Tachycardia Detection in ImplantableAntitachycardia Devices

JANICE JENKINS,* THOMAS BUMP, FRAN MUNKENBECK, JEFFREYBROWN, and ROBERT ARZBAECHERFrom the Illinois Institute of Technology and University of Chicago, Chicago, Illinois, andtJniversity of Michigan, Ann Arbor. Michigan USA

'The

La detection des tachycardies par Ies stimuiateurs antitachycardie implantables. Nous avonsun algorithme permettant Videntification des tachycardies par un microcomputer utilisable dans unpacemaker antiiachycardie. Celui ci distingue Ies tachycardies auriculaires, ventriculaires et de reentr^edu noeud AV et ne reconnait pas Ies fachycordies sinusaies. De meme a ei6 developpe une nouvellemethode permettant au pacemaker d I'aide d'un extrastimuJus auriculaire de distinguer Ies tachycardiesparoxystiques 1/1 des tachycardies sinusaies. Celle ci est actuellement testee au laboratoire d'^Iectro-physiologie. Les resullats preiiminaires indiquent que ces deux tachycardies peuvent etre distingueesde /agon fiable par un extrastimulus delivre avec un couplage correct dans J'oreiilette droite.

JENKINS, J.. BUMP, T., MUNKENBECK, F., ETAL.: Tachycardia detection in implantable antitachycardiadevices. We have developed a microcomputer algorithm for tachycardia identification suitable for usein an implanted antitachycardia pacemaker. The algorithm distinguishes atrial, ventricuJar, and AVreentranf tachycardia, while ignoring high rates due to sinus tachycardia. A new method whereby ihepacemaker delivers an atrial extrastimulus has been developed/or distinguishing paroxysmal 1:1 tachy-cardia/rom sinus tachycardia, and has undergone preliminary testing in the eleclrophysiology laboratory.Preliminary results indicate sinus tachycardia can be reliably distinguished from paroxysmal 1:1 tachy-cardia by a properly timed extrastimulus delivered to the right atrium.

tachycardias, deieciion algorithms

Introduction

The development of antitachycardia pace-makers designed for con version of malignant tachy-cardias has created a demand for more accuratetachycardia recognition schemes than those pres-ently available. Fortunately, rate microelectronictechnology now provides the mechanism for in-corporating a microprocessor within the pace-maker, and thus computer pattern recognitionschemes can be included within the device. Thispaper describes an algorithm we have developedthat is suitable for use within such a microproc-essor-based antitachycardia pacemaker. The al-gorithm distinguishes atrial, ventricular, and AV

Address for reprints: Janice Jenkins. Ph.D., Department of Elec-trical Engineering and Computer Science, The University ofMichigan, Ann Arbor, Ml 48109 USA

reentrant tachycardia and uses a criterion of rapidonset to distinguish pace-terminable tachycardiasfrom sinus tachycardia.

Antitachycardia pacemakers that are at presentundergoing clinical testing detect tachycardia bysensing a high rate in the chamber to be paced.The rate must exceed a preset threshold and thetachycardia must persist for a preselected numberof heats. Such simple criteria may result in inac-curate identification of a tachycardia and result inundesirable pacing. Other schemes^"^ have beenadvanced for using criteria other than rate but mosthave not yet been tested.

There is a particular problem with sinus tachy-cardia which is often indistinguishable from pace-terminable 1:1 tachycardias on the basis of ratealone, and these onset measures have been pro-posed. It is commonly held that sinus tachycardiasaccelerate gradually while pace-terminable tachy-cardias are sudden in onset. However, exceptions

PACE, Vol. 7 November-December 1984, Part II 1273

JENKINS, ET AL.

do occur sincn sinus tachycardia resulting fromfright, excitement, sudden effort, or postural changecan occasionally develop very quickly, while par-oxysmal tachycardias may begin quite hesitantly,with couplets and triplets separated hy long pauses.Also, premature beats and their compensatorypauses can occur just before or just after the be-ginning of a tachycardia. Thus a successful algo-rithm for assessing suddenness of onset must ex-amine more than just the two intervals which definethe transition from below the tachycardia ratethreshold to above it.

One newly developed aigorithm has decisionrules for measuring the suddenness of onset forpurposes of separating sinus tachycardia from par-oxysmal tachycardia. This algorithm, called IN-TERTACH was developed for tiie antitachycardiapacemaker* of the same name. It compares the firstinterval of the tachycardia run (the first to fallbelow the tachycardia threshold) to the first pre-vious (pre-tach) cycle length, two cycle lengthsjust preceding the tachycardia. If the first intervalin the tachycardia is shorter than either by a presetamount (and if the third interval preceding thetachycardia is not short) the tachycardia is parox-ysmal; otherwise, it is sinus. The algorithm per-forms reasonably well in discriminating sinus frompace-terminable tachycardias. Nevertheless, thereare paroxysmal tachycardias that start more slowlythan sudden sinus tachycardias and these cannotalways he separated by onset criteria alone. Fischeret al."* examined 50 spontaneous ventricular tachy-cardias and 50 sinus tachycardias and found thatwhile Ihe rate of tachycardia development is sig-nificantly faster with ventricular tachycardia, someoverlap still exists for individual subjects.

In addressing the ahove problem, we have de-veloped our own onset criterion and have alsoexplored an extrastimulus test, performed by thepacemaker itself, to identify sinus tachycardia, Insinus tachycardia, the ventricular rhythm is per-turhed by the atrial extrastimulus. In the otherrhythms shown, the ventricular response is notperturbed by the extrastimuli.

Methods

Our computer algorithm for tachycardia recog-nition is shown in Figure 1. Full details have been

EXAMINE SEQUENCE

OF

RECENT INTERVALS

ATRIAL TACH,

FLUTTER. OR

FIBRILLATION

.A»VI OF

A S V EVENTS"

AN TACH.

'SUDDEN.ONSET

SINUS TACH

V V»A

YES

V. TACH

OR

V. FIB

PAROX 1:1

TACH

'Intermedics, Freeport, Texas, USA

Figure 1. Flow chart of the main diagnostic program.Dual-chamber .sensing ullows separation of atrial andventricular tachycardia, and the analysis of onset dis-tinguishes tachycardia.

previously published.^ Following the detection ofa sustained run of short intervals in either cham-ber, the program determines whether the eventsin the tachycardia are predominantly atrial or ven-tricular. If they are atrial, the left path of the flowchart is followed and atrial tachycardia, flutter, orfibrillation is diagnosed, depending on rate. If eventsare predominantly ventricular, the right path istaken and ventricuiar tachycardia or fibrillation isdiagnosed. If the numher of atrial and ventricularevents is roughly equal, 1:1 tachycardia is diag-nosed, and a measure of suddenness of onset ismade in order to distinguish ordinary sinus tachy-cardia from other 1:1 tachycardias; atrial, AV reen-trant, AV nodal reentrant, and ventricular with ret-rograde conduction to the atria. In this test, thefirst four beats that satisfy the high rate conditionthat invoke the program are compared to the fourbeats preceding them. If they have an interheatinterval 25% shorter than the average of thepreceding four, the tachycardia is paroxysmaland presumably pace-terminable; otherwise, thetachycardia is sinus tachycardia.

For evaluation of the algorithm, we selected 22electrically induced tachycardias from patientswhose electrograms were tape recorded during in-tracardiac electrophysiologic study. All record-ings contain at least three surface and four hipolarcatheter leads, including electrograms from the rightatrium, right ventricle, bundle of His, and coro-nary sinus. Only the atrial and ventricular elec-

1274 November-December 1984, Part II PACE, Vol. 7

TACHYCARDIA DETECTION IN IMPLANTABLE ANTITACHYCARDIA DEVICES

trograms were used by the computer, but all wereavailable to the electrophysiologist, who inde-pendently diagnosed each recording for compar-ison with the computer diagnosis.

Following evaluation of the above-described al-gorithm, and motivated by a concern that gradualonset may not be a completely sensitive and spe-cific test for sinus tachycardia, we developed andtested an extrastimulus technique for confirmingthe diagnosis of sinus versus pace-terminable 1:1tachycardias. In this new concept, we hypothe-sized that a premature extra stimulus would resultin a premature ventricular response during sinustachycardia but not during many forms of 1:1 pace-terminable tachycardias. If a tachycardia is ven-tricular, AV reentrant, or AV nodal reentrant, thewavefront of the stimulated atrial extrasystole willcollide with the retrogradely conducted impulseof the tachycardia and the rhythm of the ventriclewill not be perturbed. This is illustrated in Figure2. The ladder diagram represents the spontaneousrhythm in which a stimulated atrial extrasystoleand its propagated wavefront is imposed. The bro-ken line represents what would have depolarizedif the atrial extrastimulus had not occurred.

Eleven patients were tested during electro-physiologic studies to evaluate our concept forautomatic atrial extrastimuius testing during

ATWU. EXnuanMLLAHON OUWQ TACHVCAROM

SWU3 TACHVCARDM VBITnCULAR TACHYCAncu

A-V HEEKTRAMT TACHYCARMA

Ul

A-V NCOAL flEEKniANT TACHYCAfOA

Figure 2. Ladder diagrams represenling a sponta-neous rhythm in which a hypothelical slimtilated atrialextrasystole is introduced. The broken line representswhat would have depolarized if the afrial extrastimuiushad not occurred. The premature extra stimulus willresult in a premature ventricuiar response during sinustachycardia bu( not during many forms of 1:\ pace-terminable tachycardias.

tachycardia. Seven patients had 10 different 1:1paroxysmal tachycardias. There were 5 ortho-dromic AV reentrant tachycardias all using left-sided bypass tracts, 3 AV nodal reentrant tachy-cardias, and 2 atrial tachycardias. Seven of these11 were tested during sinus tachycardia (inducedby atropine and/or isoproterenol) as well as duringtheir pace-terminable tachycardia. Three patientshad both sinus tachycardia and their respectiveparoxysmal 1:1 tachycardias. Recording elec-trodes were placed in the high right atrium andright ventricular apex. After induction and stabi-lization of the tachycardia, single extrastimuli weredelivered to the high right atrium beginning at theend of diastole and decrementing by 10 ms untilatrial refractoriness was encountered. The ven-tricular response was carefully measured from 100mm/s strip-chart recordings and AV conductioncurves were drawn.

Results

Twenty-two tape recorded electrograms wereprocessed automatically by the computer fortachycardia recognition and 21 were diagnosedcorrectly. These included: atrial fihrillation [2),atrial flutter (2), atrial tachycardia (2), ventriculartachycardia (4), 1:1 paroxysmal tachycardia [9),and sinus tachycardia (3). The single error oc-curred in diagnosing an almost 1:1 ventriculartachycardia with retrograde activation and occa-sional ventriculoatrial block. Figure 3 shows anexample of an AV nodal tachycardia which wasinitiated by two atrial stimuli. The atrial prema-ture beats conduct with delay to the ventricles andinitiate paroxysmal AV nodal tachycardia with acycle length of 350 ms. The tachycardia has one-to-one atrioventricular relationship and is diag-nosed correctly by the computer algorithm as par-oxysmal 1:1 tachycardia.

The atrial extra stimulus method was tested on11 patients. Increasingly early atrial extrastimulicaused the suhsequent ventricular heat to be morepremature in all 7 patients with sinus tachycardiaand in none of the 10 patients with paroxysmal1:1 tachycardias. In particular, when the atrial ex-trastimulus was premature by 80 ms, the subse-quent ventricular response was at least 30 ms earlyin all 7 patients with sinus tachycardia and less

PACE, Vol, 7 November-December 1984, Part H 1275

JENKINS, ET AL.

V I

RV

Figure 3. The first four complexes are sinus beats with the cycle length of 780 ms followed hytwo atrial extrastimuli that capture the atrium. The atrial premoture beats conducl with delayto the ventricles and initiate paroxysmal AV nodal tachycardia at (he cycle length of 350 ms.The first complex is narrow and subsaqueni complexes are aberrantly conducted with LBBBconfiguration. The tachycardia has one In one afrio-ventriculnr conduction. (Computer: parox-ysmal 1:1 tachycardia)

AES DURING TACHYCARDIA

STcndAT STQnd>VRT

900-

400-

300-

JOO-

no-

0-

A

A ST

vv

400-

300-

700-

no-

0- ' 1 • 1

A SI

200 300 400 500 no IDO 300 400 MO

STdidA/NRT ST and A/NRT

ASTT «MtT

—<—>-—1—•—I—'—I200 300 400 300

— I ' 1 • r • 1 . ,

no 200 SOO 400 300

Figure 4. Conduction curves obtained during both the patient's sinus tachycardia and parox-ysmal tachycardia in four cases. Open triangles represent results obtained during sinus tachy-cardia and soJid triangles represent data obtained during paroxysmal 1:1 tachycardia.

1276 November-December 1984, Part II PACE, Vol. 7

TACHYCARDIA DETECTION IN IMPLANTABLE ANTITACHYCARDIA DEVICES

than or equal to 10 ms early in all 10 cases ofparoxysmal tachycardia. When the atrial extra-stimulus was premature hy 100 ms, the next ven-tricular beat was at least 50 ms premature in all 7patients with sinus tachycardia and less than orequal to 10 ms premature in all 10 cases of parox-ysmal 1:1 tachycardia.

Figure 4 presents data from four cases. Each graphshows conduction curves ohtained during both thepatient's sinus tachycardia and his paroxysmal 1:1tachycardia. Results obtained during sinus tachy-cardia are indicated by the open triangles and thoseobtained during paroxysmal 1:1 tachycardia areindicated by the solid triangles. In sinus tachy-cardia, as the atrial extrastimulus becomes morepremature the ventricular response becomes in-creasingly premature in all panels of Figure 4. Inatrial tachycardia [upper left), the ventricular re-sponse does not become more premature when theatrial extrastimulus becomes more premature. Thesame phenomenon is seen in the case of AV reen-trant tachycardia [upper right) and in two cases ofAV nodal reentrant tachycardia [lower left andright). Single atrial extrastimuli during tachy-cardia did not provoke faster tachycardias in anyof our patients, but did terminate one paroxysmaltachycardia.

Conclusions

We have developed dual-chamber recognitionschemes that differentiate among most types oftachycardias and have demonstrated accuracy in21/22 tachycardias. In an effort to further refinethe distinction of sinus from paroxysmal 1:1tachycardia, we have tested an additional conceptin which we deliver an atrial extrastimulus andexamine the relationship of ventricular response.In all cases of sinus tachycardia the prematureatrial stimulus evoked a similar premature ven-tricular response. In all cases of paroxysmal tachy-cardia the premature atrial stimulus led to no ven-tricular response more premature than 10 ms.

It must he pointed out that interpretation of theresponse to the atrial extrastimulus would be dif-ficult if the tachycardia were not regular. Also themethod may fail to distinguish sinus tachycardiafrom AV reentrant tachycardia using a right-sidedpathway or from the uncommon form of AV nodaltachycardia. With these reservations we believethe atrial extrastimulus shows promise in safelydifferentiating sinus tachycardia from paroxysmal1:1 tachycardia in most patients and would proveuseful in an antitachycardia device. Further eval-uation of the method in a larger number of patientswith a wider variety ot tachycardias is indicated.

References

1. Camin, A.I., Nathan, A.W., Ward. D.E.: Responsivepacing for tachycardia termination. Vectors, 9:2.1981.

2. Bexton, R., Ward, D.E., Hellestrand, K., et al.: QTinterval during paroxysmal functional tachycardia.£ur. Heart /.. 2:186, 1981.

3. Furman. S., Fischer. J.D.. Pannizo. R: Necessity ofsignal processing in tachycardia detection, lii S.S.Baroid and J. Mugica (Eds.): The Third Decade ofCardiac Pacing, New York, Futura Publishing Com-pany, 1982, p. 265.

Fischer. J.D., Goldstein, M., Ostrow, E., et al.: Max-imal rate of tachycardia development: Sinus tachy-cardia with sudden exercise vs. spontaneous ven-tricular tachycardia. PACE, 6:221. 1983.Arzbaet:her, R., Bump, T., lenkins, J., et al.: Auto-matic tachycardia recognition. PAGE, 7. 1984 (inpress).

PACE, Vol. 7 November-December 1984, Part II 1277