The concept of schizophrenia is contentious. Although most psychiatrists accept without question the notion that schizophrenia is a disease – a psychiatric disorder brought about by changes in brain function, scepticism reigns outside the profession, especially in many of those to whom the label is attached. So-called ‘schizophrenics’ may choose a variety of non-medical ways to understand and explain the nature of their experiences, and this leads us to the heart of a conflict in which the subject’s explanation of his or her experiences lies at odds with the psychiatrists. Inevitably it is the psychiatrist’s account which will turn out to be the most powerful, as can be seen in the famous case of Judge Daniel Schreber (see, for example Leudar & Thomas 1999). But it need not always be this way. In this paper, I shall try to demonstrate some of the shortcomings of the way in which psychiatrists interpret the Kraepelinian concept of schizophrenia, especially the relationship between the acute ‘symptoms’ of the condition and outcome. I shall then challenge the way this interpretation is used to justify the use of long-term neuroleptic medication, with the conclusion that psychiatrists must jettison the medical model when working with people in acute psychoses, in favour of humanistic models which recognise the importance of the relationship between the individual and the experience of psychosis. But first, we must consider the origins of the concept of schizophrenia.
Kraepelin and Dementia Praecox Morel provided one of the earliest descriptions of the condition we now call schizophrenia. In 1860 he described a case of severe intellectual deterioration in an adolescent which he called demence precoce (dementia of early onset). There followed a series of descriptions in the neurological literature of the day, the most influential of which was provided by Kraepelin in the fourth edition of his textbook in 1893. Here he described an illnesses called dementia praecox, under the heading of ‘Psychic Degenerative Processes’. This group of illnesses was characterised by a progressive and irreversible deterioration of intellectual function, typically in young people. He gave the following definition of the illness in the eighth edition of his book (Kraepelin, 1913): “Dementia Praecox consists of a series of clinical states which have as their common characteristic a peculiar destruction of the internal connections of the psychic personality with the most marked damage of the emotional life and volition.”
Early psychiatrists, influenced by Griesinger, believed that medical science would reveal the cause of all forms of insanity. But basic clinical sciences such as pathology, physiology and anatomy, which had been so successful in revealing the cause of neuro-syphilis, failed to reveal the cause of dementia praecox so Kraepelin had to rely on systematic clinical observation. Hoenig (1983) has described Kraepelin’s katamnestic method, which relied on the detailed description of individual cases, in an attempt to delineate the specific symptoms and signs of dementia praecox, in the hope that such an approach would confirm that dementia praecox was a disease entity with a specific anatomical basis and pathology. The issue here concerns the extent to which he succeeded in describing a single disease, dementia praecox with poor outcome?
The first point concerns the great variability of symptoms to be found in schizophrenia. Even Kraepelin (1913) admitted this, going so far as to say that: ‘Unfortunately in the field of psychic disturbances there is not a single symptom which is pathognomonic for any particular illness. On the other hand we can expect that the composition of the individual characteristics which form the total picture, and in particular the changes which develop in the course of the illness, will not be produced in exactly the same way by any of the other diseases.’ (Vol.II, p. 945, quoted in Hoenig, 1983; my italics).
Kraepelin failed to provide a “pathognomonic” symptom for schizophrenia, and, in lieu of this, he chose to emphasise the importance of the course of the illness. This position has been immensely influential. After Kraepelin, psychiatrists believed that a deteriorating course was a cardinal feature of schizophrenia. Deterioration here refers to the “changes which develop in the course of the illness” in the quotation from Kraepelin above. It is this that forms the basis of the belief that schizophrenia is a poor prognosis illness. Furthermore, there is an implicit assumption that deterioration is related to the putative disease process.
The problem is that Kraepelin not only found it difficult to be prescriptive about the symptoms required to diagnose schizophrenia and thus predict outcome, but he was also unable to show that all of the 127 cases of schizophrenia he studied in detail actually turned out to have poor outcome and deteriorating course. Kraepelin himself found that complete recovery was possible in 12.5% of cases (Vol. II, p. 865, quoted in Hoenig, 1983). Bleuler remained unconvinced that the condition was a single disease entity, because of the diversity of symptoms and variation in outcome (Bleuler, 1911): ‘Kraepelin’s concept continues to be opposed by many, who because of the wide diversity of the clinical pictures, cannot accept it as a single entity, which originally appeared to be based on the uniform course of the illness, and yet could include cases with a good as well as a bad outcome.’
The Course of Schizophrenia There have been two ways of investigating this. The first depends upon long-term studies of large numbers of patients over many years. The second involves smaller studies over shorter periods of time, but has the advantage that standardised diagnostic and outcome criteria are used. In the first approach Huber et al (1979) in Bonn studied the outcome of over 500 people diagnosed with schizophrenia over 21 years. In 57% of cases outcome was favourable, although there was great complexity in the patterns of the condition over time. In Zurich, Bleuler(1978) followed up over 200 patients for 22 years. He found good outcome in 53% of subjects. Ciompi (1980) identified nearly three hundred people admitted to a psychiatric hospital in Lausanne whose histories were documented until the age of 65. The average follow-up period was almost 37 years. Like Huber, he found considerable variation in the pattern of the symptoms with time. Overall, 49% of patients had a favourable outcome, and in comparison with the situation on admission, in two thirds of cases mental health was completely or partially improved. These studies show remarkably similar outcomes for schizophrenia, with approximately 50% having good outcome, but there are weaknesses relating to the retrospective design of these studies, and the reliability of the diagnoses. Ciompi used Bleuler’s criteria, which may include some people who are not suffering from schizophrenia and who are destined to have a good outcome, thus artificially improving the apparent outcome in schizophrenia. The best way around this is to use a prospective design. Carpenter et al (1978) followed up over 130 patients in the American IPSS research centre after five years, and split their sample into two groups, based on the twenty patients with the best and twenty patients with the worst outcome scores. Only restricted affect discriminated between the two groups of subjects on initial assessment. The authors concluded that the diagnosis of schizophrenia, when based on the presence of acute (positive) symptoms was of little value in predicting the course and outcome of the condition. This study indicates the difficulties in predicting outcome when people present with acute symptoms of schizophrenia. The main purpose of diagnosis and recognising diseases is to enable us to make such predictions. Hays (1984) has pointed out that despite the efforts of thousands of researchers over many years and the development of many different sets of diagnostic criteria, the cause of schizophrenia remains unknown. Psychiatrists must, he argues, face up to the unpalatable fact that schizophrenia as a disease entity conceived of by Kraepelin does not exist.
“… (Kraepelin) is the sole, flawed, authority for this extraordinary but influential theory.” (Hays, 1984)
This forces us to the conclusion that there is little evidence to support the Kraepelinian model of schizophrenia. There is neither the evidence to support biological models of the condition, nor the evidence to justify the existence of the condition as a diagnostic category, characterised by specific acute symptoms which predict outcome. I want to make one final point concerning poor outcome. Ciompi found that in two thirds of cases social outcome was either intermediate or poor. His figures suggest that the main field of impairment in schizophrenia is not symptoms, but social functioning. But in view of the fact that all the patients studied had been institutionalised for decades, this is hardly surprising, and may have little to do with the progress of an illness. Zubin et al (1983; see chapter two, p. 38) have suggested that over the last fifty years schizophrenia has become a less severe condition with a better outcome. Over this period the management of the condition has changed profoundly. People who have the condition are no longer incarcerated in institutions for years as was the case when Kraepelin wrote about dementia praecox. This raises the possibility that the poor outcome described by Kraepelin was an artefact of institutionalisation.
Medication and Outcome Research indicates that between 60 to 80 percent of subjects whose symptoms are controlled by depot neuroleptics experience a relapse and hospitalisation if medication is discontinued. Such evidence provides strong support for the role of neuroleptics in the long-term management of schizophrenia. The difficulty is that there is also much evidence that medication fails to help many people diagnosed with schizophrenia. There are four issues at stake here. The first concerns the efficacy of neuroleptics in controlling positive symptoms of schizophrenia. The second concerns their claimed effect in favourably changing the long-term outcome of schizophrenia. The third concerns the validity of this evidence, much of it drawn from specially designed research protocols (drug trials), in relation to the reality of the day to day practice of psychiatry. Finally, and most worrying of all, there is evidence that relapse, should it occur, arises through a supersensitivity reaction similar to that thought to be responsible for TD.
1. Efficacy of neuroleptics Davis and Caspar (1977) noticed that 25 percent of people continued to have psychotic symptoms despite taking neuroleptics. The persistence of their symptoms was not explained by the presence of physical health problems that might cause them to be suffering from organic brain disorders. Davies et al (1980) reviewed a large data base of clinical trial data from the 1960’s generated by the National Institute of Mental Health Psychopharmacology Research Branch in the USA. About 30% of subjects in these studies were rated as either minimally improved or clinically worse, following treatment with neuroleptics. Kane et al (1988) suggested that a conservative estimate indicates that the symptoms of a fifth of all patients suffering from schizophrenia fail to respond to neuroleptic drugs. The most striking evidence of a group of people whose psychotic symptoms failed to respond to neuroleptic medication came from a British study. Curson et al (1988) found that 46% of their sample of over 200 long-stay patients had persistent delusions, and 32% had persistent auditory hallucinations. The point here is this group of people had been exposed to what the authors described as “energetic pharmacological … treatments”. These studies indicate clearly that some people diagnosed as suffering from schizophrenia do not respond to neuroleptic medicationand continue to experience so-called positive symptoms. But what about the effects of neuroleptics on the course of schizophrenia?
2. Neuroleptics and long-term course Psychiatrists, prompted by the pharmacological industry, have believed that maintenance medication taken for many years not only controls positive symptoms, but also improves social function. Ciompi (1980) was also interested in the extent to which neuroleptic medication influenced the course of schizophrenia. He divided his subjects into three groups by date of first admission, corresponding to the period prior to the introduction of electro-convulsive therapy (pre 1933), the period prior to the introduction of the neuroleptics (1933 to 1953), and the period following the introduction of neuroleptics (post 1953). If neuroleptics have a beneficial effect on outcome, then we would expect the latter group of people to fare better. Ciompi was surprised and disappointed to find that over very long periods of observation, no statistically significant differences could be found in outcomes of people admitted in these different therapeutic eras. He concluded that people admitted with schizophrenia in the nineteen fifties, following the introduction of the neuroleptics did no better in terms of social adjustment than those admitted in the first three decades of the century.
Kane and Freeman (1994) have reviewed the advantages and drawbacks of neuroleptic treatment. First, they make the point that a number of authorities have doubted whether neuroleptics have a beneficial effect on outcome. “…it remains debatable whether or not long-term neuroleptic treatment substantially alters the course and outcome of this disorder…” (Kane & Freeman, 1994, p.22)
Karl Leonhard (1980) argued that psychiatrists should only consider using long term therapy once it was established without doubt that the condition from which the individual was suffering was a phasic disorder, that is a condition characterised by recurrent episodes of positive, or other acute psychotic symptoms. This was based on his view that such a symptom pattern tended to respond to neuroleptics, whereas the defect state did not. Schooler (1991) in her study of dose reduction commented that the results of naturalistic long-term follow up studies in schizophrenia show clearly the great diversity of outcomes of the condition. Despite long term medication, some people appear to have a very poor outcome. On the other hand some people do very well with little or no medication. The word naturalistic is very important here. Most of the accepted evidence in support of the value of neuroleptics in the short-term and long-term management of schizophrenia comes from drug trials, which are special studies resourced and supported by the pharmacological industry. Such studies can hardly be described as independent. In addition, they are performed under ideal conditions which are hardly representative of the circumstances under which these drugs are used by psychiatrists in the daily work. In this sense, the ecological validity of these studies must be called into question.
The most naturalistic study in this sense was undertaken by Johnstone et al (1986) working at Northwick Park in London. Patients were under the care of forty psychiatrists working within a thirty five mile radius of Harrow. Over 460 patients met the criteria for inclusion in the study over a period of almost two and a half years, yet 14% of these patients could not be assessed for a number of reasons. These included refusal of either patient or consultant to consent to study, self-discharge before assessment could be made, administrative problems such as patient not admitted, or ‘conversation’ barriers. These problems are typical of those commonly encountered by psychiatrists in their day to day work. Of the patients who could be assessed, only 120 could be entered into a randomised placebo-controlled drug trial of maintenance neuroleptic medication (Crow et al, 1986). They found surprisingly high rates of relapse in the two years following discharge both in the placebo group and, surprisingly, the group who received active medication. Although 70 percent of patients on placebo relapsed, 58 percent of those taking active medication did so over the same period. Overall these results suggest that at best, the benefits of neuroleptic medication in preventing relapse are marginal.
The situation is summarised well by another quote from Kane and Freeman (1994): ‘Thus, current neuroleptics, despite their great value in controlling schizophrenic symptoms, have a number of major drawbacks. They are far from consistently effective, and a substantial subgroup of patients derive little if any benefit from them. They remain ineffective against negative symptoms. They cause adverse neurological effects which may interfere with psychosocial and vocational rehabilitation and are associated with major problems of drug compliance.’ (p.29)
3. Relapse – drug discontinuation or illness? It is worth pausing at this point to consider an important assumption present in much of the research that we have considered here. When psychiatrists talk about the relationship between relapse and discontinuation of neuroleptic medication, the relapse has always been viewed as a function of the illness process which, in some way or another, was being damped down or held at bay by medication. Nobody has ever produced a satisfactory explanation of how this might happen. But there are psychopharmacological models that can explain the appearance of abnormal mental states following discontinuation of drugs. Most people are familiar with the minor withdrawal symptoms such as anxiety, insomnia and mild depression which occur when someone stops minor tranquillisers after taking them for any length of time. Could the ‘relapse’ of schizophrenic symptoms on stopping neuroleptics in reality be a neuroleptic withdrawal syndrome rather than a true relapse of the illness? The model of Tardive Dyskinesia is relevant here. This condition is thought to arise because of an increased number and sensitivity of dopamine receptors in the striatum. Presumably there is no reason why such a mechanism should not occur in the meso-limbic system, the pathway thought to be important in relation to the symptoms of acute schizophrenia. If this is so, then reduction or cessation of neuroleptic medication in the presence of increased receptor sensitivity in this area should result in a relative increase in dopaminergic transmission in this system with the appearance of acute psychotic symptoms. In these circumstances the occurrence of ‘relapse’ is hardly an intrinsic feature of the disease process, but an iatrogenic (brought about by medical intervention) phenomenon arising as a consequence of the effects of dopaminergic blocking agents on the brain.
There is evidence that this indeed may be the case. A few years ago, Chouinard and Jones (1980) published a paper in the American Journal of Psychiatry, which, in its day, was widely quoted as evidence in support of the dopamine theory of schizophrenia. In it, they described in detail the drug treatment histories of ten patients treated with neuroleptic drugs. In each case a recurrence of psychotic symptoms was noticed almost immediately following reduction or discontinuation of medication. This runs contrary to expectation. There is nothing in our knowledge of the course of schizophrenia that would lead us to expect a sudden reappearance of symptoms on discontinuing medication. All the subjects had had lengthy periods of exposure to neuroleptic, and in addition, most had evidence of tardive dyskinesia. The dyskinesia became more apparent on discontinuation of medication, and with the appearance of psychotic symptoms. The authors note that:
‘An implication of neuroleptic-induced mesolimbic supersensitivity is that the tendency toward psychotic relapse in such patients is determined by more than just the normal course of the illness.’ (Chouinard & Jones, 1980: p. 16)
Further evidence in support of the supersensitivity psychosis comes from one of the most thorough reviews of neuroleptic withdrawal ever made. Gilbert et al (1995) reviewed 66 studies of neuroleptic withdrawal in schizophrenia involving over 4,000 patients. She found that amongst other things ‘relapse’ was more likely to occur in people who had been taking higher doses of neuroleptic medication prior to discontinuation, or in people who had experienced rapid discontinuation (over two weeks) as opposed to gradual discontinuation (over eight weeks). These are exactly the features we might expect to find in a disturbance that occurs simply as part of a drug withdrawal state.
Of course, no research study has seriously addressed the possibility that symptom recurrence in schizophrenia on stopping neuroleptics is effectively a drug withdrawal syndrome. To examine this possibility you would have to examine the symptoms of a large number of people in their first episodes of schizophrenia, and then compare these symptoms in great detail with those seen in the same people after discontinuing neuroleptics. Such a study has yet to be performed. It is interesting to note how the drug withdrawal explanation for ‘relapse’ has been overwhelmingly neglected in the psychiatric literature. Descriptions of drug withdrawal states dominate clinical accounts of the effects of psychoactive drugs and substances. All doctors are familiar with the phenomenon of drug withdrawal syndrome, and although Chouinard’s work has been in the literature for 16 years, this explanation for the link between relapse and medication has been glossed over. The implications of this are too frightening for psychiatry to contemplate. It means that for many patients ‘relapse’ may be iatrogenic, a consequence of drug treatment and not an inevitable part of the process of a disease.
Leudar, I. & Thomas, P. (1999) Conversations with Voices. Forthcoming, Routledge: London.
There are enormous difficulties in undertaking a study such as this. For example, there are different ways of defining relapse. For some people relapse is defined simply as the reappearance of psychotic symptoms. For others it is necessary for you to be readmitted to hospital. This makes it very difficult to compare the outcome in different studies.