The management of vasovagal syncope

Syncope, defined as transient loss of consciousness with loss of postural tone but with spontaneous recovery,1 accounts for 3% of Accident and Emergency attendances2 and 1% to 6% of hospital admissions annually,3 and has a prevalence of up to 23% in the institutionalized elderly.4 In last month's review,5 we described the head-up tilt-table test and its use in the diagnosis of the vasovagal syndrome, the commonest cause of syncope. This article aims to provide an overview of the management of the vasovagal syndrome, on the basis of current evidence and our experience in this field. Throughout the article, the terms `vasovagal syncope' and `vasovagal syndrome' will be used interchangeably. Syncope is the most common presenting symptom of the vasovagal syndrome, though patients may present with pre-syncope and dizziness without loss of consciousness. A prodromal syndrome of variable duration is common, with light-headedness, weakness, a sensation of air-hunger or hyperventilation, detachment from surroundings, palpitations, blurring of vision and visual field disturbance, nausea, diaphoresis and ultimately pre-syncope or syncope being reported by the majority of patients.6–8 Recovery is usually prompt with the assumption of the supine position, but some subjects experience nausea, dizziness and diaphoresis for a variable period following an episode.6 Elderly subjects may experience little or no prodrome,7–9 and not infrequently present with serious injury following an unheralded syncopal event. Indeed, syncope may overlap with unexplained falls in older patients,10 with some elderly subjects recalling only the index fall rather than the syncopal attack precipitating it.10 In many cases, a diagnosis can be made from history alone,6 particularly if a witness account is available. However, there are a subset of patients who require further investigation to determine a cause of symptoms because they are atypical, because comorbidity suggests another probable cause of syncope, or because a clear attributable diagnosis is necessary for social or other reasons. Clinical examination and cardiovascular and neurological investigations are generally helpful only in differentiating other causes of syncope, and the ultimate diagnosis is based on head-up tilt-table testing, as described last month.5 The multiple treatments promoted for the vasovagal syndrome are a powerful indicator both of the complexity of the disease and the lack of a single, well-evaluated therapeutic option for its management. The following section will explain some of the methodological difficulties associated with the evaluation of interventions in vasovagal syncope, before going on to discuss the rationale and evidence for the various pharmacological and pacing therapies, and finally recommendations for treatment. Definitive guidelines on therapy for the vasovagal syndrome are hampered by a dearth of randomized controlled trials, the numerous therapeutic approaches explored, and the small numbers of subjects studied. The multiple tilt protocols used at different laboratories serve to cloud the issue further. More specifically, the evaluation of therapeutic interventions in vasovagal syncope is troubled by three major methodological difficulties peculiar to this issue, namely the variable reproducibility of the head-up tilt table test, the beneficial effect of diagnosis on syncope recurrence and the tendency for vasovagal syncope to ameliorate with time. The medium- to long-term reproducibility of positive responses to tilt-table testing varies from 62% to 85% of subjects,11–14 though one-day reproducibility may be as low as 35%.15 Such variability in concordance raises grave doubts regarding the utility of a negative second tilt-test one or more days after initial testing on therapy as a primary outcome measure, a frequent, and in our view, mistaken investigational approach. Symptom recurrence is a far more robust outcome measure in this context. Many patients with vasovagal syncope experience dramatic resolution or diminution of symptoms following a positive head-up tilt-table test. This has been attributed by some to a powerful placebo effect,16 although a more likely explanation is that recurrent episodes are avoided or ameliorated by counselling regarding the avoidance of precipitating situations and recognition of premonitory symptoms to allow the patient to abort an imminent attack.17,18 Both aspects complicate therapy evaluation, particularly as there are few double-blind randomized controlled trials in this area which include a placebo arm (see Table 1). Several studies suggest that the course of vasovagal syncope is benign, with progressive alleviation of symptoms over time,16–19 although whether this is due to the effects of counselling (as outlined above) or to the natural history of the disorder is unclear. Whichever explanation is correct, with the exception of a few randomized controlled trials, most studies have yet to account for this. As mentioned above, the provision of the diagnosis of vasovagal syncope, coupled with reassurance of its benign course and counselling regarding recognition of prodromal symptoms and subsequent manoeuvres to abort attacks, are central to the management of this disorder. The avoidance of precipitating factors, such as prolonged standing or sitting, excess alcohol intake and exercising after large meals or in a hot environment should be emphasized, although one report noted an improvement in orthostatic tolerance and blood volumes following exercise training.20 Stratagems to increase blood volume by increasing fluid and salt intake may also be helpful, as one of the main triggers for vasovagal syncope (and the pathophysiological cornerstone of the head-up tilt-table test) is venous pooling in the lower limbs and relative central hypovolaemia.5 A recent uncontrolled, retrospective report found that oral fluid therapy (1920 ml on rising followed by sufficient non-caffeinated fluids to keep urine clear) significantly reduced the number of syncopal episodes in a cohort of adolescent sufferers (mean age 14.7 years),21 while El-Sayed and Hainsworth elegantly demonstrated the utility of salt therapy in a small (20 subjects) randomized placebo-controlled study.22 Patients with vasovagal syncope diagnosed by lower-body negative pressure (LBNP) tilt underwent plasma volume measurements, 24-h urinary sodium, tilt- and baroreceptor sensitivity testing before and 8 weeks after treatment with 120 mmol salt or placebo daily. Plasma volume and orthostatic tolerance (assessed as time to onset of syncope during LBNP tilt) increased while baroreceptor sensitivity decreased in 8/10 of the salt treatment group but in only 3/10 of the placebo group, with the main determinant of response to salt treatment being 24-h urinary sodium excretion of <170 mmol.22 While this study did not assess symptomatic outcome, a retrospective study in adolescents found that a positive response to normal saline challenge during tilt testing (i.e. prevention of syncope following 10 ml/kg saline bolus) predicted symptomatic benefit from salt therapy with or without fludrocortisone treatment.23 Further minimization of relative hypovolaemia may be achieved by withdrawal of culprit medications. Nitrates and other vasodilators and diuretics are particularly implicated in older patients, and may either cause vasovagal syncope de novo or uncover a previously latent tendency to the disorder. Complete resolution of symptoms may ensue following discontinuation of therapy with these agents. Finally, venous pooling may be minimized by compression hosiery,9 though these are often poorly tolerated, particularly in older subjects. These are the most commonly prescribed drugs in vasovagal syncope and will be discussed in depth. Beta-adrenergic receptor antagonists have been used for several years in the management of the vasovagal syndrome. Their action is thought to be manifold, with the primary therapeutic benefit being a reduction in myocardial inotropy, which prevents the stimulation of left ventricular mechanoreceptors culminating in the Bezold-Jarisch–type reflex thought to be responsible for the vasovagal episode.24–26 A further benefit may include modification of the increases in circulating epinephrine27–29 and heart rate29 found prior to sympathetic withdrawal30 and syncope in many patients. From a high-sensitivity Medline search,31 along with scrutiny of non-Medline-listed journals, we were able to identify only six randomized controlled trials of β-blocker therapy in the treatment of vasovagal syncope,16,32–36 only two of which were placebo-controlled and blinded.32,36 None reported power calculations to ensure adequate numbers of participants, the maximum being 48 subjects,35 the minimum 13.36 While symptoms during follow-up were reported in several of the trials,33,35 the primary outcome measure in all six studies was repeat tilt on medication from 7 days34 to 6 months35 on therapy. Metoprolol was found to be superior to clonidine33 and verapamil,34 while atenolol was superior to placebo,32 but no more effective than fludrocortisone35 in reducing the number of positive tilt tests on therapy. In addition, there were no syncopal episodes during a short 15-day follow-up period on metoprolol.33 Unfortunately, the placebo-controlled study of Mahanonda et al.32 suffers from a host of methodological problems, not least of which were the loose inclusion criteria for the study, which allowed the enrolment of 23 (55%) subjects with precipitating factors including neck movements, cough, and postural change, in whom alternative causes of symptoms which may be exacerbated or improved by β-blockers (such as carotid sinus syndrome and orthostatic hypotension) had not been excluded. The trials of Fitzpatrick et al.36 and Brignole et al.16 used a variety of pharmacological therapies which included atenolol,16,36 scopolamine,36 clonidine,36 cafedrine,16 domperidone16 and dihydroergotamine,16 and found no differences between placebo and treatment arms,16,36 although few firm conclusions can be drawn because of the small numbers of patients and the even smaller number on each treatment. Several reports have provided positive evidence for the benefits of β-blockade in the vasovagal syndrome, though the methodological confounders discussed above are of course pertinent here. Metoprolol,30,37–42 pindolol,42 propranolol,43,44 esmolol,40,41 betaxolol,45 and atenolol46–48 have all shown benefit in uncontrolled studies, though again negative repeat tilt testing was the main outcome measure in the majority of these. Two reports show no benefit,49,50 with one describing worsening of symptoms on atenolol treatment.50 Thus, while the evidence for the use of β-blockers is scant and of dubious quality, the attractiveness of the pathophysiological rationale for their use, as well as their relative tolerability, ensures that these agents continue to be widely prescribed in the vasovagal syndrome. Disopyramide is a type IA anti-arrhythmic agent with anti-cholinergic and negative inotropic effects. It was thus postulated that the drug would prevent vasovagal syncope through a combination of diminishing left ventricular contraction (see above) and the exertion of a vagolytic effect, preventing the bradycardic/asystolic component of the attack. Uncontrolled studies showed great promise,51–53 but the only placebo-controlled, randomized trial of disopyramide showed no benefit, either during repeat tilt testing or during an average 29 months of follow-up.54 Its use cannot therefore be recommended. More potent anti-cholinergics including propantheline bromide55 and transdermal scopolamine56,57 have both shown promise in small-scale uncontrolled series, but there is insufficient hard evidence of efficacy to advocate their routine use in the vasovagal syndrome. Midodrine is a powerful α-agonist and vasoconstrictor, exhibiting 80% of norepinephrine-induced contraction of human peripheral veins.58 It is used in the management of orthostatic hypotension,59 and has recently received attention in the treatment of vasovagal syncope,60–63 in which it is thought to exert systemic pressor effects while impeding orthostatic blood pooling via peripheral venoconstriction.60 Only one randomized double-blind placebo-controlled study has been reported to date,60 with the midodrine-treated patients experiencing significantly less syncopal episodes and a better quality of life than those on placebo.60 Positive experiences with the drug in case reports61,62 and a series of eleven patients refractory to `usual measures'63 further suggest benefit with this agent. Midodrine is generally well tolerated, but its main side-effects (supine systolic hypertension, urinary frequency and urgency, piloerection, worsening of angina and cerebrovascular disease) necessitate discontinuation in up to 25% of older subjects after a year's treatment.64 It is currently available in the UK on a named patient basis only. Uncontrolled studies using ephedrine,65,66 pseudephedrine,67 dextro-amphetamine,68 methylphenidate69 and etilefrine70 all showed benefit in this disorder, but the small numbers of patients and methodologies preclude any firm conclusions, particularly as the only randomized controlled study (using etilefrine) found no benefits over placebo.54 Fludrocortisone is a synthetic mineralocorticoid which acts on the distal tubules of the kidney to retain sodium and hence expand blood volume and prevent relative central hypovolaemia.71 It also increases pressor sensitivity to circulating catecholamines.71 The only randomized trial of its use in vasovagal syncope was in adolescents, and showed equal benefit with atenolol, though there was no placebo arm to the study.35 Uncontrolled studies provide further modest evidence of efficacy and tolerability, though again only small numbers of patients have been studied.47,56,57 The most common side effects include hypertension, hypokalaemia and symptomatic fluid retention, and cardiac failure. Theophylline is a methylxanthine with several pharmacological modes of action depending on serum concentration, one of which includes antagonism of adenosine receptors shown to mediate vasodilatation.73 Only two small uncontrolled reports have been published, both of which show modest benefit with theophylline, albeit with minimal follow-up.66,74 While there is undoubtedly central serotonergic neuronal involvement in blood pressure regulation, the exact mechanism and mode of action remains uncertain. In animal models, 5HT2 receptor agonism raises blood pressure,75 while 5HT1A receptor stimulation decreases blood pressure.76 Experimental studies in vasovagal syncope using clomipramine (which enhances central serotonergic activity) show a higher responsiveness of central serotonergic mechanisms in vasovagal subjects than in normal controls.77 The theoretical tenets of treatment with SSRIs are thus uncertain, as reflected in the paucity of studies on these agents in vasovagal syncope. Case reports and small series from one unit claim benefit,78–80 though a recent case series showed exacerbation of symptoms in three patients.81 A more recent randomized placebo-controlled study on 68 patients with refractory vasovagal syncope demonstrated a negative one-month tilt test in 61.8% of patients on paroxetine and 38.2% of placebo-treated controls (p<0.001), while 17.6% of the paroxetine group reported syncope during a mean of 25.4 months of follow-up vs. 52.9% of controls (p<0.0001)97 SSRIs thus have a role in patients with refractory symptoms, and in others where depressive symptoms would prompt pharmacological intervention. Angiotensin-converting-enzyme inhibitors (ACEI) were until recently avoided in vasovagal syncope because of concerns regarding exacerbation of hypotension, but a recent double-blind, randomized, placebo-controlled trial in 24 elderly patients with vasovagal syncope showed that none of the enalapril-treated group (n=12) had a second positive tilt test after one week, compared to 10 who were syncopal in the placebo group.82 The enalapril-treated group also had lower plasma catecholamine concentrations, prompting the hypothesis that the ACEI blunted the triggering effects of the catecholamine surge prior to syncope.82 Further work is awaited with interest. Vasovagal syncope has cardio-inhibitory, mixed and vasodepressor components, classified into four sub-types (VASIS Types I, IIa, IIb and III) depending on the relative contribution of each59 with Types IIa and IIb representing the cardio-inhibitory sub-types (Table 2). Permanent pacemaker implantation has been proposed as a therapeutic option particularly in the so-called `malignant vasovagal syndrome', where there is a predominant asystolic/bradycardic component (VASIS Types IIa and IIb) in association with unheralded, injurious, disablingly frequent syncope.83,84 Concerns remain over this course of action for several reasons, including the paucity of well-conducted studies, the profound placebo effect of pacemaker implantation, the benign nature of the syndrome, the problems associated with implanting a life-long foreign device in otherwise well (and often young) subjects, and the inability of pacing therapy to counter the vasodepressor component of the vasovagal response. Most of the evidence to date has been based on small retrospective series,86 case reports78,87 and uncontrolled trials,88,89 all of which showed dramatic resolution of symptoms with pacing. One small trial of pacing therapy versus conventional drug treatment (but no placebo) found no fall in number of syncopal episodes post-pacing intervention,90 while a study of tilt responses with dual-chamber pacemaker implanted and switched on and off in separate studies (within individual subjects) found no difference in the time to syncope.91 A recent randomized controlled trial of pacing in vasovagal syncope was stopped at 54 out of the planned 284 patients enrolled92 because of the dramatic benefit seen in the pacing arm of the study group, though the small numbers used and the use of time to first syncope as the primary outcome measure are controversial. Despite this confusion, `neurally mediated syncope with significant bradycardia reproduced by a head-up tilt'93 remains an American College of Cardiology/American Heart Association Guidelines indication for pacing,93 but should be reserved for selected patients as described above. The results of ongoing European and Canadian randomized controlled trials are awaited. Biofeedback-assisted relaxation has been shown to help improve symptoms in several patients with syncopal disorders and concomitant headache,94 though this technique has not been formally evaluated in purely vasovagal subjects. A recent retrospective uncontrolled report noted resolution of symptoms in subjects undergoing repeated tilt testing (to syncope) and subsequent `tilt training' at home by standing against a wall for 30 min daily.95 Further studies are needed before this practice is widely adopted. Recommendations based on the scant evidence available and our experience are summarised in Figure 1. The following are intended as a guide only, and drug doses, cautions, contra-indications should be observed as per usual practice. Patient education, and withdrawal of culprit should be the primary therapeutic and should be and at all of management. should be 120 mmol while β-blockade may be achieved with or atenolol increased as Fludrocortisone should at increasing to a maximum with of blood pressure and serum Midodrine should be prescribed in only, at or to a maximum of with of supine blood pressure and routine and The last should be no than to prevent systolic SSRIs previously reported on include paroxetine and and should be at the and should be prescribed as per the Permanent pacing therapy may be reserved for patients with serious frequent and unheralded syncope who have prolonged or bradycardia on tilt-table testing who have therapy. The of device should be with a response or Such relative in heart as well as of and at a higher initial per in an to counter the vasodepressor component of the controlled trials of therapeutic interventions in vasovagal syncope controlled trials of therapeutic interventions in vasovagal syncope of head-up vasovagal syncope of head-up vasovagal syncope for the management of vasovagal syncope.