Transcript St John’s wort for major depression (Review)

St John’s wort for major
depression (Review)
Linde, Berner & Kriston (2008/9)
Depression


Everyday term but specific diagnostic criteria
for “major” or “clinical” depression – DSM IV
of APA
Drug treatments raise serotonin (5HT) and/or
noradrenalin levels in CNS




Monoamine oxidase (MAO) inhibitors
Tricyclic antidepressants
Selective serotonin re-uptake inhibitors (SSRIs)
High placebo effect – implications?
St John’s wort – Hypericum
perforatum





Flowering plant used for food(?), herbal
medicine and food supplement
Traditionally used as “mood enhancing” & to
treat clinical depression
Hypericin said in 1980s to be MAO inhibitor –
often used to “standardise”
Hyperforin now thought to be key component
– non selective re-uptake inhibitor in CNS?
Studies in UK, Germany, & USA – supplements
often low in active substances (much less than
claimed)
Systematic reviews and metaanalyses



Technique for identifying and combining
inconclusive trials into 1 large trial of greater
statistical power
Finding trials – electronic search engine but
may go further - why?
Publication bias (also get multi-publication)


Positive trials more likely to be published
Detection - simple method is funnel plot
Funnel plots (hypothetical
data)
No Caption Found
Sutton, A J et al. BMJ 2000;320:1574-1577
Copyright ©2000 BMJ Publishing Group Ltd.
Publication bias (real example)


Effect of sponsorship?
Turner et al (2008)



74 antidepressant trials registered with
FDA
38 positive outcome – 37 published
36 negative/? – 14 published but 11 made
to seem positive
Inclusion criteria - typical





Type of subjects
Placebo controlled? Ethical? – alternative?
Dose and duration of trial
Specified outcome measures
“Quality” – e.g. Jadad score (0→5)



Double blinding (0→2)
Random allocation (0→2)
Dropouts recorded & explained (0→1)
Inclusion criteria – Linde et al





Double blind randomised trials
Adult subjects meeting DSM-IV criteria for
major depression
SJW versus placebo or SJW vs synthetic
antidepressant (4 weeks minimum and
adequate dose of “recognised” antidepressant
drug)
Outcome assessed clinically using depression
scales or symptoms
Main safety assessment – dropouts (sideeffects)
Searching




Several electronic databases searched
Bibliographies of articles checked
Direct contact manufacturers, experts
and authors
18 eligible placebo trials & 17 vs
synthetic drug (total 5500 patients)
Outcome measure(s)



Responder rate ratios (RR) = %
responders in SJW group/% in placebo
group (or synthetic drug group) and
95% CI (i.e. RR<1, SJW better)
Data presented as a series of “Forest
plots”
Funnel plots – not significantly
asymmetric
SJW vs placebo (by precision)
Comments

Less precise trials give bigger overall
effect:





Less RR 1.87 (1.22→2.87)
More RR 1.28 (1.10→1.49)
Total RR 1.48 (1.23→1.77)
Significant heterogeneity in both groups
but greater in less precise
A typical finding from a meta-analysis
SJW vs placebo (by country)
Comments

Much larger effect in trials from German
speaking countries:



RR 1.78 (1.42→2.25) German
RR 1.07 (0.88→1.31) R.O.W. (non sig)
German trials significant heterogeneity, less
and non sig. in ROW
SJW vs drug - older/SSRI
SJW vs drug by country
Comments



No significant differences in any 4 splits
or overall
No significant heterogeneity
Studies from German speaking
countries still significantly favour
hypericum
Authors’ conclusions
(Germans)



5 “German” studies pre 2000 all other 2000+
Could not rule out “some smaller studies from
German speaking countries flawed and
reported overoptimistic results”
Geoff’s comment – outcome measures seem
to be insensitive and to show poor
discrimination
Hypericin and hyperforin