Hemodynamic importance of preserving the normal sequence of ventricular activation in permanent...

9
Hemodynamic importance of preserving the normal sequence of ventricular activation in permanent cardiac pacing Pacing the right ventricle in the apex profoundly modifies the sequence of activation and thus the sequence of contraction and relaxation of the left ventricle. To evaluate the relative importance of preserving normal ventricular activation sequence and optimal atrioventricular (AV) synchrony in permanent pacing, we compared the effects of three pacing modes: AAI, preserving both normal AV synchrony and normal activation sequence; DDD, with complete ventricular capture that preserves only AV synchrony; and VVI, disrupting both, at rest and during exercise. Hemodynamic and radionuclide studies were performed in 11 patients who had normal intrinsic conduction and who were implanted on a long-term basis with a DDDR pacemaker for isolated sinus node dysfunction. AAI versus DDD and VVI significantly increased cardiac output at rest (6.6 ± 1.3 L/min vs 6 ± 0.9 L/min vs 5 ± 1 L/min; p < 0.01) and during exercise (13.5 ± 2 L/min vs 12.1 ± 2.2 L/min vs 14.4 ± 2.1 L/min; p < 0.01). Pulmonary capillary wedge pressure was lowest with AAI (15.4 ± 4.5 mm Hg), with an average reduction of 17% compared with DDD (19.6 ± 5 mm Hg; p < 0.01) and of 30% compared with VVI (25.8 ± 7 mm Hg; p < 0.01) during exercise. Identical benefits were observed for all other hemodynamic parameters: right atrial pressure, pulmonary artery pressure, left ventricular (LV) stroke work index, and systemic vascular resistances. LV ejection fraction was significantly higher in AAI than in DDD at rest (61% vs 68%, respectively; p < 0.05) and during exercise (65% vs 60%, respectively; p < 0.05). This improvement in LV systolic function resulted principally from the increase in septal ejection fraction. LV filling also was improved in AAI as demonstrated by a significant increase in peak filling rate at rest and during exercise. These data show the importance of preserving, whenever possible, not only normal AV synchrony but also normal ventricular activation sequence in permanent cardiac pacing. (AM HEART J 1995; 129:1133-41 .) Christophe Leclercq, MD, a Daniel Gras, MD, a Alain Le Helloco, MD,b Luc Nicol, MD, b Philippe Mabo, MD, a and Claude Daubert, MD a Rennes, France Right apical ventricular pacing as usually used in permanent cardiac pacing (VVI and dual-chamber pacing modes) profoundly alters the sequence of ventricular activation. It results in major interven- tricular asynchrony, delaying left ventricular (LV) activation from 30 to 60 msec in normal hearts and to 180 msec in diseased hearts) Furthermore it inverses the direction of LV activation, starting from the apex and ending at the base. These changes pro- duce significant impairment in LV systolic function, especially septal motion, and in LV filling. 24 Until From athe Department of Cardiology and bthe Department of Nuclear Medicine, Hotel Dieu/Centre Hospitalier et Universitaire de Rennes. Received for publication Aug. 15, 1994; accepted Oct. 1, 1994. Reprint requests: Christophe Leclercq, MD, Service de Cardiologie A, Ho- tel Dieu/CHRU, 35033 Rennes Cedex, France. Copyright ® 1995 by Mosby-Year Book, Inc. 0002-8703/95/$3.00 + 0 4/1/61975 the past few years, it was postulated that these adverse effects could be concealed, provided optimal atrioventricular (AV) synchrony was preserved in DDD pacing. However, this classic postulate has been questioned by the results of clinical studies 57 sug- gesting that preserving a normal ventricular activa- tion sequence could significantly improve LV systolic and diastolic function in patients paced on a long- term basis. However, some conclusions from these studies could be misleading because of methods or inclusion criteria. In particular it is not clear whether AV synchrony was optimal in all patients in the AAI mode and in the DDD mode. The purpose of the current study was to evaluate the relative benefit of preserving optimal AV syn- chrony and normal ventricular activation sequence by comparing three modes of pacing: AAI, preserving both normal AV synchrony and normal activation sequence; DDD, with permanent ventricular capture, 1133

Transcript of Hemodynamic importance of preserving the normal sequence of ventricular activation in permanent...

Hemodynamic importance of preserving the normal sequence of ventricular activation in permanent cardiac pacing

Pacing the right ventricle in the apex profoundly modifies the sequence of activation and thus the sequence of contraction and relaxation of the left ventricle. To evaluate the relative importance of preserving normal ventricular activation sequence and optimal atrioventricular (AV) synchrony in permanent pacing, we compared the effects of three pacing modes: AAI, preserving both normal AV synchrony and normal activation sequence; DDD, with complete ventricular capture that preserves only AV synchrony; and VVI, disrupting both, at rest and during exercise. Hemodynamic and radionuclide studies were performed in 11 patients who had normal intrinsic conduction and who were implanted on a long-term basis with a DDDR pacemaker for isolated sinus node dysfunction. AAI versus DDD and VVI significantly increased cardiac output at rest (6.6 ± 1.3 L/min vs 6 ± 0.9 L/min vs 5 ± 1 L/min; p < 0.01) and during exercise (13.5 ± 2 L/min vs 12.1 ± 2.2 L/min vs 14.4 ± 2.1 L/min; p < 0.01). Pulmonary capillary wedge pressure was lowest with AAI (15.4 ± 4.5 mm Hg), with an average reduction of 17% compared with DDD (19.6 ± 5 mm Hg; p < 0.01) and of 30% compared with VVI (25.8 ± 7 mm Hg; p < 0.01) during exercise. Identical benefits were observed for all other hemodynamic parameters: right atrial pressure, pulmonary artery pressure, left ventricular (LV) stroke work index, and systemic vascular resistances. LV ejection fraction was significantly higher in AAI than in DDD at rest (61% vs 68%, respectively; p < 0.05) and during exercise (65% vs 60%, respectively; p < 0.05). This improvement in LV systolic function resulted principally from the increase in septal ejection fraction. LV filling also was improved in AAI as demonstrated by a significant increase in peak filling rate at rest and during exercise. These data show the importance of preserving, whenever possible, not only normal AV synchrony but also normal ventricular activation sequence in permanent cardiac pacing. (AM HEART J 1995; 129:1133-41 .)

Christophe Leclercq, MD, a Daniel Gras, MD, a Alain Le Helloco, MD, b Luc Nicol, MD, b Philippe Mabo, MD, a and Claude Daubert, MD a Rennes, France

Right apical ventricular pacing as usually used in permanent cardiac pacing (VVI and dual-chamber pacing modes) profoundly alters the sequence of ventricular activation. It results in major interven- tricular asynchrony, delaying left ventricular (LV) activation from 30 to 60 msec in normal hearts and to 180 msec in diseased hearts) Furthermore it inverses the direction of LV activation, starting from the apex and ending at the base. These changes pro- duce significant impairment in LV systolic function, especially septal motion, and in LV filling. 24 Until

From athe Department of Cardiology and bthe Department of Nuclear Medicine, Hotel Dieu/Centre Hospitalier et Universitaire de Rennes.

Received for publication Aug. 15, 1994; accepted Oct. 1, 1994.

Reprint requests: Christophe Leclercq, MD, Service de Cardiologie A, Ho- tel Dieu/CHRU, 35033 Rennes Cedex, France.

Copyright ® 1995 by Mosby-Year Book, Inc. 0002-8703/95/$3.00 + 0 4 /1 /61975

the past few years, it was postulated that these adverse effects could be concealed, provided optimal atrioventricular (AV) synchrony was preserved in DDD pacing. However, this classic postulate has been questioned by the results of clinical studies 57 sug- gesting that preserving a normal ventricular activa- tion sequence could significantly improve LV systolic and diastolic function in patients paced on a long- term basis. However, some conclusions from these studies could be misleading because of methods or inclusion criteria. In particular it is not clear whether AV synchrony was optimal in all patients in the AAI mode and in the DDD mode.

The purpose of the current study was to evaluate the relative benefit of preserving optimal AV syn- chrony and normal ventricular activation sequence by comparing three modes of pacing: AAI, preserving both normal AV synchrony and normal activation sequence; DDD, with permanent ventricular capture,

1133

• J u n e 1 9 9 5

1134 Leclercq et al. American Heart Journal

preserving normal AV synchrony alone; and VVI in carefully selected D D D - p a c e d pat ients . Unlike the previous studies, this work used not only noninvasive me thods (radionuclide angiography) bu t also inva- sive h e m o d y n a m i c me thods at res t and dur ing sub- max ima l exercise.

METHODS Inclusion criteria. The following criteria were required

for inclusion. (1) Patients were implanted on a long-term basis with a DDD rate-responsive pacemaker for symp- tomatic sinus node dysfunction (SND) with atrial chron0- tropic incompetence (ACI), defined as a peak heart rate <75% of the maximal predicted heart rate, during a symp- tom-limited exercise test. (2) Before implantation, AV and intraventricular conductions were normal as demonstrated by a PR interval --<200 msec on a resting electrocardiogram (ECG), a Wenckebach point >130 beats/min during incre- mental atrial pacing, and permanently narrow QRS com- plex (<110 msec). (3) All antiarrhythmic drugs including ~-adrenergic receptor blocking agents were stopped for at least five half-lives before inclusion. This study was ap- proved by our institutional Ethics Committee, and every patient gave informed consent.

Study protocol. Complete ventricular capture in the DDD mode was defined as a QRS duration identical to the QRS duration in the VVI mode at rest and during exercise. An initial training exercise test was performed in the DDD mode to determine the longest AV delay resulting in com- plete ventricular capture. The next day, patients under- went a hemodynamic study and a radionnclide study at rest and during submaximal exercise. The AAI, DDD, and VVI pacing modes were successively tested in random order at 1-hour intervals. The AAI mode preserved both AV syn- chrony and normal activation sequence; the DDD mode preserved only AV synchrony; and the VVI mode preserved neither AV synchrony nor normal activation sequence.

Exercise test. Patients exercised on a cycloergometer (Rodby Elektronik AB, Germany) in the supine position. The workload was increased progressively by 10 W in 1-minute steps from 30 W to a plateau of 60 W. During ex: ercise the surface ECG was continuously recorded (Sie- mens Elema Mingograph, Solna, Sweden). Because of the inclusion criteria, the patients were paced continuously throughout the tests. To facilitate comparison between patients, the pacing rate was systematically set at 70 beats/min at rest and at 120 beats/min at the beginning of the 60 W plateau. Thus all the hemodynamic parameters were obtained at the same pacing rate at rest and during exercise in all patients and in all tested modes.

Hemodynamic study. The hemodynamic study was performed with a double-lumen 7F catheter (Swan-Ganz, Edwards Critical Care, Irvine, Calif.) introduced through the subclavian vein. The catheter tip was positioned in the right pulmonary artery to measure right heart pressures (right atrial pressure [RAP]; systolic, diastolic, and mean pulmonary arterial pressures [PAP]; and pulmonary cap-

illary wedge pressure [PCWP]). Cardiac output (CO) was determined by the thermodilution technique (9520 mod- ule, Edwards). Pressures and CO were measured at rest, at the end of each 1-minute exercise step (30, 40, and 50 W), and during the first 5 minutes of the 60 W plateau. At rest and during the 60 W plateau, three consecutive measure- ments of CO were made. The value considered for analysis was the average of these three determinations, provided there was no value outside the limit of 10 % compared with the mean. For the intermediate steps (30, 40, and 50 W) only one measurement was performed. Blood pressure (BP) was measured by cuff each minute. These data were used to calculate the cardiac index (CI), the LV stroke work index (LVSWI = mean B P - mean PCWP X 0.136 X 1.05 x CI/heart rate [gram beats per squared meter]), and the systemic vascular resistance (SVR -- mean BP - mean RAP/CO).

Radionuclide study. After in vivo erythrocyte labeling with 740 MBq (20 mCi) of technetium 99m-pertechnetate, equilibrium multiple-gated acquisition studies were per- formed with a large-field DSX camera (Sopha Medical, Buc, France) equipped with a high-resolution collimator, s Gated equilibrium radionuclide cardiac images were ac- quired in the 45-degree left anterior oblique view for each pacing mode for 4 minutes at rest before exercise and dur- ing the 60 W exercise plateau. Images were collected in a 64 x 64 pixel matrix with 16 frames per RR interval. The acquisition was continued at rest for 8 minutes, providing approximately 250,000 counts within the left ventricle, and at the end of exercise for 4 minutes.

The ECG was monitored continuously to ensure gating of the QRS complex. In all patients the R wave served as the gating trigger. The operator chose the dominant R-R interval with a 10 % to 15 % window, excluding premature ventricular beats. Quantitative cardiac parameters were derived from the time-activity curve. The regions of inter- est were automatically drawn by the Sopha computer in systole and diastole after background correction. If the separation of the left atrium from the left ventricle was controversial, two independent operators could draw re- gions of interest manually. Global LV ejection fraction (LVEF) was calculated for each pacing mode from the time-activity curve according to the formula LVEF = EDc - EDs/EDc, where EDc is counts at end-diastole and ESc is counts at end-systole. All counts were corrected for background. Regional wall motion was calculated in the three pacing modes and represented count changes in sep- tal, inferoapical, and posterolateral segments of the left ventricle. The peak filling rate (PFR) was calculated by taking the first derivative of the time-activity curve. 9

Statistics. Each patient was his or her own control. Data are presented as means + standard deviation. One-way analysis of variance was used for each parameter measured at rest and during exercise. If the analysis of variance was significant, the three pacing modes were compared two by two with Student's t test for paired series. A p value <0.05 was considered significant.

Volume 129, Number 6 American Heart Journal Leclercq et al. 1135

Table I. Patient data

Atrial stimulus-R interval (msec) Cardio- A VD

Patient Age (yr) Sex pathy PR (msec) QRS (msec) Pacemaker (msec) Rest Exercise

1 65 F No 190 90 M Elite 7075 150 190 180 2 64 F ARVD 170 85 M Elite 7075 130 170 140 3 64 F No 200 90 T Meta DDDR 150 200 230 4 60 M DCM 200 90 M Elite 7075 190 230 260 5 70 M No 180 85 M Elite 7075 125 180 170 6 66 F No 180 IRBBB/115 M Elite 7075 150 180 170 7 34 M ASD 170 IRBBB/110 SP AFP 2020T 125 170 180 8 72 M No 180 85 SP AFP 2020T 165 190 210 9 60 F No 180 90 SP AFP 2020T 125 180 175

10 65 F No 160 85 M Elite 7075 120 160 150 11 63 M No 180 90 M Elite 7075 130 180 170

Mean_+ SD 62_+3 180_+ 12 142_+ 21 184-- 18 186+_63

ARVD, Arrhythmogenic right ventricular disease; ASD, atrial septal defect; AVD, atrioventricular delay; DCM, dilated cardiomyopathy; IRBBB, incom- plete RBBB; M, Medtronic, Minneapolis, Minn.; SP, Siemens Pacesetter, Sylmar, Calif. T, Telectronics, Englewood, Colo.

RESULTS Patient data (Table I). Eleven patients (6 women

and 5 men, mean age 62 _+ 3 years) were included in this study. Eight patients had an apparently normal heart. The other 3 patients had organic heart disease: dilated cardiomyopathy (LVEF 45 % ) in 1, surgically corrected atrial septal defect in 1, and a mild form of right ventricular dysplasia in 1. The latter 2 patients had incomplete right bundle branch block, slightly widening the QRS complex (100 and 110 msec). There was no case of complete or incomplete left bundle branch block (LBBB).

The implanted pacemakers were DDDR units, programmable in AAIR, VVIR, and DDDR modes. In the AAIR mode, the average st imulus-R interval was 184 + 18 msec at rest, without significant modifica- tion at peak exercise (186 _+ 63 msec). In three patients the st imulus-R interval paradoxically lengthened during exercise (210, 230, and 260 msec); these patients had the longest st imulus-R interval at rest (190,200, and 230 msec). In the DDDR mode, the mean programmed AV delay was 142 + 24 (range 125 to 190) msec. Echo-Doppler study of the transmittal blood flow showed in all cases the presence of a cor- rectly synchronized A wave. In the VVIR mode, 1:1 ventriculoatrial conduction was observed in nine pa- tients at rest and in eight during exercise.

Hemodynamic study (Table II) RAP. Analysis of variance revealed a significant

difference among the three modes (p < 0.01), without significant difference between the AAI mode (3.8 + 1.8 mm Hg) and the DDD mode (4.3 _+ 2.8 mm Hg) at rest. The frequent presence of "canon" A waves resulting from 1:1 ventriculoatrial conduction

explained the significantly higher average value in the VVI mode (7.2 +_ 3 mm Hg). During exercise the three modes were significantly different (p < 0.001): mean RAP was significantly lower in the AAI mode (8.4 _+ 2.7 mm Hg) with an average difference of 27 % compared with the DDD mode (11.5 _+ 4.1 mm Hg; p < 0.01) and of 45% compared with the VVI (15.4 +_ 4.3 mm Hg; p < 0.01) (Fig. 1).

PAP. No significant difference was observed among the three pacing modes at rest. The difference was significant during exercise (p < 0.03): mean PAP was significantly lower in the AAI mode (26.7 +_ 7.8 mm Hg) than in DDD (31.6 _+ 7.1 mm Hg; +12%) and than in VVI (37 _+ 10.9 mm Hg; +28%).

PC WP. The three pacing modes were significantly different at rest and during exercise (p < 0.01). At rest, mean PCWP was the lowest in AAI mode (7.6 + 1.8 mm Hg) compared with DDD (9.2 _+ 3.3 mm Hg; p < 0 . 0 5 ) and VVI (11.6_+3.5 mm Hg; p < 0.01). The benefit became much more important during exercise, with an average decrease of 17% compared with DDD (p < 0.01) and of 30% com- pared with VVI (p < 0.01) (Fig. 2).

BP. No significant difference was observed among the three pacing modes at rest or during exercise.

CO. A significant difference was found among the three pacing modes at rest and during exercise (p < 0.01). At rest CO was significantly higher in the AAImode (6.6 _+ 1 .3L/min) thanintheDDD (6 _+ 0.9 L/min; p = 0.03) and VVI (5 _ 1 L/rain; p < 0.01) modes. At the end of exercise, the average increase produced by the AAI mode was 12% (p < 0.01) in comparison with the DDD mode and 30% (p < 0.01) in comparison with the VVI mode (Fig. 3).

June 1995 1136 Leclercq et al. American Heart Journal

Table II. Hemodynamic data at rest and at end of 60 W plateau

CO (L/min) CI (L/min) RAP (mm Hg) PAP (ram Hg)

AAI DDD VVI AAI DDD VVI AAI DDD VVI AAI DDD VVI

Rest Mean _+ SD 6.6 6 5 3.85 3.55 2.9

1.3 0.9 1 0.8 0 0.65 p <0.05 <0.01 <0.05 <0.01

p, AAI vs VVI <0.01 <0.01 Exercise

Mean _+ SD 13.5 12.1 10.4 7.8 7 6.2 2.1 2.2 2.1 1 1 1.35

p <0.01 <0.01 <0.01 <0.05 p, AAI vs VVI <0.01 <0.01

3.8 4.3 7.2 14.6 16 20 1.8 2.8 3 3.8 3.6 6.9

NS <0.01 NS NS <0.01 NS

8.4 11.5 15.4 26.7 31.6 37 2.7 4.1 4.3 7.8 7.1 10.9

<0.01 <0.01 <0.05 NS <0.01 <0.02

Table I I I . Radionuclide data at rest and at end of 60 W plateau

LVEF (%) Septal LVEF (%) Inferior LVEF (%) Lateral LVEF (%) PFR EDV/S

AAI DDD VVI AAI DDD VVI AAI DDD VVI AAI DDD VVI AAI DDD VVI

Rest Mean _+ SD

P p, AAI vs VVI

Exercise Mean + SD

P p, AAI vs VVI

61 58 52.5 66.5 58 51 67 65 67 69 69 73 2.1 1.61 1.7 7 6.5 8 11 11 8 12.5 11 10 12 14 11 0.2 0.3 0.3 <0.05 <0.01 <0.01 NS NS NS NS NS <0.05 NS

<0.01 <0.01 NS NS <0.01

65 60 55.5 70.5 63 59 70.6 66.4 72.4 74 70 75 4.3 3.7 3.6 6 4 7 14 12 15 6.8 4 8.6 13 9.5 9.5 0.6 0.8 0.8 <0.05 <0.01 <0.03 NS NS NS NS NS <0.05 NS

<0.01 <0.03 NS NS <0.01

EDV/S, End-diastolic volume per second.

LVSWI . The three pacing modes were signifi- cantly different only during exercise (p < 0.015). The AAI mode produced a significant increase in LVSWI only during exercise (p < 0.01 vs DDD and vs VVI).

SVR. SVR values were significantly different for the three pacing modes at rest and during exercise (p < 0.02). SVR was significantly lower in the AAI mode at rest (p < 0.01 vs DDD and vs VVI) and dur- ing exercise (p < 0.01 vs DDD and vs VVI) (Fig. 4).

Radionuclide study (Table III) L VEF. L V E F could be ca l cu la t ed in all s i t u a t i o n s

in 10 pa t i en t s . S ign i f i can t d i f fe rences were f o u n d a m o n g the t h r ee pac ing modes a t res t a n d d u r i n g ex- ercise (p < 0.05 a n d p < 0.01, respect ive ly) . A t rest , L V E F was s ign i f i can t ly g rea te r in the AAI m o d e t h a n in D D D (p < 0.05) a n d VVI (p < 0.01) modes . T h e bene f i t was con f i rmed d u r i n g exercise (p < 0.01). T h e D D D m o d e p r o d u c e d s ign i f i can t ly h igher va lues t h a n d id the VVI m o d e a t res t a n d d u r i n g exercise.

Reg iona l wall mot ion . Reg iona l L V E F s could be ca l cu l a t ed in on ly seven pa t i en t s . I n the o the r four p a t i e n t s d a t a were n o t re l iab le e n o u g h to be ana lyzed .

The inferior and lateral LVEFs were similar with the three pacing modes at rest and during exercise. The septal LVEF was greater with AAI compared with DDD and VVI at rest (p < 0.01) and during exercise (p < 0.03). No statistically significant difference was found between DDD and VVI.

L V filling. The three pacing modes were signifi- cantly different (p <'0.01); LV filling assessed by the PFR was significantly improved in the AAI mode compared with DDD and VVI at rest and duringex- ercise (p < 0.01). Differences between DDD and VVI were not statistically significant.

DISCUSSION Hemodynamic consequences of asynchronous ven-

tricular contraction. By de l ay ing LV a c t i v a t i o n a n d i n v e r t i n g v e n t r i c u l a r d e p o l a r i z a t i o n sequence , pac- ing the r igh t ven t r i c l e in the apex i nduc e s a synchro - nous v e n t r i c u l a r c o n t r a c t i o n a n d re laxa t ion . T h e re- s u l t is a s ign i f ican t mod i f i ca t i on in LV systol ic func- t ion , as d e m o n s t r a t e d wi th e x p e r i m e n t a l mode l s as ear ly as 1925,1°-13 a n d in end-sys to l i c r e l axa t i on a n d

Volume 129, Number 6 American Heart Journal Leclercq et al. 1137

PCWP (mm Hg) SVR (Wood U) LVSWI (gm/rn 2) BP (mm Hg)

AAI DDD VVI AAI DDD VVI AAI DDD VVI AAI DDD VVI

7.6 9.2 11.6 14.4 16 19 1.8 3.3 3.5 2.3 2.6 3

<0.05 NS <0.01 <0.01 <0.01 <0.01

15.4 19.6 25.8 8.7 9.6 10.6 4.5 5 7 1.1 1.6 1.6

<0.01 <0.05 <0.01 <0.03 <0.01 <0.01

63 63 53 98 100 100 23 12 15 15 12 11

NS NS NS NS NS NS

102 88.5 72.3 125 125 123 22 19 24 12 10 14

<0.01 <0.01 NS NS <0.01 NS

ventricular filling. 14-16 Similarly, myocardial oxygen consumption is reduced by 18% on average and car- diac efficiency is improved by 41% in AAI compared with rate-equivalent DDD pacing. 17 Finally, histo- pathologic studies conducted in mature or immature dogs have shown that chronic right apical ventricu- lar pacing can lead to cellular lesions of the myocar- dium, including myofibrillar rearrangements, mitch- ondrial disorganization, prominent Purkinje cells, and dystrophic myocardial calcification, ~s, 19 which might alter performance or perhaps lead to long-term impairment of LV function in the "pacing-induced cardiomyopathies."19

In the same way, studies of patients with intermit- tent LBBB 2°22 have shown that asynchronous con- traction and relaxation significantly decrease LV contractility indexes (rate of change in pressure [dP/ dt]) and pump function (mean decrease of LVEF by 10%). These modifications result from (1) an alter- ation of septal motion (septal LVEF reduced by 40 % on average) 22 without significant modification in the other walls and (2) an increased ventricle relaxation time, which impairs the quality of ventricular filling. However, the mechanical consequences of LBBB may not be exactly the same as the effects of apical ventricular pacing, simply because the activation se- quence is different within the LV. 23

Relative importance of ventricular activation sequence and AV synchrony in cardiac pacing, The earliest stud- ies comparing the effects of atrial pacing and rate- equivalent sequential AV pacing (with complete ventricular capture) did not find any significant dif- ference in CO and end-diastolic LV pressure in ani- mal models 3 or in human beings. 24 The authors con- cluded that the deleterious effects of asynochronous ventricular activation were negligible as long as nor- mal AV synchrony was preserved. This classic hy- pothesis was questioned, however, in the mid-1980s.

Indeed, other studies 5,14 showed that compared with the AAI mode, DDD pacing significantly lowered dP/dt as well as stroke volume and LVEF (mean stroke volume decrease of 15% in the study by Askenazi et al.).5 Other studies showed that end-sys- tolic relaxation and ventricular filling also were modified. 2527 These relaxation and filling abnormal- ities were particularly important in patients with coronary artery d i s ea se s

Not until the past few years have dynamic studies shown the effects of DDD-DDDR pacing at rest and during exercise in patients implanted on a long-term basis for isolated sinus node dysfunctionfi 7 Provided normal AV and intraventricular conduction were present, it was possible to compare the effects of AAI and DDD or even VVI pacing in optimal condi- tionsfi 7 However, methodologic biais can easily oc- cur in this type of study and must be prevented to obtain reliable data. In VVI and DDD modes, yen- tricular capture must be complete to induce truly asynchronous LV activation. This requirement im- plies that the pacing rate must be rapid enough to prevent fusion beats, and sinus recapture in VVI mode and the AV delay must be sufficiently short in the DDD mode because the PR interval shortens in- versely with increasing heart rate during exercise. 2s Consequently the AV delay must be determined in- dividually so it can be short enough to guarantee complete ventricular capture not only at rest but also at peak exercise. In our study this precaution was taken in each patient during the training exercise test in the DDD mode.

Another important problem in method is the qual- ity of AV synchrony in the AAI-AAIR and in DDD- DDDR modes. In AAI mode, the deleterious hemo- dynamic effects of excessively long stimulus-R in- tervals are well known. 29 They are similar to the effects of long spontaneous PR intervals in nonpaced

June 1995 1 1 3 8 ~ "l~eclercq e~ G~. American Head Journal

16' RAP (mmHg)14.

12'

10"

** 8" 7,2+3 +

6"

4,3:~2,8 i 3,8+1,8 41

2

15,4±4,3"*

11,5+.4,1"*

8,4±2,7

**:p<O,Ol

AAI ; ODD ; WI

- . , • , • , • . , • . , T I M E ( M N ) 2 4 6 8 10

F ig . 1. Mean RAP: comparison of three pacing modes (AAI, DDD, and VVI) at rest and during exercise. Arrow, Beginning of 60 W plateau. MN, Minutes.

14- CO (I/ran) ,

12"

1 0 - ~ 8

8,65±1,35 6+0,9" 6 ~ 5+1

~ 13,5±2,1.,

12,1±2,2

10,4±2,1

*:p<O,05 **:p<O,Ol

----o---* AAI = DDD

" .~ VVl

. . . . 4 9 " . . . . . 8' • T I M E ( m n )

F i g . 3 . MeanCO:comparisonofthreepacingmodes(AAI, DDD, and VVD at rest and during exercise. Arrow, Begin- ning of 60 W plateau. MN, Minutes.

29' PCWP * *

(ram Hg; ~ 21,3±s,3 24.

19' 18+3 **

15±4,5 14 ; *: p<O,OS

• *: p<O,Ol 11,6+3,5 ~ AAI 9,2+3,3 9' - DDD / 7,6+1,8 • z VW

4 . . . . . , . . . . T I M E ( r a n ) o ~ 4 ; ;

F i g . 2 , M e a n P C W P : c o m p a r i s o n o f t h r e e p a c i n g m o d e s (AAI, DDD, and VVI) at rest and during exercise. Arrow, Beginning of 60 W plateau. MN, Minutes.

S V R 20'

19+3

16+2,6"* 16"

14,4±2,3

12

~ - 60W ~ = AAI

: v v7 10,5-+1,6 *

.... • ,- - , ' , , " . , T I M E ( r a n ) 4 6 8 10

F ig . 4, Mean SVR: comparison of three pacing modes (AAI, DDD, and VVI) at rest and during exercise. Arrow, Beginning of 60 W plateau. MN, Minutes.

patients. 3° In both cases, restoring optimal AV syn- chrony by DDD pacing significantly improves cardiac performance despite asynchronous ventricular acti- vation. In AAIR mode, approximately 30% of pa- tients have a paradoxical lengthening of the stimu- lus-R interval during exercisea; consequently the P wave may occur after or within the R wave of the preceding cycle. Thus the contribution of the atrial systole is completely lost because the atria contract against closed AV valves. In the study by Rosenqvist et al. 6 the PR interval in sinus rhythm was >200 msec at rest in 8 of 12 patients; the st imulus-R interval in AAI mode was 248 + 57 (range 200 to 320) msec at rest. These mainly pathologic values may have min- imized the relative benefit of the AAI mode in that study. In our study only 1 patient had a relatively long PR interval (200 reset) in sinus rhythm. The st imulus-R interval during AAIR pacing was moder- ately lengthened in 2 patients at rest and in 3 during exercise.

It has been demonstrated clearly that in DDD- DDDR pacing AV delays should be optimized indi- vidually and that there is a difference of 70 msec on average between the optimal AV delay in sensed atrial cycles and the optimal AV delay in paced atrial cycles. 32 To prevent interferences between sensed and paced atrial cycles, it is fundamental in this type of study to impose a permanently paced atrial rhythm. This condition appears to have been met in the Rosenqvist et al. study s because the pacing rate at rest and during exercise was 10 beats/min greater than the intrinsic heart rate. However, there was a wide interindividual variability at rest (range 60 to 100 beats/rain) and during exercise (range 80 to 20 beats/min). Because of this variability it is somewhat difficult to interpret some results, notably the Dop- pler-echocardiographic and radionuclide data. In our study we selected patients with severe ACI. This cri- terion allowed us to obtain permanent atrial capture and to program identical pacing rates (i.e., 70 beats/

Volume 129, Number 6

American Heart Journal Leclercq et al. 1139

min at rest and 120 beats/min at peak exercise) in all patients, thus facilitating data analysis. Nevertheless the AV delay could not be optimized in any of these three studies. However, this criticism of method ap- pears to be secondary in our study because the mean programmed AV delay (142 msec) lies within the usual range of optimal AV delays measured in healthy hearts. In addition, neither very short delays (mini- mum 125 msec) nor very long delays (maximum 190 msec) were programmed.

These methodologic problems probably explain some discrepancies between our results and those previously published, especially those from the study by Rosenqvist et al., 6 which was based on two inva- sive methods: Doppler echocardiography with aortic flow velocity measurements at rest and during low- level exercise testing in the supine position and radi- onuclide angiography performed only at rest. The Doppler-echocardiographic study showed a signifi- cant increase in peak aortic blood velocity (10 % on average), in mean aortic blood acceleration (20 % on average), and in systolic ventricular time integral (14 % on average) during AAI pacing compared with DDD, at rest and during exercise (13%, 16%, and 17 %, respectively). However, there was no signifi- cant difference between the DDD and the VVI modes and for LV systolic function indexes (global and re- gional ejection fractions) and for diastolic function (PFR).

These observations are in disagreement with those of our radionuclide study because the mean values for each of these parameters were significantly higher in DDD mode than in VVI mode at rest and during ex- ercise. These conflicting results may be explained by the lack of strict adherence to AV synchrony in the Rosenqvist et al. G study. Another more likely expla- nation would be poorer tolerance to VVI pacing in patients selected in our study. Rosenqvist et al. 6 ob- served 1:1 ventriculoatrial conduction in only one third of patients, whereas in our study 1:1 conduction was found in 9 of 11 patients at rest and in 8 of 11 during exercise. This high prevalence is not surpris- ing: it is a usual characteristic of patients with isolated SND, especially during exercise. 33 Very poor hemodynamic tolerance also is known in this situa- tion. 34

The same favorable differences between AAI and DDD modes and between DDD and VVI modes were found in the hemodynamic study. Indeed the hemo- dynamic study produced functional data that prob- ably better correlate with patient symptoms than would the data furnished by the other investigational methods cited earlier, s, 7 For all of the parameters

studied but in particular for cardiac filling pressures (RAP and PCWP), the AAI mode was significantly more efficient than the DDD mode, and the DDD mode was better tolerated than the VVI mode. The differences were proportionally greater during exer- cise than at rest.

Clinical implications. Except in hypertrophic ob- structive cardiomyopathy ~5, 36 and possibly in dilated cardiomyopathyS in which completed ventricular capture and short AV delay are required to obtain optimal hemodynamic benefit, it appears that it is important to preserve a normal ventricular activa- tion sequence during permanent cardiac pacing when- ever possible in all patients with permanently (or at least most often) normal AV and intraventricular conduction. This requirement would affect almost all patients receiving implantations for isolated sinus node dysfunction 3s (35% to 50% of the indications for permanent cardiac pacing 39) or for the carotid si- nus syndromes and the malignant vasovagal syn- dromes or even some paroxysmal AV blocks with narrow QRS.

Several technical solutions may preserve optimal AV synchrony and normal ventricular activation se- quence. Most require the correct choice of pacemaker with an individualized pacing mode and program- ming. In some cases this objective can be achieved with a simple VVI pacemaker "intelligently" pro- grammed, that is, with an escape rate (hysteresis on) and a pacing rate low enough to avoid interference between the spontaneous sinus rhythm and the paced rhythm. In the case of sinus node dysfunction, the ideal mode is of course AAI, a mode that is un- derused ~s for no logical reason. The risk of secondary AV block is low, 0.7% to 3% per year, 35 and the blocks that do occur are always slowly progressive low-grade nodal blocks that present no great danger. Nevertheless this low risk could be a justification for preferential use of DDD pacemakers although it would be logical to program them in AAI mode after implantation. This approach is much more logical than programming the DDD or DDI mode with a very long AV delay to preserve intrinsic conduction; that type of programming would induce a major length- ening of the pacemaker's total atrial refractory pe- riod and could result in aberrant behavior during ex- ercise. Another elegant technical solution now avail- able with modern DDD units is automatic mode switching from AAI, as long as the AV conduction remains within normal limits, to DDD when the AV interval overruns a threshold level or when a non- conducted P wave is sensed.

Another route for research involves the site of ven-

June 1995 1140 Leclercq et al. American Heart Journal

tricular pacing rather than the pacemaker itself or the pacing mode. For example, in selected patients without distal conduction impairment, the distal His bundle could be stimulated by implanting a screw-in lead within the proximal septum above the tricuspid annulus. Promising results have been reported in an- imal experiments with a narrow paced QRS com- plex. 39

Conclusions. This study confirms the hemody- namic importance of preservation of normal ventric- ular sequence in all patients in whom it is technically compatible with permanent cardiac pacing. This finding implies that in sinus rhythm, AV and intra- Ventricular conduction must be constantly (or at least most often) normal. Several technical solutions are already available to reach this objective, includ- ing wider use of single-chamber atrial pacing modes or the algorithms for automatic switching between AAI and DDD modes when abnormal AV conduction is detected. Other areas of research to be investigated include more physiologic stimulation sites to recreate normal activation and thus synchronous ventricular contraction and relaxation.

REFERENCES

1. Vassalo J, Cassidy D, Miller J, Buxton A, Marchlinsky F, Josephson M. Left ventricular endocardial activation during right ventricular pacing: effect of underlying heart disease. J Am Coll Cardiol 1986;7:1228-33.

2. Samet P, Castillo C, Bernstein W. Hemodynamic sequelae of atrial, ventricular and sequential atrioventricular pacing. AM HEART J 1966;76:725-9.

3. Dagget W, Bianco J, Powell W, Austen W. Relative contributions of the atrial systole-ventricular systole and of patterns of ventricular activa- tion to ventricular function during electrical pacing of the dog heart. Circ Res 1970;27:69-79.

4. Greenberg B, Chatterjee K, Parmley W, Werner J, Holly A. The influ- ence of left ventricular filling pressure on atrial contribution to cardiac output. AM HEART J 1979;98:742-51.

5. Askenazi J, Alexander J, Koenigsberg D, Belic N, Lesch M. Alteration of left ventricular performance by left bundle branch block simulated with AV sequential pacing. J Am Coll Cardiol 1984;53:99-104.

6. Rosenqvist M, Isaaz K, Botvinick E, Dae M, Cockrell J, Abbot J, Schiller H, Griffin J. Relative importance of activation sequence compared to atrioventricular synchrony in left ventricular function. Am J Cardiol 1991;67:148-56.

7. Harper G, Pina I, Kutalek S. Intrinsic conduction maximizes cardio- pulmonary performance in patients with dual chamber pacemakers. PACE 1991;14:1787-91.

8. Smith T, Richards P. A simple kit for the preparation of 99m-Tc labeled red blood cells. J Nucl Med 1976;17:126-32.

9. Links J, Becker L, Shindlecker G, Guzman P, Barow R, Nickoloff E. Measurement of absolute left ventricle volume from gated blood pool studies. Circulation 1982;65:82-90.

10. Wiggers C. The muscular reactions of the mammalian ventricles to ar- tificial surface stimuli. Am J Physiol 1925;73:C346-78.

11. Finney J. Hemodynamic alterations in left ventricular function conse- quent to ventricular pacing. Am J Physiol 1965;208:H275-82.

12. Badke F, Boinay P, Covell J. Effects of ventricular pacing on regional left ventricular performance in the dog. Am J Physiol 1980;238:H858- 67.

13. Waltson A, Starr J, Greenfield J. Effect of different epicardial ventric-

ular pacing sites on left ventricular function in awake dogs. Am J Car- diol 1973;32:291-4.

14. Grover M, Glantz SA. Endocardial pacing sites affects left ventricular end-diastolic volume and performance in the intact anesthetized dog. Circ Res 1983;53:72-85.

15. Park RC, Little LC, O'Rourke RA. Effect of alteration of left ventric- ular activation sequence on the left ventricular end-systolic pressure- volume relation in closed chest dogs. Circ Res 1985;57:706-17.

16. Litwin W, Gorman G, Huang S. Effects of different pacing modes on left ventricular relaxation in closed-chest dogs. PACE 1989;12:1070-6.

17. Bailer D, Wolpers HG, Zipfel J, Brestscheider HJ, Hetlige G. Compar- ison of the effects of right atrial, right ventricular apex and atrioven- tricular sequential pacing on myocardial oxygen consumption and car- diac efficiency: a laboratory investigation. PACE 1988;11:394-403.

18. Adomian GE, Beazell J. Myofibrillar dissaray produced in normal hearts by electrical cardiac pacing. AM HEART J 1986;112:79-84.

19. Karpawich P, Justice C, Cavitt D, Chang C. Developmental sequelae of fixed-rate ventricular pacing in the immature canine heart. An electro- physiologic, hemodynamic and histopathologic evaluation. J Am Coll Cardiol 1990;119:1077-82.

20. Takeshita A, Basra L, Kioschos M. Effect of intermittent left bundle branch block on left ventricular performance. Am J Med 1974;56:251-5.

21. BramIet D, Morris K, Coleman E, Albert D, Cobb F. Effects of rate-de- pendent left bundle branch block on global and regional left ventricu- lar function. Circulation 1983;67:1059-65.

22. Grines C, Bashore T, Boudoulas H, Olson S, Shafer P, Wooley C. Func- tional abnormalities in isolated left bundle branch block: the effect of interventricular asynchrony. Circulation 1989;79:845-53.

23. Xiao HB, Brecker SJD, Gibson DG. Differing effects of right ventricu- lar pacing and left bundle branch block on left ventricular function. Br Heart J 1993;63:166-73.

24. Sheffer A, Rosenmab Y, Ben David Y, Flugelman M, Gotsman M, Lewis B. Left ventricular function during physiological pacing: relation to rate, pacing mode and underlying myocardial disease. PACE 1987; 10:315-25.

25. Zile MR, Blaunstein AS, Shimizu G, Gaasch WH. Right ventricular pacing reduces the rate of left ventricular relaxation and filling. J Am Coll Cardiol 1987;10:702-9.

26. Bedotto JB, Grayburn PA, Black WH. Alterations in left ventricular pacing in humans. J Am Coll Cardiol 1990;15:658-64.

27. Betocchi S, Piscione F, Villari B. Effects of induced asynchrony on left ventricular diastolic function in patients with coronary artery disease. J Am Coll Cardiol 1993;21:1124-31.

28. Daubert C, Ritter P, Mabo P, Ollitranlt J, Descaves C, Gouffault J. Physiological relationship between AV interval and heart rate in healthy subjects: applications to dual chamber pacing. PACE 1986;9: 1032-9.

29. Jutzy R, Feenstra L, Pal R, Florio J, Bansal R, Levine P. Comparison of intrinsic versus paced ventricular function. PACE 1992;15:1919-22.

30. Mabo, Cazeau S, Forrer A, Varin C, De Place C, Paillard F, Daubert C. Isolated long PR interval as only indication of permanent DDD pacing [Abstract]. J Am Coll Cardiol 1992;19:66A.

31. Mabo P, Pouillot C, Kermarrec A, Lelong B, Le Breton H, Daubert C. Lack of physiological adaptation of the atrioventricular interval to heart rate in patients chronically paced in the AAIR mode. PACE 1991; 14:2133-42.

32. Daubert C, Ritter P, Mabo P, Varin C, Leclercq C. AV delay optimiza- tion in DDD and DDDR pacing. In: Barold S, Mugica J, eds. New per- spectives in cardiac pacing. Vol. 3. New York: Futura, 1992:259-87.

33. Cazeau S, Daubert C, Mabo P, Ritter P, Lelong B, Pouillot C, Paillard F. Dynamic electrophysiology of ventriculoatriaI conduction: implica- tions for DDD and DDDR pacing. PACE 1990;13:1646-55.

34. Danbert C, Roussel A, Langella B, De Place C, Besson C, Gouffault J. Etude h~modynamique et echocardiographique TM des consequences de la conduction ventriculo-auriculaire chez l'homme. Arch Mal Coeur 1984;77:413-25.

35. Sutton R, Kenny R. The natural history of sick sinus syndrome. PACE 1986;9:1110-4.

36. Feruglio G, Rickards A, Steinbach K, Feldman S, Parsonnet V. Cardiac pacing in the world; a survey of the state of the art in 1986. PACE 1987;10:768-77.

Volume 129, Number 6

American Heart Journal Limas et al.

37. Ryden L, Karlsson O. The importance of different AV intervals for ex- ercise capacity. PACE 1988;11:1051-62.

38. Rosenqvist M, Obel P. Atrial pacing and the risk for AV block: is there a time for change in attitude? PACE 1989;12:97-101.

39. Karpawich P, Gates J, Stokes K. Septal His-Purkinje ventricular pac- ing in canines: a new endocardial electrode approach. PACE 1992; 15:2011-5.

Possible involvement of the HLA-DQB1 gene in susceptibility and resistance to human dilated cardiomyopathy

There is evidence that autoimmunity plays a role in the pathogenesis of dilated cardiomyopathy and that susceptibility to the disease is related to products of human leukocyte antigens (HLA) class II genes. We compared the distribution of HLA-DQA1 and -DQB1 alleles and haplotypes in 44 normal controls and 34 patients with idiopathic dilated cardiomyopathy patients. The distribution of two DQA1-DQB1 haplotypes ('0102-'0604 and "0102-'0501) were more frequent in the patients. Histidine at position 30 of the HLA-DQB1 gene was associated with disease (62% of patients compared to 36% of controls), whereas homozygosity for leucine at position 26 was more frequent in controls (36% vs 18% of patients). There was no correlation between HLA-DQA1-DQB1 haplotypes and the presence of anti-/3-receptor antibodies. These results suggest that the HLA-DQB1 gene is involved in the pathogenesis of human dilated cardiomyopathy. (AM HEART J 1995;129:1141-4.)

Constantinos J. Limas, MD, Catherine Limas, MD, Irvin F. Goldenberg, MD, and

Robert Blair, BS Minneapolis, Minn.

Despite considerable effort, the pathogenesis of idio- pathic dilated cardiomyopathy remains poorly de- fined. In recent years, increasing attention has been given to abnormalities in immune function that sug- gest an autoimmune cause. These abnormalities in- volve both humoral and cellular responses, but most of the studies have focused on the presence of a va- riety of autoantibodies, some of which may have functional implications. 1-4 It is uncertain, however, whether these autoantibodies are only markers for an ongoing autoreactive immune process or whether they also contribute to the initiation of the disease.

It is very likely that both genetic and environmen- tal factors are involved in the pathogenesis of dilated

From the Cardiovascular Division, Department of Medicine, Department of Laboratory Medicine and Pathology, University of Minnesota School of Medicine; the Department of Veterans Affairs Medical Center; and the Minneapolis Heart Institute.

Received for publication June 20, 1994; accepted Oct. 3, 1994.

Reprint requests: Constantinos J. Limas, MD, Department of Medicine, University of Minnesota Medical School, Box 19, UMHC, 420 Delaware St. SE, Minneapolis, MN 55455.

Copyright © 1995 by Mosby-Year Book, Inc. 0002-8703/95/$3.00 + 0 4/1/61984

cardiomyopathy. In addition to the well-recognized familial forms of the disease, it is estimated that a substantial subset of presumed sporadic (nonfamil- ial) cases may actually have a strong genetic compo- nent. 5 Genetic factors may also play a permissive role in the expression of the immune dysregulation.

We 1 and others 6 have previously noted an associ- ation between components of the HLA class II system and dilated cardiomyopathy. In particular, HLA-DR and -DQ gene products appear to be related to the propensity to develop both autoantibodies and clinically manifested disease. In the present study, we extend these observations to a comparison of the HLA-DQA1 and -DQB1 alleles in dilated cardiomy- opathy patients and normal controls. We find that the distribution of the HLA-DQ A1-DQB1 haplo- types differs in this disease and that the presence of histidine at position 30 of the HLA-DQB1 may con- tribute to disease susceptibility.

METHODS

Studies were carried out with 34 dilated cardiomyopathy patients (25 men and 9 women), aged 26 to 63 (mean 46 +_ 7 years). The diagnosis was made according to World Health

1141