Depression is unfortunately a very common illness; it is extremely unpleasant for sufferers and for their relatives and friends, and is associated with a substantial morbidity and mortality. The lifetime risk of developing depression was estimated by Boyd & Weissman (1981) as between 8% and 12% for men and between 20% and 26% for women. Estimates of point-prevalence vary, but a review of published studies suggests a level of 8.4% for men and 14.8% for women. A difference in prevalence of depression between men and women, with a twofold increase in women, appears to be well established. Recent evidence suggests, however, that the incidence of depression in men is rather more common than was previously thought: some apparent sex differences in frequency may be accounted for by a difference in coping with, and expression of, depressive symptoms (Angst & Dobler-Mikola, 1984). What is clear from the epidemiological studies in ‘normal’ populations is that much depressive illness remains untreated.

All existing treatments for depression, in particular the tricyclic antidepressants (TCAs), have drawbacks. Since the TCAs, for example, are associated with rather unpleasant and sometimes dangerous side-effects, which reduce compliance and compromise treatment, the search for new antidepressants with less unwanted effects and if possible greater efficacy remains an important goal. A recent trend in antidepressant research has been to develop compounds with specific pharmacological actions, which are selective for one of the amine systems. Currently, a number of specific 5-HT reuptake inhibitors are being developed, and some of these are now in clinical use.

The publication by the American Psychiatric Association in 1980 of the Diagnostic and Statistical Manual of Mental Disorders, Third Edition (DSM-III) - a text whose influence has been felt throughout the entire world and whose system of classification, particularly within the context of depressive disorders, has now been adopted by the majority of research work, no matter in what country it is implemented or published - has given the impression that an ‘American perspective’ has been substituted for the ‘European perspective’ that was predominant up to that point. However, this is a simplistic view of a complex history. Certainly, the basic methods of DSM-III originated from within those traditions which are particularly rooted in the USA, such as the quantitative approach to diagnostic criteria; these derive in the final analysis from the ‘statistical psychology’ devised by James McKeen Cattell at the University of Pennsylvania in 1887, and subsequently taught by him at Columbia University. But even among those factors which appear to be characteristically American, European origins can be detected: when he was in London, Cattell was the pupil of Galton, the founder of biometry. Furthermore, there has never existed in Europe any unanimity of opinion with regard to the problems of nosology: concepts supported in the German-speaking countries, in France, in the UK, or in the Scandinavian countries have only been partially adopted elsewhere and occasionally have remained specific to their national tradition. This paper will examine two of the fundamental points in the classification of depression which are linked to European perspectives - the notion of affective disorder and the endogenous/non-endogenous dichotomy - and will discuss the present situation created by the discordance which exists between the European and American approaches.

Recent progress in molecular genetics and neurobiology has attracted much attention and provided a new insight into the nature-nurture controversy in respect of the aetiology of psychiatric disorders. Affective illness, including the various subtypes of depressive and manic syndromes, has been the subject of a considerable amount of research on the relative importance of hereditary and environmental factors. This paper provides a critical review of the most relevant literature on the genetics of depression and mania.

There is evidence that serotonin (5-HT) is involved both in a wide range of psychiatric conditions and in specific behaviours which, in themselves, cannot be considered pathological. Besides the affective disorders (both depressive and manic, and both in full-blown episodes and during the intervals between these, which are seen as vulnerability to relapse) several other conditions need to be considered (Table I). There are also claims that serotonin metabolites or structurally related substances are involved, in their role as false neurotransmitters, in delusions and psychedelic phenomena, as in the case of LSD (Mantegazzini, 1960) or 5-hydroxytryptolines in delusions related to alcoholism (Davis et al, 1967).

This wide range of conditions raises the questions of how specific is the involvement of serotonin, and of what artifacts are present in attempts to investigate its actions. Page, who isolated 5-HT in 1948, stated (1968) that “serotonin has many functions but few have been sharply defined. The great variety of suggested roles can be said to be a tribute to man's ingenuity and his unquestionable willingness to write papers”. Since then, many more roles have been proposed for the involvement of 5-HT in functions of the central nervous system. Nevertheless, good evidence does exist that serotonin has some specific relationships to psychiatric disorders which are worth pursuing, in spite of great methodological difficulties.

The chemical structure of fluoxetine, (±)-N-methyl-3-phenyl-3-[(α,α,α-trifluoro-p-tolyl)oxy]propylamine, as shown in Fig. 1, lacks the three-fused ring system contained in tricyclic antidepressant drugs (TCAs) such as imipramine and amitriptyline. The p-trifluoromethyl substituent on the phenoxy ring of fluoxetine is an important determinant of its potency and its specificity as a serotonin-uptake inhibitor, e.g. the analogue having an o-trifluoromethyl substituent on that ring is only about one-hundredth as potent as fluoxetine in inhibiting serotonin uptake (Wong et al, 1975a). Nisoxetine (Wong et al, 1975b) and tomoxetine (Wong et al, 1982) are analogues differing from fluoxetine only in having an o-methoxy or an o-methyl substituent respectively, in place of the p-trifluoromethyl substituent on the phenoxy ring. Nisoxetine (LY94939) and tomoxetine (LY 139603) are potent and highly selective inhibitors of norepinephrine uptake (Wong et al, 1975b, 1982), differing strikingly from fluoxetine in specificity of uptake inhibition.

Fluoxetine has been shown to be a potent and selective inhibitor of serotonin uptake in laboratory animals; it is orally effective and has a long duration of action. This compound has been a valuable pharmacological tool to study the mechanisms of serotonergic neurotransmission and physiological functions of brain serotonin neurons (Fuller & Wong, 1977; Wong et al, 1985a; Fuller & Wong, 1987). The present paper summarises some of the pre-clinical studies which have characterised fluoxetine as a selective inhibitor of serotonin uptake.

Drugs are discovered by a variety of approaches, which include a systematic evaluation of synthetic compounds, using extensive structure activity relationships. After a lead compound which possesses desirable pharmacological properties is found, through the exhaustive screening of a variety of compounds, a number of congeners are then synthesised and tested, using specific physiological or biochemical models of therapeutic agents. The discovery of fluoxetine involved this complex process, and led to the description in the literature of one of the first selective serotonin uptake inhibitors (Wong et al, 1974); it was subsequently tested clinically, and these studies demonstrated its effectiveness and safety as an antidepressant drug (Chouinard, 1985). This paper reviews the published information relevant to fluoxetine's pharmacology and pharmacokinetics in man.

The tricyclic antidepressants (TCAs) were discovered accidentally by pharmaceutical chemists seeking first, better antihistamines and then antipsychotic agents. Careful clinical assessment revealed the antidepressant properties and suggested that the closer the patient resembled the classical textbook description of ‘endogenous’ depression, the more likely was an adequate clinical response to occur (Kuhn, 1958). However, it was quickly realised that the TCAs possess a plethora of side-effects, particularly sedation and symptoms related to their anticholinergic effects.

Despite much research over the past 25 years, it is unclear how TCAs effect clinical improvement. The two main neurotransmitters involved in some way are noradrenaline (Schildkraut, 1965) and 5-hydroxytryptamine (5-HT; serotonin; Van Praag, 1977). It was proposed that TCAs acted by blocking the reuptake of one or other or both of these neurotransmitters, thereby increasing their concentration in the synaptic cleft. However, as well as these acute effects, chronic effects, such as a decrease in the number of central beta-adrenoceptors (‘down-regulation’) occur, and these are probably more relevant to the clinical action.

The most unequivocal demonstration of efficacy with putative or reference antidepressants is normally obtained from placebo-controlled studies; heavy reliance is therefore placed on such investigations, particularly in the early stages of development of any new therapeutic agent. However, many clinicians are understandably reluctant to allow some patients to be treated with placebo, which may lead to a sampling bias. It is equally understandable that doctors taking part in such studies are anxious about clinical deterioration or lack of response, and this in turn may result in a higher drop-out rate than would be seen in studies using reference antidepressants without a placebo treatment group. However, in spite of these and other drawbacks, it is still generally accepted that the efficacy of potential antidepressants is established in a more scientific and unequivocal way using placebo-controlled methodology. Fluoxetine has been the subject of a number of such investigations (Table I); their findings are reviewed in this paper.

As a result of research in recent years, therapeutic possibilities for the treatment of depressive illnesses have been significantly expanded. Monoamine oxidase inhibitors and selective noradrenaline (e.g. oxaprotiline) or serotonin-reuptake inhibiting (e.g. fluoxetine, fluvoxamine) substances have been added to the classical tricyclic antidepressants, whose mode of action is primarily based upon combined noradrenaline and serotonin reuptake inhibition. Furthermore, the extent to which benzodiazepine derivatives or neuroleptics can be successfully used at low dosages for the treatment of depressive illness is currently being investigated.

For these compounds to be used in an individualised manner, however, agreed diagnostic criteria must be taken into consideration, as well as other factors such as the severity and type of depressive syndrome (mild, marked, or severe, retarded or agitated). For example, the benzodiazepine derivative alprazolam has a therapeutic efficacy comparable to amitriptyline for the treatment of milder depressive syndromes, but is significantly less effective for a severe depressive disorder.

We undertook a study of fluoxetine, a selective serotonin-reuptake inhibitor (Wong et al, 1974, 1975), in a double-blind comparison with amitriptyline. We particularly wanted to know if the efficacy of fluoxetine was comparable to that of amitriptyline, and whether fluoxetine caused less adverse effects as a result of its selectivity, which remains unchanged even after its metabolisation to desmethyl fluoxetine (Fuller et al, 1978).

The efficacy of antidepressants is measured primarily by their ability to treat the acute symptoms of depression yet the underlying process may take longer to resolve. A period of emotional frailty follows resolution of the acute stage, during which the patient, although apparently symptom-free, is prone to suffer relapse of the original symptoms. Antidepressants should not be stopped immediately a response is observed, but should be given for a longer period to prevent early relapse. In those suffering from recurrent depression, long-term treatment may prevent later recurrences of new episodes of depression.

The distinction between relapse of old symptoms and recurrence of new episodes is of particular importance, both in assessing efficacy of antidepressants during the continuation phase of acute treatment and in the separate assessment of their prophylactic efficacy. Unfortunately, though, many investigations have failed to distinguish between these two distinct phenomena. Klerman & Paykel (1970) emphasised the need to distinguish between early return of depressive symptoms, which they label ‘relapses’, from the later new episodes which they term ‘recurrence’. However, there has been a general lack of systematic investigation of this question, as well as a failure to use sufficiently precise methodology to distinguish between relapse and recurrence. Some studies which purported to be of prophylaxis were in effect only examining the continuation phase of acute treatment.

Fluoxetine is a new antidepressant agent which is a selective inhibitor of neuronal serotonin uptake; it has minimal affinity for muscarinic, dopaminergic, histaminic, serotonergic, or noradrenergic receptors (Stark et al, 1985). This specificity of activity suggests that fluoxetine may have a side-effect profile which is different from previously available antidepressants.

The safety of fluoxetine has been extensively studied: several hundred patients have received the drug continuously for more than one year - some have had therapy for 5 years or more. Previous reviews of the safety of fluoxetine (Wernicke, 1985; Zerbe, 1986) have described a smaller population of patients than is now available.

This review is drawn from data pooled from comparative clinical trials, which included 4336 patients: fluoxetine - 2938, tricyclic antidepressants (TCAs) - 599, and placebo - 799 patients. The TCAs studied were amitriptyline, imipramine, and doxepin. Most patients were adults with major depressive disorder, and the most common study design was a 6-week comparative double-blind phase, followed by unblinded long-term treatment. While the core of this study is a data base of 2938 fluoxetine-treated patients, all serious adverse events reported in over 7500 fluoxetine-treated patients worldwide, as of mid-1987, have been included.

Antidepressants are associated with a variety of side-effects, ranging in severity from the merely annoying to those with significant morbidity and even potential mortality. Prominent among the latter are cardiovascular effects, which have provided much of the impetus for the development of newer antidepressants such as fluoxetine. In this paper, we will review these cardiovascular effects, but as this has been previously undertaken elsewhere (e.g. Glassman & Bigger, 1981; Marshall & Forker, 1882; Glassman, 1984a), our consideration will be selective, with emphasis on methodological issues of relevance to the evaluation of new antidepressants. Since mechanisms of cardiac side-effects are best understood in relation to tricyclic antidepressants (TCAs), these will be a major focus, and will provide a background for the consideration of newer agents.

While less is known regarding the side-effects of the newer antidepressants, examination of the evolution of knowledge and opinions regarding their safety is particularly pertinent in that it illustrates the waxing and waning of enthusiasm that typically accompanies the introduction of new therapeutic agents. We review below the growing body of data regarding fluoxetine and some of the other serotonin-selective antidepressants, as well as discussing serotonergic effects on the cardiovascular system and their implications for the clinical use of these agents. Since the interactions between antidepressants and drugs used to treat cardiovascular disorders have been reviewed by Risch et al (1982) and Glassman & Salzman (1987), these will not be considered here.

The measurement of the effects of an antidepressant on psychomotor ability and cognitive processing tests is important for three reasons. Firstly, because a change in psychomotor and cognitive function is one of the truly objective assessments of psychotropic activity of a drug. Secondly, since batteries of psychological tests are often analogues of the essential aspects of real-life behaviour, performance changes can reflect the potential impairment of the activities of everyday living such as car driving, operating industrial machinery, and domestic tasks. Thirdly, since Widlocher (1983a,b) has proposed that mental and psychomotor retardation are a primary expression of depressive illness, psychomotor and cognitive tests can be used to measure the extent to which a putative antidepressant affects these psychological functions. Psychopharmacological testing can thus indicate the ‘safety’ of the drug in clinical usage by establishing an index of the behavioural ‘toxicity’ of the substance. It can also objectively determine the magnitude of the drug's psychoactive properties, and the extent to which a particular compound affects information processing and cognitive function. Substances which have a negative action on these aspects of psychological integrity could be regarded as counter-therapeutic, as they are making psychomotor and mental functions worse.

During the past 75 years, the proportion of elderly individuals in the USA has grown twice as fast as the general population. Depression in this age-group occurs four times more frequently than in the general population (Butler, 1975), and the suicide rate for people over 65 years of age is 15 times greater than that of the general population (Lehman, 1980).

The elderly may be more susceptible to depression due to biological and/or psychosocial variables. Elderly people experience significant losses associated with increasing age, including death of spouse and friends, loss of work, social status, and physical and mental abilities (Lehman, 1980). The biogenic amine hypothesis suggests that the aging brain may experience a decrease in the functional availability of neurotransmitters (Lehman, 1980); this decrease may also play a role in the aetiology of depression.

Due to age-related changes in the body, the elderly can be more sensitive to drug therapy. Older patients may require careful dosage adjustments and may also be more prone to experiencing drug-related adverse events. The elderly often receive medication for various indications, and drug interactions are a concern (Thompson et al, 1983). Therefore, efficacy and safety studies of new antidepressants in elderly patients are particularly important. We pooled data from both double-blind and open-label studies to evaluate the efficacy and safety of fluoxetine in geriatric outpatients with DSM-III major depression. Positive results of fluoxetine in the treatment of geriatric depression were reported in one of these studies (Feighner & Cohn, 1985). The favourable safety and side-effect profile of fluoxetine in the general population has been discussed elsewhere (Wernicke, 1985). Plasma concentrations of fluoxetine in elderly subjects are similar to those in younger individuals (I.emberger et al, 1985). These findings, combined with a lack of cardiovascular effects (Fisch, 1985), and low lethality with overdose, indicated promise for fluoxetine as a geriatric antidepressant.

Fluoxetine, a selective serotonin-uptake inhibitor, has been shown to be effective in the treatment of depression. A series of studies found fluoxetine to be more effective than placebo and to be associated with equivalent efficacy to standard reference tricyclic antidepressants (Stark & Hardison, 1985; Feighner & Cohn, 1985).

The selection of a dosage regime for fluoxetine in these studies was based on the information gathered from open dose-ranging investigations. In these studies, treatment starts with one dose level, which is then increased as indicated in the protocol or until increased doses are not tolerated due to side-effects. One of the weaknesses in this method of choosing the likely standard dose of a new antidepressant lies in the inherent bias towards the selection of the higher dose levels. This is particularly true for drugs which are well tolerated.