Tilt‐Table Testing in Patients with Syncope: What Does It Really Tell Us?

In this issue of the Journal, Del Rosso et al.1 report on the sensitivity, specificity, and tolerability of a shortened head-up tilt (HUT) protocol using sublingual glyceryl trinitrate (GTN). In their introduction, the authors correctly point out the increasing prevalence of syncope with advancing age. Additionally, they note that neurally mediated syncope, once thought to be largely restricted to younger individuals, has increasingly been recognized in older adults. In an arena full of different methods for performing tilt-table testing, their study makes a significant contribution to the literature by providing data on a well-standardized tilt table protocol in a population of older subjects. Nevertheless, a few limitations deserve emphasis. In their introduction, the authors identify the aims of this study as evaluation of the “sensitivity” and “specificity” of the proposed protocol in older patients with unexplained syncope. Nevertheless, as others have pointed out, there is no gold standard for vasovagal or neurally mediated syncope. Determining the “sensitivity” of a test requires the presence of a criterion standard: sensitivity = (the number of persons in the tested group with a positive test)/(all the persons in the tested group with the condition (as identified by the criterion standard)). In the absence of a criterion standard for diagnosing neurally mediated syncope, many authors have made the assumption that a positive tilt test equals neurally mediated syncope. Although this is a tempting assumption, given the relatively high positivity rates of tilt table testing in patients with typical vasovagal features and recurrent episodes of syncope, it is circular logic to propose a test as useful to detect a condition and simultaneously define the condition as being present when the test is positive. This would be akin to evaluating the sensitivity and specifity of ventilation-perfusion (VQ) scans for diagnosing pulmonary embolism while simultaneously using a positive VQ scan as the gold standard to determine whether a pulmonary embolism was indeed present. In fairness to the authors, the Methods section more appropriately refers to “positivity” rates rather than “sensitivity” rates. To interpret a positive tilt test, the wide variation of positivity rates with tilt table testing must be appreciated.2 Factors reported to affect positivity rates include the clinical presentation (typical vasovagal symptoms vs unexplained syncope vs unexplained falls), the age of the patient, the duration of tilting, the angle of tilting used in the protocol, saddle versus footplate support, and the use of pharmacological augmentation. Hargreaves et al. noted that false-positive rates may be as high as 46% if subjects are exposed to 80° tilting augmented by isoprenaline infusion and as high as 67% when saddle support is used instead of the normal footplate.3 Hackel et al. reported that all five of their older control subjects experienced syncope during a 60° tilt protocol.4 Given the modest rates of positive tilt table testing in syncope patients who also have carotid sinus hypersensitivity (but whose syncopal episodes were not associated with maneuvers that were likely to have caused direct stimulation of the carotid sinus), one must consider the possibility that the positive tilt test and the positive carotid massage are epiphenomenon, much the same as the impaired heart rate responses to cough and deep breathing observed in older syncope patients.5 Another point that deserves emphasis is the exclusion of patients with carotid sinus syndrome, organic heart disease, sick-sinus syndrome, intraventricular conduction defects, orthostatic hypotension, chronic and paroxysmal atrial fibrillation, and permanent pacemakers. Although these exclusion criteria are appropriate for defining the “normal response” to HUT, given the increasing prevalence of each of these conditions with advancing age, the clinician is left to wonder what percentage of older patients with syncope is actually represented by this cohort. Studies suggest that orthostatic hypotension is present in 5% to 25% of community-dwelling older people, and, as noted above, some authors have contended that carotid sinus hypersensitivity may represent a form of a neurally mediated syncope. In an effort to define a homogeneous study group, they also have excluded syncope patients with sick sinus syndrome, although Brignole et al.6 have suggested that “the autonomic disorder” appears to be unrelated to the sinus node dysfunction. Although methodologically more “pure,” this limits the generalizability of the results. The reader should not interpret the above remarks as casting aspersions on the value of the current study. On the contrary, the current investigation appears methodologically sound and provides valuable information about the use of the described protocol in older patients, particularly because beta-agonist augmentation may decrease specificity, and its use is discouraged in older patients unless the presence of coronary artery disease has been excluded.2,7 Given that tilt table testing has reportedly been useful in evaluation of patients with recurrent vertigo of unknown origin with near syncope,8 differentiation of convulsive syncope from epilepsy,9 and distinguishing neurocardiogenic syncope from Pacemaker Syndrome,10 in addition to its wide use in evaluating unexplained syncope, it is clear that many regard a positive tilt test as equivalent to neurally mediated syncope. The reader is simply cautioned that equating a positive tilt test with vasovagal or neurally mediated syncope may be an overly simplistic generalization. The most popular current theory about the pathogenesis of neurally mediated syncope holds that most cases result from stimulation of the Bezold-Jarisch reflex. (Stimulation of stretch receptors in the ventricle causes signals carried by unmyelinated C fibers in the vagus nerve to stimulate vasomotor centers in the brainstem, in turn producing vagal efferent activation, cardiac sympathetic withdrawal, and peripheral sympathetic withdrawal (or cholinergic sympathetic vasodilatory stimulation.)) Nevertheless, reports of neurally mediated syncope in heart transplant recipients11,12 (with concomitant cardiac autonomic denervation) have prompted some to suggest that inward invagination of the walls of the underfilled great veins (when their intraluminal pressure no longer exceeds intrathoracic pressure) may result in “collapse firing” of the veno-atrial stretch receptors and trigger reflex systemic vasodilatation and bradycardia.13 This may explain the unexpected bradycardia often observed with hemorrhagic hypotension.14 The importance of the distinction is that strategies such as beta-blockade, which decrease the stimulus to ventricular stretch receptors, may have no role if the latter mechanism is operative. The next caveat about interpreting tilt table testing relates to reproducibility. Wide variations of reproducibility have been reported in the literature. Although negative tests are largely reproducible, positive responses have been reported to progressively decrease with subsequent testing and with increasing interval between the spontaneous syncopal episode and the tilt table test.15 Brooks et al. reported that only 31% of patients with unexplained syncope had reproducible vasodepressor responses on both days, when tilt testing was repeated on 2 consecutive days.16 Morillo et al. have reported that as many as 82% of patients with a positive test who underwent repeat tilt testing after 1 week of treatment with a placebo had a negative test.17 Similar results (50% negative tests upon repeat tilt testing in the placebo group) have been reported by others.18 Additionally, even when “positive,” the predominant mechanism (vasodepression vs cardioinhibition) may differ from test to test within the same subject. It is worth noting that the natural history of patients with neurally mediated syncope diagnosed by HUT testing is recurrent syncope in only 10% to 30%.19 Finally, those choosing to use HUT as part of their diagnostic armamentarium need to be aware of the false-positive rates in patients with other potential causes of syncope. Sneddon et al. reported the results of tilt testing in patients with hypertrophic cardiomyopathy (HCM).20 Positive tilt tests occurred in three of 20 (15%) HCM patients with syncope that was unexplained after extensive evaluation, one of six (17%) HCM patients with syncope proven to be related to ventricular arrhythmias, and six of 20 (30%) HCM patients with no history of syncope. Linzer et al.21 have reported a case of a positive tilt test in an apparently otherwise healthy patient ultimately documented to have ventricular tachycardia as the cause of her symptoms, and Almquist et al.22 have reported positivity rates as high as 22% in syncope patients with alternative explanations provided on the basis of electrophysiological testing. Finally, Englund et al.23 have recommended caution about interpreting tilt testing in patients with bifascicular block. Positive test results occurred in 32% of patients with no history of syncope and in 28% of patients with syncope that remained unexplained after an extensive invasive and noninvasive investigation. Ultimately, the clinician must appreciate that cerebral hypoperfusion mediated by any mechanism can ultimately cause syncope, and systemic blood pressure and heart rate may only be modest surrogates for cerebral perfusion. Observations of diminished cerebral blood flow and increased cerebral resistance in patients with positive tilt table tests despite the absence of hypotension or bradycardia emphasize this point.24 With these limitations in mind, the knowledgeable clinician can be prepared to use and interpret the results of tilt table testing in older patients with syncope.