trials. We conclude that electrodermal testing cannot
diagnose allergy to common aeroallergens such as cat
dander and house dust mite—allergens that have a
strong association with atopic respiratory disorders
such as asthma, rhinitis, eczema, and conjunctivitis.
We thank Jackie Burnham for her help in preparing the manu-
script, all the volunteers and the operators who gave up their
time so willingly,and the doctors and practice manager at Hythe
Medical Centre who made space and time available to us.
Contributors: GTL, JNK, and STH conceived the study,
developed the protocol, and obtained funding. JB was the trial
manager. JG was responsible for the statistical design and
randomisation, and PP for the analysis. All authors were jointly
involved in writing the paper. STH will act as guarantor.
Funding: The study was supported by funding from Vega
Grieshaber, the Asthma Allergy and Inflammation Research
Trust, and the Foundation for Integrated Medicine.
Competing interests:GTL and JNK formerly had interests in
a company involved in the distribution of Vega Grieshaber
products and both have received fees for speaking at a seminar
on the Vegatest.
1 Kenyon J,ed.Short manual of the Vegatest method.Schiltach,Germany:Vega
Van Wijk R, Wilegart FAC. Homoeopathic remedies and pressure induced
changes and the galvanic resistance of the skin. Utrecht: State University of
Utrecht, Research Unit for Complementary Medicine, 1989.
Ho MW. The rainbow and the worm—the physics of organisms.
Singapore and London: World Scientific, 1993.
Tsuei J and Madill P. A food allergy study using the EAV acupuncture
technique. Am J Acupunct 1984;12:105-16.
Fuller Royal F, Fuller Royal D. Scientific support for electrodiagnosis. Br
Homoeopathic J 1991;18:166-78.
Fox A. Determination of neutralisation point for allergic hypersensitivity.
Br Homoeopathic J 1987;76:230-4.
Ali M.Correlation of IgE antibodies with specificity for pollen and mould
allergy changes in electrodermal skin responses following exposure to
allergens. Am J Clin Pathol 1989;91:357-9.
Krop J, Swiesczek J, Wood A. Comparison of ecological testing with the
Vegatest method in identifying sensitivities to chemicals, foods and inha-
lants. Am J Acupunct 1985;13:253-9.
Krop J, Lewith, G, Gziut W, Radulescu C. A double-blind, randomised,
controlled investigation of electrodermal testing in the diagnosis of aller-
gies. J Altern Complement Med 1997;3:241-8.
10 Dreborg A, Frew A. Allergen standardisation and skin tests. EAACI ’93
subcommittee on skin tests. Allergy 1993;14:48-82.
11 Gallant SP, Maibach HI. Reproducibility of allergy epicutaneous
techniques. J Allergy Clin Immunol 1973;51:245-50.
12 Basomba A. Evaluation of changes in skin sensitivity by means of skin
tests. EAACI ’93 subcommittee on skin tests. Allergy 1993;14:71-5.
(Accepted 20 October 2000)
Treatment for the premenstrual syndrome with agnus
castus fruit extract: prospective, randomised, placebo
R Schellenberg for the study group
Objectives To compare the efficacy and tolerability of
agnus castus fruit (Vitex agnus castus L extract Ze 440)
with placebo for women with the premenstrual
Design Randomised, double blind, placebo
controlled, parallel group comparison over three
Setting General medicine community clinics.
Participants 178 women were screened and 170 were
evaluated (active 86; placebo 84). Mean age was 36
years, mean cycle length was 28 days, mean duration
of menses was 4.5 days.
Interventions Agnus castus (dry extract tablets) one
tablet daily or matching placebo, given for three
Main outcome measures Main efficacy variable:
change from baseline to end point (end of third cycle)
in women’s self assessment of irritability, mood
alteration, anger, headache, breast fullness, and other
menstrual symptoms including bloating. Secondary
efficacy variables: changes in clinical global
impression (severity of condition, global
improvement, and risk or benefit) and responder rate
(50% reduction in symptoms).
Results Improvement in the main variable was
greater in the active group compared with placebo
group (P < 0.001). Analysis of the secondary variables
showed significant (P < 0.001) superiority of active
treatment in each of the three global impression
items. Responder rates were 52% and 24% for active
and placebo, respectively. Seven women reported mild
adverse events (four active; three placebo), none of
which caused discontinuation of treatment.
Conclusions Dry extract of agnus castus fruit is an
effective and well tolerated treatment for the relief of
symptoms of the premenstrual syndrome.
The premenstrual syndrome is a complex combination
of psychological symptoms,including irritability,aggres-
sion, tension, anxiety, and depression, and somatic
changes such as fluid retention,breast tenderness,head-
ache, feeling of bloating, and weight increase.1Women
are affected irrespective of socioeconomic status,race,or
cultural background, and family clusters are well
The causes of the premenstrual
syndrome have not been clearly elucidated4but have
been attributed to hormonal change,neurotransmitters,
What is already known on this topic
Unconventional testing for allergy, such as
electrodermal testing, is widely available
Few clinical trials have evaluated unconventional
What this study adds
Electrodermal testing cannot distinguish between
atopic and non-atopic individuals as previously
determined by skin prick tests
Institute for Health
Care and Science,
BMJ VOLUME 32220 JANUARY 2001bmj.com
prostaglandins, diet, drugs, and lifestyle, so causal
treatment is difficult.3
The fruits of Vitex agnus castus (the chaste tree)
contain a mixture of iridoids and flavonoids, and some
compounds similar in structure to the sex hormones
have been isolated from the leaves and flowers.5The
effects of agnus castus have been described as similar to
those of the corpus luteum.6 7The mechanism of
action may also be related to modulation of stress
induced prolactin secretion via dopamine, without
directly affecting luteinising hormone or follicle stimu-
lating hormone.8–15Binding to opioid receptors,16?
endorphins,17and neuroactive flavonoids18 19may also
have a role. The plant has been used traditionally to
relieve the symptoms of the premenstrual syndrome,
although systematic evaluation of its efficacy is
relatively recent.2 8 20–22
We evaluated the effects of agnus castus fruit (Ze
440) in a large, prospective, randomised, placebo con-
trolled study over three consecutive menstrual cycles.
The study and analysis were planned and conducted
under strict methodology.
Recruitment and blinding
All women were outpatients attending six general
medicine clinics from April to December 1998. All
physicians were experienced in the use the instruments
of assessment and underwent training before the study
on the scales to enhance consistency within and
between centres. In each clinic all assessments were
made by the same individual.
Study medication consisted of either active (fruit
extract ZE 440: 60% ethanol m/m, extract ratio 6-12:1;
standardised for casticin; one 20 mg tablet once daily)
or placebo (matched for appearance, size, colour, taste,
The study was performed according to current
European Union and International Conference on
Harmonisation guidelines on Good Clinical Practice
and the Declaration of Helsinki and was approved by
the ethics committees of the Hessen and Baden-
Württemberg Regional Orders of Physicians.
Selection of participants
All women were aged >18 years,had the premenstrual
syndrome diagnosed according to the Diagnostic and
Statistical Manual of Mental Disorders, third edition,
and gave written informed
consent before entry. Exclusion criteria were participa-
tion in other trials, concomitant psychotherapy,
pregnancy or breast feeding, inadequate contracep-
tion, dementia, alcohol or drug dependence, concomi-
tant serious medical condition, hypersensitivity to
agnus castus, fever, pituitary disease, and concomitant
use of sex hormones except oral contraceptives for
which the doses were kept unchanged.
Baseline assessment was at the start of the first cycle.
Women’s self assessment at baseline consisted of the
combined scores of the following six symptoms relative
to their previousthree
alteration, anger, headache, other menstrual symptoms
(including bloating), breast fullness. Patients also under-
(haematology,biochemistry,pregnancy test),and evalua-
cycles: irritability, mood
tion of clinical global impression (item 1—severity of
condition).24A visit at the start of the second cycle was
allowed but not encouraged so as to minimise the
chances of closer contacts with the doctor influencing
the condition. A visit at the end of the third cycle (end
point) was mandatory. Women then underwent full
medical examination, laboratory tests (haematology,
biochemistry), self assessment, clinical global impres-
sion,compliance checks,and adverse events monitoring.
Statistics, assignment, and analysis
All data were
department of medical information technology at the
University of Giessen. Randomisation was provided
centrally in blocks of four. The main efficacy variable
prospectively sought in the protocol was change from
baseline to end point in the combined scores of the six
self assessment items (irritability, mood alteration,
anger, headache, other menstrual symptoms including
bloating, and breast fullness). Women rated each item
using a visual analogue scale25validated for the assess-
ment of the premenstrual syndrome, ranging from 0
(no symptoms) to 10 (unbearable), measured in
millimetres on the linear scale. We chose the scale
because it was validated and also because it correlates
processed and analysedby the
Diagnostic criteria for the premenstrual syndrome
A: In most menstrual cycles in the past year the symptoms listed under (B)
occurred during the last week of the luteal phase and disappeared again a
few days after onset of the follicular phase. In menstruating women this
phase corresponds to the week before and a few days after the start of
B: At least four of the following symptoms must be present in the late luteal
phase of most cycles, with at least one symptom being from groups 1, 2, 3,
1 Clearly unstable mood (for example, sudden sadness)
2 Clear and persistent irritability or anger
3 Clear anxiety, nervous tension, or feeling of being overwrought
4 Clearly depressive mood, feeling of hopelessness, or reduction in self
5 Reduced interest in activities such as work, family, friends, hobbies
6 Rapid tiring and clear lack of energy
7 Subjective feeling of not being able to concentrate
8 Clear change in eating behaviour, such as increased appetite or craving
for sweet foods
9 Sleep disorders such as too much sleep or sleeplessness
10 Other physical symptoms such as tender breasts, feeling of bloating,
headache, joint or muscle pain, weight gain
C: The symptoms are such that they are invariably noticeable and create
difficulties in the normal course of daily life.
D: The symptoms are not merely an aggravated expression of the
symptoms of other diseases such as depression.
Screened and randomised (n=178)
Placebo (n=87) Active (n=91)
Baseline only (n=3)
Baseline only (n=6)
Intention to treat analysis
Intention to treat analysis
Flow of participants through study
BMJ VOLUME 322 20 JANUARY 2001bmj.com
closely with other more complex tools and is a rapid
and straightforward assessment of effect for use in
community practice.21 25 26Secondary variables were
condition), 2 (global improvement or deterioration),
and 3 (overall treatment assessment, risk or benefit)
items1 (severity of
and responder rate, defined as >50% reduction in self
assessed symptoms from baseline.
Our working hypothesis was that active medication
is superior to placebo.We defined the intention to treat
population as all randomised women who had at least
one baseline value and one value after baseline and
took any medication. Secondary
evaluated descriptively, according to Abt.27We calcu-
lated means, SD, medians, 95% confidence intervals,
minima-maxima, and absolute and relative frequen-
cies. We calculated that a difference in mean values of
12 points and 2.5-fold SD would be clinically meaning-
ful.25The expected withdrawal rate was 10%. We calcu-
lated that a sample size of 80 per group would give a
statistical power of 80%.28
Women’s characteristics and flow through study
A total of 178 women were initially screened and ran-
domised, of whom 170 had at least one baseline and
one post-baseline value recorded (intention to treat
population, figure). Characteristics at entry were
balanced between the two groups (table 1). Twenty
three women (13%) were taking oral contraceptives (11
in the active group).
Table 2 shows the results at the end of the treatment
period (three cycles). Analysis of the main efficacy vari-
able (difference from baseline to end point between the
two groups in self assessed symptoms) rejected the null
hypothesis of the two treatments being equal. Patients
who received agnus castus had a significant improve-
ment in combined symptoms compared with those on
placebo. The self assessment scores were corroborated
by the physicians’ own evaluation of all three items of
the clinical global impression scale. Five of the six self
assessment items indicated significant superiority for
agnus castus (irritability, mood alteration, anger, head-
ache, and breast fullness), other symptoms including
bloating being unaffected by treatment. The responder
rate was also favourable for the treated group relative
to placebo (52% v 24%, respectively). Subgroup analy-
ses that excluded women taking oral contraceptives
and included the eight women who were screened but
did not have any post-baseline values (assigned mean
of combined groups) did not alter the results.
There were few adverse events (table 3). Because of the
low incidence of events we could not ascertain the
commonest events related to agnus castus, despite the
large sample sizes of the two groups.One woman with-
drew because of pregnancy (placebo group) 55 days
into the trial.
We have shown that agnus castus (Ze 440 extract
tablets) is an effective treatment for women with the
premenstrual syndrome. Treatment of this condition is
a challenge for doctors, who most often treat these
women in the community.2 8 20–22Although the condi-
tion tends to be mild to moderate, the effects on
women can be debilitating.3We planned the study with
an open minded approach to the efficacy of agnus cas-
agnus castus extract as treatment for the premenstrual
symdrome (active treatment n=86; placebo n=84)
Characteristics of 170 women at entry into study of
Cycle length (days):
Age at menarche (years):
Menses duration (days):
Patient self assessment on visual analogue score (mm):
Clinical global impression item 1 (severity of condition):
the premenstrual syndrome (active treatment n=86; placebo n=84)
Results at end point (end of three consecutive cycles) in 170 women with
Difference in mean
reduction (active minus
placebo) (95% CI)P value*
Primary efficacy variable:
Differences in self assessment†
Secondary efficacy variables :
−50.5 (−23.5 to −77.5)0.001
Irritability (self assessment)
Mood alteration (self assessment)
Anger (self assessment)
Headache (self assessment)
Others and bloating (self assessment)
Breast fullness (self assessment)
Severity of condition (CGI item 1)
Improvement/deterioration (CGI item 2)
Overall assessment risk or benefit (CGI item 3)
Responder rate (%)§
−10.7 (−3.4 to −18.0)
−11.1 (−4.4 to −17.8)
−10.3 (−3.1 to −17.5)
−11.9 (−4.3 to −19.6)
1.3 (−5.5 to 8.1)
−9.2 (−2.8 to −15.6)
−0.5 (− 0.1 to −0.8)
−1.0 (−0.7 to −1.3)‡
0.7 (0.4 to 1.0)‡
CGI=clinical global impression.
*Mann-Whitney rank sum test.
†Visual analogue scale combined for six items.
‡Difference in means with 95% CI.
§Response >50% improvement in combined visual analogue scale from baseline to end point.
Table 3 Adverse events and withdrawals from treatment
Agnus castus (n=86)
Acne, multiple abscesses,
intermenstrual bleeding, urticaria
Acne, early menstrual period,
No reporting event
All events (regardless of relation to
*Each event was reported once, severity was mild for all events. All events resolved without treatment
BMJ VOLUME 32220 JANUARY 2001bmj.com
tus. Although other investigators used active compara- Download full-text
tors,21we set out to evaluate whether or not this herbal
treatment had any relevant effect on the condition. By
choosing placebo rather than an active control and by
informing patients of the likelihood of receiving
placebo we reduced the chances of any psychological
influence that an active control would have on the two
groups. We reduced the clinic visits to a minimum to
reassurance that sometimes can be gained from
frequent medical check ups. Careful screening of the
women resulted in relatively few losses to follow up,
which can be a problem in studies of this condition.
We chose the duration of treatment of three cycles to
overcome the variability of the symptoms over time,and
we set the diagnosis strictly according to established cri-
teria. The instruments of assessment were robust and
validated, and all participating physicians underwent
prior training with the scales so the results were consist-
ent across the investigating clinics (no centre effect).
Patient acceptance is crucial in the treatment of this
condition. The effects of agnus castus were clear to
both the women and physicians. We have no data on
the progress of symptoms after cessation of treatment,
but other researchers have shown that though the
effects of treatment gradually decrease over time, they
are still felt for as long as three cycles after treatment
(Berger et al, personal communication).
Tolerability of agnus castus was good in this study.
Although herbal treatments are sometimes misguid-
edly considered to be completely safe, we did not iden-
tify any event related to treatment that could be
considered prevalent in this population. In this regard,
the fact that patients were aware of the chances of
receiving either a herbal or placebo tablet may have
played a part in the low incidence of reported events.
as possible fromthe
Agnus castus is a well tolerated and effective for the
treatment for the premenstrual syndrome, the effects
being confirmed by physicians and patients alike. The
effects are detected in most main symptoms of the syn-
drome. This herbal remedy ought to be considered a
therapeutic option in women in whom a causal origin
for this syndrome cannot be established.
Contributors: E Schrader, Pohlheim, was the study coordinator.
The investigators were R Schellenberg (Hüttenberg), G Kunze
(Rottweil), E R Pfaff (Mörfelden), A Massing (Offenbach), H
Wältner (Albstadt-Tailfingen), Dr M Kirschbaum (Universitäts-
Frauenklinik Giessen), R H Boedecker (statistics, Giessen). J
Dudeck (statistics, Giessen) was responsible for statistics and
data management. RS is guarantor.
Funding: Zeller AG, CH-8590, Switzerland supplied the
study medication and sponsored the study through the Clinical
Research Organisation Praxis Klinische Arzneimittelforschung,
Competing interests: None declared.
1 Chuong CJ, Coulam CB. Current views and the beta-endorphin hypoth-
esis. In: Gies LH, Kase NG, Berkowith C, eds. The premenstrual syndromes.
New York: Churchill Livingstone, 1988:75-95.
Peters-Welte C,Albrecht M.Regeltempostörungen und PMS.Vitex agnus
castus in einer Anwendungsbeobachtung. TW Gynäkologie 1994;7:49-52.
Küpper C, Loch E-G. Prämenstruelles Syndrom. Deutsche Apotheker
Sondheimer S J. Etiology of premenstrual syndrome. In: Smith S, Schiff I,
eds. Modern management of premenstrual syndrome. New York: Norton
Medical Books, 1994:46-54.
Brickell C, ed. Royal Horticultural Society encyclopaedia of plants and flowers.
London: Dorling Kindersley, 1989.
Du Mee C. Vitex agnus castus. Aust J Med Herbalism 1993;5:63-5.
7 Hobbs C. The chaste tree: Vitex agnus castus. Pharm History 1991;23:
Milewicz A, Gejdel E, Sworen H, Sienkiewicz K, Jedrzejak J, Teucher T, et
al. Vitex agnus castus—Extrakt zur Behandlung von Regeltempoanomal-
ien infolge latenter Hyperprolaktinämie [Vitex agnus castus extract in
the treatment of luteal phase defects due to latent hyperprolactinemia.
Results of a randomized placebo-controlled double-blind study]. Drug
Sliutz G, Speiser P, Schultz AM, Spona J, Zeillinger R. Agnus castus
extracts inhibit prolactin secretion of rat pituitary cells. Hormone Metab
10 Winterhoff H, Gorkow C, Behr B. Die Hemmung der Laktation bei Rat-
ten als indirekter Beweis für die Senkung von Prolaktin durch Agnus cas-
tus. Z Phytotherapie 1991;12:175-9.
11 Wuttke W, Gorkow S, Jarry H. Dopaminergic compounds in Vitex agnus
castus. In: Loew D, Rietbrock N, eds. Phytopharmaka in Forschung und kli-
nischer Anwendung, Darmstadt: Steinkopff, Verlag, 1995:81-9.
12 Jarry H, Leonhardt S, Gorkow C, Wuttke W. In vitro prolactin but not LH
and FSH release is inhibited by compounds in extracts of Agnus castus:
direct evidence for a dopaminergic principle by the dopamine receptor
assay. Exp Clin Endocrinol 1994;102:448-54.
13 Merz PG, Gorkow C, Schrödter A, Rietbrock S, Sieder C, Loew D, et al.
The effects of a special Agnus castus extract (BP 1095E1) on prolactin
secretion in healthy male subjects. Exp Clin Endocrinol Diabetes
14 Berger D, Burkard W, Schaffner W, Meier B. Rezeptorbindungsstudien
mit daraus isolierten Substanzen. In: Meier B, Hoberg E, eds. Agni-casti-
fructus—Neue Erkenntnisse zur Qualität und Wirksamkeit; Z Phytothera-
15 Carroll BJ, Steiner M. The psychobiology of premenstrual dysphoria: the
role of prolactin. Psychoneuroendocrinol 1978;3:171-80.
16 Brugisser R, Burkard W, Simmen U, Schaffner W. Untersuchungen an
Opioid-Rezeptoren mit Vitex agnus-castus L. In: Meier B, Hoberg E, eds.
Agni-casti-fructus—Neue Erkenntnisse zur Qualität und Wirksamkeit. Z
17 Samochowiec L, Glaesmer R, Samochowiec J. Einfluss von Möchspfeffer
auf die Konzentration von beta-Endorphin im Serum weiblicher Ratten.
Ärztezeitschr Naturheilverfahren 1998;39:213-5.
18 Medina JH, Viola H, Wolfman C, Marder M, Wasowski C, Calvo D, et al.
Neuroactive flavonoids: new ligands for the benzodiazepine receptors.
19 Shiah IS, Yatham LN. GABA function in mood disorders: an update and
critical review. Life Sciences 1998;63:1289-303.
20 Jarry H, Leonhardt S, Wuttke W, Behr B, Gorkow C. Agnus castus als
dopaminerges Wirkungsprinzip in Mastodynon N. Z Phytotherapie
21 Lauritzen C, Reuter HD, Repges R, Böhnert KJ, Schmidt U. Treatment of
premenstrual tension syndrome with Vitex agnus castus—controlled,
double-blind study versus pyridoxine.Phytomedicine 1997;4:183-9.
22 Roeder D. Therapie von Zyklusstörungen mit Vitex agnus-castus. Z Phy-
23 American Psychiatric Association. Diagnostic and statistical manual of
mental disorders. 3rd ed, rev. Washington, DC: American Psychiatric
24 National Institute of Mental Health. 028 CGI clinical global impressions.
In:Guy W,ed.ECDEU assessment for psychopharmacology.Rev ed.Rockville,
25 Casper RF, Powell AM. Premenstrual syndrome: documentation by a lin-
ear analog scale compared with two descriptive scales. Am J Obstet Gynecol
26 Steiner M, Haskett RF, Carroll BJ. Premenstrual tension syndrome: the
development of research diagnostic criteria and new rating scales. Acta
Psychiatr Scand 1980;62:177-90.
27 Abt K. Descriptive data analysis: a concept between confirmatory and
exploratory data analysis. Meth Inform 1987;26:77-88.
28 Machin D, Campbell MJ. Statistical tables for the design of clinical trials.
Oxford: Blackwell Scientific Publications, 1987:81-3.
(Accepted 17 November 2000)
What is already known on this topic
Women with the premenstrual syndrome are usually seen primary care
and are a therapeutic challenge to general practitioners
Treatment of the causes is often impossible, and current treatment of
the full range of symptoms is inadequate
What this study adds
Agnus castus is effective in the treatment of the premenstrual syndrome
The effects were confirmed by women’s self assessment and by the
Over half the women had a 50% or greater improvement in their
Patient acceptance was high and side effects were few and mild
BMJ VOLUME 32220 JANUARY 2001bmj.com