Br Heartj1 1995;74:403-407

Atrial electromechanical sequence in normalsubjects and patients with DDD pacemakers

Ke Wang, Han B Xiao, Shinichi Fujimoto, Derek G Gibson

AbstractObjective-To assess the effect of rightatrial appendage pacing on atrial electro-mechanical interrelations in patientswith DDD pacemakers.Design-Prospective study by M modeechocardiogram, Doppler echocardio-gram, and apexcardiogram, along withelectrocardiogram and phonocardio-gram.Setting-Tertiary cardiac referral centre.Patients-20 patients with DDD pace-makers and 20 age matched normalcontrols.Results-Age, RR interval, atrial size, leftventricular size, and fractional shorten-ing were similar in the two groups. Atrialelectromechanical delay (the time fromthe onset ofP wave or atrial pacing spikeon ECG to the onset of atrial contractionon M mode echogram) was 68 (SD 7) msat the lateral site of right atrium, 82 (9)ms at the central fibrous body, 93 (11) msat the lateral site of left atrium in nor-mals. In patients with DDD pacing, how-ever, this delay increased to 85 (22) ms,117 (23) ms, and 138 (25) ms respectively(all P < 0.01). Interatrial mechanicaldelay (the time from the onset of rightatrial motion to the onset of the left)increased from 25 (6) ms in normal con-trols to 53 (18) ms in patients (P < 0.01).Intra-atrial mechanical dispersion (thetime from the earliest to the latest onsetof regional atrial motion around theatrioventricular ring) in the right atriumincreased from 6 (2) ms in normals to 19(2) ms in patients (P < 0.01), but itremained unchanged in the left atrium (6(2) ms in normal controls v 7 (2) ms inpatients, P > 0.05). Peak atrial shorten-ing rate was not different between the twogroups. Differences of atrial electro-mechanical activity between the twogroups were also reflected on Dopplerechocardiogram and apexcardiogram.Conclusions-Right atrial appendagepacing disturbs the normal coordinatesequence of right atrial mechanicalactivity and leads to a striking and vari-able increase in intra-atrial conductiontime as well as in interatrial conductiontime. Left atrial contraction remains syn-chronous although the timing of the startof its contraction was delayed. These val-ues can be determined in individualpatients to allow optimal setting of DDDpacemakers.

(Br Heart_j 1995;74:403-407)

Keywords: DDD pacemaker; right atrial appendagepacing; atrial electromechanical interrelations

Dual chamber pacemakers have been usedclinically for more than 20 years.1 They wereconceived with the idea of maintainingatrioventricular (A-V) synchrony and atrialsystolic function, and of providing a physio-logical rate response in patients with normalsinus node function and A-V block. For thesereasons, dual chamber pacemakers, especiallyof the DDD or DDDR type, have beendescribed as "physiological", and are consid-ered superior to conventional VVI pacing.23However, pacing the right atrial appendage,the usual site for the atrial lead, is known todelay left atrial contraction, which may pro-long the left heart mechanical A-V intervaland thus affect left ventricular function.4 Wetherefore undertook this detailed study of theelectromechanical effects of right atrialappendage pacing using a series of non-invasive techniques, in order to clarify under-lying mechanisms.

MethodsPATIENTSWe studied 36 patients with DDD pace-makers implanted, in whom the atrial elec-trode tip was positioned in the right atrialappendage. In 20, complete data wererecorded, and these form the study group.Eight patients had complete A-V block orMobitz type II second degree A-V block, eighthad sick sinus syndrome, and four had sinusbradycardia. Underlying coronary artery dis-ease was also present in three patients, andaortic valve disease in one patient. None wasreceiving drugs known to influence cardiacelectrophysiology. Twenty age matched nor-mals were used as controls, in whom therewas no clinical, electrocardiographic (ECG),radiographic, or echocardiographic evidenceof heart disease.

ECHOCARDIOGRAPHYA standardised series ofM mode echocardio-grams, guided from the cross sectional dis-play, was recorded from right and left A-Vrings with the patients in left lateral position,by a 2 5 MHz phased array transducer with aHewlett Packard Sonos 1000 system. Echo-cardiograms were recorded photographicallyat a paper speed of 100 mm/s, with simultane-ous lead II ECG showing a clear P wave or

Cardiac Department,Royal BromptonHospital, LondonK WangH B XiaoS FujimotoD G GibsonCorrespondence to:Dr D G Gibson, CardiacDepartment, RoyalBrompton Hospital, SydneyStreet, London SW3 6NP.Accepted for publication31 May 1995

403

on July 21, 2021 by guest. Protected by copyright.

http://heart.bmj.com

/B

r Heart J: first published as 10.1136/hrt.74.4.403 on 1 O

ctober 1995. Dow

nloaded from

Wang, Xiao, Fujimoto, Gibson

Figure I Diagram ofapicalfour chamber crosssectional view showing theM mode cursor positionedthrough the lateraltricuspid ring, the centralfibrous body, and thelateral mitral ring.

atrial pacing spike, along with a phonocardio-gram. Motion of the A-V rings during the car-diac cycle was obtained with the M modecursor directed from the apical four chamberview.7 The cursor was longitudinally placedthrough the anterior, lateral, and posteriorsites of right and left A-V rings and throughthe central fibrous body (figs 1 and 2). Ateach site, the beam was oriented so that it wasparallel to the longitudinal component of theA-V ring motion. In addition, a standardminor axis left ventricular M mode echocar-diogram, pulsed Doppler records of trans-

Figure 2 (A) M modeechocardiogram at thelateral site of tricuspid ringin a normal subject.Movements of theatrioventricular ringrepresent mechanicalactivity of right atrium (a)and right ventnicle (v).(B) The upper panel showsthe digitised plot of thetrace. The lower panelshows the peak atrialshortening rate (r). ECG,electrocardiogram; PCG,phonocardiogram; 1, atrialelectromechanical delay, itwas measuredfrom theonset of the P wave onECG to the onset of atrialshortening; 2, the timeintervalfrom the onset ofthe P wave on ECG to themaximum atrial shorteningamplitude

A

B

3.0(cm)

1.5

0.0

-15.0

I I

mitral and transtricuspid blood flow, and aleft ventricular apexcardiogram (ACG) werealso recorded. The M mode echograms of theA-V rings were digitised8 to quantify theextent and velocity of movement (fig 2).A typical M mode echocardiogram of the

right A-V ring is shown in fig 2. During ven-tricular systole the A-V ring moves towardsthe apex of the ventricle. During ventriculardiastole, there is a rapid movement in thereverse direction corresponding to ventricularearly diastole. There is in addition a clearsecond reverse movement which correspondsto atrial systole. During atrial systole, move-ments at the different sites ofA-V rings reflectmechanical activity of different regions of thetwo atria, while movement of central fibrousbody (CFB) reflects mechanical activity of theinteratrial septum.9The following measurements were made

from these records:On A-V ring echograms:(1) Regional atrial electromechanical delay:This interval was measured from the onset ofthe P wave or atrial pacing spike on ECG tothe onset of atrial mechanical activity phase ofA-V ring motion (fig 2). Electromechanicaldelay was measured at each of the seven sites.(2) The peak atrial shortening rate, from thecorresponding digitised trace (fig 2).(3) The time interval from the onset of the Pwave or atrial pacing spike to the maximalatrial shortening amplitude (P maximumshortening) (fig 2).On the Doppler record and apexcardiogram:(1) The time interval from the onset of the Pwave or atrial pacing spike to the onset of the Awave on transmitral and transtricuspid flow (Ponset ofA wave).(2) A wave peak velocity on transmitral andtranstricuspid blood flow.(3) The time interval from the onset of the Pwave or atrial pacing spike to the onset of the Awave on the apexcardiogram (P onset of Awave on ACG).We also measured left ventricular end dias-

tolic and end systolic dimensions and frac-tional shortening, and left atrial dimension onthe standard transverse M mode echocardio-gram.

In the normal group, the onset of P waveon ECG was derived by determining the PRinterval on a standard 12 leads ECG, usingHewlett Packard pagewriter software.Individual measurements were made from theQ wave of the ECG, which was a more repro-ducible landmark, and the computed PRinterval added. In the DDD pacing group,only the heart beats with both atrial and ven-tricular pacing were used.From these measurements, the following

quantities were calculated:(1) Interatrial mechanical delay: the timeinterval from the onset of atrial mechanicalmotion at the lateral site of the right A-V ringto that of the lateral site of the left.(2) Intra-atrial mechanical dispersion. Thiswas defined as the time difference from theearliest to the latest onset of regional atrialmotion among the three sites on each atrium.

0 -n ad

404

on July 21, 2021 by guest. Protected by copyright.

http://heart.bmj.com

/B

r Heart J: first published as 10.1136/hrt.74.4.403 on 1 O

ctober 1995. Dow

nloaded from

Atrial electromechanical sequence in normal subjects and patients with DDD pacemakers

It represents the spread over time of the onsetof nmechanical activity in each atrium.(3) The time delay from the onset ofmechanical atrial contraction at the A-V ringto the onset of the A wave on the two Dopplerrecords or on the apexcardiogram. We mea-

sured the time from the onset of atrial con-

traction at the lateral site of left side A-V ringto the onset ofA wave on transmitral flow andon ACG, and the time from the onset of atrialcontraction at the lateral site of right side A-Vring to the onset of A wave on transtricuspidflow.

REPRODUCIBILITYTo assess reproducibility of the variables on

which our major conclusions depend, electro-mechanical intervals in 18 patients were

measured by two blinded observers. Variablesincluded atrial electromechanical delay at theseven sites of A-V ring on M mode, and thetime intervals on Doppler and apexcardiogram.

DATA ANALYSISThe average values were taken from three suc-

cessive beats. They were expressed as mean

(SD) for group measurements. Paired and

Table 1 General data. Values are mean (SD)

Normal controls PatientsIndex (n = 20) (n = 20)

Age (years) 57 (18) 65 (11)Sex (male:female) 11:9 12:8P-R interval or A-V

delay (ms) 175 (22) 189 (37)R-R interval (ms) 878 (117) 812 (91)LA dimension (mm) 32 (6) 31 (6)LVEDD (mm) 50 (4) 51 (10)LVESD (mm) 34 (5) 36 (9)FS (%) 33 (6) 30 (7)

LA, left atrium; LVEDD, left ventricular end diastolicdimension; LVESD, left ventricular end systolic dimension;FS, left ventricular fractional shortening. No significantdifference between the two groups.

Table 2 Atrial electromechanical delay. Values are mean (SD)

Site Normal controls (n = 20) Patients (n = 20)

R anterior (ms) 67 (7) ) 79 (28)R lateral (ms) 68 (7) t)j85 (22) *R posterior (ms) 69 (7) J tt96 (25)jj

CFB (ms) 82 (9) J 1t117 (23) J

L antenror (ms) 92 (10) t136 (25)L lateral (ms) 93 (11) - JJ t138 (25) J JL posterior (ms) 92 (11)J t136 (23) J

R, right atnrum; L, left atnium; CFB, central fibrous body.tP < 0-01 v normal controls; *P < 0-01 by intragroup comparison.

Table 3 Atrial electromechanical measurements onM mode echogram. Values are mean (SD)

P to maximum shortening (ms) Peak shortening rate (cmls)

Normal controls Patients Normal controls Patients

Site (n = 20) (n = 20) (n = 20) (n = 20)

Ranterior 188 (16) t223 (23) 8-4 (2-1) 8-1 (2-9)R lateral 188 (16) t220 (26) 9-4 (2-3) 10-5 (3-3)Rposterior 186 (16) t220 (24) 8-6 (2-6) 8-9 (2-8)

CFB 184 (19) t214 (24) 6-2 (1 7)* 6-6 (1 6)*

Lanterior 182 (19) t220 (26) 6-3 (1-5)) 6-8 (1-5))L lateral 183 (19) t220 (24) 6-5 (1-5) * 7-3 (1-7) F*Lposterior 184 (19) t221 (23) 6-3 (1-7)J 7-1 (1-5)J

R, right atrium; L, left atrium; CFB, central fibrous body.tP < 0-01 v normal controls; *P < 0-01 v right side in intragroup comparison.

unpaired Student t tests were used as appro-priate. P < 005 was considered to be statisti-cally significant. Reproducibility was assessedfrom root mean square difference of duplicatedeterminations.

ResultsGENERAL DATAThere were no significant differences betweentwo groups with regard to age, RR interval,left ventricular cavity size, fractional shorten-ing, and left atrial dimension (table 1).

NORMAL SUBJECTSThe onset of mechanical wall motion in eachatrium was effectively synchronous, that ofthe right atrium following approximately 68ms after the onset of the P wave and that ofthe left atrium approximately 93 ms (table 2).Interatrial mechanical delay was 25 (6) ms,and intra-atrial mechanical dispersion was 6(2) ms for both atria. In spite of the consistentdelay in the onset of shortening on the leftside, the timing of maximum shortening wassynchronous in the two atria, occurring about185 ms after the start of the P wave (table 3).Peak shortening rate was rather lower on theleft than the right (table 3), compatible withsmaller overall amplitude of motion.

Other electromechanical intervals on theDoppler and ACG corresponded to that on Mmode (table 4). The onset of A wave on theDoppler and apexcardiogram were both con-sistently delayed with respect to that directlymeasured from A-V ring motion, by 19 ms forthe mitral, 27 ms for the tricuspid Doppler,and 13 ms for apexcardiogram (table 5).

DDD PACING GROUPElectromechanical delay was prolonged (table

Table 4 Doppler and apexcardiogram. Values are mean(SD)

Normal controls PatientsIndex (n = 20) (n = 20)

Transmitral flowP to onset ofA wave (ms) 112 (16) 131 (21)tA wave peak velocity (m/s) 0-57 (0-17) 0-73 (0-20)t

Transtricuspid flowP to onset ofA wave (ms) 95 (14) 102 (23)A wave peak velocity (mis) 0-34 (0 08) 0-42 (0-15)

ApexcardiogramP to onset ofA wave (ms) 106 (16) 149 (20)t

tP < 0-01 v normal controls.

Table 5 Atrial interrelations. Values are mean (SD)Normal controls Patients

Index (n = 20) (n = 20)

Time from onset of leftatrial contraction toonset ofA wave on MV (ms) 19 (12) -7 (12)t

Time from onset of rightatr-al contraction toonset of A wave on TV (ms) 27 (15) 17 (21)

Time from onset of leftatrial contraction toonset of A wave on ACG (ms) 13 (10) 11 (11)

tP < 0-01 v normals controls.MV, mitral valve; TV, tricuspid valve; ACG, apexcardiogram.

405

on July 21, 2021 by guest. Protected by copyright.

http://heart.bmj.com

/B

r Heart J: first published as 10.1136/hrt.74.4.403 on 1 O

ctober 1995. Dow

nloaded from

Wang, Xiao, Fujimoto, Gibson

Figure 3 (A) M modeechocardiograms at theanterior, lateral, andposterior sites of tricuspidring in a normal subject,which show that the timingof right atrial shortening atthese three sites issynchronous, and therange of atrialelectromechanical delay is80-85 ms at the three sites.(B) M modeechocardiograms at theanterior, lateral, andposteior sites of tricuspidring in a patient withDDD pacemaker, whichshow that the timing ofright atrial shortening atthese three sites isasynchronous, and therange of atrialelectromechanical delay is70-115 ms at the threesites. Abbreviations as infig 2.

2), and right atrial mechanical dispersion con-siderably increased to 19 (2) ms (P < 0-01compared to normals). Furthermore, a char-acteristic sequence appeared, with atrial

A

yy.ECG A

- Anterior

BOms 7

-N.~~~~~~~~~~~~~~~~0'

movement of the A-V ring beginning at theright side, followed by the anterior, lateral,and posterior walls (fig 3), and finally the sep-tum and left side (table 2). The mean value ofinteratrial mechanical delay was approxi-mately double the normal value (25 (6) ms innormals versus 53 (18) ms in pacing group, P< 0-01). In spite of this, the contraction pat-tern of the left atrium remained coordinate,though with some increase in scatter amongpatients (table 2), so that the overall value formechanical dispersion was not significantlydifferent from normals (6 (2) ms in normalsversus 7 (2) ms in pacing group, P > 005).The onset of the A wave on the Doppler

and apexcardiogram were also delayed withrespect to normals, by 7 ms for the trans-tricuspid, 19 ms for the transmitral Doppler,and 43 ms for the apexcardiogram (table 4).However, the intervals from the onset of atrialmechanical shortening to the onset ofA wavewere similar to those in the normal controls,apart from what appeared to be the prematureonset of the A wave on the transmitralDoppler (table 5). Once atrial contractionbegan, the timing of maximum shorteningwas synchronous not only within but alsobetween the two atria, and peak atrial short-ening rate did not differ significantly from thatin the normal controls (table 3).

REPRODUCIBILITYThere were no consistent differences betweenpairs of duplicate determinations. The rangeof root mean square difference was from 44to 5-4 ms on M mode echogram, and from5-2 to 7 0 ms on Doppler and apexcardiogram.

BPCG ' ' i-Aw~~~~ECGI

: Anterior

7Oms - -_ 0m ;/"'w ''d.;~~~9.-

Lateral

19o1s

Posterior

11j5ms .c_t~~~~~~~~ , kw.

1A I4t.41 "!L

.............'I

s~~~:.;.

DiscussionOur technique depends on recording atrialwall movement directly rather than its indirect

_______'''' consequences in terms of blood flow acrossL the A-V valves'0 and atrial pressure pattern."

: Since electromechanical interval is unlikely tochange, delay in local activation will be fol-

- ,f lowed by a similar delay in the onset of short-> g ening. Several atrial sites can be examined,

N-3\< and the coordination, the extent, and thevelocity of local shortening assessed. Thisinformation has been correlated with Doppler.'@ | and apexcardiogram, representing the overall

Aftunction of the atriums.Normal atrial wall motion is effectively syn-

: chronous, with a consistent sequence in the* onset of motion involving the right atrium, the

septum, and the left atrium. During rightatrial appendage pacing, the contraction

§ ~.-s. sequence of the right atrium became asyn-chronous (fig 3), starting at the anterior site ofthe tricuspid ring, followed by the lateral, andsubsequently the posterior site, with an overalldispersion of nearly 20 ms. The indirecteffects of atrial contraction with respect to the

- normals were also delayed. Left atrial contrac-tion, though delayed, was as synchronous asin the normal controls.

- 8 of The differences between the relative timesof onset of mechanical events in sinus rhythm

~-A--^ and atrial pacing shed light on normal and

406

. . I . . . . I . . . . I . . . . I . . . . I . . .

on July 21, 2021 by guest. Protected by copyright.

http://heart.bmj.com

/B

r Heart J: first published as 10.1136/hrt.74.4.403 on 1 O

ctober 1995. Dow

nloaded from

Atrial electromechanical sequence in normal subjects and patients with DDD pacemakers

abnormal atrial activation pathways. There isno interatrial conduction pathway analogousto the His-Purkinje system in the ventricles.'2 13Instead, conduction occurs preferentially inregions of increased muscle thickness.'4 15 Theterminal crest, between the sinus node andthe pectinate muscles supporting the right A-Vring, is likely to be responsible for the coordi-nate onset of right atrial motion. When it isactivated from the right atrial appendage, con-duction velocity is much slower, and thespread of activation takes nearly 20 ms to passaround the A-V ring. By contrast, left atrialcontraction is synchronous, strongly suggest-ing that a preferential pathway has nowbecome involved, althoughf the increasedinteratrial mechanical delay shows that theroute by which the activity wave front hasreached the left atrium is clearly abnormal.These findings stress a fundamental differencebetween preferential pathways in the atrium,to which access can be gained from themyocardium, and the His-Purkinje system ofthe ventricle where this is not possible.'213 16Our study clearly had limitations. The most

important was that a significant number of thepatients with pacemakers had a diagnosis ofsinus node disease, and might thus have hadintrinsic atrial disease. In suitable patients,this possibility might have been circumventedby studying the same patients before and afterpacemaker insertion. However, the consis-tency of the effects of atrial pacing makes itmost unlikely that atrial disease was respon-sible. We used different landmarks to assessthe onset of atrial electrical activity in the twogroups: the onset of the P wave is clearly notidentical with the pacing artefact, so thatabsolute estimates of atrial electromechanicaldelay cannot be compared. This differencedoes not, of course, preclude studying disper-sion of mechanical events in individualpatients. Our observations of atrial mechan-ical activity were confined to the two A-Vrings. It would obviously have been desirableto have made a more extensive study of theatriums, but we are unaware of other repro-ducible landmarks suitable for study by trans-thoracic echo. It is possible that their numbermight have been extended by trans-oesophageal echo, which allows access to thetwo atrial appendages.

Physiological "asynchrony" between thetwo atria has been well recognised for manyyears. The right atrial contraction has beenshown to precede the left by approximately 13ms in the dog'6 and by 20 ms in man.11Interatrial electrical conduction delay has beenmeasured in man using catheters in the rightatrium and coronary sinus,'7 18 normal valuesbeing 77 (8) ms with the range 62-88 ms.'7These values are significantly greater thanthose of mechanical interatrial delay weobtained in our study. Obviously, this differ-ence cannot be ascribed to any peculiarity ofmotion of the two A-V rings, since their rela-tive onset of motion was directly reflected inthe relative times of onset of tricuspid andmitral Doppler A waves. It seems veryunlikely that the electromechanical coupling

interval differs between the two atria. If theextent of left sided mechanical delay was asgreat as that predicted from electrical mea-surement, it would have profound haemo-dynamic consequences in that the normal leftatrial contribution to left ventricular fillingwould not occur. The likeliest reason for thelong electrical interatrial delay appears to bethat left atrial potentials measured from thecoronary sinus or oesophagus are significantlydelayed with respect to those activating theatrial muscles which underlie left atrial con-traction, while those on the right are likely to berecorded high in the atrium near the sinus node.The methods we have described here may

form the basis of an additional method ofstudying atrial function, and of deriving basicinformation about preferential conductionwithin the atria. This approach might be usedin association with Doppler studies, to opti-mise pacemaker settings, and to investigateatrial function in patients with sinus node dis-ease or following atrial surgery.

We are grateful to Dr Siew Yen Ho, Department ofPaediatrics, National Heart and Lung Institute, for her help inpreparing the manuscript. Ke Wang and Han B Xiao are sup-ported by The Royal Brompton Hospital Special Cardiac Fund.

1 Sutton R, Perrins JE, Citron P. Physiological cardiac pacing[review]. PACE 1980;3:207-19.

2 Forfang K, Otterstad JE, Ihlen H. Optimal atrioventriculardelay in physiological pacing determined by Dopplerechocardiography. PACE 1986;9: 17-21.

3 Iwase M, Sotobata I, Yokota M, Takagi S, Jing HX, KawaiN, et al. Evaluation by pulsed Doppler echocardiographyof the atrial contribution to left ventricular filling inpatients with DDD pacemakers. Am J Cardiol 1986;58:104-9.

4 Ausubel K, Klementowicz P, Furman S. Interatrial con-duction during cardiac pacing. PACE 1986;9:1026-31.

5 Chirife R, Ortega DF, Salazar AI. Nonphysiological leftheart AV intervals as a result of DDD and AAI "physio-logical" pacing. PACE 199 1;14: 1752-6.

6 Camous JP, Raybaud F, Dolisi C, Schenowitz A, VarenneA, Baudouy M. Interatrial conduction in patients under-going AV stimulation: effects of increasing right atrialstimulation rate. PACE 1993;16:2082-6.

7 Jones CJH, Song GJ, Gibson DG. An echocardiographicassessment of atrial mechanical behaviour. Br Heart J1991;65:31-6.

8 Gibson DG, Brown D. Measurement of instantaneous leftventricular dimension and filling rate in man, usingechocardiography. BrHeart3J 1973;35:1141-9.

9 Wang K, Ho SY, Gibson DG, Anderson RH. Architectureof atrial musculature in humans. Br Heart J 1995;73:559-65.

10 Abe H, Yokouchi M, Deguchi F, Saitoh F, Yoshimi H,Arakaki Y, et al. Measurement of left atrial systolic timeintervals in hypertensive patients using Doppler echocar-diography: relation to fourth heart sound and left ven-tricular wall thickness. JAm Coil Cardiol 1988;11:800-5.

11 Braunwald E, Fishman AP, Cournand A. Time relation-ship of dynamic events in the cardiac chambers, pul-monary artery and aorta in man. Circ Res 1956;4:100-7.

12 Janse MJ, Anderson RH, Van Capelle FJL, Durrer D. Acombined electrophysiological and anatomical study ofthe human fetal heart. Am Heart3 1976;91:556-62.

13 Anderson RH, Becker AE, Tranum-Jensen J, Janse MJ.Anatomico-electrophysiological correlations in the con-duction system-a review. Br HeartJ 1981;45:67-82.

14 Spach MS, King TD, Barr RC, Boaz DE, Morrow MN,Herman-Giddens S. Electrical potential distribution sur-rounding the atna during depolarization and repolariza-tion in the dog. Circ Res 1969;23:857-73.

15 Spach MS, Lieberman M, Scott JG, Barr RC, JohnsonEA, Kootsey JM. Excitation sequences of the atrialseptum and the AV node in isolated hearts of the dogand rabbit. Circ Res 1971;29:156-72.

16 Bachmann G. The inter-auricular time interval. Am JPhysiol 1916;41:309-20.

17 Josephson ME, Scharf DL, Kastor JA, Kitchen JG. Atrialendocardial activation in man: electrode catheter tech-nique for endocardial mapping. Am _J Cardiol 1977;39:972-81.

18 Leier CV, Jewell GM, Magorien RD, Wepsic RA, SchaalSF. Interatrial conduction (activation) times. Am JfCardiol 1979;44:442-6.

407

on July 21, 2021 by guest. Protected by copyright.

http://heart.bmj.com

/B

r Heart J: first published as 10.1136/hrt.74.4.403 on 1 O

ctober 1995. Dow

nloaded from