Diagnostic criteria for vasovagal syncope based on a quantitative history

Transcription

1 European Heart Journal (2006) 27, doi: /eurheartj/ehi584 Clinical research Diagnostic criteria for vasovagal syncope based on a quantitative history Robert Sheldon 1 *, Sarah Rose 1, Stuart Connolly 2, Debbie Ritchie 1, Mary-Lou Koshman 1, and Michael Frenneaux 3 for the Syncope Symptom Study Investigators { 1 Cardiovascular Research Group, University of Calgary, 3330 Hospital Drive NW, Calgary, Alberta T2N 4N2, Canada; 2 McMaster University, Hamilton, Ontario, Canada; and 3 Wales Heart Institute, Cardiff, Wales, UK Received 7 June 2005; revised 18 August 2005; accepted 22 September 2005; online publish-ahead-of-print 13 October 2005 See page 253 for the editorial comment on this article (doi: /eurheartj/ehi663) KEYWORDS Vasovagal; Syncope; Diagnosis; Tilt test; Faint; Point scores; History Aims Our goal was to develop historical criteria for the diagnosis of vasovagal syncope. Methods and results We administered a 118-item historical questionnaire to 418 patients with syncope and no apparent structural heart disease. The prevalence of each item was compared between patients with positive tilt tests and those with syncope of other, known causes. The contributions of symptoms to diagnoses were estimated with logistic regression, point scores were developed, and the scores were tested using receiver operator characteristic analysis. The accuracy of the decision rule was assessed with bootstrapping. Data sets were complete for all subjects. The causes of syncope were known in 323 patients and included tilt-positive vasovagal syncope (235 patients) and other diagnoses such as complete heart block and supraventricular tachycardias (88 patients). The point score correctly classified 90% of patients, diagnosing vasovagal syncope with 89% sensitivity and 91% specificity. The decision rule suggested that 68% of an additional 95 patients with syncope of unknown cause and a negative tilt test have vasovagal syncope. Conclusion A simple point score of historical features distinguishes vasovagal syncope from syncope of other causes with very high sensitivity and specificity. Introduction Vasovagal syncope is the most common cause of syncope and can be troublesome to diagnose. As with other causes of total loss of consciousness, patients have no recollection of their unconsciousness, and the reliability of the diagnosis of the first loss of consciousness is quite low 1. Many causes of syncope, although uncommon, have a treatable cause and/ or an eventual poor outcome, while vasovagal syncope is not associated with excess mortality. The concern about identifying treatable arrhythmic and other causes, and for some causes, an eventual poor outcome 2 coupled with the diagnostic uncertainty based upon the history has led to aggressive investigations of patients with syncope. Frequently, these are expensive, 3,4 invasive, and all too commonly inaccurate or inefficient. 5 8 Although the cause of syncope is eventually discovered in most patients who are referred to specialists, there is a need for a simple diagnostic tool for other settings. These might include primary care offices and emergency wards, and in epidemiologic, therapeutic, and health services studies. In these contexts, the current diagnostic tool, tilt testing, may not be feasible. * Corresponding author. Tel: þ ; fax: þ address: sheldon@ucalgary.ca { The Syncope Symptom Study Investigators are listed in the Appendix. We aimed to develop simple diagnostic criteria for distinguishing vasovagal syncope from other causes of syncope in patients with structurally normal hearts, on the basis of a careful history targeted at specific features. We began with patients having positive tilt tests, as is the current conventional practice. We prospectively administered a structured questionnaire to syncope patients with a positive tilt test and quantitatively compared the responses with those of patients with syncope due to other known causes. 9 From this, we developed a diagnostic point score to discriminate vasovagal syncope from syncope with other known causes. Methods Syncope symptom study The University of Calgary Conjoint Medical Ethics Review Committee and similar committees in all participating centres approved this study. Patients were enrolled in Calgary, Canada; Hamilton, Canada; and Cardiff, United Kingdom. Inclusion criteria Consenting patients were eligible if they had had one or more loss of consciousness. They were recruited from sequentially identified patients in university and private practice cardiology clinics, pacemaker clinics, arrhythmia and syncope clinics, and hospital cardiology wards. They were included in the Syncope Symptom Study if they had a diagnosis established according to preset criteria & The European Society of Cardiology All rights reserved. For Permissions, please journals.permissions@oxfordjournals.org

2 Diagnostic criteria for vasovagal syncope 345 (discussed subsequently), if there was no reasonable diagnostic confusion, or if reasonable investigations failed to elicit a diagnosis. That is, they were included if the diagnosis was known, or unknown despite reasonable attempts to elucidate one. Exclusion criteria Patients were excluded if they refused consent, were incapable of completing the questionnaire, or had more than one plausible cause of one syncopal spell. Patients with a history of known or suspected cardiomyopathy, or prior myocardial infarction, had the cardiac diagnosis confirmed by echocardiography, gated angiography, or cardiac catheterization, and were then excluded. The purpose of the larger study, the Syncope Symptom Study, was to develop historical diagnostic criteria for several classes of loss of consciousness. 9 There were 671 subjects, of whom 102 had epileptic seizures and 151 had syncope in the setting of structural heart disease. The remaining 418 patients with syncope in the apparent absence of structural heart disease are the subjects of this report. Gold standard diagnostic criteria Patients had vasovagal syncope if they had a positive tilt test performed according to one of several currently acceptable methods We chose not to specify that a particular tilt test protocol be used. This might have increased the internal validity of the study, but because each protocol has its own specificity and positive yield, we left the particular protocol up to the participating centre to increase external validity. It was deemed positive if it induced clinically reminiscent pre-syncope or syncope, and hypotension, bradycardia, or both. The specific degree of hypotension and bradycardia were pre-specified and depended on the protocol. These and other gold standard diagnostic criteria are listed in Table 1. All patients without an otherwise proven cause of syncope had an ECG and a tilt table test. Patients 60 years old usually underwent ambulatory electrocardiography. Syncope symptom questionnaire All patients completed a structured questionnaire 9 with 118 items developed from that of Calkins et al. 17 It assessed symptom burden, provocative situations, peri-syncopal symptoms, symptoms thought to be diagnostic of seizures, signs observed by bystanders, and relevant medical history. Completed questionnaires were checked for completion by study co-ordinators, and incomplete questionnaires were returned for revision. We cross-checked for incompatible entries to assess the accuracy of questionnaire completion. Study definitions In this article, we use working definitions, rather than pre-emptorily linking tilt test results to the term vasovagal syncope. Primary syncope denotes syncope of no apparent other cause in patients with either a positive tilt test (tilt-positive primary syncope) or a negative tilt test (tilt-negative primary syncope). Secondary syncope denotes syncope due to a known cause other than vasovagal syncope. We developed a point score that discriminated patients with primary tilt-positive syncope from patients with secondary syncope. Statistical analysis 9 First, we compared the prevalence of each variable in the tiltpositive primary syncope and secondary syncope groups using a Fisher s exact test and calculated the likelihood ratio for predicting the diagnosis of tilt-positive primary syncope. The likelihood ratio of each variable is its prevalence in the tilt-positive primary syncope group divided by its prevalence in the secondary syncope group. A variable with a likelihood ratio.1 is predictive of tilt-positive Table 1 Gold standard diagnostic criteria for patient inclusion and working diagnostic definitions used in this study during analysis Diagnosis Definition Gold standard diagnoses Vasovagal syncope Positive tilt test using a currently acceptable method Ventricular Sustained VT documented during tachycardia syncope, or 48 h of admission for syncope, or if haemodynamically unstable, sustained monomorphic VT was induced during an electrophysiologic study 15 Torsade de Pointes Documented at the time of syncope or shortly afterwards with classic clinical features Supraventricular Established diagnosis of arrhythmia, tachycardia and typical palpitations immediately preceding a syncopal spell, or awoke from syncope with those symptoms Complete heart block Complete heart block with a wide QRS escape rhythm documented during syncope or shortly afterwards Orthostatic hypotension Documented autonomic neuropathy with significant orthostatic hypotension and presyncope or syncope during tilt table testing 16 Working diagnostic definitions during analysis Primary syncope Syncope of no apparent other cause with either a positive tilt test (tilt-positive primary syncope) or negative tilt test (tilt-negative primary syncope) Secondary syncope Syncope due to a known cause other than vasovagal syncope primary syncope and a variable with a ratio,1 is predictive of secondary syncope. Binary variables were created from continuous variables using a cut-point, which was the mean of the entire sample for symmetrically distributed variables and the median for variable that were very skewed. Variables with a very low prevalence were combined into a composite variable provided that both the magnitude and significance of the likelihood ratio were about the same and that the combination made clinical sense. Variables with similar meanings and similar descriptive statistics were also combined in order to reduce the number of variables eligible for inclusion in the following logistic regression analysis. We then developed a manual stepwise logistic regression model that predicted tilt-positive primary syncope. Variables were eligible for inclusion in the logistic regression model if they were significant predictors (P, 0.01) in the univariate analysis and were retained in the multivariable model if P, 0.05 for the Wald statistic. No interactions between the variables were assessed. To derive a practical diagnostic decision rule, a point score was developed by assigning +1, +2, +3 (and so on) points to each of the factors based on the relative magnitude of the estimated regression coefficient. Each coefficient was divided by the smallest absolute value of the coefficient retained in the model, then rounded to the nearest integer. The points were summed and a diagnostic threshold chosen using receiver operating characteristic analysis. 18 Using the diagnostic threshold, the apparent sensitivity, specificity, and overall accuracy in this test sample were estimated. When the predictive accuracy of a diagnostic decision rule is assessed using the same patients from whom it was developed, an

3 346 R. Sheldon et al. unduly optimistic classification rate might occur. 19 This is referred to as the apparent accuracy. Ideally, the true accuracy should be evaluated using a split sample, 9 but in this case, our sample size for the secondary syncope group was not large enough. Accordingly, we used a bootstrap analysis to estimate the excess error that would be detected if an independent sample was used to evaluate the diagnostic accuracy. 19 Five thousand bootstrap samples were drawn from the original sample. For each bootstrap sample, the logistic regression model with the same variables as those in the adopted model was fitted to re-estimate the regression coefficients and the diagnostic decision rule developed. The accuracy of the new decision rule from the bootstrap sample was then assessed in the original sample. The expected excess error rate was calculated from the 5000 bootstrap samples, then subtracted from the apparent accuracy to obtain estimates of the true sensitivity, specificity, and overall accuracy of the diagnostic decision rule. We derived classification rules both with and without the variables of symptom burden and durations. Finally, we compared the symptoms of patients with gold standard diagnoses with those with no identifiable cause of syncope by two methods. First, we determined the diagnostic performance of the point score for tiltpositive primary syncope in patients with tilt-negative primary syncope. Secondly, we compared the tilt-positive primary syncope patients with tilt-negative primary syncope patients, both of whom fit the derived diagnostic criteria for tilt-positive primary syncope. Results Patient population (Table 2) Patient enrolment occurred between January 1995 and July All patients who were approached consented, and only two were disqualified, because they had more than one plausible cause of syncope. There were 235 patients with syncope and positive tilt tests (tilt-positive primary syncope), 95 patients with no apparent cause of syncope and negative tilt tests (tilt-negative primary syncope), and 88 patients with other identified causes of syncope (secondary syncope). Of the latter, 42 had complete heart block, 21 had a supraventricular tachycardia, six had idiopathic ventricular tachycardia, five had aortic stenosis, three had Torsades de Pointe ventricular tachycardia, cough syncope, or hypersensitive carotid sinus syncope, two had pulmonary emboli, and one had ischaemic ventricular tachycardia, hyperventilation syncope, or autonomic neuropathy. The tilt-positive primary syncope patients were younger than the secondary syncope patients, and more were women. The tilt-positive primary syncope patients had had more syncopal spells than the secondary syncope patients (P, 0.001). Symptoms were present in the tilt-positive primary syncope patients over a longer duration of time (P, 0.001) compared with the secondary syncope patients. The mean age for the tilt-negative primary syncope patients was years, and 54% were women. They had a median five syncope spells (IQR 3 12) over a median 29 months (IQR 3 132). Classification score The most important univariate features that distinguish between patients with tilt-positive primary syncope and with secondary syncope are presented in Table 3. Of the original 118 items, 71 were significantly predictors of primary vs. secondary syncope at P, These 71 variables were reduced to 34 variables by appropriate combination as described earlier, and descriptive information is presented in Table 3. These 34 variables were eligible for inclusion in the logistic regression model, seven of which were included in the final model. After including these seven variables in the multivariable logistic regression model, none of the remaining variables were significant at P, Features were included in the table if the likelihood ratio (see Methods) was.1.0 (predictive of tilt-positive primary syncope) or,1.0 (predictive of secondary syncope) with P A diagnostic point score was derived from the relative weighting of the regression coefficients without inclusion of estimates of symptom burden (Table 4). The diagnostic behaviour of this point score is illustrated in Figure 1. Patients were classified as having tilt-positive primary syncope if the point score was 22. This resulted in an apparent overall accuracy of 91%, apparent sensitivity of 90%, and apparent specificity of 93% (Figure 2). When the bootstrap estimates of excess error were included, the true overall accuracy was 89.7%, with a true sensitivity of 89.3% and a true specificity of 90.8%. Features of the history that contributed to the diagnosis of tilt-positive primary syncope included pre-syncope or syncope with exposure to pain or a medical procedure, diaphoresis or perceived warmth before a spell, and pre-syncope or syncope with prolonged sitting or standing. Factors that were strongly predictive of syncope secondary to other causes included known bifascicular block, asystole, supraventricular tachycardia, or diabetes; observed cyanosis; a memory of the spell; and older age at onset of syncope. The measures of symptom burden were nonsignificant after including the aforementioned variables in the model and did not improve the diagnostic ability of the model. Syncope of unknown cause The diagnostic scheme classified 68% of patients with tiltnegative primary syncope as though they had tilt-positive primary syncope. The similar distributions of the point Table 2 Clinical and demographic characteristics of three patient groups Characteristic Tilt-positive primary syncope Secondary syncope P-value, tilt-positive primary vs. secondary Tilt-negative primary syncope Number Age, years (mean, SD) , Females 61% 45% % Syncopal spells, n (Median, IQR) 6 (3 20) 2 (1 5), (3 12) Symptom duration, months (Median, IQR) 100 (13 268) 1 (0 16), (3 132)

4 Diagnostic criteria for vasovagal syncope 347 Table 3 Comparison of the most significant historical features in patients with tilt-positive primary syncope compared with other known causes of syncope (secondary syncope) Variable Prevalence (%) Sensitivity (%) Specificity (%) LHR P-value Age at first syncopal spell , Less than 5 s warning , Tired after syncopal spell , Provocative situations Syncope/pre-syncope with pain or , medical procedure Syncope/pre-syncope with hot or , warm environments Syncope/pre-syncope with stress , Syncope/pre-syncope with headaches , Syncope/pre-syncope with prolonged , sitting/standing Syncope/pre-syncope on the way to the toilet Syncope/pre-syncope after using the toilet a Pre-syncope after exercise Behaviour during syncope Cannot remember behaviour during syncope Unresponsive during syncope , No memory about syncope Confusion after a spell Cyanotic during syncope White or pale colour noted by bystander Medical history Valvular heart disease Atrial fibrillation or flutter Any one of bifascicular block, diabetes, , supraventricular tachycardia, diabetes Hypertension ,0.001 Perisyncopal symptoms Nausea or vomiting before a spell Sweating or warm feeling before a spell , Headache before a spell Visual distortion before a spell Auditory distortion before a spell Heart racing before a spell Abdominal rising sensation Numbness or tingling before a spell Mood changes or pre-occupation , before a spell Sweating or warm feeling after a spell Mood changes after a spell Numbness or tingling after a spell Nausea or vomiting after a spell , The behaviour of each of the variables is expressed in terms of the prevalence in the sample, sensitivity, specificity, and likelihood ratio for tilt-positive primary syncope, as diagnosed by a positive tilt table test. LHR, likelihood ratio. a The likelihood ratio cannot be calculated because the specificity is equal to 100%. scores in the tilt-negative and tilt-positive primary syncope patients when compared with the patients with secondary syncope are shown in Figure 3. Finally, we assessed whether any of the 71 original significant predictors discriminated between patients whose point scores were 22, and who had either positive or negative tilt tests. Only four out of 71 variables were statistically significant, with P ¼ When 71 independent tests are performed, the probability of four or more being significant due to chance alone at the 5% level of significance is 0.48, indicating that there is little evidence of any significant difference between these two groups. Similarly, only two of the 34 variables presented in Table 3 were significant at P, Discussion Clinical implications In patients with a transient loss of consciousness, there is substantial diagnostic disagreement among physicians, and

5 348 R. Sheldon et al. Table 4 Point scores for the diagnosis of tilt-positive primary syncope, in the absence of knowledge of the numbers and historical duration of syncope and pre-syncope Regression coefficient (SE) P-value Points Any one of bifascicular block, (0.76), asystole, supraventricular tachycardia, diabetes Blue colour noted by bystander (1.26), Age at first syncope 35 years (0.63), Remembers something about the spell (0.53), Pre-syncope or syncope with prolonged 0.95 (0.49) sitting or standing Sweating or warm feeling before 1.95 (0.56), a spell Pre-syncope or syncope with pain or medical procedure 2.90 (0.85), Classify as tilt-positive primary syncope for points 22. Figure 1 Sensitivity and specificity for the diagnosis of tilt-positive primary syncope compared with secondary syncope using the point scores reported in Table 2. Figure 2 Receiver operator characteristic analysis of the diagnostic score developed in Table 2. The populations are those of patients with primary tilt-positive syncope and patients with secondary syncope. written diagnostic criteria improve inter-rater agreement. 1,20 Although tilt table testing has provided important diagnostic and physiologic insights, it is not readily available in emergency wards and family physicians offices, and by codifying diagnosis, the score may streamline management. Similarly, tilt testing may not be feasible in larger community-based studies of people with syncope who are not patients. Varying the diagnostic cut-off level would allow investigators to vary the sensitivity, specificity, and predictive values of the score depending on the need of the study. The point score was developed from structured questionnaires prospectively administered to patients with apparently structurally normal hearts, and not from a post hoc chart review. Tilt tests were selected as the entry point for patients with presumed vasovagal syncope, because the tilt table test was developed to mimic some of the physiology of vasovagal syncope; they provoke symptoms reminiscent of the patients syncopal spells; and because positive tilt tests are accompanied by hypotension and bradycardia, as is vasovagal syncope. Features of questionnaire The classification scheme is simple, and to some extent confirms previous empiric diagnostic criteria. We present derived diagnostic questions and their points in Table 5. The patient has vasovagal syncope if the point score is 22 and syncope due to other causes if the point score is,22. Some of the diagnostic factors in favour of vasovagal syncope in the model reflect the situations in which vasovagal syncope occurs. These include prolonged sitting or standing and exposure to medical situations. The age criterion reflects the young age at which vasovagal syncope usually first occurs. However, these used alone were not necessarily sufficient. Four risk factors for other causes of syncope, asystole or bifascicular block on an ECG, supraventricular tachycardia, and diabetes, usually need to be absent to establish a diagnosis of vasovagal syncope. The classification scheme was not improved with the addition of the criteria for the duration of the history or number of syncopal spells. This suggests that patients can be accurately diagnosed at first presentation and that health care wait lists do not confound the diagnosis. Other studies Calkins et al. 17 reported a prospective, evidence-based study of the symptoms of patients with ventricular tachycardia, complete heart block, or tilt test-positive vasovagal

6 Diagnostic criteria for vasovagal syncope 349 following the tilt test, and behave similarly in a multivariable predictive model of pre-test variables for post-test outcome Figure 3 A comparison of point score distributions for populations of tiltpositive and tilt-negative primary syncope patients. Table 5 Diagnostic questions to determine whether syncope is due to vasovagal syncope or to another cause of syncope Question syncope. Most patients had structural heart disease. Alboni et al. 21 reported symptoms in 341 unselected syncope patients, many of whom had structural heart disease. The study was aimed at diagnosing cardiac causes of syncope, and a history of heart disease was the most important factor. Although there were highly significant differences in symptom distribution among the diagnostic groups in the preceding two reports, the results did not culminate in diagnostic point scores. Negative and positive tilts Points (if yes) Is there a history of at least one of 25 bifascicular block, asystole, supraventricular tachycardia, diabetes? At times have bystanders noted you to be blue 24 during your faint? Did your syncope start when you were 35 years 23 of age or older? Do you remember anything about 22 being unconscious? Do you have lightheaded spells or faint 1 with prolonged sitting or standing? Do you sweat or feel warm before a faint? 2 Do you have lightheaded spells or faint with 3 pain or in medical settings? The patient has vasovagal syncope if the point score is 22. The close symptomatic similarity between most patients with negative and positive tilt tests is consistent with the idea that most patients with negative tilt tests have falsely negative tests. Given that the score is 90% sensitive for tilt-positive primary syncope, and that 68% of tiltnegative primary syncope patients fall in the same diagnostic category, it may be that up to 76% of tilt-negative primary syncope patients truly have vasovagal syncope. Patients with negative and positive tests have similar symptom burdens, similar clinical outcomes in the 3 years Limitations Patients were accrued from tertiary care clinics and acute care facilities, and these findings need to be tested in a broadly based population. There may be some patients with rare presentations of common disorders who went undiagnosed. We did not specify a particular tilt test protocol to be used. Although this might have increased the internal validity of the study, we recognized that each protocol has its own specificity and positive yield, and therefore, left the particular protocol up to the participating centre to increase external validity. Similarly, we did not specify the investigations necessary to diagnose the absence of structural heart disease. By leaving this to individual centres, we hoped to derive diagnostic criteria driven by and relevant to conventional clinical practice. We assessed the optimism error of the diagnostic schemes with bootstrapping, and independent confirmation is desirable. These criteria should be tested in relatively uncommon syndromes such as long QT syndrome and Brugada syndrome. Another strategy to the development of a diagnostic point score would have been to develop it using subjects with what are assumed to be classic histories of vasovagal syncope, but these criteria are not evidence-based and might be self-fulfilling. Accordingly, we started with tilt test criteria and worked independently towards a historical definition. We hoped this would avoid excessive bias. The point score may reflect regional and referral idiosyncrasies of the study population tested, as well as the particular tilt tests used. The limitations of tilt tests as diagnostic tools were reviewed recently. 8 The questionnaire was administered to some patients after their relevant diagnostic studies, and this might have reinforced recollection of symptoms specific to the test outcomes. The effect of age on recollection of symptoms is unknown, as is the use of the point score in elderly patients. We do emphasize that the point score is applicable only to patients without structural heart disease. These concerns underlie the importance of assessing the accuracy and utility of the point score in other populations and settings. Indeed, it has been used in a population-based survey of syncope, is an inclusion measure in an ongoing multinational randomized clinical trial, and will be used in upcoming epidemiological studies in community and acute care sectors. Clinicians should use these criteria only as a guide to diagnosis. Acknowledgement This study was funded by Medtronic, Toronto, Ontario, Canada; the Calgary General Hospital Volunteers Foundation; and the Canadian Institutes for Health Research, Ottawa, Ontario, Canada. Conflict of interest: none declared. Appendix The following investigators participated: University of Calgary, Calgary, Alberta: R. Sheldon, D. Ritchie, M. Koshman, B. Mitchell, D. Wyse, H.J. Duff, A. Gillis, M.A. Lee, W.F. Murphy; Hamilton Health Sciences Corporation, Hamilton, Ontario: S. Connolly;

7 350 R. Sheldon et al. Wales Heart Institute, Cardiff, Wales: M. Frenneaux, J. Pascual. St Michael s Hospital, Toronto, Ontario: D. Newman. The steering committee consisted of Dr Sheldon, Dr S Rose, and Ms Ritchie. References 1. Hoefnagels WAJ, Padberg GW, Overweg J, Roos RAC. Syncope or seizure? A matter of opinion. Clin Neurol Neurosurg 1992;94: Kapoor WN, Karpf M, Wieand S, Peterson JR, Levey GS. A prospective evaluation and follow-up of patients with syncope. N Engl J Med 1983;309: Calkins H, Byrne M, El-Tassi R, Kalbfleisch S, Langberg JJ, Morady F. The economic burden of unrecognized vasodepressor syncope. Am J Med 1993;95: Sutton R, Petersen M. The economics of treating vasovagal syncope. Pacing Clin Electrophysiol 1997;20: Linzer M, Yang EH, Estes III MA, Wang P, Vorperian VR, Kapoor W. Diagnosing Syncope. Part I. Value of history, physical examination, and electrocardiography. Ann Intern Med 1997;126: Linzer M, Yang EH, Estes III MA, Wang P, Vorperian VR, Kapoor W. Diagnosing Syncope. Part II. Unexplained syncope. Ann Intern Med 1997;127: Brignole M, Alboni P, Benditt DG, Bergfeldt L, Blanc JJ, Thomsen PE, Gert van Dijk J, Fitzpatrick A, Hohnloser S, Janousek J, Kapoor W, Kenny RA, Kulakowski P, Masotti G, Moya A, Raviele A, Sutton R, Theodorakis G, Ungar A, Wieling W, Priori SG, Garcia MA, Budaj A, Cowie M, Deckers J, Burgos EF, Lekakis J, Lindhal B, Mazzotta G, Morais J, Oto A, Smiseth O, Menozzi C, Ector H, Vardas P; Task Force on Syncope, European Society of Cardiology. Guidelines on management (diagnosis and treatment) of syncope-update Executive Summary. Eur Heart J 2004;25: Sheldon R. Tilt testing for syncope: a reappraisal. Curr Opin Cardiol 2005;20: Sheldon R, Rose S, Ritchie D, Connolly SJ, Koshman M, Lee MA, Frenneaux M, Fisher M, Murphy W. Historical criteria that distinguish syncope from seizures. J Am Coll Cardiol 2002;40: Sheldon R. Evaluation of a single-stage isoproterenol-tilt table test in patients with syncope. J Am Coll Cardiol 1993;22: Morillo CA, Klein GJ, Zandri S, Yee R. Diagnostic accuracy of a low-dose isoproterenol head-up tilt protocol. Am Heart J 1995;129: Benditt DG, Ferguson DW, Grubb BP, Kapoor WN, Kugler J, Lerman BB, Maloney JD, Raviele A, Ross B, Sutton R, Wolk MJ, Wood DL. Tilt table testing for assessing syncope. J Am Coll Cardiol 1996;28: Fitzpatrick AP, Theodorakis G, Vardas P, Sutton R. Methodology of headup tilt testing in patients with unexplained syncope. J Am Coll Cardiol 1991;17: Sheldon R, Killam S. Methodology of isoproterenol-tilt table testing in patients with syncope. J Am Coll Cardiol 1992;19: Sheldon RS, Rose S, Koshman ML. Isoproterenol tilt-table testing in patients with syncope and structural heart disease. Am J Cardiol 1996; 78: Grubb BP, Karas B. Clinical disorders of the autonomic nervous system associated with orthostatic intolerance. An overview of classification, clinical evaluation, and management. Pacing Clin Electrophysiol 1999; 22: Calkins H, Shyr Y, Frumin H, Schork A, Morady. The value of the clinical history in the differentiation of syncope due to ventricular tachycardia, atrioventricular block, and neurocardiogenic syncope. Am J Med 1995; 98: Metz CE. Basic principles of ROC analysis. Semin Nucl Med 1978;8: Efron B. The Jacknife. The Bootstrap and Other Resampling Plans. Philadelphia: Society for Industrial and Applied Mathematics; Van Donselaar CA, Geerts AT, Meulstee J, Habbema JDF, Staal A. Reliability of the diagnosis of a first seizure. Neurology 1989;39: Alboni P, Brignole M, Menozzi C, Raviele A, Del Rosso A, Dinelli M Solano A, Bottoni N. Diagnostic value of history in patients with syncope with or without heart disease. J Am Coll Cardiol 2001;37: Sheldon R, Rose S, Koshman ML. Comparison of patients with syncope of unknown cause having negative or positive tilt-table tests. Am J Cardiol 1997;80: Grimm W, Degenhardt M, Hoffman J, Menz V, Wirths A, Maisch B. Syncope recurrence can better be predicted by history than by head-up tilt testing in untreated patients with suspected neurally mediated syncope. Eur Heart J 1997;9: Moya A, Brignole M, Menozzi C, Garcia-Civera R, Tognarini S, Mont L, Botto G, Giada F, Cornacchia D. International Study on Syncope of Uncertain Etiology (ISSUE) Investigators. Mechanism of syncope in patients with isolated syncope and in patients with tilt-positive syncope. Circulation 2001;104: