New peds.2011-0643 - Parent Heart Watch | Sudden Cardiac Arrest … · 2012. 3. 5. · Cause Sudden Cardiac Death in Asymptomatic Children: A Meta-analysis abstract BACKGROUNDANDOBJECTIVES:Pediatricsuddencardiacdeath(SCD)

DOI: 10.1542/peds.2011-0643; originally published online March 5, 2012;Pediatrics

Laurel K. LeslieIp, Jane W. Newburger, Susan K. Parsons, Thomas A. Trikalinos, John B. Wong and Angie Mae Rodday, John K. Triedman, Mark E. Alexander, Joshua T. Cohen, Stanley

in Asymptomatic Children: A Meta-analysisElectrocardiogram Screening for Disorders That Cause Sudden Cardiac Death

http://pediatrics.aappublications.org/content/early/2012/02/29/peds.2011-0643

located on the World Wide Web at: The online version of this article, along with updated information and services, is

of Pediatrics. All rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.Boulevard, Elk Grove Village, Illinois, 60007. Copyright © 2012 by the American Academy published, and trademarked by the American Academy of Pediatrics, 141 Northwest Pointpublication, it has been published continuously since 1948. PEDIATRICS is owned, PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly

at University of Missouri-Columbia on March 6, 2012pediatrics.aappublications.orgDownloaded from

http://pediatrics.aappublications.org/content/early/2012/02/29/peds.2011-0643http://pediatrics.aappublications.org/

Electrocardiogram Screening for Disorders ThatCause Sudden Cardiac Death in AsymptomaticChildren: A Meta-analysis

abstractBACKGROUND AND OBJECTIVES: Pediatric sudden cardiac death (SCD)occurs in an estimated 0.8 to 6.2 per 100 000 children annually.Screening for cardiac disorders causing SCD in asymptomatic chil-dren has public appeal because of its apparent potential to averttragedy; however, performance of the electrocardiogram (ECG) as ascreening tool is unknown. We estimated (1) phenotypic (ECG- or echo-cardiogram [ECHO]-based) prevalence of selected pediatric disordersassociated with SCD, and (2) sensitivity, specificity, and predictive valueof ECG, alone or with ECHO.

METHODS: We systematically reviewed literature on hypertrophic car-diomyopathy (HCM), long QT syndrome (LQTS), and Wolff-Parkinson-White syndrome, the 3 most common disorders associated with SCDand detectable by ECG.

RESULTS: We identified and screened 6954 abstracts, yielding 396articles, and extracted data from 30. Summary phenotypic prevalencesper 100 000 asymptomatic children were 45 (95% confidence interval[CI]: 10–79) for HCM, 7 (95% CI: 0–14) for LQTS, and 136 (95% CI: 55–218) for Wolff-Parkinson-White. The areas under the receiver operatingcharacteristic curves for ECG were 0.91 for detecting HCM and 0.92 forLQTS. The negative predictive value of detecting either HCM or LQTS byusing ECG was high; however, the positive predictive value varied bydifferent sensitivity and specificity cut-points and the true prevalence ofthe conditions.

CONCLUSIONS: Results provide an evidence base for evaluating pedi-atric screening for these disorders. ECG, alone or with ECHO, was a sen-sitive test for mass screening and negative predictive value was high,but positive predictive value and false-positive rates varied. Pediatrics2012;129:1–12

AUTHORS: Angie Mae Rodday, MS,a John K. Triedman, MD,b,c

Mark E. Alexander, MD,b,c Joshua T. Cohen, PhD,a,d Stanley Ip,MD,a,d Jane W. Newburger, MD, MPH,b,c Susan K. Parsons, MD,MRP,a,d Thomas A. Trikalinos, MD, PhD,a,d John B. Wong, MD,a,d

and Laurel K. Leslie, MD, MPHa,d

aTufts Medical Center, Boston, Massachusetts; bChildren’sHospital Boston, Boston, Massachusetts; cHarvard MedicalSchool, Boston, Massachusetts; and dTufts University School ofMedicine, Boston, Massachusetts

KEY WORDSsudden cardiac death, ECG screening, hypertrophiccardiomyopathy, long QT syndrome, Wolff-Parkinson-Whitesyndrome

ABBREVIATIONSAUC—area under the HSROC curveCI—confidence intervalECG—electrocardiogramECHO—echocardiogramHCM—hypertrophic cardiomyopathyHSROC—hierarchical summary receiver operating characteristicLQTS—long QT syndromeNPV—negative predictive valuePPV—positive predictive valueSCD—sudden cardiac deathWPW—Wolff-Parkinson-White syndrome

www.pediatrics.org/cgi/doi/10.1542/peds.2011-0643

doi:10.1542/peds.2011-0643

Accepted for publication Nov 22, 2011

Address correspondence to Laurel K. Leslie, MD, MPH, 800Washington St, #345, Boston, MA 02111. E-mail:[email protected]

PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275).

Copyright © 2012 by the American Academy of Pediatrics

FINANCIAL DISCLOSURE: Dr Triedman is a consultant forBiosense Webster, Inc, and has received a speaker’s honorariafrom St Jude Medical; the other authors have indicated theyhave no financial relationships relevant to this article todisclose.

FUNDING: Supported by grant 1RC1HL100546-01 from theNational Heart, Lung, and Blood Institute. Consultation from theTufts Clinical and Translational Science Institute was supportedby grant RR025752 from the National Center for ResearchResources. Funded by the National Institutes of Health (NIH).

PEDIATRICS Volume 129, Number 4, April 2012 1

REVIEW ARTICLE


mailto:[email protected]://pediatrics.aappublications.org/

Although sudden cardiac death (SCD) inchildren and adolescents (hereafter“children”) is rare (0.8–6.2 per 100 000annual incidence1,2), the sudden deathof a child is tragic and has widespreadrepercussions. Concern about SCD hasraised calls for screening in primarycare or school-based settings for allchildren; others have recommendedscreening for subgroups of childrenstarting stimulants or participatingin competitive athletics, both of whichincrease heart rate and may theoreti-cally precipitate SCD.1,3,4

Population-based screening programsthat identify children at risk for SCDhave broad public appeal, as commonsense suggests that presymptomaticdiagnosis saves lives, and the societalcost is presumed to be the cost of thescreening test itself. Japan is the solecountry with published data on massscreening of school-aged children, in-cluding a targeted cardiac history andphysical and electrocardiogram (ECG).5

No data regarding mass pediatricscreening and associated costs areavailable in the United States.6

In 2008, the American Heart Associationreleased a statement7 broadly inter-preted as recommending an ECG beforeinitiating stimulants for children withattention-deficit/hyperactivity disorder,estimated at 4% to 12% of children.8

The American Academy of Pediatricslater released a statement, in collabo-ration with the American Heart Asso-ciation, recommending that childrenwith attention-deficit/hyperactivity dis-order be assessed with a targeted his-tory and cardiac examination but thatfurther evaluation, including an ECG, beobtained only if indicated.9 A recentdecision analysis recommended thatchildren participating in competitivesports undergo mass screening.3 Sev-eral studies have described screeningprograms for athletes. Italy uses pre-participation screening, includingECGs, for athletes aged 12 to 351 and

some American universities use pre-participation screening and ECGs forcollege athletes. Because 10 millionpeople in the United States may beclassified as “young competitiveathletes,”10 calls for screening havefar-reaching implications.

Screening programs are most effectiveif (1) preclinical prevalence is suffi-ciently high in the screened population,(2) a highly discriminatory screeningtest is available, (3) the disease or dis-order is serious, and (4) treatmentwhile asymptomatic decreases mor-bidity and mortality more than treat-ment after symptoms develop.11 Thesecriteria enable evaluation of the effi-ciency of ECG to detect the disordersthat may cause SCD in asymptomaticchildren.

Several rare disorders cause pediatricSCD, but not all have ECG findings.12 Themost common disorders detectable byECG are hypertrophic cardiomyopathy(HCM), long QT syndrome (LQTS), andWolff-Parkinson-White syndrome (WPW).Their estimated prevalence rates arelow in otherwise healthy, asymptomaticchildren; moreover, the value of the ECGas a “highly discriminatory” test is notwell established. The ECG should se-lectively identify disorders respon-sible for SCD in all affected patients(ie, sensitivity) and rule out these dis-orders in healthy children (ie, specificity).Together, low prevalence and imperfectsensitivity and specificity estimatescould result in inefficient screeningstrategies with unanticipated societaland economic costs.

We undertook a systematic review andmeta-analysisof the literatureof these3disorders that cause SCD. Our first aimwas to summarize how often ECG- orechocardiogram (ECHO)-based testing(phenotypic prevalence) suggests HCM,LQTS, orWPWamongasymptomatic andundiagnosed children who could beidentified by mass screening. We fo-cused onphenotypic prevalence, rather

than genetic prevalence, because ge-netic testing is currently impractical inmass screening programs and is lim-ited to diagnosis or risk stratification.Our second aim was to examine thereported sensitivity and specificity ofthe ECG, alone or with ECHO, to detectthese disorders and calculate predictivevalues. Together, this information onphenotypic (ECG- or ECHO-based) prev-alence, sensitivity, specificity, and pre-dictive value form an evidence base thatwill facilitate further evaluation of theefficiency and downstream implicationsof ECG screening programs for SCD.

METHODS

We focused on HCM, LQTS, and WPWbecause they are the most commondisorders potentially detectable by ECGamong children. Because ECG findingsin HCM are age sensitive (ie, may not bedetected until late adolescence or earlyadulthood)andcanbenonspecific, thusrequiring an ECHO for diagnostic guid-ance, we chose to include articles thatexamined test characteristics of ECGalone, ECHO alone, or ECG combinedwithECHO (ECG/ECHO).

Literature Searches

We performed a systematic review andsearched the Medline database (1950to December 2010) for studies report-ing on HCM, LQTS, WPW, and SCD or onECG and/or ECHO detection of thesedisorders. We combined keywords andMedical Subject Heading terms forhypertrophic cardiomyopathy, longQT syndrome, Wolff-Parkinson-Whitesyndrome, sudden cardiac death, elec-trocardiography, echocardiography, sen-sitivity, and specificity. The search waslimited to English-language publica-tions of primary studies in humanswith no geographic restrictions. Sixreviewers screened titles and ab-stracts to identify relevant studiesand then examined full-text articlesfor eligibility.

2 RODDAY et al at University of Missouri-Columbia on March 6, 2012pediatrics.aappublications.orgDownloaded from

http://pediatrics.aappublications.org/

Eligibility Criteria

To summarize how often ECG- or ECHO-based testing (phenotypic prevalence)suggests HCM, LQTS, or WPW in asymp-tomatic children or young adults (3–25years old), we included cross-sectionalor cohort studies from the general pop-ulation that used ECG or ECHO diagnosticcriteria for each disorder consistentwith clinical standards. Studies in whichthemean age was.2 SDs from 25 yearswere excluded, including a recent studyfocused on neonates.13 We also excludedstudies of highly selected subgroups thatwere not representative of the generalpopulation. For example, “elite” athleteswho competed in competitive regional,national, or international events wereexcluded, but studies of normally activehigh school athletes were included.Studies that used sampling techniquesthat might result in a nonrepresentativesample (eg, convenience sampling,studies requiring informed consent fromparticipants) were excluded. Studiesassessing the frequency of genetic var-iations related to HCM, LQTS, or WPWwere excluded, given our focus on ECGscreening in asymptomatic and pre-viously undiagnosed children.

For our second aimwe included studieswithdataon thesensitivity or specificityof ECG (with orwithout ECHO) to identifychildren who would have a diagnosis ofHCM, LQTS, or WPWaccording to clinicalcriteria. Specifically, wedeemed that anadequate reference (“gold”) standardfor HCM is ECHO, genotyping, or a well-documented HCM diagnosis. For LQTS,we accepted as reference standardtesting for pathogenic variations (eg,the KCNQ1, KCNH2, and SCN5A genes) ora combination of personal and familyhistory, clinical follow-up, and ECG. ForWPW, ECG is the reference standard, sosensitivity and specificity informationwere not collected. Based on thesecriteria, studies with incorporation bi-as (where the index test comprisespart of the reference standard against

which it is measured) were eligible. Weincluded studies in which only those withpositive ECG and/or ECHO were verifiedwith the reference standard (verificationbias, which may overestimate sensitivityand underestimate the specificity of theindex test). For those studies that hadmultiple alternative sets of ECG and/orECHO criteria, we selected the mostwidely used or most sensitive criteriato avoid duplication of information.

Data Extraction

Four reviewers extracted data with atleast 2 independently extracting or re-viewing each article. All 4 reviewers dis-cussedandresolvedanydiscrepanciesbyconsensus. From studies informing onphenotypic (ECG- or ECHO-based) preva-lence, we extracted information on studypopulation (description, country), studydesign (prospective, retrospective),sampling technique (representative ornot), age of the study sample, samplesize, diagnostic criteria, and number ofparticipants with each disorder.

From studies on the sensitivity andspecificity of ECG and/or ECHO to identifyHCMorLQTS,weextracted informationonstudy population (description, country),age of the study sample, type of test (ECGand/or ECHO), diagnostic criterion andthresholds for the test, reference stan-dard definition, true-positive, false-negative, false-positive, true-negative,and presence of verification bias. If thestudy provided only sensitivity andspecificity, we used this information tocalculate thetrue-positive, false-negative,false-positive, and true-negative values.

Analysis

Because of the complexities of ourmethods, we briefly discuss analysesin the following paragraphs and pro-vide detailed Supplemental Informa-tion that discusses characteristics ofscreening tests in general (eg, sensi-tivity, specificity, predictive value) andanalytic methods used (eg, creation of

hierarchical summary receiver oper-ating characteristic [HSROC] curve).

Analysis of Phenotypic Prevalence

Phenotypic (ECG- or ECHO-based) preva-lence (per 100 000) estimates and 95%confidence intervals (CIs) for HCM, LQTS,and WPW were calculated by using theexact binomial distribution. We obtainedsummary estimates of phenotypicprevalence by using random effectsmeta-analysis of logit-transformedphenotypic prevalence.14 To assess theextent to which variation in the reportedoutcomes may be a result of chancealone, we used Cochran Q to test forheterogeneity (significant when P ,.10) and quantified its magnitude interms of I2,15 which ranges between0% and 100% and expresses the pro-portion of between-study variabilityattributable to heterogeneity ratherthan chance. We considered I2 valuesexceeding 75% suggestive of substantialheterogeneity. These calculations wereperformed by using Stata, version 11(StataCorp LP, College Station, TX).

Analysis of Sensitivity andSpecificity

For each disorder, we summarized therelationship between sensitivity (ie, theprobability of having a positive testamong those with the disorder) andspecificity (ie, the probability of havinganegative test among thosewithout thedisorder) of ECG and/or ECHO with anextension of the HSROC model.16–18 Foreach disorder and screening tool com-bination, we plotted the HSROC curve forindividual studies and the HSROC curvefor the summary of the reviewed stud-ies. These curves allow visual compari-son between individual studies and thesummary curve. Points along the sum-mary curves incorporate different di-agnostic criteria and do not corresponddirectly to specific observed study cutpoints for ECG and/or ECHO. Restrictingthe range to that observed in the data,the area under the posterior estimate

REVIEW ARTICLE

PEDIATRICS Volume 129, Number 4, April 2012 3 at University of Missouri-Columbia on March 6, 2012pediatrics.aappublications.orgDownloaded from

http://pediatrics.aappublications.org/lookup/suppl/doi:10.1542/peds.2011-0643/-/DCSupplementalhttp://pediatrics.aappublications.org/lookup/suppl/doi:10.1542/peds.2011-0643/-/DCSupplementalhttp://pediatrics.aappublications.org/

of the HSROC curve (AUC) calculatedby numeric integration indicates testperformance. An AUC of 1.0 representsa perfect test, whereas an AUC of 0.5represents a test that performs nobetter than chance.

For each summary curve, we identified2 illustrative examples to demonstratehowchanges in sensitivity andspecificityresulted in different predictive values,numberneeded toscreen, false-positives,and false-negatives. We selected 2 pointson the HSROC curve: (1) the point with“maximal accuracy” (ie, maximizing thesum of the sensitivity and specificity),thereby giving equal weight to ruling inpeople with disease (sensitivity) andruling out those without disease (speci-ficity); and (2) the point with “maximalspecificity” where specificity was near 1and the corresponding sensitivity,thereby giving more weight to rulingout those without the disease (speci-ficity). We did not select a point on thecurve where sensitivity was maximizedbecause the corresponding specificitywas low (0.001). By using the 2 illus-trative points, we calculated 5 param-eters: (1) positive predictive value (PPV,ie, the probability of having the disor-der given a positive test), (2) negativepredictive value (NPV, ie, the probabilityof not having the disorder given a neg-ative test), (3) number needed toscreen to detect 1 case, (4) number offalse-positives when detecting 1 case,and (5) number of false-negatives per100 000 children screened. To explorethe effect of alternative prevalencerates, we repeated these calculationsby using oft-cited prevalences of 200per 100 000 for HCM,19 50 per 100 000for LQTS,13 and 200 per 100 000 for WPW(See Supplemental Information, Sup-plemental Figures 4-8, and Supple-mental Tables 1-3 for more informationon screening trade-offs in general.).20

Sensitivity Analysis

To determine whether alternativeassumptions substantially affected the

meta-analysis results, we performedextensive sensitivity analyses.21 For keyquestions related to phenotypic (ECG-or ECHO-based) prevalence, we re-peated the analyses excluding studieswhere (1) diagnostic criteria were notspecified, (2) reported phenotypic prev-alence exceeded the range of often-citedprevalence rates, or (3) phenotypicprevalence was based on previouslydiagnosed cases and not asymptom-atic cases. For key questions address-ing the ability of ECG- or ECHO-basedtesting to diagnose people with theconditions of interest, we repeated theanalyses by excluding studies that didnot apply the reference standard toparticipants with a nonsuggestive ECGand/or ECHO (ie, verification bias).

For additional sensitivity analyses, weback-calculated disease prevalencewhen applying a non-ECG referencestandard to define disease. A positivescreening ECG (eg, a result suggestingLQTS) can be either a true-positive (thepersonhasLQTS)ora false-positive (thepersondoes not andwill not have LQTS).Thus, the frequency of “suggestiveECGs” is not the same as the preva-lence of the disease. One can back-calculate the prevalence of LQTS froman acceptable alternative non-ECG ref-

erence standard to diagnose LQTS (eg,genetic testing for deleterious mutationsin LQTS genes13), and from the pro-portion of positive ECG tests in a pop-ulation. We performed such analysesonly for LQTS, as an example to con-textualize our discussion comments. Asdescribed in the Supplemental In-formation, we extended the Bayesianmethod of Joseph and colleagues.22

RESULTS

From 6954 titles and abstracts screenedforeligibility,we retrieved andevaluatedthe full text in 396 articles, with 30meetingeligibility criteria (Fig 1).1,19,23–50

Characteristics of ReviewedStudies

In the 11 primary studies1,19,23–31 thatreported phenotypic (ECG- or ECHO-based) prevalence findings, studypopulations ranged from general tosubgroups of high school athletes andmilitary conscripts. Studies were con-ducted in North America, Europe, andAsia and had sample sizes rangingfrom 1369 to 1 336 377 (Table 1).

Twenty primary studies28,32–50 reportedsensitivity and specificity estimates byusing ECG to detect LQTS and/or HCM,

FIGURE 1Literature search strategy.


http://pediatrics.aappublications.org/lookup/suppl/doi:10.1542/peds.2011-0643/-/DCSupplementalhttp://pediatrics.aappublications.org/lookup/suppl/doi:10.1542/peds.2011-0643/-/DCSupplementalhttp://pediatrics.aappublications.org/

ECHO to detect HCM, or ECG/ECHO todetect HCM. Populations in these stud-ies were mainly disease probands andtheir relatives. The studies were con-ducted in North America, Europe, andAsia and had sample sizes ranging from23 to 2770 (Table 2).

Hypertrophic Cardiomyopathy

Based on 7 studies,1,19,24,25,27–29 HCMphenotypic (ECG- or ECHO-based)prevalence ranged from 0 to 170per 100 000 (Fig 2) with a summaryphenotypic prevalence rate of 45 per100 000 (95% CI: 10–79) but with sub-stantial variation between studies (I2 =91%, P , .001). Inclusion of the studywith a phenotypic prevalence estimateof zero29 helped inform the upperbounds of the estimate. Although 1study27 did not specify diagnostic cri-teria, we included it because its exclu-sion had no effect on phenotypicprevalence (remained at 45 per 100000) and this study was based on awell-established screening program inJapan. When adding 2 nonrepresenta-tive studies23,26 for sensitivity analysis,the summary phenotypic prevalencerate decreased to 13 per 100 000 (95%CI: 7–19) with substantial heterogene-ity (I2 = 92%, P , .001). Turning toscreening for HCM, Fig 3 illustratesa set of HSROC curves for detectionby ECG (10 studies),28,32–40 ECHO(6 studies),33–35,37,40,41 and ECG/ECHO (4 studies).33–35,40 Based onthe summary HSROC curves, the AUCvalues were high.

To provide clinical context for inter-preting these results, we used thesummary phenotypic prevalence esti-mate for HCM and the 2 previouslydescribed illustrative points (the max-imal accuracy point and the maximalspecificity point) on the HSROC curvesfordetection of HCMbyusing ECG, ECHO,and ECG/ECHO (Table 3). Regardless ofwhether an ECG, ECHO, or ECG/ECHO wasused, both illustrative points yielded anTA

BLE1

ArticlesReportingPhenotypic(ECG-or

ECHO

-based)Prevalence

ofHCM,LQTS,andWPW

Author

PopulationandSource

Location

StudyDesign

SamplingTechnique

Age,y

SampleSize

DiagnosticCriteria

HCM,n

=9

Arola(1997)

23a

Medicalchartreviewwithinhospitals

Finland

RCRepresentative

Range:0–20

1336377

Interventricular

septum

orLV

wall

thickness$2SD

ofnorm

alColivicchi(2004)24

Pre-participationathleticscreening

Italy

PCRepresentative

Mean=

16.2SD=2.4

7568

LVwallthickness

$13

mm

Corrado(1998)

25Pre-participationathleticscreening

Italy

PCRepresentative

Mean=

19SD=5

33735

LVwallthickness

$13

mm

Corrado(2006)


Italy

PCRepresentative

Range:12–35

42386

LVwallthickness

$13

mm

Maron

(1995)

19Epidem

iology

studywith

subjects

selected

from

generalpopulation

USA

PCRepresentative

Range:23–35

4111

LVwallthickness

$15

mm

Maron

(1999)

26a

Diagnostictestingrequestedby

primaryphysicianinruralcom

munity

USA

PCRepresentative

Range:16–87

15137

LVwallthickness

.13

mm

Niimura(1989)

27Screeningof“presumablyhealthy”

nursery

schoolandjunior

high

schoolchildren

Japan

PCRepresentative

Ranges:3–5,12–14

930939

Notspecified

Nistri(2003)

28Screeningofmilitary

recruits(m

ales

only)

before

mandatory

military

service

Italy

RCRepresentative

Mean=

19SD=2

34910

LVwallthickness

$15

mm

Zou(2004)

29Epidem

iology

studyinrandom

sample

ofgeneralpopulation

China

PCRepresentative

Range:18–29

1369

LVwallthickness

$13

mm

LQTS,n

=4

Chiu(2008)

30Citywidesurvey

ofgeneralpopulation

Taiwan

PCRepresentative

Range:6–20

430391

QTc.450ms

Corrado(2006)


Italy

PCRepresentative

Range:12–35

42386

Male:QTc.440ms,Female:QTc.460ms

Kobza(2009)

31a

Screeningofmilitary

recruits(m

ostly

male)

before

mandatory

military

service

Switzerland

PCRepresentative

Mean=

19.2SD=1.4

40917

Male:QTc.450ms,Female:QTc.460ms

Niimura(1989)

27Screeningof“presumablyhealthy”

nursery

schoolandjunior

high

schoolchildren

Japan

PCRepresentative

Ranges:3–5,12–14

930939

Notspecified

WPW

,n=3

Chiu(2008)

30Citywidesurvey

ofgeneralpopulation

Taiwan

PCRepresentative

Range:6–20

430391

PRinterval,120ms;slurredupstroke

oftheQR

Scomplex;QRS

.120ms

Corrado(1998)


Italy

PCRepresentative

Mean=

19SD=5

33735

PRinterval,0.12

s;QR

S$0.12

sCorrado(2006)


Italy

PCRepresentative

Range:12–35

42386

PRinterval,0.12

s;QR

S$0.12

s

LV,leftventricular;PC,prospectivecohort;QTc,corrected

QTinterval;RC,retrospectivecohort.

aThesestudieswereincluded

insensitivityanalysisonly.

REVIEW ARTICLE



TABLE2

ArticlesReportingSensitivity

andSpecificity

forScreeningECGand/or

ECHO

todetect

HCM

andLQTS

Disorder

&Screening

Test

Author

(year)

Sample

Location

Age,y

ScreeningTestCriteria

Definitionof

Reference

(“Gold”)Standard

Verification

Bias

Sample

Size

HCM,n=11

ECG,n=

10Autore

(1988)

32First-degreerelatives

ofpatientswith

HCM

Italy

Mean=

36SD=20

LVhypertrophy;abnorm

alQwaves;negativeTwaves;

atrialfibrillation;leftor

rightb

undlebranch

block

ECHO

No72

Charron(1997)

33Genotypedprobands

and

first-degreerelatives

France

Range:18–29

Qwaves;LVhypertrophy,repolarizationalterations;

isolated

leftatrialenlargem

ent;shortP

Rinterval;

microvoltage;m

inor

Qwaves;bundle-branch

blockor

hemiblock

Genotyping

No58

Charron(1998)

34ChildrenofHCMgenotyped

families

France

,18

Qwaves;voltage;repolarizationalterations;abnormal

PRinterval,leftand/orrightatrialenlargement;

atrialfibrillation;abnorm

alQR

Saxis;increased

QRSduration;increasedventricularactivationtim

e;Twaves;R/S

ratio,rSr’aspect;bundlebranch

blockor

hemiblock;m

icrovoltage

Genotyping

No35

Charron(2003)

35Genotypedprobands

and


France

Mean=

37.7SD=17.9

abnorm

alQwaves;T-waveinversion;LV

hypertrophy

Genotyping

No109

Dipchand

(1999)

36Childrenwith

HCMandhealthy

controls

Canada

Median=

3,Range:

0–19

Qwaves;R

waves;S

waves;T

waves;QTc

interval;voltage

ECHO

,LV

angiography

Yes

73

Fragola(1993)


ofpatientswith

HCM

Italy

Mean=

34SD=19

LVandRV

hypertrophy;atrialenlargem

ent;rhythm

disturbances;atrioventricularandintraventricular

conduction;ST-Tdisplacement;Qwaves;R

waves;QRS

ECHO

No116

Konno(2004)

38Genotypedrelatives

ofpatientswith

HCM

Japan

,30

Qwave;LV

hypertrophy;ST-segmentd

epression;

T-waveinversion

Genotyping

No45

Nistri(2003)

28Screeningofmilitary

recruits(m

ales

only)

Italy

$17

LVwallthickness;Q

waves;ST-Twaves

ECHO

No2770

Potter

(2010)

39Patientswith

HCM

andhealthycontrols

UK,Sweden,US

Mean=

48.7SD=14.0

RR,PR,P-wave,QR

SandQT

andJT

intervals;P,QR

S,andT-waveam

plitudes;frontalplane

QRS

andT-waveaxes;and

ST-segmentlevels

ECHO

Yes

181

Ryan

(1995)

40Probands

andrelatives

UK,Poland

Mean=

47SD=19

Rwaves;S

waves;Q

waves;ST-Twaves

Genotyping

orclinicaldiagnosis

No506

ECHO

,n=6

Charron(1997)

33Genotypedprobands

andfirst-degreerelatives

France

Range:18–29

MWT

Genotyping

No58

Charron(1998)

34ChildrenofHCMgenotyped

families

France

,18

MWT;intraventricular

septum

/posterior

wall;left

atrium

diam

eter;systolic

anterior

motionofmitral

valve;mid-systolic

aorticclosure;gradient

.30

mmHg;

mitralvalveregurgitation;E/Awaveratio

Genotyping

No35

Charron(2003)

35Genotypedprobands

and


France

Mean=

37.7SD=17.9

MWTinanterior

septum

orposteriorwall;MWT

inposteriorseptum

orfree

wall;systolicanterior

motionofthemitralvalve,redundantleaflets

Genotyping

No109

Fragola(1993)


ofpatientswith

HCM

Italy

Mean=

34SD=19

Increasedinterventricular

septalthickness;posterior

wallthickness

ECHO

No122

Ho(2002)


ofpatientswith

HCM

andhealthycontrols

USA

Mean=

35.6SD=12.6

LVejectionfraction;earlydiastolic

myocardialvelocities

Genotyping

No72



NPV that was near 100%, but PPV,number needed to screen, false-positives, and false-negatives differedsubstantially. At the maximal accuracypoint, the PPVs fell below 1% comparedwith PPVs from 2% to 21% at the max-imal specificity point. The maximalaccuracy point led to fewer false-negatives (16%) and a lower numberneeded to screen to detect 1 case ofHCM (2600). In contrast, the maximalspecificity point led to 40% to 96%false-negative rates and 4000 to 57 000needed to screen to detect 1 case ofHCM. Last, the maximal accuracy pointled to more false-positives per trueHCM case detected (400 vs 4–57) thanthe maximal specificity point. By usingthe often-cited prevalence of 200 per100 00019 (4 times our estimate) re-sulted in similar NPV, a fourfold in-crease in PPV and false-negatives per100 000 screened, and a decrease inthe number needed to screen to detect1 case and number of false-positiveswhen detecting 1 case.

Long QT Syndrome

Phenotypic (ECG-based) prevalencerates of the 3 studies reporting on LQTSranged from1 to 12 per 100 000 (Fig 2),1,27,30 with a summary phenotypicprevalence rate of 7 per 100 000 (95%CI: 0–14) and substantial heterogeneity(I2 = 93%, P, .001). Although 1 study27

did not specify diagnostic criteria, weincluded it because its exclusion in-creased phenotypic prevalence to only9 per 100 000 and this study was basedon a well-established screening pro-gram in Japan. The Kobza study31 wasexcluded because its prevalence rate(550 per 100 000) exceeded often-citedprevalence rates of LQTS (40–50 per100 000).13 When incorporating Kobza,the summary prevalence increasedfivefold to 38 per 100 000 (95% CI: 19–58)and heterogeneity increased (I2 = 99%, P, .001).42–50 The Bayesian sensitivityanalysis giving the Schwartz13 priora low weight (1/2500) resulted in similarTA

BLE2

Continued

Disorder

&Screening

Test

Author

(year)

Sample

Location

Age,y

ScreeningTestCriteria

Definitionof

Reference

(“Gold”)Standard

Verification

Bias

Sample

Size

Ryan

(1995)

40Probands

andrelatives

UK,Poland

Mean=

47SD=19

LVwallthickness

Genotyping

orclinicaldiagnosis

No506

ECG/ECHO

,n=

4Charron(1997)

33Genotypedprobands


France

Range:18–29

SeeaboveCharron1997

ECGandECHO

criteria

Genotyping

No58

Charron(1998)

34ChildrenofHCMgenotypedfamilies

France

,18

SeeaboveCharron1998

ECGandECHO

criteria

Genotyping

No35

Charron(2003)

35Genotypedprobands


France

Mean=

37.7SD=17.9

SeeaboveCharron2003

ECGandECHO

criteria

Genotyping

No109

Ryan

(1995)

40Probands

andrelatives

UK,Poland

Mean=

47SD=19

SeeaboveRyan

1995

ECGandECHO

criteria

Genotyping

orclinicaldiagnosis

No506

LQTS,n=9

ECG,n=

9Benhorin(1990)

42ParticipantsintheInternational

LQTS

Registry

andhealthy

controls

USA,Italy

Range:17–60

STsegm

ent;repolarizationarea;T

wavearea

symmetry

ECG:QTc.

440ms

Yes

352

Kaufman

(2001)


ofpatientswith

LQTS

USA

#13

QTc

Genotyping

No38

Miller

(2001)

44Genotypedproband


USA

Range:2–85

QTc

Genotyping

No23

Moennig(2001)

45Genotypedprobands

andrelatives

Germ

any

Mean=

38Range:

31–50

QTc

Genotyping

No116

Neyroud(1998)

46Genotypedpatientswith

LQTS

andmatched

controls

France

Mean=

31SD=17

QTc

Genotyping

Yes

50

Swan

(1998)

47Genotypedprobands

andrelatives

Finland

Range:7–72

QTc

Genotyping

No73

Vincent(1992)48

Genotypedprobands

andrelatives

Notspecified

Range:1.5–39.0

QTc

Genotyping

No198

Viskin(2010)

49Patientswith

LQTS

andhealthycontrols

Notspecified

Mean=

32SD=15

QTc

LQTS

registry

orgenotyping

Yes

150

Wong(2010)

50Genotypedprobands

andrelatives

UKMean=

26SD=31

QTc

Genotyping

No159

ECG/ECHO

,ECG

combinedwith

ECHO

;LV,leftventricular;MWT,maximalwallthickness;QTc,corrected

QTinterval;RV,rightventricular.

REVIEW ARTICLE



phenotypic prevalence of 7 per 100 000(95% CI: 1–30), whereas giving theSchwartz prior more weight increasedthe phenotypic prevalence to 34 per100 000 (95% CI: 20–54). For detectingLQTS (9 studies), Fig 3 illustrates a set

of HSROC curves for ECG with a sum-mary AUC of 0.92.

To provide clinical context, we used thesummary phenotypic prevalence esti-mate for LQTS and 2 illustrative points(the maximal accuracy point and the

maximal specificity point) on theHSROCcurve for detection of LQTS by using ECG(Table 3). For both points, the NPV wasnear 100%. The PPV was very low(0.04% [1 in 2324]) at the maximal ac-curacy point, but increased slightly atthe maximal specificity point (0.7%[1 in 136]). With maximal accuracy, thenumber needed to screen to detect 1case of LQTS with an ECG was morethan 16 000 with only 14% of those withLQTS missed (false-negatives), butmore than 2000 false-positives per LQTScase detected. With maximal specificity,the number needed to screen to detect 1case increased to 135 000 and 91% ofthose with LQTS would be missed, butthere would be only 135 false-positives per LQTS case detected.By using the often-cited prevalence of50 per 100 00013 (7 times our estimate)resulted in a similar NPV, a sevenfoldincrease in PPV and false-negatives per100 000 screened, and a decrease innumber needed to screen to detect 1case and number of false-positives whendetecting 1 case.

WPW Syndrome

Three studies1,25,30 reported pheno-typic (ECG-based) prevalence rates forWPW ranging from 68 to 222 per 100000 with a summary phenotypic prev-alence rate of 136 per 100 000 (95% CI:55–218) (Fig 2) and substantial het-erogeneity (I2 = 95%, P , .001). Be-cause ECG is considered the referencestandard and no studies reportedestimates of sensitivity and specificityfor any other screening tests to detectWPW, we assumed that its sensitivityand specificity were one and the PPVand NPV estimates were perfect andare not discussed further. By using theoften-cited prevalence estimate of 200per 100 00020 did not substantially alterthe number needed to screen.

DISCUSSION

Byusingpublished literature,we reporton phenotypic (ECG- or ECHO-based)

FIGURE 2Forestplot of phenotypic (ECG- or ECHO-based) prevalenceofHCM, LQTS, andWPW fromreviewedstudies.ES, effect size.

FIGURE 3HSROC and AUC of reviewed studies for HCM and ECG, HCM and ECHO, HCM and ECG/ECHO (ECG combinedwith ECHO), and LQTSandECG. Solid black lines represent summaries of reviewed studiesandother linesrepresent individual studies. The lines describe the relationship between (average) sensitivity and(average) specificity for varying diagnostic thresholds (eg, increasingly stringent ECGor ECHOcriteria inthe 2 top panels, respectively). These lines describe average test performance. It is generally notstraightforward to correspond specific points on the curve to specific cut points for ECG or ECHO. Thex-axis is restricted to the range of the data.



prevalence ratesofHCM, LQTS, andWPWin asymptomatic children and the testcharacteristics of ECG and/or ECHO indetecting thesedisorders. Basedonourprespecified inclusion/exclusion crite-ria and methodology, phenotypic prev-alence estimates demonstrated widevariation across studies and werelower than those in neonates (eg,Schwartz et al13) or studies examininggenotypic prevalence. Consequently,we explored the effects of alternativeprevalence estimates in our results.Although the AUC ranged from 0.88 to0.92, indicating that ECG and/or ECHOare statistically acceptable screeningtests for detecting the most commondisorders that cause SCD, the low phe-notypic prevalence substantially affectedthe predictive value.

Because these disorders have a verylow phenotypic prevalence, choosinga point on the HSROC curve that max-imizes accuracy or maximizes speci-ficity had little impact on NPV (nearly

100% NPV for HCM and LQTS); however,themaximal specificity point resulted inimproved PPV (0.74% [1 in 136] to 21%)at the cost of needing to screen moreindividuals to detect 1 case andmissingmore diseased individuals because ofreduced sensitivity. With maximal ac-curacy, the number needed to screen todetect 1 case fell and fewer cases weremissed, but at the cost of lower PPV(0.04% [1 in 2324] to 0.26% [1 in 390])and more false-positives per casedetected. These findings help defineboundaries of the theoretical utility ofECG and/or ECHO as screening tests forthese disorders, but are difficult tocomprehend in isolation. Unlike “typi-cal” screening programs that valueruling in those who may have the dis-ease (ie, sensitivity), these illustrativepoints demonstrate that when pheno-typic prevalence is low, prioritizingspecificity over sensitivity can improvePPV while not affecting the NPV (similarto HIV screening51).

We performed calculations to under-stand how these estimates might applyto population-based ECG screening.First, assuming independence, thecombined prevalence estimate of HCM,LQTS, and WPW from our meta-analysisis 188 per 100 000. When maximizingaccuracy, the NPV approaches 100%,indicating a low false-reassurancerate. However, the PPV of using anECG to screen for any of the 3 disordersis 1%, indicating a high false-alarmrate (99% of children with a positiveECG would not have any of the dis-orders). Conversely, when maximizingspecificity, the NPV still approaches100% (false-reassurance rate remainsnear 0%), but the PPV is 41% (false-alarm rate decreases to 59%). Asensitivity analysis using often-citedprevalence rates (40–50 per 100 000for LQTS,13 200 per 100 000 for HCM,19

100–200 per 100 000 for WPW20) showeda more favorable outlook for screening(higher PPV, fewer need to screen to

TABLE 3 Implications of Screening ECG and/or ECHO on PPV, NPV, Number Needed to Screen, False-Positives, and False-Negatives for Illustrative Pointson the HSROC Curve

Prevalence per100 000

Sensitivity Specificity PPV NPV Number Neededto Screen toDetect 1 Case

Number ofFalse-Positives WhenDetecting 1 Case

Number ofFalse-Negatives per100 000 Screened

Illustrative point where sensitivity and specificity are equally weighted (maximal accuracy) with prevalence from meta-analysisHCM & ECG 45 0.847 0.848 0.0025 (1/400) 0.9999 2624 399 7HCM & ECHO 45 0.851 0.851 0.0026 (1/390) 0.9999 2611 389 7HCM & ECG/ECHO 45 0.837 0.837 0.0023 (1/434) 0.9999 2655 433 7LQTS & ECG 7 0.861 0.860 0.0004 (1/2324) 0.9999 16 592 2323 1WPW & ECG 136 1.000 1.000 1.0000 1.0000 735 0 0Illustrative point where specificity is given more weight (maximal specificity) with prevalence from meta-analysisHCM & ECG 45 0.039 0.999 0.0173 (1/58) 0.9996 56 980 57 43HCM & ECHO 45 0.607 0.999 0.2146 (1/5) 0.9998 3661 4 18HCM & ECG/ECHO 45 0.514 0.999 0.1879 (1/5) 0.9998 4323 4 22LQTS & ECG 7 0.106 0.999 0.0074 (1/136) 0.9999 134 771 135 6WPW & ECG 136 1.000 1.000 1.0000 1.0000 735 0 0Illustrative point where sensitivity and specificity are equally weighted (maximal accuracy) with oft-cited prevalenceHCM & ECG 20019 0.847 0.848 0.0110 (1/91) 0.9996 590 90 31HCM & ECHO 20019 0.851 0.851 0.0113 (1/88) 0.9996 588 87 30HCM & ECG/ECHO 20019 0.837 0.837 0.0102 (1/98) 0.9996 597 97 33LQTS & ECG 5013 0.861 0.860 0.0031 (1/326) 0.9999 2323 325 7WPW & ECG 20020 1.000 1.000 1.0000 1.0000 500 0 0Illustrative point where specificity is given more weight (maximal specificity) with oft-cited prevalenceHCM & ECG 20019 0.039 0.999 0.0725 (1/14) 0.9981 12 821 13 192HCM & ECHO 20019 0.607 0.999 0.5488(1/2) 0.9992 824 1 79HCM & ECG/ECHO 20019 0.514 0.999 0.5074 (1/2) 0.9990 973 1 97LQTS & ECG 5013 0.106 0.999 0.0504 (1/20) 0.9996 18 868 19 45WPW & ECG 20020 1.000 1.000 1.0000 1.0000 500 0 0

ECG/ECHO, ECG combined with ECHO.

REVIEW ARTICLE



detect 1 case, fewer false-positiveswhen detecting 1 case, but more false-negatives per 100 000 screened).

Although these results suggest that ECGmay be considered for mass screeningfrom a statistical perspective and fromthe US Preventive Services Task Forcecriteria,52 it does not address othercomponents of screening programs,including changes in mortality, mor-bidity, cost, quality of life, and func-tioning that need to be weighed.Because of the very low phenotypicprevalence and inherent inaccuracy innearly all medical tests (including pe-diatric cardiologists reviewing ECGs53),screening for rare disorders will leadto many false-positive tests that triggeradditional diagnostic evaluations and,possibly, unnecessary therapies andphysical activity restrictions. In addi-tion, false-positives may lead to un-warranted child and parent anxiety;previous work suggests this anxietymay not dissipate immediately follow-ing a cardiac evaluation and may in-fluence lifelong lifestyle decisions.54

Concern has been raised about in-creased rates of diagnosed heart dis-ease where diagnosis may not behelpful (resulting in diagnosis of car-diac nondisease and overtreatment55).Some children may ultimately be di-agnosed by other means (eg, familyhistory, emergence of sublethal symp-toms, diagnostic testing for unrelatedindications) even in the absence ofmass screening, and earlier diagnosisof the disorder may not provide sur-vival benefit. Among those childrendetected by screening, the safety, effi-cacy, and acceptability of specifictherapies and recommendations forprophylaxis of SCD in asymptomaticchildren is sometimes uncertain and

often based on expert consensus asopposed to clinical evidence.6

In addition, families need to be awarethat a negative ECG does not definitivelyrule out risk for SCD. Other rare cardiacdisorders cause SCD, such as anomalousorigin of the coronary artery, arrhyth-mogenic right ventricular cardiomy-opathy, catecholaminergic ventriculartachycardia, and Brugada syndrome.With the exception of Brugada syndrome,which manifests in late adolescence,these disorders are not typically di-agnosed by ECG. Also, given that clinicaland ECG findings may not manifest inHCM until adolescence, programs thatscreen young children may miss thosegenetically predisposed to developingHCM, which may necessitate repeatECGs during adolescence.

This study has several limitations. First,our search was restricted to literaturecataloged by Medline. Medline indexesmost biomedical articles, making itunlikely that we omitted importantfindings. Second, our estimates mayreflect publication bias, as it is con-ceivable that “unsuccessful” studies ofdiagnostic or detection interventionsmay not have been published. Thisphenomenon, if it took place, wouldinflate our test accuracy estimates.Third, heterogeneity existed betweenstudies. Populations varied by age andstudies varied in their screening ap-proach and their diagnostic criteria.Fourth, our calculations omitted a tar-geted history and physical. Althoughwe included medical history, physicalexamination, and family history in oursearch, we found insufficient data forfurther analyses. Published studiessuggest that history and physical ex-amination have low sensitivity,25 lowPPV,56 and limited value from a health

economics perspective.3,4 Fifth, we de-fined phenotypic prevalence basedon results from ECG or ECHO (not ge-notyping) because these were thescreening options considered in ouranalysis and because genotyping hasonly recently become available and isstill evolving. In estimating the sensi-tivity and specificity of ECG and/orECHO, however, we allowed genotypi-cally identified cohorts because ge-netic testing will likely play a moreprominent role in screening and diag-nosing disorders that cause SCD. Byusing genotyping as the referencestandard allows us to incorporatesome variability in penetrance (leadingto false-positives), which will likely beimportant for screening test inter-pretation. Finally, this study is limitedby considering only 3 disorders, butthese are the 3 most common dis-orders detectable by ECG and/or ECHO.

Despite these limitations, this studyprovides an important starting pointforevaluatingSCDscreeningprograms.Screening programs may be gainingpopularity because of availability biasin risk perception (ie, recent publicizedevents result in the overestimatedlikelihood of a similar event occurring).Given our results on the low phenotypic(ECG- or ECHO-based) prevalence andthevariation in false-positiveratebasedon different sensitivities and specific-ities, further cost- or comparative-effectiveness analyses will be necessaryto determine whether screening pro-grams to detect SCD in asymptomaticchildren should be promoted as publichealth policy.

ACKNOWLEDGMENTWe thank Tully S. Saunders for his assis-tance in manuscript preparation.

REFERENCES

1. Corrado D, Basso C, Pavei A, Michieli P,Schiavon M, Thiene G. Trends in sudden

cardiovascular death in young competitiveathletes after implementation of a pre-

participation screening program. JAMA.2006;296(13):1593–1601



2. Berger S, Utech L, Fran Hazinski M. Suddendeath in children and adolescents. PediatrClin North Am. 2004;51(6):1653–1677, ix–x

3. Wheeler MT, Heidenreich PA, Froelicher VF,Hlatky MA, Ashley EA. Cost-effectiveness ofpreparticipation screening for preventionof sudden cardiac death in young athletes.Ann Intern Med. 2010;152(5):276–286

4. Denchev P, Kaltman JR, Schoenbaum M,Vitiello B. Modeled economic evaluation ofalternative strategies to reduce suddencardiac death among children treated forattention deficit/hyperactivity disorder. Cir-culation. 2010;121(11):1329–1337

5. Tanaka Y, Yoshinaga M, Anan R, et al. Use-fulness and cost effectiveness of cardio-vascular screening of young adolescents.Med Sci Sports Exerc. 2006;38(1):2–6

6. Kaltman JR, Thompson PD, Lantos J, et al.Screening for sudden cardiac death in theyoung: report from a National Heart, Lung,and Blood Institute working group. Circu-lation. 2011;123(17):1911–1918

7. Vetter VL, Elia J, Erickson C, et al; AmericanHeart Association Council on Cardiovascu-lar Disease in the Young Congenital Car-diac Defects Committee; American HeartAssociation Council on Cardiovascular Nurs-ing. Cardiovascular monitoring of childrenand adolescents with heart disease re-ceiving medications for attention deficit/hyperactivity disorder [corrected]: a scien-tific statement from the American HeartAssociation Council on Cardiovascular Dis-ease in the Young Congenital Cardiac DefectsCommittee and the Council on Cardiovascu-lar Nursing. Circulation. 2008;117(18):2407–2423

8. Pastor PN, Reuben CA. Diagnosed attentiondeficit hyperactivity disorder and learningdisability: United States, 2004 -2006. Na-tional Center for Health Statistics. VitalHealth Stat. 2008; 10(237)

9. Perrin JM, Friedman RA, Knilans TK; BlackBox Working Group; Section on Cardiologyand Cardiac Surgery. Cardiovascular mon-itoring and stimulant drugs for attention-deficit/hyperactivity disorder. Pediatrics.2008;122(2):451–453

10. Maron BJ, Thompson PD, Ackerman MJ,et al; American Heart Association Councilon Nutrition, Physical Activity, and Metabo-lism. Recommendations and considerationsrelated to preparticipation screening forcardiovascular abnormalities in competitiveathletes: 2007 update: a scientific statementfrom the American Heart Association Coun-cil on Nutrition, Physical Activity, and Me-tabolism: endorsed by the American Collegeof Cardiology Foundation. Circulation. 2007;115(12):1643–1655

11. Hennekens CH, Buring JE, Mayrent SL. Epi-demiology in Medicine. 1st ed. Philadelphia,PA: Lippincott Williams & Wilkins; 1987

12. Berger S, Dhala A, Friedberg DZ. Suddencardiac death in infants, children, and ado-lescents. Pediatr Clin North Am. 1999;46(2):221–234

13. Schwartz PJ, Stramba-Badiale M, Crotti L, et al.Prevalence of the congenital long-QT syn-drome. Circulation. 2009;120(18):1761–1767

14. DerSimonian R, Laird N. Meta-analysis inclinical trials. Control Clin Trials. 1986;7(3):177–188

15. Higgins JP, Thompson SG. Quantifying het-erogeneity in a meta-analysis. Stat Med.2002;21(11):1539–1558

16. Rutter CM, Gatsonis CA. A hierarchical re-gression approach to meta-analysis of di-agnostic test accuracy evaluations. StatMed. 2001;20(19):2865–2884

17. Gatsonis C, Paliwal P. Meta-analysis of di-agnostic and screening test accuracy eval-uations: methodologic primer. AJR Am JRoentgenol. 2006;187(2):271–281

18. Dukic V, Gatsonis C. Meta-analysis of di-agnostic test accuracy assessment studieswith varying number of thresholds. Bio-metrics. 2003;59(4):936–946

19. Maron BJ, Gardin JM, Flack JM, Gidding SS,Kurosaki TT, Bild DE. Prevalence of hypertro-phic cardiomyopathy in a general populationof young adults. Echocardiographic analysisof 4111 subjects in the CARDIA Study. Cor-onary Artery Risk Development in (Young)Adults. Circulation. 1995;92(4):785–789

20. Triedman JK. Management of asymptomaticWolff-Parkinson-White syndrome. Heart. 2009;95(19):1628–1634

21. Higgins J, Green S, eds. Cochrane Handbookfor Systematic Reviews of InterventionsVersion 5.0.2 [updated September 2009]. TheCochrane Collaboration. West Sussex, England:John Wiley & Sons Ltd; 2009

22. Joseph L, Gyorkos TW, Coupal L. Bayesianestimation of disease prevalence and theparameters of diagnostic tests in the ab-sence of a gold standard. Am J Epidemiol.1995;141(3):263–272

23. Arola A, Jokinen E, Ruuskanen O, et al. Ep-idemiology of idiopathic cardiomyopathiesin children and adolescents. A nationwidestudy in Finland. Am J Epidemiol. 1997;146(5):385–393

24. Colivicchi F, Ammirati F, Santini M. Epide-miology and prognostic implications ofsyncope in young competing athletes. EurHeart J. 2004;25(19):1749–1753

25. Corrado D, Basso C, Schiavon M, Thiene G.Screening for hypertrophic cardiomyopa-thy in young athletes. N Engl J Med. 1998;339(6):364–369

26. Maron BJ, Mathenge R, Casey SA, Poliac LC,Longe TF. Clinical profile of hypertrophiccardiomyopathy identified de novo in ruralcommunities. J Am Coll Cardiol. 1999;33(6):1590–1595

27. Niimura I, Maki T. Sudden cardiac death inchildhood. Jpn Circ J. 1989;53(12):1571–1580

28. Nistri S, Thiene G, Basso C, Corrado D,Vitolo A, Maron BJ. Screening for hyper-trophic cardiomyopathy in a young malemilitary population. Am J Cardiol. 2003;91(8):1021–1023, A8

29. Zou Y, Song L, Wang Z, et al. Prevalence ofidiopathic hypertrophic cardiomyopathy inChina: a population-based echocardiographicanalysis of 8080 adults. Am J Med. 2004;116(1):14–18

30. Chiu SN, Wang JK, Wu MH, et al; Taipei Pe-diatric Cardiology Working Group. Cardiacconduction disturbance detected in a pediat-ric population. J Pediatr. 2008;152(1):85–89

31. Kobza R, Roos M, Niggli B, et al. Prevalenceof long and short QT in a young populationof 41,767 predominantly male Swiss con-scripts. Heart Rhythm. 2009;6(5):652–657

32. Autore C, Fragola PV, Picelli A, et al. Equivo-cal and borderline myocardial hypertrophyin relatives of patients with hypertrophiccardiomyopathy: possible implications ingenetics of the disease. Cardiology. 1988;75(5):348–356

33. Charron P, Dubourg O, Desnos M, et al. Di-agnostic value of electrocardiography andechocardiography for familial hypertrophiccardiomyopathy in a genotyped adult pop-ulation. Circulation. 1997;96(1):214–219

34. Charron P, Dubourg O, Desnos M, et al. Di-agnostic value of electrocardiography andechocardiography for familial hypertrophiccardiomyopathy in genotyped children. EurHeart J. 1998;19(9):1377–1382

35. Charron P, Forissier JF, Amara ME, et al.Accuracy of European diagnostic criteriafor familial hypertrophic cardiomyopathyin a genotyped population. Int J Cardiol.2003;90(1):33–38, discussion 38–40

36. Dipchand AI, McCrindle BW, Gow RM, FreedomRM, Hamilton RM. Accuracy of surface elec-trocardiograms for differentiating childrenwith hypertrophic cardiomyopathy fromnormal children. Am J Cardiol. 1999;83(4):628–630, A10

37. Fragola PV, Borzi M, Cannata D. The spec-trum of echocardiographic and electrocardi-ographic abnormalities in nonaffectedrelatives of patients with hypertrophic car-diomyopathy: a transverse and longitudinalstudy. Cardiology. 1993;83(5-6):289–297

38. Konno T, Shimizu M, Ino H, et al. Diagnosticvalue of abnormal Q waves for identificationof preclinical carriers of hypertrophic

REVIEW ARTICLE



cardiomyopathy based on a moleculargenetic diagnosis. Eur Heart J. 2004;25(3):246–251

39. Potter SL, Holmqvist F, Platonov PG, et al.Detection of hypertrophic cardiomyopathyis improved when using advanced ratherthan strictly conventional 12-lead electro-cardiogram. J Electrocardiol. 2010;43(6):713–718

40. Ryan MP, Cleland JG, French JA, et al. Thestandard electrocardiogram as a screen-ing test for hypertrophic cardiomyopathy.Am J Cardiol. 1995;76(10):689–694

41. Ho CY, Sweitzer NK, McDonough B, et al.Assessment of diastolic function withDoppler tissue imaging to predict genotypein preclinical hypertrophic cardiomyopa-thy. Circulation. 2002;105(25):2992–2997

42. Benhorin J, Merri M, Alberti M, et al. LongQT syndrome. New electrocardiographiccharacteristics. Circulation. 1990;82(2):521–527

43. Kaufman ES, Priori SG, Napolitano C, et al.Electrocardiographic prediction of abnor-mal genotype in congenital long QT syn-drome: experience in 101 related familymembers. J Cardiovasc Electrophysiol. 2001;12(4):455–461

44. Miller MD, Porter CB , Ackerman MJ. Diagnos-tic accuracy of screening electrocardiograms

in long QT syndrome I. Pediatrics. 2001;108(1):8–12

45. Moennig G, Schulze-Bahr E, Wedekind H,et al. Clinical value of electrocardiographicparameters in genotyped individuals withfamilial long QT syndrome. Pacing ClinElectrophysiol. 2001;24(4 pt 1):406–415

46. Neyroud N, Maison-Blanche P, Denjoy I, et al.Diagnostic performance of QT intervalvariables from 24-h electrocardiography inthe long QT syndrome. Eur Heart J. 1998;19(1):158–165

47. Swan H, Saarinen K, Kontula K, Toivonen L,Viitasalo M. Evaluation of QT interval du-ration and dispersion and proposed clin-ical criteria in diagnosis of long QTsyndrome in patients with a geneticallyuniform type of LQT1. J Am Coll Cardiol.1998;32(2):486–491

48. Vincent GM, Timothy KW, Leppert M, KeatingM. The spectrum of symptoms and QTintervals in carriers of the gene for thelong-QT syndrome. N Engl J Med. 1992;327(12):846–852

49. Viskin S, Postema PG, Bhuiyan ZA, et al.The response of the QT interval to thebrief tachycardia provoked by standing:a bedside test for diagnosing long QTsyndrome. J Am Coll Cardiol. 2010;55(18):1955–1961

50. Wong JA, Gula LJ, Klein GJ, Yee R, SkanesAC, Krahn AD. Utility of treadmill testing inidentification and genotype prediction inlong-QT syndrome. Circ Arrhythm Electro-physiol. 2010;3(2):120–125

51. Meyer KB, Pauker SG. Screening for HIV:can we afford the false positive rate?N Engl J Med. 1987;317(4):238–241

52. Harris RP, Helfand M, Woolf SH, et al;Methods Work Group, Third US PreventiveServices Task Force. Current methods ofthe US Preventive Services Task Force:a review of the process. Am J Prev Med.2001;20(suppl 3):21–35

53. Hill AC, Miyake CY, Grady S, Dubin AM. Ac-curacy of interpretation of preparticipationscreening electrocardiograms. J Pediatr.2011;159(5):783–788

54. Young PC. The morbidity of cardiac non-disease revisited. Is there lingering con-cern associated with an innocent murmur?Am J Dis Child. 1993;147(9):975–977

55. Bergman AB, Stamm SJ. The morbidity ofcardiac nondisease in schoolchildren.N Engl J Med. 1967;276(18):1008–1013

56. Maron BJ, Shirani J, Poliac LC, Mathenge R,Roberts WC, Mueller FO. Sudden death inyoung competitive athletes. Clinical, de-mographic, and pathological profiles. JAMA.1996;276(3):199–204



DOI: 10.1542/peds.2011-0643; originally published online March 5, 2012;Pediatrics

Laurel K. LeslieIp, Jane W. Newburger, Susan K. Parsons, Thomas A. Trikalinos, John B. Wong and Angie Mae Rodday, John K. Triedman, Mark E. Alexander, Joshua T. Cohen, Stanley

in Asymptomatic Children: A Meta-analysisElectrocardiogram Screening for Disorders That Cause Sudden Cardiac Death

ServicesUpdated Information &

/peds.2011-0643http://pediatrics.aappublications.org/content/early/2012/02/29including high resolution figures, can be found at:

Supplementary Material

9/peds.2011-0643.DCSupplemental.htmlhttp://pediatrics.aappublications.org/content/suppl/2012/02/2Supplementary material can be found at:

Permissions & Licensing

tmlhttp://pediatrics.aappublications.org/site/misc/Permissions.xhtables) or in its entirety can be found online at: Information about reproducing this article in parts (figures,

Reprints http://pediatrics.aappublications.org/site/misc/reprints.xhtml

Information about ordering reprints can be found online:

rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.Grove Village, Illinois, 60007. Copyright © 2012 by the American Academy of Pediatrics. All and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elkpublication, it has been published continuously since 1948. PEDIATRICS is owned, published, PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly

http://pediatrics.aappublications.org/content/early/2012/02/29/peds.2011-0643http://pediatrics.aappublications.org/content/suppl/2012/02/29/peds.2011-0643.DCSupplemental.htmlhttp://pediatrics.aappublications.org/site/misc/Permissions.xhtmlhttp://pediatrics.aappublications.org/site/misc/reprints.xhtmlhttp://pediatrics.aappublications.org/

Documents

New peds.2011-0643 - Parent Heart Watch | Sudden Cardiac Arrest … · 2012. 3. 5. · Cause Sudden Cardiac Death in Asymptomatic Children: A Meta-analysis abstract BACKGROUNDANDOBJECTIVES:Pediatricsuddencardiacdeath(SCD)