Acute treatment of moderate to severe depression with hypericum extract WS 5570 (St John's wort): randomised controlled double blind non-inferiority trial versus paroxetine
BMJ 2005; 330 doi: https://doi.org/10.1136/bmj.38356.655266.82 (Published 03 March 2005) Cite this as: BMJ 2005;330:503All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
Sirs:
The reader of a study final report can only assess its internal consistency, but cannot determine whether the original data had been altered, intentionally or by mistake.
In the case of clinical trials sponsored by pharmaceutical companies, who most often utilize Contract Research Organizations, all data undergo several controls. Without listing all the obligations, it is worth mentioning that during the course of the study clinical sites are regularly inspected by professional clinical monitors and that during data analysis every single piece of information is double-checked to ensure that it is correctly transcribed from the clinical record, onto the case report form, and into the computer.
Potential from bias is particularly strong when a commercial interest is involved. Formal declarations by investigators, monitors and company sponsors might not be worth the paper on which they are written, were it not for the fact that the whole process is reviewed by Regulatory Authorities, who may disqualify investigators and penalize sponsors and monitors.
Are clinical trials on herbal remedies subjected to the same scrutiny?
Competing interests:
None declared
Competing interests: No competing interests
Having taken Hypericum extract myself for a mild depressive illness I feel one of the side-effects should be better known.
After being on 1-2 300mg tablets for a couple of months (and feeling its benefit in terms of the depression) I developed episodes of severe tachycardia. A 24-hour ECG showed runs of atrial fibrillation with ventricular rates of up to 168/min. Full investigation was (fortunately) normal and the episodes ceased completely shortly after I stopped taking the tablets.
I believe that this side-effect has been the subject of several reviews.
Competing interests:
None declared
Competing interests: No competing interests
The published Rapid Responses so far to this paper express
a variety of concerns aboutmethodologic
and ethical aspects of the randomized controlled trial (RCT) reported by
Szegedi and colleagues. Respondents have questioned the choice of the
comparator intervention (active agent rather than a placebo)
; the use of a placebo run-in period; the dose of the comparator agent
(paroxetine) ; the extract of Hypericum used; and potentially wide variations
in the amount of active agents (hyperforin and hypericin) in the Hypericum
preparation. As is typically the case with debate of this kind, some
contentions are pertinent, while others betray substantive misunderstandings of
the conduct and interpretation of RCTs.
It is unfortunate that Szegedi et al. did not use the
Discussion section of their report to discuss the trial’s results in the
context of what is known about the merits of Hypericum in the treatment of
depression. Fortunately, Ruhe et al. [1] alerted us to the presence of an
up-to-date Systematic Review [2] relevant to this issue. Has the uncertainty
regarding Hypericum’s clinical utility been resolved? What may we learn from
all this?
In reading the cited review by Linde et al., one is struck
by the Abstract’s conclusions :
Current
evidence regarding Hypericum extracts isinconsistent and confusing. In
patients who meet criteria for major depression, several recent
placebo-controlled trials suggest that Hypericum has minimal beneficialwhile other trial suggest that
Hypericum and standard antidepressantshave similar beneficial effects.
This, in spite of the fact that the
reviewers identified 37 suitable trials for analysis. The
authors identify prevalent methodologic weaknesses such as inadequate sample
sizes, improper dosages of control drugs, and inadequate quality control of
administered herbal extracts.Like
Szegedi et al., they finally express a distressingly familiar refrain : that“…more
trials that compare specific extracts with both placebo and synthetic
antidepressants in clearly defined patient populations with and without major
depression are needed.”Why has all the
research performed to date not resolved the therapeutic research question ? The answer, I believe, derives in large part from
the prevalence of poor quality clinical research performed by the psychiatric
community.
In addressing the recent controversy surrounding the merits
of anti-depressants in childhood depression, Peter Kramer, psychiatrist and
author, has written[3] :
It may just be that
Prozac is more effective than Paxil in young patients. But it is also possible
that the difference in the research results has less to do with the medication
than with the nature of the studies. It turns out that drug companies are
shockingly inept at testing their own products.
When independent
psychiatrists and psychologists look at data submitted to the Food and Drug
Administration—the results of corporations' early studies on their own
antidepressants—the medications look only modestly effective. But these
findings always strike me as suspect, since the same antidepressants have
proven highly useful in treating recalcitrant kinds of depression, such as
depression after strokes and post-partum depression. Why would the medications
fail only in research on the run-of-the-mill mood disorders?
There are a number
of answers to this puzzling question, but mostly it comes down to this: The
pharmaceutical companies do shoddy research on the
drugs' efficacy. Because the patents on medication have a limited duration, the
corporations are always in a rush to bring drugs to market. The companies
pressure the subcontractors that perform the studies, demanding that they
gather research subjects fast. The recruiters then stretch diagnostic criteria,
signing up patients who may not have the disease in question. Studies often
include people with a host of shifting complaints, many of which are based less
on acute illness than on personality style. The result is a group with poorly
defined conditions and high placebo response rates—enough static to drown out
whatever effects the medications have on substantial disease.
Drug company
studies can have other problems. Sometimes they are conducted before the proper
dosage or length of treatment has been established. Inexact dosing leads to low
rates of response, or to excessive side effects—and a tendency for subjects to
drop out of a study. These and other study design defects tend to make
medications appear less effective than they prove in clinical practice or in
later research.
Notably,
other active members of the psychiatric community have voiced similar concerns.
Two examples should suffice. Safer [4] has detailed the myriad ways in which
the validity and informativeness of psychiatric research is undermined by poor
trial methodology and reporting, and conflicting interests. Klein and
colleagues (cited by Kramer) made many similar points in a 2002 article [5],
and even called for the establishment of a putatively independent“Federal Proactive Pharmaceutical
Agency” to remedy this distressing situation.
It
is time for national professional psychiatric societies to lay aside any and
all parochial and proprietorial interests, and collaborate with each other and
other independent actors to promote the performance of scientifically valid and
ethical research. Psychiatric investigators need to develop a standard
methodology for conducting psychiatric RCTs that will produce internally valid
results in which patients and their physicians will have confidence. In this,
they would do well to collaborate with members of the CONSORT group, taking notice also
of recently published and upcoming CONSORT guidance for the conduct and
interpretation of non-inferiority trials, the proper elicitation and reporting
of harms in RCTs, and the conduct oftrials using herbal interventions.
In
formulating such guidance, psychiatric investigators also need to consult with
knowledgeable representatives of the patient and research ethics communities.
Patient
representatives (secured, for example, through collaboration with groups such
as the James Lind Alliance)
should clarify which research is most relevant to them; which interventions are
suitable comparators for particular trials; which clinical endpoints are
important to them; which patients should be recruited for a particular trial
(for example, patients in whom “standard” drugtherapy has been tried and has failed may ethically be recruited for a
placebo-controlled trial); how long trials should be run; and should help
define clinically relevant non-inferiority and superiority margins for the
analysis of a trial’s results.
Research
ethics committees should not acquiesce to this lamentable state of affairs by
approving trials that do not meet universal requirements for the ethical
conduct of research : social and scientific value; scientific validity; fair
subject selection; a favorable balance of harms and benefits; informed consent;
and respect for potential and enrolled volunteers [6]. Among the ethical
disputes that still need to be addressed are the unethical conduct of
single-blind placebo run-in periods [7], and the circumstances in which a
placebo instead of an active comparator intervention may ethically be used [8].
A collaboration should produce a checklist for
psychiatric RCTs that research ethics committees should uniformly use in the
evaluation of proposed trials.
References
1.Ruhe HG, et
al. Acute
treatment of depression with hypericum versus paroxetine reconsidered
2. Linde K, Berner M, Egger M, Mulrow C. St John's
Wort for depression: meta-analysis of randomised controlled trials. Br J
Psychiatry 2005; 186:99-107.
3. Kramer PD. Should
teenagers take drugs ?Slate.
June 4, 2004.
4. Safer DJ.Design and reporting modifications in industry-sponsored comparative
psychopharmacology trials J Nerv Ment Dis 2002; 190: 583-592
5. Kline DF, et al. Improving
clinical trials Arch Gen Psychiatry 2002; 59: 272-2786.
6. Emanuel EJ, et al. What makes clinical research ethical ?JAMA2000; 283: 2701-2711
7. Senn S. Are placebo run ins justified
? BMJ 1997; 314: 1191
Competing interests:
None declared
Competing interests: The published Rapid Responses so far to this paper expressa variety of concerns about
The authors state that the extract of St Jonh's wort, WS 5570, was
standardized to contain 3-6% hyperforin and 0.12-0.28% hypericin. This
means a variation of 100% for hyperforin and 133% for hypericin. Would we
accept a similar variation in the potency of paroxetine tablets, namely,
20-40 mg? I doubt it. The conclusions of the study, therefore, are
untenable.
Competing interests:
None declared
Competing interests: No competing interests
Wort a day!
Some time ago, the New Scientist carried an article on the benefits of
Omega 3 fatty acids in preventing depression. It quoted the example of
Japanese who have a very low risk of depression. It was believed that the
low risk is due to their high consumption of oily fish. Studies were then
done with omega 3 dietary replacement and showed a drop in depression. It
makes me wonder if increasing the consumption of oily fish in diet would
reduce the risk of depression in society and add a new slogan of 'fishy
five a day'.
Competing interests:
None declared
Competing interests: No competing interests
I would like to congratulate Szegedi et al on an excellent article,
which I found a pleasure to read. The consistency of their results across
different outcome measures and analysis populations makes it hard to
disagree with their main conclusion that 6 weeks of treatment with St
John's wort is at least as effective as paroxetine. I look forward with
interest to reading the report of their follow-up study and learning
whether that efficacy is maintained in the long-term.
However, I do have one minor criticism of the statistical methods
used for the adverse event data. The confidence interval for the incidence
rate ratio appears to have been calculated based on the assumption that
adverse events were statistically independent, but since a single patient
can have more than one adverse event, that assumption is unwarranted. A
correct calculation of the confidence interval around the incidence rate
ratio would take account of the non-independence of adverse events within
patients, and would doubtless produce a wider confidence interval than the
one they report. My guess is that it would still show a significant
difference between the treatments, but it would be interesting if Szegedi
et al could re-calculate the confidence interval taking account of the
clustering by patient and report it as a rapid response.
Competing interests:
My company has provided consultancy services to GlaxoSmithKline, although not on any projects relating to paroxetine.
Competing interests: No competing interests
Dear Editor,
I read with interest Szegedi et al and several on line rapid
responses to the article which have followed since. I would like to
highlight certain issues of clinical importance in relation to St. John’s
Wort.
First of all, it is not licensed in the UK for treatment of
depression. However St. John’s Wort is widely used as self-medication.
In addition, St. John’s Wort presents potentially serious drug
interactions. It can induce metabolism of Anti convulsants, Anti HIV
drugs, Cyclosporin, Oral contraceptives, Theophylline, Digoxin and
Warfarin, reducing their efficacy and therefore should not be taken in
combination with any of these (CSM warning 2000).
Due to minor serotonin, nor adrenaline and dopamine reuptake
inhibitory activity St. John’s Wort potentiates existing MAOI (Mono
amine oxidase inhibitor), SSRI (selective serotonin re uptake inhibitor)
and Tricyclic Antidepressant therapy and therefore, concurrent use of St.
John’s Wort is best avoided, particularly at higher doses.
Serotonin syndrome and other serious adverse effects including
toxicity have been reported with St. John’s Wort.
St.John’s Wort may exacerbate bipolar disorder by inducing mania.
Safety of St. John’s Wort in children, pregnancy and in the elderly
is not well established.
Antidepressant effect of St. John’s Wort in severe depression is
still disputed.
Clinicians should be fully aware of these issues whilst treating
patients taking St. John’s Wort.
Refernces:
1. Psychotropic Drug Directory (2003 –04), Stephen Bazire 313-315.
2. Drug Information Handbook 2000, Fuller 958.
3. Szegedi et al. Acute treatment of moderate to severe depression with
St. John’s Wort. BMJ Mar 2005.
Competing interests:
None declared
Competing interests: No competing interests
To the editor
Dear Sir,
The randomized controlled trial by Szegedi et al. (1) probably
promises more than can be expected from Hypericum The authors demonstrate
non-inferiority of Hypericum compared to paroxetine, with a significant
larger decrease in Hamilton (HDRS) score at study endpoint for Hypericum
(difference of 3 points (0.8-5.2)), a nonsignificant increased response
rate (NNT= 10 (4.4-infinity)) and a significant remisson rate (NNT= 7
(3.7-40.3)). We think the design of the study may have contributed to the
findings.
First problem is the absence of a placebo-arm. As found in earlier
Hypericum trials and summarized in a recent meta-analysis of Linde et al.
(2) the superiority of Hypericum against placebo is equivocal, especially
in more precise trials restricted to patients with major depressive
disorder (pooled RR 1.15 (1.02-1.29). The high inclusion criterion of a
Hamilton Depression Rating Scale (HDRS) >=22 in the trial by Szegedi is
reassuring, however a closer inspection of these criteria and table 1
provides doubt as patients with a relatively short duration of current
episodes (>=2 weeks) were included, and the median duration of their
depressive episodes indeed was 4.7 months. Due to the natural course of
major depressive disorder, 63% of patients recover in 6 months (3) in an
unselected population. This and previous findings (4;5) should have urged
the authors to include a placebo-arm, despite their ethical concerns.
Second problem with a potentially negative bias against paroxetine is
the decision to double the dose of paroxetine 2 weeks after the start of
treatment. Although often applied this is an irrational strategy (6), and
this might increase drop-out due to side-effects (7). Maybe for Hypericum
a high dose is warranted, but for paroxetine this may have biased the
results as (more disadvantageous) last observations of those dropped-out
are carried forward in the analyses. Maybe this effect fully explains the
significant difference in remission rates between Hypericum and
paroxetine. Unfortunately the authors did not state the number of drop-
outs to overcome this major concern.
Finally the importance of this trial should be reconsidered with the
recent meta-analysis by Linde et al. (2) in mind. If the results of this
trial are added to their meta-analysis (fig. 5), the RR slightly changes
from 0.98 (0.85-1.12) to 1.03 (0.92-1.15).
All together we think Hypericum extracts should first be tested
against the same standards that are requested for conventional drugs
before their registration by the FDA and the EMAE. Further fixed dose
studies against placebo are warranted, and in future head-to-head
comparisons a placebo-arm is obligatory.
References
1. Szegedi A, Kohnen R, Dienel A, Kieser M. Acute treatment of
moderate to severe depression with hypericum extract WS 5570 (St John's
wort): randomised controlled double blind non-inferiority trial versus
paroxetine. BMJ 2005;330:503-507.
2. Linde K, Berner M, Egger M, Mulrow C. St John's wort for
depression: meta-analysis of randomised controlled trials. Br J Psychiatry
2005; 186:99-107.
3. Spijker J, de Graaf R, Bijl RV, Beekman AT, Ormel J, Nolen WA.
Duration of major depressive episodes in the general population: results
from The Netherlands Mental Health Survey and Incidence Study (NEMESIS).
Br J Psychiatry 2002; 181:208-213.
4. Hypericum Depression Trial Study Group. Effect of Hypericum
perforatum (St John's wort) in major depressive disorder: a randomized
controlled trial. JAMA 2002; 287:1807-1814.
5. Storosum JG, van Zwieten BJ, van den BW, Gersons BP, Broekmans
AW. Suicide risk in placebo-controlled studies of major depression. Am J
Psychiatry 2001; 158:1271-1275.
6. Benkert O, Szegedi A, Wetzel H, Staab HJ, Meister W, Philipp M.
Dose escalation vs. continued doses of paroxetine and maprotiline: a
prospective study in depressed out-patients with inadequate treatment
response. Acta Psychiatr Scand 1997; 95:288-296.
7. Dunner DL, Dunbar GC. Optimal dose regimen for paroxetine. J Clin
Psychiatry 1992; 53 Suppl:21-26.
Competing interests:
None declared
Competing interests: No competing interests
I believe the run-in period used placebo - not active St. John's
Wort. In fact, the paper describes the use of both a St. John's Wort
placebo and a paroxetine placebo during the 3-7 day run-in. I assume the
purpose of the run-in was to exclude early responders that may not, in
fact, have a moderate to severe depression. It would also exclude those
who would be unlikely to be compliant early in the trial, causing an
unnecessary sample size slippage.
The exclusion criterion re: amitriptyline seems to be an effort to
exclude those suffering from a refractory depression that may be unlikely
to respond to either therapy. Results, therefore, not applicable to
refractory depression. I don't quite see how this selects St. John's Wort
responders.
Competing interests:
None declared
Competing interests: No competing interests
Re: Standards of WS 5570 are unacceptable
I believe that the dose of the hyperforin in the tablets was
misunderstood by the author of a previous Rapid Response. It appears from
my reading of the article that the content did not vary from 0.3 to 0.6%,
but rather there were tablets of two potencies (0.3 and 0.6%) so that if
an increase in dose of hypericum were made at two weeks, as described in
the protocol, the change would be blinded (i.e. no change in the number
and size of tablets).
Competing interests:
None declared
Competing interests: No competing interests