007)andisanindependent risk marker for, and predictor of, ventricular tachycardia and ventricular fibrillation (Haigney et al., 2004). Little is known about whether chronic anger traits and acute anger responses have effects onthemagnitudeofrepolarizationinstabilityasindicatedinQTvariability, and whether these effects would be limited to patients with increasedriskofarrhythmiaandsuddendeath.Therefore,thepurpose of this study was to evaluate the link between chronic anger and acute anger reactions and repolarization lability as measured via QT variability index (QTVI). It was hypothesized that both chronic anger traits and anger induced by mental stress would be related to increased QT variability in patients with CAD and arrhythmic vulnerability.  (Rozanskietal.,1984)andnoevidenceofelectrocardiogram(ECG) abnormalities were also tested.
To optimize the ECG assessment of QTVI, calcium antagonists andACEinhibitorswerewithheldfor24hrandlong-actingnitrates withheld for 6 hr. Beta-blockers were withheld for >36 hr in 7 patients, 7 patients were not on beta-blockers as part of their medical management, and 27 did not discontinue beta-blockers. Patients tested on beta-blockers (n =27)hadsimilarQTVIatrest(p =.89)and during anger recall testing (p =.10)comparedwithpatientstested off beta-blockers (n =14).ThisstudyconformedtotheUSFederal Policy for the Protection of Human Subjects and was approved by theInstitutionalReviewBoardsattheparticipatinginstitutions.All participants provided written informed consent.
Aspartofthelargerstudy,participantscompleteda2-daylaboratory mental stress procedure that included testing with the anger recalltask.Aftera15-minrestingbaseline,participantswereasked to recall a recent incident in which they felt irritated, frustrated, angry, or upset and instructed to deliver a 4-min speech about this anger-provokingsituationtotheresearchteam.Angerrecallisapotent task capable of producing significant hemodynamic responses and impairments in ventricular function and myocardial ischemia in CAD patients (Jain et al., 2001;Jiang et al., 2013;Steptoe & Kivimäki,2013).Participantsratedtheircurrent(state)levelofanger at baseline and during anger recall task using a Likert scale from 0 = not at all to 7 = very much.
During rest and anger recall, blood pressure was obtained every 60 s, and continuous digitized ECGs were obtained during restandangerrecall.RecordinganddigitalizationofECGoccurred at 1,000 Hz with 16-bit resolution using the CH2000 (Cambridge Heart) with high-resolution silver-silver chloride electrodes for noise reduction. Data were exported for blinded off-line QT variability analyses by a trained reader using the method developed by Berger and colleagues (Berger et al., 1997). Each QT interval and heart rate value were measured over equal time periods obtained fromtheleadwiththebestvisualizationoftheQTinterval(gener-allyleadIII).AQTintervaltemplateisselectedfromarepresentative beat and compared with each subsequent beat and stretched or compressed in time to achieve a high degree of fit. The stretch factorforeachbeatisusedtoderiveaQTandheartratetimeseries for the epoch. Beats preceding and following ectopy, and all artifacts wereexcludedandremovedpriortoanalyses.QTvariabilityindex (QTVI) was calculated as the log ratio of normalized QT variability (QTVN;QTintervalvariancedividedbyQTintervalmean-squared) to normalized HR variability (heart rate variance divided by heart ratevariancemean-squared).QTVIscoresarequantifiedasnegative numbers,withQTVIindicatinghigherarrhythmicvulnerabilityhav-inganegativevaluecloserto0(i.e.,smallernegativenumber),and QTVIindicatinglowervulnerabilityhavinganegativevaluefurther from0(i.e.,largernegativenumber).
Respondents rate each statement as true or false as applied to them, with higher scores indicating greater hostility. The Hostility scale hasgoodpsychometricproperties (Cook&Medley,1954;Williams etal.,1980). Data are presented as mean ± standard deviation (SD)orfrequencies(N)andpercentages(%),andt tests and chi-squared tests used to evaluate demographics. Differences between ICD patientsandhealthycontrolsintraitvariables(AngerSuppression, AngerExpression,AngerControl,andHostility),andbaselineand anger recall values of QTVI, hemodynamics, and anger ratings were evaluated with analysis of covariance (ANCOVA), adjusted for age, sex, and race (white vs. nonwhite). Repeated measures ANCOVA was used to evaluate group differences in QTVI, hemodynamics, and self-rated anger rating levels during baseline and anger recall, with time as a within-subjects factor (baseline vs.angerrecall),group(ICDpatientsvs.controls)asabetweensubjects factor, and covariates of age, sex, and race. Paired samples t tests were used to examine changes from baseline to anger recallwithineachgroupofQTVI,hemodynamics,andstateanger ratings. Effect size was calculated using partial eta squared (η p 2 ) andCohen'sd.

call,andangerrecalldidnotprovokeasignificantresponseinSDNN
in either group (all p >.09).Therewerenogroupdifferencesinblood pressure or heart rate at rest or during anger recall, and within both groups, blood pressure and heart rate significantly increased from baseline to anger recall (all p <.001).
Resting baseline self-ratings of current (state) anger were low and comparable among groups and significantly increased within Although levels of QTVI were consistently elevated in ICD patients compared with controls, there was no time by ICD group interaction (Figure 1), and the magnitude of QTVI change(Δ) from baseline to anger recall was comparable in both groups (p =.85),with a similar pattern observed for the ΔinQTnumerator,QTdenominator,andSDNN(allp >.08).Themagnitudeofsystolicanddiastolic blood pressure and heart rate Δ from baseline to anger recall was alsocomparableamongICDpatientsandcontrols(allp >.66).ICD patients and controls reported similar increases in state anger ratings in response to anger recall (p =.38).
In the overall sample (n =67),traitAngerandHostilitymeasures and self-rated anger were not associated with QTVI at baseline or during anger recall (all r < .20 and p >.11).Similarly,Δ in self-rated anger in the overall sample from baseline to anger recall were not associated with the magnitude of ΔQTVIfrombaselinetoangerrecall(p =.38).

Thissuggeststhat,comparedwithhealthycontrols,angerincreases
repolarization instability out of proportion to the effects on autonomic tone. Increased QTVI predicts sudden death (Dobson et al., 2011;Piccirillo et al., 2007) and ventricular tachycardia or fibrillation (Haigney et al., 2004). An increase in QTVI indicates that the QT interval is varying out of proportion to the heart rate, implying loss of "repolarization reserve," the capacity of the myocardium to regulateexcitabilityinresponsetochangingautonomictone.Conditions manifesting reduced repolarization reserve are associated with an increased risk of re-entrant arrhythmias such as polymorphic ventricular tachycardia or fibrillation.

| Study limitations
The present results add to a body of knowledge regarding dynamic changes in cardiac repolarization that occur during mental stress and suggest that, in vulnerable patients, abnormalities in ventricular repolarization and cardiac repolarization lability may be one mechanism linking acute anger to arrhythmias and/or sudden cardiacdeath (Atigaetal.,1998;Burgetal.,2004;Dobsonetal.,2011;Lampertetal.,2002).Furthermore,inpatientswithpreviouslyexist-ingCADandarrhythmicvulnerability,chronicangertraitspotentiate the effects of acute anger and may be particularly important triggers of cardiac repolarization lability and susceptibility to malignant arrhythmias.

E TH I C A L A PPROVA L
This study conforms to the US Federal policy for protection of human subjects and was reviewed and approved by Institutional

ACK N OWLED G EM ENTS
ThisworkwassupportedbyNationalHeartLungandBloodInstitute grant R01HL47337. The opinions and assertions expressed herein are those of the authors and do not necessarily express the views of USUHS or the US Department of Defense.