Depression is effectively treated with a number of antidepressants. However, all of these drugs are associated with specific adverse effects that can range from mildly unpleasant or mild discomfort to unbearable distress and even life-threatening conditions. For some patients, the treatment may seem worse than the disease, and this can significantly affect the quality of life and the course of treatment. Results from a study of the herb Hypericum perforatum indicate that it may be as effective as conventional antidepressants for treating mild to moderate depressive disorders, but with a lower incidence of adverse effects. There have been no attempts to compare the general safety of hypericum with that of conventional antidepressants. Safety is critical in risk-benefit or cost-benefit analysis. This review presents an attempt to systematically compare the existing safety data for Hypericum perforatum with several conventional antidepressants.
Comparison of antidepressants
Choice of antidepressants for comparison
The purpose of this review was to compare St. John’s wort with different types of conventional antidepressants. For this, typical drugs were selected from the following classes of substances: tricyclic antidepressants (TCAs); selective serotonin reuptake inhibitors (5-hydroxytryptamine; 5-HT) (SSRIs); monoamine oxidase inhibitors (MAOIs) and new types of antidepressants. The popularity of the prescription served as a criterion for drug selection. Data on the number of appointments made during 1996 were obtained from the Department of Health, England. Dotyepine was chosen as a typical TCA because it was the most frequently prescribed of all antidepressants (3,266,900 prescriptions made in England in 1996). Fluoxetine was the most commonly prescribed SSRI (2,514,000), while moclobemide was the most commonly prescribed MAOI (47,800). Mirtazapine, a new therapeutic agent introduced in England following the publication of these prescriptions by the Ministry of Health of England, was selected to represent a new class of antidepressants: noradrenergic and serotonergic specific antidepressants (NaSSA).
Table 1. Percentage of patients reporting specific PEL in clinical trials that included antidepressants
| PEL | St. John’s wort | Fluoxetine | Moclobemide | Mirtazapine | Dotyepin | ||||
| V plac | V drug | V plac | V drug | V plac | V drug | V plac | V drug | all studies | |
| (n = 408) | (n = 243) | (n = 1322) | (n = 781) | (n = 834) | (n = 1291) | (n = 359) | (n = 463) | (n = 13834) | |
| Nausea | 0.2 | 22.9 | 24.1 | 7.2 | 8.6 | 3.0 | 6.0 | ||
| Diarrhea | 11.1 | 9.2 | 4.1 | 3.3 | 1.5 | ||||
| Constipation | 5.3 | 9.3 | 6.4 | 13.0 | 16.0 | 4.2 | |||
| Gastroenteric symptoms | 0.2 | 8.5 | 6.2 | 8.2 | 6.0 | 1.6 | |||
| Insomnia | 0.2 | 0.9 | 15.5 | 13.8 | 7.2 | 9.8 | 10.0 | 8.0 | 2.2 |
| Drowsiness | 0.2 | 13.2 | 19.2 | 5.4 | 23.0 | 20.0 | 10.5 | ||
| Sedation | 19.0 | 13.0 | |||||||
| Fatigue | 0.2 | 4.6 | 16.0 | 0.7 | |||||
| Anxiety | 2.6 | 5.0 | 1,2 | ||||||
| Agitating | 12.2 | 3.6 | 9.0 | 6.0 | |||||
| Nervousness | 14.5 | 15.0 | |||||||
| Anxiety | 11.6 | 10.5 | 2.5 | 1.9 | |||||
| Dizziness | 4.5 | 9.1 | 12.2 | 7.9 | 7.8 | 6.0 | 8.0 | 4.9 | |
| Headache | 1.7 | 15.4 | 18.6 | 11.3 | 13.1 | 5.0 | 10.0 | 1.9 | |
| Photophobia | 0.6 | ||||||||
| Skin reaction | 0.7 | 0.9 | 0.2 | ||||||
| Tremor | 0.6 | 10.4 | 13.2 | 5.4 | 6.0 | 4.7 | |||
| Dry mouth | 4.0 | 11.0 | 19.7 | 10.3 | 16.6 | 25.0 | 34.0 | 15.9 | |
| Visual disturbance | 6,3 | 2.6 | 3.9 | 7.0 | 2.9 | ||||
| Sweating | 9.1 | 10.5 | 6.5 | 5.0 | 2.6 | ||||
| Anorexia | 11.1 | 7.2 | 1.0 | 0.3 | |||||
| Body weight gain | 0.1 | 10.0 | 11.0 | 1.4 | |||||
| Reducing weight | 1.1 | 2.0 | |||||||
| Note. Here and in table. 2: V drug versus antidepressant; V plac versus placebo. | |||||||||
Table 2. Percentage of Patients Dropout Due to PEL (ADRs)
| St. John’s wort | Fluoxetine | Moclobemide | Mirtazapine | Dotyepin | ||
| V plac (n = 408) | V drug (n = 243) | V plac (n = 1322) | V drug (n = 781) | all studies (n = 1235) | all studies (n = 822) |
all studies (n = 13834) |
| 0,4 | 0.8 | 14.1 | 16.6 | 2-13 | 4.9 | 3.6 |
Sources of safety data The
information on side effects of drugs (PEL) reported in clinical trials was collected from 2 sources. The first is Medline’s search for meta-analyzes and reviews that provided PEL data for selected antidepressants. The second is contacting relevant manufacturers and requesting information about clinical trials.
Clinical Trials Data
Medline’s search and contact with drug manufacturers produced a proposed clinical trial program report for each antidepressant, which reflected data on PEL reported by patients in the clinical trials. This information is presented in table. 1.
Hypericum had the lowest percentage of different PELs reported by fewer patients. The number of reports for most PEL was very low for dotiepine and moclobemide, and the highest for fluoxetine and mirtazapine.
Table 2 shows the proportion of patients withdrew from clinical trials due to PEL. For hypericum, the data are lower than for other drugs.
Interpreting the data
Evidence for PEL in clinical trials is difficult to interpret. Because trials typically involve a relatively small number of patients over a short period of time, they are unlikely to find rare or delayed adverse effects. While they should be theoretically useful for identifying the scope of frequent, predictable PEL, the method of data collection and the way the data was obtained tend to be inconsistent across trials. Tests that are primarily aimed at establishing efficacy may not be rigorous in identifying the PEL participants. Therefore, results may reflect varying dosages, sample sizes and methods larger than they provide accurate data on PEL. In addition, tests were carried out in several different countries; for example, most trials of St. John’s wort were conducted in Germany, while most trials of conventional antidepressants were conducted in other European countries or the United States.
Medicinal monitoring trials
The safety data for hypericum are obtained from monitoring trials in Germany. In the largest trial of 3250 patients followed up for 4 weeks of hypericum therapy, PEL was reported by 79 (2.4%) patients and 48 (1.5%) patients withdrew from the trial. In another trial (n = 1060), PEL was tested in 21 (2%) patients and there were 8 (0.8%) exits due to PEL. In two smaller trials of 114 and 163 patients, the number of PEL patients was 7 (6.1%) and 9 (5.5%), respectively. In all of these trials, the most commonly reported PELs were gastrointestinal symptoms, allergic reactions, fatigue, and anxiety.
PEL overdose toxicity reported by patients taking antidepressants at therapeutic dosages provides some indication of their relative tolerability. However, it is their toxicity in overdose that is important to assess the overall safety.
There is no evidence that hypericum is dangerous in overdose. The risk of photosensitization is believed to be associated with high doses and is observed in white-skinned animals consuming large quantities of plants in pastures in bright sunlight. There were only 2 cases of a photosensitive reaction in humans that were associated with taking hypericum, and the symptoms disappeared after the drug was discontinued. It has been estimated that doses of hypericum 30 to 50 times the recommended daily dose will be required to develop serious phototoxic reactions in humans, and that serious complications can be avoided by protecting the patient from ultraviolet radiation for a week.
Fluoxetine, moclobemide, and mirtazapine appear to be acceptably safe in overdose. Both fluoxetine and moclobemide were classified as “safe” according to the lethality criterion: if a dose 14 times the adequate daily dose (2 weeks supply) does not result in a life-threatening situation in a healthy adult. In some cases of overdoses of mirtazapine and moclobemide, no fatal consequences have been reported to date, all symptoms disappeared after the drug was discontinued.
Dotyepine is much less safe to overdose. Symptoms may include tachycardia, hallucinations, mydriasis, coma, convulsions, hypotension, and cardiac arrhythmias.
Drug interactions
There are no known interactions between hypericum and other drugs.
Most of the clinical trials designed to test the efficacy of hypericum have excluded patients who are already taking psychotropic or psychoactive drugs, so that interactions between them and hypericum have not been tested. However, the use of other drugs (eg, for hypertension, circulatory disorders, bronchial asthma, and menopausal symptoms) has been resolved in many trials of hypericum, and there has been no evidence of any interaction between these drugs and St. John’s wort.
All other antidepressants included in this analysis had interactions with other drugs. A rare but potentially life-threatening serotonin syndrome resulting from overstimulation of central serotonin receptors is a risk of SSRI drugs when used in combination with other antidepressants. The long half-life of fluoxetine, in particular, means that interactions can take place for some time after its withdrawal. Medicines that increase the levels of norepinephrine at cell receptors can be dangerous if taken with TCAs. Although moclobemide does not carry the same risk of hypertensive crisis as older MAOIs, patients are still advised to avoid consuming large amounts of tyramine-rich foods and using sympathomimetic drugs such as ephedrine, pseudoephedrine, and phenylpropanolamine.
Conclusion
Based on the data currently available, hypericum appears to be at least as safe as all conventional antidepressants studied here, or possibly safer. The proportion of patients experiencing PEL due to hypericum in clinical trials was small and PEL was moderate. In addition, ad-hoc regimens have received few reports of PEL and drug monitoring trials show low levels of moderate PEL in patients receiving hypericum. However, there is currently uncertainty about the effects of hypericum in overdose and about its interactions with other drugs, and there is also uncertainty about the mechanisms of its action. New comparative trials of hypericum and conventional antidepressants will provide additional useful information about their relative tolerability.