The American Journal of Medicine
Volume 120, Issue 1 , Pages 54-62, January 2007

The Expectation Effect and Cardiac Pacing for Refractory Vasovagal Syncope

Division of Cardiology, Department of Medicine, University of Western Ontario, London, Ontario, Canada.

Article Outline

Abstract 

Background

Clinical trials of pacing for vasovagal syncope have shown conflicting results. We performed a meta-analysis to determine whether permanent pacemaker therapy prevents refractory vasovagal syncope.

Methods

Randomized trials comparing pacemaker therapy with medical therapy, usual care, placebo, or different pacing algorithms in the prevention of recurrent vasovagal syncope were considered. The primary endpoint was first recurrence of syncope.

Results

Nine randomized trials (2 double blind, 7 open label or single blind) were included. There was significant heterogeneity when all 9 trials were pooled (P=.0009 and I2=69.6%), reflecting methodological diversity in blinding and the nature of the control therapy. When pooled by trial methodology, heterogeneity was no longer apparent. Permanent pacing reduced the risk of recurrent syncope in unblinded studies (odds ratio [OR] 0.09, 95% confidence interval [CI], 0.04 to 0.22) and in studies comparing pacemaker algorithms (OR 0.04, 95% CI, 0.0 to 0.23). No effect was seen in double-blinded trials (OR 0.83, 95% CI, 0.41 to 1.70). Awareness that a permanent pacemaker was implanted and functional was associated with a significant ‘expectation’ effect, which itself reduced the risk of recurrent syncope (OR 0.16, 95% CI, 0.06 to 0.40, P=.0001). Results were similar when restricted to patients with a marked cardioinhibitory response on baseline tilt table testing.

Conclusion

The results of small, preliminary trials have overestimated the treatment effect of pacemakers due to a lack of blinding of physicians and patients. Blinded trials suggest that the apparent response is due to a strong expectation response to pacing.

Keywords: Vasovagal syncope, Pacemakers, Meta-analysis, Placebo

 

Vasovagal syncope is the most frequent etiology of syncope where a cause can be identified.1 In the majority of cases, vasovagal syncope follows a benign course.1 Nonetheless, some patients experience recurrent syncope or presyncope despite lifestyle changes and empiric trials of pharmacological therapy.2 Permanent pacemakers have intuitive appeal as a means of counteracting the cardioinhibitory component of vasovagal syncope, but randomized controlled trials have shown conflicting results.3, 4, 5, 6, 7, 8, 9, 10, 11, 12

Clinical Significance

 


•Vasovagal syncope is the most common form of syncope.

•Pacemaker therapy has been proposed to treat the cardioinhibitory component of recurrent medically refractory vasovagal syncope.

•Unblinded trials suggest a substantial benefit from pacing, which is absent in trials involving rigorous blinding.

•A strong expectation effect is the best explanation for the substantial response in the unblinded studies.

We performed a systematic review and meta-analysis of randomized trials in which pacemakers were used to prevent syncope in patients with recurrent vasovagal syncope. The purpose of our study was to determine the efficacy of permanent pacing in preventing recurrent syncope, to determine the extent of a pacemaker “placebo effect,” and whether any effect modifiers exist that might identify a subgroup that benefits from pacemaker insertion.

Back to Article Outline

Methods 

Study Identification 

We used systematic methods to identify all published randomized controlled trials (RCT) comparing pacemaker therapy with medical or no therapy, or comparing one type of pacing algorithm with another, for the prevention of vasovagal syncope.13, 14, 15 The search strategy involved the MEDLINE, EMBASE, and Cochrane databases using the search terms: syncope, fainting, vasovagal syncope, neurocardiogenic syncope, neurally mediated syncope, neurogenic syncope, reflex syncope, vasodepressor syncope, cardioinhibitory syncope, artificial pacemaker, cardiac pacemaker, cardiac pacing, and randomized-controlled trial. We further searched PubMed using the “related article” feature, augmented by manual searches of reference lists from individual reports, review articles, and contact with clinical experts and authors. Unpublished data were searched through grey literature databases including SIGLE, NRR, and HSRProj, and ongoing trials via www.clinicaltrials.gov and www.controlled-trials.com. We included reports published as manuscripts or described in abstract form in any language from 1966 to January 2005.16, 17, 18

Study Eligibility 

Two investigators (S.S., P.L.S.) independently evaluated studies for possible inclusion and resolved any differences by consensus. Investigators were not blinded to the author, journal, or type of publication.19 Studies were included in the meta-analysis if: patients (age >18 years) with refractory vasovagal syncope were enrolled; there was a comparison of 2 or more groups, at least 1 of which was treated with a pacemaker; treatment groups were assigned through random allocation; and methods were used to assess for the development of recurrent vasovagal syncope or adverse events related to pacemaker therapy.

Study Quality 

We used a modification of the study-quality criteria outlined by Schultz and colleagues16 to assess each study that included: proper generation of treatment allocation sequence; proper concealment of the allocation sequence; patients lost to follow-up; and double-blinding.

Data Extraction 

Data were abstracted in duplicate (S.S., P.L.S.) with use of a standardized protocol and reporting form, and disagreements were resolved by consensus. The major outcomes included: the recurrence of syncope after enrollment and complications directly related to pacemaker therapy. The primary authors’ definitions were accepted.

Statistical Analysis 

Statistical calculations were performed with Review Manager (RevMan version 4.2.1, Oxford, England: The Cochrane Collaboration, 2003). The odds ratio was used as the measure of effect in the overall comparison test, because it is the most satisfactory metric to combine across trials with discrete outcomes. For assessment of internal consistency of the observations on the treatment effect, 3 separate methods of pooling the trial results were used: Peto method comparing observed minus expected events; Mantel-Haenszel method for combining information from a series of 2×2 tables; and the random-effects model of Der Simonian and Laird.20, 21, 22, 23, 24, 25

Different trials or groups of trials were compared using standard tests for heterogeneity or, where appropriate, tests for trend between the observed effects on outcome events.23 As these tests have low sensitivity for detecting heterogeneity, a more liberal level of statistical significance (P <.1) was assumed.26 In addition, an I2 test of heterogeneity, which describes the percentage of total variation across studies that is due to heterogeneity rather than due to chance, was performed for each of the comparisons.27 Heterogeneity has been described as low, moderate, and high based on I2 values of 25%, 50%, and 75%, respectively.27 If a test of heterogeneity was significant, a random-effect analysis was reported.25 A 2-sided P-value <.05 was considered to be statistically significant. Results for individual trials and summary results are expressed as an odds ratio (OR) with 95% confidence intervals (CIs).

Sensitivity Analyses 

Sensitivity analyses were conducted to explore the robustness of our results. To identify studies that might have a disproportionate influence on the summary treatment effect, we deleted each study one at a time. To assess for the possibility of publication bias, we conducted a visual examination of inverted funnel plots and statistically using Begg’s rank correlation test and Egger’s method.28, 29, 30

A priori, we explored the effect among patients with a cardioinhibitory response on tilt-table testing. Studies in which all patients were required to demonstrate a baseline cardioinhibitory response were pooled with patients from The North American Vasovagal Pacemaker Study (VPS1) and the Second Vasovagal Pacemaker Study (VPS2), who showed a cardioinhibitory tilt table response defined as a heart rate <40 beats per minute or asystole >3 seconds.5, 8

Estimation of the ‘Expectation’ Response 

Four trials were comparisons of pacemaker implantation with active programming4, 5, 6, 7 versus medical therapy, and 2 trials8, 11 were double-blind comparisons where all patients received a pacemaker but neither patients nor physicians knew if the programming was to active pacing or only sensing. The first group of trials described “active pacemaker versus no treatment” and the second group “active pacemaker versus ‘inactive’ pacemaker.” Using the method of Fisher et al, we estimated the effect of “inactive” pacemaker versus “no treatment” or what we have called the “expectation response” of pacemaker insertion.31 In the absence of any treatment effect in the double-blind trials, the observed effect among the pacemaker versus medical therapy trials could represent an “expectation” response. That is, patients who have undergone surgical pacemaker implant might expect to have fewer syncopal recurrences. Moreover, physicians assessing outcome events are not immune to this phenomenon, and it is possible that outcome ascertainment could be influenced as well.

Back to Article Outline

Results 

Study Identification and Selection 

Nine prospective randomized controlled trials enrolled 430 subjects (mean age=55±16, female=231) and met the inclusion criteria (Table 1).3, 4, 5, 6, 7, 8, 9, 10, 11 One trial was excluded due to enrollment of a pediatric population only.12 Of these 9 trials, 4 were open-label trials comparing pacemaker with medical therapy or no therapy (Group 1),4, 5, 6, 7 3 were open-label pacemaker trials (all patients received a pacemaker and programming was to active pacing of different pacemaker algorithms, Group 2),3, 9, 10 and 2 were double-blind pacemaker trials (all patients received a pacemaker but neither patients nor physicians were aware whether programming was to active or inactive pacing, Group 3).8, 11 In the Inotropy Controlled Pacing in Vasovagal Syncope (INVASY) trial, participants were initially randomized, but after the first year, randomization was discontinued and all participants crossed over to the treatment arm.10 For the purposes of this meta-analysis, only the results from the randomized phase were included. In 8 studies, the diagnosis of vasovagal syncope was made by history and positive tilt table test; in the remaining study, the adenosine triphosphate test was used in place of tilt table test.4

Table 1. Study Characteristics
TrialReferenceYearStudy Quality⁎Tilt Table Criteria at Entry (% cardioinhibitory)†Followup Duration (months) Age (mean, yrs)Male (%)
Allocation ConcealmentBlindingRandomized Patients
PersonnelPatientsInterventionnControlN
Ammirati71998UnspecifiedNoNoPositive TTT with bradycardia after failing medical therapy (100)‡17.7DDD-RDR12DDI-RH PPM861.160.0
Flammang81999UnspecifiedNoNoNot performed (100)§62DDD10Usual care1072.315.0
VPS91999YesNoNoPositive TTT with HR <60 (100)16DDD-RDR27Usual care2743.027.8
VASIS102000YesNoNoVASIS 2A or 2B (100)44DDI-RH19No therapy permitted2359.657.1
SYDIT112001YesNoNoPositive TTT with HR <60 (100)17.3DDD-RDR46Atenolol 100 mg4758.140.9
VPS 2122003YesYesYesPositive TTT with HRxBP <6000 (19)6DDD-RDR48ODO PPM5249.240.0
Deharo132003UnspecifiedNoYesPositive TTT with HR <60 (100)12DDDR with PEA Sensor23DDI PPM2361.882.8
INVASY⁎⁎142004YesNoYesVASIS 2A or 2B (100)12DDD-CLS PPM17DDI PPM959.057.7
SYNPACE††152004YesYesYesPositive TTT with asystolic or mixed (asystole and bradycardia) response (100)23.8DDD-RDR16OOO PPM1353.034.5

CLS=closed loop stimulation; PPM=permanent pacemaker; R=rate response; RDR=rate-drop response; RH=rate hysteresis; TTT=tilt table test.

⁎All studies were randomized, analyzed according to the intention-to-treat principle, and claimed no losses to follow-up.

†Numbers in brackets are the % cardioinhibitory syncope.

‡Patients had to have a positive TTT with VASIS 2B response, followed by treatment failure after atenolol or etilepherine treatment, followed by a second positive TTT showing cardioinhibition.

§ATP test for cardioinhibitory syncope was used in lieu of tilt table testing.

HR <40 or asystole >3 sec on tilt table testing.

⁎⁎For the purposes of the meta-analysis, only the results from the 12-month randomized controlled phase of INVASY were used.

††The trial was terminated early, after 6 months, due to failure to demonstrate viability of intervention.

Data Synthesis - Recurrent Vasovagal Syncope 

When all 9 trials were pooled, there was statistical evidence of heterogeneity using both the Peto (P=.0009, I2=69.6%) and the Mantel-Haenszel (P=.002, I2= 66.7%) methods. There was no evidence that any single trial was the source of that heterogeneity. Given the evidence of heterogeneity, the summary effect was described using a random effects analysis; the rate of recurrent syncope was significantly reduced with pacemaker therapy (OR 0.15, 95% CI, 0.05 to 0.42, P=.0004). Notwithstanding, the clinical diversity of the trials and the evidence of heterogeneity suggest that a common effect is highly implausible.

When trials were grouped by control group therapy and whether patients and physicians were aware of group allocation (Figure 1), there was no longer evidence of heterogeneity (P>.10 and I2=0% within each group; Figure 1, Table 2). The risk of recurrent syncope was reduced when pacemaker therapy was compared with medical or no therapy, and in single blind trials comparing active pacing algorithms. There was no evidence of benefit of pacemaker therapy in Group 3 studies involving double blinding (OR 0.83, 95% CI, 0.41 to 1.70). The results for Group 1 and 2 were similar (P=.15), while a test of interaction showed differences between Groups 1 and 3 (P=.0001) and Groups 2 and 3 (P=.00005).

  • View full-size image.
  • Figure 1. 

    Forest plot of primary study and summary odds ratios for recurrent syncope. Studies are ordered, based on the control therapy and based on double-blind trial methodology. For each study, the random effects odds ratios are indicated by black squares, and horizontal lines represent the 95% confidence intervals. The area of the black square is proportional to the amount of ‘statistical information’ contributed by the trial. Odds ratios less than unity favor the strategy of permanent cardiac pacing. The diamond at the bottom of each category and overall represents the summary odds ratio (OR) with 95% confidence intervals. The results favor cardiac pacing in Groups 1 and 2, but not Group 3.

Table 2. Random Effects Meta-analysis on Recurrent Syncope
Treatment Groups Results
ActiveControl OR95% CIP
All patients
Group 1Active pacemakerMedical/no therapyOpen label0.090.04to0.22<.00001
Group 2⁎Active pacemakerActive pacemakerOpen label0.040.01to0.230.0004
Group 3Active pacemakerInactive pacemakerBlinded0.830.41to1.700.61
All studies, favoring PPM implantation 0.150.05to0.420.0004
Cardioinhibitory Subgroup
Group 1PacemakerMedical/no therapyOpen label0.080.03to0.24<.0001
Group 3Active pacemakerInactive pacemakerBlinded1.490.41to5.420.61
All studies, favoring PPM implantation 0.210.05to0.880.03

⁎Comparison of different pacemaker algorithms.

Publication Bias 

A funnel plot was used to assess for the possibility of publication bias. Neither Begg’s rank correlation test nor Egger’s test were statistically significant. Visual inspection revealed some asymmetry, with a gap seen to the right side of the funnel plot, which can be seen when smaller nonsignificant studies remain unpublished. Funnel plot asymmetry observed in this study, however, was linked to methodologic differences, with blinded trials to the right and unblinded trials to the left. It can be estimated that 257 studies would have to exist with “negative” results (average observed effect of zero) to nullify this meta-analysis and render it conventionally nonsignificant, according to the fail safe ‘N’ method of Rosenthal,32 and 187 trials according to the method of Klein.33

Patients with a Baseline Cardioinhibitory Response 

Data were available from 6 trials and included unpublished data from patients enrolled in VPS1 and VPS2 (R. Roberts, personal communication). The results in this subgroup are in keeping with the overall results (Figure 2, Table 2). For all trials combined, there was evidence of statistical heterogeneity (P=.02, I2=61.6%). Recurrent syncope was reduced by pacemaker therapy among Group 1 but not Group 3 patients (OR 0.08 vs OR 1.49, test of interaction P=.001). There was no evidence of heterogeneity within these groups of trials.

  • View full-size image.
  • Figure 2. 

    Forest plot of primary study and summary odds ratios for recurrent syncope, restricted to patients with cardioinhibitory response to tilt table testing. The random effects odds ratios (OR) and 95% confidence intervals for the development of recurrent syncope in patients with a cardioinhibitory tilt table response are based on reports from 6 trials involving patients from Groups 1 and 3 only. The results mirror those of the overall meta-analysis with benefit of cardiac pacing seen in Group 1 but not Group 3. The test of interaction between groups was significant (P=.001). The confidence intervals for Group 3 are wide, and modest treatment effects cannot be excluded.

Expectation Response 

The rate of recurrent syncope was similar and high among the controls in the trials of pacemaker versus medical therapy (Group 1, 47%) and the double-blind pacemaker trials (Group 3, 42%). The recurrence rate was much lower when patients and physicians were aware of the active pacemaker status (8% vs 38%). The expectation response (Figure 3) of pacemaker implantation appears to explain the majority of the observed pacemaker response (OR 0.16, 95% CI, 0.06 to 0.40, P=.0001).

  • View full-size image.
  • Figure 3. 

    Estimation of the expectation response of cardiac pacing in preventing recurrent syncope. Shown are the odds ratios (ORs) (based on the Peto Method) and 95% confidence intervals (CI) for the comparisons of active cardiac pacing versus no therapy (Group 1), “inactive” versus active cardiac pacing (Group 3), and the synthesized results for “inactive” cardiac pacing versus no therapy. Inactive cardiac pacing refers to pacemakers programmed to sensing only. Note, the results for Group 3 are simply the reciprocal of the results shown in Table 2 as the emphasis of the comparison is reversed. The horizontal line drawn at the OR of 1.0 indicates the null hypothesis of no difference between groups. Estimates to the left of this line represent treatment better and estimates to the right represent control better. Note that the controls will be no therapy for the first and third comparison and active pacing for the middle comparison. The bottom comparison represents the synthesized “expectation response” as this is the effect of inactive cardiac pacing (functioning in sensing mode only) versus no therapy. The expectation response alone reduces the odds of syncope by 84% (OR 0.16, 95% CI, 0.06 to 0.40, P=.0001). PPM=permanent pacemaker.

Complications of Therapy 

Safety data were available for 7 of the 9 randomized trials (n=391) (Table 3). Among patients who received a pacemaker (n=286), there were 20 major and minor procedure-related complications, an overall rate of 7.0% (95% CI, 4.6% to 10.6%). Two major complications occurred (OR 0.7%, 95% CI, 0.2% to 2.5%): one pericardial tamponade and a pacemaker infection requiring re-implantation.

Table 3. Safety Data for Patients with Pacemaker⁎
Complications
TrialNumber with PPMMinor†Major‡
Ammirati---
Flammang---
VPS2610
VASIS2200
SYDIT4500
VPS 2100192
Deharo2300
INVASY4100
SYNPACE2900

PPM=permanent pacemaker; -=not reported.

⁎This includes all patients with pacemaker implantation whether programmed to active and inactive modes.

†Minor complications included lead dislodgement, infection treated with antiobiotics, hematoma, pain related to pacemaker generator.

‡Major complications include pericardial tamponade (1) and infection requiring reimplantation of pacemaker (1).

Back to Article Outline

Discussion 

The apparent benefit of pacing in vasovagal syncope from the current analysis is associated with significant statistical heterogeneity, suggesting that a single common estimate of effect is highly improbable. The benefit of pacing was evident when compared against medical or no therapy, or when the trials were not rigorously blinded. There was no evidence of benefit when both patients and physicians were unaware whether pacemaker programming was to an active or inactive mode. Our results strongly suggest that the primary mechanism of benefit of pacemaker insertion is an expectation response, suggesting the physiological benefit is small. Moreover, we were unable to find evidence of benefit among patients with a cardioinhibitory response on tilt table testing.

Heterogeneity was addressed using a random-effects rather than a fixed-effects analysis,15 although this is insufficient to explain heterogeneity. Thus, we looked for clinical or methodological causes of heterogeneity.34, 35 Statistical heterogeneity was best explained by differences in trial methodology given heterogeneity did not persist when double blinding was specified. If methodological diversity is accepted as the source of heterogeneity, the reduction of syncope associated with pacemaker insertion in early nonblinded randomized trials may have been the result of an expectation response.4, 5, 6, 7 We estimate the relative reduction in risk of recurrent syncope due to the expectation response to be 74%; this accounts for virtually all of the observed effect of cardiac pacing. The association of a surgical placebo effect with permanent pacemaker insertion has also been suggested in patients with hypertrophic cardiomyopathy.36, 37, 38

While the Vasovagal Syncope International Study (VASIS) showed a reduction in syncope with DDI pacing, 3 other trials showed DDD pacing to be superior to DDI programming mode.6, 7, 9, 10 Interestingly, the rate of recurrence in patients treated with DDI pacing in VASIS (1/19) was considerably less than when DDI pacing was used as a control group in the other trials (14/40). In INVASY and the Deharo study, physicians but not participants were aware of treatment allocation, while in the Syncope Diagnosis and Treatment Study (SYDIT), both patients and physicians were aware.6, 7, 9, 10 These findings support the existence of an expectation effect and suggest an ascertainment bias where patients but not physicians were blinded.

The total number of patients studied to date in randomized double-blind placebo-controlled trials is only 129. Assuming a syncopal recurrence rate of 50% and a relative risk reduction of 25% (which would give a number needed to treat of 8), a randomized blinded trial would require 500 patients to show significant benefit (2-tailed P=.05 and 80% power). As such, reliable conclusions regarding pacing in these patients cannot be reached in the absence of a larger study.39 While a large treatment effect is unlikely, a smaller incremental benefit beyond the expectation response has not been ruled out.

The duration of follow-up in VPS2 and the Vasovagal Syncope and Pacing Trial (SYNPACE) was 6 and 23.8 months respectively.8, 11 Observational studies have suggested the median time to syncope after pacemaker insertion was as long as 3 or 4 years.40, 41 Thus, the expectation response may carry a lasting benefit so long as the ritual of insertion, programming, and follow-up is continued, or confer additional benefit once the expectation effect has been exhausted. A pragmatic approach might be to advocate pacing in select patients with problematic symptoms, accepting that the primary benefit would be due to the expectation response. However, justifying the risks and costs of pacemakers in this population based on an expectation effect, may be difficult for some clinicians. The question is philosophical, depending on the role clinicians feel the expectation response, which has long been known to play a role in the practice of medicine, should have in their practice.42, 43

Clinical diversity, notably the highly variable tilt table inclusion criteria and differences in tilt table protocols, has been suggested as a cause for the conflicting results.2 In VPS2, only 58/100 patients had a trough heart rate <60 beats per minute (bpm) on tilt table testing, which was the qualifying criterion for inclusion in the VPS and SYDIT trials.5, 7, 8 Furthermore, only 19/100 patients had a trough heart rate <40 bpm or asystole >3 seconds, which would have been required to meet the VASIS 2A or 2B tilt criteria required for inclusion in VASIS.6 Thus, one might argue that VPS2, although the largest and methodologically most rigorous trial to date, did not focus on the appropriate study population with severe cardioinhibitory syncope.2 Furthermore, the International Study of Syncope of Uncertain Etiology (ISSUE) has demonstrated that response to tilt table testing does not predict a cardioinhibitory pattern recorded during subsequent spontaneous syncope.44 The reliability of using tilt testing to identify candidates for pacemaker therapy is thus somewhat debatable.

If one excludes the VPS2 trial, the remaining data consist of patients with cardioinhibitory tilt table responses and favor the use of cardiac pacing. In order to evaluate patients with bradycardia during tilt testing, we repeated the analysis using patients from VPS1 and VPS2 who had severe bradycardia on tilt testing.5, 8 The results still demonstrated moderate heterogeneity, mirroring the overall results, suggesting that even patients with a severe cardioinhibitory response on tilt testing may experience a strong expectation effect. Conversely, recently presented data suggest that documentation of spontaneous bradycardia during vasovagal syncope may select a group of responders to pacing.44 A large, randomized, double-blind study that specifically targets this patient population is needed.

We selected trials in which patients were at least 18 years of age. We reasoned that patients younger than this would not be the target for cardiac pacing, given the risks and invasive nature of pacing, and that spontaneous remission is common. Younger adult patients (aged 18-23 years) represented less then 2.5% of the patient population in the present study. Vasovagal syncope typically follows a benign course and often spontaneously resolves. Thus, lifelong therapy, in light of the present study, is unlikely to be necessary.

Back to Article Outline

Summary 

The results of small, preliminary trials overestimated the treatment effect of pacemakers due to a lack of blinding of physicians and patients. VPS2, while the largest blinded randomized trial to date, was underpowered and could have missed a moderate treatment benefit.8 These data do not support the use of permanent cardiac pacing as a first-line therapy for vasovagal syncope. Longer and adequately powered double-blinded trials are required to examine whether a sustained benefit exists.

Back to Article Outline

Acknowledgment 

The authors thank Professor Robin Roberts and the VPS 2 investigators for providing unpublished data.

Back to Article Outline

References 

  1. Soteriades ES, Evans JC, Larson MG, et al. Incidence and prognosis of syncope. N Engl J Med. 2002;347(12):878–885
  2. Trim GM, Krahn AD, Klein GJ, Skanes AC, Yee R. Pacing for vasovagal syncope after the second vasovagal pacemaker study (VPS II): a matter of judgement. Card Electrophysiol Rev. 2003;7(4):416–420
  3. Ammirati F, Colivicchi F, Toscano S, et al. DDD pacing with rate drop response function versus DDI with rate hysteresis pacing for cardioinhibitory vasovagal syncope. Pacing Clin Electrophysiol. 1998;21:2178–2181
  4. Flammang D, Antiel M, Church T, et al. Is a pacemaker indicated for vasovagal patients with severe cardioinhibitory reflex as identified by the ATP test? (A preliminary randomized trial). Europace. 1999;1(2):140–145
  5. Connolly SJ, Sheldon R, Roberts RS, Gent M. The North American Vasovagal Pacemaker Study (VPS) (A randomized trial of permanent cardiac pacing for the prevention of vasovagal syncope). J Am Coll Cardiol. 1999;33(1):16–20[see comments]
  6. 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. 2000;102(3):294–299
  7. Ammirati F, Colivicchi F, Santini M. Permanent cardiac pacing versus medical treatment for the prevention of recurrent vasovagal syncope: a multicenter, randomized, controlled trial. Circulation. 2001;104(1):52–57
  8. Connolly SJ, Sheldon R, Thorpe KE, et al. The Second Vasovagal Pacemaker Study (VPSII). Pacing Clin Electrophysiol. 2003;26:1565
  9. Deharo JC, Brunetto AB, Bellocci F, et al. DDDR pacing driven by contractility versus DDI pacing in vasovagal syncope: a multicenter, randomized study. Pacing Clin Electrophysiol. 2003;26:447–450
  10. 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. 2004;6(6):538–547
  11. Raviele A, Giada F, Menozzi C, et al. A randomized, double-blind, placebo-controlled study of permanent cardiac pacing for the treatment of recurrent tilt-induced vasovagal syncope (The vasovagal syncope and pacing trial (SYNPACE)). Eur Heart J. 2004;25(19):1741–1748
  12. McLeod KA, Wilson N, Hewitt J, Norrie J, Stephenson JB. Cardiac pacing for severe childhood neurally mediated syncope with reflex anoxic seizures. Heart. 1999;82(6):721–725
  13. Counsell CE, Salinas R, Naylor R, Warlow CP. A systematic review of the randomised trials of carotid patch angioplasty in carotid endarterectomy. Eur J Vasc Endovasc Surg. 1997;13(4):345–354
  14. Meade MO, Richardson WS. Selecting and appraising studies for a systematic review. Ann Intern Med. 1997;127(7):531–537
  15. Lau J, Ioannidis JP, Schmid CH. Quantitative synthesis in systematic reviews. Ann Intern Med. 1997;127(9):820–826
  16. Schulz KF, Chalmers I, Hayes RJ, Altman DG. Empirical evidence of bias (Dimensions of methodological quality associated with estimates of treatment effects in controlled trials). JAMA. 1995;273(5):408–412
  17. Egger M, Zellweger-Zahner T, Schneider M, Junker C, Lengeler C, Antes G. Language bias in randomised controlled trials published in English and German. Lancet. 1997;350(9074):326–329
  18. Cook DJ, Guyatt GH, Ryan G, et al. Should unpublished data be included in meta-analyses? (Current convictions and controversies). JAMA. 1993;269(21):2749–2753
  19. Berlin JA University of Pennsylvania Meta-analysis Blinding Study Group. Does blinding of readers affect the results of meta-analyses?. Lancet. 1997;350(9072):185–186
  20. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst. 1959;22(4):719–748
  21. Robins J, Breslow N, Greenland S. Estimators of the Mantel-Haenszel variance consistent in both sparse data and large-strata limiting models. Biometrics. 1986;42(2):311–323
  22. Laird NM, Mosteller F. Some statistical methods for combining experimental results. Int J Technol Assess Health Care. 1990;6(1):5–30
  23. Yusuf S, Peto R, Lewis J, Collins R, Sleight P. Beta blockade during and after myocardial infarction: an overview of the randomized trials. Prog Cardiovasc Dis. 1985;27(5):335–371
  24. Fleiss JL, Gross AJ. Meta-analysis in epidemiology, with special reference to studies of the association between exposure to environmental tobacco smoke and lung cancer: a critique. J Clin Epidemiol. 1991;44(2):127–139
  25. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7(3):177–188
  26. Fleiss JL. Statistical Methods for Rates and Proportions. 2nd ed.. New York, New York: J. Wiley; 1981;
  27. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327(7414):557–560
  28. Light RJ, Pillemer DB. Summing Up: The Science of Reviewing Research. Cambridge, Massachusetts: Harvard University Press; 1984;
  29. Begg CB, Berlin JA. Publication bias: A problem in interpreting medical data. J R Stat Soc Ser A Stat Soc. 1988;151:419–463
  30. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997;315(7109):629–634
  31. Fisher LD, Gent M, Buller HR. Active-control trials: how would a new agent compare with placebo? (A method illustrated with clopidogrel, aspirin, and placebo). Am Heart J. 2001;141(1):26–32
  32. Rosenthal R. The file drawer problem and tolerance for null results. Psychol Bull. 1979;86:638–641
  33. Klein S, Simes J, Blackburn GL. Total parenteral nutrition and cancer clinical trials. Cancer. 1986;58(6):1378–1386
  34. Thompson SG. Why sources of heterogeneity in meta-analysis should be investigated. BMJ. 1994;309(6965):1351–1355
  35. Pettiti DB. Approaches to heterogeneity in meta-analysis. Stat Med. 2001;20:3625–3633
  36. Fananapazir L, Epstein ND, Curiel RV, Panza JA, Tripodi D, McAreavey D. Long-term results of dual-chamber (DDD) pacing in obstructive hypertrophic cardiomyopathy (Evidence for progressive symptomatic and hemodynamic improvement and reduction of left ventricular hypertrophy). [see comments] Circulation. 1994;90:2731–2742
  37. Maron BJ, Nishimura RA, McKenna WJ, Rakowski H, Josephson ME, Kieval RS. Assessment of permanent dual-chamber pacing as a treatment for drug-refractory symptomatic patients with obstructive hypertrophic cardiomyopathy (A randomized, double-blind, crossover study (M-PATHY)). Circulation. 1999;99(22):2927–2933
  38. Linde C, Gadler F, Kappenberger L, Ryden L PIC Study Group. Placebo effect of pacemaker implantation in obstructive hypertrophic cardiomyopathy (Pacing In Cardiomyopathy). Am J Cardiol. 1999;83(6):903–907
  39. Pogue J, Yusuf S. Overcoming the limitations of current meta-analysis of randomised controlled trials. Lancet. 1998;351(9095):47–52
  40. Raj SR, Koshman ML, Sheldon RS. Outcome of patients with dual-chamber pacemakers implanted for the prevention of neurally mediated syncope. Am J Cardiol. 2003;91(5):565–569
  41. Moya A, Brignole M, Menozzi C, et al. Mechanism of syncope in patients with isolated syncope and in patients with tilt-positive syncope. Circulation. 2001;104(11):1261–1267
  42. Moerman DE. The meaning response and the ethics of avoiding placebos. Eval Health Prof. 2002;25(4):399–409
  43. Moerman DE, Jonas WB. Deconstructing the placebo effect and finding the meaning response. Ann Intern Med. 2002;136(6):471–476
  44. Steering Committee of the ISSUE 2 Study. International Study on Syncope of Uncertain Etiology 2: the management of patients with suspected or certain neurally mediated syncope after the initial evaluation rationale and study design. Europace. 2003;5(3):317–321

PII: S0002-9343(06)00670-X

doi:10.1016/j.amjmed.2006.05.046

The American Journal of Medicine
Volume 120, Issue 1 , Pages 54-62, January 2007

Article Tools

* Image Usage & Resolution