Role of Autonomic Reflexes in Syncope Associated with Paroxysmal Atrial Fibrillation

JACC Vol . 22, No. 4 1123 October 1993 :1123-9

Role of Autonomic Reflexes in Syncope Associated With Paroxysmal Atrial Fibrillation

MICHELE BRIGNOLE, MD, FESC, LORELLA GIANFRANCHI, MD, CARLO MENOZZI, MD,* ANTONIO RAVIELE, MD,t DANIELE ODDONE, MD, GINO LOLLI, MD,* NICOLA BOTTONI, MD* Lat'agna, Reggio Ernilia anmid Mestre, Italy

Objectives, The purpose of this study was to evaluate the role of and in 2 (12%) of 16 control subjects (p = 0.0004) . The Induction autonomic reflexes in the genesis of syncope associated with the of atrial fibrillation in the upright position eligited syncope In 16 onset of paroxysmal atrial fibrillation . (42%) of 38 patients but in none of 16 control subjects (p = 0 .001). Background. Syncope associated with paroxysmal atrial fibril- At The beginning of atrial fibrillation, systolic blood pressure was lation has been Interpreted as an ominous finding predictive of lower in patients than in control subjects (88 t 32 vs . 127 t rapid ventricular rates. However, various mechanisms may be 32 aim Hg), whereas mean heart rate was similar (142 t 35 vs . Involved when heart rate is not particularly high . 134 t 25 beats/mint) . The correlation between heart rate and Methods. Forty patients (age 60 s: 14 years, 21) men, 20 systolic blood pressure was weak (r = 0 .35), and In five patients women) with syncope and atrial fibrillation were compared with syncope occurred at a heart rate !5130 beats/min. At the time of 16 control subjects (age 66 t 13 years, 8 men, 8 women) with syncope, heart rate decreased (-12 t 21 beats/min) in patients atrial fibrillation without syncope. Carotid sinus massage and with induced syncope, whereas it remained unchanged in patients head-up tilt testing (at 60° for 60 min at baseline and during without indu'. -.4 ,scope (+1 ± 17 beats/min, p = 0.04) or slightly isoproterenol infusion) were performed during sinus rhythm . A increased in control subjects (+9 t 21 beats/min, p = 0 .009). positive response was defined as the induction of syncope . Atria] Conclusions . Patients with syncope associated with paroxysmal fibrillation was also induced on a tilt table at 60° by means of short atrial fibrillation are predisposed to an abnormal neural response bursts of atrial pacing. during both sinus rhythm and arrhythmia . In some patients the Results. Results of carotid sinus massage were positive in 15 onset at atrial fibrillation triggers vasovagal syncope . (37%) of 40 patients but in no control subjects (p = 0 .002) . (J Am Coll Cardiol 1993;22 :1123-9) Head-up tilt test findings were positive in 25 (66%) of 38 patients

Syncope is an unusual clinical presentation of paroxysmal Nonhemodynamic mechanisms may be even more readly atrial fibrillation, and its cause remains uncertain . It has been suspected when heart rate is not particularly high . It is well claimed that a fast ventricular rate at the onset of atrial known that the onset of paroxysmal tachycardia produces a fibrillation is the main factor responsible for syncope as a hemodynamic disturbance that evokes a compensatory car- consequence of hemodynamic deterioration secondary to diovascular response (2,3) . This response, which is mediated the critical reduction in diastolic filling time . Fast ventricular mainly by reflex changes in autonomic nervous tone, is very rates are common during paroxysmal atrial fibrillation in often effective in restoring both blood pressure and cardiac patients affected by enhanced atrioventricular (AV) conduc- output during tachycardia to levels similar to those existing tion or accessory bypass tract . Nevertheless, Auricchio et before the onset of arrhythmia. Moreover, intense vagal al. (1) did not find any difference in heart rate among patients stimulation can cause atrial fibrillation (4,5), and it has with and without a history of syncope when atrial fibrillation recently been suggested that atrial fibrillation may result was induced during electrophysiologic study . They therefore from vagal stimulation occurring at the time of vasovagal suggested that other factors may play a role in the genesis of syncope (6) . In this study we therefore evaluated the role of the syncope . autonomic reflexes in the genesis of syncope associated with the onset of paroxysmal atrial fibrillation .

From the Laboratory of Electrophysiology and Pacing . Service of Cardi- ology, Ospedali Riuniti, Lavagna ; *Laboratory of Elecirophysiology and Methods Pacing, Service of Cardiology, Ospedale SM Nuova. Reggio Emiliu; and . 40 tDepartment of Cardiology, Ospedale Umberto I, Mestre. Italy. Selection of patients. From January 1991 to June 1992 Manuscript received December 8 . 1992; revised manuscript received patients affected by syncope associated with paroxysmal March 26.1993 . accepted March 31, 1993 . atrial fibrillation met the inclusion criteria and were enrolled Address for corresoondence: Michele Brignole. MD, FESC. Via A. Grilli 164,1-16041 Borzonasca (GE), Italy. in the study . They accounted for 6% of the overall popula-

0735.1097!931$6.00 01993 by the American College of Cardiology

1124 BRIGNOLE ET AL . JACC Vol. 22, No . 4 SYNCOPE AT THE ONSET OF ATRIAL FIBRILLATION October 1"3-1123-9

Table 1. Clinical Characteristics of the Study Groups between 8:00 AM and 1:00 PM . All cardioactive and vasoac- Patients Control Subjects tive drugs were withdrawn . Informed consent was obtained (n = 40) (n = 16) from all participants. The method of execution of carotid sinus massage and the head-up tilt test have been previously 60 ± 14 66± 13 Age In) . Male gender 2000) 800) described and validated (7-9) ECO abnormalities Carotid sinus massage was performed on both right and Overall 1600) 7 (44) left carotid sinuses in the supine and erect positions for 10 s . Sinus bradycardia 4 I Vasodepressor reflex was evaluated after atropine infusion Ist degree AV block 2 0 when indicated. A positive response was defined as the wave abnormalities 6 3 ST-T development of syncope in association with bradycardia or Q wave 3 0 hypotension, or both . Bundle branch block 2 4 Bclwcardlographk abnormalities 21(521 7 (44) The baseline head-up tilt test protocol consisted of a 60° Fractional shortening (%) 36 t 10 4018 tilt for 60 min or until the occurrence of syncope . If the test Structural heart disease did not induce syncope, the patient or control subject was Overall 25021 1109) lowered into the supine position for 30 min before under- Atherosclerotic 14 7 going the test during isoproterenol infusion . Previous myocardial infarction 3 0 Head-up tilt testing during isoproterenol infusion con- Valvviw S 4 sisted of a 60° tilt for 10 min repeated in successive stages (at Atrial seplal aneurysm 3 0 5-min intervals) with use of graded isoproterenol infusion t SD or number of subjects . AV Values presented are mean value ('%) rates of I to 5 pg/min. aidoventrkular : ECO . electrocardiographic. The upright induction of atria) fibrillation was attempted on a tilt table at 60° immediately after the upright position had been achieved for a maximum of 3 min . Short-duration tion of 672 patients referred to us for investigation of (<5 s) atrial bursts at rates ranging from 300 to 800 beats/min syncope from the emergency room, inpatient service and were delivered until atria) fibrillation L-30 s in duration was ambulatory program. induced. The duration of tilt before initiation of atrial fibril- To be included in the study, the patients had to fulfill the lation was 64 ± 47 s . When atrial fibrillation was induced, the following criteria: 1) unexpected syncope that was immedi- patient was maintained in the upright position for 2 min . If ately preceded and followed by palpitation; 2) documenta- syncope occurred, the patient was rapidly returned to the tion, within 6 h of the event, of paroxysmal atrial fibrillation supine position . Ducting induction procedures, beat to beat lasting <24 h ; 3) absence of electrocardiographic (ECO) finger arterial pressure was monitored continuously by or electrophysiologic evidence of enhanced AV conduction the Penile: volume-clamp method (10) using a finger cuff or accessory bypass tract ; 4) exclusion of other enuses for (Ohmeda Monitoring System) . The arm was supported with syncope on the basis of a standardized evaluation used in a mobile arm plate to keep the finger at heart level in the our departments that has been previously described (7); upright position. Blood pressure was recorded simulta- 5) absence of severe heart failure (New York Heart Associ- neously with the ECG on a paper recorder. For each patient, ation functional class I or II) and of recent acute illness . mean heart rate and systolic pressure were calculated for the The clinical characteristics of the patients are summa- initial 30 s from the onset of atrial fibrillation or until the dud in Table 1 . Syncope invariably occurred while the occurrence of syncope. Changes in mean finger arterial patient was standing or sitting, at rest or during mild activity. pressure, as measured by the Finapres technique, have been At the first examination after the syncopal attack, atrial shown to match closely changes in arterial pressure mea- fibrillation was well tolerated. No signs of cardiovascular sured by invasive techniques during syncope induced by collapse were detected, and no patient reported symptoms passive tilt testing or during induced tachycardia (11,12) . other than palpitation . The average heart rate was 140 ± The upright position was chosen for the induction of atrial 29 beatslmin, fibrillation for two reasons: 1) because spontaneous syncope COW group. The control group consisted of 16 patients always occurred in the sitting or standing position, and without syncope or presyncope who were affected by par- 2) because previous studies have demonstrated that syncope oxysmal atria) fibrillation (Table 1). The average heart rate due to tachyarrhythmia is more likely to be reproduced during the last episode was 144 ± 22 beatslmin . They were during electrophysiologic study in the upright rather than the recruited during the same period as the syncope group . supine position (13,14) . I ductIon IML Patients and control subjects underwent Statistical analysis. Continuous variables were compared carotid sinus massage, head-up tilt testing in the baseline using analysis of variance for paired and unpaired data . slate and during isoproterenol infusion and electrophysio- Frequencies were compared by means of the Fisher exact logic study, including the induction of atrial fibrillation in the test. The relation between heart rate and blood pressure was upright position. The tests were always performed in an examined by the Pearson correlation coefficient . Continuous isolated tow (when patieais were in the nonfasting state) variables were reported as mean value ± I SD . JACC Vol . 2-?. No. 4 BRIGNOLE ET AL. 1125 October 199, :1123-9 SYNCOPE AT THE ONSET OF ATRIAL FIBRILLATION

I able 2. Results of the Provocative Tests in Patients and Control Subjects Control Test Patients Subjects p Value Positive carotid sinus massage 15!401371 0/16 (0) 0.002 Positive baseline head-up tilt test* 17138145) 1116 (61 0.005 Positive isoproterenol head-up tilt test* 8/17 (44) 1115 0) 0.01 Overall positive head-up lilt test" 25138 (661 2/16 412.) 0.0004 Upright induction of AF Patients Ino.) 38 16 Syncope at the onset o1 AF 16/38(42) 0/1610) 0.001 Mean HR Ibeatslmin)t 142 . 35 134 t 25 NS HR distribution (beatshrin) s 130 16/38 (42) 611607) 131-179 1(/38 (0 811600) NS _e 180 6!38116) 2116112) Systolic blood pressure (mm Hg) Upright preinduction of AF 140 ± 19 162 ± 31 0.003 Upright AP 88 ± 32 127 ± 32 0.00005 *Two of 40 patients refused to undergo baseline head-up tilt testing, and 4 of 21 patients with a negative baseline head-up tilt test response refused to undergo isoproterenol head-up tilt testing. All orthese patients had had a positive response to carotid sinus massage . ')Mean heart ride and systolic blood pressure are calculated for the iuiliai 30s front the onset of airial fibrillation or until the occurrence of syncope . Values presented are mean value t SI) or number I%) of subjects. AF = atrial fibrillation ; HR = heart rate .

Results lations between heart rate and sy,,tolic blood pressure and between heart rate and syncope were weak (Fig . 2). Indeed, The clinical characteristics of the patients and control syncope also occurred in some patients with a low heart rate subjects were similar, except for the presence or absence of (Table 3). syncope (Table 1) . Heart rate was evaluated in study group patients with and Carotid sinus massage and head-up tilttesting. The results without induction of syncope and in control subjects (Table of the provocative tests are listed in Table 2. Positive 4, Fig. 1) . Discordant features are evident. In fact, heart rate responses to carotid sinus massage or head-up tilt testing decreased in patients with induced syncope, remained un- were recorded in 31 (77%) of 40 patients but only in 2 (12%) changed in patients without induced syncope and increased of 16 control subjects (p = 0.00005) . A vasodepressor reflex in control subjects . >50 mm Hg was present in all positive test results . During Positive responses to carotid sinus massage or head-up up carotid sinus massage, a cardioinhibitory reflex (defined as tilt testing were recorded in 14 (87%) of 16 patients in the ?3-s ventricular asystole) was present in 14 of 15 positive subgroup with induced syncope and in 15 (68%) of 22 tests, and during head-up tilt testing, a cardioinhibitory patients in the subgroup without induced syncope (difference reflex (defined as a decrease in heart rate to <40 beatslmin) statistically not significant, p = 0 .15). was also present in 7 of 17 positive test results . Syncope was induced in 24 ± 17 min during baseline tilt testing and in 22 10 min during isoproterenol tilt testing . Reproducible, short- duration atrial fibrillation occurred in one patient during Discussion carotid sinus massage . In a selected inpatient population, paroxysmal atrial Upright electrophysiologic study. Atrial fibrillation lasting fibrillation associated with syncope is probably not as rare as >30 s was induced in 38 of 40 patients with syncope and in expected . Indeed, we observed this feature in 6% of patients 16 of 16 control subjects. The onset of atrial fibrillation referred for syncope . The frequency would presumably be caused syncope in 16 study group patients (42%) but in no even higher if we took into account those cases in which control subjects. Syncope occurred in all subjects within 30 s atrial fibrillation disappeared before it could be demon- after the end of burst pacing (mean time 22 ± 7 s) . Mean strated . Admittedly, at the time of syncope, heart rhythm systolic blood pressure was lower in patients than in control was not monitored . Therefore it is possible, although un- subjects. By contrast, mean heart rate and heart rate distri- likely, that in some instances atrial fibrillation began not with bution were similar in the two groups (Table 2) . syncope but a short time before or after the attack . The 16 patients with induced syncope had lower values of Principal findings. Our results demonstrate that in the systolic blood pressure and higher values of heart rate than absence of enhanced AV conduction or an AV bypass tract, the remaining 22 patients who remained asymptomatic and syncope at the onset of paroxysmal atrial fibrillation is the control subjects (Table 3, Fig . 1). However, the corre- associated with a propensity toward a disturbance of the JACC Vol. 22. No. 4 1126 BRtONOLE ET AL. SYNCOPE AT THE ONSET OF ATRIAL FIBRILLATION October 1993:1123-9

Table 3. Comparison Between Study Group Patients With and Without Induction of Syncope During the Upright Electrophysiologic Study

Syncope No Syncope Test (n = 16) (n = 22) p Value

Positive carotid sinus massage 5/16(31) 8/22 (36) NS Positive baseline head-up tilt test 911500) 7121(32) NS (p = 0 .12) Positive isoproterenol head-up tilt test 2/5(40) 6/1100) NS Upright induction of AF Mean HR (bests/min) (*) 158 ± 40 131 ± 29 0.003 HR distribution (beatslmin) 5130 511601) 11/22(50) 131 to 179 6116 (37) 10122 (45) 0 .007 also 5116011 1122(s) Systolic blood pressure (mm Hg) Upright preinduction of AF 135 m 17 143 ± 19 NS Upright AF* 59 ± 15 108 x 23 0 .00000

*Mean heart rate and systolic blood pressure are calculated as in 'table 2 . Values presented are mean value ± SD or number (%) of subjects . Abbreviations as in Table 2 .

autonomic nervous system, and it is not directly related to also not particularly high or different from that observed in tachyarrhythmia . Indeed, syncope was induced by either control subjects . Admittedly, in those patients with atrial carotid sinus massage or head-up tilt testing in the majority fibrillation-induced syncope, the average heart rate was of our patients, whereas it was rarely observed in control significantly higher. Nevertheless, the correlation between subjects affected by paroxysmal atrial fibrillation without heart rate and systolic blood pressure or syncope was weak, syncope (Table 2) . and, in some cases, syncope occurred at heart rates that This association suggests that patients with paroxysmal were so low that syncope clearly could not be explained on atria) fibrillation presenting with syncope may be predis- a hemodynamic basis alone (Table 3, Fig . 2). It therefore posed to an abnormal neurally mediated response during seems likely that the hemodynamic consequences of the both sinus rhythm and tachycardia. If an abnormal neural onset of atrial fibrillation contribute to the initiation of an susceptibility is absent, as in the control subjects, atria) abnormal neural response and that some correlation exists fibrillation is invariably unable to cause syncope . between the magnitude of the initial hemodynamic changes The diagnostic value of carotid sinus massage and and the magnitude of the subsequent reflex responses. head-up tilt testing in patients affected by unexplained syn- Proposed pathophysiologic classification . On the basis of cope is well known . The tests are highly sensitive, and the electrophysiologic study results, two different pathophysio- rate of false positive results is low (7-9). The overall 77% logic patterns of paroxysmal atria) fibrillation associated rate of positive responses to the tests in our present patients with syncope can be proposed. with paroxysmal atrial fibrillation and syncope was sur- L Abnormal neural regulation triggered by the onset of prisingly similar to the 71% rate observed by our group in atrial fibrillation . This mechanism was observed during patients with unexplained syncope studied with a similar upright electrophysiologic study in 42% of patients . In these protocol (7). Moreover, the 12% rate of false positive re- patients, the hemodynamic consequences of the onset of sponses observed in our present control subjects was similar atrial fibrillation probably triggered a vasovagal reflex or to the 8% rate previously observed in healthy subjects (7). unbalanced the compensatory cardiovascular responses . Leitch et al . (11) found similar results in patients affected Vasovagal bradycardia-hypotension reflex has been pos- by supraventricular tachycardia. In their study, tilt testing tulated to occur because of inappropriate stimulation of left during sinus rhythm elicited syncope in 6 of 7 patients with ventricular stretch receptors as a result of hypovolemia and tachycardia-induced syncope (during upright electrophysio- vigorous ventricular contraction (8,9). An increase in sym- logic study) compared with only 4 of 15 patients without pathetic tone may precipitate this reflex and is the theoretic tachy± a-induced syncope . They found no differenc+, in justification for the use of isoproterenol infusion during the tachycardia cycle length between the two groups ; of passive tilt testing (9). The onset of atrial fibrillation reduces patients, They therefore concluded that syncope during left ventricular filling through loss of atrial contraction, supraventricular tachycardia is related to vasomotor factors abbreviation of the diastolic filling time and irregularity of and does not predict a more rapid tachycardia rate . These cycle length, resulting in a sudden decrease in cardiac output data support the results of a previous retrospective study (3,15) . An increase in sympathetic tone occurs at the onset of that failed to find an association between syncope and tachyarrhythmia (2,16) . These changes may be magnified Mimed atria) fibrillation cycle :ength (1). when atrial fibrillation occurs in the upright position, which The a., heart rate of Lhe present syncope group was in itself reduces left ventricular volume and increases sym-

3ACC Vol . 22 . No. 4 BRIGNOLE ET AL . 1127 OLlober 1'`93 :1123--9 SYNCOPE AT THE ONSET OF ATRIAL FIBRILLATION

Figure 1. Electrocardiographic and blood pressure recording during upright induction of atlial fibrillation . The mean heart rate (HR) and blood pressure values are reported in con- secutive 6-s periods for the 1st 30 s s a+ 9oas after the end of burst pacing . A, Atrial fibrillation-induced syncope . After induction of atrial fibrillation, there is an immediate decrease in systolic blood pressure (SRP) that sta- bilizes at approximately 55 mm Hg; 11wuuuuaA' . after 24 s, blood pressure decreases to a mean of 47 mm fig, with syncope . At this time the rate of utrial fibrilla- W Uw "amm"s tion slightly decreases to 130 beats/ min. B, Induction of asymptomatic utrial fibrillation in a patientt with spontaneous syncope. After atlial fibrillation is induced, blood pressure ...kkk°~ immediately decreases but to a lesser a a a apses degree than that observed in A. a Thereafter, it increases without

changes in heart rate . C, Induction Wasi n of asymptomatic atrial fibrillation in a control subject . After atrial fibrillation is induced, blood pressure

changes are smaller than in A and B ; M a slight increase in heart rate is observed. bpm = beats per minute.

pathetic tone (13,17) . Consequently, there may be activation explain, although it has previously been observed by our of cardiac mechanoceptors during atrial fibrillation as a group (14). This association probably means that there result of diminished ventricular volume and vigorous ven- is a common autonomic nervous system disease or related tricular contraction. Similar mechanisms have been postu- etiologies . The association of positive responses to carotid lated for head-up tilt-induced syncope (8,9). In the present sinus massage and head-up tilt testing has been previously study, 11 (73%) of 15 patients who had syncope induced by reported both in patients affected by unexplained syncope the onset of atrial fibrillation also had a positive response to (7) and in patients affected by sick sinus syndrome (18). either baseline or isoproterenol head-up tilt testing (Table 3) . Carotid sinus reflex can be abnormally increased during The tendency toward a decrease in heart rate at the time of atrial fibrillation but not during sinus rhythm (19) . It is likely syncope during indua :-ed atrial fibrillation (Table 4, Fig. IA) that the mechanism of central neural summation is involved also suggests a vagal influence and a diminished sympathetic in causing abnormal reflexes (20) . tone. Moreover, a reflex mechanism is suggested by the An alternative mechanism for the induction of syncope by relatively prolonged time to syncope (mean time 22 s after atrial fibrillation is a diminished or delayed activation of the the onset of the arrhythmia) . A delayed occurrence of compensatory cardiovascular response to initial hemody- syncope was also observed in four of seven patients in the namtc disturbances caused by the onset of atrial fibrillation . study of Leitch et al . (11). Experimental animal studies of selective sympathetic dener- The association between carotid sinus hypersensitivity vation have demonstrated that both alpha- and beta- and atrial fibrillation-induced syncope is more difficult to adrenergic blockade cause increased persistence of tachy-

JACC Vol. 22, No I12S BRIGNOLE ET AL . . 4 SYNCOPE AT THE ONSET OF ATRIAL FIBRILLATION October 1993 :1123--9

fibrillation could be partly responsible for the less severe neural response observed in these patients . It is well known that intense vagal stimulation can cause paroxysmal atrial fibrillation (4,5). It is not infrequent during head-up tilt testing or carotid sinus massage to observe "idiopathic" atrial fibrillation at the time of reflex syncope (6,25). We also observed one such case in the present study . Vagal stimulation shortens atrial refractoriness (6); atrial fibrillation may result from vagal stimulation because changes in atrial refractoriness are not distributed uniformly throughout the atria (26). Clinical implications . This study demonstrates the useful- so re 00 110 130 160 Systolic blood pressure (mmHg) ness of upright electrophysiologic study in association with vasovagal maneuvers as a diagnostic tool and as a guide for ii gore 2. Scattcrplots of systolic blood pressure and heart rate in therapy in patients with syncope associated with paroxysmal induced atrial fibrillation. Mail .,tuares show patients with syn- atrial fibrillation . Indeed, this study may have therapeutic cope associated with induced atrial fibrillation ; open equarea show implications, Treatment should be aimed at preventing the patients without syncope . The correlation between blood pressure recurrence of either atrial fibrillation or abnormal neurogenic and heart rate is weak (r - 0.33), with a particularly wide dispersion of heart rate values in patients with syncope . bpm = beats per reflexes . The administration of antiarrhythmic drugs seems minute. to be justified only when, during upright electrophysiologic study, syncope is shown to be the consequence of the induction of atrial fibrillation. Conversely, when atria] fibrillation is secondary to syncope, measure that will prevent the cardia-induced hypotension (16) . It is well known that vasovagal reaction may be preferable . However, the utility elderly people and patients with heart failure show an of such an approach should be investigated further in pro- impairment of beroref ex-mediated vasoconstrictor re . spective clinical studies. sponses to the unloading of baroreceptors by orthostatic stress (21-23) and a decreased responsiveness to endoge- nous catecholamines (24). The mean age of our patients with References atrial fibrillation was 60 years and mild structural heart disease was present in 6 of patients, 1. Auricchio A, Klein H. Trappe HJ, Wenzleff P . Lack of prognostic value a neurally of syncope in patients with Wolfl=Parkinson-White syndrome . I Am Coll 2. Atrial fbr1llation as a consequence of Cardlol 1991 ;17:152-8. mediated syncope. This possible mechanism may be indi- 2. Curry PVL . The hemodynamic and electrophysiological effects of parox- rectly supposed in those patients (58% of total) in whom, ysmal tachycardia. In: Nanda OS. ed. Cardiac Arrhythmies. Electrophys- during the electrophysiologic study, the induction of atrial iology, Diagnosis and Management. Baltimore: Williams & Wilkins, 1979:364-81. fibrillation was unable to cause syncope . That these patients 3. Goldreyer BN . Kastor JA, Kershbaum KL. The hemodynamic effects of also had a high neurogenic susceptibility is suggested by the induced supraventricular tachycardia In man. Circulation 1976;54:783-9. high rate 0%) of positive responses to vasovagal maneu- 4. Lewis T, Drury AN, Bulger HA. Observations upon flutter and atrial vers. Yet atrial fibrillation was not able to trigger a reflex fibrillation. VII. The effects of vagal stimulation. Heart 1921 :8:141-69. 5. Coumel P, Leclercq IF. Auuol P. Lavallee JP, Flammang D. Autonomic response of sufficient magnitude to cause syncope. The influences in the genesis of atrial fibrillation : atrial flutter andfibrillation of lower average value of the heart rate during induced atrial vagal origin. In Ref 2:243-55. 6. Leitch J. Klein G. Yet R. Murdock C. Teo WS. Ncurally-mediated syncope and atrial fibrillation. (letter). N Engl I Med 1991;324:495-6. 7. Brignole M, Menozzi C. Gianfranchi L, Oddone D, Lolli G, Bertulla A . Tile 4. Difference in Heart Rate After the Onset of Induced Carotid sinus massage, eyeball compression and head-up till test in Attial Fibrillation in Patients With and Without Syncope and in patients with syncope of uncertain origin and in healthy control subjects. Control Subjects Ain Heart J 1991 ;123 :1644-51. 8. Kenny RA, Ingram A, Bayliss J . Sutton R . Head-up tilt : a useful test ror Patients patients im esligating unexplained syncope . Lancet 1986;2:1352-4. With Without 9. Almquist A. Goldberger JF, Miletein S, et al . Provocation of bradycardia Induced induced control and hypotension by isoproterenol and upright posture in patients with 'hate From Onset of Atria] Syncope Syncope Subjects unexplained syncope . N Engl J Med 1989;320:346-51. Fibrillation (0-16) (n - 22) (n = 16) 10. Molhoek GP . Wesseling KH, Settels JJ, et al . Evaluation of the Pendz servo-plethysmo-manometer for the continuous, non-invasive measure- Men HR Cbeatslmin) ment of finger blood pressure . Basic Res Cardiol 1984;79:518-609. At0-6s 164 ± 38" 129 :t 29" 127 ± 33 It. Leitch IW, Klein GJ, Yee R, Leather RA, Kim YH . Syncope associated At 24-?0 s or at syncope 152 :t 44 130 ± 36 135 ± 24 with supraventricular tachycardia. An expression of tachycardia rate or Didbrence (beatslminl -12 ± 214 +) ± 17t +9 ± 24 vasomotor response? Circulation 1992 ;85:1064-71. 12. Friedman DB. Jensen FB, Matzen S, Secher NH. Non-invasive blood "p= 0094. to= 0.04. tp = 0.009. Values presented are mean value ± SD. pressure monitoring during head-up tilt using the Penhz principle. Acta HR = heart rate. Anaesthesiol Scand 1990;34:519-22.

BRIGNOLE ET AL . 1129 SYNCOPE AT THE ONSET OF ATRIAL FIBRILLATION

13 . Hammill SC, Holmes DR. Wood DL, et al . Electrophysiologic testing in 19. Hellestrand KJ, Nathan AW, Camm AL Differential response to carotid the upright position: improved evaluation of patients with rhythm distur- sinus pressure during sinus rhythm and atrial fibrillation . Br Heart J bances using a tilt table . J Am Coll Cardiol 1984 ;4:65-71 . 1982 ;47:504-6. 14. Gianfranchi L, Brignole M . Menozzi C. Oddone D, Lolli G . Petillo A . 20. Greene AS. Brunner MJ . Concerning reflex summation . Am J Physiol : usefulness of The Syncope due to paroxysmal atrial fibrillaiionlflutter 1988 .255 :H]244-6. electrophysiological study in the supine and upright position . G ltal 21 . Ferguson DW, Abboud FM. Mark AL . Selective impairment of barorefiex- Cardiol 1990 ;20:828-33 . mediated vasoconstrictor responses in patients with ventricular dysfunc- 15 . Naito M, David D, Michelson EL, Schaffenburg M, Dreifus LS. The hemodynamic consequences of cardiac arrhythmias : evaluation of the tion . Circulation 1984 ;69 :451-60. relative roles of abnormal atrioventricular sequencing, irregularity of 22 . Piha SJ . Cardiovascular autonomic reflex tests : normal responses and ventricular rhythm and atrial fibrillation in a canine model . Am Heart J age-related reference value . Clin Physiol 1991 ;11 :277-10 . 198300084-91 . 23 . KaUser L, Sachs C. Autonomic cardiovascular responses in old age . 16. Feldman T, Caroll JD, Munkenbeck F, et al . Hemodynamic recovery Chn Physiol 1985;5:347-57. : role of adrenergic receptor during simulated ventricular tachycardia 24. Packer M . Neurohormonal interactions and adaptations in congestive activation . Am Heart J 1 8 ;115 :576-87. heart failure . Circulation 1988;77:721-30. 17 . Hashimoto T, Fukatani M, Mod M, Hashiba K . Effects of standing on the 25 . Strasberg B, Sagie A, Erdman S, Kusniec J, Sclarovsky S, Agmon J. induction of paroxysmal supraventricular tachycardia . J Am Coll Cardiol Carotid sinus hypersensitivity and the carotid sinus syndrome . Prog 1991 ;1 :690-5 . Cardiovas•c Dis 1989 ;31 :379-91 . 18 . Brignole M, Menozzi C, Gianfranchi L. Oddone D, Lolli G . Bertulla A. Neurally-mediated syncope detected by carotid sinus massage and 26 . Ninomiya 1 . Direct evidence of nonuniform distribution of vagal effects on OctoberJACChead-up tilt.1123-9. test in sick sinus syndrome . Am J Cardiol 1991 ;68:1032-6 . dog alria . Circ Res 1966 ;19:576-83 . 22No.4 Vol. 1993