Heel.es

Treatment of Vertigo with aHomeopathic Complex RemedyCompared with Usual Treatments A meta-analysis of clinical trials
Berthold Schneider a, Peter Kleinb, and Michael Weiserc
Institut für Biometrie, Medizinische Hochschule Hannovera, Hannover (Germany), d.s.h. statistical servicesb,Rohrbach (Germany), and Institut für Antihomotoxische Medizin und Grundregulationc, Baden-Baden (Germany) The increasing interest in alternative
groups); mean reduction of the intensity
medical practices has led to a number of
(on a scale 0−4) for VH 1.8 and for the
controlled studies on herbal and
control 1.8 (standard error 0.03 for both
homeopathic agents. This paper presents
groups). In the non-inferiority analysis
the results of a meta-analysis of four re-
from all trials, VH was non-inferior in all
cent clinical trials evaluating the
variables. The results show the applicabil-
homeopathic preparation Vertigoheel
ity of meta-analyses on the data from
(VH) compared with usual therapies (be-
studies with homeopathic drugs and sup-
tahistine, Ginkgo biloba extract, di-
port the results from the individual stud-
menhydrinate) for vertigo in a total of
ies indicating good efficacy and tolerabil-
1388 patients. Two trials were observa-
ity of VH in patients with vertigo.
tional studies and the other two were ran-
domised double-blind controlled trials.
The duration of treatment (6−8 weeks)
and dosage were comparable in all stud-
ies. Treatments were evaluated for the
variables “number of vertigo episodes”,
“intensity of episodes” and “duration of
episodes”. As the studies differed in the
age of patients and in the baseline values
of vertigo, the individual reductions of
number, intensity and duration of epis-
odes were adjusted on equal age and
baseline values (total means). An analysis
of variance (with studies as random ef-
fects) showed no relevant influence of
studies on the adjusted reductions and
no relevant interaction between studies
Key words
and treatment effects. The meta-analysis
of all four trials showed equivalent reduc-
᭿ Homeopathy
tions with VH and with control treat-
᭿ Vertigo
ment: mean reduction of the number of
᭿ Vertigoheel, efficacy,
daily episodes 4.0 for VH and 3.9 for con-
meta-analysis, tolerability
trol (standard error 0.11 for both
groups); mean reduction of the duration
Arzneim.-Forsch./Drug Res.
(on a scale 0−4) for VH 1.1 and for the
55, No. 1, 23−29 (2005)
control 1.0 (standard error 0.03 for both
Arzneim.-Forsch./Drug Res. 55, No. 1, 23−29 (2005)
 ECV · Editio Cantor Verlag, Aulendorf (Germany) Zusammenfassung
Behandlung des Schwindels mit einem
bei allen Studien als vergleichbar angese-
wert auf einer Skala von 0−4) bei VH 1,1
homöopathischen Komplexpräparat im
hen werden. Die Studien unterscheiden
und bei der Kontrolle 1,0 (Standardfehler
Vergleich mit anderen Therapieformen /
sich im Alter der Patienten und der Aus-
0,03); Reduktion der mittleren Intensität
Eine Meta-Analyse klinischer Studien
gangslage (Zahl der täglichen Attacken).
(Scorewert auf einer Skala von 0−4) bei
Um diese Unterschiede auszugleichen,
VH 1,8 und bei der Kontrolle 1,8 (Stan-
Es werden die Ergebnisse einer Meta-
wurden in der Meta-Analyse die indivi-
dardfehler 0,03). Die Hypothese der
Analyse von vier klinischen Studien dar-
duellen Änderungen von Anzahl, Intensi-
Nichtunterlegenheit von VH konnte bei
gestellt, in denen die Wirksamkeit und
tät und Dauer der Attacken auf gleiches
allen Studien und in der Meta-Analyse
Verträglichkeit des homöopathischen
Alter und gleiche Ausgangslage (Gesamt-
mit einer Wahrscheinlichkeit von 97.5 %
Arzneimittels Vertigoheel (VH) bei der
mittel) adjustiert. Eine Varianzanalyse
angenommen werden. Die Analyse bestä-
Behandlung des Schwindels mit der von
(mit den Studien als Zufallsfaktor) ergab
tigt somit die Ergebnisse der einzelnen
anderen gebräuchlichen Arzneistoffen
für die adjustierten Änderungen keinen
Studien, die eine klinisch relevante Wirk-
(Betahistin, Ginkgo biloba Extrakt, Di-
bedeutsamen Studieneinfluß und auch
samkeit und Verträglichkeit von VH bei
menhydrinat) bei insgesamt 1388 Patien-
keine Wechselwirkung zwischen Studien-
Patienten mit Schwindel gezeigt haben.
ten verglichen werden. Zwei dieser Stu-
und Behandlungseinfluβ. Die Ergebnisse
dien waren randomisierte, doppelblinde,
der vier Studien konnten somit zusam-
kontrollierte Studien, die beiden anderen
mengefaßt werden. Dabei ergab sich für
offene Beobachtungsstudien. Primäre
alle drei Zielgröβen eine äquivalente Bes-
Zielgröβen der Wirksamkeit waren in al-
serung unter VH und der jeweiligen Kon-
len Studien die Verbesserung der Anzahl,
trollbehandlung: Reduktion der mittle-
Intensität und Dauer der täglichen
ren Zahl der Episoden bei VH 4,0 und bei
Schwindelattacken. Die Behandlungs-
der Kontrolle 3,9 (Standardfehler 0,11);
dauer (6−8 Wochen) und Dosierung kann
Reduktion der mittleren Dauer (Score-
1. Introduction
homeopathically attenuated petroleum that has longbeen available over-the-counter in several countries Vertigo is the most common form of dizziness; a feeling with an established record of general use in the treat- of unsteadiness, spinning, whirling, or exaggerated mo- ment of vertigo [5]. The recent interest in evaluating tion when stationary. It is usually accompanied by alternative medical practices in a more rigorous man- nausea and loss of balance. Sweating, tinnitus and col- ner is reflected in a number of controlled studies on the lapse are commonly associated phenomena. Vertigo effects of VH since 1998. This situation is in marked may be caused by several factors, including head injury, contrast to that for many other agents used in comple- viral upper respiratory infection or cerebrovascular dis- ease. Other causes include tumours, inflammation of or The presence of a number of studies on VH in damage to nerves, or the use of drugs that affect the patients with the same indications allows for a system- inner ear, including aminoglycoside antibiotics, acetyl- atic review and meta-analysis of the data. Meta-analy- salicylic acid, cisplatin and furosemide. Most cases of ses are a common tool in the evaluation of treatment vertigo occur with nystagmus, an abnormal, rhythmic, effects in clinical studies [6] and have found a large ap- plication in recent decades across a wide range of indi- Common pharmacological interventions for vertigo cations [7, 8]. The outcomes of meta-analyses show are meclizine, dimenhydrinate, promethazine, scopol- smaller random errors and increased precision com- amine, atropine and diazepam. These drugs have been associated with side effects, such as drowsiness, mal- The current work presents a systematic review and aise, visual problems and dry mouth [1], which may be meta-analysis of four studies on VH with different con- one reason why alternative medical practices are often trol agents. The studies comprise a total of 1368 used to treat vertigo. The interest in complementary patients. In this meta-analysis we focused on studies medicine is increasing worldwide [2, 3], but in spite of with active controls and did not consider placebo-con- this growing attention, the possible benefits of treat- trolled trials. The four trials used three different com- ments are often not assessed in proper controlled stud- Ginkgo biloba extract and dimenhydrinate (CAS 523-87- Vertigoheel1) (VH) is a homeopathic preparation of 5), which provided a spectrum of controls reflecting the diluted plant and mineral extracts (listed in Table 1) and varied approaches to vertigo therapy in everyday clin-ical practice.
1) Manufacturer: Biologische Heilmittel Heel GmbH, Baden- Arzneim.-Forsch./Drug Res. 55, No. 1, 23−29 (2005)
24 Schneider et al. − Homeopathic complex  ECV · Editio Cantor Verlag, Aulendorf (Germany) Table 1: Constituents of VH.
used as variables, adjusted for age and baseline values. Mean differences between the VH and control treatment groups and their 95 % confidence intervals were calculated for all variables.
Randomised controlled trials and OSs were analysed separa- Petroleum rectificatum (attenuated petroleum) tely, as well as in one overall analysis of all four trials. Thehomogeneity of the adjusted mean differences between VHand control between the studies was tested using an analysisof variance with the studies as random effects. All analyseswere done using SPSS 11.5 for Windows (SPSS Software, Mun- 2. Methods
Four trials on VH in vertigo have been published and were used As all studies included active comparators, the outcomes in as the basis for this analysis [9−12]. Of these, two were random- the individual trials were not analysed for superiority. Instead, ised controlled trials (RCTs) [9, 12] and two were observational the overall outcomes were analysed for non-inferiority of VH studies (OSs). For inclusion in the analysis, trials had to meet versus active control. This approach was followed for the meta- the criteria of comparing VH with an active treatment, have a analysis. The criterion for asserting non-inferiority was that the minimum duration of 6 weeks, and to have completed before 95 % confidence interval for treatment differences between the VH group and the control group remained above the value− Data on baseline characteristics and on effects of treatment 1.0 for the variable “number of episodes” and above the value −0.5 for the variables “duration of episodes” and “inten- during follow-up were obtained from published sources and sity of episodes”. This limit of non-inferiority corresponds to from investigators or trial sponsors as needed. Data were 10 % of the maximal range of each scale. The level of signifi- checked for completeness and accuracy.
All studies evaluated the effects on the three variables No analysis was undertaken on the tolerability or the occur- “number of vertigo episodes”, “intensity of episodes” and rence of adverse events. However, descriptive data were cap- “duration of episodes”. The variable “number of episodes” was tured in the respective publications to provide a general de- quantified as numbers per day. Intensity and duration of epis- scription of the tolerability of VH compared with the respective odes were quantified on a scale from 0−4. For intensity of epis- odes, the levels were: 0 = no symptoms; 1 = mild symptoms;2 = moderate; 3 = moderate-to severe and 4 = severe symptoms.
The variable “duration of episodes” was graded as 0 = no ver- 3. Results
tigo or an episode lasting less than 2 min; 1 = duration 2 to 10min; 2 = duration 11 to 60 min; 3 = 1−6 h and 4 = vertigo Four trials were included in the meta-analysis, two RCTs episodes lasting longer than 6 h. These scales were used for all and two OSs. Two studies, one RCT and one OS, used betahistine as comparator. The other RCT used Gingko For the meta-analysis, the reductions in mean numbers of biloba extract and the other OS dimenhydrinate as con- episodes and intensity and duration scores, respectively, were Table 2: Age and baseline episodes of the four trials included in the meta-analysis.
VH = Vertigoheel, SD = standard deviation, OS = observational study, RCT = randomised controlled trial, n = number of patients.
Arzneim.-Forsch./Drug Res. 55, No. 1, 23−29 (2005)
 ECV · Editio Cantor Verlag, Aulendorf (Germany) Fig. 1: Reduction in number, intensity and duration or vertigo
Fig. 2: Reduction in number, intensity and duration or vertigo
episodes with VH and control treatments in randomised control-
episodes with VH and control treatments in observational studies
led trials (RTCs) with a) betahistine and b) Ginkgo biloba; c)
(OSs) with a) betahistine and b) dimenhydrinate; c) meta-analy-
meta-analysis of both trials. Means and standard errors of the
sis of both trials. Means and standard errors of the mean are
mean are shown, adjusted for age and baseline levels.
shown, adjusted for age and baseline levels.
A summary of the demographics and baselines of 259 patients was enrolled, in OSs 1129. Whereas the 4 number, intensity and duration of daily vertigo epis- studies differed in mean patient age (50 to 70 years) and odes of the trials is given in Table 2. The total number baseline number of daily vertigo episodes (4.3 to 6.0), of patients included was 1388; 623 of whom received there were no significant differences between the treat- VH and 753 the control medication. In RCTs a total of ment groups within the studies. The studies were Arzneim.-Forsch./Drug Res. 55, No. 1, 23−29 (2005)
26 Schneider et al. − Homeopathic complex  ECV · Editio Cantor Verlag, Aulendorf (Germany) Number of episodes (top
Intensity of episodes
(bottom scale)
Duration of episodes
(bottom scale)
Fig. 3: Meta-analysis of all four studies.
homogenous in intensity and duration scores (2.4 to 2.6 and 1.4 to 1.7, respectively). The mean duration of fol-low-up ranged from 6 weeks in the RCTs and 8 weeks Fig. 4: Non-inferiority analysis of VH versus other treatments for
vertigo. Note that the scales are different: for the variable “num-
ber of episodes” the top scale is used; for the other variables the
All trials provided information on all 3 outcomes in- bottom scale applies. The dotted line indicates the limit for
cluded in the analysis. Primary endpoints were the re- non-inferiority.
ductions of number, intensity and duration of the dailyvertigo episodes between baseline and end of the study.
confidence intervals falls within the limits of 10 % of As there were difference in age and baseline values of the maximal range of each scale: > −1 in the number of the episodes between the studies, the individual reduc- episodes and > −0.5 in the intensity and duration tions were adjusted to the mean age and baseline values scores, respectively. These conditions were fulfilled both of the study using an analysis of covariance. The ad- in the meta-analysis and in the analyses of the respec- justed mean reductions in the VH and control groups tive individual studies. Thus, the efficacy of VH in the are shown in Fig. 1 for the RCTs and in Fig. 2 for the OSs.
treatment of vertigo (i.e. reduction of number, intensity The homogeneity of the mean reductions between and duration of vertigo episodes) can be considered as the studies and a possible interaction between studies non-inferior to that of the control medications used in and treatment differences were tested using a mixed model with the studies as random and treatments as The tolerability was not included in the meta-analy- fixed effects. No significant (p > 0.01) differences in sis but published data from 4 studies describe a gener- mean reductions between the studies and no interac- ally very favourable tolerability profile for all therapies.
tions between studies and treatment differences were The total number of reported treatment-related adverse found. This analysis verified the methodological sound- events in all 4 studies was 10, of which 5 were seen with ness of combining the reductions in the individual VH, 2 with Ginkgo biloba, 2 with betahistine and 1 with dimenhydrinate. Treatment-satisfaction scores were The results of the meta analysis are shown in Fig.
generally very good, with the percentage highest satis- 3. On all three endpoints mean reductions were highly faction scores ranging from 80 % (Ginkgo biloba) [12] similar in both treatment groups: mean reduction of the to more than 90 % (all other therapies).
number of episodes: VH 4.0, control 3.9 (SEM 0.11);mean reduction of duration score: VH 1.1, control 1.0(SEM 0.03); mean reduction of intensity score: VH 1.8, 4. Discussion
This meta-analysis of 4 studies on the efficacy of the To test for non-inferiority, the 95 % confidence inter- homeopathic drug VH compared with common treat- vals for the differences in mean reduction between VH ments for vertigo shows that the homeopathic remedy and the control group was calculated. The intervals are is not inferior to the comparator medications measured shown in Fig. 4. The hypothesis of non-inferiority is ac- on the number of vertigo episodes, their duration and cepted (with probability 97.5 %), if the lower limit of the Arzneim.-Forsch./Drug Res. 55, No. 1, 23−29 (2005)
 ECV · Editio Cantor Verlag, Aulendorf (Germany) The approach of meta-analysis of data from multiple intervals comfortably within the limit defining non-in- trials in patients with the same indication and address- feriority (Fig. 4). Indeed, the border of the confidence ing the same question, in this case the non-inferiority of interval for the reduction in the duration of episodes VH versus other treatments for vertigo, reduces random with VH versus Gingko biloba does not cross the line errors and increases the precision of estimates com- of unity, indicating superiority of VH on this variable, pared with the individual trials. In the current case, all although it should be noted that none of the studies individual trials reported similar outcomes and the was designed to prove superiority. This possible superi- main added value of the meta-analysis was to increase ority of VH on reducing the intensity of vertigo could the power of the non-inferiority analysis, where the not be verified in the meta-analysis and must be con- 95 % confidence intervals of the treatment differences sidered as either a specific advantage over Ginkgo bi- were markedly smaller in the combined analysis com- loba or as a statistical play of chance in the respective pared with individual studies (Fig. 4). The relative treat- ment effects were similar for all variables, although the The trials included were all non-inferiority trials.
most prominent absolute reductions were seen in the Such analyses are commonly carried out when there are number of episodes. The results from the meta-analysis ethical obstacles to a placebo-controlled design [18]. In make the individual results seem very unlikely to have the case of the VH trials, there would be no ethical ob- jections to placebo-controlled design, however, the in- There were many similarities between the included vestigators wanted to capture a situation closer to clin- studies which make them suitable for a meta-analysis.
ical reality than placebo-controlled trials are able to All trials studied the same variables using the same provide [9−12]. Non-inferiority analyses differ from tri- means of quantification. All variables were graded on a als designed to show differences between treatments in scale from 0 to 4 except for the variable “number of that the null hypothesis in equivalence trials is that out- episodes” which was simply given as numbers of ver- comes between treatment arms are different [18]. One tigo attacks per day. A difference between the trials lay consequence of the design of non-inferiority analyses in the comparator substances used: betahistine, Ginkgo is a greater dependence on high treatment adherence biloba extract and dimenhydrinate, respectively. Betahi- for a reliable outcome. Non-inferiority analyses in trials stine is commonly used for treatment of vertigo and with a large number of discontinuations will show re- diseases such as Menier’s disease [13]. Ginkgo biloba sults biased towards non-inferiority, as the statistical extract is a widely used alternative medication for ver- power to show possible differences is reduced and the tigo, and benefits compared with placebo have been overall differences in outcomes shift towards zero [19].
demonstrated in studies of both vestibular and non- Thus, it is important to have low discontinuation rates vestibular vertigo [14] Dimenhydrinate is a commonly and high persistence with the medication. All 4 trials in available over-the-counter combination of two agents, the current meta-analysis had very high retention rates diphenhydramine and chlorotheophylline [15]. This use and in addition the relative short duration of only 6−8 of different control substances in the analysed trials is weeks contributed to low numbers of patients dropping not commonly seen as an obstacle to reliable meta- out of the studies. Thus, the conclusions from the meta- analyses and indeed, some authors expressly recom-mend the inclusion of differently designed trials in a proper meta-analysis [16]. Further, all studies used Although no specific analysis of tolerability was car- comparable doses, as recommended by the respective ried out, the published data in all 4 trials supported the manufacturers, for VH and the different control agents.
general good tolerability of complementary medica- Thus, treatment efficacies can be expected to be similar tions, specifically VH in this case. Tolerability scores in all 4 studies. This is an important point as the out- may well vary between RCTs and observational studies comes of meta-analyses can be skewed by the inclusion as both investigators and patients in RCTs are more of very large trials with inappropriately given treatment likely to look for, adjudicate and report adverse events doses in one of the treatment arms, which reduces the than patients in studies closer to clinical reality. This is overall reported effect of treatment in this arm. Such reflected in the numbers of reported adverse events in effects can be seen, e.g. in analyses of antihypertensive the studies analysed: the two RCTs, although compris- treatments [17]. Moreover, since most of the studies at- ing some 20 % of the total number of patients, reported tempted to mirror clinical practice in vertigo therapy, 90 % of all treatment-related adverse events.
which is characterised by a lack of consensus as to pre- In conclusion, the increased interest in alternative ferred regimen, the variety of comparator substances medical practices and the growing number of con- reflects the current practices. Thus, the results of this trolled trials with homeopathic and other complement- meta-analysis should be relevant to a wide spectrum of ary medications are opening the way for conducting meta-analyses and systematic reviews of published The individual non-inferiority analyses consistently data. Such analyses will add additional value to the indicated non-inferiority of VH over the control therap- study of non-standard medical practices and, it is ho- ies, with all left-hand borders of the 95 % confidence ped, help resolve the issues of sustainability of claims Arzneim.-Forsch./Drug Res. 55, No. 1, 23−29 (2005)
28 Schneider et al. − Homeopathic complex  ECV · Editio Cantor Verlag, Aulendorf (Germany) made in the individual studies. In the case of the VH [15] Pyykko, I., Magnusson, M., Schalen, L. et al., Pharma- trials in the present work, the meta-analysis supports cological treatment of vertigo. Acta Otolaryngol. Suppl. 455,
the consistently demonstrated efficacy and tolerability of this homeopathic preparation, effects that seem to [16] Meinert, C., Meta-analysis: Science or religion? Con- be at least as good as for standard therapies.
trolled Clin Trials. 10, 257 (1989)
[17] Staessen, J. A., Wang, J. G., Thijs, L., Cardiovascular pre- vention and blood pressure reduction: a quantitative overview
updated until 1 March 2003. J. Hypertens. 21, 1055 (2003)
5. References
[18] Temple, R., Ellenberg, S. S., Placebo-Controlled Trials [1] Schmid, R., Schick, T., Steffen, R. et al., Comparison of Seven and Active-Control Trials in the Evaluation of New Treatments: Commonly Used Agents for Prophylaxis of Seasickness. J.
Part 1: Ethical and Scientific Issues. Ann. Intern. Med. 133,
Travel Med. 1, 203 (1994)
[2] Haltenhof, H., Hesse, B., Bühler, K., Beurteilung und Ver- [19] Pfeffer, M. A., McMurray, J. J., Velazquez, E. J. et al., breitung komplementärmedizinischer Verfahren − eine Befra- Valsartan, captopril, or both in myocardial infarction compli- gung von 793 Ärzten in Deutschland. Gesundh. Wes. 57, 192
cated by heart failure, left ventricular dysfunction, or both. N.
Engl. J. Med. 349, 1893 (2003)
[3] Schneider, B., Hanisch, J., Weiser, M., Complementary Medicine Prescription Patterns in Germany. Ann. Pharmac- Acknowledgement
other. 38, 502 (2004)
This work was supported by an unrestricted research grant [4] De Smet, P. A., Herbal remedies. N. Engl. J. Med. 347,
from Biologische Heilmittel Heel GmbH, Baden-Baden (Ger- [5] Metzger, J., Gesichtete Homöopathische Arzneimittel- lehre, 8 ed. Karl F. Haug − Verlag, Heidelberg (1964) [6] Hedges, L., Olkin, I., Statistical methods for meta-analy- [7] Schneider, B., Ginkgo-biloba-Extrakt bei peripheren art- eriellen Verschluβkrankheiten / Meta-Analyse von kontrolli-
erten klinischen Studien. Arzneim.-Forsch./Drug Res. 42 (I), 3
(1992)
[8] Turnbull, F., Effects of different blood-pressure-lowering regimens on major cardiovascular events: results of prospec-
tively-designed overviews of randomised trials. Lancet 362,
1527 (2003)
[9] Weiser, M., Strösser, W., Klein, P., Homeopathic vs Con- ventional Treatment of Vertigo. Arch. Otolaryngol. Head Neck
Surg. 124, 879 (1998)
[10] Weiser, M., Strösser, W., Behandlung des Schwindels: Vergleichsstudie Homöopathikum vs Betahistin. Allgemeinarzt
13, 692 (2000)
[11] Wolschner, U., Strösser, W., Weiser, M. et al., Treating Vertigo − Combination Remedy Therapeutically Equivalent to
Dimenhydrinate: Results of a Reference-Controlled Cohort
Study. Biol. Med. 30, 184 (2001)
[12] Issing, W., Klein, P., Weiser, M., The Homeopathic Pre- paration Vertigoheel Versus Ginkgo Biloba in the Treatment of Correspondence:
Vertigo in an Elderly Population: A Double-Blind, Randomized, Controlled Clinical Trial. J. Altern. Complement. Med., inpress (2004) [13] Ballester, M., Liard, P., Vibert, D. et al., Meniere’s disease in the elderly. Otol. Neurotol. 23, 73 (2002)
[14] Hamann, K.-F., Physikalische Therapie des vestibulären Schwindels in Verbindung mit Ginkgo Biloba Extrakt. Therapie- woche 35, 4586 (1985)
Arzneim.-Forsch./Drug Res. 55, No. 1, 23−29 (2005)
 ECV · Editio Cantor Verlag, Aulendorf (Germany)

Source: http://www.heel.es/upload/AC_E152_Vertigoheel_Meta_analysis.pdf