Factor	Description
Goal	Prevent reflex bradycardia and compensate for reflex hypotension
Goal	Prevent syncope
Level of evidence for success	Observational studies and open-label, randomized controlled trials (RCTs)
Level of evidence for success	No double-blind studies
Consensus recommendations	Class I: Recurrent syncope, with syncope induced by carotid sinus massage
Consensus recommendations	Class IIa: Recurrent syncope, with profound bradycardia induced by carotid sinus massage
Patient selection	Syncope and positive carotid sinus massage
Programming considerations	Atrioventricular sequential pacing

The carotid sinus reflex is an integral component of the homeostatic mechanisms of blood pressure regulation.¹⁵ Increases in intrasinus pressure stimulate mechanoreceptors, which participate in an afferent arc terminating in the brainstem. The efferent arc travels to peripheral end organs through vagal efferents, which augment cardiac vagal input and slow heart rate, and through the spinal cord to inhibit peripheral sympathetic activity in skeletal vasculature, resulting in peripheral vasodilatation. This reflex maintains blood pressure within a narrow range.

An abnormal carotid sinus reflex can cause exaggerated responses of heart rate and blood pressure. Some evidence suggests that the major defect in carotid sinus hypersensitivity does not reside in the carotid sinus or in its neural efferents,^16,¹⁷ or in the brainstem. Rather, the neuromuscular structures surrounding the carotid sinus may be involved in CSS. Blanc et al.¹⁸ found similar results in 30 patients without known carotid sinus hypersensitivity or syncope. Abnormal sternocleidomastoid electromyograms were associated with abnormal responses to CSM. Because the denervated sternocleidomastoid muscle cannot provide or contribute information to the central nervous system baroreflex centers, any output from the carotid sinus is inappropriately interpreted as increased blood pressure.

Other Therapies

When CSS is the likely cause of syncopal episodes, the initial treatment recommendation should be simple elimination of any recognized maneuvers that may precipitate an event. Discontinuation of wearing tight collars and ties and shaving more carefully may help. Hypovolemia should be corrected. The addition of high salt intake, volume expanders such as fludrocortisone acetate (Florinef),¹⁹ or oral vasopressors such as midodrine (ProAmatine) may be helpful, but these interventions are frequently limited in older patients by comorbidities (e.g., hypertension, heart failure).

In the pre–pacemaker era, recalcitrant cases of carotid sinus syncope were treated with carotid sinus denervation by surgical technique.²⁰ Surgical sinus denervation currently is reserved for cases secondary to head or neck tumors or lymphadenopathy, or it is performed in conjunction with carotid endarterectomy or in patients with severe refractory carotid sinus syncope of the purely vasodepressor type.

Evidence of Clinical Benefit

Observational Studies

Table 15-2 summarizes studies of pacing for CSS.^13,²¹^–²⁵ Earlier studies tended to be retrospective reports of pacing practices for CSS and therefore were inherently biased toward patients with a clear diagnosis of CSS who would truly benefit from pacing.

TABLE 15-2 Clinical Studies of Pacing for Carotid Sinus Syncope*

Randomized Trials

More recently, prospective, randomized trials have examined outcomes on the basis of presence of pacing and mode. A prospective, randomized Italian trial reaffirmed the important role of permanent pacing; 60 patients with CSS were randomized to pacing (32) or no pacing (28) therapy.¹⁴ During a follow-up of about 3 years, syncope recurred in 16 patients of the no-pacing group (51%) and in three (9%) of the pacing group (P = .002). This finding somewhat confirms the usefulness of pacing for the prevention of carotid sinus syncope, although potent placebo effects cannot be ruled out (see later).

Falls

Pacing of patients with CSS has been associated with a reduction in falls.²⁶ Furthermore, among patients with unexplained and recurrent falls, carotid sinus hypersensitivity may be an important risk factor.^27,²⁸ In 2001, Kenny et al.²⁹ reported the open-label SAFE PACE trial, designed to determine whether cardiac pacing reduces falls in older patients with unexplained falls and a cardioinhibitory response to CSM. The 187 patients were randomized to receive a dual-chamber pacemaker, with rate-drop responsiveness, or no intervention. Patients who received a pacemaker had a highly significant 58% reduction in falls and a 40% reduction in syncope. Although these results suggest that many unexplained falls in the elderly are caused by carotid sinus syncope, and that these can be prevented with pacing, one must remember that this was an open-label trial.

In 2010, Ryan et al.³⁰ reported the SAFEPACE-2 study, in which 141 older patients with unexplained falls and cardioinhibitory carotid sinus hypersensitivity were randomized to receive a rate-drop responsive dual-chamber pacemaker or an implantable loop recorder (ILR). Although the relative risk (RR) of reporting a fall after implantation of a device decreased (0.23; 95% confidence interval [CI] = 0.15 to 0.37), there were no significant differences in falls reported between the paced group (67%) and the ILR group (53%) (difference RR = 1.25; 95% CI = 0.93 to 1.67). These results are at odds with the findings in SAFE PACE; possibly because of differences in patient population (patients were older and frailer). Furthermore, due to recruitment difficulties, SAFEPACE-2 may have been underpowered; according to sample size calculations, 226 patients were needed to detect a 20% difference with 80% power.

Only one double-blind, placebo-controlled crossover trial has reported pacing in patients with recurrent, unexplained falls and carotid sinus hypersensitivity.³¹ All study participants received a dual-chamber pacemaker with rate-drop response programmer and were randomized to DDD/RDR mode (on) versus ODO mode (off); after 6 months, patients switched to the opposite mode. Only 25 of the 34 recruited patients completed the study. The total number of falls decreased after pacemaker implantation, but risk of falling with the pacemaker on versus off was not statistically significant (RR = 0.82; 95% CI = 0.62 to 1.10). Although this suggests a placebo effect, similar to that seen in pacing of vasovagal syncope patients, the results must be interpreted with caution; the study was underpowered because of an unexpectedly high attrition rate of 26%.

Patient Selection

Careful patient selection may help provide effective and efficient therapy for CSS. Permanent pacemaker therapy is indicated for patients with recurrent, frequent, or severe CSS, particularly for predominant cardioinhibitory syncope.^32,³³ Predictors of success with permanent pacing include multiple episodes before implantation; episodes that occur while upright or sitting; and episodes preceded by a recognized stimulus.³⁴ Syncope recurrence after implantation of a permanent pacemaker may be caused by a prominent vasodepressor component.

Physical Diagnosis with Carotid Sinus Massage

The carotid sinus is located high in the neck below the angle of the mandible. Carotid massage is contraindicated in the presence of bruits or a history of cerebrovascular disease, transient ischemic attacks (TIAs), or endarterectomy. Sequential application of CSM to the left and right carotid arteries should be performed with at least 10 to 20 seconds between applications. The duration of carotid massage should be 5 to 10 seconds, and it should be terminated with the onset of characteristic asystole or severe presyncope. In most series, the predominant responses to CSM were obtained on the right side.⁵ CSM should be performed while the patient is both supine and upright, either sitting or while secured safely on a tilt table. It may be difficult to document transient hypotension with standard sphygmomanometric methods, and noninvasive continuous digital plethysmography is often used.

Physiologic Responses

Carotid sinus massage elicits both cardioinhibitory and vasodepressor responses (Fig. 15-1). A cardioinhibitory response to CSM is defined as 3 seconds or longer of ventricular standstill or asystole. Ventricular asystole usually results from a sinus pause caused by sinus node exit block,³⁵ but it can result from atrioventricular (AV) block as well. A vasodepressor response to CSM is defined as a drop in systolic blood pressure of 50 mm Hg or more during massage; this may be difficult to demonstrate in patients who have a significant cardioinhibitory component. In contrast to the induced cardioinhibitory component of carotid sinus hypersensitivity, the vasodepressor response may have a slower, more insidious onset and a more prolonged resolution.

Figure 15-1 Combined cardioinhibitory and vasodepressor response to carotid sinus massage (CSM). Note slow return of blood pressure despite resolution of asystole.

Carotid Sinus Hypersensitivity and Carotid Sinus Syncope

Carotid sinus hypersensitivity denotes abnormal physiologic responses to CSM, either cardioinhibitory or vasodepressor (or both). The presence of asymptomatic carotid sinus hypersensitivity is quite common in older adults. For example, a positive cardioinhibitory response to CSM was noted in 32% of patients undergoing coronary angiography.³⁶ CSS is the syndrome of syncope in association with carotid sinus hypersensitivity, and in the absence of other apparent causes of syncope.

Complications

Carotid sinus massage is safe if done carefully. CSM is contraindicated in patients with a history of cerebrovascular disease or carotid bruits, because it can cause cerebrovascular accident (CVA, stroke). In a review of 3100 episodes of CSM performed on 1600 patients, the seven complications (0.14%) were neurologic and transient.³⁷ In another review of CSM on 4000 patients, complications were observed in 11 patients (0.28%);³⁸ all were neurologic. After 1 month, nine patients had made a full recovery; neurologic symptoms persisted in two patients. Rare, arrhythmic complications include asystole and ventricular fibrillation.³⁹

Programming

Pacing in AAI mode is contraindicated because many patients may eventually demonstrate associated reflex AV block.⁴⁰ In general, patients appear to benefit most from AV sequential pacing, even when a significant component of vasodepressor CSS is present. VVI pacing should not be used in patients with intact ventriculoatrial (VA) conduction,⁴¹ because of possible pacemaker syndrome. Lack of VA conduction at a given time, however, does not ensure against its future development. Therefore, we recommend dual-chamber pacemakers for patients with CSS and normal sinus rhythm.

Few studies have examined the role of rate-responsive pacing in CSS. Patients are generally older and therefore may have bradycardic comorbidities such as sick sinus syndrome or chronotropic incompetence, either intrinsic or pharmacologic. Therefore, rate-responsive pacing might be beneficial. Similarly, few studies have prospectively examined pacing with rate-drop or hysteresis capabilities, which has the theoretical advantage of providing rapid, higher-rate AV sequential pacing to counteract the vasodepressor component during CSS attacks.⁴²

Recommendations

Evidence is weak to support use of pacing in patients with carotid sinus syncope. No persuasive evidence supports pacing in patients with falls and carotid sinus hypersensitivity.

Vasovagal Syncope

Clinical Perspective

Vasovagal syncope is the most common of the neurally mediated syncopal syndromes. Most people who faint probably do not seek medical attention for isolated events.² Prolonged standing, sight of blood, pain, and fear are common precipitating stimuli for this, the common faint. Patients develop nausea, diaphoresis, pallor, and loss of consciousness from hypotension with or without significant bradycardia. Return to consciousness typically occurs after seconds or 1 to 2 minutes. Those with adequate warning may be able to use physical counterpressure maneuvers, or simply sit or lie down, to prevent a full faint. However, some patients have little or no prodrome, no recognized precipitating stimulus, or marked bradycardia accompanying the faint.⁴³^–⁴⁷ These patients have sparked interest in permanent pacing as a therapy.

Epidemiology

About 40% of people faint at least once in their life, and at least 20% of adults faint more than once.^48,⁴⁹ Fainters usually present first in their teenage years and 20s and may faint sporadically for decades. This long, usually benign, and sporadic history can make for difficult decisions about therapy. Syncope is responsible for 1% to 6% of emergency room visits and 1% to 3% of hospital admissions.⁵⁰^–⁵² Tilt tests are often used as a diagnostic tool, although they are limited by difficulties with sensitivity, specificity, and reproducibility, and with little evidence-based agreement on methodologic details and outcome criteria. Positive tilt tests are characterized by presyncope, syncope, bradycardia, and hypotension, as well as a reproduction of the patient’s perisyncopal symptoms^53,⁵⁴ (Fig. 15-2).

Figure 15-2 Hypotension and bradycardia induced during a positive drug-free passive tilt-table test. HR, Heart rate; MAP, mean arterial pressure.

Although many patients of all ages simply have vasovagal syncope, clinicians need to remain vigilant and look for other causes, including valvular and structural heart disease, sick sinus syndrome, CSS, and orthostatic hypotension.

Symptom Burden and Quality of Life

The vasovagal syncope syndrome has an extremely wide range of symptoms. The symptom burden varies from a single syncopal spell in a lifetime to daily faints. Some patients have very sporadic presentations, with periods of intense symptoms interspersed with long periods of quiescence. Other patients faint at more regular intervals, although they may be months or years apart. Several observational studies and randomized clinical trials reported that patients have a median of 5 to 15 syncopal spells, with fainting episodes occurring over 2 to 60 years.⁵⁵^–⁵⁸ Patients with recurrent syncope are impaired similar to those with severe rheumatoid arthritis or chronic low back pain and psychiatric inpatients.⁵⁵ The quality of life decreases as the frequency of syncopal spells increases.⁵⁶

After clinical assessment, many patients continue to do poorly. After 1, 2, and 3 years, 28%, 38%, and 49% of patients faint again, respectively.⁵⁹ Interestingly, several groups reported a 90% reduction in the total number of faints in this patient population after the tilt test. The reason for this apparently great reduction in syncope frequency after assessment is unknown, but it does lead to a large number of patients who request further treatment. Therefore, when assessing syncope patients, clinicians need to be alert to the surprising impairment of quality of life that many patients endure, to provide a perspective that lasts decades, and to remember that the patient’s clinical state will probably fluctuate.

Rationale for Pacing

Table 15-3 lists recommendations and other considerations in pacing for vasovagal syncope.

TABLE 15-3 Summary of Pacing for Vasovagal Syncope

Factor	Description
Goal	Prevent reflex bradycardia and compensate for reflex hypotension
Goal	Prevent syncope
Level of evidence for success	Limited evidence for benefit based on double-blind RCTs
Level of evidence for success	May be a subset of patients with proved bradycardia who benefit
Consensus recommendations	Class IIa: Recurrent vasovagal syncope with clinically documented bradycardia, or bradycardia induced on tilt test
Patient selection	Medically refractory, frequent, disabling vasovagal syncope
	Documented pauses during syncope
	Tilt test results not helpful
Programming considerations	Dual-chamber pacemaker
Programming considerations	Benefit from specific sensor to drive rate-response or pacing algorithm (rate-drop response, ventricular impedance) not proved

Physiology

Syncope is a transient loss of neurologic function caused by a global reduction of cerebral blood flow (CBF). Sudden cessation of CBF results in loss of consciousness within 4 to 10 seconds.⁶⁰ Lesser CBF reductions may result in presyncope. Almost all vasovagal syncope occurs while the patient is in an upright position. Syncope is usually associated with heightened physiologic or psychological stress, such as prolonged orthostatic stress; arising quickly and walking; pain, fear, emotion, or seeing blood or medical procedures; and strenuous exercise.

Transient hypotension is the most common hemodynamic manifestation of vasovagal syncope. Many patients have inappropriate peripheral sympathetic responses to physiologic and psychological stressors. Under conditions in which the normal response may be vasoconstriction, syncope patients usually fail to do so. Abnormalities have been documented in arteriolar vasoconstriction, splenic venoconstriction, and venous capacitance. The ultimate cause of hypotension is an abrupt cessation of vascular sympathetic traffic, causing withdrawal of α-adrenergic tone.

Evidence for Bradycardia

Permanent pacemaker therapy could be effective if bradycardia is a common and symptomatically important feature of vasovagal syncope. The evidence for clinically important bradycardia comes from studies that have used tilt tests, pacemaker memory, and implantable loop recorders to record heart rate during syncopal spells.

Tilt-Table Tests

Bradycardia frequently occurs during vasovagal syncope induced by tilt-table testing.^61,⁶² The mean heart rate during syncope induced by passive head-up tilt tests is 30 beats/min (bpm), and asystole longer than 3 seconds is often documented. However, uncertainty surrounds the relationship between the hemodynamics of tilt testing and clinical vasovagal syncope. For example, the ISSUE investigators found no relationship between the heart rate during syncope on tilt testing and during syncope in the community population. Patients with tilt test–induced bradycardia frequently do not have bradycardia during clinical syncope.^63,⁶⁴ Therefore, although bradycardia is the rule rather than the exception during a positive tilt test, the bradycardia evoked on a tilt test may not resemble the hemodynamics during syncope in that patient in the community.

Pacemaker Memory

Is asystole in patients in the community frequently associated with syncope? Evaluation of frequent fainters with pacemakers programmed to act as event recorders demonstrated that while transient asystole is common during documented syncope, many other asymptomatic asystolic episodes also occurred. Only 0.7% of asystolic events lasting 3 to 6 seconds and 43% of events lasting longer than 6 seconds resulted in symptoms of presyncope or syncope.⁶⁵ Therefore, even asystole of several seconds’ duration does not necessarily cause syncope.

Implantable Recorders

The ILR permits prolonged electrocardiographic monitoring and is a reasonable approach to diagnosing patients with infrequent syncope. Current ILRs weigh only 17 grams and have battery life of 14 months. The ECG signal is stored in a buffer that can be frozen with a manual activator. The ILR has programmable, automatic detection parameters for high and low rates and pause. In a Canadian study of 206 patients, symptoms recurred in 69% of patients. Bradycardia was detected more frequently than tachycardia (17% vs. 6%).^66,⁶⁷

The European International Study on Syncope of Uncertain Aetiology (ISSUE) enrolled 111 patients with syncope and previous tilt-table testing; not all tilt tests were positive.⁶⁸ Patients with positive or with negative tilt tests both had events in 34% of each group over a follow-up of 3 to 15 months. Marked sinus bradycardia (46%) or asystole (62%) were detected during syncope. Therefore, bradycardia reported during syncope varies widely; 17% to 62% of patients with vasovagal syncope had significant bradycardia during syncope in the ILR studies. The heart rate response during tilt testing did not predict spontaneous heart rate response, with more frequent asystole than expected based on tilt response. Thus, many patients with positive tilt tests may develop some degree of bradycardia at presyncope or syncope, and pacing may be a plausible treatment.

Conservative Therapy

Pacing should be tried only in patients with vasovagal syncope who have not responded to, or who are not candidates for, other treatments. Currently, most clinicians first teach patients about the causes of syncope, encourage fluid and salt intake, and coach physical counterpressure maneuvers. If this initial approach is unsuccessful, pharmacologic therapy is used (fludrocortisone, midodrine, β-blockers, SSRIs). Only after these options have been explored should permanent pacing be considered (Table 15-4).

TABLE 15-4 Principles of Management of Vasovagal Syncope

Area	Intervention
Diagnosis and prognosis	Confirm diagnosis with history, tilt-table tests, and loop recorder.
Diagnosis and prognosis	Assess likelihood of syncope recurrence (>2 spells or recent worsening).
Assessment of patient needs	Insight into diagnosis
	Cause of syncope
	Probability of syncope recurrences
	Treatment options
Conservative advice	Maximizing salt and fluid intake
	Physical counterpressure maneuvers
	Driving and reporting to authorities
	Avoidance and management of triggers
Medical options	Fludrocortisone (weak evidence)
	Midodrine (good evidence)
	Serotonin reuptake inhibitors (weak evidence)
	β-Blockers in patients age >42 (modest evidence)
Permanent pacing	Weak evidence

Reassurance and Education

Most patients with vasovagal syncope simply require reassurance and education. Many of the patients with more frequent syncope also benefit from these conservative measures. In particular, patients should be taught to recognize their prodromal symptoms and to sit or lie down as quickly as possible to minimize injury during recurrences and lessen the severity of attacks.

Salt and Fluids

Blood volume is an important factor in the pathophysiology of vasovagal syncope. Syncope almost only occurs in the upright position, and many patients faint only from orthostatic stress in situations such as attending religious services, parades, or outdoor events and taking showers. Prolonged drug-free head-up tilt provokes syncope in a large number of syncope patients; this depends on the duration and angle of head-up tilt.⁶⁹ Finally, many patients report avoiding dietary salt and have low daily urinary sodium excretion.⁷⁰

Physical Counterpressure Maneuvers

Physical counterpressure maneuvers may be quite helpful, although no blinded, controlled studies have been reported. Patients must have a prodrome long enough that they can react by isometrically tightening muscles using maneuvers such as squatting, leg crossing, and fist clenching. Van Dijk et al.⁷¹ reported the Physical Counterpressure Maneuvers Trial, a multicenter, randomized clinical trial to evaluate the effectiveness of physical counterpressure in preventing syncope recurrence. A total of 223 patients were randomized to receive optimal conventional therapy (lifestyle changes; e.g., avoiding triggers, increasing salt intake) with or without additional training in physical counterpressure. During the follow-up of up to 18 months, 50.9% of conventional therapy patients and 31.6% of conventional therapy with counterpressure patients experienced syncope recurrence, a relative risk reduction of 0.36 (95% CI = 0.11 to 0.53). These results provide sufficient evidence to support including the low-risk, low-cost therapy of physical counterpressure in the recommended first-line treatment of vasovagal syncope.

Medical Therapy

Table 15-5 summarizes major randomized clinical trials of treatment for vasovagal syncope. No therapies have proved effective in large, randomized clinical trials in preventing vasovagal syncope. Few have been subjected to rigorous clinical trials, and when interpreting open-label studies, one should remember that most patients appear to improve after assessment. There is an estimated 90% reduction in syncope in the population after tilt testing.⁷² The four major drug classes used are α₁-adrenergic agonists, β-blockers, selective serotonin reuptake inhibitors (SSRIs), and salt-retaining mineralocorticoids.

TABLE 15-5 Summary of Major RCTs of Treatment for Vasovagal Syncope

Vasopressors (Midodrone)

Reasonably good evidence exists for the effectiveness of the α₁-adrenergic agonist midodrine (ProAmatine). Midodrine, a prodrug, reduced symptoms of syncope and presyncope in three small, randomized clinical trials;⁷³^–⁷⁵ overall, midodrine is probably helpful.⁷⁶ The major limitations of midodrine are the need for frequent dosing and its tendency to increase supine blood pressure. The latter side effect is usually seen at higher doses (>30 mg/day). Midodrine should not be used in patients with hypertension. Side effects include piloerection and “crawling” paresthesias in the scalp.

Selective Serotonin Reuptake Inhibitors

Numerous small, open-label studies in the early 1990s reported that SSRIs prevented the induction of syncope on tilt tests and reduced symptoms in patients in the community. Paroxetine was effective in preventing syncope ⁷⁷ in one randomized placebo-controlled study and numerous case-report series. In contrast, Takata et al.⁷⁸ reported that the same drug did not block the vasovagal reaction elicited by lower-body negative pressure. SSRIs do not appear to be used widely for the prevention of syncope, and the treatment effect is debatable.⁷⁶

Beta-Adrenergic Blockers

The evidence for the effectiveness of β-blocker therapy is mixed. It has a strong physiologic rationale, and two positive and one negative open-label studies involving 42 to 153 patients.⁷⁹^–⁸¹ Five randomized clinical trials studied the efficacy or effectiveness of β-blockers for the prevention of syncope.⁸²^–⁸⁶ Although not completely consistent, these studies indicate that metoprolol and atenolol, and possibly β-adrenergic receptor blockade in general, are ineffective in preventing vasovagal syncope in the broad patient population. A substudy showed a possible benefit in middle-aged and elderly patients.

Fludrocortisone Acetate

Fludrocortisone has mineralocorticoid activity without appreciable glucocorticoid effect at doses up to 0.2 mg, the usual clinical doses for various disorders.⁸⁷ The acute actions of fludrocortisone acetate are sodium and water retention, at the expense of urinary potassium excretion. Two open-label trials examined fludrocortisone in patients with “neurocardiogenic syncope.” Both revealed clinical improvement but neither were placebo-controlled.^88,⁸⁹ Recently, Salim and Di Sessa⁹⁰ reported a small, placebo-controlled, randomized clinical trial (RCT) of fludrocortisone (Florinef) in children with vasovagal syncope. Patients taking fludrocortisone did significantly worse (P < .04) than those taking placebo. An RCT powered to detect a 40% relative risk reduction is nearing conclusion.⁹¹ Other drugs have been studied, but not in adequately powered, placebo RCTs.

Evidence of Clinical Benefit

Four groups assessed the effect of pacing in preventing syncope induced by tilt-table testing (Table 15-6). Forty-one patients with a positive initial tilt test and a marked bradycardia underwent a second tilt test with temporary pacing at rates of 85 to 100 bpm. Taken together,^61,^62,^65,^90,⁹²^–⁹⁶ the studies showed that temporary dual-chamber pacing prevented the development of syncope in 24 of the 41 patients (57%).⁹²^–⁹⁴ However, almost all the conscious patients developed the vasovagal reaction and had significant presyncope. Temporary pacing may be partly effective in preventing vasovagal syncope, but it does not prevent presyncope.

TABLE 15-6 Clinical Studies of Rate Drop–Responsive Pacing in Vasovagal Syncope

Rate Drop–Responsive Pacemakers

The strategy that emerged in the early 1990s was to attempt to use cardiac pacing not only to overcome transient bradycardia during syncope, but also to provide enough heart rate support to compensate for the transient hypotension that is part of the vasovagal reflex. The major issues in developing pacing strategies are (1) detection of the onset of syncope and (2) whether pacing is effective during the episode. Given the modest efficacy of simple bradycardia support in preventing syncope induced by tilt tests, investigators soon turned to more flexible and sophisticated modes of sensing and pacing.

Sensors of Transient Heart Rate Decrease

Three combined sensing-pacing options have been assessed clinically: rate hysteresis, rate smoothing, and rate-drop sensing. In rate hysteresis, pacing at a relatively high rate is triggered by a drop in heart rate below a low detection rate. An example of settings is a sensing rate of 40 bpm and response rate of 90 bpm. Typically, the high-rate pacing might be interrupted periodically to permit the pacemaker to detect an underlying rate, as well as cease pacing if the rate is high enough. Rate smoothing is used to avoid abrupt and sustained changes in heart rate. Here, an abrupt drop in heart rate over only one to three beats elicits a pacing response at a slightly lower rate, and this pacing rate gradually slows until it is below the accelerating intrinsic heart rate. Rate drop–responsive pacing is a more sensitive and sophisticated variant of rate hysteresis. An abrupt drop in heart rate of a programmable magnitude over a programmable duration elicits a burst of high-rate pacing for 1 to 2 minutes. The drop in heart rate needs to be only about 10 bpm, over a few seconds, with the high rate about 90 to 110 bpm.

Observational Studies

Three groups reported studies of the usefulness of chronic pacing in the prevention of vasovagal syncope (see Table 15-6). Petersen et al.⁹⁷ reported the first clinical study of dual-chamber pacing with rate hysteresis in 37 syncope patients. The patients had had a median of six syncopal spells, and a positive tilt test with bradycardia. Of the 37 patients, 31 received pacemakers with rate hysteresis. Over a mean follow-up of 50 months 62% of the patients remained free of syncope, and the number of syncopal spells in the total population fell from an expected number of 136 to only 11.

Benditt et al.⁹⁸ reported equally encouraging results in a study of 36 patients with predominantly vasovagal syncope. The patients were very symptomatic, with a median of 10 syncopal spells over about 2 years, or about five spells annually. All patients received a pacemaker with rate-drop responsiveness. The patients were followed for a mean of 6 months. During this time, syncope recurred in only six patients, compared to expected recurrences in about 30 patients. Therefore, in this relatively short-term study, pacing may have benefited about 80% of patients.

We studied 12 extremely symptomatic patients who had had a median syncope frequency of three spells monthly.⁹⁹ All had a positive tilt test and recurrent syncope while receiving medical therapy. All received a pacemaker with a rate-smoothing feature but without a high rate response. Following implantation of the pacemaker, the actuarial syncope-free survival increased 20-fold, the syncope frequency dropped by 93%, and quality of life improved highly significantly.

All the previous trials were sequential design studies, with no control for time-dependent effects or for the placebo effect.

Open-Label Randomized Studies of Rate-Drop Responsiveness

North American Vasovagal Pacemaker Study

The VPS-I Study tested whether permanent pacing with rate-drop responsiveness would reduce the likelihood of syncope in patients with frequent vasovagal syncope.¹⁰⁰ Patients were eligible if they had fainted six or more times before tilt testing, or fainted within the first year after a positive tilt test, and had a predefined degree of bradycardia. Fifty-four patients were randomized evenly to receive a pacemaker with automatic rate-drop responsiveness or the best medical therapy (no pacemaker), according to their treating physicians. Syncope recurrence was lower in the pacemaker patients (6 of 27) than in the medical patients (19 of 27). The hazard ratio for a recurrence of syncope in the paced versus medically treated patients was 0.087 (P = .000016). Figure 15-3 shows the likelihood of a first syncope recurrence among patients randomized to receive a pacemaker (or not) in VPS-I.

Figure 15-3 Cumulative likelihood of a recurrence of syncope in patients randomized to receive or not to receive a pacemaker in VPS-I.

(From Connolly SJ, Sheldon RS, Roberts RS, Gent M: The North American Vasovagal Pacemaker Study. A randomized trial of permanent cardiac pacing for the prevention of vasovagal syncope. J Am Coll Cardiol 33:16-20, 1999.)

Although the first randomized controlled trial (RCT) to show benefit from pacing, VPS-I appeared to have several limitations. First, with insufficient evidence for pacemaker effectiveness to implant a device in all subjects, the investigators selected an open-label design. Second, the patients were highly select; all had fainted frequently, had positive tilt tests with development of bradycardia, and agreed to participate in a study with only a 50% chance of receiving new therapy (i.e., pacemaker). The pacemaker could have benefited these patients by either conventional bradycardia support or sophisticated rate drop–responsiveness algorithm. Also, medical therapy was not standardized.

Ammirati et al.¹⁰⁰ performed a small, randomized clinical trial in 20 patients with moderately frequent syncope who received a pacemaker with either rate hysteresis or rate-drop responsiveness. Three patients with rate hysteresis fainted while no patients with rate-drop responsiveness fainted (P < .05). This study suggested that rate-drop responsiveness is superior to rate hysteresis in preventing syncope, and therefore not all the pacemaker effect was caused by placebo.

Vasovagal Syncope International Study

The VASIS Investigators randomly assigned 19 patients to receive a dual-chamber pacemaker with rate hysteresis and 23 patients to no pacemaker implant.¹⁰¹ The patients all had three or more (median six) syncopal spells over the previous 2 years and a cardioinhibitory response to tilt testing. Patients had a lower syncope burden than those of VPS-I. During a mean follow-up of 3.7 ± 2.2 years, the pacemaker group had a lower likelihood of a syncope recurrence than the no-pacemaker group (5% vs. 61%; P = .0006). Figure 15-4 shows the intent-to-treat results. Similar to VPS-I, VASIS was an open-label study that included highly select patients and could not eliminate the impact of placebo effect.

Figure 15-4 Cumulative likelihood of a recurrence of syncope in patients randomized to receive or not receive a pacemaker in VASIS.

(From Sutton R, Brignole M, Menozzi C, et al: 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 102:294-299, 2000.)

Syncope Diagnosis and Treatment Study

The SYDAT trial tested whether pacemakers or atenolol best prevented vasovagal syncope.¹⁰² Randomized to receive a DDD pacemaker with rate-drop response (46) or atenolol (47), the 93 patients were older than 35, had three or more syncopal spells in the preceding 2 years, and had a positive tilt test, with a trough heart rate of 60 bpm or less. At least one syncope recurred in 4.3% of the pacing group versus 26% in the atenolol group (odds ratio [OR] = 0.13; P = .004). This was another open-label study of pacing in vasovagal syncope using a highly select population. One confounding issue is a possible deleterious effect from the atenolol, rather than a beneficial effect from the pacemaker implantation. This seems unlikely given the overall neutral effect of β-blockers in the randomized trials previously summarized.

Summary

Observational reports ^42,^97,⁹⁹ and open-label RCTs^100,^101,¹⁰³ strongly suggested that patients have less syncope after they receive a permanent pacemaker. The question is whether this effect is real. All these studies were unblinded to both patients and physicians. Syncope is an outcome that can be difficult to verify objectively. Also, surgical procedures can have a placebo-effect.¹⁰⁴^–¹⁰⁶ Patients receiving a pacemaker may have benefited from the psychological effects of receiving a surgical procedure from enthusiastic health professionals. Given these uncertainties, the invasiveness and cost of pacing mandated that placebo-controlled or blinded RCTs be used to determine the true benefit of pacing.

Blinded Randomized Studies of Rate-Drop Responsiveness

Three blinded studies have compared the benefit of rate drop–responsive pacemakers with therapy anticipated to be of lesser benefit.

McLeod Crossover Study

McLeod et al.¹⁰⁷ reported the efficacy of rate drop–responsive dual-chamber pacing in the prevention of vasovagal syncope in 12 highly symptomatic young children who had frequent syncope associated with asystolic pauses longer than 4 seconds. In this three-way, double-blind, randomized crossover study, the pacemakers were programmed to no active pacing, ventricular pacing with rate hysteresis, or dual-chamber pacing with rate-drop responsiveness. Both pacing modes were equivalently more effective than no pacing in preventing syncope, and dual-chamber pacing was superior to ventricular pacing in preventing presyncope. This small study concluded that rate drop–responsive pacing was more efficacious than no pacing in preventing vasovagal syncope in children.

Second Vasovagal Pacemaker Study

To ascertain the therapeutic effect of permanent pacing in vasovagal syncope, we performed the Second Vasovagal Pacemaker Study (VPS-II).¹⁰⁸ We expected that the risk of syncope in the control group would be reduced to some extent by the placebo effect of receiving a device, and we increased the study sample size accordingly. VPS-II was a multicenter, double-blind, placebo-controlled, randomized clinical trial. Patients were eligible if they had recurrent vasovagal syncope with at least six lifetime syncope spells, or at least three spells in the 2 years before enrollment, and a positive tilt-table test performed according to the protocol in each center. A requirement for a specific degree of bradycardia during tilt testing was not included because trough heart rate during tilt testing did not correlate in patients with heart rate during clinical syncope,⁶⁹ and because trough heart rate during tilt testing did not appear to predict response to pacing.

All 100 patients received a dual-chamber pacemaker and were randomized to either rate-drop responsiveness or sensing without pacing. The care providers remained blinded to treatment allocation, except for an unblinded nurse or physician who did all the programming but did not disclose any details. The study was designed to have 80% power to detect a 50% relative reduction in the risk of recurrent syncope from a rate of 60% in the control group to 30% in the treatment group. A total of 38 patients had recurrent syncope during the 6-month follow-up: 22 of 52 patients in the sensing-only group and 16 of 48 in the active pacing group. The cumulative risk of syncope at 6 months was 40% (95% Cl = 25% to 52%) for the sensing-only group and 31% (95% CI = 17% to 43%) for the rate drop–responsive group (Fig. 15-5). The relative risk reduction in time to syncope with active pacing was 30% (95% CI = −33% to 63%; P = 0.14 one-sided). A retrospective analysis found no variable that predicted benefit from pacing, other than patients who received isoproterenol during the tilt test.

Figure 15-5 Kaplan-Meier plots of time to first recurrence of syncope in 48 patients randomized to receive active dual-chamber pacing and 52 patients randomized to receive a pacemaker in sense-only mode by intention-to-treat analysis in VPS-II.

(From Connolly SJ, Sheldon R, Thorpe KE, et al: The Second Vasovagal Pacemaker Study (VPS-II). A double-blind randomized controlled trial of pacemaker therapy for the prevention of syncope in patients with recurrent severe vasovagal syncope. JAMA 289:2224-2249, 2003.)

Most importantly, VPS-II found no statistically significant benefit in favor of pacemaker therapy for prevention of syncope. The most important difference between VPS-I and VPS-II results is the observed risk of syncope in the nonpaced group. In VPS-I, almost 80% of control patients fainted within 6 months, versus only 41% in VPS-II. In contrast, the 6-month likelihood of syncope in the patients receiving active pacing therapy was similar: 20% in VPS-I and 31% in VPS-II.

Vasovagal Syncope and Pacing Trial

SYNPACE involved 29 patients who had had a median of 12 lifetime syncopal spells, a positive tilt test, and bradycardia during the syncope induced by the tilt test.¹⁰⁹ They received a dual-chamber rate drop–responsive pacemaker and were randomized to either active pacing or no pacing. The trial was stopped early after the VPS-II results were released. Thirteen patients had at least one syncope recurrence, and there was no benefit from active pacing with rate-drop responsiveness. Although extremely underpowered, the study did not provide any support for the usefulness of pacing in preventing vasovagal syncope.

Why Were VPS-II and SYNPACE Not Positive?

Both VPS-II and SYNPACE failed to demonstrate a statistically significant benefit of pacemaker therapy for prevention of vasovagal syncope. Although SYNPACE was underpowered because of early termination, VPS-II was the largest RCT of pacemaker therapy for vasovagal syncope, whether open label or double blind. There was considerable effort to maintain blindedness and no known protocol violation. This strict adherence sets the standard for RCTs of treatment for vasovagal syncope. Reasons for the negative outcomes include early termination of previous studies, inability to overcome vasodepression with high-rate pacing, and a placebo effect in open-label studies.

Early Termination of Open-Label Studies

An important difference between VPS-II and the four open-label trials was that three of the four previous trials were stopped prematurely. Early termination of a trial for unexpected efficacy tends to overestimate the treatment effect.

Vasodepression Cannot Be Paced

Prevention of bradycardia is the main physiologic mechanism by which a pacemaker can prevent attacks of syncope. During positive tilt tests, however, reductions in blood pressure begin earlier than the development of bradycardia.^93,⁹⁴ Pacing therapy might not help patients with hypotension caused by vasodepression even if bradycardia or asystole also occurs at syncope. The results of VPS-II and SYNPACE suggest that most episodes of vasovagal syncope may be associated with profound vasodepression as the cause of syncope, rather than simply bradycardia. In this light, pacing may simply be ineffective in the setting of profound vasodepression, and future progress in devices might best target implantable drug delivery systems.

Placebo Effect

The history of attempts to treat patients with implanted devices has other examples of initial promises of therapeutic success being followed by subsequent well-controlled, negative studies. For example, open-label studies suggested that dual-chamber pacing causes a marked improvement in the hemodynamics and functional status of patients with hypertrophic cardiomyopathy. Later blinded RCTs revealed evidence of a much smaller effect size. Similarly, preliminary open-label studies suggested that atrial-based pacing might prevent atrial fibrillation, but a well-controlled randomized crossover trial showed much less benefit of conventional atrial pacing in prevention of atrial fibrillation.^110,¹¹¹ Large open-label studies also provided strong evidence for the ability of atrial pacing to reduce stroke and death in patients with pacemakers. A large, blinded RCT showed that patients with atrial-based pacemakers versus those with single-lead ventricular pacemakers had no benefit with respect to death, stroke, quality of life, or exercise tolerance for several years after implantation.¹¹² Therefore, care should be taken in the assessment of open-label or nonrandomized pacemaker studies. The placebo effect can be substantial.

Pacemaker therapy may be associated with an initial spurious benefit for two major reasons. First, patients receiving expensive or invasive therapy may not want to admit that such therapy may be ineffective. The placebo effect can be pronounced, particularly with surgical procedures.¹⁰⁴^–¹⁰⁶ Second, many patients with vasovagal syncope appear to improve spontaneously after tilt testing.^52,^57,⁵⁸ This effect may account for up to 90% of the apparent benefit. The mechanism is unknown but may involve the counseling received at the clinic visit, a regression to the mean, and the sporadic timing of vasovagal syncope. This effect is of a similar magnitude to the beneficial effect of pacing in sequential design trials. In the unblinded studies, patients who hoped to receive a pacemaker and were disappointed not to receive one may have been more prone to report syncope. Conversely, patients receiving a pacemaker may have benefited from the psychological effects of receiving a surgical procedure from enthusiastic health professionals. The double-blind trial design to a considerable extent removes this type of potential bias.

Clinical Trials of Pacing in Vasovagal Syncope: A Meta-Analysis

To assess the efficacy of pacing in vasovagal syndromes based on clinical trials, Sud et al.¹¹³ conducted a meta-analysis of nine randomized trials: four active pacemaker versus medical or no therapy,^101,^102,^114,¹¹⁵ three active pacemaker comparisons (open label or single blind),^103,^116,¹¹⁷ and two double-blind active pacemaker versus sensing only or pacemaker off.^109,¹¹⁸ When pooled, the nine trials showed significant heterogeneity; when based on trial methodology, however, the trials no longer showed evidence of heterogeneity. In open-label single-blind studies, risk of syncope recurrence was reduced with pacing; double-blind studies showed no effect (Figure 15-6). These results suggest that the main benefit of pacemaker insertion is an expectation effect; the expectation response appeared to explain most of the observed pacemaker response (OR = 0.16; 95% CI = 0.06 to 0.40; P = .0001).¹¹³ However, the two double-blind placebo-controlled RCTs conducted to date were small (129 total patients). A larger study is needed for definite conclusions on pacing in vasovagal syncope patients.

Figure 15-6 Forest plot of primary study and summary odds ratios (OR) for recurrent syncope in meta-analysis of nine studies. For each study, the red squares and horizontal lines represent the random effects OR and 95% confidence intervals (CI). Area of red square is proportional to amount of statistical information contributed by the trial. Diamond at the bottom of each category and overall represents OR with 95% CI.

(From Sud S, Massel D, Klein GJ, et al: The expectation effect and cardiac pacing for refractory vasovagal syncope. Am J Med 120:54-62, 2007.)

International Study on Syncope of Uncertain Aetiology

A multicenter double-blind RCT is in progress to study the effectiveness of pacing for the prevention of asystolic neurally mediated syncope.¹¹⁹ Patients 40 or older with at least three syncopal spells (suspected or confirmed as neurally mediated) over the preceding 2 years and with a severe clinical presentation requiring treatment initiation are eligible for the International Study on Syncope of Uncertain Aetiology (ISSUE-3). The exclusion criteria include carotid sinus hypersensitivity. In Phase I, participants undergo ILR until (a) first syncope with asystole of 3 seconds or longer in the loop recording; (b) asystole of 6 seconds or longer without reported syncope; or (c) 24 months of recording without (a) or (b). In Phase II, patients with (a) or (b) are randomized to receive a dual-chamber pacemaker switched on (active therapy) or off (placebo therapy). Patients, follow-up physicians, and study personnel are blinded to treatment allocation. The primary outcome is the time to first syncope recurrence in both study arms. The study will enroll up to 710 patients to randomize 60 patients to each study arm. The total study duration is estimated at 4 years.

Patient Selection

Patients should have a definite history of vasovagal syncope based on a positive tilt test, suggestive ILR findings, or scrupulous history. Given the weakness of the evidence supporting the efficacy of pacing, it is reasonable to pace only patients with documented profound bradycardia or asystole during syncope. Only patients at high risk of syncope recurrence should be paced. Any patient with at least one syncopal spell in the preceding year has almost a 50% risk of at least one syncopal spell in the next year. Otherwise, similar syncope patients with either negative or positive tilt-table tests have similar likelihood of syncope after assessment.^120,¹²¹ The outcome of the tilt test (negative vs. positive) does not predict subsequent clinical outcome. Similarly, the lowest heart rate (including asystole) during tilt testing does not predict the eventual likelihood of syncope in clinical follow-up.^69,⁹⁷

Utility of Loop Recordings

Sud et al.¹²² and Brignole et al.¹²⁴ assessed spontaneous bradycardia patterns from internal and external loop recordings of patients with unexplained syncope who later underwent pacemaker implantation for bradycardia. The ISSUE classification of detected rhythm from ILRs was used. Although severity of bradycardia or asystole during the loop recording did not predict response to pacemaker therapy, mechanism of bradycardia did; syncope associated with abrupt bradycardia was associated with better response to pacing therapy than syncope with gradual-onset bradycardia. These results suggest that the ISSUE classification is useful in distinguishing between bradycardia as part of vasovagal syncope and primary bradycardia, the latter of which is more responsive to pacing therapy.

Programming

Rate-drop response is only present with dual-chamber programming. There is an inexorable trade-off between sensitivity and specificity in programming rate-responsive pacemakers to detect the early stages of syncope. The main detection features include the range over which heart rate must fall, the time interval during which the fall must take place, and the number of confirmation beats below the minimum-detection heart rate. Decreasing the specified heart rate range or number of confirmation beats increases the sensitivity, as does increasing the time in which the heart rate fall must occur. Generally, the pacemakers greatly overdetect, with therapies delivered many times daily. This is usually well tolerated, particularly if the patient perceives a benefit in syncope prevention, but in some patients the palpitations are intolerable. This can be particularly true at night, when respiratory sinus arrhythmia is greater and the patient is quieter. Although the understandable tendency is to program the rate-drop feature off at night, this affords patients no potential protection should they arise. This can be a problem because patients usually have had no fluids for hours and can have syncope provoked by either orthostatic changes or micturition.

Pacing therapy is usually a relatively high-rate burst. The pacemaker should be programmed to deliver dual-chamber pacing at rates of 90 to 110 bpm for 1 to 2 minutes. There is no evidence of increased benefit from any particular rate or duration.

Evolving Paradigms and Technology

Contractility Sensors

The search continues for alternate sensing strategies such as QT intervals, respiratory volumes or frequency, right ventricular pressure transduction (dP/dt), and indices of contractility. These are intended to sense either early hypovolemia or early rises in sympathetic activity that may precede frank syncope. Some evidence indicates that contractility can be estimated with measures of endocardial acceleration, using a microaccelerometer in the pacemaker lead to estimate right ventricular myocardial contractility,^124,¹²⁵ or with intracardiac impedance measurements.^126,¹²⁷ The theory behind contractility sensors is that vasovagal syncope might be preceded by small but significant increases in contractility caused by a sympathetic surge. These devices increase pacing rates in response to increases in apparent contractility, then slowly decrease rates after contractility subsides toward baseline. Discouragingly, Brignole et al.¹²⁴ found that endocardial acceleration did not predict the occurrence of tilt table–induced syncope.

Griesbach et al.¹²⁷ evaluated the use of closed-loop stimulation (CLS) in a preliminary study in 2002. CLS pacemaker technology reacts to a change in the right ventricular intracardiac impedance, felt to be a surrogate measure of contractility and therefore sympathetic tone. This study of 22 patients demonstrated that syncope sensed with this modality could be prevented on tilt testing of patients with cardioinhibitory syncope. A subsequent study used CLS prospectively in 34 patients with recurrent vasovagal syncope. Over 12 to 50 months of follow-up, 30 of 34 patients had not experienced a recurrent event.

Based on this pilot study, a larger, multicenter RCT, the Inotropy Controlled Pacing in Vasovagal Syncope (INVASY), was partly carried out in 2004.^128,¹²⁹ In this open-label trial, 26 patients with recurrent vasovagal syncope and a positive tilt test with induced bradycardia received a dual-chamber CLS pacemaker, then were asymmetrically randomized to simple DDI pacing (9) or dual-chamber CLS pacing (17). INVASY was stopped early after a preliminary analysis showed that, after a mean of 19 months, seven of the nine DDI patients and none of the 17 CLS patients had syncope recurrences (P < .0001) (Fig. 15-7). These positive results suggest three possible conclusions: (1) pacing can be useful in patients with vasovagal syncope; (3) CLS based on right ventricular impedance changes is effective sensing for syncope; and (3) this may be the placebo effect, again. A larger and properly blinded study is being planned.

Figure 15-7 Probability of remaining free of syncope recurrence by Kaplan-Meier estimation in 41 patients in closed-loop stimulation (CLS) arm and nine patients in control group in INVASY study.

(From Occhetta E, Bortnik M, Audoglio R, Vassanelli C: Closed-loop stimulation in prevention of vasovagal syncope. Inotropy Controlled Pacing in Vasovagal Syncope (INVASY): a multicentre randomized single-blind controlled study. Europace 6:538-547, 2004.)

Automatic Drug Delivery Systems

Current pacemaker therapies focus only on heart rate support. This might not be as useful in patients with a predominantly vasodepressor response. Giada et al.¹³⁰ described an acute study assessing a novel implantable system that delivers phenylephrine when activated at the onset of syncope (prodrome with blood pressure drop). When treated with the phenylephrine, 15 of 16 patients had an immediate blood pressure rise and their tilt-induced syncope terminated, despite ongoing tilt testing. In contrast, none of the patients was able to abort the episodes when a placebo infusion was delivered. The one patient who fainted despite the phenylephrine experienced a severe cardioinhibitory response to the tilt test. This small study suggests the promise of combined approach using pacing for the cardioinhibitory component, as well as acute pharmacologic support for the vasodepressor component of syncope.