Abstract
Keywords
Recommendations for the use of antipsychotic drugs for those with depressive disorders are unable to be found in standard textbooks, or in evidence-based or opinion-driven treatment guidelines – apart from the management of delusional or psychotic depression and where there exists a supportive meta-analytic database [1]. Yet Australasian psychiatrists have and do prescribe antispychotic drugs for those with depression, with an Australasian Data Base (ADB) study of 369 depressed patients establishing that 17% had received such a medication on a previous occasion for their depression [2]. Here we will suggest that the atypical antipsychotic drugs may have an antidepressant action
In approaching the proposition we should first reject the view that there is no evidence to support the roleof ‘typical’ antipsychotic drugs for depression. Robertson and Trimble [3] considered 34 double-blind trials of a number of the early major tranquillizers. They concluded, ‘some neuroleptics have antidepressant properties’ that they were comparatively free of side-effects and that they had an earlier onset of action than tricyclic antidepressants. That review would appear to have been uninfluential clinically, perhaps reflecting concerns about side-effects of the older antipsychotic drugs. For instance, Prien [4] noted that antipsychotic drugs are not a treatment of choice for maintenance therapy, despite their utility in controlling acute manic symptoms, asthe ‘risk of tardive dyskinesia and other adverse reactions limits their use on an extended basis’.
‘Atypicality’ (of the antipsychotic drugs) has been defined on the basis of favourable serotonin and dopamine receptor affinity ratios (5-HT2: D2 and D4: D2), and is thought to be associated with greater efficacy in treating the negative symptoms of schizophrenia and to have fewer extrapyramidal side-effects [5]. Despite some therapeutic overlap, the receptor interactions of the atypical antipsychotics actually differ substantially from each other. For instance, olanzapine possesses marked affinity for serotonin receptors (in particular 5-HT2A), while its affinity for dopamine receptors (especially D2) is much greater than that of clozapine. Similarly, olanzapine's affinity for the muscarinic M1 receptor sets it apart from risperidone which instead possesses appreciable α2 adrenergic affinity [6]. The greater variety of receptor interactions and broader pharmacological profile of the atypical antipsychotic drugs theoretically raises the possibility of a more extensive therapeutic portfolio.
A weak argument stems from observations of their antidepressant effects in patients with schizophrenia and schizoaffective disorder. Retrospective studies of clozapine indicated that those with schizoaffective disorder showed better response rates than those withschizophrenia [7–9] suggesting a thymoleptic action. A recent review supported such a role for risperidone, suggesting that up to 70% of patients with bipolar disorder and schizoaffective disorder may experience clinically significant improvement [10]. In a controlled trial of patients with coexisting psychotic and depressive symptoms (and receiving diagnoses of schizoaffective disorder, depressive type; major depression with psychotic features; or non-residual schizophrenia with major depression), risperidone alone appeared to show some antidepressant effect [11]. Risperidone has also been reported to be useful as an adjunct in the treatment of mania [12], psychotic depression [13] chronicdepression [14] and as an augmentation strategy intreatmentresistant [15] and agitated depression [16].
Turning to olanzapine, two controlled studies [17], [18] have examined its utility in the treatment of depressive signs and symptoms co-occurring with schizophrenia, reporting improvement, and with benefits persisting with maintenance therapy. Olanzapine has been reported to be of benefit in the treatment of psychoticdepression [19], [20] mania [10] and, importantly in regard to our proposition, of some assistance in the management of non-psychotic treatment-resistant depression [21], [22].
The second argument emerges from clinical observation. Over the last 2 years, we have trialled olanzapine in some 20, and risperidone in three, treatment-resistant depressed patients. We observed a range of outcomes, including seemingly striking benefit, apparent augmentation benefits, and no benefit. After detailing the clinical profile of the seven improved patients, we suggest that – and consider how – atypical antipsychotic medications may have a role in the management of non-psychotic melancholic depression.
Case summaries
Case 1
A 50-year-old woman described mild mood swings from late childhood, and the development of significant mood swings (including full manic episodes) requiring psychotropic medication from her mid-thirties. Her depressive episodes were most likely to occur in autumn and spring. For 6 years, her depressive episodes had responded to a tricyclic antidepressant (TCA). Episodes then appeared non-responsive to TCAs alone or when augmented with T3 (biothyronine). For a period, she then responded to fluoxetine alone. Subsequently, a combination of fluoxetine 20 mg and desipramine 50–75 mg kept her relatively free of depressive episodes for about 3 years, before severe ‘break through’ episodes of depression occurred. She was then treated with other selective serotonin re-uptake inhibitors (SSRIs) and venlafaxine, mood stabilizers (lithium and carbamazepine) and two courses of electroconvulsive therapy (ECT).
Depressive episodes were marked by profound insomnia, appetite and weight loss, severe psychomotor retardation, a deeply depressed and non-reactive mood and abnormal taste sensations and hyperacusis.
In the 6 previous years, once established, a depressive episode would tend to be treatment resistantfor 2–3 months and the patient was of the view that no treatment was then clearly efficacious – and that episodes ‘just ran their course’. In September (spring), while on venlafaxine 225 mg and valproate 800 mg (serum level 400 μmol/L), a non-psychotic depressive episode commenced, with cardinal early features of depressed mood, insomnia, amotivation, appetite loss and noise sensitivity, and olanzapine 2.5 mg nocte was prescribed. She stated that the next day she felt ‘zonked, as if drunk’, but that she had slept and that her mood had normalized. She remained euthymic over thenext 3 days, after which olanzapine was ceased without any recurrence of depression over the next few months.
The following March (autumn) a similar episode commenced despite the patient remaining compliant with venlafaxine and valproate, and this too was aborted within a day after taking olanzapine 2.5 mg nocte. On this occasion, olanzapine was maintained for 10 days and then ceased without any recurrence of symptoms over the next month (and when last reviewed). No previous treatment had aborted a depressive episode so promptly. In the following spring, a further episode while on maintenance medication failed to respond so clearly or so quickly to olanzapine. A significant medical problem may have confounded treatment response on the third occasion or the olanzapine effect may have ‘pooped out’. On a subsequent fourth occasion, and when the medical problem was corrected, she remitted rapidly again with olanzapine augmentation.
Case 2
A 51-year-old man described mood swings commencing in adolescence and his first clinical depressive episode in his early thirties. Following a second episode in his mid-forties, he experienced relatively persistent daily mood swings. Over the next few years he received a range of treatments and initially responded well to imipramine, but improvement was not maintained and he developed significant side-effects. He subsequently experienced severe serotonergic reactions to fluoxetine and venlafaxine, and was unable to tolerate lithium, carbamazepine or valproate. In early 1998 he elected not to take any more psychotropic drugs as he felt they offered no more than transient benefits.
In August 1998, after being ‘high’ for nearly a year, he developed a severe melancholic depression, marked by a profoundly depressed, non-reactive, guilty and suicidal mood, psychomotor retardation, appetite and weight loss, diffuse insomnia and his voice became husky. He was commenced on olanzapine which was steadily increased over a period of several days to 10 mg daily and he requested that fluoxetine be re-started. This too was increased over weeks to a daily dose of 40 mg and the combination was associated with some levelling of his mood swings for the next 8 months. He then experienced another severe depressive episode; the dose of olanzapine was increased to 15 mg daily, and the depression once again slowly remitted.
Six months later the olanzapine was stopped because of leg oedema, and the depression returned, being marked by diffuse insomnia and suicidal ideation. At this point, risperidone was commenced (2 mg nocte), and that night he slept for 10 h, stating that his depression was twothirds better the next day. Over the subsequent week the dose of risperidone was adjusted to 4 mg daily (reducing his sleep to 8 h/night), and his mood state remained considerably improved a month later (by which time he had reduced the risperidone further to 2 mg/day).
Case 3
A 30-year-old man developed an agitated melancholic depression 2 years previously, and reported a family history of bipolar disorder. His previous psychiatrist had sequentially prescribed fluoxetine 20 mg, paroxetine up to 40 mg (augmented also with thyroxine),nortriptyline 100 mg (the maximum dose tolerable) and lithium supplementation (ceased because of renal complications). Each treatment produced at best a partial response at best.
On assessment, he had slight psychomotor disturbance, both retardation and agitation, and described a moderately depressed mood with diurnal variation worse in the morning, anhedonia, amotivation, appetite loss and poor memory and concentration.
Venlafaxine was prescribed and the dose titratedto 300 mg/day but, because of headaches, tapered and substituted with phenelzine (60 mg/day), which resulted in sedation. In response to both these drugs, he judged himself some 80% improved but noted significant ongoing suicidal ideation, as well as troubling psychomotor agitation. Olanzapine 2.5 mg nocte was commenced and he remarked that it made him feel as if he had been ‘hit over the head with a lump of wood’. He gradually reduced the dose to 1 mg and described a distinct improvement in his mood. His agitation settled for the first time and he showed a clearly reactive mood, smiling appropriately at interview. Over the next few weeks he was mildly hypomanic (his first ever such episode), the olanzapine was ceased after several months and (until last review) he has remained euthymic while taking low-dose phenelzine (30 mg/day) only.
Case 4
A 44-year-old female described four significant lifetime episodes of depression commencing when shewas 15 and resulting in psychiatric hospitalization for a total of 18 months, where she had been treated with TCAsand ‘typical’ antipsychotics. She sought no treatment for the second 8-month episode, and was treated withfluoxetine 20 mg for the third episode, which resolved over a few weeks. She then remained well on fluoxetine for 5 years before stopping all medication.
At age 43, following a marital stressor, she experienced another episode of depression that failed to respond to fluoxetine, even at a higher dose (40 mg). At assessment, she had a severe mood disorder with anhedonia, amotivation, suicidal ideation and guilt, poor concentration, diffuse insomnia and psychomotor retardation. She reported transient perceptual illusions, but did not describe any clear delusions or hallucinations.
Olanzapine was commenced at 2.5 mg, with immediate benefit in her mood and sleep. The fluoxetine was assumed to have lost its effectiveness and was therefore substituted with citalopram. Meanwhile the olanzapine was increased to 5 mg and the combination was effective for a period of 2 months after which she became ‘flat’ and, once again, quite depressed. She was therefore switched from citalopram to phenelzine and the dose of this was increased to 45 mg daily and that of olanzapine to 10 mg daily, but she described significant side-effects such as sweating and ‘feeling stoned’. Both medications were therefore gradually withdrawn and her depression worsened. Olanzapine 5 mg daily was then reintroduced in conjunction with desipramine (titrated up to a doseof 250 mg without any significant side-effects nor benefit). Hence, augmentation with lithium was attempted. A low serum desipramine level (0.2 μmol/L; therapeuticrange 0.5–1.1) and a lack of any improvement prompted the use of ECT (10 treatments) to treat her profound depression associated with psychomotor retardation and social impairment. Objectively to her friends, she appeared to improve, however, she disagreed and because of concerns about her memory, ceased the course of ECT. Continuing with olanzapine 5 mg, sertraline was added and when increased to 300 mg daily, she reported some improvement, however, side-effects eventually necessitated a reduction to 200 mg daily and, after 4 weeks, she became euthymic.
When reviewed, 3 months later, she had stopped the olanzapine 1 month after episode recovery, was being maintained on sertraline 150 mg daily, and had no depression over the interval.
Case 5
A 45-year-old male had had mild mood swings begin in adult life and had developed a severe agitated melancholic depression following a stressful work situation. In addition to a profound mood disturbance with suicidal and guilt preoccupations, he lost 10 kg in weight and had diffuse insomnia, while showing distinct psychomotor agitation. His treating psychiatrist sequentially prescribed sertraline, fluvoxamine and aurorix without improvement.
On referral, he was commenced on desipramine, which was progressively increased to 250 mg daily, and haloperidol added because of overvalued guilty preoccupations (increased from 2.5 mg to 7.5 mg/day). Partial improvement led to augmentation with lithium, after which he judged himself to be 90% better, but noted distinct periods of depression and agitation. Subsequently, haloperidol was substituted by olanzapine, 2.5 mg initially and then 5 mg, with his condition rapidlyresolving – and leading to cessation of the olanzapine. Two months later, a reoccurrence, again following work stress, prompted the use of phenelzine 60 mg daily and the episode eventually remitted on this alone. However, he rapidly developed a manic episode, which was then followed by a severe melancholic depression resulting in hospitalization in a near catatonic state. He was treated with 30 mg of phenelzine and lithium (plasma levelof 0.7 mmol/L), but responded only partially, still experiencing a depressed mood and anergia, and manifesting agitation. The reintroduction of olanzapine, 2.5 mg nocte, was associated with an improvement inmood 3 days later as well as mitigation of his agitation. Three months later, he developed side-effects from lithium, and this was substituted with valproate. He then remained euthymic on this combination of olanzapine, valproate and phenelzine.
Case 6
A 58-year-old man had one episode of depression in his early twenties, but developed a severe agitated melancholic depression following a series of stressful events, with suicidal ideation, appetite and weight loss, anergia, anhedonia, early morning wakening and diurnal mood variation. He was initially commenced on citalopram, and then, because of overvalued ideas involving guilty ruminations, olanzapine 5 mg daily was added. His agitation settled but there was little mood improvement, so that the citalopram was substituted with desipramine and the olanzapine increased to 10 mg daily. Both he and his family reported a modest improvement. Over the next few months, lithium, valproate and T3 augmentation were trialled, while the desipramine was substituted with phenelzine, but his mood remained unimproved preventing him from returning to work.
Case 7
A 40-year-old female presented with a historyof ‘highs’ since her early teens, and clear melancholic depressive episodes since her early thirties. During the latter, she reported having a severe, non-reactive and guilt-dominated mood disturbance, characterized by patchy sleep, suicidal ideation, poor concentration and memory, and both psychomotor retardation and agitation. Her psychiatrist sequentially treated her with SSRIs (sertraline and fluoxetine), several TCAs, irreversible monoamine oxidase inhibitors (MAOIs; phenelzine and tranylcypromine) and lithium. She reported that most of these treatments had helped partially but that none had had a sustained effect, and that her episodes of depression were usually brought to a close by a period of elated mood.
For the presenting episode, 150 mg desipramine was progressively prescribed, but with slight benefit. Olanzapine 5 mg was then added, producing some modest improvement before being stopped after 5 weeks. At this point, her depression immediately worsened, and olanzapine was recommenced in conjunction with T3, associated with an 80% improvement over the subsequent week. The olanzapine was once again stopped after 2 weeks, and she again described a deterioration of mood, becoming tearful and having ‘black thoughts’. The recommencement of olanzapine for a third time (at a dose of 10 mg) was associated with a distinct improvement, while a reduction to 5 mg led to a recrudescence of depressive symptoms. Over the subsequent 18 months, several ‘new’ antidepressants such as nefazadone were tried, and previous treatments (e.g. lithium, valproate, phenelzine) recommenced, but without any distinct benefit. As in the past, her depressive episode resolved with the onset of mild hypomania, which she judged as her ‘usual pattern’ and not as a specific consequence of any treatment.
Discussion
In this case series, all seven patients evidencing improvement with an atypical antipsychotic had a treatment-resistant melancholic depression with evident psychomotor disturbance – with resistance eitherseemingly ‘intrinsic’ or emerging over time when previously effective treatments appeared to become ineffective. Four had a bipolar, and three a unipolar longitudinal depressive pattern. One (case 3) had his first ever hypomanic episode following olanzapine, which, while transient was followed by euthymia rather than a recurrence of depression. For those with an established bipolar pattern, improvement might merely reflect intrinsic mood shifts experienced by bipolar patients, a possibility that cannot be discounted. Alternately, the atypical antipsychotic drugs may act on or otherwise influence the ‘switch’ mechanism responsible for mood shifts in bipolar patients, an issue addressed in our review of the pharmacology of the antipsychotic drugs below. It is important to note that, although one patient (case 4) experienced perceptual illusions and two others (cases 5 and 6) had overvalued ideas of guilt, none of the patients were ascribed a diagnosis of psychotic depression, although such a possibility must also be conceded.
Two antidepressant effects were noted. In cases 1and 2, the implementation of olanzapine and risperidone, respectively, was associated with immediate improvement following a single dose, while in cases 3 and 4, rapid and early improvement was also evident following the introduction of olanzapine. In cases 5, 6 and 7 the atypical drug appeared to have been beneficial through augmentation, and in case 7, in particular, there was a close relationship between its addition and removal, and changes in mood. The ‘on-off-on-off-on’ sequence associated with relevant mood improvement and worsening sequences supports, but naturally does not prove, a true contribution.
As we principally prescribed olanzapine, we now examine its pharmacodynamic profile and possible antidepressant mechanisms. Olanazapine is an antagonist both in vitro and in vivo at many neuronal receptors including dopaminergic (D1–D5), serotonergic (5-HT2A, 5-HT2B and 5-HT2C), adrenergic (α1), histaminergic (H1) and muscarinic (M1–5) receptor subtypes [23], [24]. Clinically, it is an effective antipsychotic and, like many typical antipsychotics, it has moderately high D2 receptor occupancy [25–27] but, ‘atypically’, a low incidence of extrapyramidal side-effects (EPSs). High D2 occupancy can therefore be tolerated in the case of olanzapine in the absence of EPSs, however, what is unknown is whether this is due to concomitant 5-HT2 receptor occupancy or another mechanism masking EPSs. It is of note that, in most of our vignettes, patients described experiencing an improvement in mood within a relatively short period of time and that in some cases the effect was almost ‘immediate’. This suggests that the improvement in mood is unlikely to be secondary to olanzapine's antipsychotic effect as generally this occurs over a much longer period of time.
Major depression is broadly considered to be associated with the perturbation of one or more monoaminergic (5-HT, DA, NA) neuromodulatory systems. Studies examining the treatment of depressive signs and symptoms in those with psychotic disorders have found that olanzapine-treated patients exhibit statistically significant greater improvement in depression compared with risperidone-treated patients [17] and that olanzapine was superior to haloperidol in reducing depressive symptoms in schizoaffective disorder [28]. Such actions suggest then that it is important to examine how olanzapine may impact on the monoaminergic systems and exert an antidepressant effect.
First, the serotonin system. Some antidepressants such as nefazodone, mianserin and mirtazapine arepotent 5-HT2 receptor antagonists and, like β-1 adrenoceptors, 5-HT2 receptors are also down-regulatedfollowing 10–20 days of antidepressant treatment [29], [30]. Olanzapine occupancy of these receptors is high (74–92%) [25] and this may contribute to both its antipsychotic efficacy and more tolerable side-effect profile [31], [32] but also provide a means by which it modulates mood. However, 5-HT2 interaction is unlikely to be the primary, or indeed only, explanation for olanzapine's putative antidepressant properties as several clinical case studies report that it can also induce mania [33] and yet in one controlled study it was used successfully to treat the symptoms of acute mania [10]. Mechanistically, this might provide indirect support for it having some action at the ‘switch’ site for manic episodes.
The role of dopamine in depression, or at least certain depressive subtypes, is often overlooked, and yet certain characteristic symptoms of depression, such as anhedonia and emotional blunting, are very sensitive to DA agonists. Furthermore, subcortical DA systems have been shown to play a crucial role in the processing of rewardrelated information and in the selection and elaboration of motivated behaviour [34]. DA theories favour stimulation, rather than inhibition, of the limbic system to achieve antidepressant activity [35] and this is shown in the chronic mild stress-induced anhedonia model in rats in which DA agonists reverse anhedonia [36] but DA antagonists are able to reverse the effects of an antidepressant without having an effect themselves [35], [37].
Animal studies have also shown that antagonismof 5-HT2A receptors increases dopamine and serotonin levels in the limbic system [38] and that, in the medial prefrontal cortex of the rat, olanzapine robustly is able to increase the extracellular levels of dopamine andnoradrenaline [39]. It is possible that these changes also occur in the human prefrontal cortex, a brain region essential for cognition, affect and social behaviour [40], [41]. Given that 5-HT2 antagonists decrease the reactivity of DA neurones to pharmacological stimulation but increase their basal activity [42], it is possible then that olanzapine produces an antidepressant effect through a complex interaction between these systems. This would be in keeping with its suggested clinical efficacy in a percentage of our patients who all had melancholic depression with distinct psychomotor disturbance. The psychomotor disturbance in melancholic depression [43] may reflect a dopaminergic contribution and this is supported by the fact that depressed patients, especially those with psychomotor disturbance have reduced cerebrospinal fluid (CSF) homovanillic acid (HVA) [44]. However, these dopamine metabolite changes may simply reflect motor activity rather than the mood disturbance [45].
Dopaminergic dysfunction in depression, or at least in some subtypes, may explain the suggestion that SSRIs have reduced efficacy compared with TCAs and irreversible MAOIs in treating melancholic depression [2] and it may also explain why TCAs and MAOIs fail in a percentage of those with melancholia. Such limitations again suggest a potential role for dopamine-active antidepressants. Indeed, dopamine uptake inhibition is a prominent feature of a number of relatively new antidepressants such as amineptine and minaprine, both of which have been shown to be more effective than clomipramine in the treatment of retardeddepression [46], [47].
However, there is also some evidence of antidepressant benefit with dopamine antagonists. Low doses of the atypical antipsychotic sulpiride [48] have been shown to be antidepressant, with the seemingly paradoxical effect perhaps being mediated via dopamine autoreceptor antagonism, which increases dopamine turnover [3].
Clearly the role of dopamine remains to be considered carefully. Willner [45], when reviewing dopaminergic mechanisms in depression some years ago, cited reports suggesting that both dopamine agonists (e.g. nomifensine, buproprion and amineptine) and dopamine receptor antagonists (e.g. sulpiride) have antidepressant effects. He concluded that it is ‘questionable whether neuroleptics are truly antidepressant’ and went on to statethat ‘there is no evidence that neuroleptics can improve either psychomotor retardation or anhedonia’. However, whether such a profile extends to the atypical antipsychotic drugs and, in particular to olanzapine, remains unclear.
Interestingly, in conjunction with a seemingly immediate improvement in mood, some of our subjects described a marked and immediate restoration of sleep upon starting olanzapine. Olanzapine is an extremely potent histamine H1 receptor antagonist and, in the brain, these receptors are involved in the regulation of arousal and appetite [49]. This effect of olanzapine on sleep warrants further exploration, particularly as sleep disturbance is a robust biological marker of depression and any effect on sleep architecture may play a key role in its antidepressant effect. Also of interest is the fact that the muscarinic M1 receptor, at which olanzapine is an antagonist, is the most abundant in the human brain [50] and that olanzapine is also an agonist at muscarinic M4 receptors [51]. This is important because specific neurons modulate information processing and these pathways involve acetylcholine, GABA and neuropeptides, rather than NA, DA and 5-HT.
In their early review suggesting efficacy of the older neuroleptics, Robertson and Trimble [3] judged that they might be more efficacious for mixed anxiety-depression and depressive neurosis (than for ‘endogenous depression’), although they conceded that any suchdifferential – and their early onset of action – might reflect an anxiolytic action. On the basis of our clinical observations we favour pursuit of any antidepressant role of the atypical antipsychotic drugs in those with melancholic depression and where there is distinct psychomotor disturbance. Such disorders are generally severe and often associated with suicidal risk. They not infrequently result in considerable disability and may necessitate ECT: either because of an urgent need to achieve an antidepressant response or following the unsuccessful use of standard antidepressants. Such studies could well trial the atypical antispychotic drugs as (i) a single-line treatment for acute onset of such disorders in those whohave ‘break out’ exacerbations while receiving maintenance medication, and (ii) as an augmentation strategy for partial responders and for treatment-resistant depression. rIn relation to the last, Shelton and colleagues [52] have recently reported a controlled study demonstrating that the combination of olanzapine and fluoexetine was superior to each drug alone in the management of treatmentresistant depression, suggesting that our experience may not be idiosyncratic. As described for the older antipsychotic medications [53] we observed evidence of rapid onset of action in a number of instances, and it would be important to evaluate this component in future controlled efficacy studies.
Footnotes
Acknowledgements
We thank Kerrie Eyers and staff at our Mood Disorders Unit, Prince of Wales Hospital for study assistance. The study was supported by an NHMRC Program Grant (993208) and a New South Wales Department of Health Infrastructure Grant. No financial assistance was sought or received from any pharmaceutical company to undertake this case series.