Schizophrenia Research 123 (2010) 3036

Contents lists available at ScienceDirect

Schizophrenia Research
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s c h r e s

Review

Randomized-controlled trials in people at ultra high risk of psychosis: A review of treatment effectiveness
Antonio Preti a,, Matteo Cella b,1
a b

Centro Medico Genneruxi, Via Costantinopoli 42, 09129 Cagliari, Italy King's College London, Institute of Psychiatry, Department of Psychological Medicine, Weston Education Centre, Cutcombe Rd, London SE5 9RJ, United Kingdom

a r t i c l e

i n f o

a b s t r a c t
As an extension of the early intervention in psychosis paradigm, different focused treatments are now offered to individuals at ultra high risk of psychosis (UHR) to prevent transition to schizophrenia, however the effectiveness of these treatments is unclear. A systematic literature search in PubMed/Medline and PsycINFO was performed to derive information on randomized control trials (RCTs) in UHR samples. Seven reports were identied detailing results from ve independent RCT studies. Two studies used antipsychotic drugs (one in combination with cognitive behavior therapy); one study employed cognitive therapy; one study used a two-year program of intensive community care with family psychoeducation; one study assessed the effectiveness of 3-months omega-3 polyunsaturated fatty acids (Omega-3 PUFAs) supplementation. Intensive community care and the Omega-3 PUFAs supplementation were effective in reducing the transition to psychosis at 12 months. Overall, rates of transition to psychosis at 1 year were 11% for focused treatment groups (n = 180) and 31.6% for control UHR groups (n = 157). Receiving any of the focused treatment was associated with a lower risk of developing psychosis if compared with no treatment or treatment as usual (Relative Risk = 0.36; 95%CI: 0.220.59). The available evidence at 2/3 years follow-up indicates that the effects of focused treatments are not stable after intervention cessation and when treatment is delivered over a restricted time (e.g. 6 months or less), it may achieve only a delay in psychosis onset. Due to the heterogeneity in the interventions considered, the current results do not allow recommendation for any specic treatment. 2010 Elsevier B.V. All rights reserved.

Article history: Received 8 March 2010 Received in revised form 16 July 2010 Accepted 26 July 2010 Available online 21 August 2010

Keywords: Schizophrenia Early intervention High risk Meta-analysis Randomized-controlled trial

1. Introduction Over the last fteen years, development of early, focused treatment protocols for psychosis resurrected interest in the prodromal phase of the disorder (for a review: Gross, 1997; Raballo et al., 2009). The proposed generalization of the clinical staging paradigm to psychiatry, which is producing excellent results in oncology and cardiovascular diseases, led to the establishment of protocols of diagnosis and treatment aimed at individuals presenting pre-psychotic symptoms that might evolve into full-blown psychosis (McGorry, 2007; Raballo and Lari, 2009). Yung and McGorry (1996) initially
Corresponding author. Tel.: + 39 070 480922. E-mail addresses: apreti@tin.it (A. Preti), matteo.cella@kcl.ac.uk (M. Cella). 1 Tel.: +44 20 322 83191. 0920-9964/$ see front matter 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.schres.2010.07.026

characterized prodromic features of psychosis as an admixture of anxiety, symptoms of depression and social difculties but it was soon recognized that these symptoms were much too generic to be highly predictive of the future onset of psychosis. In Australia, the Personal Assessment and Crisis Evaluation (PACE) Clinic in Melbourne advanced a more precise set of criteria to identify the so-called ultra high risk (UHR) individuals. Currently, the prole of UHR individuals corresponds to: a) Attenuated psychotic symptoms: i.e. subthreshold or attenuated psychotic symptoms during the past year; or, b) Brief limited intermittent psychotic symptoms: i.e. recurring brief episodes of frank psychotic symptoms lasting no more than a week and spontaneously abating; or, c) Genetic and familial risk factors: i.e. being diagnosed with schizotypal personality disorder or having a rst-degree relative with a psychotic disorder, and having experienced a

A. Preti, M. Cella / Schizophrenia Research 123 (2010) 3036

31

signicant functional decrease in the last year (Yung et al., 2003, 2004). Specic interviews were developed to diagnose these prodromic symptoms, such as the Comprehensive Assessment of At-Risk Mental States (Yung et al., 2005) and the Structured Interview for Prodromal Syndromes/Scale of Prodromal Symptoms (Miller et al., 2003). In Germany, Gross and Huber developed alternative criteria based on the socalled Basic Symptoms (BS), a list of anomalous subjective experiences reported by patients during an active episode of psychosis within the spectrum of schizophrenia (Huber et al., 1979; Huber and Gross, 1989). After extensive psychometric validation, a list of nine symptoms predictive of the risk of psychosis was identied, and implemented in protocols of early diagnosis (Klosterktter et al., 2005). Individuals identied with UHR or BS criteria are currently studied and their transition to psychosis investigated at follow-up (Bechdolf et al., 2007; Klosterktter et al., 2001; Yung et al., 2008). Different specic treatments have been trialed for those recognized at higher risk (Edwards and McGorry, 2002; Larsen et al., 2001; Cocchi et al., 2008). Past reviews on the effectiveness of these treatments arrived at inconclusive results (Olsen and Rosenbaum, 2006; Marshall and Rathbone, 2006; de Koning et al., 2009), in part due to the limited number of completed randomized control trials (RCTs). However, more recently new RCT studies of people at high risk of transition to psychosis have been published, thus justifying a reconsideration of treatment effectiveness. 2. Methods A literature search was conducted for RCTs evaluating treatments aimed at preventing the development of a diagnosable psychotic episode in UHR or BS positive patients. Studies were considered relevant if: they tested interventions' effectiveness in reducing transition to psychosis; the specic intervention was compared to a contrast group; assignment to treatment was blind and randomized; transition to psychosis was considered over a period of 12 months or more. Studies on treatments and/or symptoms management interventions designed for people in the early stages of psychosis were excluded. Therefore all studies including patients with one or more clinically relevant psychotic episode were excluded. These studies differ in a signicant way from prevention studies since they offer timely intervention after a rst episode of psychosis has already been experienced. The main outcome considered was transition rate to psychosis in the high risk prospective cohort, dened on the basis of the criteria used in each study (see Results). 2.1. Search criteria and results Both PubMed/Medline and PsycINFO were searched from January 1990 to February 2010 (last update: July 10th, 2010) by MC and AP using the following keywords: psych* or schizo* and risk or ultra and random* or trial or control* or RCT. Only publications in English were considered. Abstracts of retrieved papers were scanned to identify studies matching the inclusion criteria. References of retrieved articles and review on the topic (e.g. Olsen and Rosenbaum, 2006; de Koning et al., 2009; McGorry et al., 2009) were also examined to identify other possible relevant studies.

The search strategy conducted on PsychInfo identied 484 abstracts of which 56 referred to potentially eligible studies in the at-risk population. All non-empirical studies, those not delivering an intervention and those with a design different from RCT were excluded, resulting in six articles that were included in the analysis. Using the same terms we conducted a search in PubMed/Medline and identied an initial pool of 581 studies. Using the same inclusion and exclusion criteria, seven articles were identied and included in the nal analysis. There was a complete overlap between the studies identied with PsychInfo and PubMed/Medline. Further the data presented in McGorry et al. (2002) and Phillips et al. (2007) and those in Morrison et al. (2004, 2007) referred to the same study, therefore ve independent studies were included in the meta-analysis (Table 1). One additional study, referenced in a topical review on psychosis proneness by de Koning et al. (2009) was not included due to lack of information (e.g. sample size). This study investigated cognitive behavioral therapy (CBT) vs. supportive counseling in high risk individuals diagnosed according to BS criteria (Bechdolf et al., 2007). A further study, aimed at evaluating the effectiveness of amisulpiride in individuals at late prodromal state according to the BS criteria, was excluded because reported only data on a 12-week followup and did not assess transition to psychosis as an outcome (Ruhrmann et al., 2007). 2.2. Analysis Intervention effectiveness was investigated using metaanalytic techniques. The meta-analysis is a technique used to amalgamate and review previous quantitative research to answer the question of whether the interventions considered make a difference on a dependent variable. Lipsey and Wilson (2001) suggest that for a good meta-analysis it is essential that the analyst have a denition of the domain of interest and a rationale for inclusion and exclusion of studies. This study wants to test if proposing a specic intervention to people who comply with the criteria for ultra high risk of psychosis is better than offering treatment as usual or no treatment. The conation of different treatment methods in the same analysis has been conducted in previous research and deemed valid. A classic example is the seminal metaanalysis paper by Smith and Glass (1977) investigating the effectiveness of different types of psychotherapy (e.g. psychodynamic, gestalt). Meta-analysis was carried out with Comprehensive MetaAnalysis (version 2.2) software (www.meta-analysis.com/). In the pooled analysis, the Relative Risk (RR) was reported with 95% Condence Interval (CI); for a result to be statistically signicant, extremes of CI should not include the unit (i.e. both of them should be above or below 1). Measures of heterogeneity were the Cochran's Q and the I2 statistic, which assesses the extent of inconsistency among the studies' results and is interpreted as approximately the proportion of total variation in each study estimates independent from sampling errors; I2 statistic of 50% or more was taken to be indicative of moderate heterogeneity (Higgins et al., 2003). According to Egger et al. (1997) funnel plots were also inspected for evidence of asymmetry as a proxy measure of heterogeneity. Heterogeneity was further assessed with

32

A. Preti, M. Cella / Schizophrenia Research 123 (2010) 3036

Table 1 RCT of focused treatment aimed at reducing the risk of transition to psychosis in people at high risk of psychosis. Study McGorry et al., 2002; Phillips et al., 2007 Criteria for diagnosis Criteria for outcome Focused treatment (FT) Contrast group (C) Needs-based intervention (?) Duration = 6 months N = 28 Dropout = none Monitoring Duration = 6 months N = 23 Dropout = 7 Transition to psychosis Transition to psychosis at 1 year* at more than 1 year* FT = 6/31 (19.3%) C = 10/28 (35.7%) Within 3/4 years FT = 10/31 (32.2%) C = 12/28 (42.8%) Dropout = 18 (7/11) Within 3 years FT = 7/35 (20.0%) C = 7/23 (30.4%) Dropout = 31 (18/13)

UHR criteria according to PACE criteria based on the CAARMS Morrison et al., 2004; UHR criteria equivalent Morrison et al., to PACE 2007 criteria based on PANSS McGlashan et al. UHR criteria (2006) based on the SIPS

Suprathreshold Risperidone 12 mg + levels of psychosis CBT Duration = 6 months N = 31 Dropout = none CT Transition to psychosis using Duration = 6 months cut-off points on N = 35 PANSS Dropout = 9

FT = 2/35 (5.7%) C = 5/23 (21.7%)

Nordentoft et al. (2006)

ICD-10 criteria for Schizotypal disorder

Amminger et al. (2010)

UHR criteria equivalent to PACE criteria based on PANSS

Conversion to psychosis according to the Presence of Psychosis Scale ICD-10 diagnosis of a psychotic disorder within the F2 spectrum Raters were not blind Transition to psychosis using cut-off points on PANSS

Olanzapine 515 mg Duration = 12 months N = 31 Dropout = 14 Intensive treatment with family intervention Duration = 24 months N = 42 Dropout = 5 Omega-3 PUFAs 1.2 g Duration = 3 months N = 41 Dropout = 3

Placebo FT = 5/31 (16.1%) Duration = 12 months C = 11/29 (37.9%) N = 29 Dropout = 19 Standard care FT = 3/37 (8.1%) Duration = 24 months C = 10/30 (33.3%) N = 37 Dropout = 7

Within 2 years FT = 8/31 (25.8%) C = 13/29 (44.8%) Dropout = none Within 2 years FT = 9/36 (25.0%) C = 14/29 (48.2%) Dropout = 14 (6/8)

Placebo Duration = 3 months N = 40 Dropout = 2

FT = 2/41 (4.8%) C = 11/40 (27.5%)

Abbreviations and explanations: PACE = Personal Assessment and Crisis Evaluation (PACE) Clinic in Melbourne. CAARMS = Comprehensive Assessment of At-Risk Mental States. PANSS = Positive and Negative Syndrome Scale. SIPS = Structured Interview for Prodromal Syndromes, which operationally denes the PACE UHR criteria. ICD-10 = International Classication of Diseases, tenth edition, World Health Organization. CBT = Cognitive Behavioral Therapy. CT = Cognitive Therapy. PUFAs = Polyunsaturated Fatty Acids. Suprathreshold levels of psychosis: a score of 3 or more on the hallucinations subscale, a score of 4 or more on the unusual thought content subscale (plus a score 3 for delusional conviction on the Comprehensive Assessment of Symptoms and History), or a score of 4 or more on the conceptual disorganization subscale of the Brief Psychiatric Rating Scale; all for a duration greater than 1 week. Transition to psychosis using cut-off points on PANSS: 4 or more on hallucinations, 4 or more on delusions and 5 or more on conceptual disorganization; all for a duration greater than 1 week. Conversion to psychosis according to the Presence of Psychosis Scale: any psychotic disorder in the DSM-IV schizophrenia spectrum on the basis of scores on the Presence of Psychosis Scale (unspecied threshold). *Data on transition to psychosis are on a intention-to-treat basis.

the trim and ll method (Duval and Tweedie, 2000a,b). The trim and ll method assumes the most extreme results go unpublished, and recalculates the effect size by imputation of those missing studies, producing a symmetric funnel plot. A smaller shift in the trim and ll adjusted effect size indicates a good accuracy of the initial effect size. In the case of statistically signicant result, the classical fail-safe n was calculated: this is the minimum number of additional null studies necessary to make the result no longer signicant on a statistical ground (Rosenthal, 1979; Carson et al., 1990).

There were two trials based on medications: - McGlashan et al. (2006) administered 515 mg olanzapine daily for 1 year, compared to placebo; - McGorry et al. (2002) administered low dosage antipsychotic medication (12 mg of risperidone daily) in combination with a six-month course of cognitive therapy. Phillips et al. (2007) reported further follow-up data from this study. One trial was based on cognitive behavioral therapy: - Morrison et al. (2004, 2007) developed a protocol to treat UHR individuals, in line with the basic principles of CBT, and compared this protocol to monitoring (i.e. no active intervention). This six-month long intervention aimed to develop a set of strategies to enhance symptom control and to reduce associate distress by evaluating alternative

3. Results The trials selected for this study involved a variety and combination of interventions and control conditions.

A. Preti, M. Cella / Schizophrenia Research 123 (2010) 3036

33

explanations, decatastrophize fears and testing hypotheses using behavioral experiments. One trial was based on intensive community care: - Nordentoft et al. (2006) used a two-year community treatment intervention based on: home visits, regular assessments, psycho-educational components, social skills training and substance abuse aid. In this study the comparison group was provided with treatment as usual. One trial was based on dietary supplementation: - Amminger et al. (2010) administered a daily dose of 1.2 g omega-3 polyunsaturated fatty acids (Omega-3 PUFAs) for 12 weeks. The omega-3 group was compared with a placebo group prescribed with coconut oil (polyunsaturated fatty acids free). The placebo also contained vitamin E and 1% sh oil to mimic taste of the active treatment. Denition of the transition to psychosis was based on cutoff points on previously validated symptom scales. Threshold levels based on the Brief Psychiatric Rating Scale were used in the PACE study (McGorry et al., 2002; Phillips et al., 2007). Morrison et al. (2004, 2007) and Amminger et al. (2010) used an adaptation of these threshold levels to the Positive and Negative Syndrome Scale. Finally, McGlashan et al. (2006) used an ad hoc scale, the Presence of Psychosis Scale (McGlashan et al., 2003) developed specically to full the lack of DSM-IV-dened criteria for psychosis onset (p. 791). In addition, the PACE group (McGorry et al., 2002; Phillips et al., 2007) and the Early Detection and Intervention Evaluation study (Morrison et al., 2004, 2007) used formal diagnostic criteria for psychosis according the DSM-IV (APA, 1994), while the OPUS trial (Nordentoft et al., 2006) used ICD-10 diagnostic criteria (WHO, 1992). Data on dropout were reported in all the studies considered. A priori power analysis was included in 2 studies only: in one case without details on parameters (Nordentoft et al., 2006) and in the other case (Amminger et al., 2010) using unusual parameters (1-beta, i.e. power= 0.70 rather than 0.80; a too high expected reduction of the risk, i.e. 50%, for the specic treatment). In all studies, blinding to assessors was difcult to maintain, since differences in treatment were evident, for example, on the basis of side effects.

3.1. Treatment effectiveness Twenty-four months program of intensive treatment with family intervention (Nordentoft et al., 2006) and a 3 months program of Omega-3 PUFAs supplementation (Amminger et al., 2010) were found equally effective in reducing transition to psychosis rate after 12 months. Pharmacological interventions with antipsychotic drugs combined or not with CBT were not found to be effective in reducing the transition rate to psychosis in UHR people at 12 months. However there was a trend for less people to be diagnosed with psychosis in all the focused treatment compared with the contrast group (see Table 1 for details). Data on longer follow-up were available only for four studies. There was a trend for intensive community care and olanzapine to produce lower rates of transition to psychosis compared with the contrast group at 24 month follow-up: 25.0% vs. 48.2% (Nordentoft et al., 2006) and 25.8% vs. 44.8% (McGlashan et al., 2006), respectively. For CT and CBT in combination with low dosage risperidone the effects were less evident: 32.2% vs. 42.8% at 3/4 year follow-up (Morrison et al., 2007) and 20.0% vs. 30.4% at 3 year follow-up (Phillips et al., 2007). 3.2. Results of the meta-analysis Table 2 shows the results of the pooled analysis. In all studies the focused treatment yielded a reduced risk of transition to psychosis over 12 months but in three of the ve studies the results were not signicant (on the basis of 95% CI). When considered together, the pooled studies indicated a signicant effect of the focused treatment over the contrast group. This was evident in both the xed and random model RR = 0.364 (95%CI: 0.2220.599), favoring the focused specic intervention. Application of the trim and ll method did not modify this estimate. Heterogeneity was absent: Q = 2.296, df = 4, p = 0.682; I2 = 0%. Classical fail-safe value from the studies considered was 18, which is rather strong (considering that it derives only from ve studies). The exclusion of the Nordentoft et al. (2006) study, due to the inclusion of patients with schizotypal disorder only, did not change the nding favoring the focused specic intervention in both the xed and random model RR = 0.282 (95% CI: 0.1420.560; Q = 1.447, df = 3, p = 0.694; I2 = 0%); however, the classical fail-safe number dropped to 11, which is still moderately good.

Table 2 Meta-analytic results for prospective randomized-controlled studies on transition to psychosis in people at high risk of psychosis. Study Type of treatment Within 12 months RR (95%CI) McGorry et al., 2002; Phillips et al., 2007 Morrison et al., 2004; Morrison et al., 2007 McGlashan et al. (2006) Nordentoft et al. (2006) Amminger et al. (2010) Fixed Random Trim and ll estimate Risperidone + CBT CT Olanzapine Intensive therapy Omega-3 PUFAs 0.542 0.263 0.425 0.264 0.177 0.364 0.364 0.364 (0.2261.298) (0.0561.242) (0.1681.076) (0.0790.888) (0.0420.750) (0.2220.599) (0.2220.599) (0.2210.599) z-value 1.374 1.686 1.806 2.152 2.350 3.981 3.981 p-value 0.169 0.092 0.071 0.031 0.019 0.0001 0.0001 Within 24/36 months RR (95%CI) 0.753 0.657 0.576 0.566 (0.3871.465) (0.2661.626) (0.2801.183) (0.2781.153) z-value 0.836 0.908 1.502 1.562 2.410 2.410 p-value 0.403 0.364 0.133 0.117 0.016 0.016

0.636 (0.4400.919) 0.636 (0.4400.919) 0.635 (0.4400.918)

34

A. Preti, M. Cella / Schizophrenia Research 123 (2010) 3036

As noted, intervention effectiveness was not maintained at the longer follow-ups, when the four studies reporting data on transition to psychosis within 2 or more years were considered individually. However, the pooled analysis still found a statistically signicant effect favoring the focused specic interventions: in both the xed and random model RR = 0.636 (95%CI = 0.4400.919), again with no effect after application of the trim and ll method, and no indication of heterogeneity (Q = 0.427, df = 3, p = 0.935; I2 = 0%). However on the basis of classical fail-safe value of 3 this is a very unstable estimate and means that only three additional studies with no signicant difference between active treatment and contrast group would invalidate this nding. Point estimates of transition to psychosis in the control group, assuming that this is the effect of a non-specic treatment unlikely to modify risk, was 31.6% (95%CI: 24.6% 39.6%) at 12 months and 42.0% (33.1%51.5%) at 24/36 months, as against 11% (6.4%18.5%) and 25.8% (19.0%34.0%), respectively, in the treated group. 4. Discussion Focused treatments with people at high risk of psychosis are effective in reducing the risk of transition to full-blown psychosis over a 12 month period. This nding is robust, with no heterogeneity across studies. However, over a longer period (i.e. 2 years), interventions provided for a limited time interval are less effective, resulting merely in a delay of transition to psychosis. Our ndings are in line with the result of Bechdolf et al.'s (2006, 2008) RCT study, which was not included in this metaanalysis due to lack of information on sample size in the two contrasting groups. These authors found BS positive patients who received a CBT intervention to be less likely than BS positive patients who received supportive counseling to exacerbate from early to late psychosis prodromal state (i.e. a condition with attenuated psychotic symptoms) at both 12 (3.2% vs. 16.9%) and 24 month (6.3% vs. 20.0%) follow-up. The heterogeneity of interventions considered, each with distinct mechanisms by which a therapeutic effect might occur, precludes any clear treatment recommendation. However our results provide an empirical account for strengthening the diagnostic criteria in early intervention programs. The DSM-V will, most probably, include criteria for the diagnosis of the prodromal phase of psychosis (Corcoran et al., 2010). The DSM inclusion of formal criteria to diagnose the at-risk state of developing psychosis may improve the comparability of studies and foster empirical challenges to the denition that will be adopted. When considering the current results some limitations have to be considered. In all but the Nordentoft et al. (2006) study raters are described as blind to interventions. However, in these studies blinding to assessors was difcult to maintain and this is a serious limitation in randomized trials, frequently exposing the results to bias. Also none of the studies considered had a clear description of the method used to generate the random allocation sequence or the procedure used to test concealment (Schulz et al., 2010). Pharmacotherapy based interventions with antipsychotic drugs resulted in a higher proportion of patients developing psychosis, when discontinued, compared to non-pharmaco-

logical intervention: 16.1% for olanzapine and 19.3% for risperidone vs. 5.7% for cognitive therapy alone and 4.8% for Omega-3 PUFAs supplementation. Two naturalistic studies had already reported antipsychotic treatment to be associated with higher rates of transition to psychosis than other interventions (e.g. antidepressants, Cornblatt et al., 2007; Fusar-Poli et al., 2007). The results of the reviewed RCTs suggest caution in prescribing antipsychotic medication to individuals at-risk of developing psychosis, due to noticeable (and potentially permanent) side effects (e.g. Corcoran et al., 2005; Haddad and Dursun, 2008; Odagaki, 2009). Also stigmatization and self-stigmatization can arise by merely receiving a psychiatric prescription or diagnosis, and evidence suggests that antipsychotic prescription can be considered more stigmatizing than other interventions (Corcoran et al., 2005; Jenkins and Carpenter-Song, 2009; Yang et al., 2010). However, various research and clinical accounts showed that individuals at high risk of transition to psychosis, dened with UHR or the BS criteria, are generally suffering from multiple mental and functional disturbances, and can benet from appropriate psychiatric/psychological help (e.g. Ruhrmann et al., 2010a,b). One major advantage in treating UHR individuals would be the timely recognition and treatment of the rst episode, resulting in a substantially shortened duration of untreated psychosis, which can ultimately reduce the worst outcomes of schizophrenia in term of personal and social costs, morbidity and mortality (McGorry et al., 2009). In the studies herewith reviewed, transition to psychosis in the control UHR patients groups was 31% over 12 months and 42% over 24/ 36 months. However, over a longer period (i.e. 9.5 years followup), almost 50% of the patients positive to the BS criteria received a diagnosis of schizophrenia (Klosterktter et al., 2001). The creation of shared criteria for diagnosis of the high risk status and transition to psychosis is a prerequisite to achieve more reliable and consistent results in this area (Yung et al., 2010; Ruhrmann et al., 2010a,b). Most studies included in this review used the UHR criteria, but were variable as far as criteria for transition to psychosis were concerned. Another important aspect worth considering is that no study used hard indicators of social functioning as outcome, such as independent living vs. living with the birth family; being in paid employment or studying vs. unemployment; having a partner vs. being single. Moreover, the use of reliable and valid scales with patients with full-blown psychosis (i.e. the Brief Psychiatric Rating Scale or the Positive and Negative Syndrome Scale) may not be adequate to measure symptoms in high risk individuals. Also dropouts should be more consistently reported and the use of ad hoc parameters for power analysis should be discouraged. The evaluation of costs and implementation feasibility would allow a more accurate comparison of the different therapeutic approaches and will be valuable to inform service delivery. Our results showed that a two-year long community treatment program (Nordentoft et al., 2006) was effective in reducing transition to psychosis rates, nevertheless this intervention it is expected to be more costly than other treatments delivered for shorter periods (e.g. 3 to 6 months). Long lasting and more demanding interventions can be required only for individuals with a severe prodromal presentation. In less severe cases, CBT based approach may prove to be a successful and

A. Preti, M. Cella / Schizophrenia Research 123 (2010) 3036

35

cost-effective strategy. Additional treatments may include glycine-mediate supplementation (Javitt, 2009), antidepressant (Cornblatt et al., 2007) and combination of pharmacological and non-pharmacological interventions. Some of these new intervention approaches are currently under investigation and more evidence will be available to update the results of this review in the future. Replication of positive ndings is necessary to extend the evidence on treatment effectiveness and improve the quality of interventions aimed at preventing the risk of transition to psychosis.
Role of funding source Research was supported by internal funds. Contributors Preti and Cella jointly had the idea and organized the study. Conict of interest Dr. Preti is a scientic advisor and consultant to the Programma 2000 project, a comprehensive program targeted at the early detection and intervention for persons at the onset and for those with a high risk of psychosis, active in Milan (Italy) since 1999. As for Dr. Cella, to the best of his knowledge no conict of interest is foreseeable concerning the data and results described in this article. Acknowledgement None.

References
Amminger, G.P., Schfer, M.R., Papageorgiou, K., et al., 2010. Long-chain omega-3 fatty acids for indicated prevention of psychotic disorders: a randomized, placebo-controlled trial. Arch. Gen. Psychiatry 67, 146154. APA (American Psychiatric Association), 1994. DSMIV. Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition. American Psychiatric Association Press, Washington, DC. Bechdolf, A., Wagner, M., Veith, V., et al., 2006. A randomized controlled multicenter trial of cognitive behaviour therapy in the early initial prodromal state of psychosis. Schizophr. Res. 86, S8. Bechdolf, A., Wagner, M., Veith, V., et al., 2007. Randomized controlled multicentre trial of cognitive behaviour therapy in the early initial prodromal state: effects on social adjustment post treatment. Early Interv. Psychiatry 1, 7178. Bechdolf, A., Wagner, M., Ruhrmann, S., et al., 2008. CBT in the early initial prodromal state: 24-months results. Schizophr. Res. 102, S33. Carson, K.P., Schriesheim, C.A., Kinicki, A.J., 1990. The usefulness of the failsafe statistic in meta-analysis. Educ. Psychol. Meas. 50, 233243. Cocchi, A., Meneghelli, A., Preti, A., 2008. Programma 2000: celebrating ten years of activity of an Italian pilot program on early intervention in psychosis. Aust. N. Z. J. Psychiatry 42, 10031012. Corcoran, C., Malaspina, D., Hercher, L., 2005. Prodromal interventions for schizophrenia vulnerability: the risks of being at risk. Schizophr. Res. 73, 173184. Corcoran, C.M., First, M.B., Cornblatt, B., 2010. The psychosis risk syndrome and its proposed inclusion in the DSM-V: a riskbenet analysis. Schizophr. Res. 120, 1622. Cornblatt, B.A., Lencz, T., Smith, C.W., et al., 2007. Can antidepressants be used to treat the schizophrenia prodrome? Results of a prospective, naturalistic treatment study of adolescents. J. Clin. Psychiatry 68, 546557. de Koning, M.B., Bloemen, O.J.N., van Amelsvoort, T.A.M.J., et al., 2009. Early intervention in patients at ultra high risk of psychosis: benets and risks. Acta Psychiatr. Scand. 119, 426442. Duval, S., Tweedie, R., 2000a. A nonparametric trim and ll method of accounting for publication bias in meta-analysis. J. Am. Stat. Ass. 95, 8998. Duval, S., Tweedie, R., 2000b. Trim and ll: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics 56, 455463. Edwards, J., McGorry, P.D., 2002. Implementing Early Intervention in Psychosis. A Guide to Establishing Early Psychosis Services. Dunitz, London, UK.

Egger, M., Davey Smith, G., Schneider, M., et al., 1997. Bias in meta-analysis detected by a simple, graphical test. BMJ 315, 629634. Fusar-Poli, P., Valmaggia, L., McGuire, P., 2007. Can antidepressants prevent psychosis? Lancet 370, 17461748. Gross, G., 1997. The onset of schizophrenia. Schizophr. Res. 28, 187198. Haddad, P.M., Dursun, S.M., 2008. Neurological complications of psychiatric drugs: clinical features and management. Hum. Psychopharmacol. 23 (Suppl 1), 1526. Higgins, J.P., Thompson, S.G., Deeks, J.J., et al., 2003. Measuring inconsistency in meta-analyses. BMJ 327, 557560. Huber, G., Gross, G., 1989. The concept of basic symptoms in schizophrenic and schizoaffective psychoses. Recenti Prog. Med. 80, 646652. Huber, G., Gross, G., Schttler, R., 1979. Schizophrenie. Eine Verlaufsund sozialpsychiatrische Langzeitstudie. Springer, Berlin, Germany. Javitt, D.C., 2009. Glycine transport inhibitors for the treatment of schizophrenia: symptom and disease modication. Curr. Opin. Drug Discov. Dev. 12, 468478. Jenkins, J.H., Carpenter-Song, E.A., 2009. Awareness of stigma among persons with schizophrenia: marking the contexts of lived experience. J. Nerv. Ment. Dis. 197, 520529. Klosterktter, J., Hellmich, M., Steinmeyer, E.M., Schultze-Lutter, F., 2001. Diagnosing schizophrenia in the initial prodromal phase. Arch. Gen. Psychiatry 58, 158164. Klosterktter, J., Ruhrmann, S., Schultze-Lutter, F., et al., 2005. The European prediction of psychosis study (EPOS): integrating early recognition and intervention in Europe. World Psychiatry 4, 161167. Larsen, T.K., Friis, S., Haahr, U., et al., 2001. Early detection and intervention in rst-episode schizophrenia: a critical review. Acta Psychiatr. Scand. 103, 323334. Lipsey, M.W., Wilson, D.B., 2001. Practical Meta-analysis. Sage Publications. Marshall, M., Rathbone, J., 2006. Early Intervention for Psychosis. Cochrane Database Syst. Rev. CD004718. McGlashan, T.H., Zipursky, R.B., Perkins, D., et al., 2003. The PRIME North America randomized double-blind clinical trial of olanzapine versus placebo in patients at risk of being prodromally symptomatic for psychosis, I: study rationale and design. Schizophr. Res. 61, 718. McGlashan, T.H., Zipursky, R.B., Perkins, D., et al., 2006. Randomized, doubleblind trial of olanzapine versus placebo in patients prodromally symptomatic for psychosis. Am. J. Psychiatry 163, 790799. McGorry, P.D., 2007. Issues for DSM-V: clinical staging: a heuristic pathway to valid nosology and safer, more effective treatment in psychiatry. Am. J. Psychiatry 164, 859860. McGorry, P.D., Yung, A.R., Phillips, L.J., et al., 2002. Randomized controlled trial of interventions designed to reduce the risk of progression to rstepisode psychosis in a clinical sample with subthreshold symptoms. Arch. Gen. Psychiatry 59, 921928. McGorry, P.D., Nelson, B., Amminger, G.P., et al., 2009. Intervention in individuals at ultra high risk for psychosis: a review and future directions. J. Clin. Psychiatry 70, 12061212. Miller, T.J., Mcglashan, T.H., Rosen, J.L., et al., 2003. Prodromal assessment with the structured interview for prodromal syndromes and the scale of prodromal symptoms: predictive validity, interrater reliability, and training to reliability. Schizophr. Bull. 29, 703715. Morrison, A.P., French, P., Walford, L., et al., 2004. Cognitive therapy for the prevention of psychosis in people at ultra-high risk. Br. J. Psychiatry 185, 291297. Morrison, A.P., French, P., Parker, S., et al., 2007. Three-year follow-up of a randomized controlled trial of cognitive therapy for the prevention of psychosis in people at ultrahigh risk. Schizophr. Bull. 33, 682687. Nordentoft, M., Thorup, A., Petersen, L., et al., 2006. Transition rates from schizotypal disorder to psychotic disorder for rst-contact patients included in the OPUS trial. A randomized clinical trial of integrated treatment and standard treatment. Schizophr. Res. 83, 2940. Odagaki, Y., 2009. Atypical neuroleptic malignant syndrome or serotonin toxicity associated with atypical antipsychotics? Curr. Drug Saf. 4, 8493. Olsen, K.A., Rosenbaum, B., 2006. Prospective investigation of the prodromal state of schizophrenia: review of studies. Acta Psychiatr. Scand. 113, 247272. Phillips, L.J., McGorry, P.D., Yuen, H.P., Ward, J., Donovan, K., Kelly, D., 2007. Medium term follow-up of a randomized controlled trial of interventions for young people at ultra high risk of psychosis. Schizophr. Res. 96, 2533. Raballo, A., Lari, F., 2009. Clinical staging: a new scenario for the treatment of psychosis. Lancet 374, 365367. Raballo, A., Sbye, D., Parnas, J., 2009. Looking at the schizophrenia spectrum through the prism of self-disorders: an empirical study. Schizophr. Bull. Jun 15. [Epub ahead of print]. doi:10.1093/schbul/sbp056. Rosenthal, R., 1979. The le drawer problem and tolerance for null results. Psychol. Bull. 86, 638641.

36

A. Preti, M. Cella / Schizophrenia Research 123 (2010) 3036 Yung, A.R., McGorry, P.D., 1996. The prodromal phase of rst-episode psychosis: past and current conceptualizations. Schizophr. Bull. 22, 353370. Yung, A.R., Phillips, L.J., Yuen, H.P., et al., 2003. Psychosis prediction: 12month follow up of a high-risk (prodromal) group. Schizophr. Res. 60, 2132. Yung, A.R., Phillips, L.J., Yuen, H.P., McGorry, P.D., 2004. Risk factors for psychosis in an ultra high-risk group: psychopathology and clinical features. Schizophr. Res. 67, 131142. Yung, A.R., Yuen, H.P., McGorry, P.D., et al., 2005. Mapping the onset of psychosis: the comprehensive assessment of at-risk mental states. Aust. N. Z. J. Psychiatry 39, 964971. Yung, A.R., Nelson, B., Stanford, C., et al., 2008. Validation of prodromal criteria to detect individuals at ultra high risk of psychosis: 2 year followup. Schizophr. Res. 105, 1017. Yung, A.R., Nelson, B., Thompson, A., Wood, S.J., 2010. The psychosis threshold in Ultra High Risk (prodromal) research: is it valid? Schizophr. Res. 120, 16.

Ruhrmann, S., Bechdolf, A., Khn, K.U., et al., LIPS study group, 2007. Acute effects of treatment for prodromal symptoms for people putatively in a late initial prodromal state of psychosis. Br. J. Psychiatry 51 (Suppl), S88S95. Ruhrmann, S., Schultze-Lutter, F., Klosterktter, J., 2010a. Probably at-risk, but certainly ill advocating the introduction of a psychosis spectrum disorder in DSM-V. Schizophr. Res. 120, 2337. Ruhrmann, S., Schultze-Lutter, F., Salokangas, R.K., et al., 2010b. Prediction of psychosis in adolescents and young adults at high risk: results from the prospective European prediction of psychosis study. Arch. Gen. Psychiatry 67, 241251. Schulz, K.F., Altman, D.G., Moher, D., CONSORT Group, 2010. CONSORT 2010 statement: updated guidelines for reporting parallel group randomized trials. Ann. Intern. Med. 152, 726732. Smith, M.L., Glass, G.V., 1977. Meta-analysis of psychotherapy outcome studies. Am. Psychol. 32, 752760. WHO (World Health Organization), 1992. Tenth Revision of the International Classication of Diseases and Related Health Problems (ICD10). WHO Press, Geneve. Yang, L.H., Wonpat-Borja, A.J., Opler, M.G., Corcoran, C.M., 2010. Potential stigma associated with inclusion of the psychosis risk syndrome in the DSM-V: an empirical question. Schizophr. Res. 120, 4248.