α-Adrenoceptor agonists for the treatment of vasovagal syncope: a meta-analysis of worldwide...

REGULAR ARTICLE

a-Adrenoceptor agonists for the treatment of vasovagal syncope:a meta-analysis of worldwide published dataYing Liao1, Xueying Li2, Yanwu Zhang3, Stella Chen4, Chaoshu Tang5, Junbao Du ([email protected])1,6

1.Department of Pediatrics, Peking University First Hospital, Beijing, China2.Department of Statistics, Peking University First Hospital, Beijing, China3.Medical Library of Peking University First Hospital, Beijing, China4.Torrey Pines High School, San Diego CA, USA5.Department of Physiology and Pathophysiology, Peking University Health Science Center, Beijing, China6.Key Laboratory of Molecular Cardiology, Ministry of Education of China, Beijing, China

Keywordsa-adrenoceptor agonists, Meta-analysis,Randomized controlled trial, Vasovagal syncope

CorrespondenceProfessor Junbao Du, Department of Pediatrics,Peking University First Hospital, Beijing100034, China.Tel: +8610-66551122 ext. 3238 |Fax: +8610-66134261 |Email: [email protected]

Received3 January 2009; revised 3 March 2009;accepted 4 March 2009.

DOI:10.1111/j.1651-2227.2009.01289.x

AbstractAim: The present study was aimed at evaluating present randomized controlled trials (RCTs)

regarding the effect of a-adrenoceptor agonists on vasovagal syncope (VVS).

Methods: According to inclusion and exclusion criteria, articles were selected from medical electronic

databases. RCTs were then assessed based on the Juni assessment, and meta-analysis was

completed using the Review Manager 4.2 software. Indication to further evaluate effects was the

recurrence of syncope during follow-up treatment or a response in the head-up tilt test (HUT) after

treatment. The results were stated as odd ratio (OR), with a 95% confidence interval (CI) and a

p < 0.05 significant level.

Results: In total, six RCTs were selected. Funnel plot analysis showed possible publication bias.

Meta-analysis of the six RCTs, including all 165 patients in the treatment group and 164 patients in

the control group, indicated that a-adrenoceptor agonists were more effective than placebos in

treating VVS (OR ¼ 0.21, 95% CI: 0.06–0.77, p ¼ 0.02). The further, weighted independent t-test

disclosed that the weighted mean percentage of responders for midodrine (76.3% ± 7.7%) was

significantly higher than that for etilefrine (65.5% ± 15.4%) (t ¼ 5.863, p < 0.001).

Conclusion: The currently published RCTs support that a-adrenoceptor agonists might be effective for VVS.

Midodrine can be regarded as a better choice compared with etilefrine.

BACKGROUNDSyncope has a cumulative lifetime incidence of 35% (1)and accounts for 1–2% of all emergency department visits(2). Vasovagal syncope (VVS), an important form of neu-rally mediated syncope, is the most common cause of unex-plained syncope, with up to 80% of children with syncopeexperience belonging to VVS (3). Although VVS is benign,recurrent syncope may lead to psychosocial and physicalproblems and usually results in poor quality of life. Conse-quently, increasingly more attention has been paid to thetherapy of VVS to prevent recurrent episodes and improvethe quality of the patient’s life. This therapy not only in-cludes cardiac pacing and non-pharmacologic treatments,including educating patients regarding tilt training, waterdrinking, salt intake and predisposing factors that may trig-ger episodes, but also often involves pharmacotherapy (4).

Various drugs used for the treatment of VVS have beenreported, such as b-adrenoceptor blockers, a-adrenoceptoragonists, selective serotonin reuptake inhibitors (SSRIs),fludrocortisone, angiotensin converting enzyme inhibitors(ACEIs), scopolamine, disopyramide, etc. Several trials havebeen held to explore the efficacy of these drugs (5–9). Gen-erally, few randomized controlled trials (RCTs) were heldfor the above medicines other than b-adrenoceptor block-

ers and a-adrenoceptor agonists. b-Adrenoceptor blockersproved to be useful in many uncontrolled studies (4) butwere unable to show such advantages in some long-termfollow-up double-blind RCTs (10,11). Although the mecha-nism of VVS is still not completely clear, failure to achieveproper vasoconstriction of the peripheral vessels is consid-ered to be common to all the forms of neurally mediated syn-cope (12). As a result, vasoconstrictive substances such asa-adrenoceptor agonists, have been used in VVS and studiedin several RCTs. Although some of the RCTs in Ward’s study(13) showed the benefits of a-adrenoceptor agonists overplacebos, other RCTs in Raviele’s study (14) have not beenable to support the superiority of a-adrenoceptor agonists.Therefore, we aimed to achieve the systemic assessmentof a-adrenoceptor agonists for treatment of VVS throughmeta-analysis, to offer more powerful evidence for clinicalpractice.

MATERIALS AND METHODSInclusion criteria of studiesThe studies taken into our analysis referred to prospectiverandomized placebo-controlled trials for the treatment ofVVS (or neurocardiogenic syncope). In the enrolled studies,the aetiologies of syncope for the patients were not definite

Acta Pædiatrica ISSN 0803–5253

1194 ª2009 The Author(s)/Journal Compilation ª2009 Foundation Acta Pædiatrica/Acta Pædiatrica 2009 98, pp. 1194–1200

by physical examination, electrocardiography, chest X-rayor biochemical screening. Patients were diagnosed with VVSaccording to the results of a head-up tilt test (HUT) with orwithout isoproterenol or nitroglycerine. a-Adrenoceptor ag-onists were used on the treatment group while no pharmaco-logical therapy was given to the control group. The protocolof the trials, the procedure of the follow-up and the generalcharacteristics of the subjects were stated in detail. Indica-tion for evaluating effects was the rate of syncope recurrenceduring follow-up or response in HUT after undergoing thetreatment.

Exclusion criteria of studiesStudies with uncertain diagnosis criteria of VVS or the useof other drugs in the control group were excluded. Someretrospective, uncontrolled or non-random trials were notselected either.

Protocol of searchingFor our meta-analysis, clinical trials were determinedby electronic reviews of published medical literature orsearches from medical libraries. The databases, includ-ing PubMed (1968–2008), EMBASE (1991–2008), Elsevier(1990–2008) and CNKI (1990–2008), were searched by twoprofessional co-workers using search terms (in English orChinese) such as ‘vasovagal syncope AND management ⁄treatment ⁄ therapy ⁄ a agonists’, ‘neurocardiogenic syn-cope AND management ⁄ treatment ⁄ therapy ⁄ a agonists’,‘neurally mediated syncope AND management ⁄ treatment⁄ therapy ⁄ a agonists’, ‘orthostatic intolerance AND man-agement ⁄ treatment ⁄ therapy ⁄ a agonists’ and ‘clinical trialAND management ⁄ treatment ⁄ therapy ⁄ a agonists’. Origi-nal articles were obtained as downloads from electronicdatabases or copies from libraries.

Assessment of studiesTwo reviewers rated study quality independently. Character-istics of subjects, design of trials, course of arrangement andanalysis were abstracted and evaluated. Studies assessed thequality of the RCTs based on the four standards of the Juni

assessment (15,16): (i) methods of randomization applied tosequence generations were suitable; (ii) allocation conceal-ment was carried out in proper means (17); (iii) blinding wasstated; and (iv) patients who dropped out during follow-upperiods were considered complete cases for the intention-to-treat analysis (18). RCTs that fulfilled all of the above wereassessed as grade ‘A’, with the least probability to introducebias. RCTs that only partially met any one or more of theabove were assessed as grade ‘B’, with moderate possibilityof bias occurrence. RCTs that did not satisfy one or moreof the four standards at all were assessed as grade ‘C’, withthe highest probability to produce bias. The publication biaswas evaluated by Funnel plot.

Statistical analysisWe used the Review Manager 4.2 software (The NordicCochrane Centre of The Cochrane Collaboration, Copen-hagen, Denmark) for meta-analysis. The test for heterogene-ity was achieved by v2 test. If p ‡ 0.05, the results of selectedstudies were deemed homogenous, in which the outcome ofsystemic analysis was stated in a fixed effects model. If p <0.05, the results were deemed heterogeneous and a randomeffects model was applied. The final results were shown inthe form of vector images, including OR and its 95% CI,since the data in our analysis were dichotomous. Weightedindependent t-test was applied to the comparison betweenthe two medicines. Additionally, p < 0.05 was consideredstatistically significant.

RESULTSBasic data of included studiesTwo hundred eighteen original articles were selected pri-marily from world-wide databases. Six RTCs (13,14,19–22)were finally accepted, based on inclusion and exclusion cri-teria. All six were published in English. The general back-grounds of the six RCTs are shown in Table 1. Four ofthe six RCTs used recurrence of syncope as the endpointof follow-up, while the other two used response in HUTto reflect the efficacy. Follow-up duration lasted from 1week to more than 4 years. Baseline characteristics of the

Table 1 Basic data of included six RCTs

Study Groups Study design HUT Intervention Endpoint Follow-up duration p value

Moya 1995 (19) T 30 RCT, db Base Etilefrine HUT 2 weeks NS

C 30 Cross Placebo

Ward 1998 (13) T 16 RCT, sb Nitro Midodrine HUT 1 month*2 0.01

C 16 Cross Placebo

Raviele 1999 (14) T 63 RCT, db Base or Etilefrine syncope 1 year NS

C 63 20cen Nitro Placebo recurrence

Perez-Lugones 2001 (20) T 31 RCT, ub Base or Midodrine syncope >6 months <0.01

C 30 Iso Placebo recurrence

Kaufmann 2002 (21) T 12 RCT, db Base Midodrine HUT 1 week <0.02

C 12 Cross Placebo

Zhang 2006 (22) T 13 RCT, ub Base or Midodrine HUT 42 ± 16 <0.05

C 13 Nitro Placebo months

T ¼ treatment; C ¼ controlled; RCT ¼ randomized controlled trail; sb ¼ single blined; db ¼ double blined; ub ¼ unblined; cross ¼ cross over; Base ¼ base HUT;

Nitro ¼ HUT induced by sublingual nitroglycerin; Iso ¼ HUT induced by isoproterenol infusion; NS ¼ not significant.

Liao et al. Meta-analysis of vasovagal syncope

ª2009 The Author(s)/Journal Compilation ª2009 Foundation Acta Pædiatrica/Acta Pædiatrica 2009 98, pp. 1194–1200 1195

subjects are listed in Table 2. The average ages of the subjectswere between 12 and 56 years. Each study involved 12–126patients, all of whom were diagnosed as VVS by HUT withor without isoproterenol or nitroglycerine.

Quality of included studiesAll of the six studies were designed as RCTs with random-ized grouping. However, allocation concealment was notdescribed in detail in any of the articles. Two studies wereunblended (20,22), and one of them did not use intention-to-treat analysis, when patients dropped out (22). As a re-sult, both of them were assessed as grade C. The other fourstudies, employing blinding and intention-to-treat analysis ifnecessary, were classified as grade B (Table 3). Funnel plotanalysis showed publication bias may exist (Fig. 1).

Result of meta-analysisRatios of patients with syncope recurrence during follow-ups in four RCTs and ratios of patients with positive re-sponse in HUT at the endpoint of the other two RCTs weresummarized as fraction for efficiency. Heterogeneity analy-sis of included RCTs showed heterogeneity among the re-sults of the studies (v2 ¼ 28.94, df ¼ 5, p < 0.0001). Thus, arandom effects model was adopted to further analysis. Meta-analysis of the six RCTs (Fig. 2), including the 165 totalpatients in the treatment group and 164 patients in thecontrol group, indicated that a-adrenoceptor agonists weresuperior to placebos for the treatment of VVS (OR ¼ 0.21,95% CI: 0.06–0.77, p¼ 0.02).

Comparison of efficacy between midodrineand etilefrineSince there were two medicines involved in the analysis,we compared the percentages of responders between thetwo using independent t-test, weighted by the number ofparticipants. The weighted mean percentage of respondersin studies of midodrine was 76.3% ± 7.7% (n ¼ 72) whilethat of etilefrine was 65.5% ± 15.4% (n ¼ 93). Since theLevene’s test showed the variance was not equal (p £ 0.01),the independent t-test was corrected and the outcome isdisclosed in Table 4 that the weighted mean percentage of

Table 2 General characteristics of subjects enrolled in six RCTs

Study Age (year) Sex (male ⁄ female) SBP (mmHg) HR (beats ⁄ min)

Moya 1995 (19) 44 ± 22(T) 6 ⁄ 9(T) 137 ± 15(T) 76 ± 9(T)

48 ± 16(C) 10 ⁄ 5(C) 140 ± 16(C) 72 ± 13(C)

Ward 1998 (13) 56 ± 18 5 ⁄ 9 127 ± 18 74.6 ± 8

Raviele1999 (14) 46 ± 19(T) 23 ⁄ 40(T) 130 ± 14(T) 70 ± 11(T)

42 ± 18(C) 29 ⁄ 34(C) 131 ± 16(C) 68 ± 10(C)

Perez-Lugones 2001 (20) 42 ± 16(T) 11 ⁄ 20(T) <150 -

43 ± 18(C) 10 ⁄ 20(C)

Kaufmann 2002 (21) 42 ± 4 2 ⁄ 10 122 ± 4 65 ± 2

Zhang 2006 (22) 12.2 ± 2.9 10 ⁄ 16 - -

T ¼ treatment group; C ¼ controlled group.

Table 3 Quality assessment of six RCTs

Study Generation of random Allocation concealment Blinding Loss of follow-up ITT Grade

Moya 1995 (19) + 0 + 0 B

Ward 1998 (13) + 0 + 0 B

Raviele 1999 (14) + 0 + + + B

Perez-Lugones 2001 (20) + 0 0 + + C

Kaufmann 2002 (21) + 0 + 0 B

Zhang 2006 (22) + 0 0 + 0 C

ITT ¼ intention-to-treat analysis.

0.001 0.01 0.1 1 10 100 1000

0.0

0.2

0.4

0.6

0.8

SE(log OR)

OR (fixed)

Figure 1 Funnel plot of six RCTs. Each dot stood for a study. The distance

between each dot and the upright line indicated the bias in each study. As

the whole ‘funnel, composed by all the dots was not completely symmetric,

publication bias may exist.

Meta-analysis of vasovagal syncope Liao et al.


responders for midodrine was significantly higher than thatfor etilefrine (p £ 0.001).

DISCUSSIONFor this meta-analysis, articles were selected from pro-fessional medical electronic databases, including PubMed(1968–2008), EMBASE (1991–2008), Elsevier (1990–2008)and China National Knowledge Infrastructure (CNKI)(1990–2008). Each database was carefully searched to en-sure that the articles were accurately representative and se-lected from worldwide information.

VVS is a clinically challenging disorder, characterizedby a baroreceptor-mediated hypotensive or bradycardic re-sponse to orthostatic stress (13). The mechanism is com-

plex and remains controversial, but the Bezold–Jarisch re-flex is the most commonly accepted explanation. In healthyindividuals, upright posture causes venous pooling and anassociated decrease in cardiac output, resulting in less stim-ulation of baroreceptors. Corresponding increases of sym-pathetic activity and parasympathetic withdrawal lead toperipheral arterial vasoconstriction, venoconstriction andincreased heart rate and contractility. As a result, bloodpressure, heart rate and cardiac output increase to main-tain normal cerebral blood flow. Patients considered tohave VVS tend to have relative reductions in central bloodvolume, which is further aggravated by upright posture.Increases in circulating catecholamines and reduced ve-nous return result in increased myocardial contractility,which in turn stimulates mechanoreceptors, believed to be

Study Treatment group Control group OR (random) Weight OR (random)or sub-category n/N n/N 95% CI (%) 95% CI

01 Etilefrine Angel Moya 1995 17/30 15/30 18.73 1.31 [0.47, 3.61]

Raviele 1999 15/63 15/63 19.49 1.00 [0.44, 2.27]

Subtotal (95% CI) 93 93 38.22 1.11 [0.59, 2.10]

Total events: 32 (Treatment), 30 (Control)Test for heterogeneity: χ2= 0.16, df = 1 (P = 0.69), I2= 0%Test for overall effect: Z = 0.32 (P = 0.75)

02 Midodrine Ward 1998 6/16 14/16 15.16 0.09 [0.01, 0.52]

Perez-Lugones 2001 6/31 26/30 17.12 0.04 [0.01, 0.15]

Kaufmann 2002 2/12 8/12 14.49 0.10 [0.01, 0.69]

QY Zhang 2006 3/13 10/13 15.01 0.09 [0.01, 0.56]

Subtotal (95% CI) 72 71 61.78 0.07 [0.03, 0.15]

Total events: 17 (Treatment), 58 (Control)Test for heterogeneity: χ2= 1.05, df = 3 (P = 0.79), I2= 0%Test for overall effect: Z = 6.35 (P < 0.00001)

Total (95% CI) 165 164 100.00 0.21 [0.06, 0.77]Total events: 49 (Treatment), 88 (Control)Test for heterogeneity: χ2= 28.94, df = 5 (P < 0.0001), I2= 82.7%Test for overall effect: Z = 2.35 (P = 0.02)

0.001 0.01 0.1 1 10 100 1000

Favours treatment Favours control

Figure 2 Forest plot of the six RCTs. Section 01 includes the two studies involving etilefrine and placebo, and section 02 shows the four studies comparing

midodrine with placebo. The analysis of total effects of all the six studies states in the bottom. Ratios of patients with syncope recurrence during follow-up in four

RCTs and ratios of patients with positive response in HUT at the endpoint in the other two RCTs were summarized as fraction for efficiency (shown in the middle).

Heterogeneity analysis (shown in the bottom on the left) of included six RCTs showed significant differences among the results of the studies (p < 0.01). As a

result, a random effects model was adopted. Each little square block stood for the OR value of each study and the intersecting transverse line represented 95% CI.

The black rhombus stood for the results of synthetic analysis. Since the rhombus in the bottom was completely on the left side of the upright line, in the middle

without any overlap, our analysis indicated a-adrenoceptor agonists were more effective than placebos in the treatment of vasovagal syncope (OR ¼ 0.21, 95%

CI: 0.06–0.77, p ¼ 0.02).

Table 4 Comparison of efficacy among midodrine, etilefrine and placebo on VVS

Drugs Study Percentage of responders (%) Weighted mean percentage (%)

Midodrine Ward 1998 (13) 62.5 (n ¼ 16) 76.3 ± 7.7 (n ¼ 72)

Perez-Lugones 2001 (20) 80.6 (n ¼ 31)

Kaufmann 2002 (21) 83.3 (n ¼ 12)

Zhang 2006 (22) 76.9 (n ¼ 13)

Etilefrine Moya 1995 (19) 43.3 (n ¼ 30) 65.5 ± 15.4 (n ¼ 93)

Raviele 1999 (14) 76.2 (n ¼ 63)

Weighted independent t¢ test* t’ ¼ 5.863 p < 0.001

*Weighted by number of the participants.

The Levene’s test showed the variance was not equal (p < 0.01), the independent t-test was corrected as t¢ test.



situated in the wall of the left ventricle. The generated im-pulses running across afferent C fibres trigger a centrally me-diated withdrawal of sympathetic activity and an increase inparasympathetic response, leading to the ‘vasodepressor re-sponse’ and ‘cardioinhibitory response’ reflexes (3,23,24).Finally, hypotension, bradycardic, cerebral hypoperfusionand syncope may exist. This paradoxical reflex is believedto be a variant of the Bezold–Jarisch reflex (23). There-fore, excess venous pooling, high levels of circulating cate-cholamines, abnormal vasodilation and the Bezold–Jarischreflex are believed to play important roles in pathogenesis ofVVS.

Current pharmacotherapies of VVS are mostly based onthe theory above. Fludrocortisone, water drinking and saltintake are effective because of their ability to expand bloodvolume (21). b-Adrenoceptor blockers are also proposedto reduce the activation of mechanoreceptor, to increasevenous return and to block the effect of elevated circu-lation catecholamines (4,25). SSRIs may raise blood pres-sure by stimulating 5HT2 receptors of central serotonergicneurons, although the exact mechanism is still uncertain(26). ACEIs are believed to induce lower catecholamineconcentrations by inhibition of angiotensin II (26,27). Anti-cholinergics, such as scopolamine, are presumed to be effec-tive by decreasing left ventricular contractions and increas-ing parasympathetic response (8). Simply put, every agentused for VVS was reasonable in theory. However, most ofthem still need large-scale and well-designed RCTs to gainmore evidence for use.

a-Adrenoceptor agonists have been used for treating VVSfor several years. Our study supported that they were supe-rior to placebos by meta-analysis of recent RCTs. The mech-anisms of these improved results were undetermined. Theeffects of a-agonists include increased blood pressure dueto enhanced peripheral vascular resistance and diminishedvenous pooling (25). Therefore, investigators of a-agonistsused to believe that both arterial construction and veno-constriction take part in the improvement of VVS (26). Infact, it was reported that midodrine increased blood pres-sure only at a dose higher than 5 mg daily (28). Ward’s study(13) showed significantly increased systolic blood pressurewhen the patients were both supine and upright in mido-drine group. In contract, Kaufmann’s study (21) showedno increase in supine blood pressure with midodrine, in anacute tilt test. Although both studies confirmed the efficacyof midodrine for VVS, it was not clear whether arterial con-struction was one of the main improving factors. However,reduced pooling of blood in the lower part of the body byvenoconstriction was proved by several studies and is ac-knowledged by most investigators. One recent study con-ducted by Maxime et al. found that a low dose (5 mg/day)midodrine decreased plasma norepinephrine concentrationsand heart rate of healthy volunteers independent of systemicblood pressure while the atrial natriuretic peptide (ANP)concentrations increased, compared with placebo. These re-sults indicated that 5 mg midodrine had sympatholytic prop-erties and had effects more obviously on capacitance vesselsthan on arteries. The mechanism of this difference is unclear.

However, it was presumed that the sympatholytic propertiesof midodrine might be due to the interaction with cardiopul-monary reflex and autonomic nervous system (29). That ca-pacitance vessels and venous return are more sensitive thanarterioles and blood pressure to vasoconstrictive stimuli inanimal studies, may also give some clues (30). Both of thesympatholytic properties and effects on capacitance vesselsand venous return may be responsible for the efficacy ofmidodrine on VVS.

The six RCTs involved two a-agonists. Interestingly, fourof the six studies (13,20–22) referring to midodrine showedsuperiority of the active medicine over placebos, while theother two studies (14,19) of etilefrine failed to support theefficacy. Further analysis might disclose a higher weightedmean percentage of responders to midodrine for treatingVVS than etilefrine. There may be several reasons for theseresults. Midodrine is a selective a1-agonist that can causeconstrictive effects on both peripheral arteries and veins(13). Below a certain dose level, midodrine causes decreasesin venous capacity without significant influence on bloodpressure (21). The pharmacological effects of midodrine fortreating VVS have been stated above. While former stud-ies suggested that etilefrine has both b1- and a-adrenergiceffects (31), Mellander’s studies (32) of the cat indicatedthat etilefrine caused reflex vasodilation in low doses anda-receptor-mediated vasoconstriction of peripheral vesselsat higher doses. In Coleman’s studies (31) of man, thedose-related fall in peripheral vascular resistance associatedwith etilefrine infusion stopped when a dose of more than100 lg ⁄ kg was administered. They also found that an a-receptor-mediated rise in venous return was less importantthan b1-adrenergic stimulation in determining increases incardiac output following administration; after using etile-frine accompanied with a-blockade, cardiac output didnot decrease. It is possible that the stimulation of b1-adrenoceptor can influence the efficacy for treating VVS.Furthermore, midodrine does not cross the blood–brain bar-rier, and side effects such as anxiety, agitation and insom-nia can be avoided. Etilefrine has both central and periph-eral a-adrenergic effects. Differences in the individuals’ age,haemodynamics type, dosage of medicine and symptoms be-fore treatment (shown in Table 5) should also be taken intoconsideration when comparing the two medicines.

Although the six studies were all of RCTs, publicationbias may exist, according to the funnel plot. Study design,including number of subjects, generation of random groups,utilization of blinding, allocation concealment, protocol offollow-up and indices for investigation, may have affectedthe result (33). Additionally, studies with negative resultscannot always be published favourably and with completeaccuracy, which leads to publication bias. Of the six studies,only one enrolled over 100 subjects across multiple centres,and two of them were unblinded. In addition, there wasno golden standard for evaluating the efficacy of medicinefor VVS, and the value of HUT for estimating the effectwas suspected in recent researches (19). As a result, furtherstudies are still needed to reduce bias and produce moreforceful evidence.



CONCLUSIONSAs shown in our study, the currently published RCTs supportthat a-adrenoceptor agonists can be effective treatments forVVS. Midodrine might be a better choice compared withetilefrine for VVS. More large-scale, multifaceted and better-designed studies are still expected to give more powerfulevidence.

ACKNOWLEDGEMENTSThis work was supported by The National Tenth Five-Year Plan Research Project of China (2004BA720A10),Changjiang Scholar Program and the Major Basic ResearchProject of China (2006 CB503807) National Natural Sci-ence Foundation of China (30821001) and Capital MedicalDevelopment Fundation (2007–2003).

References

1. Ganzeboom KS, Mairuhu G, Reitsma JB, Linzer M, Wieling W,van Dijk N. Lifetime cumulative incidence of syncope inthe general population: a study of 549 Dutch subjects aged35–60 years. J Cardiovasc Electrophysiol 2006; 17: 1172–6.

2. Colman N, Nahm K, Ganzeboom KS, Shen WK, Reitsma J,Linzer M, et al. Epidemiology of reflex syncope. Clin AutonRes 2004; 14: 9–17.

3. Du JB. Mechanism and clinical diagnosis of vasovagal syncopein children. Chin J Med 2002; 37: 8–10.

4. Brignole M. Randomized clinical trials of neurally mediatedsyncope. J Cardiovasc Electropysiol 2003; 14: S64–9.

5. Di Gerolamo E, Di Iorio C, Sabatini O, Leonzio L, Barbone C,Barsotti A. Effects of paroxetine hydrochloride, a selectiveserotonin reuptake inhibitor, on refractory vasovagal syncope:a randomized, double-blinded, placebo-controlled study. J AmColl Cardiol 1999; 33: 1227–30.

6. Zeng C, Zhu Z, Liu G, Hu W, Wang X, Yang C, et al.Randomized, double-blinded, placebo-controlled trial of oralenalapril in patients with neurally mediated syncope. AmHeart J 1998; 136: 852–8.

7. Parry SW, Kenny RA. The management of vasovagal syncope.Q J Med 1999; 92: 697–705.

8. Lee TM, Su SF, Chen MF, Liau CS. Usefulness of transdemalscopolamine for vasovagal syncope. Am J Cardiol 1996; 78:480–2.

9. Morillo CA, Leitch JW, Yee R, Klein GL. A placebo-controlledtrial of intravenous and oral disopyramide or prevention ofneurally mediated syncope induced by head-up tilt. J Am CollCardiol 1993; 22: 1843–8.

10. Sheldon R, Stuart C, Sarah R, Klingenheben T, Krahn A,Morillo C, et al. Prevention of Syncope Trial (POST): arandomized, placebo-controlled study of metoprolol in the

prevention of vasovagal syncope. Circulation 2006; 113:1164–70.

11. Mahanonda N, Bhuripanyo K, Kangkagate C, Wansanit K,Kulchot B, Nademanee K, et al. Randomized double-blind,placebo-controlled trial of oral atenolol in patients withunexplained syncope and positive upright tilt table test results.Am Heart J 1995; 130: 1250–3.

12. Brignole M, Alboni P, Benditt DG, Bergfeldt L, Blanc JJ,Thomsen PE, et al. The Task Force on Syncope, EuropeanSociety of Cardiology. Guidelines on management (diagnosisand treatment) of syncope-update 2004. Executive Summary.Eur Heart J 2004; 25: 2054–72.

13. Ward CR, Gray JC, Gilroy JJ, Kenny RA. Midodrine: a role inthe management of neurocardiogenic syncope. Heart 1998;79: 45–9.

14. Raviele A, Brignole M, Sutton R, Alboni P, Giani P, MenozziC, et al. Effect of etilefrine in preventing syncopal recurrencein patients with vasovagal syncope: a double-blind,randomized, placebo-controlled trial. Circulation 1999; 99:1452–7.

15. Juni P, Altman DG, Egger M. Assessing the quality of

controlled clinical trials. BMJ 2001; 323: 42–6.16. Yu ZB, Han SP, Guo XR. Meta-analysis of inhaled nitric oxide

for hypoxaemic respiratory failure in preterm infants. Chin JEvid Based Pediatr 2007; 2: 327–37.

17. Wu TX, Liu GJ. The definition, practice and report of

allocation concealment and blinding. Chin J Evid-based Med2007; 7: 222–4.

18. Liu JP. Compliance and intention-to-treat analysis in

randomized controlled trial. Chin J TW Med 2003; 23: 884–6.19. Moya A, Permanyer-Miralda G, Sagrista-Sauleda J, Carne X,

Rius T, Mont L, et al. Limitations of head-up tilt test for

evaluating the efficacy of therapeutic interventions in patients

with vasovagal syncope: results of a controlled study of

etilefrine versus placebo. J Am Coll Cardiol 1995; 25:

65–9.20. Perez-Lugones A, Schweikert R, Pavia S, Sra J, Akhtar M,

Jaeger F, et al. Usefulness of midodrine in patients with

severely symptomatic neurocardiogenic syncope: a

randomized control study. J Cardiovasc Electrophysiol 2001;

12: 935–8.21. Kaufmann H, Saadia D, Voustianiouk A. Midodrine in

neurally mediated syncope: a double-blind, randomized,crossover study. Ann Neurol 2002; 52: 342–5.

22. Zhang QY, Du JB, Tang CS. The efficacy of midodrinehydrochloride in the treatment of children with vasovagalsyncope. J Pediatr 2006; 149: 777–80.

23. Nair N, Padder FA, Kantharia BK. Pathophysiology andmanagement of neurocardiogenic syncope. Am J Manag Care2003; 9: 327–34.

24. Blair PG. Neurocardiogenic syncope. N Engl J Med 2005; 352:1004–10.

Table 5 Dose of medicine and symptom before treatment of six RCTs

Study Medicine Dose Symptom before treatment

Moya 1995 (19) Etilefrine 10 mg Tid Experience syncope before

Ward1998 (13) Midodrine 5 mg Tid ‡2 episodes every month

Raviele1999 (14) Etilefrine 25 mg Tid ‡3 episodes in last 2 years

Perez-Lugones 2001 (20) Midodrine 5–15 mg Tid ‡1 episode every month

Kaufmann 2002 (21) Midodrine 5 mg st ‡2 episodes in last year

Zhang 2006 (22) Midodrine 1.25–2.5 mg Bid ‡3 episodes every year

HUT ¼ head-up tilt test



25. Ku BS. Pharmacology. Beijing: Peking University Med Press,2004: 105–9.

26. Parry SW, Kenny RA. The management of vasovagal syncope.Q J Med 1999; 92: 697–705.

27. Wei QM, Yao LM, FU XH. Treatment and mechanism ofvasovagal syncope. Adv Cardiovasc Dis 2005; 26: 655–9.

28. Hainsworth R. Reflexes from the heart. Pkysiol Rev 1991; 71:617–58.

29. Lamarre-Cliche M, du Souich P, de Champlain J, LarochelleP. Pharmacokinetic and pharmacodynamic effects ofmidodrine on blood pressure, the autonomic nervous system,and plasma natriuretic peptides: a prospective, randomized,single-blind, two-period, crossover, placebo-controlled study.Clin Ther 2008; 30: 1629–38.

30. Mellander S. Comparative studies on the adrenergicneuro-hormonal control of resistance and capacitance bloodvessels in the cat. Acta Physiol Scand Suppl 1960; 50:1–86.

31. Coleman AJ, Leary WP, Asmal AC. The cardiovascular effectsof etilefrine. Eur J Clin Pharmacol 1975; 8: 41–5.

32. Mellander S. Comparative effects of acetylcholine,butyl-nor-synephrine (Vasculat), noradrenaline, andethyl-adrianol (Effortil) on resistance capacitance andprecapillary sphincter vessels and capillary filtration in catskeletal muscle. Angiologia 1966; 3: 77–99.

33. Qu HQ, Lin S, Qiu MC. Brief introduction on studies bias ofsystematic reviews. Chin J Evid-based Med 2001; 1:114–16.



α-Adrenoceptor agonists for the treatment of vasovagal syncope: a meta-analysis of worldwide...

Documents

Transcript of α-Adrenoceptor agonists for the treatment of vasovagal syncope: a meta-analysis of worldwide...