Outcome of Patients With Dual-Chamber Pacemakers Implanted for the Prevention of Neurally Mediated Syncope

Outcome of Patients With Dual-Chamber Pacemakers Implanted for the Prevention of Neurally Mediated Syncope

Outcome of Patients With Dual-Chamber Pacemakers Implanted for the Prevention of Neurally Mediated Syncope Satish R. Raj, MD, Mary Lou Koshman, RN,...

147KB Sizes 0 Downloads 29 Views

Outcome of Patients With Dual-Chamber Pacemakers Implanted for the Prevention of Neurally Mediated Syncope Satish R. Raj,

MD,

Mary Lou Koshman,

RN,

and Robert S. Sheldon,

MD, PhD

Dual-chamber pacemaker insertion (PI) has been found to reduce the recurrence of neurally mediated syncope (NMS) in 3 randomized trials. However, the long-term benefits of PI are unknown. To assess the natural history of NMS, we followed a cohort of 40 patients who underwent PI for frequent NMS for 46 to 75 months. We assessed the reduction in syncope frequency after PI and the time to first recurrence of syncope. Sixty months after PI, 32.5% of patients remained free of NMS. The frequency of syncope decreased from 0.46 to 0.06 spells/ month (before to after PI, p ⴝ 0.04). Two groups of patients were identified, with responders to PI defined as having a 75% decrease in the frequency of NMS. Responders (n ⴝ 22) experienced a significant decrease in the frequency of NMS (0.31 to 0.01 spells/month,

p <0.0001), whereas nonresponders did not experience a similar reduction (p ⴝ 0.8). Responders could not be identified by either baseline or tilt-test parameters. Patients without an early recurrence of syncope after PI (within 6 months) experienced a significant reduction in the frequency of NMS (0.24 to 0.02 spells/month, p ⴝ 0.0002), although the reduction was not significant (p ⴝ 0.3) in patients with an early recurrence of syncope. Some, but not all, patients respond to permanent PI for NMS. The long-term benefit of permanent PI can be predicted by timing of the first recurrence of syncope, but not by preimplantation factors. 䊚2003 by Excerpta Medica, Inc. (Am J Cardiol 2003;91:565–569)

lmost 20% of the population experience at least 1 episode of neurally mediated syncope (NMS) in A their lifetime, and 9% have recurrent syncope. Al-

NMS on the basis of clinical presentation, tilt-test outcome, or both.12 Other causes of syncope were eliminated on the basis of clinical assessment, and no patient had structural heart disease, carotid sinus hypersensitivity, or atrioventricular block. A short-term follow-up study of the effect of PI in 12 patients was published,13 and another 18 received pacemakers while participants in the first North American Vasovagal Pacemaker Study9 (observation period ⬍15 months). Tilt-table test: Patients underwent tilt-table testing in the absence of a ␤–adrenoreceptor-blocking agent disopyramide, or in the absence of drugs with anticholinergic activity, in a quiet room after they had fasted for 4 to 8 hours. A short tilt-test protocol was used in 17 patients. This consisted of passive head-up tilt for 10 minutes, followed by up to another 10 minutes of head-up tilt during an infusion of 5 ␮g/min of isoproterenol. The remaining 23 patients underwent a tilt test similar to that of Morillo et al.14 Patients underwent passive head-up tilt for 45 minutes, followed immediately by head-up tilt with isoproterenol 30 ng/kg/ min. Tilt-test results were positive if they ended in clinically reminiscent syncope or presyncope, and a decrease in rate-systolic blood pressure product to ⱕ6,500 mm Hg/min without isoproterenol or a decrease to ⱕ9,000 mm Hg/min with an infusion of isoproterenol.15 Syncope was defined as a transient state of unconsciousness characterized by spontaneous recovery. Trough heart rate was defined as the lowest heart rate when tilt-test results became positive. Each patient was deemed to have a trough heart rate above or below the group median, stratified by whether an

1

though conservative measures2– 4 and medical therapies5– 8 are often attempted, they often fail. Insertion of a dual-chamber pacemaker was associated with a relative risk reduction of syncope recurrence of 80% to 92% in 3 randomized, prospective, open-label controlled trials.9 –11 However, because of their open-label nature, and the relatively brief duration of 2 of the studies,9,11 the true long-term benefit of permanent pacemaker insertion (PI) in NMS is not known. Furthermore, some investigators10 have suggested that a bradycardic response during a tilt-table test may identify patients more likely to benefit from pacemaker therapy. In this open-label observational study, we assessed whether the benefit of permanent PI is sustained over a period of years in patients with NMS, and whether baseline variables, such as bradycardia during tilt testing, are predictors of a successful response to permanent pacemaker therapy.

METHODS

Patients were included if they had ⱖ6 prior lifetime syncopal spells that were thought to be due to From the Cardiovascular Research Group, University of Calgary, Calgary, Alberta, Canada. This study was supported in part by grants from the Canadian Institutes of Health Research, Ottawa, Canada. Manuscript received July 11, 2002; revised manuscript received and accepted October 21, 2002. Address for reprints: Robert Sheldon, MD, PhD, Faculty of Medicine, University of Calgary, 3330 Hospital Drive NW, Calgary, Alberta T2N 4N1, Canada. E-mail: [email protected]. ©2003 by Excerpta Medica, Inc. All rights reserved. The American Journal of Cardiology Vol. 91 March 1, 2003

0002-9149/03/$–see front matter doi:10.1016/S0002-9149(02)03307-6

565

infusion of isoproterenol was used. Trough blood pressure and trough rate-systolic blood pressure product were similarly defined. Interventions: All patients were taught about the diagnosis, probable pathophysiology, and prognosis of NMS. They were taught to recognize and act on their prodromal presyncopal symptoms. In keeping with the lack of evidence at the time for drug therapy, patients were not receiving any standardized medical therapy. Patients received 1 of 2 pacemakers. Twenty patients received a Vitatron Diamond “rate smoothing” dualchamber pacemakers (Dieren, The Netherlands) in which the “Flywheel” feature was activated. With this feature, the pacemaker tracks the intrinsic heart rate and paces whether the sensed rate decreases by ⱖ15 beats/min/beat. The initial pacing rate is 15 beats/min below the rate preceding the rate decrease, and the paced rate slowly decreases until it is exceeded by the intrinsic heart rate. The remaining patients received a Medtronic Thera DR pacemaker (Minneapolis, Minnesota), with the “rate drop response” feature turned on. In response to a sudden heart rate decrease, the patient receives sustained dual-chamber pacing at an elevated rate followed by a gradual decline down to the lower rate. The pacemaker for each patient was dictated by availability and ongoing studies. Study design: The primary outcomes were the eventual frequency of syncope throughout the follow-up period, and the time to the first syncope recurrence. We compared the frequency of syncope in the periods before and after PI. The frequency of syncopal spells was calculated as the total number of syncopal spells divided by the duration of the observation period. Follow-up was completed for all patients using a combination of clinical visits and telephone contact. We arbitrarily prespecified a patient to have a successful response to PI (responder) if they experienced a ⱖ75% decrease in the frequency of syncope; otherwise patients were nonresponders to PI. This threshold was chosen because we believed that a dramatic reduction in symptom burden would be required from invasive therapy to justify its use. (Generally similar results were seen if the definition of a responder was relaxed to ⬎50% reduction in syncope frequency.) The time to the first recurrence of syncope16 after PI was displayed using the Kaplan-Meier method. Statistical analysis: Normally distributed data are presented as mean ⫾ SD, whereas non-normally distributed data are presented as median (25th percentile, 75th percentile). For continuous variables, the Mann-Whitney U test was used for unpaired 2 group comparisons, and the paired sign test was used for paired comparisons. The Wilcoxon log rank-sum test was used for comparisons between actuarial survival functions. Categoric variables were compared using Fisher’s exact test. A 2-tailed analytic approach was used.

RESULTS

Patient group: The study consisted of 40 patients (28 women and 12 men, aged 42 ⫾ 18 years) of whom 35 had positive results during the tilt-table test. Rate 566 THE AMERICAN JOURNAL OF CARDIOLOGY姞

VOL. 91

FIGURE 1. Actuarial syncope-free survival after PI. Vertical line separates patients with a first syncope recurrence within 6 months (left of line) and patients with a later first recurrence (right of line).

drop-responsive pacemakers were implanted in 20 patients, and 20 patients received rate-smoothing pacemakers. The median time from first lifetime syncope to PI was 79 months (38, 327). During this time, they experienced a median of 42.5 (19, 198) syncopal spells, with a median syncope frequency of 0.46 (0.16, 1.86) syncopal spells/month. The median duration of follow-up was 63 months (56, 67; range 46 to 75). Recurrence of syncope: The actuarial freedom from syncope recurrence is shown in Figure 1. By 60 months after PI, patients had only a 32.5% probability of remaining free of further syncope. During the follow-up period, 11 patients (27.5%) had no recurrence of syncope, 6 (15%) had 1 recurrence of syncope, and 23 (57.5%) had ⬎1 recurrence. Patients with only 1 syncope recurrence had their episode 16 months (4, 42) after PI, whereas those with ⬎1 recurrence of syncope had their first recurrence at a median 3 months (1, 14) (p ⫽ 0.2). The median frequency of syncopal spells decreased from 0.46 (0.16, 1.86) syncopal spells/month before PI, to 0.06 (0.00, 0.62) syncopal spells/month after PI (p ⫽ 0.04). Of the 29 patients who experienced ⱖ1 recurrence of syncope, the median frequency of syncopal spells decreased from 0.76 (0.17, 2.33) syncopal spells/month before PI to 0.27 (0.03, 1.15) syncopal spells/month after PI (p ⫽ 0.02). Figure 2 shows the syncope frequency for each patient before and after permanent PI. Most patients (85%) experienced a decrease in the frequency of their syncope after PI. Responders: Patients were considered “responders” to PI if they experienced a reduction of at least 75% in their syncope frequency after insertion versus before insertion. Of the 40 patients assessed, 55% were longterm responders to permanent PI. Both preimplant and postimplant factors were assessed as predictors of a successful response (Table 1). Patient ages at implant, gender, pacemaker model, the time from first syncope to implant, and the total number of syncopal spells and the syncope frequency before PI were not significantly different between responders and nonresponders. MARCH 1, 2003

the presence of 2 groups based on whether or not they experience a first recurrence of syncope within 6 months of PI. Do baseline tilt-test parameters predict patient outcome? Tilt-table tests before PI were positive in 35 of

FIGURE 2. Syncope frequency before and after pacemaker (PM) insertion are plotted for each patient on a log scale; median data are presented.

After PI (Table 1), patients who responded to PI experienced fewer syncopal spells at follow-up (median 0.5 vs 24 syncopal spells in nonresponders; p ⬍0.0001) and had a lower frequency of syncope at follow-up (median 0.01 vs 0.40 syncopal spells/month in nonresponders; p ⬍0.0001). Responders experienced a significant reduction in the frequency of syncope with PI (median 0.31 syncopal spells/month before PI vs median 0.01 syncopal spells/month after PI; p ⬍0.0001), whereas nonresponders did not experience a similar reduction (median 0.60 syncopal spells/ month before PI vs median 0.40 syncopal spells/ month after PI; p ⫽ 0.8). Pacemaker responders experienced a longer time to their first syncope recurrence (median 49 months) than did nonresponders (median 4 months; p ⫽ 0.0007). Figure 1 shows 2 groups of patients with an abrupt decline in the likelihood of a first recurrence of syncope at 6 months. Seventeen of the 23 patients with a prolonged time (⬎6 months) to first recurrence of syncope were long-term responders to PI. When dichotomized at 6 months, a time to first recurrence of syncope of ⬎6 months predicted a successful response to PI (p ⫽ 0.01). Patients who first had a syncope recurrence after ⬎6 months experienced a reduction in the frequency of syncope with PI (0.24 [0.12, 0.98] syncopal spells/month before PI vs 0.02 [0, 0.07] syncopal spells/month after PI; p ⫽ 0.0002), whereas patients with an early recurrence of syncope (⬍6 months) did not experience a similar reduction in the frequency of syncope (0.90 [0.35, 2.33] syncopal spells/month before PI vs 0.44 [0.09, 2.84] syncopal spells/month after PI; p ⫽ 0.3). To determine if these were 2 distinct groups, data were modeled to fit a monoexponential decay of the proportion free of syncope versus the time after PI (Figure 1). The slope of the natural logarithmic regression line for patients with a recurrence of syncope within 6 months of PI (⫺0.460 ⫾ 0.114 months⫺1) was significantly different from the slope for patients who did not have such a recurrence (⫺0.013 ⫾ 0.006 months⫺1; p ⬍0.0001). These data are consistent with

the 40 patients. The tilt-table test results were not different between patients who were long-term responders to PI (91% positive) and those who did not respond (83% positive, p ⫽ 0.6). To assess whether a low trough heart rate at tilt-table testing predicts a successful response to permanent PI, we dichotomized the group into patients with a trough heart rate above the median (high heart rate) or below the median (low heart rate). To avoid the confounding effect of isoproterenol on the trough heart rate, we independently determined the median heart rates for patients who received an infusion of isoproterenol (80 beats/min) and those patients who did not (43 beats/min), and then grouped patients whose trough heart rate was above or below their respective median trough heart rate. There was no difference in the time to first recurrence of syncope between these 2 groups (p ⫽ 0.6; Figure 3). The dichotomized trough heart rate could not predict long-term response to PI (p ⫽ 0.2). Similarly, patients’ trough systolic blood pressures (median 40 mm Hg with and 68 mm Hg without isoproterenol) and rate-systolic pressure products (median 6,188 mm Hg/min with and 2,000 mm Hg/min without isoproterenol) did not predict a long-term response (p ⫽ 0.2 for both).

DISCUSSION Although we found an overall 87% decrease in the median frequency of syncope, most patients had at least 1 syncope recurrence, and when pacemaker benefit was defined as a 75% decrease in syncope frequency, this benefit was found in only half the patients. Several explanations exist to explain this discrepancy with the early trials. First, the median of 42.5 lifetime syncopal spells before PI in this report is greater than in the 3 trials. There might be a pacemaker benefit in patients who have only a moderate symptom burden, but not in patients more severely afflicted. However, we did not detect an effect of baseline syncope history on the likelihood of benefit. Second, it could be that because patients in the randomized trials were not blinded to their PI, they may have experienced a substantial placebo effect that eventually diminished. Both the North American Vasovagal Pacemaker Study-I and Syncope Diagnosis and Treatment (SYNDAT) trials were terminated early owing to the observed benefit. The extended length of follow-up after PI in this study is its greatest strength. Third, half the patients received a pacemaker with a rate-smoothing algorithm, not a rate-drop responsive algorithm. Although this is lesser therapy, the likelihood of recurrences was not significantly different between patients who received either pacemaker. Finally, patients were relatively young compared with patients from the Vasovagal Syncope International Study, and it might be that some of the discrepancy is due to differences in age. Again, we

ARRHYTHMIAS AND CONDUCTION DISTURBANCES/PACEMAKER FOLLOW–UP IN NEURALLY MEDIATED SYNCOPE

567

TABLE 1 Patient Characteristics by Response to Pacemaker Insertion (PI) Overall (n ⫽ 40)

Responder (n ⫽ 22)

Nonresponder (n ⫽ 18)

p Value

Before Pacemaker Age (yrs) Women Men

38.5 (28.5, 58.5) 28 (70%) 12 (30%)

39 (29, 62) 16 (40%) 6 (15%)

37.5 (28, 56) 12 (30%) 6 (15%)

0.6 0.7

Trough Heart Rate (beats/min) Positive tilt without isoproterenol Positive tilt with isoproterenol

43 (29.5, 52.5) (n ⫽ 13) 74 (60, 103) (n ⫽ 22)

41.5 (10, 47.5) (n ⫽ 8) 72 (62.5, 101.5) (n ⫽ 12)

49 (39.8, 63.8) (n ⫽ 5) 75 (58, 110) (n ⫽ 10)

0.2

7 (18%) 11 (28%)

0.3

0.8

Pacemaker Type Rate-drop responsive Rate smoothing

20 (50%) 20 (50%)

13 (33%) 9 (22%) Syncope History

Time from 1st lifetime syncope to PI (mo) Lifetime syncopal spells before PI Syncope frequency before PI (spells/mo)

79 (38, 327) 42.5 (18.5, 198) 0.46 (0.16, 1.86)

77 (38, 450) 66 (17, 204) 0.31 (0.15, 7.61)

81 (33, 306) 25 (20, 80) 0.60 (0.16, 1.00)

0.7 0.3 0.8

After Pacemaker Length of follow-up after PI (mo) Syncopal spells after PI Syncope frequency after PI (spells/mo) Time to 1st syncope recurrence (mo)

63 3.5 0.06 14

(56, 67) (0, 34) (0, 0.62) (2, 52)

64 0.5 0.01 49

(55, 67) (0, 2) (0, 0.04) (8, 64)

61 24 0.40 4

(59, 66) (7, 131) (0.11, 2.17) (1, 11)

0.4 ⬍0.0001 ⬍0.0001 0.0007

Data are presented as median (25th percentile, 75th percentile). Continuous data are analyzed using the Mann-Whitney U test; categorical data are analyzed with the Fisher’s exact test.

FIGURE 3. Actuarial freedom from syncope after PI in patients with a trough heart rate at tilt-table testing that is greater than the median heart rate (trough heart rate greater than the median, shaded circles) and below the median heart rate (trough heart rate less than the median, open circles).

detected no age effect on the likelihood of benefit. Better predictors of benefit are needed. Our results suggest a high likelihood of recurrence of syncope in many patients despite PI. The simplest explanation is that these patients have a high frequency of vasodepressor syncope. The prevalence of vasodepressor 568 THE AMERICAN JOURNAL OF CARDIOLOGY姞

VOL. 91

syncope is unknown, although Moya et al17 noted profound bradycardia in only 50% of patients with syncope who had implanted loop recorders. Although it is conceptually appealing that a pacemaker would provide benefit only in a bradycardic milieu, we found that trough heart rate during tilt testing did not predict the long-term response to PI. Acute intervention studies suggest that a cardioinhibitory response to tilt testing predicts a response to temporary pacing for the prevention of syncope during a follow-up tilt test.18 However, a cardioinhibitory response has not been as useful in studies of permanent pacemakers.13,19 Recently Petersen et al19 reported a study in which 111 patients with a history consistent with NMS underwent implantation of a loop recorder. Syncope recurred equally whether the tilt test was positive or negative. Half the patients with syncope had an asystolic or bradycardic heart rate during recurrent syncopal spells. An asystolic response during recurrent syncope was found even if the patient had a vasodepressor response during the tilttable test. Taken together, these data show that hemodynamic responses during tilt testing do not predict hemodynamic responses during subsequent syncope, and do not predict responses to permanent PI. The use of syncope frequency as an outcome measure is a strength of our study. The 3 randomized trials9 –11 used the first syncope recurrence as the primary end point, MARCH 1, 2003

whose limitations have been reviewed.20 In contrast, the eventual frequency of syncopal spells, our primary measure, more accurate reflects of the long-term burden of syncope. We defined a “successful” response to PI as a 75% decrease in syncope frequency because we believe that a large benefit would be required for a clinician to counsel a patient to undertake an invasive therapeutic approach for NMS. Although almost all of our patients experienced a reduction in the frequency of their syncopal spells after PI (Figure 2), only 55% of the patients experienced a 75% decrease in frequency. Most patients who did not have their first recurrence of syncope for ⬎6 months after PI were long-term responders. These data support our prior observation16 that the time to the first recurrence of syncope is a reasonable surrogate for eventual syncope frequency. Study limitations: Therapy was not randomized, and there was no control group. Thus, we do not know what the natural history of this group would have been without PI, and it could be that responders were going to improve without intervention. However, this study, the equivalent of a phase IV study, does provide valuable information about the natural history of patients with recurrent NMS after PI. One would expect biased results from the lack of a control group to magnify the benefits of PI. However, we found the opposite result. Finally, this was a single-center study. 1. Chen LY, Shen WK, Mahoney DW, Jacobsen SJ, Rodeheffer RJ. Prevalence

of self-reported syncope: an epidemiologic study from Olmsted County, MN (abstr). J Am Coll Cardiol 2002;39:114A–115A. 2. Di Girolamo E, Di Iorio C, Leonzio L, Sabatini P, Barsotti A. Usefulness of a tilt training program for the prevention of refractory neurocardiogenic syncope in adolescents: a controlled study. Circulation 1999;100:1798 –1801. 3. Ector H, Reybrouck T, Heidbuchel H, Gewillig M, Van de Werf F. Tilt training: a new treatment for recurrent neurocardiogenic syncope and severe orthostatic intolerance. PACE 1998;21:193–196. 4. Mangru NN, Young ML, Mas MS, Chandar JS, Pearse LA, Wolff GS. Usefulness of tilt table test with normal saline infusion in management of neurocardiac syncope in children. Am Heart J 1996;131:953–955.

5. Di Girolamo E, Di Iorio C, Sabatini P, Leonzio L, Barbone C, Barsotti A.

Effects of paroxetine hydrochloride, a selective serotonin reuptake inhibitor, on refractory vasovagal syncope: a randomized, double-blind, placebo-controlled study. J Am Coll Cardiol 1999;33:1227–1230. 6. Cox MM, Perlman BA, Mayor MR, Silberstein TA, Levin E, Pringle L, Castellanos A, Myerburg RJ. Acute and long-term beta-adrenergic blockade for patients with neurocardiogenic syncope. J Am Coll Cardiol 1995;26:1293–1298. 7. Mahanonda N, Bhuripanyo K, Kangkagate C, Wansanit K, Kulchot B, Nademanee K, Chaithiraphan S. Randomized double-blind, placebo-controlled trial of oral atenolol in patients with unexplained syncope and positive upright tilt table test results. Am Heart J 1995;130:1250 –1253. 8. Perez-Lugones A, Schweikert R, Pavia S, Sra J, Akhtar M, Jaeger F, Tomassoni GF, Saliba W, Leonelli FM, Bash D, Beheiry S, Shewchik J, Tchou PJ, Natale A. Usefulness of midodrine in patients with severely symptomatic neurocardiogenic syncope: a randomized control study. J Cardiovasc Electrophysiol 2001;12:935–938. 9. Connolly SJ, Sheldon R, Roberts RS, Gent M. The North American Vasovagal Pacemaker Study (VPS). A randomized trial of permanent cardiac pacing for the prevention of vasovagal syncope. J Am Coll Cardiol 1999;33:16 –20. 10. Sutton R, Brignole M, Menozzi C, Raviele A, Alboni P, Giani P, Moya A. Dual-chamber pacing in the treatment of neurally mediated tilt-positive cardioinhibitory syncope: pacemaker versus no therapy: a multicenter randomized study. The Vasovagal Syncope International Study (VASIS) Investigators. Circulation 2000;102:294 –299. 11. Ammirati F, Colivicchi F, Santini M. Permanent cardiac pacing versus medical treatment for the prevention of recurrent vasovagal syncope: a multicenter, randomized, controlled trial. Circulation 2001;104:52–57. 12. Sheldon R, Rose S, Flanagan P, Koshman ML, Killam S. Risk factors for syncope recurrence after a positive tilt-table test in patients with syncope. Circulation 1996;93:973–981. 13. Sheldon R, Koshman ML, Wilson W, Kieser T, Rose S. Effect of dualchamber pacing with automatic rate-drop sensing on recurrent neurally mediated syncope. Am J Cardiol 1998;81:158 –162. 14. Morillo CA, Klein GJ, Zandri S, Yee R. Diagnostic accuracy of a low-dose isoproterenol head-up tilt protocol. Am Heart J 1995;129:901–906. 15. Sheldon R, Killam S. Methodology of isoproterenol-tilt table testing in patients with syncope. J Am Coll Cardiol 1992;19:773–779. 16. Malik P, Koshman ML, Sheldon R. Timing of first recurrence of syncope predicts syncopal frequency after a positive tilt table test result. J Am Coll Cardiol 1997;29:1284 –1289. 17. Moya A, Brignole M, Menozzi C, Garcia-Civera R, Tognarini S, Mont L, Botto G, Giada F, Cornacchia D. Mechanism of syncope in patients with isolated syncope and in patients with tilt-positive syncope. Circulation 2001;104:1261– 1267. 18. Kurbaan AS, Franzen AC, Heaven D, Mathur G, Bowker TJ, Petersen M, Sutton R. Cardioinhibition during tilt testing identifies patients who may benefit from pacing. PACE 2000;23:1792–1794. 19. Petersen ME, Chamberlain-Webber R, Fitzpatrick AP, Ingram A, Williams T, Sutton R. Permanent pacing for cardioinhibitory malignant vasovagal syndrome. Br Heart J 1994;71:274 –281. 20. Sheldon R, Rose S. Components of clinical trials for vasovagal syncope. Europace 2001;3:233–240.

ARRHYTHMIAS AND CONDUCTION DISTURBANCES/PACEMAKER FOLLOW–UP IN NEURALLY MEDIATED SYNCOPE

569