Abstract
Depression was infrequently diagnosed before the advent of the antidepressants but has now apparently become a major public health problem. National campaigns are organized aimed at increasing recognition of the condition and at commencing treatment for sufferers. Implicit in these approaches is the premise that treatment will necessarily reduce disability and ultimately lower suicide rates. This is by no means certain. The treatment effect size of many antidepressants is modest, the burden of side effects they produce has never been established, and data on the quality of life during treatment is absent. It remains possible that mild depressive disorders confer a protective effect against suicide and that injudicious or unmonitored treatment may increase that risk. In their concern to help patients, physicians appear to have systematically overlooked the risks they expose patients to as part of their therapeutic effort to minimize the risks posed by the patient's condition. Their propensity to overlook the risks posed by therapy may stem in part from the availability of antidepressant treatments on prescription only. Remedying this situation will require first of all a recognition of the biases that prescription-only status introduces into therapeutics. Current pharmacological and neuroscientific developments have the potential to make alternative health care frameworks possible. Whether these alternatives are adopted will probably depend on the capacity of all interested parties to reform the present arrangements. Future concepts of depressive disorders will probably reflect the regulatory arrangements adopted.
Hard on the heels of the discovery of chlorpromazine, and the transformation of psychiatry occasioned by the availability of effective antipsychotics, came the discovery of the antidepressants. As has happened often before in science, two completely different kinds of antidepressants, a monoamine-oxidase inhibitor and a tricyclic antidepressant, were simultaneously but independently discovered by Nathan Kline and Roland Kuhn, respectively. The efforts to reconcile the modes of action of these two distinctive groups of drugs led to a recognition that both increased functional levels of monoamines. This, in conjunction with reports that reserpine, an agent that lowered monoamines, could lead to depression and suicide, led to the monoamine theories of depression in the mid-1960s (Schildkraut, 1965).
The Sanctioned Narrative
The catecholamine hypothesis of depression outlined by Schildkraut provided a new paradigm for psychiatry. It marked the point at which the eclipse of psychoanalysis became all but inevitable. A new psychiatric language began to emerge, which culminated in the publication of DSM-III in 1980.
The development of the monoamine theories also laid the basis for rational drug development, and there has, in addition, been a significant interplay between drug development and increases in the neuroscientific knowledge-base. The development of the selective serotonin reuptake inhibitor (SSRI) fluoxetine is popularly seen as an expression of our increasing capacity to make drug development more scientific in a manner that is leading to more effective and safer therapeutic agents.
The development of safer and more effective antidepressants has, in turn, underpinned a recognition of depression as a major source of morbidity and mortality. Epidemiological studies suggest that up to 5% or even 10% of the population at any one point in time may be depressed and that 20% to 45% of the population may, at some point during their lifetime, suffer from a depressive episode (Blacker and Clare, 1987; Kessler et al., 1994). There are projections that depression will be, in the early years of the next century, the second most significant source of disability in the health care arena (Murray and Lopez, 1996).
The development of pharmacotherapy has also been associated with a development in methodologies to evaluate the efficacy of therapeutic agents. Psychiatry was quick to adopt the randomized control trial (RCT) as a method for evaluation and has introduced standardized rating scales, as well as operational criteria, to make rational the assessment of therapeutic efficacy. These methods have permitted demonstrations of efficacy in conditions that often show large spontaneous variability, and which are also set against a background of diverse constitutional types and psychosocial settings. This approach, initially developed for the pharmacotherapies, has been forced on psychotherapeutic procedures also. They must now specify the active ingredients of therapy, pinpoint what outcomes are being sought, and ideally must be able to deliver those outcomes within periods of time that are not noticeably different to those produced by pharmacotherapeutic approaches.
The adoption of modern evaluative methods in turn has underpinned the development of evidence-based medicine (EBM), which governments, third party payers, and many scientists involved in health care embrace whole-heartedly. It is expected that EBM will eliminate from medical practice procedures that are inefficacious and costly by virtue of their inefficacy. On this basis, the link of practice to evidence may, it is hoped, reduce health care expenditures.
An Alternative Narrative
The first antidepressant was discovered in 1952, 5 years before either Kuhn or Kline, when both Jean-Francois Buisson in Paris and Max Lurie in Cincinnati demonstrated the effects of isoniazid in depressed patients (Delay et al., 1952; Healy, 1997; Salzer and Lurie, 1953, 1955). Lurie was probably the person who coined the term antidepressant (Lurie, 1998). His early reports are much more persuasive than the original articles by Kline (Loomer et al., 1957) and the first reports by Kuhn (Kuhn, 1996) on the effects of iproniazid and imipramine, respectively. More importantly, isoniazid brought about its effects in community depressive disorders without having a significant action on monoamine systems (Lurie, 1998).
There was another antidepressant discovery in 1955, 2 years before either Kuhn or Kline published their work. Against a background of reports that it made people "better than well" and that it was "good psychotherapy in pill form," Michael Shepherd at the Maudsley Hospital in England published the results of the first parallel group randomized control trial in psychiatry, in which reserpine was compared with placebo for anxious depressives (Davies and Shepherd, 1955; Healy and Savage, 1998). It was found to be an antidepressant. This result should come as little surprise to clinicians, as a great number of other neuroleptics including chlorpromazine, flupentixol, haloperidol, thioridazine, sulpiride, and pimozide among others, in addition to reserpine, have been shown by RCT methods to be of clinical benefit in depression (Robertson and Trimble, 1982). Antipsychotics are widely used in low doses for just this purpose. In the early 1960s, Klein and Fink randomized all patients admitted to Hillside Hospital to either chlorpromazine, imipramine, or placebo and showed that chlorpromazine had antidepressant efficacy (Klein and Fink, 1962a, 1962b). The differential benefits of imipramine appeared in individuals with phobic features. Klein later replicated the original study.
The evidence, in fact, that reserpine caused depression is meager. It did, however, lead to suicides in a number of patients who were being treated for hypertension, probably by causing akathisia (Healy and Savage, 1998). Reserpine also led to the creation of biochemical pharmacology on the basis of observations in Bernard Brodie's laboratory of a correlation between a lowering of 5HT levels and sedation after its administration in rabbits (Pletscher et al., 1955). Even though these observations were later shown not to have a causal connection, the reserpine model became a screening test for "antidepressants." Its use produced a generation of "me-too" agents, which had, however, the benefit of not producing akathisia. Some recent antidepressants fail the reserpine test and produce akathisia. The fact that the antidepressant effects of a wide range of neuroleptics and of isoniazid could have been overlooked point to a number of developments. Scientifically, the monoamine theories should never have developed or taken the stranglehold on the field that they did. Their dominance, therefore, must in some part be owing to other factors, such as the paradigm shift occasioned in psychiatry by the emergence of effective therapeutic agents. They also point to an emerging eclipse of clinical observation by laboratory based worldviews.
The neglect of the discoveries of the antidepressant properties of isoniazid and reserpine sheds light on a number of other important issues. These discoveries did not register owing, in great part, to the fact that the concept of an antidepressant did not exist at the time and indeed depression itself barely existed in the 1950s. When Kuhn approached Geigy and Kline approached Roche with their respective discoveries, they met indifference on the part of the companies, who it seems saw little market for a new kind of compound, which Kuhn called a thymoleptic and Kline called a psychic-energizer. The very word antidepressant cannot be found in the Third Edition of Webster's International Dictionary (1966), whereas in the Second Edition of the Random House Dictionary (1987), it is suggested that the word appeared at some point in the mid-1960s. In addition to having no concept of an antidepressant, companies were faced with estimates that no more than 50 to 100 people per million suffered from the kind of disorder that these new agents would treat (Healy, 1997). This did not appear to be a large enough market to warrant the development of a new class of drug, and indeed sales of the antidepressants through the 1960s remained unspectacular. These agents were the poor cousins of the antipsychotics and the minor tranquilizers.
Although epidemiological studies from the mid-1960s onward have pointed to widespread levels of nervous conditions in the community, the assumption that these are in the main depressive disorders only began to take root in the 1980s. This coincided with a crisis in Western psychopharmacology produced by the recognition of dependence to benzodiazepines. It also coincided with the development of the SSRIs. A comparable crisis, for example, did not happen in Japan, where dependence on the benzodiazepines, whether for cultural or genetic reasons or both, did not become the problem that it had become in the West. The Japanese anxiolytic market accordingly has remained much greater than the market for antidepressants and indeed, to date, no SSRI has been licensed for depression in Japan. Worldwide, recent World Health Organization studies indicate that a tranquilizer is more likely to be given for depressive conditions whereas an antidepressant is more likely to be used for an anxiety-based condition (Ustun and Sartorius, 1995).
The SSRIs have their origin in observations by Kielholz that not all of the antidepressants that were available during the 1960s acted in the same way to get depressed people well (Kielholz, 1971). Some appeared to enhance drive and others appeared to be doing something else. Carlsson recognized that those which were more drive enhancing had preferential actions on catecholamine systems, whereas those which were doing something else to bring about improvements in depression appeared to have preferential effects on the 5HT system (Carlsson et al., 1969; Carlsson, 1996). This led to his proposal to develop a selective 5HT reuptake inhibitor, of which the first such compound patented, in 1971, was zimelidine, whereas the first released clinically was indalpine.
The SSRIs, therefore, had their origin in the fact that not all antidepressants were functionally the same. Despite this, the dominant concepts of how antidepressants work through the 1970s, 1980s, and 1990s have stressed their common effects on a specific brain target. Beta adrenergic receptors were the supposed target in the 1970s and 1980s, whereas 5HT-1a or 5HT-2 receptors have been the focus in the 1990s. These unitary notions of antidepressant efficacy have been underpinned by the clinical trial methods, which began to play an increasing part in drug development from the late 1960s. Our assessments of efficacy in recent years have relied heavily on observer-based, disease-specific rating scales such as the Hamilton Depression Rating Scale (HDRS). On instruments like this, provided mild to moderate depressions are the subject of the study, all antidepressants appear the same, in that all produce roughly comparable falls in HDRS scores.
In fact, however, there were good indications from the start that the SSRIs were quite different agents from the previous generation of drugs. The early clinical trials of a number of these compounds in hospitalized cases of depression found them to be relatively inefficacious, so that the regulatory submissions in some cases consisted exclusively of nonhospitalized depression studies. Furthermore, it had begun to become clear during the 1980s that a number of these compounds produced effects on conditions such as obsessive compulsive disorder (OCD), which antidepressants without significant actions on the serotonergic system did not produce. It has since become clear that the treatment effect size for some of the SSRIs for social phobia, OCD, and posttraumatic stress disorder, for example, are larger than their effect size in moderate to severe mood disorders. The SSRIs are clearly not "weaker" than the older generation of agents, given their effects in a number of severe conditions such as body dysmorphic disorder or OCD, but they are different.
Their designation as "antidepressants" owes more to the state of the market in the mid-1980s than to anything else. The SSRIs appear to be effective across a wide range of nervous conditions, possibly by producing a reduction in emotional reactivity. It was clear, however, in the late 1980s that these compounds could not have been marketed as "anxiolytics" given the problems that had engulfed the benzodiazepines (Healy, 1991). The presence of several large companies moving into the depression market brought depression and the antidepressants out of the shadows. It led to an increased awareness of evidence that pointed to the widespread existence of conditions in the community that could be termed depressive, at a time when depressive disorders were underrecognized and undertreated. From estimates of 100 cases per million in the 1950s, the figure for depressive prevalence rose to between 50,000 and 100,000 cases per million in the mid-1990s.
This is a figure that has the contours of a serious public health problem. When physicians are also faced with claims that depressive disorders have a 15% lifetime prevalence of suicide, it would take a brave clinician who would not reflexively reach for a prescription pad. On the basis of figures like this, the vigorous detection of community depressions was advocated in many Western countries, in the hope that national suicide rates could be lowered. But these estimates of a 15% lifetime prevalence risk are derived from early studies of untreated cases of depression and later hospitalized samples (Guze and Robins, 1970). The rate of suicide in primary care primary affective disorders has never been established. Indeed, if figures of 5% to 10% annual prevalence for affective disorders are correct and if secondary affective disorders are much more likely to commit suicide than primary affective disorders, then the figures are such that it becomes conceivable that untreated uncomplicated mild depressive disorder may even confer a protective effect against suicide. On this basis, although the treatment of primary care depressions may reduce the burden of disability caused by this disorder, it is much less clear that increased rates of detection and treatment will reduce rates of suicide. Indeed, increased rates of detection and indiscriminate treatment along with, in particular, failures to monitor the impact of treatment may increase the risk of suicide (Healy et al., 1999) and impair quality of life. There would appear to have been an astute marketing of the evidence here.
Practitioners have perhaps been lulled into a false sense of security by the repetition of the safety in overdose and tolerability mantras for many of the newer antidepressants. In fact, the picture as regards safety and tolerability is ambiguous. The frequency of side effects of current psychotropic drugs has never been established. Estimates of the rates of sexual side effects with antidepressants provides an instructive example. Early clinical trials with SSRIs, based on the collection of spontaneous reports, suggested rates of sexual dysfunction of 5% (Stark and Hardison, 1985), whereas later systematic studies indicated rates in excess of 50% (Patterson, 1993). This points to two factors. One is the inadequacy of spontaneous reports as the method of assessing the incidence of side effects. Second, the revised figures for the incidence of side effects have come about in part with the development of other antidepressants, which have not had the same effects on sexual functioning as the SSRIs. For data on an issue to be generated and to come into play, therefore, it increasingly seems necessary that these data have a commercial value and a sponsor who will market the evidence.
At present, we are faced with a situation where authoritative compendiums such as Ayd's Lexicon (Ayd, 1996) can state that the propensity of fluoxetine to cause akathisia is widely recognized but the company can list 42 possible side effects of the compound, without mentioning akathisia (Plewes et al., 1997). There are further problems. In addition to companies citing frequencies of side effects based on the collection of spontaneous reports of adverse events, in recent years clinicians have allowed a situation to develop where many of these reports have been collected by relatively junior and untrained operatives working for clinical trial agencies (Stecklow and Johannes, 1997). This is a situation where either clinicians or the Food and Drug Administration should have insisted on the incorporation of comprehensive check lists of possible side effects to be completed by patients. It has, in fact, been left to a number of pharmaceutical companies to make moves toward remedying the problem.
In the belief that some of the newer agents would produce fewer side effects and hence a better quality of life, many of the companies bringing compounds to market during the 1990s produced quality of life (QoL) rating scales or used other QoL instruments in clinical trials. The results of very few of these assessments have been reported (possibly fewer than 10%). There is no obligation for companies to publish either negative trials or the results of negative outcomes on particular rating scales within otherwise published studies. If this point is taken in conjunction with the fact that the vast majority of rating clinical trials conducted within psychiatry are conducted on patient samples of convenience and essentially for business rather than for scientific purposes, and the fact that the epidemiological studies that have been conducted, although establishing the distribution of various symptoms and syndromes, have not shed light on the impact of treatment, it becomes clear that there is very little medicine-based evidence (MBE) to which psychiatric practitioners can appeal.
The upshot of this is that although all of the currently available antidepressants can be shown to have a treatment effect in mood disorders in clinical trials, it is less clear how much an individual clinician faced with a depressed patient should be compelled by this data toward a particular line of treatment. The data on placebo response in mood disorders, indeed, suggests strongly that a large element of clinical discretion should remain as to the correct moment to introduce pharmacotherapy into the management of a depressive condition. It also points to a need to monitor for psychosocial inputs into the genesis and maintenance of depressive conditions.
A Hidden Narrative
There is a further factor operative in the developments outlined above. In the 1950s, the newly emerging drug therapies became prescription-only medications (POM). There was opposition to this, but the thalidomide disaster copper-fastened prescription-only arrangements in place and with them encouragement to the pharmaceutical industry to develop compounds for disease indications only, in the hope that this would enhance the risk-benefit ratios of drug use. The antibiotics were the exemplary developments of the period.
These new arrangements produced a backdrop against which a premium was put on lesion models in psychiatry and on notions that pharmacotherapy corrected lesions, such as monoamine depletion, rather than produced changes in physiological or behavioral functioning that might have a range of applications of modest treatment effect size (Healy, 1998). For 30 years, the dominant views of how antidepressants act have seen them as all but antibacterial in their capacity to work regardless of constitutional type or psychosocial setting. We have accepted claims that the agents sold as antidepressants, which have diverse preclinical pharmacologies, are essentially all the same, differing only in their side-effect profiles and safety in overdose. We have neglected evidence that temperament or personality type may account for up to 50% of the variance in responsiveness to selective agents (Joyce et al., 1994).
Against this prescription-only background, the treatment of depressive disease with antidepressants was a better long-term bet for the pharmaceutical industry than the management of problems of living with minor tranquilizers. The market became a market of physicians and categorical models of nervous disorders rather than patients and lay or dimensional models of distress. The availability of St. John's wort as an over the counter "antidepressant" offers an instructive counterpoint. This is an agent, with a treatment effect size for depression as great as that of the SSRIs, that is marketed for stress, burnout, or problems of living. Individuals can take it without stigma and without giving themselves a depressive disease. The prescription of an antidepressant, in contrast, necessitates the concomitant prescription of a disease.
Retrospectively, then it should be no surprise that the restriction of sales to a disease market should have been associated with an apparent increase in the amount of disease in the community. Depression is not the only condition to have undergone an explosive increase. OCD, panic-disorder, social phobia, and a range of other conditions have also increased dramatically in apparent frequency (Baldessarini, 1970; Stoll et al., 1992; Zarate et al., 1997). This need not be a problem if physicians balance risk-benefit ratios appropriately. Their capacity to do this, however, will be hampered if managed care arrangements restrict their reimbursement to acts of prescribing only. If this happens, then one of the side effects of POM arrangements will be to reduce physicians to prescribers only. There may be other side effects. It is not clear, for example, what the social effects of the mass marketing of disease will be.
It may be that these side effects are an inevitable price to pay for the regulation of new agents. But if these arrangements, which give considerable power to medical practitioners, are to persist, then the relationship between those physicians and the pharmaceutical industry must necessarily be different from the relationship that exists for example between physicians and the tobacco industry. The original basis for prescription-only arrangements was to control access to addictive agents such as heroin and cocaine. The extension of these arrangements in the 1950s, in contrast, was on the basis that the populace at large was ignorant on health care matters and the new agents were complex. Making the drugs available through physicians only was undertaken on the basis that physicians would explain the indications for and hazards of these genuine miracles to those who might benefit from them. Given that drugs like the antibiotics and others are not addictive, the justification of POM arrangements for agents like the antidepressants is on the basis that physicians speak on the part of the public rather than act as a legitimate outlet for drug supply.
If they are to fulfill this role, physicians need better-quality data than they are currently provided with. For 30 years, we have been prepared to allow large amounts of pertinent data to remain unpublished. This data indicates that many antidepressants, although getting the core features of a mood disturbance well, do not contribute to a sense of well-being in a significant number of patients who take them. We have been persuaded to support campaigns to increase the detection rate of depressive disorders, which would be laudable had we also campaigned to improve the standards of monitoring outcomes. We appear not to recognize that, in the current market place, data only emerge into scientific debate when they coincide with the interests of significant parties. This applies as much to psychotherapy as it does to pharmacotherapy; indeed, it may be impossible in the therapeutic arena to have disinterested information. If so, clinicians must perhaps recognize this more clearly than they have done and be able to take account of the biases that are introduced into therapy accordingly.
There is no clamor from physicians for a change to current regulatory arrangements, but the situation may nevertheless be about to change as a consequence of neuroscientific developments and the emergence of "lifestyle" agents. Efforts to constrain the use of Viagra to medically certified erectile dysfunction seem Canute-like. In addition to being used outside a disease model, it looks as though Viagra will be available through the Internet, bypassing the need to access physicians. Similarly, although there have been vigorous efforts to make obesity into a disease, so that an emerging group of weight-reducing agents can be contained with the current framework, these agents are also likely to slip through the regulatory net. Against this background, consider the SSRIs. The effect size of a number of SSRIs for premature ejaculation is larger than their effect size for mood disorders. When these come off patent, what is the likelihood of their repositioning in this lifestyle market? In some European countries, agents such as the H-2 antagonists have moved from POM to OTC status. Given this there would seem little reason to block an application for the SSRIs to be OTC-unless of course these agents are not as safe or tolerable as we routinely claim they are. Once positioned there, or sold through the Internet for lifestyle purposes, the SSRIs could of course also be used for other nervous conditions. At this point, will it still be in the interest of companies to continue to market a disease model, or will they begin to endorse lay models as in the marketing of St. John's wort?
The developments in neuroscience may also begin to impact on practice in the near future. Hitherto, our neuroimaging capacities have given no better resolutions that those that can be obtained from the handheld post-mortem brain-no more than 1 cubic millimolar volume taking 12 seconds of real time to realize. In the near future, a new generation of machines will be capable of micromolar volumes within 1 second, permitting in vivo functional histology. Clinicians will be able to see whether their treatments are working in the expected manner-particularly when they are not producing the expected clinical change. In addition, although there have been clear indications for three decades that responses to psychotropic drugs breed true in families (Pare et al., 1962), this has had little effect on clinical practice. Pharmacogenetic developments, however, will provide simple blood tests within the next few years to predict the likelihood of adverse effects to therapeutic agents. Legal and financial considerations will drive the early and comprehensive implementation of these tests. This implementation in turn will necessarily reveal differences in constitutional types, pushing us back from our current highly categorical views of nervous disorders toward a more dimensional framework.
Combined, these developments will allow us to attain a better quality of outcome in clinical situations, where quality is defined in the industrial sense of reproducibility. From the point of view of the pharmaceutical industry, one of the benefits of the disease model to date has been that disease models are very forgiving on quality-crudely put, doing something has been better than doing nothing. That may soon not be good enough, and models other than the disease model may be needed to reflect the new possibilities.
References
Ayd F Jr (1996) Lexicon for psychiatry, neurology and neuroscience (p 268). Baltimore: Williams & Wilkins. [Context Link]
Baldessarini RJ (1970) Frequency of diagnoses of schizophrenia versus affective disorders from 1944 to 1968. Am J Psychiatry 127:59-63. [Context Link]
Blacker R, Clare A (1987) Depressive disorder in primary care. Br J Psychiatry 150:737-751. ExternalResolverBasic Bibliographic Links [Context Link]
Carlsson A (1996) The rise of neuropsychopharmacology: Impact on basic and clinical neuroscience. In D Healy (Ed), The psychopharmacologists (Vol 1, pp 51-80) London: Chapman & Hall. [Context Link]
Carlsson A, Corrodi H, Fuxe K, Hokfelt T (1969) Effect of antidepressant drugs on the depletion of intraneuronal brain 5-hydroxytryptamine stores caused by 4-methyl-alpha-ethylmeta-tyramine. Eur J Pharmacol 5:357-366 ExternalResolverBasic Bibliographic Links [Context Link]
Davies DL, Shepherd M (1955) Reserpine in the treatment of anxious and depressed patients. Lancet ii:117-120. [Context Link]
Delay J, Laine R, Buisson J-F (1952) Note concernant l'action de l'isonicotinyl-hydrazide dans le traitment des etats depressifs. Ann Medico-psychol 110:689-692. [Context Link]
Guze S, Robins E (1970) Suicide and primary affective disorder; A study of 95 cases. Br J Psychiatry 117:437-438. ExternalResolverBasic Bibliographic Links [Context Link]
Healy D (1991) The marketing of 5HT: Anxiety or depression? Br J Psychiatry 158:737-742. ExternalResolverBasic Bibliographic Links [Context Link]
Healy D (1997) The antidepressant era. Cambridge, MA: Harvard University Press. [Context Link]
Healy D (1998) Meta-analysis of trials comparing antidepressants with active placebos. Commentary. Br J Psychiatry 172:232-234. ExternalResolverBasic Bibliographic Links [Context Link]
Healy D, Langmaack C, Savage M (1999). Suicide in the course of the treatment of depression. J Psychopharmacol 13:106-111. [Context Link]
Healy D, Savage M (1998) Reserpine exhumed. Br J Psychiatry 172:376-378. ExternalResolverBasic Bibliographic Links [Context Link]
Joyce PR, Mulder RT, Cloninger CR (1994) Temperament predicts clomipramine and desipramine response in major depression. J Affect Disord 30:35-46. [Context Link]
Kessler RC, McGonagle KA, Zhao S, Nelson CB, Hughes M, Eshleman S, Wittchen HU, Kendler KS (1994) Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States: Results from the National Comorbidity study. Arch Gen Psychiatry 51:8-19. ExternalResolverBasic Bibliographic Links [Context Link]
Kielholz P (1971) Diagnose und Therapie der Depressionen für Praktiker. Münich: J. F. Lehmanns. [Context Link]
Klein DF, Fink M (1962a) Psychiatric reaction patterns to imipramine. Am J Psychiatry 119:449-459. [Context Link]
Klein DF, Fink M (1962b) Behavioral reaction patterns with phenothiazines. Arch Gen Psychiatry 7:449-459. ExternalResolverBasic Bibliographic Links [Context Link]
Kuhn R (1996) The first patient treated with imipramine. In TA Barn, OS Ray (Eds), A history of the CINP (p 436). Brentwood, TN: JM Productions. [Context Link]
Loomer HP, Saunders JC, Kline NS (1957) A clinical and pharmacodynamic evaluation of iproniazid as a psychic energizer. Psychiatr Res Rep 8:129-141. [Context Link]
Lurie M (1998) The enigma of isoniazid. In D Healy, The psychopharmacologists (Vol 2, pp 119-134). London: Chapman & Hall. [Context Link]
Murray C, Lopez A (1996) The global burden of disease. Cambridge, MA: Harvard University Press. [Context Link]
Pare CMB, Rees WL, Sainsbury MJ (1962) Differentiation of two genetically specific types of depression by the response to antidepressants. Lancet 2:1340-1343. ExternalResolverBasic [Context Link]
Patterson WM (1993) Fluoxetine-induced sexual dysfunction. J Clin Psychiatry 54:71. ExternalResolverBasic Bibliographic Links [Context Link]
Pletscher A, Shore PA, Brodie BB (1955) Serotonin release as a possible mechanism of reserpine action. Science 122:374-375. ExternalResolverBasic Bibliographic Links [Context Link]
Plewes JM, Koke SC, Sayler ME (1997) Adverse events and treatment discontinuations in fluoxetine clinical trials: An updated meta-analysis. Eur Neuropsychopharmacol 7:S169. [Context Link]
Random House Dictionary of the English Language (1987) Second edition. New York: Random House. [Context Link]
Robertson MN, Trimble MR (1982) Major tranquillisers used as antidepressants. J Affect Disord 4:173-193. ExternalResolverBasic Bibliographic Links [Context Link]
Salzer HM, Lurie ML (1953) Anxiety and depressive states treated with isonicotinyl hydrazide (isoniazid). Arch Neurol Psychiatry 70:317-324. ExternalResolverBasic Bibliographic Links [Context Link]
Salzer HM, Lurie ML (1955) Depressive states treated with isonicotinyl hydrazide (Isoniazid): A follow up study. Ohio State Med J 51:437-441. ExternalResolverBasic Bibliographic Links [Context Link]
Schildkraut JJ (1965) The catecholamine hypothesis of affective disorders: A review of supporting evidence. Am J Psychiatry 122:509-522. ExternalResolverBasic Bibliographic Links [Context Link]
Stark P, Hardison CD (1985) A review of multicenter controlled studies of fluoxetine versus imipramine and placebo in outpatients with major depressive disorder. J Clin Psychiatry 46:53-58. ExternalResolverBasic Bibliographic Links [Context Link]
Stecklow S, Johannes L (1997) Questions arise on new drug testing: Drug Makers relied on clinical researchers who now await trial. Wall Street Journal Aug 15th. [Context Link]
Stoll AL, Tohen M, Baldessarini RJ (1992) Increasing frequency of the diagnosis of obsessive-compulsive disorder. Am J Psychiatry 149:638-640. ExternalResolverBasic Bibliographic Links [Context Link]
Ustun T, Sartorius N (1995) Mental illness in general health care. New York: J. Wiley & Sons. [Context Link]
Webster's Third New International Dictionary (1996) Spring-field MA: G&C Merriam Co. [Context Link]
Zarate CA, Tohen M, Baraibar G, Zarate SB, Baldessarini RJ (1997) Shifts in hospital diagnostic frequencies: Bipolar disorder subtypes, 1981-1993. J Affect Disord 43:79-84. ExternalResolverBasic Bibliographic Links [Context Link]
Accession Number: 00005053-199903000-00007