Show Summary Details
Page of

Treatment of Schizophrenia 

Treatment of Schizophrenia
Treatment of Schizophrenia
Page of

PRINTED FROM OXFORD MEDICINE ONLINE ( © Oxford University Press, 2021. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Medicine Online for personal use (for details see Privacy Policy and Legal Notice).

date: 24 September 2021

Interest in psychological treatments for schizophrenia has increased in recent years, particularly in Europe and Australia, driven by understandable patient dissatisfaction with purely pharmacological approaches. The recognition that 40% of patients do not achieve symptom resolution with drug treatment (Kane, 1996) has added impetus to the search for alternatives and adjuncts. Unfortunately, the paucity of data concerning the application of such approaches to adolescents suffering from schizophrenia means that, at least for the time being, inferences have to be drawn mainly from studies of adult populations.

Nevertheless, the similar lack of data regarding efficacy and safety of antipsychotic drug treatment in adolescents and the observation that adolescents may be especially sensitive to the adverse effects of typical antipsychotics (e.g., extrapyramidal side effects; Lewis, 1998) and clozapine (e.g., neutropenia and seizures; Kumra et al., 1996) mean that there is a real need for alternative or supplementary interventions. Indeed, one could hypothesize that psychological treatments might be more effective in adolescents than in adults. This is a group with a greater degree of neural plasticity and a still evolving personality, and who are especially likely to have an ongoing system of support in the form of family and educational input. Furthermore, the latter presents opportunities for early detection. Alongside early treatment with antipsychotics, there is potential for psychotherapeutic interventions to lessen the impact of positive symptoms, improve coping strategies, and potentially to reduce the cognitive deficits, which so impair psychosocial function.


Like those who go on to develop other serious mental disorders, individuals who develop schizophrenia in later life often, but not invariably, demonstrate interpersonal and emotional difficulties during childhood and adolescence. Cannon et al. (2002) reported that schizophrenia was specifically predicted by the presence of deficits in receptive language, neuromotor function, and cognitive development between 3 and 11 years of age.


Because such abilities are already observed and assessed, to an extent, as part of a child's schooling, there is potential for the development of predictors and the identification of targets for psychological intervention. Studies examining the ability of teachers and other educational professionals, in day-to-day contact with adolescents, to predict future sufferers of schizophrenia do show some statistical power, but this is at the cost of many false positives and even more false negatives (Isohanni, et al., 2004) unless restricted to high-risk populations (Kravariti, Dazzan, Fearon, & Murray, 2003). One possible exception was the study by Davidson et al. (1999) linking the Israeli Draft Board Registry with the National Psychiatric Hospitalization Case Registry. Adolescent boys, aged 16 to 17 years, underwent preinduction assessments to determine suitability for military service. Those admitted to a psychiatric hospital with schizophrenia 4 to 10 years later were matched with control individuals from their school class at the time of original assessment. Identified predictors for schizophrenia in the male adolescents included deficits in social and intellectual functioning and organizational activity. The predictive model derived by the authors had 75% sensitivity, 100% specificity, and a positive predictive value of 72%. However, as highlighted in a commentary by Jones and van Os (2000), the predictive power of this model was achieved by excluding from the sample those school classes without individuals who later became schizophrenic.

So far the best prediction has come not from teachers but from a psychiatric interview. In the Dunedin study (Poulton et al., 2000), 11-year-old children were asked about experiences of quasi-psychotic symptoms. The questions were as follows:

  • 1. Do you believe in mind reading or being psychic? Have other people ever read your mind?

  • 2. Have you ever had messages sent just to you through the television or radio?

  • 3. Have you ever thought that people are following you or spying on you?

  • 4. Have you heard voices other people can't hear?

  • 5. Has something ever gotten inside your body or has your body changed in some way?

Those who answered positively to one of the five questions or possibly positively to two were 16 times more likely to develop a schizophrenia-like psychosis by age 26 years. Teachers or school nursing staff could be advised of the value of such questions. However, little work aiming at primary prevention (i.e., preventing progression from premorbid abnormalities to prodrome) is in progress.

Early Treatment Projects

Falloon initiated an early intervention for “prodromal” symptoms in adolescents and adults, in the form of the Buckingham project, a “shared care model” between primary and secondary care (using low doses of medication, interventions designed to reduce stress, psychoeducation, and follow-up for 2 years after the symptoms had occurred). He claimed that this reduced the annual incidence of first-episode psychosis from 7.4 per 100,000/year, as measured by the same group in 1989, to 0.75 per 100,000/year during the 4-year study period (Falloon, Kydd, Coverdale, & Laidlaw, 1996).

Unfortunately, adolescents developing schizo-phrenia suffer the same delays as their older counterparts, often not receiving diagnosis or treatment for a prolonged period. Consequently, a number of projects have been developed to reduce this unnecessary period of untreated psychosis.

Projects such as the Personal Assessment and Crisis Evaluation (PACE) clinic in Melbourne, Australia, and the early Treatment and Intervention in Psychosis (TIPs) project in Norway aim at secondary prevention (i.e., preventing progression from prodrome into syndrome) in adolescents and adults. Such programs are motivated by the belief that the chronicity of schizophrenia may develop in the early stages of the illness and that long-term outcomes may be linked to the duration of untreated psychosis, or DUP that is, the time period between onset of symptoms and initiation of treatment. However, the latter hypothesis remains theoretical. Norman and Malla (2001) reviewed the concept of DUP and while they confirmed that there was evidence to suggest a relationship between DUP and the ease with which first remission of symptoms is achieved, they could not find evidence to support a link with disease progression. Although the concept of DUP focuses on initiation of pharmacological treatment, de Haan, Linszen, Lenoir, de Win, and Gorsira (2003) suggest that delay in initiating intensive psychosocial treatment may have similar implications for outcome, particularly in relation to negative symptoms at follow-up.

The TIPS project is a prospective clinical trial that started in Norway in 1997, comparing an experimental sector with two other control sectors (age range 15 to 65 years). The experimental sector developed a system for early detection and also established a comprehensive information, service, and education program aimed at both the general public and professionals involved in health care and education (Johannessen et al., 2001). In the 2 years after the initiation of the TIPs project, mean DUP decreased from 1.5 years to 0.5 years.

The PACE was established in Melbourne, Australia, in 1994; the aim was to evaluate the prodromal phase, develop interventions that prevent further deterioration and maximize function, and set up a clinical service to identify and engage young people experiencing potential early psychosis. Preliminary results, which are further discussed in Cognitive Behavior Therapy, below, show that early intervention in the prodrome can at least delay onset of first-episode psychosis.

It seems likely that the wider-ranging public and educative measures occurring in the TIPS project and PACE clinic are vital to any health service initiatives or collaboration, especially in view of adolescents' poor primary care attendance. It is important to note, however, that while a number of the early-intervention services cover the adolescent age range, their goal is actually to intervene in an early stage of the disorder rather than to specifically target those who develop schizophrenia in adolescence.


Cognitive Behavior Therapy

Cognitive behavior therapy (CBT) addresses problems in the here-and-now by targeting dysfunctional thoughts and behaviors within a collaborative therapeutic relationship. Efficacy of CBT and acceptability by those with disorders such as depression and anxiety have been demonstrated (Beck, Sokol, Clark, Berchick, & Wright, 1992; Kovacs, Rush, Beck, & Hollon, 1981). In support of the growing (but relatively underevaluated) practice of CBT in adults with schizophrenia, Rector and Beck (2002) suggest that since the inferential errors and faulty logic in hallucinations and delusions are similar to those seen in other disorders, CBT should also work in schizophrenia. They describe CBT for psychosis as an active, structured therapy, usually of 6 to 9 months duration, given individually.

Adolescent Studies

Unfortunately, we could not find any reports of the use of definitive CBT in specifically adolescent individuals. However, data are beginning to accumulate from the early-intervention programs, reporting on the use of CBT in early-onset psychosis. McGorry et al. (2002) conducted a randomized controlled trial (RCT) with 14-to 30-year-olds at the Early Psychosis Prevention and Intervention Centre, Melbourne, which is associated with the PACE clinic discussed earlier. This study compared interventions designed to reduce the risk of progression to first-episode psychosis, in a clinical sample aged 14 to 30 years, termed ultra-high risk (first-degree family history of schizophrenia and subthreshold symptoms). The interventions comprised a needs-based intervention and a specific preventive intervention (low-dose risperidone and CBT) for 6 months with assessments at baseline and 6 and 12 months, using a defined threshold outcome rather than a formal diagnosis of schizophrenia. Ten of 28 in the needs-based intervention versus 3 of the 31 in the specific intervention had reached the defined outcome by the end of treatment; however, there was no significant difference at 6-month follow-up. When the data were assessed, taking into account drug adherence, a significant difference was found between the fully adherent specific intervention group and the needs-based group. It seemed that the specific intervention delayed onset.

Adult Studies

Dickerson (2000) reviewed all studies investigating CBT for schizophrenia in adult populations (although age data are not supplied) between 1990 and 1999. She examined the available data for seven different CBT approaches, some focusing on acute psychosis, others on persisting positive symptoms. Her conclusion was that there were some CBT strategies that reduced positive symptomatology, especially in individuals with clearly defined symptoms that they themselves viewed as problematic. The most beneficial outcome appeared to be in reducing conviction in and distress about delusions; there was little evidence to suggest that CBT was efficacious in negative symptoms or social functioning. The overall superiority of CBT was reduced when the control condition was matched for therapist input.

Even more disappointing, a Cochrane meta-analysis of CBT in schizophrenia (Cormac, Jones, & Campbell, 2002) found no evidence that CBT, in addition to standard care, reduces the relapse and readmission rate in the short or longer term (1 year) any more than standard care alone. Moreover, there was no overall difference between CBT and supportive psychotherapy, with respect to relapse rate or improvements in mental state.

Pilling et al. (2002) conducted a systematic review of RCTs of CBT (and family interventions, social skills training, and cognitive remediation). The trials reviewed compared CBT with supportive therapies and standard care in predominantly chronic psychosis. There was no evidence for increased effectiveness of CBT during treat ment, although CBT showed a clear advantage over the comparison treatment at follow-up when measured continuously in terms of “important improvement.” This superiority persisted for up to 18 months after treatment. Furthermore, CBT groups had lower dropout rates. It was not possible to identify any particular responder characteristics or the optimum frequency or length of treatment.

Although the Study of Cognitive Reality Alignment Therapy in Early Schizophrenia (SoCRATES; Lewis et al., 2002) was conducted among individuals with an average age of 27, it is one of the more relevant adult studies to this discussion because the patients recruited were in the early phases of their illness. The RCT compared CBT with routine care, supportive counseling and routine care, and routine care alone, all for 5 weeks duration only, in individuals suffering their first or second episode of schizophrenia. Use of CBT showed only transient advantages over the other two intervention conditions in speeding remission from acute symptoms in this group of individuals.

In summary, the adult literature suggests that, although CBT might have some beneficial effects, questions such as defining the length of treatment required, which patients would benefit, and at which stages in their illness it should be implemented still remain unanswered. Until these questions are answered in adult populations, it would be foolish to commence large trials in adolescent individuals, especially without adequately defined aims. This caution should not prevent CBT techniques from being borrowed in part in the development of specific interventions targeted at the special needs of adolescents with schizophrenia.

Cognitive Remediation Therapy

Deficits in cognitive function such as working memory, attention, and executive functioning are core features of schizophrenia. Cognitive remediation therapy (CRT) aims to teach people thinking skills. More specifically, it uses material that is not personal to the individual and targets the specific domains affected in schizophrenia. Cognitive remediation therapy can be characterized into three generic approaches (Bellack & Brown, 2001):

  • 1. Practice and brief training on neurocognitive tests or computer tasks to improve a single domain of functioning

  • 2. Repetitive practice on a battery of computer tasks aimed at multiple domains

  • 3. Strategies to improve cognitive functioning in general by increasing self-confidence, interest, and initiative

The assessment of efficacy of CRT has again been hampered by methodological constraints, in particular, the heterogeneity of intervention packages used and then inappropriately compared. Wykes et al. (2001) have made attempts to reduce this phenomenon by describing a typology for classification of methods.

Adolescent Populations

Techniques specifically targeted at cognitive differentiation, attention, memory, and social perception are being evaluated in adolescent populations by Rund and colleagues in Oslo. Ueland and Rund (2004) carried out an RCT comparing the effects of psychoeducation and cognitive training with those of psychoeducation in a small group of adolescents with early-onset psychosis. They did not detect any significant between-group differences in any of their treatment scores, although there were some specific significant improvements in visual long-term memory and early visual information processing and Brief Psychiatric Rating Scale (BPRS) scores, limited to the remediation group. This group has also focused on the remediation of more specific cognitive deficits. Ueland et al. (2004) assessed the effect of enhanced instructions and contingent monetary reinforcement on attentional skills. A group of adolescents with early-onset psychosis received the span of apprehension performance (SPAN) at baseline, three times as an intervention, and then after testing and at 10-day follow-up. Improvements in performance were evident at the end of the intervention, diminishing slightly after testing but recovering at 10-day follow-up.

Adult Populations

Research into CRT in adults, based on the third category listed above, has recently shown an effect on memory durable to 6 months of follow-up, which, if large enough, gives rise to associated improvement in social performance (Wykes et al., 2001). We await with interest the ongoing study this group is conducting of patients between 16 and 21 years of age. Furthermore, a study involving “cognitive enhancement therapy” in early-course patients, some of whom will be within the adolescent age range, is being undertaken by G. Hogarty's team in Pittsburgh (personal communication).

Interpersonal Psychotherapy

Very little evaluative work has been published in the area of interpersonal therapy (IPT) since the 3-year trial conducted by Hogarty et al. (1997). Although this study was aimed at reducing the “late relapse” observed to occur in the second year after psychotherapeutic intervention, and was conducted in a mixed adult and adolescent cohort (16 to 55 years old), it warrants discussion because its conclusions seem particularly pertinent to the adolescent population. This study examined relapse and noncompliance in outpatients with schizophrenia and schizoaffective disorder in two concurrent trials. One trial concerned individuals who resided with their families who were randomly assigned to personal therapy, family therapy, personal therapy and family therapy, or supportive therapy; the other trial studied those individuals who lived alone who were randomly assigned to either personal or supportive therapy.

Personal therapy occurred in three stages, and although it focused on internal, personal responses to stress, and not the regulation of external triggers, it did not use symbolic interpretation or analysis of unconscious factors. It aimed more specifically to identify and manage the “affect dysregulation” that might mediate relapse or inappropriate behavior. All patients were on the minimum effective dose of medication. The overall rate of relapse over the 3-year period was 29% over all groups, which was lower than expected. For individuals residing with their family, the group receiving personal therapy relapsed less than the other groups, although this was only significant for family intervention. For individuals living alone, personal therapy significantly increased relapse rates, compared to those for supportive treatment, even though the latter was particularly rich in its provisions. The authors suggested that the group living away from their family may not have reached a stable independence; they may have been too distracted to prioritize the therapy or the intervention itself may have overloaded them. Perhaps this conclusion should serve as a more global caution: full assessment of levels of independence and subjective perceptions of stability may be required before initiating any form of psychotherapy.

Social Skills

Social skills training is a structured educative program that involves modeling, role-play, and reinforcement (Bellack & Mueser, 1993). It is based on the hope that the improved social skills generalize to real-life situations and might even improve symptomatology and reduce relapse. These interventions are targeted at the profound impairments in social functioning that characterize schizophrenia and affect life in the workplace, family, and wider community. There are three forms of social skills training as detailed by Bellack and Mueser (1993): the basic model, the social problem-solving model, and the cognitive remediation model (considered in the previous section).

The basic model involves breaking down complex social interactions into smaller elements. The patients are therefore taught the steps and then the combined elements by means of role-play and areas such as self-care, medication management, and conversation are targeted. Bustillo, Lauriello, Horan, and Keith (2001) reviewed reports from 1996 to 2001 and found that the basic model was repeatedly efficacious in improving social skills and that this effect continued for up to 12 months. However, there was not much to suggest that this treatment affected overall social performance. Evidence from studies of the problem-solving approach suggests modest benefits on very discrete areas of social functioning that appear to have some durability.

Although such research seems old-fashioned and potentially out of sync with the dilemmas and challenges facing an adolescent suffering from schizophrenia, the approach may have merits. For a child, the transition into adolescence and then adulthood is a difficult process requiring support and structure. Development of schizophrenia during this process can disrupt personal development in an infinite number of ways. Therefore, there is a real need for research among groups of adolescents with schizophrenia into developing ways to teach them social skills that generalize to real life.

Family Therapies

Successful family therapies have psychoeducation at their heart, taking the form of a collaborative respectful relationship with the family, provision of information, and teaching of family members less stressful ways of communicating and solving problems. It seems intuitive that family interventions, in a nonspecific sense, would be particularly beneficial for adolescents with schizophrenia, especially those who remain still dependent on their parents.

The need for intervention is supported by reports that children with schizotypal personality disorder or schizophrenia and the parents of such children tend to show increased rates of thought disorder during direct family interactions (Tompson, Asarnow, Hamilton, Newell, & Goldstein, 1997). Some reports claim that the parents also show increased communication deviance (an index of difficulties associated with a failure to establish and maintain a shared focus of attention; Asarnow, 1994). It is not possible to know at this stage whether this represents a shared genetic vulnerability to psychosis or a parental response to their child's illness.

Additional support comes from the work on expressed emotion (EE). Although this was originally investigated in adult populations, most data assessing family interventions in adolescents are based on EE. The concept of EE, which encompasses critical comments, hostility, and overinvolvement, arose out of a body of research focused on the effect of the family environment on the maintenance of schizophrenia and other severe mental disorders (Brown, Monck, Carstairs, & Wing, 1962). Family interventions evolved, again targeted at adults, that aimed to reduce high EE and thus reduce relapse rates (Leff, Kuipers, Berkowitz, Eberlein-Vries, & Sturgeon, 1982).


The applicability of therapies directed at high EE in adolescent schizophrenia is hampered by three problems. First, adolescent families have less EE or different EE. Asarnow, Tompson, Hamilton, Goldstein, and Guthrie (1994) found relatively low parental EE (compared to families of adults with schizophrenia) when measuring criticism and overinvolvement on the Five-minute Speech Sample Expressed Emotion (FMSS-EE) scale. Twenty-three percent of families of adolescents with schizophrenia or schizotypal personality disorder were rated as having high EE, compared to 44% of families of adults with schizophrenia in a similar study by Miklowitz et al. (1989).

The second problem in using EE-related therapies for adolescents is the lack of stability over time and lack of response to treatment. Lenior, Dingemans, Schene, Hart, and Linszen (2002) conducted a longitudinal study to analyze the stability of parental EE in individuals with recent-onset schizophrenia, aged 15 to 26 years, in the 8 years following discharge from two interventions from inpatient care: community care alone and community care with additional family intervention according to the model of Falloon, Boyd, and McGill (1984). The families were stratified according to high and low EE before allocation and EE was measured over the follow-up period using the FMSS. According to these measurements, EE did change over time, although the study failed to detect any overall treatment effect on EE levels. In addition, this group found no intervention effect on the number of months of psychosis during 5-year follow-up.

Nugter, Dingemans, van der Does, Linszen, and Gersons et al. (1997) studied individuals, aged 16 to 25, with recent-onset schizophrenia and related disorders who were randomly allocated to individual treatment with or without a family intervention (modeled on Falloon et al., 1984). At the end of treatment (1 year), there were no significant between-group differences in EE (as assessed by the FMSS). There were no detectable relationships between EE and relapse, except that in the individual group, changeable EE (in whichever direction) was correlated with relapse rate.

The third difficulty with using EE in treatment for adolescents is its lack of specificity to schizophrenia. A meta-analysis of EE-outcome relationships in mental disorders (Butzlaff & Hooley, 1998), though confirming EE as a significant predictor of relapse in schizophrenia, found significantly larger effect sizes for EE in mood and eating disorders. This study did not specify age ranges covered by the reviewed studies. Thus, attempts to evolve therapies targeted at underlying problems, particularly in younger populations, have been complicated.

Linszen et al. (1996) found that adding a behavioral family intervention (after Falloon et al., 1984) to an individual psychosocial intervention in patients aged 15 to 26 years made no difference to rates of psychotic relapse in the 15 months following first-episode psychosis. In fact, Linszen's group found a near-significant increase in relapse rate in low-EE families subjected to the intervention, possibly because of the families' perception of the therapy as artificial, critical, or interfering in their reactions to their offspring's illness. However, during that 15-month intervention period, the relapse rate was 15%, which suggests that early intervention improved outcome. The cohort was then referred for care by other agencies and at 5-year follow-up, the low relapse rate had not been maintained. The authors suggest that sustained intervention above and beyond regular services might be required to improve outcome in the longer term. This result, taken in conjunction with the increase in relapse of low-EE families, calls into question the value and possibly the ethics of using this intervention in early psychosis.

Asarnow et al. (2001) concluded that although available data support the use of family interven-tions in the treatment of adolescents with schizophrenia, it is not yet possible to determine which model is most effective. There is, however, an encouraging shift from reducing putative risk factors to empowering and channeling this resource. We concur with this viewpoint but would welcome more work into alternative means of monitoring response to family therapy.

Adult Studies

The preliminary results from the multicentered National Institute of Mental Health (NIMH) Treatment Strategies Study (Schooler et al., 1997), which compared variable medication strategies in conjunction with a “supportive family management” (psychoeducational workshop for relatives with a monthly support group for 2 years) and “applied family management” (which included the former and an intensive at-home family intervention based on the Falloon et al. behavioral program), suggest that the latter, more intensive program does not yield better results and does not permit the use of lower or intermittent medication regimes.

McFarlane et al. (1995) conducted a pilot study in which psychoeducation administered in single-family groups was compared with that in multiple-family groups for individuals aged 18 to 45 years. There were lower relapse rates (12.5% at 12 months and 25% at 24 months) in the multiple-family groups (compared to 23.5% and 47.1%, respectively). Both of these programs resulted in lower relapse rates than those achieved in a multiple-family program without a psychoeducational model.

In Pilling et al.'s (2002) systematic review, family intervention data were also analyzed. The mean age of studied individuals was 31 years, and the mean number of admissions per individual was 2.7. All family interventions (single and multiple) were more effective than the control condition at reducing relapse in the first 12 months of treatment, especially when it consisted of standard care. Only single-family interventions reduced readmission during this time. At 1 to 2 years only, single-family interventions were still reducing relapse, although all treatments were effective at reducing readmission. In- terestingly, all family interventions studied had higher rates of treatment compliance to both the family intervention and concurrently prescribed medication. It is important to note, however, that not all the studies included in this review used a supportive individual program as the control condition. Some control conditions actually comprised a family intervention itself, as noted in a response to the review (Bentsen, 2003).

The Family to Family Education Program developed by the National Alliance for the Mentally Ill, detailed on their Web site (, involves a highly structured program conducted by trained family members for 2-to 3-hour sessions over 12 weeks. Participants report decreases in family members' “worry and displeasure” and “subjective burden,” with increased empowerment and knowledge and improved coping strategies. This program is cheap, popular with family members, and can be widely disseminated, thus aiding implementation.

Compliance Therapy

Compliance therapy is a brief, pragmatic intervention in which cognitive behavioral techniques, very closely linked with motivational interviewing, are used to focus on improving treatment adherence. This therapy evolved out of initial work in programs using psychoeducational and behavioral techniques (e.g., Eckman et al., 1992), and its further development was encouraged by the UK National Health Service, which declared noncompliance a research priority. It was tested in a pilot study, modified, and then used by Kemp, Hayward, Applewhaite, Everitt, and David (1996) in an RCT of inpatients with psychotic illnesses with follow-up to six months. Eighteen-month follow-up was reported (Kemp, Kirov, Everitt, Hayward, & David, 1998) using an expanded sample. The therapy itself takes place over four to six sessions of 20 to 60 minutes each, approximately biweekly and involves the following:

  • 1. Review of the individual's treatment history and his or her views and understanding of the illness and treatment

  • 2. Exploration of symptoms and side effects and thus evaluation of pros and cons of treatment

  • 3. Discussion of stigma

Despite the fact that the therapy was biased to encourage only positive attitudes to treatment, the therapy also aimed to reframe the use of medication as a decision, freely chosen to enhance quality of life, referred to as an “insurance policy” or “protective layer.” Immediately after the intervention, the compliance therapy had significantly improved insight and compliance, compared to a supportive counseling matched for therapist time, with the same therapists. This effect was maintained at 6-month follow-up with a 23% difference between groups, with those with higher IQ achieving better results. It was not possible to determine whether these gains resulted in increased function or diminished relapse rate. The 18-month follow-up confirmed that compared to the control condition, compliance therapy improved compliance and insight in the intervention group by 19% and improved global functioning, especially as time progressed. It had no overall effect on improving symptomatology or reducing time spent in the hospital over the follow-up period. The positive effects on compliance were not replicated by O'Donnell et al. (2003) in an RCT of adult inpatients with schizophrenia. Compliance therapy, administered according to Kemp et al. (1996, 1998), conferred no advantage in compliance, symptomatology, or overall function outcomes at 1 year post-therapy, compared to a control condition of equivalent duration with nonspecific counseling. Furthermore, one struggles to see how such a therapy could exert such a persuasive effect in a teenager in the throes of puberty and a battle for independence. Yet again, there is a real need for adaptation of such techniques to adolescent populations, followed by RCTs evaluating their effect.

Illicit Drugs

Although many illicit substances are used by individuals with psychosis, our discussion of such substances will be limited to cannabis because it is by far the most common illicit substance used by adolescents.


There is much evidence to suggest that cannabis consumption among those already schizophrenic has a detrimental effect. The strength of this effect is not yet certain, however (Johns, 2001). Unfortunately, there is even less evidence concerning what can be done about it.


Zammit, Allebeck, Andreasson, Lunberg, & Lewis (2002) conducted a further analysis of the Swedish conscript data for 1969–1970 (>97% of the country's male population aged 18 to 20). They separated the cohort into those with psychosis onset < 5 years and those with onset > 5 years after data collection to rule out the possibility of prodrome at the time of conscription. They found a significant dose-related relationship between cannabis and increased risk of developing schizophrenia, strongest in those who had onset of psychosis within 5 years of conscription and present in both those who used cannabis alone and cannabis with other drugs. Similar results were obtained when only those with onset >5 years were analyzed.

This study could not determine, however, whether cannabis use in adolescence was a result of preexisting psychotic symptomatology or a cause in itself for psychosis. This issue was addressed by the Dunedin study (Arsenault, Cannon, Witton, & Murray, 2003), a longitudinal, prospective study that assessed preexisting psychotic symptoms at age 11, drug use at 15 and 18 years of age, and psychiatric outcome measures at 26 years. This study showed that adolescents who used cannabis at 15 and 18 years of age had significantly more symptoms of schizophrenia than controls at 26 years of age. Furthermore, these results remained significant when quasi-psychotic symptoms at 11 years were controlled for. Use of cannabis at 15 years increased adult risk of schizophreniform disorder by a factor of 4, although these results did not remain significant when psychotic symptoms at age 11 were controlled for. Thus cannabis use, especially earlier in life, increases the risk of schizophrenia symptoms. This effect is not explained by use secondary to psychosis and this effect appears to be specific to cannabis use. Although cannabis is not thought to be necessary or sufficient to cause the onset of psychosis, it is estimated that 8% of schizophrenia cases in New Zealand could be prevented by the cessation of cannabis use in the general population (Arsenault et al., 2003).

Therefore, in addition to discouraging the legalization and supply of cannabis, mental health initiatives should educate adolescents on the previously unrecognized risks of cannabis misuse. Such initiatives could occur concurrently with the campaigns already implemented by early-intervention services.


Studies attempting to assess efficacy of psychological treatments have been hampered by a multitude of methodological problems. Often very different versions of a treatment model are described as one and the same. They are often applied to heterogenous populations (e.g., different ages or illness stages) and compared with a control condition unmatched for therapist time and attention.

Debate also continues about the best way to assess the outcome of such interventions. Bellack and Brown (2001) feel that judging psychosocial treatments on the same outcome measures as pharmacological intervention might not be appropriate. Rather than evaluating therapies on whether they reduce symptoms, induce remission, or prevent relapse, they recommend focusing on their effect in reducing impairments in social role functioning and improving overall quality of life and treatment adherence. So they recommend rehabilitation rather than treatment, taking into account the confounding effect of cognitive deficits and the need for newly learned skills to generalize to real life in the community. They liken psychosocial therapies to the use of Braille in visually impaired individuals.

Outcome studies in schizophrenia tend to focus on categorical measures, such as hospital admission, and on professional observations of relapse, symptomatology, or cognitive impairment. Very few assess psychosocial outcome or quality of life, especially as rated by users of services and their caregivers. Such measures of outcome are particularly important for adolescent populations. Lay, Blanz, Hartmann, and Schmidt (2000) conducted a 12-year follow-up study of 96 consecutively admitted individuals with schizophrenia (aged 11 to 17). Of the 68% reassessed at 12 years, 66% had serious social disability, which was predicted by severity of positive symptoms in the early stages and by admissions numbering more than two; 75% were financially dependent. Jarbin, Yngve, & Von Knorring (2003) conducted a 10-year follow-up of adolescents (age <19 years) who were diagnosed with first-episode early-onset psychosis in the 1980s and early 1990s; 79% of those with early-onset schizophrenia spectrum disorders suffered a chronic course with poor outcome.

The collaborative and empowering nature of many of the psychotherapeutic options, though incompletely evaluated and not always available, switches our focus from such sobering outcomes and inflicted choices to a more patient-driven framework. It is encouraging to see that the quest for alternatives and adjuncts to pharmacological treatments has been stepped up and that attempts are being made to improve the methodological quality of their evaluation.

However, a clinician faced with an adolescent newly diagnosed as having schizophrenia would find it extremely difficult to tease out which interventions might be helpful and of these, which might be a cost-effective use of available resources. Adolescents continue, as before, to fall between child and adult services in terms of service provision (National Institute of Clinical Excellence [NICE], 2002). Perhaps mental health initiatives, building on the continuing outcomes of early-intervention programs, can shape devoted services to be targeted at problems that beset adolescents with schizophrenia by virtue of their developmental stage. Such initiatives need to be designed specifically for young minds and hearts.


As noted in Chapter 5, schizophrenia is a severe mental illness characterized by abnormalities of thought and perception that affects about 1% of the population worldwide over the course of a lifetime (Bourdon, Rae, Locke, Narrow, & Regier, 1992; Eaton, 1985; Hare, 1987; Helgason, 1964; Jablensky, 1986; Kramer, 1969; Robins et al., 1984). The optimal time to treat this illness with the currently available therapeutic agents is as early in the course and as close to the onset as possible. Often the onset of the illness precedes the manifestation of symptoms diagnosable at the syndromal level by a considerable period of time. As also summarized in Chapter 5, the onset of the formal symptoms of schizophrenia is generally preceded by a prodromal phase. So-called prodromal symptoms and behaviors (i.e., those that herald the approaching onset of the illness) include attenuated positive symptoms (e.g., illusions, ideas of reference, magical thinking, superstitiousness), mood symptoms (e.g., anxiety, dysphoria, irritability), cognitive symptoms (e.g., distractibility, concentration difficulties), social withdrawal, or obsessive behaviors, to name a few (McGlashan, 1996; Yung & McGorry, 1996). Because many of these prodromal phenomena extensively overlap with the range of mental experiences and behaviors of persons in the ages of risk who do not subsequently develop schizophrenia, prodromal symptoms cannot be considered diagnostic. It is precisely their nonspecificity and lack of high predictive validity that limits their utility for the purposes of early intervention (Gottesman & Erlenmeyer-Kimling, 2001; Schaffner & McGorry, 2001).

The development of frank psychotic symptoms marks the formal onset of first-episode schizophrenia, although this is usually not diagnosed for some time, until the patient seeks or is brought to medical attention. Indeed, the duration of psychotic symptoms prior to diagnosis and treatment averages about 1 year, and if time since prodromal symptoms first appeared is considered, the average duration is about 3 years (McGlashan, 1996). Despite this, most individuals recover symptomatically from the first episode. However, most patients proceed to have one or more subsequent episodes in the form of psychotic relapses from which some proportion fail to recover, at least to the same degree as they had during their first or prior episode (Lieberman et al., 1993, 1996; Robinson et al., 1999a). This process of psychotic relapses, treatment failure, and incomplete recovery leads many patients to a chronic course of illness. Finally, persistent disturbances and deficits in perceptions, thought processes, and cognition affect development (Lieberman, 1999; McGlashan, 1988). In this way, patients accumulate morbidity in the form of residual or persistent symptoms and decrements in function from their premorbid status. The process of accruing morbidity in the context of exacerbations and (relative) remissions has been attributed to progression of the illness (Kraepelin, 1919) and described as “clinical deterioration” (Bleuler, 1980). Interestingly, the deterioration process occurs predominantly in the early phases of the illness, in the prepsychotic prodromal period and during the first 5 to 10 years after the initial episode (see Figure 5.1). For these reasons, early intervention is highly indicated. In this context, treatment serves two purposes: first, it is remedial for active symptoms of whatever level of severity or syndromal criteria; second, it may be preventive of the deterioration, which can occur and is the most devastating consequence of the illness. However, there are several challenges for treating patients optimally in the earliest stages of schizophrenia. First, patients usually do not seek (or are not brought in for) treatment until they have had the symptoms for often lengthy periods of time. Second, the symptoms by which persons in the prodromal stage and at imminent risk for psychosis are identified are not highly specific or sufficiently validated to use in clinical practice. Third, the optimal treatment agents and strategies for prodromal patients have not been determined. Fourth, patients in the prodromal stages and experiencing first episodes of schizophrenia are reluctant to take medications for sustained periods of time, as they have limited insight into the nature of their illness and are sensitive to and often object to side effects.

Acute Treatment

There have been relatively few studies on first-episode schizophrenia and very few on patients in the prodromal phase. There are currently two published studies of controlled or standardized acute treatment of patients with prodromal symptoms of schizophrenia and eight published studies of controlled or standardized acute treatment of patients with first-episode schizophrenia (summarized in Tables 6.1 and 6.2; Emsley, 1999; Kopala et al., 1996; Lieberman et al., 1993; May, Tuma, Yale, Potepan, & Dixon, 1976; Sanger et al., 1999; Scottish Schizophrenia Research Group, 1987; Szymanski et al., 1994). From these data and those from some uncontrolled studies and secondary analyses, several principles can be suggested to provide guidance for the treatment of patients experiencing a first episode of schizophrenia and for future research on improving the standard of care for this critical stage of psychotic disorders. With such limited data available on the management of prodromal and first-episode schizophrenia, practice will be guided by a hybrid of clinical trials evidence, real-life studies, and clinical experience.

Table 6.1 Studies of Acute Treatment in Prodromal Stages of Schizophrenia





Response Criteria

Response Rates

McGorry et al., 2002

59 patients at incipient risk of progression to first-episode psychosis (“high risk”)

Ages 14–30, living in Melbourne metropolitan area; met criteria for 1 or more of 3 operationally defined “high-risk” categories

Single-blind, randomized, controlled trial with blinded interviews. Subjects were randomized to need-based intervention (according to presenting or existing problems) and specific preventive intervention, which included 1–2 mg risperidone and a modified CBT.

Progression to psychosis at 6 months (postintervention) and 12 months (follow-up) after study entry, defined by a predetermined threshold of positive symptoms sustained for 1 week or more (based on BPRS and comprehensive assessment of symptoms and history)

Needs-based intervention: 36% postintervention, 36% at 12 months

Specific prevention intervention: 10% postintervention, 19% at 12 months

Woods et al., 2002

60 patients from 4 North American centers diagnosed by means of the Structured Interview for Prodromal Symptoms

Not available in published abstract

Randomized to 5–15 mg olanzapine daily or placebo

SOPS at 8 weeks; weight gain

Olanzapine: least-squares SOPS mean = −14.0 ± 3.3 versus −2.1 ± 3.4. Olanzapine patients gained significantly more weight (p = .001).

BPRS, Brief Psychiatric Rating Scale; CBT, cognitive behavior therapy; SOPS, Scale of Prodromal Symptoms.

Table 6.2 Studies of Acute Treatment in First-Episode Schizophrenia





Response Criteria

Response Rates

May et al., 1976

228 patients selected from consecutive admissions to a state psychiatric hospital between 1959 and 1962

Inclusions: first-admission patients with diagnosis of schizophrenia and “no significant prior treatment”

Exclusions: those judged unlikely to be discharged and those who remitted fairly quickly (within 18 days)

Randomization to either individual psychotherapy, trifluoperazine, psychotherapy in combination with trifluoperazine, electroconvulsive therapy, or milieu therapy only

Release from hospital after a “fair trial” (6–12 months) of the assigned treatment

Individual psychotherapy (65%), trifluoperazine (96%), psychotherapy in combination with trifluoperazine (95%), ECT (79%), and milieu therapy only (58%)

Scottish Schizophrenia Research Group, 1987

46 patients with first-episode schizophrenia admitted to hospital

Inclusion: diagnosis of first-episode criteria on clinician's ICD-9

5-week, double-blind, randomized trial of flupenthixol versus pimozide

Adjunctive medications allowed

“Responders”: able to enter maintenance treatment on their assigned drug therapy

“Non-responders”: those who received further treatment, either ECT or another antipsychotic

“Non-completers”: those who did not proceed to maintenance therapy but were clearly not “non-responders”

63% were “responders” overall

Positive symptoms improved significantly (p < .01) for both drugs during the study period, but negative symptoms did not change

Lieberman et al., 1993; Robinson et al., 1999b

70 RDC-diagnosed patients with schizophrenia (N = 54) and schizoaffective disorder (N = 16)

1. No prior psychotic episodes

2. Age 16–40 years

3. No history of neurological or general medical illness that could influence diagnosis or biological variables being studied

Open, prospective study using a standardized antipsychotic protocol of sequential trials until response criteria were met with fluphenazine, haloperidol, molindone, and clozapine

Operationally defined as a CGI rating of “much” or “very much” improved and a rating of mild or less on specified SADS-C+PDI items with response sustained for at least 8 weeks

83% of patients remitted by 1 year with mean and median times to remission of 35.7 and 11 weeks, respectively

Szymanski et al., 1994

10 patients in the Lieberman et al. (1993) study

Patients in the above study who had failed treatment with a standardized protocol of 3 typical antipsychotics

After a 2-week washout period, patients were treated for 12 weeks on clozapine

20% reduction in BPRS score and a CGI severity of illness score of ≤3

30% (3 of 10)

Kopala et al., 1996

22 neuroleptic naive patients consecutively admitted for the first time; mean age of 25 years

DSM-IV diagnosis of a first episode of schizophrenia

Open trial of risperidone monotherapy for a mean duration of treatment was 7.1 weeks (SD = 3.2, range 1.8–14.1).

Benzotropine for EPS and lorazepam or clonazepam for insomnia were the only adjunctives allowed.

20% reduction in total PANSS score


Negative symptoms improved less than positive symptoms.

Sanger et al., 1999

83 first-episode patients out of 1,996 patients enrolled in a multicenter trial of olanzapine versus haloperidol in psychotic disorders

1. First episode of psychosis with a DSM-II-R diagnosis of schizophrenia, schizophreniform, or schizoaffective disorder

2. Duration of episode of <5 years

3. No more than 45 years old at episode onset

4.  Minimum BPRS score of 18 or intolerant to current antipsychotic therapy

6-week, double-blind, randomized trial of olanzapine (N = 59) or haloperidol (N = 24) at mean modal doses of 11.6 (SD = 5.9) and 10.8 (SD = 4.8) mg/day, respectively

Defined a priori as a ≥40% reduction in total BPRS from baseline, also calculated for a 20% reduction

40% BPRS reduction in total BPRS: olanzapine 67.2% response rate, haloperidol 29.2% (Fisher's exact p = .003)

20% BPRS reduction: olanzapine 82.8% response rate, haloperidol 58.3% (Fisher's exact p = .03)

Emsley, 1999

183 patients recruited in multiple international sites

1. Ages 15–40 years

2. Diagnosis of provisional schizophreniform disorder or schizophrenia according to DSM-III-R

3. No prior treatment beyond 3 days of emergency antipsychotics

4. No clinically relevant medical abnormalities

6-week, double-blind study of risperidone versus haloperidol 2–16 mg/day

Antiparkinsonian drugs or benzodiazepines administered only if essential

50% improvement in total PANSS scores was defined a priori as clinical response

Risperidone: 63% response rate

Haloperidol: 56% response rate

Yap et al., 2001

24 patients recruited from Woodbridge Hospital and Geylang Psychiatric Outpatient Clinic

Previously untreated male and female patients aged 18–65 with DSM-IV schizophreniform disorder or DSM-IV schizophrenia for no longer than 12 months

Open-label, 8-week study of risperidone

20% reduction in total PANSS score, response for 50% reduction in total PANSS score was also calculated

Responder rate (≥20% reduction in total PANSS score) was 87.5%

13 patients (54.2%) exhibited ≥50% reduction in total PANSS score

Lieberman, Gu, et al., 2003

160 Chinese patients in first-episode schizophrenia

1. Ages 16–45 years

2. Diagnosis of provisional schizophreniform disorder or schizophrenia according to DSM-IV

3. No prior antipsychotic treatment

4. No clinically relevant medical abnormalities

52-week, double-blind RCT of clozapine and chlorpromazine and trihexyphenidyl

Lieberman, Tollefson, et al., 2003

263 patients recruited from 14 sites in N. America and W. Europe

1. Ages 16–40 years

2. Diagnosis of provisional schizophreniform disorder, schizophrenia, or schizoaffective disorder according to DSM-IV

3. Prior lifetime treatment <16 weeks

4. No clinically relevant medical abnormalities

12-week, acute treatment results of 2-year double-blind RCT of olanzapine (5–15 mg/day) and haloperidol (2–20 mg/day).

Antiparkinsonian drugs or benzodiazepines administered only if essential

Total PANSS response defined as 30% reduction of PANSS and CGI severity <4 (moderately ill)

Significantly greater reduction in total PANSS and PANSS Negative Scale with olanzapine on mixed models but not LOCF analysis

55% of olanzapine and 46% of haloperidol met response criteria by week 12

BPRS, Brief Psychiatric Rating Scale; CGI, Clinical Global Impression; ECT, electroconvulsive therapy; LOCF, last-observation-carried-forward; PANSS, Positive and Negative Syndrome Scale for schizophrenia; RCT, randomized control trial; RDC, Research Diagnostic Criteria; SADS-C+PDI, Schedule for Affective Disorders and Schizophrenia Change Version with Psychosis and Disorganization Items rating scale; SD, standard deviation.

Dosing and Selection of Pharmacologic Agents for Early Stages of Schizophrenia

The pharmacology of treating the prodromal stages of schizophrenia and related psychotic disorders has not been sufficiently well developed. Many agents have been suggested (by theories of pathogenesis) as preventive agents for schizophrenia to be used in the prodromal stage. These include a wide variety of treatments (antioxidants, benzodiazepines, phospholipids, lithium and mood-stabilizers, antidepressants, glutamatergic and nicotinic agents, selective DRD 1, 3, 4 dopamine receptor antagonists, and antipsychotic drugs). For both prodromal and first-episode patients, if an antipsychotic drug is to be used, the general consensus is that this be one of the second-generation antipsychotic drugs (Addington, 2002; Bhana, Foster, Olney, & Plosker, 2001; Bustillo, Lauriello, & Keith, 1999; Green & Schildkraut, 1995; Lieberman, 1996; NICE, 2002; Sartorius et al., 2002). Currently, there are no specific guidelines or sufficient evidence to determine which second-generation antipsychotic to use. Side effects are the primary distinguishing features among the various drugs.

Young patients without prior exposure to antipsychotic drugs may be more sensitive to the antipsychotic side effects than patients in other stages of the illness. In a sample of 70 treatment-naive patients who received fluphenazine at 20 to 40 mg/day for the first 10 weeks of treatment, 34% developed parkinsonism, 18% developed akathisia, and 36% developed dystonia (Chakos, Mayerhoff, Loebel, Alvir, & Lieberman, 1992). Lower doses of antipsychotics may be adequate to achieve positive symptom remission, but less likely to cause side effects (Cullberg, 1999; Zhang-Wong, Zipursky, Beiser, & Bean, 1999). For example, in a post hoc analysis, low-dose risperidone (maximum of ≤6 mg/day) was more effective and better tolerated than high-dose risperidone (maximum of ≥6 mg/day; Emsley, 1999). Another recent study of 49 acutely psychotic, neuroleptic-naive patients with schizophrenia, schizophreniform disorder, or schizoaffective disorder treated with either 2 mg or 4 mg daily of risperidone showed the two doses to be comparable in efficacy with an advantage for the lower dose in fine motor functioning (Merlo et al., 2002). The greater sensitivity to side effects of first-episode patients than that of chronic patients was dramatically demonstrated by McEvoy, Hogarty, and Steingard (1991) in comparing their neuroleptic thresholds for extrapyramidal symptoms. In the context of a gradual dose titration paradigm, first-episode patients exhibited lower thresholds to develop signs of extrapyramidal symptoms than previously treated more chronic patients. Younger patients are also more susceptible to other side effects, such as weight gain (Kumra et al., 1998; Lieberman, Gu, et al., 2003). Consistent with these studies, the Schizophrenia Patient Outcomes Research Team recommended that patients in a first psychotic episode be treated with relatively lower doses (300 to 500 mg chlorpromazine equivalents per day) of antipsychotics than for patients with schizophrenia in general (300 to 1,000 mg chlorpromazine equivalents per day; Lehman & Steinwachs, 1998). Clinical experience and available research findings suggest that lower doses of antipsychotics are as effective as higher doses in patients experiencing a first episode of schizophrenia, with superior tolerability.

Since the first episode is a time when patients form their attitudes about treatment, efforts to minimize unpleasant side effects may influence patients' willingness to take medications over the long term. In a study of first-episode patients, the only variable that predicted whether patients would attend a follow-up assessment was antipsychotic dosage, with those on higher doses less likely to comply (Jackson et al., 2001).

Treatment of Positive Symptoms

Available data indicate that in prodromal and first-episode patients, positive symptoms, including hallucinations and delusions, will usually remit with antipsychotic treatment. Thus, clinicians should expect remission of positive symptoms in prodromal and first-episode patients. A series of adequate trials of available agents should be employed with this goal in mind. If residual symptoms persist, a trial of clozapine or the addition of adjunctive treatments should be considered.

Treatment of Negative and Cognitive Symptoms

While positive symptoms in first-episode patients tend to respond well to antipsychotic drug treatment, negative and cognitive symptoms of schizophrenia generally take longer to respond or are less responsive to antipsychotic medications (Kopala et al., 1996; Sanger et al., 1999; Scottish Schizophrenia Research Group, 1987). This indicates that negative and cognitive symptoms may have a different time course for response than positive symptoms or that the relative refractoriness of negative and cognitive symptoms may contribute to the less than optimal functional recovery that is often observed in first-episode patients. Improving treatments for negative and cognitive symptoms in the first episode of schizophrenia is an area of major importance in future research and drug development efforts, especially since these symptoms likely affect these patients' functional abilities.

Adjunctive Treatments of Residual Symptoms and Comorbid Syndromes

Often antipsychotic medications are insufficient by themselves to achieve full symptom remission and functional recovery in early-stage schizophrenia patients. For these reasons, a variety of adjunctive treatments both pharmacologic and nonpharmacologic can be used to enhance and optimize treatment response. Adjunctive treatments have different roles in the management of first-episode schizophrenia, targeting residual symptoms and treating comorbid syndromes. A clinical approach to treatment-refractory, first-episode schizophrenia, as described in a recent manual on first-episode schizophrenia, should include the strategies that promote medication adherence, attention to substance abuse, sequential trials of antipsychotic agents, or dose adjustment if clinical improvement is not seen by 6 to 12 weeks of treatment, and consideration of clozapine even early in the course of treatment (Edwards & McGorry, 2002).

While not yet systematically studied in prodromal or first-episode patients, data from studies in chronic patients suggest that cognitive therapy also may benefit residual symptoms (Cormac et al., 2002). Even given the lack of specific trials in first-episode patients, it is likely that CBT may have a role in the treatment of prodromal and first-episode schizophrenia, especially in helping patients to comply with treatment and the transition to outpatient care.

Adjunctive treatments that have been studied in the treatment of schizophrenia include benzodiazepines, anticonvulsants, and antidepressants. Benzodiazepines are often used as adjuncts to antipsychotics in acute schizophrenia. There have been no controlled studies of benzodiazepines in groups of first-episode patients with schizophrenia. Literature on mixed populations of patients with schizophrenia suggest overall that benzodiazepines have a role in treating agitation, anxiety, and aggression in patients with acute psychosis (Barbee, Mancuso, Freed, & Todorov, 1992; Battaglia et al., 1997; Kellner, Wilson, Muldawer, & Pathak, 1975; Salzman et al., 1991) and in preventing an impending psychotic relapse (Carpenter, Buchanan, Kirkpatrick, & Breier, 1999).

Mood stabilizers, including anticonvulsants and lithium, are widely used in patients with schizophrenia (Citrome, Levine, & Allingham, 2000), and there is some evidence that they may also have a role in reducing aggression and agitation (Christison, Kirch, & Wyatt, 1991; Ko, Korpi, Freed, Zalcman, & Bigelow, 1985; Leucht, McGrath, White, & Kissling, 2002; Linnoila & Viukari, 1979; Wassef et al., 2000). Although there is no evidence as to their efficacy in early stage patients, mood stabilizers should be considered for patients who have mood symptoms of excitement and mood lability during residual to the prodromal and first episodes of schizophrenia.

Depressive syndromes are common in prodromal and first-episode patients. Patients who ultimately manifest symptoms of schizophrenia often report a previous depressive episode (Hafner, Loffler, Maurer, Hambrecht, & an der Heiden, 1999) or suicide attempt in their prodromal period (Cohen, Lavelle, Rich, & Bromet, 1994). On presentation with an acute psychotic episode, first-episode patients often have mood symptoms (Addington, Addington, & Patten, 1998). Depressive symptoms will often resolve as psychotic symptoms remit (Koreen et al., 1993), however, in some cases they may persist or occur in the episode's aftermath (“postpsychotic depression”). Antidepressants should be used cautiously in prodromal and first-episode schizophrenia as they could possibly provoke or exacerbate psychotic symptoms. In addition, residual negative symptoms that persist after the stabilization of the acute episode may respond to antidepressant treatment (Berk, Ichim, & Brook, 2001; Hogarty et al., 1995; Silver, Barash, Aharon, Kaplan, & Poyurovsky, 2000; Silver & Nassar, 1992).

Suicidal behavior can occur in prodromal and first-episode schizophrenia patients (Steinert, Wiebe, & Gebhardt, 1999). While depression in the presenting psychotic episode or in the postpsychotic period is an important risk factor for suicide (Axelsson & Lagerkvist-Briggs, 1992), prodromal and first-episode patients with schizophrenia may attempt suicide in the ab sence of prominent depressive symptoms as a result of hallucinations, paranoia, disorganization, or other symptoms considered more primary to psychosis or other factors. Mounting literature supports the use of clozapine (Meltzer et al., 2003; Meltzer & Okayli, 1995; Reid, Mason, & Hogan, 1998; Sernyak, Desai, Stolar, & Rosenheck, 2001) in patients with psychotic disorder and suicidal behaviors. Although use of clozapine in first-episode schizophrenia has been studied recently (Lieberman, Gu, et al., 2003), it is not considered at this time a first-line drug for first-episode schizophrenia. It should be considered early in the course of treatment only in patients who are unresponsive to other second-generation antipsychotic drugs.

Another important comorbid syndrome in the treatment of first-episode schizophrenia is substance abuse with its possible role in lowering initial vulnerability to the onset or recurrence of psychosis (Chouljian et al., 1995; DeQuardo, Carpenter, & Tandon, 1994; Gupta, Hendricks, Kenkel, Bhatia, & Haffke, 1996; Hambrecht & Hafner, 2000; Linszen, Dingemans, & Lenior, 1994; Rabinowitz et al., 1998). Thorough evaluations of substance abuse habits of prodromal and first-episode schizophrenia patients are critical to directing appropriate clinical attention to this issue throughout care of the patient.

Treatment After Remission of Symptoms in Prodromal and First-Episode Schizophrenia

Continuation and Maintenance Treatment

As the symptomatic response to treatment of young early stage patients is generally very good, clinicians should expect and aim for achieving maximal remission of symptoms, recognizing that psychosis may resolve first and then negative and cognitive symptoms. Residual symptoms should be targeted with adjunctive therapies as needed. After achieving optimal treatment response, the next goal of treatment is to maximize the functional recovery of patients. To an extent, this is dependent on the resolution of symptoms but may not necessarily occur fully concurrently with symptom remission. Maximization of functional recovery involves the use of adjunctive pharmacologic treatments for any residual symptoms but then various nonpharmacologic treatments to enhance functional recovery. These may include psychoeducation about the nature of the illness, supportive psychotherapy, supported employment, social and vocational rehabilitation, and case management. We think of many of these psychosocial interventions as being associated with the care of chronic patients; however, they can and should also be adapted to young early-stage patients. It is also important to keep in mind that it is necessary to be patient and not rush patients prematurely into the activities with which they may have been previously engaged.

The need for interventions aimed at achieving functional recovery is reflected by the results of outcome studies in first-episode patients. Although patients typically recover from a first episode of schizophrenia, the long-term course for most patients is still characterized by chronic illness, disability, and relapse. Studies report that a minority of patients, about 15% to 20%, will maintain good symptomatic and functional recovery from a first episode. For example, in a study of 349 patients followed up to 15 years after their first onset of schizophrenia, 17% had no disability at follow-up, whereas 24% still suffered from severe disability, and the remaining 69% had varying degrees of disability (Wiersma et al., 2000; Wiersma, Nienhuis, Sloof, & Giel, 1998). The long-term prognosis of patients in the pre-antipsychotic era is similar, with about 20% of patients having good symptomatic and functional recovery (Bleuler, 1978). In a study of first-episode patients with operationally defined criteria for recovery (a period of 2 years of remission of positive and negative symptoms, fulfillment of age-appropriate role expectations, performance of daily living tasks without supervision, and engagement in social interactions), 16.8% of the 118 patients achieved full functional recovery (Robinson et al., 1999b). However, a recent study of 1,633 patients with psychotic disorders from diverse cultures found more optimistic rates of favorable outcome, with nearly half of the patients with schizophrenia considered to be recovered (Harrison et al., 2001).

The next question in the clinical management of prodromal and first-episode patients is, after achieving maximal therapeutic response, how long should treatment (particularly pharmacologic treatment) be continued? There are not sufficient data to answer that question in prodromal patients. Thus, treatment of prodromal symptoms should be considered time limited and aimed at alleviating current symptoms and stabilizing the patient. There is a growing body of evidence on first-episode patients, however, that suggests the value of continuing medication for a sustained and possibly indefinite period. Ideally, this decision would be informed by data from prospective, controlled studies answering the following questions: (1) How likely is relapse with and without antipsychotic medication? (2) Is it possible to predict who will relapse after remission of a first episode? (3) Will antipsychotic therapy improve the course and long-term outcome of the illness or merely suppress symptoms in the short to medium term? Table 6.3 summarizes the six controlled studies of maintenance antipsychotic treatment of remitted first-episode schizophrenia (Crow, MacMillan, Johnson, & Johnstone, 1986; Gitlin et al., 2001; Kane, Rifkin, Quitkin, Nayak, & Ramos-Lorenzi, 1982; McCreadie et al., 1989; Nuechterlein et al., 1994).

Table 6.3 Studies of Maintenance Treatment in First-Episode Schizophrenia





Relapse Criteria

Relapse Rates

Kane et al. 1982

28 patients referred for aftercare with a diagnosis of a single episode of schizophrenia; 19 of 28 patients met RDC criteria for schizophrenia

1. At least 4 weeks of remission

2. 1 year or less since hospitalization

3. No treatment prior to 3 months before hospitalization

4. No evidence of drug abuse, alcoholism, or important medical illness

Double-blind, 1-year duration Random assignment to fluphenazine hydrochloride (5–20 mg/day), fluphenazine decanoate (12.5–50 mg/2 weeks) or placebo. Only patients thought to have possible compliance problems were randomized to decanoate.

Procyclidine hydrocholoride for all patients, with substitution of placebo in second month for placebo patients

Substantial clinical deterioration with a potential for marked social impairment

Patients were considered dropouts only if they showed no clinical deterioration at the time they left the study

For all patients: 0% (0 out of 11) on fluphenazine and 41% (7 out of 17) on placebo relapsed

Of those with RDC schizophrenia: 0% (0 out of 6) on fluphenazine and 46% (6 out of 13) on placebo relapsed

Follow-up with 26 patients (mean interval of 3.5 years): 69% had a second relapse; 54% had a third relapse

Crow et al., 1986

120 patients diagnosed with first-episode schizophrenia, recruited from both psychiatric and district general practices in Harrow, England

1. Age 15–70 years

2. Suffering from a first psychotic episode that was “not unequivocally affective”

3. Admission to an inpatient psychiatric unit for at least 1 week

4. Clinical diagnosis of schizophrenia

5. Absence of organic disease of probable etiological significance

Drugs [fluphenthixol (40 mg/month IM), chlorpromazine (200 mg/day orally), haloperidol (3 mg/day orally), pimozide (4 mg/day), or trifluoperazine (5 mg/day)] chosen by clinicians, then patients randomized to either drug or placebo 1 month after remission of initial episode

Adjunctive medications allowed

Psychiatric readmission for any reason; readmission deemed necessary by treating clinician, but not possible; or active antipsychotic medication considered to be essential because of features of imminent relapse

Relapse determinations made by treating clinicians

Actuarial relapse rates: 6 months: placebo 43%, drug 21%; 12 months: placebo 63%, drug 38%; 18 months: placebo 67%, drug 46%; 24 months: placebo 70%, drug 58%

Scottish Schizophrenia Research Group, 1987

15 patients who had suffered a first episode of schizophrenia

Patients had to respond to acute treatment and then be relapse-free for an additional year of treatment on either once-weekly pimozide or IM fluphenthixol decanoate

Double-blind trial of active medication (either once-weekly pimozide or IM fluphenthixol decanoate) or placebo for 1 year

Deterioration in schizophrenia symptoms or behavior sufficient to warrant patient's withdrawal from study

0% (0 out of 8) of the patients who received active drug, but 57% (4 out of 7) who received placebo were readmitted in second year of study treatment

Nuechterlein et al., 1994

106 patients from four public hospitals and an outpatient clinic of an academic medical center 66% had never taken antipsychotic medication and the remainder had a mean duration of treatment of 2.7 months (SD = 3.1)

Inclusion: recent-onset psychotic symptoms lasting at least 2 weeks and not more than 2 years, age 18–45 years, and RDC diagnosis of schizophrenia or schizoaffective disorder (mainly schizophrenia)

Exclusions: known neurological disorder, recent significant substance abuse, or African-American descenta

Phase I: stabilized patients were treated with 12.5 mg fluphenazine decanoate every 2 weeks for 12 months

Phase II: those who remitted in Phase I were recruited into Phase II. They were treated for 12 weeks with either placebo or fluphenazine, followed by crossover to the opposite treatment

Operationally defined as a 2-point worsening on any three BPRS psychotic items, excluding changes where the scores remained at nonpsychotic levels; a score of 6 or 7 was obtained on any three items; or clinical deterioration warranting a change in treatment as judged by treating psychiatrist

Phase I: 11 patients needed to have dose lowered because of side effects and 6 were prescribed antidepressants

Phase II: 6% (3 of 53 subjects) had a relapse while on active medication and 13% (7 out of 53) relapsed on placebo

Robinson et al., 1999b

104 patients who responded to treatment of their index episode of schizophrenia and were at risk for relapse

1. RDC-defined diagnosis of schizophrenia or schizoaffective disorder

2. Total lifetime exposure to antipsychotic medications of ≤12 weeks

3. Rating of ≥4 (moderate) on at least one psychotic symptom item on SADS−C+PD

4. No medical contraindications to treatment with antipsychotic medications

5. No neurologic or endocrine disorder or neuromedical illness that could affect diagnosis or biological variables in study

Patients were treated openly according to a standard algorithm, progressing from one phase of the algorithm to the next until they met response criteria. The sequence was initial treatment with fluphenazine, then haloperidol, then lithium augmentation, then molindone hydrochloride or loxapine, then clozapine

At least “moderately ill” on the CGI Severity of Illness Scale, “much worse” or “very much worse” on the CGI Improvement Scale, and at least “moderate” on one or more of the SADS-C+PD psychosis items listed above; these criteria had to be sustained for a minimum of 1 week

5-year overall relapse rate: 81.9%; the second relapse rate. By 4 years after recovery from a second relapse, the cumulative third relapse rate was 86.2%. Discontinuing antipsychotic drug therapy increased the risk of relapse by almost 5 times

Gitlin et al., 2001

53 patients with RDC schizophrenia or schizoaffective disorder who had been stabilized for 1 year on fluphenzine decanoate

Patient who completed Phase II of the Nuechterlein et al. study (above) and did not relapse

Open-label discontinuation of drugs

Two-point worsening on any three BPRS psychotic items, excluding changes where the scores remained at nonpsychotic levels, a score of 6 or 7 was obtained on any three items, or the treating psychiatrist deemed that there was clinical deterioration warranting change in treatment

78% relapsed by 1 year and 96% by 2 years with low threshold for relapse. Only 13% required hospitalization

a Excluded because of differences in electrodermal conductivity from other groups because this was a biological variable being studied.

BRPS, Brief Psychiatric Rating Scale; CGI, Clinical Global Impression; IM, intramuscular; RCT, randomized control trial; RDC, Research Diagnostic Criteria; SADS-C+PDI, Schedule for Affective Disorders and Schizophrenia Change Version with Psychosis and Disorganization Items rating scale; SD, standard deviation.

The risk of eventual relapse after recovery from a first psychotic episode is very high and is greatly diminished by maintenance antipsychotic treatment. However, even with strong evidence of the risk of relapse without antipsychotic medication, there is still no clear consensus on the recommended duration of treatment for patients who have recovered from a first episode of schizophrenia. Clinicians may have a difficult time convincing patients who have recovered from one episode of schizophrenia that indefinite and possibly lifelong antipsychotic treatment is indicated because of the diagnostic uncertainty and instability associated with a first psychotic episode (Amin et al., 1999), limited patient understanding and awareness of the illness (Thompson, McGorry, & Harrigan, 2001), and risks of long-term antipsychotic therapy. Most practice guidelines for the treatment of patients with schizophrenia recommend that patients who have had only one episode of positive symptoms and have been symptom-free during the subsequent year of maintenance therapy can be considered for a trial period without medication, provided that dose reductions are made gradually over several months with frequent visits (Rose, 1997). Similarly, the Schizophrenia Patient Outcomes Research Team (Lehman & Steinwachs, 1998) recommended in their report that patients continue treatment for at least 1 year after remission of acute symptoms, with continual reassessment of the maintenance dose for possible reduction. The draft consensus statement of 26 international consultants (Addington, 2002) recommends taking into consideration the severity of the first episode when deciding how long to continue maintenance treatment. They suggest that patients who achieve full remission be offered gradual withdrawal of medication after 12 months of maintenance treatment, but that patients experiencing more severe episodes and are slow to respond be maintained for 24 months. This panel further suggests that patients who respond incompletely to medication, but clearly benefit from treatment, be maintained for 2 to 5 years on medication.


Patients in the prodromal phase or first episode of schizophrenia respond very favorably to antipsychotic treatment in terms of their positive symptoms. However, cognitive and negative symptoms are often slower to respond or are refractory to treatment, leaving many of these individuals with significant social disability. Sequential trials of adequate dose and duration of antipsychotics and adjunctive treatments should be employed to help patients achieve their optimal response. Psychosocial therapies such as CBT and education groups may be helpful in addressing residual symptoms and helping patients to adhere to their medication regimens.

In addition to lessening the morbidity of the presenting episode, early intervention in prodromal and first-episode schizophrenia may improve the long-term course of the disorder, as evidenced by studies showing an inverse correlation between the duration of untreated psychosis and outcome. The outcome of schizo phrenia is variable, and despite a large literature, we know relatively little about the predictors of outcome. Duration of untreated illness in various studies also varies widely, ranging between 22 and 166.4 weeks (Norman & Malla, 2001). Several studies have suggested that prolonged duration of untreated illness (DUI) may predict poor outcome as evidenced by longer time to and level of remission. Based on such findings, it has been suggested that decreasing DUI, perhaps by early identification and intervention, might lead to a more favorable outcome. It has also been argued that prolonged untreated illness might be causally related to poor outcomes, perhaps as a result of a neurotoxic process. However, controversy has shrouded this literature, as some studies have not found an association between DUI and outcome.

Prevention of schizophrenia through presymptomatic treatment is an exciting possibility, but future research is needed to develop the methodology by which to reliably identify those at risk before this strategy can become part of routine clinical practice. Once remission from the first episode is reached, maintenance with antipsychotic treatment is indicated for at least 1 year. The overwhelming majority of individuals who do not remain on antipsychotic therapy eventually experience a relapse. This raises the question of the optimal length of continuation and maintenance treatment for patients who have recovered from a first episode of schizophrenia or related psychoses. Clinically useful predictors of the small minority who maintain remission without pharmacotherapy have not yet been identified.

Atypical antipsychotics represent an advance in the treatment of first-episode schizophrenia, with strong evidence for greater tolerability with equal or better therapeutic efficacy. While future research will help to characterize their efficacy relative to one another and define the effect of their use on the long-term outcomes of schizophrenia, available evidence and consensus expert opinion support their use as first-line treatment in first-episode schizophrenia.