1,2) Clinical psychologists, private clinic, Aarhus (Denmark), ²) Lecturer, University of Aarhus

[Abstract]

[Key words: Hypericum (St. John’s Wort) – depression – psychotherapy]

Hypericum perforatum L. [syn. Hypericon perf. Lam.] is a low-growing herb with yellow flowers, called St. John’s Wort in English (Millepertuis Fr., Johanniskraut Ger., Prikbladet Perikon Dan.) which is common throughout Europe, including Scandinavia, and also found on other continents. Its properties as a medicinal plant have been known since antiquity, and its use goes back for centuries and probably even millenia. In fact, it was recommended by Hippocrates (5^th century BC) to treat “nervous restlessness”, since when it has been used in folk medicine, amongst many other cases, for anxiety, disturbed sleep, and melancholy.

Nowadays, only an extract of the petals is made (e.g. for St. John’s Wort wine or liqueur) though formerly, in Denmark and elsewhere, a decoction was prepared from the whole plant. Almost two centuries ago [and earlier: comp. J.C.B. Bendt: Anvisning til at indsamle, tørre og conservere…medicinske planter og plantedele (Directions on How to Gather, Dry and Preserve…Medicinal Plants and Their Parts). Copenhagen, 1810], the plant was called “Thought Grass” (Tankegræs) and it was assumed, as the name implies, that it was able to counteract melancholy and improve the capacity for thought and action in the person taking it. Another name for it was “Horsebane” (Hestemare) due to the fact that horses in pastures where a lot of Hypericum grew became restless and unmanageable. The herb is nowadays sold in some countries in the form of tablets or capsules containing an extract (via alcohol) of the plant’s active substances. We know that hypericin is one of the most important active ingredients, and a number of investigations have shown that it inhibits the uptake of serotonin by the synapses in specific cerebral nerve cells — in the same way as synthetic SRI (Serotonin Reuptake Inhibitor) in antidepressant preparations. Another possible mode of action is an inhibition of MAO (monoamine oxidase) enzymes which ensure the catabolism of specific amines (such as noradrenaline). A number of products containing hypericum are found on the market, some of them together with other natural substances, such as valerian etc. In Denmark, single preparations are indicated as being natural medicines for the treatment of “dejectedness and melancholy”, though not of “depression”. For one reason or the other, the recommended daily dose for these products is below two thirds of the dose that a number of studies have demonstrated as having a good efficacy without any side-effects (900mg), though studies are also available showing that a higher dose is able to act against serious depressions. In comparitive clinical investigations, hypericum in particular has been compared with imipramine, showing effects which are apparently equal to those of such a widely recognized antidepressant, in addition to having fewer side-effects. Only just recently, a comparative investigation has been started in the context of a new generation of antidepressants (“lykkepiller”).

Klaus Linde and coworkers have conducted a survey of all studies on the effects (or side-effects) of hypericum published in medical literature. The studies chosen included the scientifically most valid investigations according to more closely defined criteria. Their results were published in the British Medical Journal in 1996 and were based on data from 1008 participants. The conclusion was reached that hypericum clearly has a greater efficacy than placebo where mild to moderate depressions were involved. No or only slight side-effects were recorded in addition to the fact that hypericum was just as effective if not more so than currently available synthetic substances.

Effect (I). The principal aim of the study was to investigate the effect of Hypericum perforatum L. on depression and depressive conditions in persons not or only at certain times meeting the formal requirements of depression. Secondary effects (II). These included side-effects or possible adverse drug reactions. Differences in effect (III). On the tertiary level, we were interested in learning something about the relationship between individual problem cases (personality structures) and the effect of hypericum. We defined our criteria as a) Are there any specific cases in which hypericum is clearly contraindicated? Or b) can we, conversely, make any statements about persons receiving low or high doses of hypericum? c) As regards its mode of action, does it act on different conditions in different ways, and within different time limits? Furthermore, we had an idea that would not see any particular placebo effect (in contrast with former active substance versus placebo trials) as the participants had been receiving previous treatment from the trial investigators themselves, a fact which ought to reduce the ”contact/interest” type of placebo effect at least.

A total of 26 persons — all clients in discussion therapy (for those concerned also/only in group therapy) — aged between 24 and 37 years took part in the trial. Of these, 4 dropped out: two (placebo and active medication) because they lost their data (subjective report forms (TEPRA) described below), one (active) decided to discontinue after a few days and withdrew due to anxieties (placebo) after several weeks. We were thus working with a small group of relatively young test persons who had already been receiving therapy prior to starting the study. Exclusion criteria were, first and foremost: suicide risk, psychotic traits and in particular paranoid features, and ongoing medication. Inclusion criteria were: depression or depressive elements in the clinical picture, even if these were not sufficient to fulfil the diagnosis of ”(major) depression” as defined on the basis of recognized diagnostic systems. Our rationale for including nearly all of the unmet exclusion criteria was because the great majority of those presenting themselves at a clinic, whether with manifest or hidden depressive elements, are admitted as the latter type of case always manifests itself during the course of further treatment. In the final instance, we are generally dealing with the problem of deficiency or loss, even although we are confronted with a varied collection of clinical pictures, i.e.: depression, narcissistic disorders, boulimia, schizoid personality disorders, conditions of anxiety and histrionic disorders. A few decades ago, the corresponding indications would have been: depression, neurotic depression, various neuroses and perhaps ”borderline psychosis”.

Many studies performed abroad over the last 10 to 20 years (Germany, Spain and the USA in recent years, among others) may all together be called convincing in that they are able to demonstrate the antidepressant effect of hypericum. In addition, a considerable number of studies suggesting the neurophysiological modes of action (the active ingredients of hypericum probably act via the transmitters noradrenaline and serotonin, though other simultaneous paths are also possible) [ref.]. The present study is thus the first in which its participants had been receiving previous, non-medical psychotherapy and its investigators, also therapists, were able to maintain a thorough and continuous investigation of the patients before, during and even after the study. Theoretically, this was expected to limit the placebo effect, insofar as we may assume that part of it has to do with the attention the participants receive from the investigators, and insofar as they are already involved in a process where they are in contact with a number of emotions in a controlled way. This last factor means that the participants were trained to be both observant and critical towards their own emotions and moods, whereby the self-suggestive element is presumedly reduced.

AM = active medication

We selected what is called the ”observation group” or the ”control group as investigative group” design, i.e. the control group, in this case also the placebo group, received the active substance (verum) at one point in time.

Prior to starting the study, all groups were informed about its aim, and that their participation was purely on a voluntary basis, strict anonymity being maintained towards all persons outside the clinic. Randomisation was carried out by persons other than the trial investigators (in this case by members of the staff at Jemo-Pharm), neither party knowing who received active medication and who not. The verum substance contained hypericum extract (Modigen, manuf. By Jemo-Pharm A/S, Stege Danmark), standardised as regards it content of total hypericin (0.3%) and containing a total 2.5% hyperforins. 300 mg extract was filled into hard gelatine capsules. The placebo consisted of 300 mg coloured microcrystalline cellulose filled in the same way into hard gelatine capsules. The placebo was identical with the active medication, both when the capsules were sealed and when they were opened.

The actively medicated (verum) group was given 300mg x 3 capsules standardised hypericum extract, to be used over 14 days each time, with 900mg per day for 5 times i.e. all together for 10 weeks, and the reference group was given placebo for 6 weeks, followed by active medication for 4 weeks. The leading investigators saw the participants f at least every 14^th day, often combined with a therapy session which, however, always took place after the study-related evaluation.

The most important rationale comprises what is understood or ought to be understood as ”depression” or being ”depressive”. We hereby understand an increase in specific unpleasant conditions, in this study investigated under the categories A: anxiety and restlessness, S: self-dissatisfaction or self-hatred, G: anger (hatred) towards others, H: dejectedness or a feeling of being upset, and T: a feeling of being under compulsion due to external circumstances – or a sensation of being abandoned and empty. This last category per se refers to an ambivalent condition and may in short be described as: an absence of pleasurable motives to act although, in the study, it should only be understood as being a contribution to a sensation of aversion or dislike. If, from these parameters, we construe a total ”depression parameter” we are for the time being in a position to evaluate developments in the actively medicated (verum) versus the placebo group in the context of this one parameter. As we made use of an ”observation group” design, We were also able to evaluate 4 weeks of verum in both groups (verum4group). Finally, though not least in importance, seen from a clinical psychological perspective, we were able to study the effects of hypericum in the individual person or, inversely, to study the individual client with the help of hypericum, provided we have some ideas about its general effect. In order to do this, we have to look at the change in the individual category or parameter through time in each client/participant. This retrograde aspect in itself meets the aim of the study, i.e. to discover whether hypericum has a beneficial effect on depression and depressive conditions, and also to discover whether adverse reactions occur. On the other hand, what applies to both as regards the overall aim of the study and is of psychological interest is: how do the various conditions (categories) change in the population as a whole over a specific period of time? To the extent that hypericum does have a favourable effect, measured via depression parameters, is it first and foremost a reduction in anxiety level and dejectedness we are concerned with, or is it either the feeling of being under compulsion or the sensation of emptiness which subside first? It is conceivable that certain parameters are subject to an initial increase before subsiding to a lower level, such as the sensation of anxiety or anger (as experienced in therapy), as well as there being a phenomenologically functional necessity for a general improvement.

The patients’ subjective reporting was conducted using TEPRA (Temporal Psychological Recording and Analysis). Every evening for 10 weeks the participant was to mark a point on a scale from 0 to 4 (raw scores) in 15 categories, as well as for a number of extra categories (physical symptoms, sleep disturbances, sexual problems etc.). Only 5 of all these categories are used in the depression parameter, though a number of the unused categories may be regarded as reflecting these 5.For example, the category ”content with one’s own performance” is reciprocal with ”self-contentedness” (S), thus providing us with the possibility of cross-checking the ”internal coherence” factor in the individual subject’s data. If the value 0 is given this should correspond to the absence of, for example, anxiety or restlessness, so that high values correspond to a very high level of anxiety. Use of the scale was to be individual and was deliberately not guided or objectivized in relation to the ”subjective calibration” (quantification) as completed by the test person himself/herself. These were the individual variations in the single categories we were interested in, requiring a quantification not agreeing with a desire to objectify subjective conditions in an empirically objective manner of understanding. The amount of data collected in the form of raw scores was close on 7000 for the 5 categories selected, in spite of the limited number of persons in the trial. To make data processing easier, the following procedure was used to transform raw scores into scores: for each period of 14 days, a score is defined as the rounded-off sum of two days’ r-score from the first week, and two days’ r-score from the second week (days 4 and 5). The individual scores for each category and for each 14 day period, as the material for statistical processing amounts to, are thus within an interval of 0-16. The depression index (Di) is accordingly defined as the sum: A+G+H+T, i.e. the sum of scores for the 5 categories, and is situated in a scale from 0 to 80. In the case of active medication (verum) the same procedure is used for each week.

A one-tailed t-test was used for statistical processing. It is one-tailed, because our scale was broken down into intervals from no effect to defined effect. The rationale behind this was that none of the many investigations undertaken up to now have demonstrated negative effects worth any consideration (i.e. that patients suffer more from depression). Based on the wide range of inclusion criteria among other factors, we may assume that the population has a normal distribution and, as a population average for the Di (depression index) we used the mean value for both groups up to one week into the study, i.e. Diμ = (A+S+G+H+T)μ = 34.2. The reason why we did not include day 1 was because, in our experience, it takes the clients a few days to acquaint themselves with the scale (subjective calibration), even if this decision meant that the population average, perhaps a little lower in reality, in turn meaning that it is somewhat more difficult to reject H₀. The level of significance is 0.05, which means rejecting H₀ (no significant effect) for t > 1.84. N(Vgr)=10, N(Pgr)=12, N([V+P]gr)=22.

A comparison of the two groups (Vgr and Pgr) already shows a clear difference after three weeks, which is marked, though not significant as we had assumed prior to the study.

In the case of Vgr, Di<30 is the expression for a significantly positive change (S). After week 4, H₀ was discarded, in other words verum has a positive effect. In the case of Pgr, a significant level of effect is approached in weeks 8-10. Looking at the verum4gr, i.e. the ”virtual” (N=22) group consisting of Vgr for the first 4 and of Pgr for the last 4 weeks, we already obtain a significant result in week 4. This is naturally due to a higher population (N=22) and, consequently, a rejection of H₀ for a t-value at 1.73 (temp. Vgr, N=10: 1.84).

Effect (I). As regards the primary aim of the study, we found that hypericum has a positive effect on depressive conditions, as defined in the categories: A anxiety and restlessness, S self-dissatisfaction or self-hatred, G anger (hatred) towards others, H dejectedness or a feeling of being upset, and T a feeling of being under compulsion due to external circumstances – or a sensation of being abandoned and empty. In contrast to a number of other studies, our approach in this investigation has been phenomenological, i.e. a subjective assessment of how the participants themselves experienced treatment was what mattered — although, in this case, a clinical assessment aiming at objectivity combined with an evaluation of the clients’ level of performance — fully agreed with the results.

Secondary effects (II). As regards the secondary aim of the study, i.e. as to whether side-effects occur, it is generally possible to say that no serious events occurred during treatment. A few participants reported a slight laxative reaction, which they only saw as positive, though this is a possible indication that it is best to take hypericum together with meals. Other symptoms, such as those often permissible in connection with the intake of psychopharmaceuticals (nausea, perspiration, dryness of the mouth etc.), were absent. A single participant developed a pronounced headache for a few days, though this was followed by feelings of anger and sadness (headache is often the result of suppressed anger). Correspondingly, with some individual clients, there was an increase in anxiety and thoughts producing anxiety, which may be viewed as a result of the general mode of action of hypericum: an emotional disturbance or something approximate. Contrary to what often happens in the case of synthetic antidepressants, we found no deprivation of sexual desire or capability of any considerable extent. Many participants reported having more intensive dreams than otherwise, and a few said they even woke up during the night. The old name “Thought Grass” (Tankegræs) is thus not quite so far off the mark, as an amplification and (the possibility of) clarification apparently takes place on the emotional plane. The other old name for the herb, i.e. “Horsebane” (Hestemare) refers to the physical and mental restlessness it produces in horses taking (too much of) it. Additionally, in the context of anxiety, we were able to determine a certain initial increase in clinical development in a number of subjects, this being so marked in one case, that the person concerned had to discontinue participation in the study. From a permanent viewpoint, this augmentation on the emotional plane may be considered potentially favourable — such an augmentation accelerating the therapeutic process and thus having a positive effect — though we must emphasize the fact that persons suffering from pronounced anxiety and especially paranoia ought not to take hypericum unless they are undergoing psychiatric treatment at the same time.

Finally, reports on physiological adverse reactions exist which give us reason to take them seriously. We are here dealing with interactions where other pharmaceutical preparations are concerned, such as for AIDS or heart conditions in addition to contraceptive drugs. These are perhaps reported by the media as being more serious than they actually are. It is a known fact that most forms of drug influence others in the context of effect, but since so many women take contraceptive drugs, we have reason to investigate this aspect more thoroughly. In general, we may say that hypericum is a very ”gentle” in relation to synthetic preparations (antidepressants), with which it has been compared.

Differences in effect (III) a) and b). What we defined as the tertiary aim of the study, i.e. how hypericum acts in the individual and how far its effect may be of diagnostic importance, is of great psychological interest, though it is obviously not particularly decisive from an empirical viewpoint, due to the low number of persons included. In our opinion, the experience we have gained with synthetic antidepressants and other drugs in treating our clients has been confirmed. According to our clinical experience, the class of antidepressants called ”happy pills” (lykkepiller) in Denmark (SRI) act in greatly differing ways in different personality types — or diagnostic groups, if one so wishes — and the same applies to hypericum, with which we here have a certain empirical basis worth taking into account. As regards the the following statements, we must underline the fact that they are the result of a particularly wide clinical experience; as such, therefore, they ought not to be understood as actual study results. Thus, the following remarks may at best be taken as applying to a ”pilot study”, meaning that further investigation is necessary to provide interesting results (see indications and contraindications below).

Pronounced anxiety and especially paranoia, particularly if the person(s) concerned are not undergoing current psychiatric treatment.

In general, psychotic or markedly confused patients/clients.

Patients having received previous medication (except small doses of neuroleptics, see below), including suicide hazard cases.

Pronouncedly depressive cases of a schizoid character (schizothymic depression, anaclitic depression).

Emotional instability, histrionic or highly fluctuating narcissistic personalities.

The majority of cases manifesting moderately depressive symptoms, including a number of nutritional disturbances and schizoid types (anaclitic depression), provided they are not paranoid and treatment is longer than 8 weeks

Persons who are relatively competent, socially and emotionally, though experiencing crises implying that they are already confronted with with many of the feelings which persons with severe depression or of the pronounced schizoid type are not capable of dealing with (due to or with pronounced physical pain or insufficient emotional development). This group may include many persons not fitting into the preceding three by virtue of ongoing non-medicinal psychotherapy.

No measurable placebo effect (III,d) was found, a fact giving us little surprise, even though we knew that most ”placebo-creating” factors were absent or reduced. The participants were amazingly good at ”guessing” whether they had received the active substance or not, in just the same way as the investigators were able to do.

As regards III c), as to how hypericum acts in different conditions through time, unpleasant conditions such as anger, irritation, self-discontentedness (G and S) appeared practically to increase during the first 4 weeks before dropping below the initial level. However, dejectedness or a feeling of being upset (T and H) experienced an initial decrease before rising again (week 5) only to drop below the initial level in the final run. Anxieties showed a pattern corresponding to that of G and S, increasing simultaneously during week 4, in advance of the increases of T and H during week 5. These forms of fluctuation are known to occur in psychotherapy, and can be explained psychodynamically. Hypericum appears to affect or produce feelings which had been suppressed or inhibited, thus producing a certain relief or anxiety in the first instance. The anxiety is generally linked up with a manifestation of aggressive feelings, although the threat of contact with pronounced mental pains (worry reactions) is also able to create anxiety (increase in advance of week 5). The graph shows the results for the verum group over 10 weeks, for the sum G+S (the ”aggressive” element) and T+H (the ”melancholic” element), together with A (anxiety; multiplied by two for the sake of comparison). In addition, certain disturbances exist (Di/10)-1, showing a transformation of the depression index.

Inasmuch as these results are the expression of a general tendency, we are therefore able to recommend that clients or patients continue treatment with hypericum for longer than 4 weeks — i.e. even if no mentionable change is taken felt — insofar as not dealing with a direct negative reaction or a condition which is so serious that another form of medication must be considered.

Although not directly related to the study, another experience we would particularly like to communicate to our colleagues, psychologists, physicians and psychiatrists is that certain patients found an advantage in combined medication: hypericum with a low dose of neuroleptics such as1 mg per day of Fluanxol (flupenthixol decanoate). Fluanxol reduces the discomfort directly related to anxiety and restlessness, for many meaning an increased working capacity so that, in our eyes, a number of the clients involved want to move from their group to the ”good and rapid effect” group. In this case, our experience agrees with what we have found with combining (the more recent) synthetic preparations with neuroleptics. For some clients, this combination is conclusive in a dramatic way. We have directly observed, for example, that Seroxat (paroxetine hydrochloride) by itself over a number of weeks has no or only a slight effect in persons who manifestly suffered from pronounced depression but with great anxiety as well. The addition of a little Fluanxol is able to change such a condition radically.

These last remarks may be given to illustrate that the effects of psychopharmaceuticals do not comprise a one-dimensional affair such as when we discuss either synthetic or natural medicine. In the clinical pictures discussed, hypericum acts in various highly individual ways, though they are happily positive or at least harmless in by far the majority of cases. Both knowledge acquisition level and accessibility to psychotherapy are not reduced, thus creating an ideal situation in the context of discussion therapy.

We are therefore able to recommend St John’s Wort to many persons suffering from mental problems of the milder kind, often including sadness, worry or reduced vitality; among other factors, therapy with hypericum is characterised by: few side-effects, low tolerance, good general effect, no drug dependency and the absence of a long-term/continuous effect.