Consensus was reached on the five following points: (1) “Treatment resistance” is a term which is widely used, but not adequately defined. In particular, definitions of “responders” and “non-responders” need to be more precise. (2) There is a need to extend the assessment of the results of treatment to include factors such as quality of life and the subjective evaluation by the patient and relatives. (3) Clozapine (Clozaril®/Leponex®) has become the standard for atypical antipsychotic drugs. The adjective “atypical” is not an exact descriptor. The vague nature of the term leads to confusion with other drugs that are putative atypical antipsychotic agents but have different properties. D2 dopamine receptor antagonism is the single factor common to all antipsychotic drugs. Antagonism at a specific D2-like receptor may underlie the atypical antipsychotic actions of clozapine. Drugs described as being atypical antipsychotics share a relatively low incidence of extrapyramidal side effects. (4) Clozapine is the only drug that produces antipsychotic (therapeutic) effects at doses that do not result, in significant extrapyramidal side effects. This may be due to the fact that antipsychotic efficacy is obtained with clozapine in doses that result in significantly less striatal D2 occupancy. Other factors including concomitant 5-HT2 antagonism, may contribute to the lack of motor side effects observed with clozapine treatment. (5) Clozapine is the only drug currently well proven in therapy-resistant schizophrenia, and there is a case for its earlier use in patients who can be anticipated to become therapy resistant. Other antipsychotic therapies remain unproven in this group of patients.

This paper argues for a broader definition of treatment refractoriness that would include patients with partial drug responses, a relatively recent onset, less dramatic symptoms and intolerance to conventional drugs. The drug trials that define their responsiveness should be carefully conducted over a prolonged period of time, but need not include high doses or supplemental medications. If the consensus conference can agree on a definition of responsiveness, it will provide important guidance to clinicians as well as researchers.

The treatment of nonresponsive schizophrenia poses perennial problems. The situation is made yet more complex when the great variety of patient types, their social background, family circumstances and psychosocial requirements are taken into account. The comparative lack of response to neuroleptic therapy seen in some patients, especially those with minimal structural brain changes, further complicates the problem, provoking—as such therapy may—more EPS than are normally seen in patients without more obvious physical changes in brain structure. Until recently, the treatment of nonresponsive schizophrenia has tended to involve switching from one neuroleptic (usually a “highpotent” agent) to another of different chemical structure. However, as these “typical” neuroleptics occupy D2 receptors, the rationale for using a less conventional antipsychotic, such as clozapine, is increasingly appreciated in the management of the treatment-refractory schizophrenic.

The outcome of treatment in schizophrenia is best considered as a multidimensional construct. This report lists thirteen categories of outcome measures that have been studied in schizophrenia research, among the most important of which are psychopathology, cognitive function, interpersonal social function, quality of life, extrapyramidal function, the suicide tendeccy and the need for hospitalisation. Clozapine is effective in achieving improved outcome in all these measures in a high percentage of neuroleptic-resistant patients. Ninety-two of 180 (51.1%) clozapine-treated neuroleptic-resistant patients achieved a decrease of 20% in Brief Psychiatric Rating Scale total scores at six weeks using the 0–6 scaling system. Cognitive function contributes importantly to overall function for schizophrenia. Clozapine has been shown to improve some types of cognition in schizophrenia, especially semantic memory and some tests of executive function. Clozapine has also been shown to decrease the incidence of suicide and hospitalisation. The task of deciding which therapies to use in schizophrenia should encompass the spectrum of benefits of each therapy, their risks and costs, in relation to all alternatives.

Akathisia, dystonia, dyskinesia and parkinsonism, the four main categories of neuroleptic-induced extrapyramidal syndromes (EPS), represent major disadvantages in antipsychotic therapy. In vulnerable patients, acute EPS may progress into potentially irreversible forms such as tardive dystonia and tardive dyskinesia. In the psychiatric clinic, these EPS are often insufficiently recognised or permitted to exist without treatment. In order to ensure a better EPS diagnosis, a simple examination procedure is described. EPS rating scales may serve as an aid in this process. Guidelines are given to prevent and treat EPS. Thus, EPS are best prevented by a course of neuroleptic medication involving as little antidopaminergic D2 effect as possible, including the use of the lowest effective dose (sometimes obtained by addition of a benzodiazepine or carbamazepine) and with antipsychotic drugs which produce low D2 receptor blockade. Treating EPS also consists of using the lowest effective dose and antipsychotics with a low D2 dopamine receptor occupancy. At present, clozapine is the only drug that produces antipsychotic benefits at doses that cause far less D2 receptor antagonism in the basal ganglia of the brain than that seen with standard neuroleptics; however, newer drugs, such as olanzepine, seroquel and sertindole, are on the way.

Antipsychotic drugs are used in a range of schizophrenic patients from the acutely disturbed to the chronically inert. Acute psychotic breakdowns usually involve secondary phenomena such as anxiety, depression and hostility superimposed on a worsening of the psychotic illness: a threshold of arousal is exceeded. This sets up a vicious circle which requires a major therapeutic effort to interrupt. Unfortunately, patients still relapse despite apparently adequate medication, suggesting only partial protection. Relapse is influenced by drug-related factors such as compliance and substance abuse, psychological factors such as cultural context, family dynamics and life events, and illness features such as poor insight. Treatment strategies, such as low-dose and intermittent medication, have proved disappointing. The most pragmatic model is the classic stressor-vulnerability model, but this does not explain the enhanced efficacy of clozapine.

The mechanisms of action which account for the effectiveness of clozapine as a pharmacotherapy for the treatment of neuroleptic non-responders and neuroleptic intolerant schizophrenic subjects remain elusive. We review recent data concerning the actions of clozapine in laboratory animals, and discuss the likely sites of action of clozapine and the receptors through which clozapine acts. We suggest that actions at dopamine D2 receptors in the caudate nucleus and putamen underlie the extrapyramidal side effects of conventional neuroleptics. In contrast, we propose that clozapine acts in the prefrontal cortex, specifically targeting an as yet unidentified DA receptor of the D2 family, to exert therapeutic actions in neuroleptic non-responders. We suggest that the ability of clozapine to augment extracellular dopamine levels in the prefrontal cortex may represent a key mechanism contributing to the therapeutic effects of this drug, and suggest some alternative approaches which might be expected to result in effects similar to those of clozapine.