The adenosine triphosphate test in the diagnosis of unexplained syncope: a test looking for a home

Adenosine is a ubiquitous purine nucleoside which acts on a group of specific G-protein-coupled adenosine receptors. In the heart, stimulation of predominantly the A1 adenosine results in potent chronotropic and dromotropic effects. These actions have led to its established role as a therapeutic agent in the management of supraventricular tachycardia.1 However, its use as a diagnostic agent in the management of syncope, principally in syncope associated with significant bradycardia, remains unclear, despite interest over the last 15 years. It is a test looking for a home. The adenosine triphosphate (ATP) test has garnered increasing interest as a potential diagnostic tool in unexplained syncope since the publication of two seminal papers in 19972,3 and involves the intravenous administration of a 20 mg bolus of ATP followed by a flush with continuous electrocardiogram (ECG) and non-invasive blood pressure monitoring. A positive test is defined as either a cardiac pause (the time from the last conducted sinus beat to the next conducted sinus beat) ≥10 s2 or asystole ≥6 s.3 These values are derived from control populations in whom a cardiac pause of 10s was the upper 94th percentile in one group (n = 51, 59% male, age 56 ± 2.4 years),2 while 6000 ms was the upper 95th percentile of the maximum RR interval distribution in the other (n = 90, 50% male, age 55 ± 17 years).3 Despite a growing and suggestive evidence base, the ATP test has yet to become a routine test in the diagnosis of unexplained syncope, with the most recent iteration of the European Society of Cardiology (ESC) syncope guidelines4 giving the test a Class III (Level of Evidence B) recommendation, despite its Class II recommendation in previous iteratations of the guidelines. This is because studies have failed to demonstrate a consistent correlation between the ATP test result and the rhythm disturbance during spontaneous syncope documented on an implantable loop recorder (ILR).5,6 The most recent ESC and European Heart Rhythm Association guidance on cardiac pacing and resynchronization therapy7 is a little more favourable, giving a Class IIb recommendation for pacing in unexplained syncope and a positive ATP test, citing more recent data (discussed below) as evidence.8,9 Thus, while the test has been upgraded in terms of evidence level, the lack of positive recommendations from other sources including the most recent European pacing guidelines4,7 serves only to muddy the waters further. One of the key issues concerning the ATP test is the uncertainty surrounding the diagnosis revealed by a positive test. There are two principal schools of thought. The first2 is that a positive test identifies underlying cardioinhibitory neurally mediated syncope. The second is that a positive test identifies latent conducting system disease resulting in syncope secondary to bradycardia.10 There is evidence to support both theories, although that which supports the latter theory is more compelling. In the paper by Flammang et al.2 that established a cardiac pause ≥10 s as indicating a positive ATP test, 41% (n = 141) patients with syncope or presyncope had a positive test compared with only 6% of the control group. Despite the absence of tilt testing to support the diagnosis and a higher likelihood of a positive ATP test with advancing age (when one has a higher index of suspicion for conducting system disease), the authors concluded that a positive test identified severe cardioinhibitory neurally mediated syncope. Brignole et al.11 compared ATP testing with tilt test results in 175 individuals with unexplained syncope. Of the 121 patients with a positive test, 64% (n = 77) were tilt test positive, 15% (n = 18) were ATP-test positive, and 21% (n = 26) were both, suggesting the ATP test does not reliably identify neurally mediated syncope. Similar to the study by Flammang et al.,2 those with a positive ATP test were older and, in addition, had more cardiovascular co-morbidity (increasing the likelihood of conducting system disease). More recently, in a cohort of 25 patients with head-up tilt test-positive vasovagal syncope, cardioinhibition was present in 8 of 25 patients (32%), and the ATP test was also positive in 9 of the 25 (36%) patients.6 However, there was no correlation between bradycardia on implantable loop recording, a cardioinhibitory tilt response or a positive ATP test. This small study was performed in a relatively young group of patients (mean 60.2 ± 17.1 years)6 with vasovagal syncope, and suggests that an ATP test is unlikely to reliably diagnose cardioinhibitory vasovagal syncope. Equally, given the small numbers of patients, it may simply more accurately reflect the difficulties in diagnosing syncope of unknown origin. Studies investigating the role of ATP in unmasking conducting system disease as a cause of syncope have examined both sinus and atrioventricular (AV) nodal disease. Brignole et al.3 studied 9 individuals with spontaneous syncope secondary to sinus arrest and 15 secondary to AV block. A positive ATP test was noted in 53% of those with AV block and 0% of those with sinus arrest; the authors postulating that a positive ATP test in those with unexplained syncope points to paroxysmal AV block as a cause. This is supported by recent work from the same group9 where an ATP test was positive in 15 of 18 (83%) individuals with unexplained syncope that subsequently had spontaneous AV block demonstrated on prolonged monitoring (mean follow-up 4 ± 4 years). An invasive electrophysiology study failed to demonstrate impaired AV conduction in any participants and while tilt testing induced syncope in seven (41%) participants, AV block was never induced. This prompted the authors to describe ‘idiopathic paroxysmal AV block’ in which ATP testing seemed to unmask with reasonable sensitivity. More definitively, in a cohort study of 50 individuals10 adenosine was 100% sensitive and 86% specific for identifying bradycardia pacing indications (a mixture of sinus node disease, AV block and cardioinhibitory carotid sinus syndrome) and a smaller study of 10 patients demonstrated adenosine to be 80% sensitive and 97% specific for sinus node disease.12 This last study is supported by recent work showing that ATP testing was comparable with the corrected sinus node recovery time (CSNRT) at EP study in identifying sinus node disease (94% sensitivity and 84% specificity for adenosine testing vs. 74% sensitivity and 100% for CSNRT).13 Given the higher likelihood of a positive ATP test in older patients, the majority of whom were women, it may be that the test will be of particular benefit in unmasking conducting system disease in this group.14 The presence of more convincing evidence of a positive ATP test implicating conducting system disease leads to the randomized controlled trials that involve pacing therapy consequent upon a positive test.8,15 In a pilot study8 of 20 patients with syncope of presumed vasovagal origin, based on history alone, and a positive ATP test, 10 patients received usual care and the other 10 dual-chamber pacing; mean follow-up was 49.3 months. Estimated recurrence-free survival at five years was 100% in the paced group and 40% in the routine care group. This promising preliminary study was followed-up by a multicentre study;15 this time of ‘placebo’ pacing (AAI 30b.p.m.) vs. dual-chamber pacing (DDD) in unexplained syncope. The two year recurrence-free survival was 31% for the AAI group and 77% for the DDD group. This well-designed study's main flaw was that the pacing allocation was only single blind (the physician knew the allocation) and the blinding integrity is not reported. This is important, as the syncope literature is littered with observational studies of pacing therapy with initially promising results that have not been borne out by adequately blinded trials.16 Moreover, it took 5 years to recruit 80 subjects in 10 centres, probably because of exhaustive selection criteria suggesting that the findings may only be applicable to a small number of individuals with syncope.6 Despite these weaknesses, the study adds considerably to the pilot work and suggests that there is an emerging role for ATP testing to identify individuals with unexplained syncope that will benefit from pacing therapy. Of additional interest, the recently published Third International Study on Syncope of Unexplained Etiology (ISSUE 3)17 demonstrates that dual-chamber pacing can be effective in cardioinhibitory vasovagal syncope with ILR-documented asystole during spontaneous syncope. The principal limitation in everyday practice is that this would involve only 9% of all sufferers of vasovagal syncope and is a resource intensive way to identify them while exposing this population to the small but significant number of adverse events related to permanent pacing. There might be a role for the ATP test to identify this 9% at considerably less cost and by less invasive means but the very similar recruitment methods of ISSUE 25 and ISSUE 316 studies make this unlikely. One of the most important aspects to establish is what happens to those with unexplained syncope who have a negative ATP test? All studies to date have focused solely on those with a positive result, so the true sensitivity and specificity of the test have never been evaluated. The test will only prove useful if it can clearly discriminate between those who will benefit from a particular intervention and those that will not. This is being addressed in an ongoing clinical trial that uses an ILR to document any rhythm disturbance associated with spontaneous syncope in those with a negative adenosine test while testing the efficacy of permanent pacing in those with a positive test in a randomized double-blind placebo-controlled trial.18 Despite 15 years of interest, the ATP test is an investigation without a home. For it to enter routine clinical practice there needs to be clear evidence of three key principles: As has been outlined, there are proposed answers to the first two of these three questions; with conducting system disease and pacing therapy being the most likely candidates; but no answer at all to the third. Current trials will hopefully be more definitive in answering all three. However, until that time, the ATP test will remain homeless. What diagnosis is unearthed by a positive test? What therapy should be introduced in light of a positive test? What should happen to those that have a negative test? Conflict of interest: none declared.