ReiKHMK

FMEHMIR

Bc(rp

Ota

MSwEQ

Rwa0Dac

FCtPgwRt

1

elationship of common candidate gene variants tolectrocardiographic T-wave peak to T-wave endnterval and T-wave morphology parametersimmo Porthan, MD,a Annukka Marjamaa, MD, PhD,b Matti Viitasalo, MD, PhD,a

eikki Väänänen, Lic Sc (Tech),c Antti Jula, MD, PhD,d Lauri Toivonen, MD, PhD, FHRS,a

arkku S. Nieminen, MD, PhD,a Christopher Newton-Cheh, MD, MPH,g,h Veikko Salomaa, MD, PhD,e

immo Kontula, MD, PhD,f Lasse Oikarinen, MD, PhDa

rom the aDepartment of Cardiology, Helsinki University Central Hospital, Helsinki, Finland, bResearch Program inolecular Medicine, Biomedicum Helsinki, University of Helsinki, Helsinki, Finland, cDepartment of Biomedicalngineering and Computational Science, Helsinki University of Technology, Espoo, Finland, dNational Institute forealth and Welfare, Turku, Finland, eNational Institute for Health and Welfare, Helsinki, Finland, fDepartment ofedicine, Helsinki University Central Hospital, Helsinki, Finland, gProgram in Medical and Population Genetics, Broad

nstitute of Harvard and MIT, Cambridge, Massachusetts, and hCardiovascular Research Center & Center for Human Genetic

esearch, Massachusetts General Hospital, Boston, Massachusetts.

.tt

Csbol

Kpt

AsgomT

(

ACKGROUND Single-nucleotide polymorphisms (SNPs) in genes en-oding cardiac ion channels and nitric oxide synthase-1 adaptor proteinNOS1AP) are associated with electrocardiographic (ECG) QT-interval du-ation, but the association of these SNPs with new, prognostically im-ortant ECG measures of ventricular repolarization is unknown.

BJECTIVE The purpose of this study was to examine the rela-ionship of SNPs to ECG T-wave peak to T-wave end (TPE) intervalnd T-wave morphology parameters.

ETHODS We studied 5,890 adults attending the Health 2000tudy, a Finnish epidemiologic survey. TPE interval and four T-ave morphology parameters were measured from digital 12-leadCGs and related to the seven SNPs showing a phenotypic effect onT-interval duration in the Health 2000 Study population.

ESULTS In multivariable analyses, the KCNH2 K897T minor alleleas associated with a 1.2-ms TPE-interval shortening (P � .00005)nd the KCNH2 intronic rs3807375 minor allele was associated with a.8-ms TPE-interval prolongation (P � .001), whereas the KCNE185N variant had no TPE-interval effect (P � .20). NOS1AP minorlleles (rs2880058, rs4657139, rs10918594, rs10494366) were asso-

Dr. Porthan was supported by grants from the Aarne and Aili Turunenoundation, the Aarne Koskelo Foundation, the Finnish Foundation forardiovascular Research, the Finnish Medical Foundation, the Maud Kuis-

ila Memorial Foundation, the Orion-Farmos Research Foundation, and theaavo and Eila Salonen Foundation. Dr. Marjamaa was supported by arant from the Emil Aaltonen Foundation. Drs. Viitasalo and Oikarinenere supported by grants from the Finnish Foundation for Cardiovascularesearch. Dr. Newton-Cheh was supported by NIH Grant K23 HL080025,

he Doris Duke Charitable Foundation, and the Burroughs Wellcome Fund.

DafwFPHk1

547-5271/$ -see front matter © 2010 Heart Rhythm Society. All rights reserved

032 to .002), which resulted from their stronger effects on QTpeak

han QTend interval. None of the SNPs showed a consistent associa-ion with T-wave morphology parameters.

ONCLUSION KCNH2 K897T and rs3807375 as well as the fourtudied NOS1AP variants have modest effects on ECG TPE intervalut are not related to T-wave morphology measures. The previ-usly observed prognostic value of T-wave morphology parametersikely is not based on these SNPs.

EYWORDS Electrocardiography; Epidemiology; Genetics; Geneticolymorphism; Ion channels; Nitric oxide synthase; Repolariza-ion; T wave

BBREVIATIONS ECG � electrocardiographic; LQTS � long QTyndrome; NOS1AP � nitric oxide synthase-1 adaptor proteinene; PCA � principal component analysis; SNP � single-nucle-tide polymorphism; TCRT � total cosine R-to-T; TMD � T-waveorphology dispersion; TPE � T-wave peak to T-wave end;WR � T-wave residuum

Heart Rhythm 2010;7:898–903) © 2010 Heart Rhythm Society. All

iated with a shorter TPE interval (effects from 0.5 to 0.8 ms, P from rights reserved.

r. Salomaa was supported by Grant 129494 from the Academy of Finlandnd by the Sigrid Juselius Foundation. Dr. Kontula was supported by grantsrom the Sigrid Juselius Foundation and the Academy of Finland. Thisork was performed at Helsinki University Central Hospital, Helsinki,inland. Address reprint requests and correspondence: Dr. Kimmoorthan, Department of Cardiology, Helsinki University Central Hospital,aartmaninkatu 4, P.O. Box 340, 00029 HUS, Finland. E-mail address:immo.porthan@fimnet.fi. (Received December 5, 2009; accepted March, 2010.)

. doi:10.1016/j.hrthm.2010.03.002

IIsdsnmvanabaldtisi

dpdrimpEppuraTprS

MSTa2ls2h(ccifdtcro

CA

SjFMd(b6seccwds

EAqitfScwopcmb

TTvwemtimiiQwwp

TTtPePA

899Porthan et al Single-Nucleotide Polymorphisms and T Wave

ntroductionn monogenic repolarization-based arrhythmia disorders,uch as long QT syndrome (LQTS), mutations in myocar-ial ion channel genes generate the arrhythmogenic sub-trate.1 Common ion channel gene variants, such as single-ucleotide polymorphisms (SNPs), are associated withodest alterations in electrocardiographic (ECG) QT-inter-

al duration and may further modify repolarization-relatedrrhythmia vulnerability.1–6 In addition, genetic variation initric oxide synthase-1 adaptor protein gene (NOS1AP) isssociated with QT interval6,7 and, more importantly, haseen associated with increased risk for clinical symptomsnd more severe disease in a LQTS type 1 founder popu-ation8 as well as with increased risk for sudden cardiaceath in a sample from the general population.9 However,he risk for sudden cardiac death was independent of QT-nterval duration in multivariable analyses,9 raising the pos-ibility that ECG repolarization measures other than QTnterval might better reveal the increased risk.

The T wave is generated by myocardial voltage gradientsuring the repolarization phase of cardiomyocyte actionotentials.10,11 QT interval is a measure of repolarizationuration but may not reveal other changes during theepolarization process. T-wave peak to T-wave end (TPE)nterval measures terminal repolarization and has experi-entally been linked to arrhythmogenic repolarization dis-

ersion in the myocardium.10,12,13 In addition, several newCG repolarization measures characterizing T-wave mor-hology have been introduced.14,15 In fact, in the generalopulation T-wave morphology parameters seem to be moreseful than other ECG repolarization measures in mortalityisk assessment.16 In the present study, we examined thessociation of several SNPs with ECG TPE interval and-wave morphology parameters to characterize more com-rehensively the effects of the defined SNPs on ventricularepolarization in the large population-based Health 2000tudy.

ethodtudy populationhe study population was derived from Health 2000 Study,n epidemiologic survey conducted in Finland from 2000 to001. The study population, drawn from the Finnish Popu-ation Information System, was a two-stage stratified clusterample of 8,028 Finnish adults aged �30 years. The Health000 Study included a home interview, a comprehensiveealth examination with questionnaires, measurementse.g., height, weight, blood pressure, ECG), and a physi-ian’s clinical examination. Definitions of hypertension,oronary heart disease, myocardial infarction, and diabetesn the current study have been published elsewhere.16 Heartailure classification in the survey was based on hospitaliagnoses, health interview, physician’s clinical examina-ion, and medication use for heart failure. Probable andertain diagnoses of heart failure were included in the cur-ent study. The Health 2000 Study protocol is available

nline17 and was approved by the Epidemiology Ethics E

ommittee of the Helsinki and Uusimaa Hospital Region.ll participants gave signed informed consent.DNA samples were collected from 6,334 Health 2000

tudy subjects, and ECGs were available from 6,292 sub-ects. In the present study, subjects carrying one of fourinnish LQTS founder mutations (n � 27), ECGs withinnesota coding showing Wolff-Parkinson-White syn-

rome (n � 1), paced rhythm (n � 4), atrial fibrillationn � 93), atrial flutter (n � 1), complete left bundle branchlock (n � 56), complete right bundle branch block (n �1), or low-quality ECG (n � 8) were excluded from thetudy. Subjects who used medication having a possibleffect on QT-interval duration (n � 151) also were ex-luded. These medications were the drugs listed in the firstategory (Drugs with Risk of Torsades de Pointes) on theebsite www.qtdrugs.org (accessed November 2007) andigoxin, which shortens QT interval. Thus, the presenttudy included 5,890 individuals.

CG measurementsdigital standard 12-lead ECG was recorded using a Mar-

uette MAC 5000 electrocardiograph (GE Marquette Med-cal Systems, Milwaukee, WI, USA). All leads were simul-aneous recordings, and a median QRS-T complex was usedor analyses.18 QT Guard software (GE Marquette Medicalystems) was used to measure heart rate and principalomponent analysis (PCA) ratio. Custom-made softwareas used for all other measurements. The reproducibility ofur repolarization measurements was good and has beenreviously published.18 Left ventricular hypertrophy wasonsidered to be present if Sokolow-Lyon voltage �3.5V, Cornell voltage-duration product �244 mV�ms, or

oth criteria were observed.

PE intervalsPE-interval measurements were obtained based on a pre-iously described and validated algorithm.19,20 The soft-are calculated TPE interval from T-wave peak to T-wave

nd in each lead. A single observer (KP) reviewed measure-ents onscreen in blinded fashion. Leads with low signal-

o-noise ratio or flat T waves were excluded from TPE-nterval analyses, and the mean number of TPE-intervaleasurements per ECG was 10.3 � 1.7. The maximum TPE

nterval from precordial leads was used for analyses. TPEntervals were not corrected for heart rate.21 SNP effects onTpeak and QTend intervals were analyzed from the leadith the maximum precordial TPE interval. QT intervalsere heart rate–adjusted with the nomogram method asreviously described.16

-wave morphology parametershe automatic calculation of T-wave morphology parame-

ers has been published in detail elsewhere.16 In brief, theCA ratio was calculated as the ratio of the second to firstigenvalues of the spatial T-wave vector, with an increase inCA ratio indicating more complex T-wave morphology.14

fter singular value decomposition of eight independent

CG leads (I, II, V1–V6), T-wave morphology dispersion

(siTtlanTcrocTo0

GWLprSbirSCSC

SAIndtiipmldvhvriwwwlpwtmh

vgammmTrrBwfF

RGSll

RTlropDNtT0m

Tr

MACBHSDHCPHDETPTTT

a

Ti

900 Heart Rhythm, Vol 7, No 7, July 2010

TMD) measures the average angle between T-wave recon-truction vector pairs, with higher values indicating increas-ng differences in T-wave morphology among the leads.16

otal cosine R-to-T (TCRT) measures the deviation be-ween potential directions during depolarization and repo-arization phases, with lower values indicating increasingngle between R- and T-wave vectors.16 The vector mag-itude of the fourth to eighth eigenvalues gives the absolute-wave residuum (TWR). Higher TWR values indicate in-reasing nondipolar ECG signal content and are expected toepresent higher degrees of ventricular repolarization heter-geneity,15 whereas PCA ratio, TCRT, and TMD reflecthanges in dipolar ECG signal content. In the present study,-wave morphology parameters showed nonsignificant ornly very weak correlations to heart rate (absolute r from.005 to 0.08).

enetic analysese recently reported the relationship between several

QTS gene and NOS1AP SNPs to QTend interval.6 Theresent study included KCNH2 rs1805123 (K897T) ands3807375, KCNE1 rs1805128 (D85N), and four NOS1APNPs (rs2880058, rs4657139, rs10918594, rs10494366),ased on their shortening or prolonging effects on QTend

nterval in the Health 2000 Study population.6 Fors1805123, genotyping was performed using a TaqManNP Genotyping Assay (Applied Biosystems, Foster City,A, USA) and for all other SNPs was performed usingequenom MALDI-TOF mass spectrometry (MassArrayompact Analyzer, Sequenom, San Diego, CA, USA).

tatistical analysisnalyses were performed with SPSS software 15.0 (SPSS,

nc., Chicago, IL, USA). The prevalence estimates of ge-otypes were derived from the weighted study population asescribed previously.17 We analyzed the association be-ween ECG repolarization parameters and clinical variables,ncluding gender, age, current smoking status, body massndex, heart rate, systolic blood pressure, diastolic bloodressure, hypertension, coronary heart disease, previousyocardial infarction, heart failure, diabetes mellitus, ECG

eft ventricular hypertrophy status, and Cornell voltage–uration product (as a continuous variable). Of these clinicalariables, gender, age, systolic blood pressure, coronaryeart disease, previous myocardial infarction, and Cornelloltage–duration product were selected as covariates foregression analyses (P �.001 for association with repolar-zation parameters for most of the variables). Using a step-ise linear regression model, each repolarization parameteras adjusted for the clinical covariates. The output residualsere saved and used as the dependent variable in further

inear regression analyses, with the genotypes as the inde-endent variable. In additive genetic models, each genotypeas transformed to a continuous variable that corresponded

o the number of minor alleles (0, 1, 2). In genotypic geneticodels, two degrees of freedom test was used, and the

eterozygote and minor homozygote genotypes were con- *

erted to two dichotomous variables and the major homozy-ote genotype was the reference. The results are shown fordditive genetic models. Appropriateness of the additiveodel was confirmed by evaluating the genotype-specificeans for residuals of repolarization parameters. Non-nor-al variable distributions of PCA ratio, TMD, TCRT, andWR were normalized in regression analyses. The PCA

atio, TMD, and TWR were normalized with natural loga-ithmic transformation. The TCRT was normalized with thelom method22 because its distribution was not normalizedith the logarithmic transformation. All SNPs were tested

or Hardy-Weinberg equilibrium using the Chi-square test.or all tests, two-tailed P �.05 was considered significant.

esultsenotyping call rates ranged from 97.6% to 99.9%, and allNPs were in Hardy-Weinberg equilibrium (P �.01). Table 1

ists the clinical characteristics and mean values of ECG repo-arization measures in the study population.

elationship of SNPs to TPE intervalable 2 lists TPE intervals by genotype in the study popu-

ation and SNP effects on TPE interval from multivariableegression analyses. A 1.2-ms TPE-interval shortening wasbserved per KCNH2 K897T minor allele and a 0.8-msrolongation per KCNH2 rs3807375 minor allele. KCNE185N did not have an effect on TPE interval. For the fourOS1AP SNPs with strong linkage disequilibrium (r2 be-

ween 0.77 and 0.97), minor alleles were associated withPE-interval shortening, with the effect varying from 0.5 to.8 ms. The results were unchanged when TPE interval waseasured from a single lead (V3) (data not shown). The

able 1 Clinical characteristics and electrocardiographicepolarization measures (n � 5,890)

en (%) 45.2ge (years) 52.1 � 14.3urrent smoking (%) 21.9ody mass index (kg/m2) 26.9 � 4.7eart rate (bpm) 63.3 � 10.7ystolic blood pressure (mmHg) 134.2 � 21.0iastolic blood pressure (mmHg) 81.9 � 11.0ypertension (%) 46.8oronary heart disease (%) 6.6revious myocardial infarction (%) 2.3eart failure (%) 1.1iabetes mellitus (%) 5.6CG left ventricular hypertrophy (%) 17.8PE interval* (ms) 83.4 � 12.5CA ratio (%) 15.5 � 8.4MD (°) 14.6 � 13.4CRT (unitless) 0.36 � 0.52WR (technical units) 16,810 � 22,749

Values are given as mean � SD for continuous variables and percent-ges for categorical variables.

PCA � principal component analysis; TCRT � total cosine R-to-T;MD � T-wave morphology dispersion; TPE � T-wave peak to T-wave endnterval; TWR � T-wave residuum.

Maximum precordial TPE interval.

rbtmh

lTattirap

RpT(raia(s(cnwbiwht0mnr

DMOiiEainodsS

T

G

K

KN

*†‡p oltage–

F1gVBsg

901Porthan et al Single-Nucleotide Polymorphisms and T Wave

esults were also tested with the genotypic genetic models,ut the results did not change. Analyses after excluding aotal of 1,355 subjects with coronary heart disease, previousyocardial infarction, heart failure, or ECG left ventricular

ypertrophy did not alter the results (data not shown).SNP effects on QTpeak and QTend intervals were ana-

yzed for the SNPs that were significantly associated withPE interval. Figure 1 shows that both KCNH2 rs3807375nd NOS1AP rs10918594 were associated with prolonga-ion of QTend interval. For NOS1AP, however, the associa-ion with QTpeak interval was stronger than with QTend

nterval, resulting in shortening of TPE interval. The resultsemained the same when QT intervals were analyzed from

single lead (V3) and from the lead with the maximumrecordial QTend interval (data not shown).

able 2 TPE intervals by genotype and effect of SNPs on electr

ene SNPAllele* and aminoacid changes

Majorhomozygotes

TPE† (ms) %

CNH2 rs1805123 A¡C, K897T 83.8 67.9rs3807375 G¡A 82.7 32.4

CNE1 rs1805128 G¡A, D85N 83.4 97.3OS1AP rs2880058 A¡G 83.7 42.3

rs4657139 T¡A 83.6 41.3rs10918594 C¡G 83.9 43.8rs10494366 T¡G 83.7 41.6

Values are given as the mean.SNP � single nucleotide polymorphism; TPE � T-wave peak to T-wave

Major¡minor allele.Maximum precordial TPE interval. Percentages refer to prevalence estimaAdditive genetic model with values representing differences from majorressure, coronary heart disease, previous myocardial infarction, Cornell v

igure 1 Effect of KCNH2 K897T, KCNH2 rs3807375, and NOS1AP0918594 single-nucleotide polymorphisms (SNPs) on electrocardio-raphic QTpeak, QTend, and T-wave peak to T-wave end (TPE) interval.alues are differences from major homozygotes per one minor allele.ecause of the similarity of results for all other NOS1AP SNPs, results are

hown only for rs10918594. *P �.05; †P �.01; ‡P �.0001. Nc � nomo-

pram corrected for heart rate; n.s. � nonsignificant.

elationship of SNPs to T-wave morphologyarametershe relationship of SNPs to T-wave morphology parameters

PCA ratio, TMD, TCRT, TWR) was studied. Multivariableegression analyses showed that the KCNH2 K897T minorllele was associated with a 0.03-unit decrease in normal-zed TWR (P � .04), NOS1AP rs4657139 minor allele wasssociated with a 0.04-unit decrease in normalized TCRTP � .04), and NOS1AP rs10494366 minor allele was as-ociated with a 0.04-unit decrease in normalized TCRTP � .03). For all other SNPs, the results were nonsignifi-ant (both allelic and genotypic genetic models tested, dataot shown). When the three SNPs with significant resultsere tested with genotypic models, only the relationshipetween KCNH2 K897T and TWR was statistically signif-cant (P � .04). Analyses after excluding study subjectsith coronary heart disease, previous myocardial infarction,eart failure, or ECG left ventricular hypertrophy showedhat KCNH2 K897T minor allele was associated with a.04-unit decrease in normalized TMD (P � .02) and nor-alized TWR (P � .02). These associations remained sig-

ificant in genotypic models, but for all other SNPs, theesults were nonsignificant (data not shown).

iscussionain findingsur study results show that KCNH2 K897T minor allele C

s associated with shorter ECG TPE interval, and KCNH2ntronic rs3807375 minor allele A is associated with longerCG TPE interval. These findings are consistent with thessociation of these alleles with shorter and longer QTend

ntervals, respectively, as previously shown.6 NOS1AP mi-or alleles are associated with shorter TPE interval. This ispposite in direction to the effects of the same alleles on theuration of QTend interval and seems to be a result of theirtronger effect on QTpeak than QTend interval. None of theNPs showed a consistent association with T-wave mor-

ographic TPE interval

terozygotesMinorhomozygotes Minor allele effect‡

E† (ms) % TPE† (ms) % �TPE† (ms) P value

.5 28.8 81.6 3.3 �1.2 .00005

.7 49.4 84.1 18.2 �0.8 .001

.7 2.6 83.9 0.1 �1.3 .20

.4 45.5 82.1 12.2 �0.6 .01

.5 46.1 82.1 12.6 �0.5 .03

.3 44.0 82.0 12.3 �0.8 .002

.4 45.7 82.3 12.7 �0.6 .02

terval.

ygotes per one minor allele (adjustment for gender, age, systolic bloodduration product).

ocardi

He

TP

82838183838383

end in

tes.homoz

hology parameters.

EDpshrst2ar(c

tsgprnw

bHdwwa

KrEsgEpipatmmrKh

tmlowoscEIr

KKiscLLpdc

uQpohm1htapi

NIpNetCwp2Nl

rrQaLcctcSdastn

ahs

902 Heart Rhythm, Vol 7, No 7, July 2010

CG TPE interval and T-wave morphology measuresispersion of ventricular repolarization predisposes to re-olarization-related arrhythmogenesis.10 In experimentaltudies using myocardial wedge preparations, TPE intervalas been shown to correlate with transmural dispersion ofepolarization.10,11,23 In other studies, TPE interval has beenhown to be a measure of global dispersion of repolariza-ion.12,13 In an ambulatory ECG study of LQTS type 1 and

patients, TPE interval behavior paralleled experimentalnd computer-simulated findings in transmural dispersion ofepolarization in loss of function of the slow (IKs) and rapidIKr) components of the delayed rectifier current, whichause LQTS type 1 and 2, respectively.24

T-wave morphology parameters, which characterize thehree-dimensional morphology of the T wave, have beenhown to provide prognostic information on survival in theeneral population.14-16 Of note, in the Health 2000 Studyopulation, T-wave morphology parameters, but not heartate–corrected QTend interval, contained independent prog-ostic information on mortality, and the prognostic valueas specifically related to cardiovascular mortality.16

Several SNPs of KCNE1, KCNH2, and NOS1AP haveeen shown to be associated with longer QTend interval.1

owever, the recently observed increased risk of suddeneath with NOS1AP variants in a general population sampleas independent of the duration of QTend interval.9 Thus,e explored the effects of these variants on TPE interval

nd T-wave morphology parameters.

CNH2 K897T and rs3807375, andepolarization measuresxperimental studies have shown that the KCNH2 K897Tubstitution has multiple effects on cellular electrophysiolo-y,2,25-27 suggesting that the net effect of these changes onCG ventricular repolarization is difficult to predict. Ourresent study shows that K897T variant shortens ECG TPEnterval but does not consistently alter T-wave morphologyarameters. In the Health 2000 Study population analyzing thege-, gender- and heart rate–adjusted QT interval, we showedhat this polymorphism shortens QTend interval by 2.6-ms perinor allele.6 Several other studies also showed that this poly-orphism slightly shortens QTend interval,2-5 but conflicting

esults do exist.28 In previous reports, the impact of the minor897T allele has been suggested to be both protective andarmful, but sample sizes in these reports were small.29,30

No experimental studies have characterized the func-ional properties of KCNH2 intronic SNP rs3807375 onyocardial repolarization. We observed that the minor al-

ele of this variant prolongs TPE interval but has no effectsn T-wave morphology parameters. Previously this variantas reported to prolong QTend interval,5 a finding alsobserved in the Health 2000 Study population, whichhowed a 1.6-ms prolonging effect per minor allele.6 Be-ause prolongation of QTend and TPE interval on the restingCG is a typical feature in LQTS type 2 caused by reduced

Kr current,23,24,31 our ECG findings suggest that rs3807375

educes the repolarization reserve. w

CNE1 D85N and repolarization measuresCNE1 D85N polymorphism causes variation in minK, which

s the regulatory subunit of the IKs channel.32 Experimentaltudies have suggested that the minor D85N allele reduces IKs

urrent.33 Thus, the 85N allele may predispose to acquiredQTS and modify the clinical expression of congenitalQTS.34 In experimental LQTS type 1 models, IKs blockrolongs action potential duration homogeneously; therefore, itoes not increase transmural dispersion at rest but does so inonjunction with �-adrenergic stimulation.31

In a previous analysis from the Health 2000 Study pop-lation, we observed a substantial 10.5-ms prolongation inTend interval per each KCNE1 D85N minor allele.6 Theresent results show that, despite the marked prolongationf QTend interval observed in our prior study,6 this variantas no effect on ECG TPE interval or T-wave morphologyeasures. In ambulatory ECG recordings from LQTS typepatients, TPE interval increased significantly only at high

eart rates.24 Thus, our findings in resting ECGs showinghat the D85N variant prolongs QTend but not TPE intervalre consistent with reduced IKs current. Whether the D85Nolymorphism is associated with a prolongation of TPEnterval during physical exercise remains to be studied.

OS1AP variants and repolarization measuresn experimental studies, overexpression of CAPON, theroduct of the NOS1AP gene, results in up-regulation ofOS1-NO signaling pathways, leading to reduced ICa,L and

nhanced IKr currents and shortened action potential.35 Pos-ulating that genetic variants in NOS1AP lead to lowerAPON levels, increased ICa,L and decreased IKr currentsith prolonged action potential are expected.8 In previousopulation studies, including our recent study of the Health000 Study population, the minor alleles of severalOS1AP variants have been shown to be associated with

onger QTend interval.6,7

Crotti et al8 recently reported that NOS1AP variantss4657139 (one of the SNPs included in the present study) ands16847548 were associated with greater prolongation ofTend interval, increased risk for symptoms, and greater prob-

bility for cardiac arrest and sudden death in a South AfricanQTS type 1 founder population. They suggested that in-reased ICa,L current might be associated with arrhythmogeniconsequences in LQTS type 1 patients.8 Kao et al9 reportedhat NOS1AP variants rs16847548 and rs12567209 were asso-iated with sudden cardiac death in a large cohort of Unitedtates white adults, but the effect was independent of theuration of QTend interval in multivariable analyses. Theseuthors suggested that either QTend interval might not be aufficient measure of ventricular repolarization or the prognos-ic value of NOS1AP variants might be mediated by mecha-isms other than repolarization.9

In the present study, the minor alleles of NOS1AP vari-nts examined were associated with shorter TPE interval butad no effects on T-wave morphology parameters. With aupposed decreased IKr current, TPE-interval shortening

as an unexpected finding. Of interest, our results suggest a

sQmm

SOdldaisSfep

COaniNQvttnHdv

R

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

903Porthan et al Single-Nucleotide Polymorphisms and T Wave

tronger prolonging effect of NOS1AP SNPs on QTpeak thanTend interval. Whether the shortening of TPE interval isediated by increased ICa,L or some other mechanism re-ains to be studied.

tudy limitationsur results may not be applicable to populations fromifferent genetic backgrounds or to other races, and rep-ication of the observed associations in another, indepen-ent cohort is warranted. The association between SNPsnd TPE interval appears weak; thus, its clinical signif-cance remains to be studied. Our analyses included fourtrongly correlating NOS1AP SNPs, and some otherNPs not examined in the present study may have dif-erent repolarization effects and prognostic value.9 Thexact electrophysiologic basis of T-wave morphologyarameters currently is unknown.

onclusionur results show that the KCNH2 K897T minor allele is

ssociated with shortening, and the KCNH2 rs3807375 mi-or allele is associated with prolongation of ECG TPEnterval. KCNE1 D85N has no effect on TPE interval. TheOS1AP minor alleles studied prolong QTpeak more thanTend interval, with the net effect being shorter TPE inter-al. None of the SNPs studied showed consistent associa-ion with ECG T-wave morphology parameters, suggestinghat these polymorphisms do not mediate the observed prog-ostic value of T-wave morphology parameters in theealth 2000 Study. The repolarization parameters studiedo not appear to clarify the association between NOS1APariants and sudden cardiac death.

eferences1. Noseworthy PA, Newton-Cheh C. Genetic determinants of sudden cardiac

death. Circulation 2008;118:1854–1863.2. Bezzina CR, Verkerk AO, Busjahn A, et al. A common polymorphism in

KCNH2 (HERG) hastens cardiac repolarization. Cardiovasc Res 2003;59:27–36.

3. Gouas L, Nicaud V, Berthet M, et al. Association of KCNQ1, KCNE1, KCNH2and SCN5A polymorphisms with QTc interval length in a healthy population.Eur J Hum Genet 2005;13:1213–1222.

4. Pfeufer A, Jalilzadeh S, Perz S, et al. Common variants in myocardial ionchannel genes modify the QT interval in the general population: results from theKORA study. Circ Res 2005;96:693–701.

5. Newton-Cheh C, Guo CY, Larson MG, et al. Common genetic variation inKCNH2 is associated with QT interval duration: the Framingham Heart Study.Circulation 2007;116:1128–1136.

6. Marjamaa A, Newton-Cheh C, Porthan K, et al. Common candidate genevariants are associated with QT interval duration in the general population.J Intern Med 2009;265:448–458.

7. Arking DE, Pfeufer A, Post W, et al. A common genetic variant in the NOS1regulator NOS1AP modulates cardiac repolarization. Nat Genet 2006;38:644–651.

8. Crotti L, Monti MC, Insolia R, et al. NOS1AP Is a Genetic Modifier of theLong-QT Syndrome. Circulation 2009;120:1657–1663.

9. Kao WH, Arking DE, Post W, et al. Genetic variations in nitric oxide synthase1 adaptor protein are associated with sudden cardiac death in US white com-munity-based populations. Circulation 2009;119:940–951.

0. Yan G-X, Antzelevitch C. Cellular basis for the normal T wave and theelectrocardiographic manifestations of the long-QT syndrome. Circulation 1998;98:1928–1936.

1. Antzelevitch C. Cellular basis for the repolarization waves of the ECG. Ann N

Y Acad Sci 2006;1080:268–281.

2. Zabel M, Portnoy S, Franz MR. Electrocardiographic indexes of dispersion of

ventricular repolarization: an isolated heart validation study. J Am Coll Cardiol1995;25:746–752.

3. Opthof T, Coronel R, Wilms-Schopman FJ, et al. Dispersion of repolarization incanine ventricle and the electrocardiographic T wave: Tp-e interval does notreflect transmural dispersion. Heart Rhythm 2007;4:341–348.

4. Okin PM, Devereux RB, Fabsitz RR, Lee ET, Galloway JM, Howard BV.Principal component analysis of the T wave and prediction of cardiovascularmortality in American Indians: the Strong Heart Study. Circulation 2002;105:714–719.

5. Zabel M, Malik M, Hnatkova K, et al. Analysis of T-wave morphology from the12-lead electrocardiogram for prediction of long-term prognosis in male USveterans. Circulation 2002;105:1066–1070.

6. Porthan K, Viitasalo M, Jula A, et al. Predictive value of electrocardio-graphic QT interval and T-wave morphology parameters for all-cause andcardiovascular mortality in a general population sample. Heart Rhythm2009;6:1202–1208.

7. Aromaa A, Koskinen S, editors. Health and functional capacity in Finland.Baseline results of the Health 2000 Health Examination Survey. PublicationNumber B12/2004. Helsinki: Publications of the National Public Health Insti-tute. Accessed September 18, 2008. Available at http://www.terveys2000.fi/julkaisut/baseline.pdf.

8. Porthan K, Virolainen J, Hiltunen TP, et al. Relationship of electrocardiographicrepolarization measures to echocardiographic left ventricular mass in men withhypertension. J Hypertens 2007;25:1951–1957.

9. Oikarinen L, Paavola M, Montonen J, et al. Magnetocardiographic QT intervaldispersion in postmyocardial infarction patients with sustained ventricular tachy-cardia: validation of automated QT measurements. Pacing Clin Electrophysiol1998;21:1934–1942.

0. Hekkala A-M, Väänänen H, Swan H, Oikarinen L, Viitasalo M, Toivonen L.Reproducibility of computerized measurements of QT interval from multiple leadsat rest and during exercise. Ann Noninvasive Electrocardiol 2006;11:318–326.

1. Andersen MP, Xue JQ, Graff C, Kanters JK, Toft E, Struijk JJ. New descriptorsof T-wave morphology are independent of heart rate. J Electrocardiol 2008;41:557–561.

2. Blom G. Statistical Estimates and Transformed Beta Variables. New York: JohnWiley & Sons, Inc.; 1958.

3. Gima K, Rudy Y. Ionic current basis of electrocardiographic waveforms: amodel study. Circ Res 2002;90:889–896.

4. Viitasalo M, Oikarinen L, Swan H, et al. Ambulatory electrocardiographicevidence of transmural dispersion of repolarization in patients with long-QTsyndrome type 1 and 2. Circulation 2002;106:2473–2478.

5. Anson BD, Ackerman MJ, Tester DJ, et al. Molecular and functional charac-terization of common polymorphism in HERG (KCNH2) potassium channels.Am J Physiol Heart Circ Physiol 2004;286:H2434–H2441.

6. Paavonen KJ, Chapman H, Laitinen PJ, et al. Functional characterization of thecommon amino acid 897 polymorphism of the cardiac potassium channelKCNH2 (HERG). Cardiovasc Res 2003;59:603–611.

7. Gentile S, Martin N, Scappini E, Williams J, Erxleben C, Armstron DL. The humanERG1 channel polymorphism, K897T, creates a phosphorylation site that inhibitschannel activity. Proc Natl Acad Sci U S A 2008;105:14704–14708.

8. Pietilä E, Fodstad H, Niskasaari E, et al. Association between HERG K897Tpolymorphism and QT interval in middle-aged Finnish women. J Am CollCardiol 2002;40:511–514.

9. Crotti L, Lundquist AL, Insolia R, et al. KCNH2-K897T is a genetic modifier oflatent congenital long-QT syndrome. Circulation 2005;112:1251–1258.

0. Zhang X, Chen S, Zhang L, et al. Protective effect of KCNH2 single nucleotidepolymorphism K897T in LQTS families and identification of novel KCNQ1 andKCNH2 mutations. BMC Med Genet 2008;9:87.

1. Shimizu W, Antzelevitch C. Cellular basis for the ECG features of the LQT1form of the long QT-syndrome: effect of �-adrenergic agonists and antagonistsand sodium channel blockers on transmural dispersion of repolarization andtorsades de pointes. Circulation 1998;98:2314–2322.

2. Lehmann-Horn F, Jurkat-Rott K. Voltage-gated ion channels and hereditarydisease. Physiol Rev 1999;79:1317–1372.

3. Westenskow P, Splawski I, Timothy KW, Keating MT, Sanguinetti MC. Com-pound mutations: a common cause of severe long-QT syndrome. Circulation2004;15:1834–1841.

4. Nishio Y, Makiyama T, Itoh H, et al. D85N, a KCNE1 polymorphism, is adisease-causing gene variant in long QT syndrome. J Am Coll Cardiol 2009;54:812–819.

5. Chang KC, Barth AS, Sasano T, et al. CAPON modulates cardiac repolar-ization via neuronal nitric oxide synthase signaling in the heart. Proc Natl

Acad Sci U S A 2008;105:4477– 4482.