CORRESPONDENCE

more explicit attention should have any synthetic antidepressant, from combinations, but do not constitute an
been paid to the risk that patients on imipramine (Tofranil) to fluoxetine embargo against any one drug used
phenprocoumon or digoxin should not (Prozac), with the exception of those singly, or in combination with others
injudiciously discontinue St John’s classed as MAO inhibitors. In one with which there are no such
wort by themselves, as this could lead review on the subject,2 Nathan points interactions. So, the clinician must
to toxic reactions in stabilised patients. out that: “It [St John’s wort] is the balance the benefits of pharma-
The interaction study with digoxin only antidepressant capable of cological interventions against any
did not confirm induction of hepatic inhibiting the re-uptake of 5-HT possible harm that they may cause,
cytochrome P450 by St John’s wort, [serotonin], NA [norepinephrine], and remote though such a likelihood may
but suggested induction of the drug DA [dopamine] with similar potencies. be. For example, in untreated resistant
transporter P-glycoprotein, which The potencies for monoamine depression, there is a high probability
mediates the intestinal absorption, reinhibition and chronic changes in of fatal outcome and it is accepted
distribution, and renal excretion of receptors are also consistent with practice to use combinations of
digoxin.3 The full report of this study changes seen with known anti- different antidepressants, though there
also mentions unpublished data that St depressants.” Cott 3 found that is a risk of interactions between them.
John’s wort can lower the plasma hypericin (a main component of the St David Wheatley is an independent consultant
concentrations of amitriptyline, John’s wort extract: “lacked significant to Lichtwer Pharma UK.
thereby providing additional evidence MAO A or MAO B inhibition at David Wheatley
that some St John’s wort interactions concentrations up to 10 ␮M.” Other Psychopharmacology Research Group,
might involve induction of cytochrome studies on the whole extract have only 10 Harley St, London W1N 1AA, UK
P450.4 shown very weak MAO inhibition,
More research into these which lead Muller and colleagues4 to 1 Ernst E. Second thoughts about safety of
St John’s wort. Lancet 1999; 354: 2014–15.
mechanisms is important for two conclude that St John’s wort does not
2 Nathan PJ. The experimental and clinical
reasons. First, because it will help to have any antidepressant activity in pharmacology of St Johns Wort (Hypericum
predict which additional drugs are human beings. perforatum L). Mol Psychiatry 1999; 4:
prone to pharmacokinetic interactions However, the similarity in 333–38.
neurotransmitter activity between St 3 Cott JM. In vitro receptor binding and
with St John’s wort. Second, because it
enzyme inhibition by Hypericum perforatum
will show whether the antidepressant John’s wort and other synthetic extract. Pharmacopsychiatry 1997; 30:
properties and interactions of St John’s antidepressant drugs, is consistent with 408–12.
wort reside in the same components. the well-established therapeutic 4 Muller WE, Rolli M, Schafer C, Hafner U.
Since hyperforin is increasingly response pattern that is observed in Effects of hypericum extract (LI 160) in
practice. For instance, there is a lag- biochemical models of antidepressant
recognised as a major active principle activity. Pharmacopsychiatry 1997; 30:
of St John’s wort, 5 it would be period of 3–4 weeks before any 102–07.
particularly interesting to test the observable clinical effect becomes 5 Markowitz JS, De Vane CL, Boulton DW,
interaction potential of this established and there is a necessity to Carson SW, Nahas Z, Risch SC. Effect of
complete a full course of treatment of St John’s Wort (Hypericum perforatum) on
constituent. cytochrome P-450 2D6 and 3A4 activity in
at least 6 months’ duration, if relapse is healthy volunteers. Life Sci (in press).
*Peter A G M De Smet, Daan J Touw
to be avoided. In achieving these two
*Scientific Institute Dutch Pharmacists,
Postbus 30460, 2500 GL, The Hague, objectives, St John’s wort provides a
Netherlands; and University Hospital Vrije distinct advantage in patient co- Sir—In Sweden, products containing
Universiteit, Amsterdam, Netherlands operation, which occurs because there extracts of St John’s wort 1 are
is almost a complete lack of any marketed as natural remedies and not
1 Ernst E. Second thoughts about safety of
appreciable side-effects. as food supplements and the Medical
St John’s wort. Lancet 1999; 354: 2014–16.
2 Bon S, Hartmann K, Kuhn M.
But it is timely that Ernst should Products Agency (MPA) is responsible
Johanniskraut: Ein Enzyminduktor? Schweiz query the potential of St John’s wort to for the approval of these products. St
Ap Ztg 1999; 16: 535–36. cause drug interactions as a result of John’s wort is contained as the sole
3 Johne A, Brockmöller J, Bauer S, Maurer A, effects on the hepatic cytochrome ingredient or in combination with
Langheinrich M, Roots I. Pharmacokinetic P450 system. These effects need to be
interaction of digoxin with an herbal extract
other active ingredients in several
from St John’s wort (Hypericum perforatum). further elucidated and put into products approved in Sweden as
Clin Pharmacol Ther 1999; 66: 338–45. perspective by the benefits that St natural remedies. Slight mood
4 Venkatakrishnan K, Greenblatt DJ, John’s Wort can confer on depressed lowering has been deemed an
Von Moltke LL, Schmider J, Harmatz JS, individuals. The benign side-effect appropriate indication for self-
Shader RI. Five distinct human ytochromes
profile of St John’s wort should not be medication which is a prerequisite for
mediate amitriptyline N-demethylation in
vitro: dominance of CYP 2C19 and 3A4. seen as a screen that hides perils, and it natural remedies.
J Clin Pharmacol 1998; 38: 112–21. is encouraging that studies on the Since 1998 the MPA has received
5 Chatterjee SS, Bhattacharya SK, effects of St John’s wort on the P450 seven case reports of a reduced
Wonnemann M, Singer A, Muller WE. system seem to confirm this. anticoagulant effect of warfarin—ie, a
Hyperforin as a possible antidepressant
component of hypericum extracts. Life Sci
Markowitz and co-workers5 studied decreased International Normalised
1998; 63: 499–510. seven normal adults over 4 days and Ratio (INR)—associated with
used dextromethorphan and concomitant use of St John’s wort
alprazolam as substrate probes. They (table). Most patients were stable on
Sir—The commentary by E Ernst1 concluded that their results: “suggest warfarin for some time before St John’s
raises important points concerning the that St John’s wort, when taken at wort was started. Although none of the
clinical use of St John’s wort in the recommended doses for depression, is patients developed thromboembolic
treatment of depression, while unlikely to inhibit CYP2D6 or complications, the decrease in INR
perpetuating a myth concerning its CYP3A4 activity”. Most drugs listed was thought to be clinically significant.
mode of action. St John’s wort does in the pharmacopia carry drug- The INR returned to target values
not act by inhibition of monoamine interaction warnings, which are very either after the warfarin dose was
oxidase (MAO), any more than does necessary to avoid dangerous increased or St John’s wort was

576 THE LANCET • Vol 355 • February 12, 2000

Patients INR before INR during Duration of Warfarin dose once St John’s wort asymptomatic until visual-field loss is
warfarin treatment started advanced enough to effect the quality of
treatment life of the patient. Risk factors for
79 years, female 2·5–3·8 1·7 2·5 years Increased from 18·5 to 21·25 mg/week steroid-induced glaucoma include
65 years, male 2·4–3·6 2·0–2·1 4 years Increased from 37·5 to 40 mg/week. INR patients with POAG, a family history of
returned to 2·7 on the ordinary dose of 37·5
mg/week once St John’s wort was stopped. POAG, and myopia. About 5% of the
76 years, male 2·3 1·1 10 days Increased. general population will have a change in
61 years, female NA 1·2 Years. INR decreased and then increased after intraocular pressure of greater than 15
St John’s wort was stopped.
mm Hg when challenged with
84 years, female 2·9–3·6 1·5 Stable 肁6 months. Stable for ⭓6 months. INR returned to
target range once St John’s wort stopped. corticosteroids. However, 90% of
56 years, female 2·6 1·5 Unknown. INR returned to previous value after St John’s patients with POAG will have a raised
wort stopped. intraocular pressure when challenged
85 years, male 2·1–4·1 1·5 Long time. INR decreased so dose increased while
with corticosteroids. Rarely, steroid-
St John’s wort was continued.
induced glaucoma can present as an
NA=not available.
acute attack with intraocular pressure in
Seven cases of decreased effect of warfarin during concomitant treatment with the range of 50–70 mm Hg. Steroid-
St John’s wort induced glaucoma is associated with
withdrawn. The reduced effect of products in Sweden and requested topical, periocular, and systemic routes
warfarin suggests an induction of studies on the extent and implications of corticosteroid administration. There
cytochrome P450 2C9.2 of drug-drug interactions. In the is a wide range in the time course of
In addition, the MPA has received meantime, the companies have been steroid-induced glaucoma with the
eight reports of intermenstrual requested to add on the packaging and onset ranging from hours to weeks and
bleedings and one report of changed patient information leaflet that St duration lasting days to years after
menstrual bleeding from manu- John’s wort products should not be cessation of steroids. The raised
facturers of St John’s wort products. used concomitantly with any medicinal intraocular pressure is a result of a
Most reports concerned women aged product. decrease in outflow of aqueous humor.
23–31 years who had been on oral *Qun-Ying Yue, Carin Bergquist, The treatment of steroid-induced
contraceptives for a long time. Time Barbro Gerdén glaucoma includes the discontinuation
between starting St John’s wort and Medical Products Agency, S-75103 Uppsala, of steroids when possible in addition to
onset of menstrual disorders varied, Sweden the medical and surgical options
and was about 1 week in five of the associated with the treatment of POAG.
patients. There was recovery after St 1 Ernst E. Second thoughts about safety of It is impossible to predict which patients
St John’s wort. Lancet 1999; 354: 2014–16. will develop steroid-induced glaucoma,
John’s wort was stopped in three
2 Kaminsky LS, Zhang ZY. Human P450
patients, for whom outcome was metabolism of warfarin. Pharmacol Ther
thus physicians must be aware of this
known. This suggests an induction of 1997; 73: 67–74. iatrogenic condition and communicate
CYP3A4, which is involved in the 3 LeBel M, Masson E, Guilbert E, et al. with the patient and ophthalmologist
metabolism of steroids. Decreased Effects of rifabutin and rifampicin on the with regards to steroid exposure and the
pharmacokinetics of ethinylestradiol and
steroid concentrations and break- norethindrone. J Clin Pharmacol 1999; 38:
risk of blindness associated with this
through bleedings have been reported 1042–50. form of glaucoma.
when oral contraceptives are combined 4 Kerb R, Bauer S, Brockmöller J, Roots I. *Maria del Pilar Bernal, Kathrine Loftfield,
with known enzme inducers. 3 Urinary 6-␤-hydroxycortisol excretion rate
Jonathan D Nussdorf
Induction of CYP3A4 by St John’s is affected by treatment with hypericum
extract. Eur J Clin Pharmacol 1997; 52: Department of Ophthalmology, Ochsner Clinic,
wort is also supported by phar- A186 (abstr). New Orleans, Louisiana 70121, USA
macokinetic data. Kerb and 5 Johne A, Brockmöller J, Bauer S, Maurer A,
colleagues 4 found an increased Langheinrich M, Roots I. Pharmacokinetic 1 Fabre-Guillevin E, Tchen N,
6-␤-hydroxycortisol excretion after interaction of digoxin with an herbal extract Anibali-Charpiat M-F, Calluaud L, Ravaud
from St John’s wort (Hypericum perforatum). A. Taxane-induced glaucoma. Lancet 1999;
repeated doses of St John’s wort and Clin Pharmacol Ther 1999; 66: 338–45. 354: 1181–82.
in a study on dextromethorphan 2 Skuta GL, Morgan RK. Corticosteroid-
presented at the New Clinical Drug induced glaucoma. In Ritch R, Sheilds MB,
Evaluation (NCDEV), 39th Annual Krupin T, eds. The Glaucomas. Volume 2.
Taxane-induced glaucoma St Louis Mosby 1996: 1177–88.
Meeting, 1999, June 1–4, Boca Raton,
Florida, USA, CYP3A4 activity was Sir—Fabre-Guillevin and colleagues Author’s reply
found to increase when patients were (Oct 2, p 1181)1 describe taxane-
also taking St John’s wort. It is worth induced glaucoma in a young patient Sir—I totally agree with the comments
pointing out that these are in-vivo and with metastatic breast cancer who made by Maria del Pilar Bernal and
not in-vitro studies, as stated in the underwent chemotherapy of docetaxel colleagues that the history of glaucoma
commentary by E Ernst.1 in combination with both low-dose and could have been induced by
The reported cases of interactions high-pulse corticosteroid treatment. corticosteroids, which was the main
together with pharmacokinetic data This patient developed symptomatic initial diagnosis. But, the recurrence of
provide a strong indication that St vision loss during the course of this episode when the patient was taking
John’s wort is an inducer of a broad treatment and had high intraocular taxanes and the fact that she had never
range of drug-metabolising enzymes. pressure, optic nerve cupping, and had such an episode when taking low or
The study by Johne and colleagues, bilateral visual-field loss. We argue that repetitive high-dose corticosteroids
which shows decreased concentrations this clinical scenario is consistent with outside chemotherapy, prompted us to
of digoxin, suggests that St John’s wort streroid-induced glaucoma. decide that taxane-induced glaucoma
might also induce the transport protein We think that steroid-induced was the most probable diagnosis.
P-glycoprotein.5 glaucoma is underdiagnosed because it Alain Ravaud
The MPA has contacted the is similar to primary open-angle Institut Bergonié, Regional Cancer Center,
companies marketing St John’s wort glaucoma (POAG) in that it is 33076 Bordeaux Cedex, France

THE LANCET • Vol 355 • February 12, 2000 577