New insight into QT dispersion

Posters From Day 3 NEW INSIGHT INTO QT DISPERSION PW Macfarlane, SC McLaughlin, JC Rodger* University Dept. of Medical Cardiology, Glasgow Royal Infirmary and *Dept. of Medicine, Monklands Hospital, Airdrie, Scotland There is currently controversy over the clinical value of measuring QT dispersion given problems associated with accuracy of measurement as well as interpretation. The present computer based study presents new data derived from a series of databases of ECGs analysed by the Glasgow Progrmn. Using ECGs from 1.501 adult nonnals, it was found that there w,-~ no significant difference in QT dispersion between m',d~ and females (24.67 ± 8.2 vs. 24.35 ± 8.2ms). The upper lhnit of normal in each case was 44 milliseconds. There was no gradient with respect to age. In a subset or the conventional 12-lead ECG. viz. 6 pmeeordial leads plus leads I and II, the QT dispersion was reduced to 20.76 + 7.Sins. h~ a group of 1.785 healthy children aged from birth to adolescence. QT dispersion varied little with age with a mean of 24.52 _+8.7ms. and an upper normal limit of ~ m s . as in adults. It is often argued that QT dispersion is in la.rg¢ measure due to the projection of cardiac electrical activity on to different lead axes. To study this problem. 1,220 normal and abnonnal r::CGs from the CSE Database were used. Both the conventional 12-lead ECG and the 3-orthogonal lead XYZ lead ECG were available from each patient. From the latter, it was possible to derive the conventional 12-1ends as lineal combinations of leads X,Y and Z. The mean QT dispersion for the conventional and derived 12-1ead ECG was 29.1 --. 10.2ms and 27.5 + t0.Sms respectively. On the other trend, the tne~ QT dispersion in the 3-lead ECG was only 17.1 _+20.0ms. The repeatability of the automated technique was shown in the same 1.220 ECGs to be excellent by using a splitting technique which created 2 ECGs from each odginal. The mean difference in QT dispersion between corresponding paks of ECGs was 0.28 _.+9.7ms. These new data provide hfformatioa on QT dispersion throughout the age spectrum and show that it is not related solely to dispersion of repoh"aisation but to the number of leads used in irs measurement.

Q T Interval Measurement of the Ambulatory ECG Robert E. Bruce, Atul P. Shah

Medicomp Inc., Melbourne, Florida, USA The EPICARDIA TM Halter System measures the QT interval on two channels of ECG for eve~ 30 seconds in an ambulatory ECG procedure. The measurements are made on a representative beat that is constructed by averaging selected beats dunng each 30 second pealed of the ambulatory procedure. The ECG of the representative beat starts 320 ms prior to the RWave apex of the PQRST complex and extends for 1000 ms. The selected beats are sampled at 250 sps and then averaged to produce the representative beat. The algorithm that determines the T-Wave offset of each of the representative bests maps out each T-Wave with line segments that are then amalgamated into a section of a polyhedron. The polyhedron is then successively refined until the T-Wave offset is approximated. Then a procedure that utilizes a tangent line segment is used to refine the offset of T-Wave, Those beats that cannot be processed by the T-Wave offset algorithm are marked as members of a special "rejected set" and are color coded accordingly. The QT values of the =rejected set" beats are ignored dudng any graphing or reporting. EPICARDIA allows the user to graph or scatter plot any of the QT data versus other ECG measurements provided by the system. This paper will examine the algorithm and its behavior in the ambulatory environment. Future work will involve testing of 15 and 20 second representative beats, exploring new techniques for deriving the representative beat, and validation of the accuracy of the algorithm.

95

Proan'hythmic effectof classIIl antiaz~ythmic agent inrulationto itseffects on both Q T intervaland Q T dispersion No,.hike Ohno, Hirokazu Saitoh, Takuya One, Atsunobu Nomura, Yos1~noti Kobayashi, Hi~tsug'u Atar~hi, Takao Kntoh, Hir~kazu Hayak~wa F i ~ Department of InternalMedicine, Nippon Medical School, Tokyo, Japan [P~rpose] Although Q T dispersion(QTd) was thought to be usefulpredictorof lethalvenmcul~r ~hy~hmia, the relationbetween preat~ythmic effectof class Ill a~iatrbyth~ic drug and QTd was not examined yet. [Method] Twelve patientswith more than 3000 ventricularpremature contraetions(VPCs)/day, ~.nd treated with M S 551(450mg or 5.00rag)wet'eincluded. Seven patientswith more than 50% t-eductionof VPCs by MS551 were classifiedintoeffectivegroup (grpup E) and 5 patientswho revealed lethalventticul~ atrhythmia with MS551 into side effectg~up(g~ap S). Q T and R R in all 12 leads were measured by hand. QTcwas than automatically calculated asing Bazett'sformula. QTc di~-p~rsion(QTcd)was dermed as differencebetween the mlnlm~,~I2~and m~xim~n QTc vahi~ among 12 leads. [Result] (1) Afr~ MS551, the lead which showed minimum Qtc differedfrom that of controlin 8 patientsout of 12. After MS551, the lead which showed mavJmum QTc also di~er~l from thatof cant~l in 8 out of 12. Consequently',thet~ were no sigx~licantchanges inmaxi~u~ QTc, rain/mum QTc, or QTcd with M S 5 5 I. (2) There were no significant differencein age, RR, minimum QTc, maximum QTc or (~Tl'c.dbetween group E and group S. In group S, QTcd reduced in 2 patientsand inc,-easedin 3 patients with MS551, In the patientwith To~ades de poimes(Tdp), QTed reduced by 60 msec, but QTc markedly inct-easedby 150 msec in the lead of mln{rnu.m QTc ill cOnLroL [Conclusion] MS551 did not make significantch.~e in QTcd probably due to change of the leads of both minimum QTc and maximum QTc. In the patientwith T6p, despireof r~uction of QTcd, marked pmlongaaon of ( ~ c ~migh; be responsiblefor Tdp. In conclusion, itwas suggestedtha~the cha~e of QTd by classIllantia~-hythm.icdrug did nut predictitsp~ar~ythn~c effect.

The Bectrocerdiologica/Mechanisms of Quinidbte-induced Dispersion of the Ventticuhr RepolarisatJon Duration on the Surface ECG : A multivadato

analysis. S. Fareh, P. Am~ud, J. Fayn, P. Nony, J.L Run, M. CucheraL J.P. 8oiasel, P. Rubel. INSERMU121 and Cardiovascular Hospital,Lyon, France. The dispersionof th9 yentricularrepotadzalion(DVR) is assessedas a new marker of the venliScular arrhy~rnogenic substrata but ~ ~ gives rise to unsolved fundamental questions. Our aim was to f~l the characteris~asof the spaljo-temperalstructure of the T loop that could pesdl:~yexplain the dispersion (QTmax-QTrrin) of the QTapex and QTend inte~s (QTdapex, QTdend), We perfom',~ a randomizedplacebo-controlledldal in 10 healey subjects,with a single oral dose of quinidtho (Q : 330 rag) and placebo (PL), Orthogonat3J,ead ECG acquisilJonand analysis (n=70 per period of trealrnent)were performedusing the LYON and CAVIAR programswhich preciselydascd0e~ spatio-temporalT loop structure. QRS onset, T apex, T offset were visually detwmined using the PC CardiologicClinical Work~tton interac~vouser interface.The correlationsbetween the spalio-tamboralT loop parameters and the two inffces of DVR have been perfonT~ with mullip~e stepwise regression analyds. Q versus PL induced a sign'nlcant increase of QTdapex (!:)=0.007) and QTdeed(,o=0.022)wh~ no QR$ modifications were found. Under PL, QTdapex was correlated (multiple r = 0.87, P<0.0001) with the maximal QRS amplitude and with the shape of the T loop. QTdendwas not correlatedwith any of the studiedparametersprobablybecauseof its small magnitude(8=4 ms). Under Q, QTdapexwas correlated(mu]l~pler : 0.82, P<0.0001) with the drug-induced alterationsof the globalT kx~ morphologyand planarity.QTdend was correlated(mul~pte r = 0.74, P<0.0C01)with the changesin morphologyand plana'ityof the terminalpart of the T bop. No con?Jattonwas found with heartrate or T waveduration.We have dearly denons~atedthat the incre,as~ DVR induced by Q is related to the spatial charactedslicschanges of the T loop whi~ dependson the myocardialrepolarizalionprocessvia the cardiacelec-~ical