Welcome to the Schizophrenia Resource Centre

Welcome, this website is intended for healthcare professionals in EMEA with an interest in the treatment of schizophrenia. By clicking the link below you are declaring and confirming that you are a healthcare professional

You are here

Are there clinically useful predictors and early warning signs for pending relapse?

Schizophrenia Research, 2-3, 152, pages 469 - 477

Abstract

Objective

Despite the availability of effective long-term treatment strategies in schizophrenia, relapse is still common. Relapse prevention is one of the major treatment objectives, because relapse represents burden and costs for patients, their environment, and society and seems to increase illness progression at the biological level. Valid predictors for relapse are urgently needed to enable more individualized recommendations and treatment decisions to be made.

Methods

Mainly recent evidence regarding predictors and early warning signs of relapse in schizophrenia was reviewed. In addition, data from the first-episode (long-term) study (FES; Gaebel et al., 2007, 2011) performed within the German Research Network on Schizophrenia were analyzed.

Results

On the basis of FES data, premorbid adjustment, residual symptoms and some side effects are significant predictors. Although a broad spectrum of potential parameters has been investigated in several other studies, only a few and rather general valid predictors were identified consistently. Data of the FES also indicated that predictive power could be enhanced by considering interacting conjunctions, as suggested by the Vulnerability–Stress-Coping model. Prospective studies, however, are rare. In addition, prodromal symptoms as course-related characteristics likewise investigated in the FES add substantially to early recognition of relapse and may serve as early warning signs, but prognosis nevertheless remains a challenge.

Conclusions

Comprehensive and well-designed studies are needed to identify and confirm valid predictors for relapse in schizophrenia. In this respect, broadly accepted and specifically defined criteria for relapse would greatly facilitate comparison of results across studies.

Keywords: Schizophrenia, Relapse, Predictors, Prodromal symptoms, Early warning signs.

1. Introduction

After a first episode in schizophrenia, the further course is mostly characterized by re-exacerbation(s) of psychotic symptoms. About 80% of patients (Watt et al, 1983, Wiersma et al, 1998, and Robinson et al, 1999) suffer from symptom recurrence or relapse. Each exacerbation represents substantial burden for patients, their environment, and society in general ( Kane, 2007 ). Besides the stress caused by the symptoms themselves, symptomatic patients are at higher risk to harm themselves or others. In most cases, inpatient treatment is required and social and occupational re-integration is interrupted or hindered and thus prevents recovery. Relapse increases direct and indirect financial costs massively (Fitzgerald et al, 2009 and Ascher-Svanum et al, 2010); one study found that costs almost doubled in a 3-year observation period ( Hong et al., 2009 ). In addition, there is some evidence that each exacerbation has a negative impact on illness course at the neurobiological level (Lieberman et al, 2001 and Emsley et al, 2012; see also Emsley, in this issue) and hence contributes to further illness progression. Accordingly, relapse prevention is among the major objectives in long-term treatment after an acute episode in schizophrenia. In this regard, valid predictors for an (impending) relapse that are based on patient or course characteristics are essential to guide appropriate and more individualized treatment decisions.

On the basis of the heuristic framework of the Vulnerability–Stress-Coping (VSC) model (e.g. Zubin and Spring, 1977 and Nuechterlein and Dawson, 1984), relapse in schizophrenia is a result of a complex interaction of vulnerability factors, stress, and (insufficient) protecting factors that could be located on the biological, psychological, and environmental or social level. In addition, before the full re-exacerbation of psychotic symptoms, prodromal symptoms (can) occur as an intermediate state. Thus, a wide spectrum of potential predictors or indicators, including prodromal symptoms, can be expected for relapse in schizophrenia. The objective of this paper was to summarize (mainly recent) evidence regarding predictors for relapse in schizophrenia, including prodromal symptoms, against the background of the heuristic VSC model. In addition, the data from a randomized controlled trial (RCT) on different long-term treatment strategies in first-episode patients conducted between 2000 and 2006 (Gaebel et al, 2007 and Gaebel et al, 2011) were re-analyzed.

2. Methods

Regarding the literature review, relevant publications published between 2005 and October 2012 were identified through a PubMed search using the following (combined) search terms: Schizophrenia OR psychosis; relapse OR exacerbation OR outcome OR course; predictors OR indicators; prodromes OR early warning signs.

The “first-episode study” (FES) was conducted at 13 university centers across Germany, as part of the German Research Network on Schizophrenia (GRNS, Wölwer et al., 2003 ), a nationwide research network funded by the German Federal Ministry of Education and Research (BMBF). After 8 weeks of (acute) treatment of their first episode in schizophrenia with randomly allocated double-blind treatment with risperidone or low-dose haloperidol in the so-called “acute study” ( Möller et al., 2008 ), patients were included in the subsequent long-term study (for details see Gaebel et al., 2007 ). In the first post-acute year of this two-year trial, patients continued with their formerly allocated, double-blind treatment (risperidone or haloperidol) and were compared regarding differences in relapse rates (primary outcome) and other measures of efficacy and side effects ( Gaebel et al., 2007 ). Because of the substantial drop-out rate in the acute study phase, the design was amended at an early stage by adding a “lateral entry” procedure that allowed first-episode patients who had completed acute treatment but had not participated in the acute study to be included in the 1-year maintenance study. Lateral entry patients were randomly allocated to double-blind treatment with risperidone or haloperidol at entry into the first study year. In addition, a psychological treatment study was performed during the 1-year maintenance phase; its results also are relevant for the analyses presented here. In five of the 13 study centers, patients were randomly allocated to psychoeducation (PE) for 8 weeks or to cognitive behavioral therapy (CBT; for more details see Klingberg et al., in preparation ) for 12 months. In the subsequent second treatment year, patients who were stable after one year on antipsychotic maintenance treatment (MT) were randomly allocated to a further year of MT with their former antipsychotic or to intermittent treatment (IT, stepwise drug discontinuation and early drug intervention in case of prodromal symptoms or early warning signs of a relapse; Gaebel et al., 2011 ).

Relapse as the primary outcome measure was predefined as an increase in the Positive and Negative Syndrome Scale (PANSS; Kay et al., 1986 ) positive score > 10, a CGI (Clinical Global Impression, Guy, 1976a and Guy, 1976b) change score ≥ 6, and a decrease in GAF score > 20 between two visits. No patient fulfilled these relapse criteria ( Gaebel et al., 2007 ), and therefore different measures of “marked clinical deterioration” were added post hoc as less conservative outcome measures. The lowest criterion relevant here was fulfillment of one of the following six criteria between two visits: (1) increase in PANSS positive sum score ≥ 7 if sum score is ≥ 17; (2) increase in PANSS positive sum score ≥ 5 if sum score is ≥ 20; (3) at least 1 PANSS positive item is ≥ 5; (4) increase in CGI severity of symptoms ≥ 1 if score is ≥ 6; (5) increase in CGI severity of symptoms ≥ 2 if score is ≥ 5; and (6) decrease in GAF score ≥ 15.

2.1. Predictors (assessed at baseline) for relapse within 1 year of post-acute maintenance treatment

The following characteristics (assessed at baseline/entry into the post acute 1-year maintenance phase) were included as potential predictors: age at study entry; gender; age at illness onset; substance abuse (yes vs. no); SCPS (Strauss–Carpenter Prognosis Scale; Strauss and Carpenter, 1978 ); SAS-II (Social Adjustment Scale II; Schooler et al., 1979 ); GAF (Global Assessment of Functioning; Frances et al., 1994 ), and SOFAS (Social and Occupational Functioning Assessment Scale; Goldman et al, 1992 and American Psychiatric Association (APA), 2000) at study entry; lowest GAF and SOFAS score in the year before study entry; PANSS positive, negative, and general score; CGI change score; HAM-D (Hamilton Rating Scale for Depression, Hamilton, 1960 ); EPS (Extrapyramidal Side Effects scale; Simpson and Angus, 1970 ); HAS (Hillside Akathisia Scale; Fleischhacker et al., 1989 ); AIMS (Abnormal Involuntary Movement Scale; Guy, 1976a and Guy, 1976b); the UKU (Udvalg for Kliniske Undersogelser Side Effect Rating Scale; Scand Soc. of Psychopharm., 1987 ) subscores “neurological,” “psychological,” “autonomous,” and “other” side effects; compliance (CRS; Compliance Rating Scale; Kemp and David, 1996 ); DAI (Drug Attitude Inventory; Hogan et al., 1983 ); mode of study entry (from acute phase vs. lateral entry); participation in the psychological study (yes vs. no); and study drug (risperidone vs. haloperidol). A logistic regression analyses was conducted with a stepwise forward procedure, with alpha levels for inclusion of p ≤ 0.05 and for exclusion of p ≤ 0.1.

2.2. Examining interactive conjunctions according to the VSC-model

Regarding the analyses of interactive conjunctions as proposed by the VSC model, different parameters are monitored and assessed within the GRNS-FES. Relevant here are stressful events, which were assessed at every visit (about every two weeks) with the Munich Event List (MEL; Maier-Diewald et al., 1983 ). This list records the occurrence of stressful events, their valence for the person (positive vs. negative), and the respective burden (from 0 = “not at all” to 5 = “very much”). Patients' coping abilities were assessed (at study entry) with the German coping questionnaire SVF ( Janke et al., 1985 ), which assesses coping abilities in 20 (sub-)domains. Another questionnaire was used to assess family atmosphere regarding criticism, disinterest or resignation, and overprotection (FEF, Feldmann et al., 1995 ). In addition, occurrence of mental disorders in patients' relatives was assessed and used as an indicator for a higher vulnerability at the biological level.

In a first step, univariate analyses were used to test whether different parameters were significant predictors for deterioration (by testing for significant differences between patients with vs. patients without deterioration in these parameters; t tests were used for metric parameters and chi-square tests for categorical measures). On the basis of these results, significant predictors were selected and combined.

2.3. Relapse predictive validity of prodromal symptoms

Another objective of the GRNS-FES was to examine and enhance the relapse predictive validity of prodromal symptoms. At every visit (scheduled every 2 weeks), a broad spectrum of 45 prodromal symptoms (see Table 2 ) was assessed on a 4-point Likert scale (from 0 = “not present” to 3 = “severe”) in the first (MT for all patients) and second (MT or IT) post-acute year. Thus, the presence of prodromal symptoms could be cross-tabulated with the presence of deterioration at the subsequent visit (scheduled two weeks later; mean interval 14.5 days; SD = 3.8) for a period of up to 2 years in long-term treatment. These cross-tabulations were used to calculate values for sensitivity (rate of visits with prodromal symptoms prior all visits with deterioration) and specificity (rate of visits with no prodromal symptoms prior all visits with no deterioration).

All analyses were conducted with IBM SPSS Statistics (V20).

All results regarding predictors for relapse, including prodromal symptoms, have not been published elsewhere and are published here for the first time.

3. Results

3.1. Relapse in schizophrenia: operational definition and prevalence rates

The (operational) definition and assessment of relapse in schizophrenia are very heterogeneous, which limits comparison of study results. The criterion or indicator most often used is hospital (re-)admission ( Gleeson et al., 2010 ); however, rates of symptom relapse and readmission to hospital differ greatly (e.g. Leucht et al., 2012 ), and influencing factors and predictors for relapse and hospital (re-)admission also are diverse ( Falloon et al., 1983 ). Early attempts to assimilate criteria for assessment of relapse ( Falloon, 1984 ) emphasized that criteria should consider: 1) an increase of positive symptoms (from a post-acute, improved, sub-clinical level) above a clinically relevant threshold (which should be defined a priori); 2) an appropriate duration of emerging clinical symptoms; and 3) assessment of symptoms with objective and standardized instruments. In a recent review on the reporting of relapse in RCTs in first-episode psychosis, Gleeson et al. (2010) added that 4) intervals for symptom assessment and the follow-up period should be appropriate [together with additional criteria addressing mainly methodological issues based on the CONSORT statement for the reporting of RCTs (e.g. Moher et al, 2001 and Schulz et al, 2010), namely that: 5) symptoms should be assessed by trained (and blinded) raters; and 6) inter-rater reliability checks should be conducted]. A very comprehensive operationalization for relapse was used by Csernansky et al. (2002) and decisive for different other studies. They defined relapse as the presence of at least one of the following: psychiatric hospitalization; an increase in the level of psychiatric care and an increase of 25% from baseline in the PANSS total score (or an increase of 10 points if the baseline score was 40 or less); deliberate self-injury; suicidal or homicidal ideation; clinically significant violent behavior; substantial clinical deterioration (as indicated by a CGI change score ≥ 6). Most RCTs in the review by Gleeson et al. (2010) do not comply with their criteria, so the authors state that there is an “urgent need for a standardized, universally adopted set of criteria for psychotic relapse” (p. 79). For example, the Remission in Schizophrenia Working Group (RSWG; Andreasen et al., 2005 ) defined and consented such a set of criteria for remission that stimulated a large amount of research and has enabled a better comparison of results across studies ( Gaebel et al., 2013 ).

Reported rates of relapse vary also mainly depending on the assessment criteria used, study type (RCTs or epidemiological studies), treatment conditions, sample characteristics (first- vs. multiple-episode patients), and observation period. A recent meta-analytic review on RCTs that compared antipsychotic (post-acute) MT with placebo regarding relapse prevention estimated 1-year relapse rates of 27% versus 64%, and 2- to 3-year rates of 44% versus 79%, respectively ( Leucht et al., 2012 ). Interestingly, results were similar for patients with symptom remission before the relapse observation period (30% vs. 63%), patients who did not fulfill such a criterion (26% vs. 64%), first-episode (26% vs. 61%) and multiple-episode patients (27% vs. 65%), and treatment with first-generation antipsychotics (FGA; 24% vs. 62%) and second-generation antipsychotics (SGA; 30% vs. 67%; MT vs. placebo respectively). However, in other meta-analyses some SGAs (olanzapine, risperidone, sertindole) showed significant advantages in relapse prevention compared to FGAs (mainly haloperidol; Leucht et al., 2009 ), and relapse rates with depot or long-acting injectable antipsychotics were significantly lower than with oral antipsychotics (22% vs. 33%; Leucht et al., 2011 ), whereas a very recent meta-analysis found no differences between oral and depot antipsychotics ( Kishimoto et al., 2014 ).

In contrast to placebo-controlled RCTs, epidemiological and observational studies correspond more closely to the natural clinical setting. A recently published review on first-episode psychosis (including schizophrenia, schizophrenia spectrum disorders, and affective disorders with psychotic symptoms) included 29 mainly epidemiological or observational studies ( Alvarez-Jimenez et al., 2012 ); estimated pooled relapse rates were 28% in the first post-acute year, 43% in the second, and 54% in the third.

3.2. Predictors for relapse in schizophrenia

Predictors in general are defined as variables or parameters that “systematically” precede an event and hence are related i.e. correlated to the occurrence of that event ( Gaebel, 2004 ). This relationship can be causal (in extreme cases “deterministic”) or, as in many cases, not causal, or the relationship even may remain unknown. Predictors may derive from theoretical assumptions regarding the illness process (e.g. according to the heuristic framework of the VSC-model, e.g. Zubin and Spring, 1977 and Nuechterlein and Dawson, 1984) or emerge in empirical studies as a result of correlation analyses. Predictors often are summarized as patient characteristics—such as more general sociodemographic variables (e.g. sex, age, ethnicity), personality traits, or premorbid characteristics—and as more illness-related clinical characteristics (e.g. symptom severity, illness duration), that include also genetic, biological (e.g. parameters from neuroimaging data), and cognitive parameters. In addition to patient variables, other predictors refer to environmental factors (e.g. stressful life events), course variables (e.g. response, remission, prior relapse), treatment characteristics (see above), or interactions between these categories (e.g. adherence to treatment).

Predictors for relapse in schizophrenia have been studied for several decades. In 1998, Doering et al. (1998) summarized predictors for relapse that had been identified at the time of their study and found the following as “favorable” (lowered risk for relapse): female gender, better (premorbid) social adjustment (being in partnership, higher level of education, better functional and occupational status), acute onset of (first) psychotic symptoms, paranoid or hallucinatory syndrome, no previous relapse, family or relatives with low scores in “expressed emotions” (EE). Indicators for a higher risk for relapse were a family history of schizophrenia, (premorbid) schizoid personality, early age at onset, (pronounced) negative symptoms, and longer duration of inpatient treatment.

Several studies on predictors for relapse have been published in recent years (since 2005). Four studies in multiple-episode patients aimed to identify significant predictors (post hoc) from a set of different variables that were assessed to examine other (primary) hypotheses (DiMichele et al, 2007, Haro et al, 2008, Ascher-Svanum et al, 2010, and Schennach et al, 2012). Only one variable (premorbid adjustment, although operationalized by different indicators) was analyzed in all four studies and evolved also as a significant predictor in all. Two studies examined (non-remitting) positive symptoms as predictors for relapse and obtained significant results (DiMichele et al, 2007 and Ascher-Svanum et al, 2010). All other (reported significant) predictors were examined in only one of the studies: Haro et al. (2008) identified (male) gender as an additional predictor for a higher relapse risk; Ascher-Svanum et al. (2010) identified lower age, prior relapse, non-adherence to drug treatment, and co-morbidity with a substance abuse disorder; and Schennach et al. (2012) identified more negative symptoms and higher depression scores, treatment with FGAs (vs. SGAs), more side effects, and a less positive attitude toward drugs.

Several other studies in multiple-episode patients evaluated the relapse-predictive potential of specific parameters. Lipkovich et al. (2007) reanalyzed two studies of olanzapine treatment to evaluate whether decreasing the drug dose (at the physician's discretion but within the allowed dose range) is predictive of relapse by the next visit (after 4 weeks). They found a significant, four-fold higher risk, but only in male patients. Levene et al. (2009) examined the interaction of patient and family characteristics in a sample of male patients and found that both (non-)remitted positive symptoms and higher emotional burden of family members (although not expressed emotions) are equally predictive and significant. Cechnicki et al. (2011) evaluated the duration of untreated psychosis (DUP) as a predictor for relapse and illness course in schizophrenia: In an observation period of up to 12 years, patients with a longer DUP showed significantly more relapse than patients with a shorter DUP.

In the last 15 years, an increasing number of studies have been conducted in first-episode or recent-onset schizophrenia, because the first years of illness are proposed as a critical period for the further course ( Birchwood et al., 1998 ). Also, at this stage the illness is less influenced by the illness course itself, including selection effects and treatment conditions. Three studies aimed to identify (general) predictors for relapse among a broad set of parameters in this specific patient group (Robinson et al, 1999, Uçok et al, 2006, and Caseiro et al, 2012). Again, the set of variables analyzed was rather heterogeneous, and only a small number of variables were included in all studies (premorbid [childhood] adjustment; gender; post-acute positive and negative symptoms; and treatment adherence). Only treatment non-adherence was a significant predictor for relapse in all three studies. Social adjustment in childhood was significant in two studies (Robinson et al, 1999 and Uçok et al, 2006), whereas premorbid adjustment in general was not significant in any of the studies. Also, positive symptoms were not significant in any study, and negative symptoms and (male) gender were significant in only one ( Uçok et al., 2006 ). Two studies evaluated DUP (Uçok et al, 2006 and Caseiro et al, 2012) but did not find it to be significant. Other parameters were tested in single studies: treatment with SGAs showed a significantly lower relapse risk than treatment with FGAs ( Uçok et al., 2006 ); brain abnormalities as indicated by MRI assessments showed no significant effects on relapse rate ( Robinson et al., 1999 ); and illness insight and comorbidity with substance abuse disorders also were not significant ( Caseiro et al., 2012 ). The importance of treatment adherence for risk of relapse was underlined in a study specifically designed to evaluate this variable ( Subotnik et al., 2011 ); the study found that even brief periods with only partial adherence to antipsychotic treatment significantly increase risk for relapse. Other studies that examined specific parameters found significant results for illness insight ( Drake et al., 2007 ) and personality traits (agreeableness and neuroticism; Gleeson et al., 2005 ), but not for brain volume ( van Haren et al., 2003 ). In addition, four studies examined the predictive value of neuropsychological or neurocognitive assessments (Chen et al, 2005, Holthausen et al, 2007, Rund et al, 2007, and Wölwer et al, 2008); three found significant results for specific parameters (Chen et al, 2005, Rund et al, 2007, and Wölwer et al, 2008; for a summary see Alvarez-Jimenez et al., 2012 et al., 2012), and one study did not ( Holthausen et al., 2007 ).

Alvarez-Jimenez et al. (2011) analyzed data from a study in first-episode psychosis (including also affective disorders with psychotic symptoms) over a 7.5-year observation period with the (opposite) objective to identify patients with no risk for relapse. They aimed to identify parameters that are indicative of the about 20% of patients who develop only one illness episode and thereafter remain stable throughout their lives. Of the 274 patients with sufficient follow-up data, 16.5% had no relapse in the subsequent 7.5 years. These patients were characterized by the following significant predictors: better premorbid (childhood) adjustment (general adjustment, higher education, no parental loss), a shorter DUP, rapid onset of psychotic symptoms, a favorable initial response, a diagnosis of an affective psychosis, and less pronounced (post-acute) negative symptoms. Among others, gender and non-remitting positive symptoms did not reach the significance level. The authors provide no information on maintained drug treatment or treatment adherence but, given the observational design, one could assume that drug treatment was at the doctors' and patients' discretion (including the decision whether or not antipsychotic treatment was maintained).

3.3. Results of the first-episode long-term study within the German Research Network on Schizophrenia (GRNS-FES)

Regarding the analyses for predictors of relapse based on the first post-acute year of the GRNS-FES, a total of 135 first-episode patients were eligible for inclusion in this analysis. Their average age was 31.7 years (SD = 9.9), 78 (57.8%) were male, 115 (85.2%) had participated in the preceding acute trial, and 84 (62.2%) had participated in the associated psychological trial (39 patients in the 8-week PE arm and 45 patients in the 12-month CBT arm). Sixty-seven patients (49.6%) were randomly allocated to risperidone treatment (mean dose of 4.2 mg/d [SD = 2.2]) at study entry; 68 (50.4%), to haloperidol treatment (mean dose 4.1 mg/d, SD = 2.2; for further details see Gaebel et al., 2007 ).

Twenty-six patients (19.3%) showed deterioration according to the above mentioned criteria. Based on logistic regression, the following predictors evolved as significant in univariate analyses: lowest score in social and occupational functioning (SOFAS) in the preceding year (OR = 0.91; 95% CI = 0.86–0.96; p < 0.001); lowest score in global functioning (GAF) in the preceding year (OR = 0.93; 95% CI = 0.89–0.96; p = 0.003); global functioning (GAF) at entry into the long-term study (OR = 0.96; 95% CI = 0.93–0.99; p = 0.012); illness severity according to CGI at entry into the long-term study (OR = 1.68; 95% CI = 1.10–2.55; p = 0.016); PANSS positive score at entry into the long-term study (OR = 1.09; 95% CI = 1.01–1.19; p = 0.033); and SOFAS score at entry into the long-term study (OR = 0.97; 95% CI = 0.94–1.0; p = 0.047). In addition, some parameters had a significance level > 0.05 but < 0.1, indicating a trend for an association: “other” side effects according to UKU (p = 0.068), social adjustment (SAS-II; 0.06), and attitude toward drugs (DAI; p = 0.085). On the basis of the multivariate (stepwise forward) regression, the following predictors were included as significant in the model: lowest SOFAS score in the preceding year (OR = 0.89; 95% CI = 0.84–0.94; p < 0.001); “other” side effects according to UKU (OR = 1.50; 95% CI = 1.04–2.16; p = 0.029), age at study entry (OR = 1.06; 95% CI = 1.01–1.12; p = 0.026), and non-participation in the accompanying psychological study (OR = 2.83; 95% CI = 1.02–7.87; p = 0.046; with no difference between PE and CBT).

3.4. Considering interaction of parameters according to the Vulnerability–Stress-Coping (VSC) model as a predictor for relapse

According to the VSC model for schizophrenia (e.g. Zubin and Spring, 1977 and Nuechterlein and Dawson, 1984), psychotic (re-)exacerbations are a result of interacting variables such as vulnerability at the biological level, stress from environmental factors (moderated by psychological processes of appraisal), and availability of protecting factors at the biological (e.g. antipsychotic drugs) or psychological level (coping abilities). Thus, prediction of relapse in schizophrenia might be enhanced by considering these parameters in an interactive manner, rather than these (or other) predictors alone. The abovementioned study by Levene et al. (2009) supports this view since both (non-)remitted positive symptoms (indicating a disturbed appraisal and higher vulnerability) and higher emotional burden of family members (indicating environmental stress) were equally predictive and significant. The authors only describe that patients were “adequately medicated throughout the study,” without addressing adherence, so it remains unclear whether adherence contributed to the study results. In a similar study, Docherty et al. (2009) examined the interactive effect of emotional reactivity and stressful life events on symptom exacerbation. They found that the occurrence of stressful life events alone increases the risk of exacerbations, whereby patients with both high stress and high reactivity showed the greatest symptom increase. In a similar study by this group ( Docherty et al., 2011 ), patient anxiety and critical relatives alone increased the risk of symptom exacerbations, although the presence of both factors together showed the greatest symptom increase.

The first-episode study performed within the GRNS examined different parameters of the VSC model, and thus different interactive conjunctions could be investigated. The respective analyses included 159 first-episode patients with a mean age of 31.7 years (SD = 10.2); 91 (57.2%) were male. Forty-one patients (25.8%) participated also in the second year of the FES long-term study, in which drug treatment was maintained or discontinued in a stepwise fashion (open, random design; prodrome-based early intervention in both groups; for details see Gaebel et al., 2011 ); 20 patients were assigned to IT (intermittent treatment after stepwise drug discontinuation). The data from this second year also were included in the analyses described below. Overall, deterioration (according to the above defined criteria) was observed in 40 patients (25.2%), whereby 12 patients deteriorated (only) in the second year (10 after drug discontinuation).

Initial univariate analyses were used to test whether different parameters were significant predictors for deterioration. Regarding (higher) vulnerability, no significant difference resulted in patients with family members having a mental disorder vs. those with no such relatives were compared. As regards to coping abilities, several domains of the coping assessment instrument (SVF) were significant predictors for deterioration (“minimizing”, “downplaying”, “defense of guilt”, and “compensatory gratification”; all p < 0.05). In a Cox regression analysis, “minimizing” (i.e. “downgrading intensity, duration, or impact of stressful events”) remained the only significant variable in a multivariate analysis. None of the three parameters of family atmosphere (assessed by FEF) reached significance. Regarding other environmental stresses, the overall score in the MEL (sum of subjective burden of all negative events) was significant (p < 0.05). Regarding protective factors, the number of visits at which patients were prescribed a (very) low dose of antipsychotics (risperidone or haloperidol ≤ 1.5 mg/d) was counted. The number of “lower-dose visits” was not found to be a significant predictor for deterioration for the total sample; however, it became significant in the sample of patients also included in the second study year (stepwise drug discontinuation vs. further MT; p < 0.05). In contrast, not participating or participating in the psychological trial (8 weeks' PE or 12 months' CBT) was not significantly related to deterioration.

In the second step, these significant predictors were selected and combined. In addition, we decided to include “higher vulnerability” in the further analyses to examine also the interaction of factors that had not been significant per se. Firstly, patients with lower and higher vulnerability were cross-tabulated with occurrence of low and high stress, resulting in cell frequencies for the respective prevalence rates for deterioration (see Fig. 1 ). The highest deterioration rate of 43.8% was found in the group characterized by higher vulnerability and high stress; however, this rate was just above the significance level (p = 0.051) when compared to the deterioration rate of all other groups (21.2%). Likewise, cross-tabulating coping abilities (low vs. high “minimizing”) with stress (low vs. high, see Fig. 2 ) resulted in the highest deterioration rate in the low coping/high stress group (52.4%); the rate was significantly higher than the other three combinations (19.1%; p = 0.003). The same picture evolved when patients treated several times with a very low dose of antipsychotics and patients who were not given such treatment were cross-tabulated with the occurrence of low vs. high stress ( Fig. 3 ). Again, the deterioration rate was significantly higher (p = 0.02) in the very low dose/high stress group (38.1%) than in the other three groups (15.6%). All these figures illustrate that one factor or condition associated with an elevated risk for relapse or deterioration results only in a small risk increase, whereas the combination of two conditions leads to noticeable and significantly higher deterioration rates. Fig. 4 summarizes all the results and gives rates of deterioration for different parameters or combinations of them. For comparison, the overall prevalence rate of deterioration of 25% also is given. Significant differences were found for patients with high stress (30%, p = 0.02), with low “minimizing” (40%, p = 0.02), with combinations of: lower dose antipsychotic treatment and high stress (38.1%; p = 0.02); higher vulnerability and high stress (43.8%, p = 0.051); and low coping and high stress (52.4%, p = 0.003). Finally, a multivariate Cox regression analysis was performed with all these parameters, with time to deterioration as the dependent variable; only the combination of low coping and high stress was significant (p < 0.001). This illustrates that studying the interaction of parameters, as suggested by the VSC model, could enhance the prediction of relapse or deterioration in schizophrenia.

gr1

Fig. 1 Rates of deterioration (%) for different patient groups characterized by “vulnerability” (higher vulnerability if patient has relatives with mental disorders, lower vulnerability if not) and stress (high or low stress if Munich Event List score is high or low, respectively).

gr2

Fig. 2 Rates of deterioration (%) in different patient groups characterized by the coping strategy “minimizing” (assessed with the German stress questionnaire “SVF”; a high score for minimizing indicates high coping skills; a low score, low coping skills) and stress (high or low stress if Munich Event List score is high or low, respectively).

gr3

Fig. 3 Rates of deterioration (%) for different patient groups characterized by administration of lower dose antipsychotic treatment (≤ 1.5 mg/d risperidone or haloperidol; in some visits yes vs. no) and stress (high or low stress if Munich Event List score is high or low, respectively).

gr4

Fig. 4 Rates of deterioration (%) in different patient groups characterized by different prognostic parameters or interactions of them.

3.5. Prodromal symptoms or early warning signs as predictors for relapse in schizophrenia

According to the VSC model, (re-)exacerbations of psychotic symptoms are regularly preceded by rather unspecific symptoms like tension, nervousness, or sleep disturbances, which are subsumed under the concept of prodromal symptoms. Hence, these symptoms likewise (may) predict relapse in schizophrenia and are considered as “early warning signs.” In addition to these symptoms, which are rather unspecific for schizophrenia, more psychosis-related (or attenuated positive) symptoms like mistrust, ideas of reference, suspiciousness, and paranoid thinking also are included ( Bustillo et al., 1995 ). Several studies have examined empirically the relapse predictive validity of prodromal symptoms, and the results were recently summarized elsewhere ( Gaebel and Riesbeck, 2007 ; see also Table 1 ). The authors concluded that relapse prediction based on prodromal symptoms remains rather uncertain and needs to be enhanced. In particular, relapses or exacerbations may not be preceded by prodromal symptoms (sensitivity noticeably lower than 100%), and prodromal symptoms are rather prevalent and often not followed by exacerbations (specificity likewise markedly lower than 100%).

Table 1 Prospective studies examining the predictive validity of prodromal symptoms or early warning signs for relapse in schizophrenia (adapted from Gaebel and Riesbeck, 2007 ).

Author(s) Sample size (n) Relapsing patients (n) Sensitivity Specificity
Subotnik and Nuechterlein (1988) 50 17 59%
Birchwood et al. (1989) 19 8 63% 82%
Hirsch and Jolley (1989) 54 10 73%
Tarrier et al. (1991) 84 16 50% 81%
Marder et al. (1994) 80 42 37%
Malla and Norman (1994) 55 24 50% 90%
Jorgensen (1998) 60 27 81% 79%
Gleeson et al. (2005) a 46 8 80% 73%
Gaebel and Riesbeck (2007) 339 153 72% b 38% b

a Early warning signs at any time point prior to a relapse.

b Overall sum score with lowest cut-off 0 vs. ≥ 1.

In the meantime, one pilot study in six patients aimed to improve prognosis by including psychophysiological measures: Dawson et al. (2010) found that patients exhibited a heightened skin conductance level within 2 weeks before a relapse or exacerbation (compared to visits prior non-relapse).

Another objective of the GRNS-FES was to examine and enhance the relapse predictive validity of prodromal symptoms. At every visit (scheduled every 2 weeks), 45 prodromal symptoms (see Table 2 ) were assessed and their presence was cross-tabulated with the presence of deterioration at the subsequent visit. These cross-tabulations were used to calculate values for sensitivity (rate of visits with prodromal symptoms prior all visits with deterioration) and specificity (rate of visits with no prodromal symptoms prior all visits with no deterioration). Overall, the 159 patients made more than 3650 visits—about 2300 in the first study year and about 1300 in the second year—and deterioration was observed at 42 visits (1.2%) of 40 patients (i.e. 25.2% of all 159 patients). The predictive validity of the single prodromal symptoms and overall prodrome scores (for different cut-off values) regarding deterioration (according to the criteria given above) at the subsequent visits is given in Table 2 .

Table 2 Predictive values of unspecific and specific prodromal symptoms or early warning signs for deterioration (the following visit, scheduled every 2 weeks); arranged in order of decreasing sensitivity.

  Sensitivity Specificity
Unspecific prodromal symptoms    
Trouble concentrating 66.7 70.7
Tense and nervous 61.9 75.3
Restlessness 52.4 78.3
Social withdrawal 47.6 69.9
Mood swings 45.2 79.2
Loss of interests 45.2 76.9
Loss of happiness 45.2 79.8
Worrying 45.2 76.5
Irritability 42.9 88.9
Depressed mood 42.9 74.4
Loss of power 40.5 84.7
Forgetfulness 38.1 88.0
Narrowing of interests 35.7 83.2
Trouble sleeping 28.6 82.5
Shyness 28.6 79.5
Neglect of outer appearance 28.6 92.7
Increased consumption of coffee, alcohol or drugs 23.8 92.4
Indisposition without reason 23.8 91.4
Anxiety 23.8 88.8
Loss of appetite 21.4 94.9
Increased activity 16.7 95.1
Feeling of worthlessness/suicidal thoughts 14.3 92.8
Sum of all unspecific prodromal symptoms:    
 Cut-off 0 vs. ≥ 1 95.2 39.6
 Cut-off ≤ 1 vs. ≥ 2 90.5 47.0
 Cut-off ≤ 2 vs. ≥ 3 85.7 53.3
 Cut-off ≤ 3 vs. ≥ 4 81.0 58.5
 Cut-off ≤ 4 vs. ≥ 5 69.0 63.5
 
Specific prodromal symptoms
Mistrust 33.3 89.4
Behavior or appearance that is odd, eccentric or peculiar 26.2 97.1
Decreased ability to discriminate between ideas and perception, fantasy and true memories 21.4 96.7
Odd thinking and speech 21.4 96.7
Fear of going mad 19.0 96.7
Ideas of reference 19.0 93.8
Unstable ideas of reference (subject-centrism) 19.0 93.7
Suspicion and paranoid thinking 16.7 93.3
Disturbances of receptive language 16.7 97.5
Sensitivity to noise 16.7 95.5
Impression of being controlled 14.3 96.2
Thought interference 14.3 97.2
Odd beliefs or magical thinking 14.3 97.9
Absent-mindedness 14.3 98.3
Compulsory perseveration 9.5 97.8
Thought pressure 9.5 98.1
Thought blocking 9.5 97.5
Acoustic perceptual disturbances 9.5 97.3
Derealization 7.1 97.6
Hallucinations 7.1 97.9
Unusual perceptual experiences 2.4 99.0
Increasing religiosity 0.0 99.3
Visual perceptual disturbances 0.0 99.6
Sum of all specific prodromal symptoms:    
 Cut-off 0 vs. ≥ 1 47.6 80.1
Sum of all (specific and unspecific) prodromal symptoms:    
 Cut-off 0 vs. ≥ 1 95.2 39.9
 Cut-off ≤ 1 vs. ≥ 2 90.5 47.7
 Cut-off ≤ 2 vs. ≥ 3 85.7 54.1
 Cut-off ≤ 3 vs. ≥ 4 81.0 59.6
 Cut-off ≤ 4 vs. ≥ 5 66.7 65.0

Overall, unspecific prodromal symptoms showed noticeably higher sensitivity (highest for “trouble concentrating”: 66.2%) than more specific (attenuated positive) symptoms, whereas specificity scores were higher for specific prodromes. Correspondingly, the sum score of all unspecific prodromes reached a sensitivity of up to 95% but a specificity of only about 40%, and the sum score of all specific prodromes reached a sensitivity of 48.8% and a specificity of 80.1%. To achieve as high a sensitivity as possible (to “detect” as many impending relapses as possible) while simultaneously being as specific as possible (to avoid too many “false alarms”), a cut-off point of ≤ 3 vs. ≥ 4 might be chosen for the sum score of all unspecific prodromal symptoms; this would result in a sensitivity of about 80% (81.0%) and a specificity of around 60% (58.5%). Nevertheless, even if the predictive validity of prodromal symptoms may be enhanced by comprehensive and continuous assessment, early detection of relapse continues to represent a challenge. Some deteriorations or relapses (about 20%) are not preceded by prodromal symptoms until about 2 weeks before. On the other hand, (in particular unspecific) prodromes are rather frequent (in our study a sum score of unspecific prodromes of ≥ 4 occurred in 40% of all visits) and deterioration or relapse is very rare (present at only about 1% of all visits in our study), resulting in a substantial number of “false alarms.”

4. Discussion and conclusions

Despite the availability of effective long-term treatment strategies in schizophrenia, relapse is still common. Because relapse represents burden and costs for patients, their environment, and society in general and seems to increase illness progression at the biological level, relapse prevention is one of the major treatment objectives. Thus, valid predictors for relapse are urgently needed to enable more individualized recommendations and treatment decisions to be made. However, although a broad spectrum of potential predictors has been investigated, only a few and rather general valid predictors have been identified. Based on the literature review, relapse rates in first- and multiple-episode patients increase steadily from 20% to 30% in the first year to 45% in the second and 55% in the third, reaching a plateau at 75% to 80% after five years (Scottish Schizophrenia Research Group, 1992, Wiersma et al, 1998, and Robinson et al, 1999). Besides assessment criteria, administration of (antipsychotic) MT and the duration of the observation period after the acute phase have evolved as major contributing factors (and likewise predictors) for relapse.

Regarding single predictors for relapse in schizophrenia, the results of the GRNS-FES are in the line with other study results (Uçok et al, 2006, DiMichele et al, 2007, and Ascher-Svanum et al, 2010). The fact that compliance was not a significant predictor in our data (p = 0.8 in univariate analysis) is possibly due to the (very) positive compliance ratings for nearly all patients, which resulted in a diminished variance. Overall, one could conclude that besides some strong and highly valid predictors (favorable i.e. lowered risk for relapse: better premorbid functioning, perhaps as early as childhood; antipsychotic maintenance treatment and good adherence to it), the many other very heterogeneous parameters have been tested less uniformly or yielded rather inconsistent results. Relapse risk can be said to be somewhat higher in male patients and in patients with a longer DUP (regardless of the underlying mechanisms), with more pronounced (post-acute) negative symptoms or non-remitting positive symptoms, more cognitive impairments, and comorbid substance use disorders. These findings correspond with the results of a recent review on predictors for relapse in first-episode or recent-onset psychosis (including affective disorders; Alvarez-Jimenez et al., 2012 ). Thus, based on our review; predictors identified so far with consistent results are rather general, while other predictors yielded single or inconsistent results. Overall, this does not meet requirements for a “valid” prognosis, in particular in individual cases, and so treatment implications and recommendations still remain rather general (e.g. the better do better; antipsychotic treatment should be maintained after the acute phase).

Prognostic validity may be enhanced by considering interacting conjunctions as suggested by the VSC model; however, respective studies are rare. Prodromal symptoms as course-related characteristics add substantially to early recognition of relapse and could serve as early warning signs, but prediction of relapse nevertheless remains a challenge. Comprehensive and well-designed studies are needed to identify and confirm valid predictors for relapse in schizophrenia. In this respect, broadly accepted and specifically defined criteria would greatly facilitate comparison of results across studies.

Some limitations of the analyses of the GRNS-FES-data have to be considered. Different statistical tests were conducted on the data from our study cohort, so multiple testing with a noticeably higher probability of a type I error is given. We did not adjust the alpha level, since all analyses were explorative, and thus we did not want to increase the risk of a type II error and to overlook significant results. Some variables that were included as predictors, like compliance and drug dose, have diminished variance, which could be a reason for the non-significant results. In addition, “relapse” according to the pre-defined criteria did not occur, so we predicted “marked clinical deterioration” instead. Nevertheless, this condition represents also a noticeable symptom re-exacerbation, and the criteria we used are in the range of those used to define relapse in other studies (see Gleeson et al., 2010 ). Also, our results are in the line with those of other studies reviewed.

Role of funding source

This study was conducted within the German Research Network on Schizophrenia, which was funded by the German Federal Ministry for Education and Research BMBF (grants 01 GI 9932 and 01 GI 0232).

Study design and execution, data acquisition and analyses and preparation of the manuscript were completely independent from funding.

Contributors

Wolfgang Gaebel and Mathias Riesbeck contributed equally to the manuscript.

Conflict of interest

Dr. Gaebel reports having received symposia support from Janssen-Cilag GmbH, Neuss, Lilly Deutschland GmbH, Bad Homburg and Servier, Munich and being a faculty member of the Lundbeck International Neuroscience Foundation (LINF), Denmark. Mr. Riesbeck reports no financial or other relationships relevant to the subject of this article.

Acknowledgment

The authors are much obliged to the members of the German Study Group on First-Episode Schizophrenia as well as to the members of the scientific advisory board of the German Research Network on Schizophrenia for all their contributions. We thank also Jacquie Klesing, ELS, for editing assistance with the manuscript.

The German Study Group on First-Episode Schizophrenia consists of: W. Gaebel (P.I.), W. Wölwer, M. von Wilmsdorff, R. Krohmer, J. Brinkmeyer, and M. Riesbeck (Duesseldorf); A. Klimke (Offenbach) and M. Eickhoff (Warstein/Lippstadt); H.-J. Moeller and M. Jäger (Munich); G. Buchkremer, S. Klingberg, and M. Mayenberger (Tuebingen); P. Hoff and F. Schneider (Aachen; recruitment until 6/2002); W. Maier, M. Lemke, B. Johannwerner and K.-U. Kühn (Bonn); I. Heuser and M.C. Jockers-Scherübl (Berlin); J. Klosterkötter, A. Bechdolf, and W. Huff (Cologne); M. Gastpar, S. Bender, and V. Reissner (Essen); E. Rüther and D. Degner (Goettingen); P. Falkai (Goettingen); H. Sauer, R. Schlösser, and G. Wagner (Jena); F. A. Henn, H. Häfner, K. Maurer, H. Salize, and A. Schmitt (Mannheim); and L.G. Schmidt (Mainz; recruitment until 2/2002).

The members of the scientific advisory board of the German Research Network on Schizophrenia are as follows: Prof. Dr. A. G. Awad, Toronto (CAN); Prof. Dr. W. Fleischhacker, Innsbruck (A); Prof. Dr. R. Holle, Neuherberg (D); Prof. Dr. S.R. Marder, Los Angeles (USA); Prof. Dr. P.D. McGorry, Melbourne (AUS); Prof. Dr. F. Müller-Spahn, Basel (CH); Prof. Dr. W. Rössler, Zürich (CH); and Prof. Dr. H. van den Bussche, Hamburg (D).

References

  • Alvarez-Jimenez et al., 2011 M. Alvarez-Jimenez, J.F. Gleeson, L.P. Henry, S.M. Harrigan, M.G. Harris, G.P. Amminger, E. Killackey, A.R. Yung, H. Herrman, H.J. Jackson, P.D. McGorry. Prediction of a single psychotic episode: a 7.5-year, prospective study in first-episode psychosis. Schizophr. Res.. 2011;125:236-246 Crossref
  • Alvarez-Jimenez et al., 2012 M. Alvarez-Jimenez, A. Priede, S.E. Hetrick, S. Bendall, E. Killackey, A.G. Parker, P.D. McGorry, J.F. Gleeson. Risk factors for relapse following treatment for first episode psychosis: a systematic review and meta-analysis of longitudinal studies. Schizophr. Res.. 2012;139:116-128 Crossref
  • American Psychiatric Association (APA), 2000 American Psychiatric Association (APA). Social and Occupational Functioning Assessment Scale (SOFAS). Diagnostic and Statistical Manual of Mental Disorders fourth edition (American Psychiatric Association, Washington, DC, 2000) 817-818 (text revision)
  • Andreasen et al., 2005 N.C. Andreasen, W.T. Carpenter, J.M. Kane, R.A. Lasser, S.R. Marder, D.R. Weinberger. Remission in schizophrenia: proposed criteria and rationale for consensus. Am. J. Psychiatry. 2005;62:441-449 Crossref
  • Ascher-Svanum et al., 2010 H. Ascher-Svanum, B. Zhu, D.E. Faries, D. Salkever, E.P. Slade, X. Peng, R.R. Conley. The cost of relapse and the predictors of relapse in the treatment of schizophrenia. BMC Psychiatry. 2010;10(2) ( www.biomedcentral.com/1471-244X/10/2 )
  • Birchwood et al., 1989 M. Birchwood, J. Smith, F. MacMillan, B. Hogg, R. Prasad, C. Harvey, S. Bering. Predicting relapse in schizophrenia: the development and implementation of an early signs monitoring system using patients and families as observers, a preliminary investigation. Psychol. Med.. 1989;19:649-656 Crossref
  • Birchwood et al., 1998 M. Birchwood, P. Todd, C. Jackson. Early intervention in psychosis. The critical period hypothesis. Br. J. Psychiatry. 1998;172:53-59 (Suppl.)
  • Bustillo et al., 1995 J. Bustillo, R.W. Buchanan, W.T. Carpenter. Prodromal symptoms vs. early warning signs and clinical action in schizophrenia. Schizophr. Bull.. 1995;21:553-559 Crossref
  • Caseiro et al., 2012 O. Caseiro, R. Pérez-Iglesias, I. Mata, O. Martínez-Garcia, J.M. Pelayo-Terán, R. Tabares-Seisdedos, V. Ortiz-García de la Foz, J.L. Vázquez-Barquero, B. Crespo-Facorro. Predicting relapse after a first episode of non-affective psychosis: a three-year follow-up study. J. Psychiatr. Res.. 2012;46(8):1099-1105 Crossref
  • Cechnicki et al., 2011 A. Cechnicki, I. Hanuszkiewicz, R. Polczyk, A. Bielańska. Prognostic value of duration of untreated psychosis in long-term outcome of schizophrenia. Med. Sci. Monit.. 2011;17(5):CR277-CR283 Crossref
  • Chen et al., 2005 E.Y. Chen, C.L. Hui, E.L. Dunn, M.Y. Miao, W.S. Yeung, C.K. Wong, W.F. Chan, W.N. Tang. A prospective 3-year longitudinal study of cognitive predictors of relapse in first-episode schizophrenic patients. Schizophr. Res.. 2005;77(1):99-104 Crossref
  • Csernansky et al., 2002 J.G. Csernansky, R. Mahmoud, R. Brenner. A comparison of risperidone and haloperidol for the prevention of relapse in patients with schizophrenia. N. Engl. J. Med.. 2002;346(1):16-22 Crossref
  • Dawson et al., 2010 M.E. Dawson, A.M. Schell, A. Rissling, J. Ventura, K.L. Subotnik, K.H. Nuechterlein. Psychophysiological prodromal signs of schizophrenic relapse: a pilot study. Schizophr. Res.. 2010;123(1):64-67 Crossref
  • DiMichele et al., 2007 V. DiMichele, F. Bolino, M. Mazza, R. Roncone, M. Casacchia. Relapsing versus non relapsing course of schizophrenia: a cohort study in a community based mental health service. Epidemiol. Psichiatr. Soc.. 2007;16(1):50-58
  • Docherty et al., 2009 N.M. Docherty, A. St-Hilaire, J.M. Aakre, J.P. Seghers. Life events and high-trait reactivity together predict psychotic symptom increases in schizophrenia. Schizophr. Bull.. 2009;35(3):638-645 Crossref
  • Docherty et al., 2011 N.M. Docherty, A. St-Hilaire, J.M. Aakre, J.P. Seghers, A. McCleery, M. Divilbiss. Anxiety interacts with expressed emotion criticism in the prediction of psychotic symptom exacerbation. Schizophr. Bull.. 2011;37(3):611-618 Crossref
  • Doering et al., 1998 S. Doering, E. Müller, W. Köpcke, A. Pietzcker, W. Gaebel, M. Linden, P. Müller, F. Müller-Spahn, J. Tegeler, G. Schüssler. Predictors of relapse and rehospitalization in schizophrenia and schizoaffective disorder. Schizophr. Bull.. 1998;24(1):87-98 Crossref
  • Drake et al., 2007 R.J. Drake, G. Dunn, N. Tarrier, R.P. Bentall, G. Haddock, S.W. Lewis. Insight as a predictor of the outcome of first-episode nonaffective psychosis in a prospective cohort study in England. J. Clin. Psychiatry. 2007;68(1):81-86 Crossref
  • Emsley et al., 2012 R. Emsley, I. Nuamah, D. Hough, S. Gopal. Treatment response after relapse in a placebo-controlled maintenance trial in schizophrenia. Schizophr. Res.. 2012;138:29-34 Crossref
  • Falloon, 1984 I.R.H. Falloon. Relapse — a reappraisal of assessment of outcome in schizophrenia. Schizophr. Bull.. 1984;10(2):293-299 Crossref
  • Falloon et al., 1983 I.R.H. Falloon, G.N. Marshall, J.L. Boyd, J. Razani, C. Woodsiverio. Relapse in schizophrenia — a review of the concept and its definitions. Psychol. Med.. 1983;13(3):469-477 Crossref
  • Feldmann et al., 1995 R. Feldmann, G. Buchkremer, E. Minneker-Huegel, P. Hornung. Fragebogen zur Erfassung der emotionalen Familienatmosphäre (FEF). Diagnostica. 1995;41:334-348
  • Fitzgerald et al., 2009 P. Fitzgerald, A. de Castella, D. Arya, W.R. Simons, A. Eggleston, S. Meere, J. Kulkarni. The cost of relapse in schizophrenia and schizoaffective disorder. Australas. Psychiatry. 2009;17:265-272 Crossref
  • Fleischhacker et al., 1989 W. Fleischhacker, K. Bergmann, R. Perovich. The Hillside Akathisia Scale (HAS): a new rating instrument for neuroleptic-induced akathisia. Psychopharmacol. Bull.. 1989;25:222-226
  • Frances et al., 1994 A. Frances, H.A. Pincus, M.B. First. The Global Assessment of Functioning Scale (GAF). Diagnostic and Statistical Manual of Mental Disorders Fourth edition (American Psychiatric Association, Washington, DC, 1994) 32
  • Gaebel, 2004 W. Gaebel. Course typologies, treatment principles, and research concepts. Pharmacopsychiatry. 2004;37(Suppl. 2):S90-S97
  • Gaebel and Riesbeck, 2007 W. Gaebel, M. Riesbeck. Revisiting the relapse predictive validity of prodromal symptoms in schizophrenia. Schizophr. Res.. 2007;95:19-29 Crossref
  • Gaebel et al., 2007 W. Gaebel, M. Riesbeck, W. Wölwer, A. Klimke, M. Eickhoff, M. von Wilmsdorff, M.C. Jockers-Scherübl, K. Kühn, M. Lemke, A. Bechdolf, S. Bender, D. Degner, R. Schlösser, L.G. Schmidt, A. Schmitt, M. Jäger, G. Buchkremer, P. Falkai, S. Klingberg, W. Köpcke, W. Maier, H. Häfner, C. Ohmann, H.J. Salize, F. Schneider, H.J. Möller. Maintenance treatment with risperidone or low-dose haloperidol in first-episode schizophrenia. One-year results of a randomized controlled trial within the German Research Network on Schizophrenia. J. Clin. Psychiatry. 2007;68(11):1763-1774 Crossref
  • Gaebel et al., 2011 W. Gaebel, M. Riesbeck, W. Wölwer, A. Klimke, M. Eickhoff, M. von Wilmsdorff, M. Lemke, I. Heuser, W. Maier, W. Huff, A. Schmitt, H. Sauer, M. Riedel, S. Klingberg, W. Köpcke, C. Ohmann, H.J. Möller, German Study Group on First-Episode Schizophrenia. Relapse prevention in first-episode schizophrenia: maintenance vs. intermittent drug treatment with prodrome-based early intervention. Results of a randomized controlled trial within the German Research Network on Schizophrenia. J. Clin. Psychiatry. 2011;72:205-218 Crossref
  • Gaebel et al., 2013 W. Gaebel, M. Riesbeck, W. Wölwer, A. Klimke, M. Eickhoff, M. von Wilmsdorff, M. Lemke, I. Heuser, W. Maier, J. Klosterkötter, A. Schmitt, H. Sauer, M. Riedel, S. Klingberg, W. Köpcke, C. Ohmann, H.J. Möller, German Study Group on First-Episode Schizophrenia. Rates and predictors of remission in first-episode schizophrenia within 1 year of antipsychotic maintenance treatment. Results of a randomized controlled trial within the German Research Network on Schizophrenia. Schizophr. Res. 2013;10.1016/j.schres.2013.04.012 (May 1, pii: S0920-9964(13)00212-0, [Epub ahead of print])
  • Gleeson et al., 2005 J.F. Gleeson, D. Rawlings, H.J. Jackson, P.D. McGorry. Early warning signs of relapse following a first episode of psychosis. Schizophr. Res.. 2005;80:107-111 Crossref
  • Gleeson et al., 2010 J.F. Gleeson, M. Alvarez-Jimenez, S.M. Cotton, A.G. Parker, S. Hetrick. A systematic review of relapse measurement in randomized controlled trials of relapse prevention in first-episode psychosis. Schizophr. Res.. 2010;119:79-88 Crossref
  • Goldman et al., 1992 H.H. Goldman, A.E. Skodol, T.R. Lave. Revising Axis V for DSM-IV. A review of measures of social functioning. Am. J. Psychiatry. 1992;149:1148-1156
  • Guy, 1976a Abnormal Involuntary Movement Scale (AIMS). W. Guy (Ed.) ECDEU Assessment Manual for Psychopharmacology, Revised (US Dept Health, Education and Welfare, Washington, DC, 1976) 534-537
  • Guy, 1976b Clinical Global Impressions (CGI) Scale. W. Guy (Ed.) ECDEU Assessment Manual for Psychopharmacology, Revised (US Dept Health, Education, and Welfare, Washington, DC, 1976) 218-222
  • Hamilton, 1960 M. Hamilton. A rating scale for depression. J. Neurol. Neurosurg. Psychiatry. 1960;23:56-62 Crossref
  • Haro et al., 2008 J.M. Haro, D. Novick, D. Suarez, S. Ochoa, M. Roca. Predictors of the course of illness in outpatients with schizophrenia: a prospective three year study. Prog. Neuropsychopharmacol. Biol. Psychiatry. 2008;32(5):1287-1292 Crossref
  • Hirsch and Jolley, 1989 S.R. Hirsch, A.G. Jolley. The dysphoric syndrome in schizophrenia and its implications for relapse. Br. J. Psychiatry. 1989;155(Suppl. 5):46-50
  • Hogan et al., 1983 T.P. Hogan, A.G. Awad, R.A. Eastwood. Self-report predictive of drug compliance in schizophrenia: reliability and discriminative ability. Psychol. Med.. 1983;13:177-183 Crossref
  • Holthausen et al., 2007 E.A. Holthausen, D. Wiersma, W. Cahn, R.S. Kahn, P.M. Dingemans, A.H. Schene, R.J. van den Bosch. Predictive value of cognition for different domains of outcome in recent-onset schizophrenia. Psychiatry Res.. 2007;149(1–3):71-80 Crossref
  • Hong et al., 2009 J. Hong, F. Windmeijer, D. Novick, J.M. Haro, J. Brown. The cost of relapse in patients with schizophrenia in the European SOHO (Schizophrenia Outpatient Health Outcomes) study. Prog. Neuropsychopharmacol. Biol. Psychiatry. 2009;33:835-841 Crossref
  • Janke et al., 1985 W. Janke, G. Erdmann, W. Boucsein. Stressverarbeitungsfragebogen (SVF). (Hogrefe, Göttingen, 1985)
  • Jorgensen, 1998 P. Jorgensen. Early signs of psychotic relapse in schizophrenia. Br. J. Psychiatry. 1998;172:327-330 Crossref
  • Kane, 2007 J.M. Kane. Treatment strategies to prevent relapse and encourage remission. J. Clin. Psychiatry. 2007;68(Suppl. 14):27-30
  • Kay et al., 1986 S.R. Kay, L.A. Opler, A. Fiszbein. The Positive and Negative Syndrome Scale (PANSS) rating manual. Soc. Behav. Sci. Doc.. 1986;17:28-29
  • Kemp and David, 1996 R. Kemp, A. David. Psychological predictors of insight and compliance in psychotic patients. Br. J. Psychiatry. 1996;169:444-450 Crossref
  • Kishimoto et al., 2014 T. Kishimoto, A. Robenzadeh, C. Leucht, S. Leucht, K. Watanabe, M. Mimura, M. Borenstein, J.M. Kane, C.U. Correll. Long-acting injectable vs oral antipsychotics for relapse prevention in schizophrenia: a meta-analysis of randomized trials. Schizophr. Bull.. 2014;40:192-213 Crossref
  • Klingberg et al., in preparation S. Klingberg, S. Schneider, B. Conradt, A. Schaub, M. Wagner, W. Wölwer, A. Bechdolf, W. Gaebel, G. Buchkremer. Individual Psychological Intervention for Relapse Prevention in Post Acute, First Episode Schizophrenia — 2-Year Follow-up of a Multicentric Randomized Clinical Trial. (, 2013) (in preparation)
  • Leucht et al., 2009 S. Leucht, C. Corves, D. Arbter, R.R. Engel, C. Li, J.M. Davis. Second-generation versus first-generation antipsychotic drugs for schizophrenia: a meta-analysis. Lancet. 2009;373:31-41 Crossref
  • Leucht et al., 2011 C. Leucht, S. Heres, J.M. Kane, W. Kissling, J.M. Davis, S. Leucht. Oral versus depot antipsychotic drugs for schizophrenia—a critical systematic review and meta-analysis of randomised long-term trials. Schizophr. Res.. 2011;127:83-92 Crossref
  • Leucht et al., 2012 S. Leucht, M. Tardy, K. Komossa, S. Heres, W. Kissling, G. Salanti, J.M. Davis. Antipsychotic drugs versus placebo for relapse prevention in schizophrenia: a systematic review and meta-analysis. Lancet. 2012;379:2063-2071 Crossref
  • Levene et al., 2009 J.E. Levene, W. Lancee, M.V. Seeman, H. Skinner, S.J. Freeman. Family and patient predictors of symptomatic status in schizophrenia. Can. J. Psychiatry. 2009;54(7):446-451
  • Lieberman et al., 2001 J.A. Lieberman, D. Perkins, A. Belger, M. Chakos, F. Jarskog, K. Boteva, J. Gilmore. The early stages of schizophrenia: speculations on pathogenesis, pathophysiology, and therapeutic approaches. Biol. Psychiatry. 2001;50:884-897 Crossref
  • Lipkovich et al., 2007 I. Lipkovich, W. Deberdt, J.G. Csernansky, P. Buckley, J. Peuskens, S. Kollack-Walker, Y. Zhang, H. Liu-Seifert, J.P. Houston. Predictors of risk for relapse in patients with schizophrenia or schizoaffective disorder during olanzapine drug therapy. J. Psychiatr. Res.. 2007;41(3–4):305-310 Crossref
  • Maier-Diewald et al., 1983 W. Maier-Diewald, H.U. Wittchen, H. Hecht, K. Werner-Eilert. Die Münchner Ereignisliste (MEL) — Anwendungsmanual. (Max-Planck-Institut für Psychiatrie, München, 1983)
  • Malla and Norman, 1994 A.K. Malla, R.M.G. Norman. Prodromal symptoms in schizophrenia: a prospective investigation. Br. J. Psychiatry. 1994;164:487-493 Crossref
  • Marder et al., 1994 S.R. Marder, W.C. Wirshing, T. Van Putten, J. Mintz, J. McKenzie, K. Johnston-Cronk, M. Lebell, R.P. Liberman. Fluphenazine vs palcebo supplementation for prodromal signs of relapse in schizophrenia. Arch. Gen. Psychiatry. 1994;51:280-287 Crossref
  • Moher et al., 2001 D. Moher, K.F. Schulz, D. Altman, CONSORT Group (Consolidated Standards of Reporting Trials). The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomized trials. JAMA. 2001;285:1987-1991 Crossref
  • Möller et al., 2008 H.J. Möller, M. Riedel, M. Jäger, F. Wickelmaier, W. Maier, K.U. Kühn, G. Buchkremer, I. Heuser, J. Klosterkötter, M. Gastpar, D.F. Braus, R. Schlösser, F. Schneider, C. Ohmann, M. Riesbeck, W. Gaebel. Short-term treatment with risperidone or haloperidol in first-episode schizophrenia: 8-week results of a randomized controlled trial within the German Research Network on Schizophrenia. Int. J. Neuropsychopharmacol.. 2008;11:985-997
  • Nuechterlein and Dawson, 1984 K.H. Nuechterlein, M.E. Dawson. A heuristic vulnerability/stress model of schizophrenic episodes. Schizophr. Bull.. 1984;10:300-312 Crossref
  • Robinson et al., 1999 D. Robinson, M.G. Woerner, J.M. Alvir, R. Bilder, R. Goldman, S. Geisler, A. Koreen, B. Sheitman, M. Chakos, D. Mayerhoff, J.A. Lieberman. Predictors of relapse following response from a first episode of schizophrenia or schizoaffective disorder. Arch. Gen. Psychiatry. 1999;56:241-247 Crossref
  • Rund et al., 2007 B.R. Rund, I. Melle, S. Friis, J.O. Johannessen, T.K. Larsen, L.J. Midbøe, S. Opjordsmoen, E. Simonsen, P. Vaglum, T. McGlashan. The course of neurocognitive functioning in first-episode psychosis and its relation to premorbid adjustment, duration of untreated psychosis, and relapse. Schizophr. Res.. 2007;91(1–3):132-140 Crossref
  • Scandinavian Society of Psychopharmacology Committee of Clinical Investigations (UKU), 1987 Scandinavian Society of Psychopharmacology Committee of Clinical Investigations (UKU). The UKU Side Effect Rating Scale: scale for the registration of unwanted effects of psychotropics. Acta Psychiatr. Scand.. 1987;76(Suppl. 334):81-94
  • Schennach et al., 2012 R. Schennach, M. Obermeier, S. Meyer, M. Jäger, M. Schmauss, G. Laux, H. Pfeiffer, D. Naber, L.G. Schmidt, W. Gaebel, J. Klosterkötter, I. Heuser, W. Maier, M.R. Lemke, E. Rüther, S. Klingberg, M. Gastpar, F. Seemüller, H.J. Möller, M. Riedel. Predictors of relapse in the year after hospital discharge among patients with schizophrenia. Psychiatr. Serv.. 2012;63(1):87-90 Crossref
  • Schooler et al., 1979 N.R. Schooler, G.E. Hogarty, M.M. Weissman. Social Adjustment Scale II (SAS II). W.A. Hargreaves, C.C. Attkisson, J.E. Sorenson (Eds.) Resource. Materials for Community Mental Health Evaluators Publication No (ADM) 290–330 (Department of Health Education and Welfare, Washington, DC, US, 1979) 79-328
  • Schulz et al., 2010 K.F. Schulz, D.G. Altman, D. Moher, CONSORT Group. CONSORT 2010 statement: updated guidelines for reporting parallel group randomized trials. Ann. Intern. Med.. 2010;152(11):726-732 Crossref
  • Scottish Schizophrenia Research Group, 1992 Scottish Schizophrenia Research Group. The Scottish First Episode Schizophrenia Study VIII: five-year follow-up: clinical and psychosocial findings. Br. J. Psychiatry. 1992;161:496-500
  • Simpson and Angus, 1970 G.M. Simpson, J.W. Angus. A rating scale for extrapyramidal side effects. Acta Psychiatr. Scand. Suppl.. 1970;212:11-19 Crossref
  • Strauss and Carpenter, 1978 J.S. Strauss, W. Carpenter. The prognosis of schizophrenia: rationale for a multidimensional concept. Schizophr. Bull.. 1978;4:56-67 Crossref
  • Subotnik and Nuechterlein, 1988 K.L. Subotnik, K.H. Nuechterlein. Prodromal signs and symptoms of schizophrenic relapse. J. Abnorm. Psychol.. 1988;97:405-412 Crossref
  • Subotnik et al., 2011 K.L. Subotnik, K.H. Nuechterlein, J. Ventura, M.J. Gitlin, S. Marder, J. Mintz, G.S. Hellemann, L.A. Thornton, I.R. Singh. Risperidone nonadherence and return of positive symptoms in the early course of schizophrenia. Am. J. Psychiatry. 2011;168(3):286-292 Crossref
  • Tarrier et al., 1991 N. Tarrier, C. Barrowclough, J.S. Bamrah. Prodromal signs of relapse in schizophrenia. Soc. Psychol. Psychiatr. Epidemiol.. 1991;26:157-161
  • Uçok et al., 2006 A. Uçok, A. Polat, S. Cakir, A. Genç. One year outcome in first episode schizophrenia. Predictors of relapse. Eur. Arch. Psychiatry Clin. Neurosci.. 2006;256(1):37-43
  • van Haren et al., 2003 N.E. van Haren, W. Cahn, H.E. Hulshoff Pol, H.G. Schnack, E. Caspers, A. Lemstra, M.M. Sitskoorn, D. Wiersma, R.J. van den Bosch, P.M. Dingemans, A.H. Schene, R.S. Kahn. Brain volumes as predictor of outcome in recent-onset schizophrenia: a multi-center MRI study. Schizophr. Res.. 2003;64(1):41-52 Crossref
  • Watt et al., 1983 D.C. Watt, K. Katz, M. Shepherd. The natural history of schizophrenia: a 5-year prospective follow-up of a representative sample of schizophrenics by means of standardized clinical and social assessment. Psychol. Med.. 1983;13:663-670 Crossref
  • Wiersma et al., 1998 D. Wiersma, F.J. Nienhuis, C. Sloof, R. Giel. Natural course of schizophrenic disorders: a 15 year follow-up of a Dutch incidence cohort. Schizophr. Bull.. 1998;24:78-85
  • Wölwer et al., 2003 W. Wölwer, G. Buchkremer, H. Häfner, J. Klosterkötter, W. Maier, H.J. Möller, W. Gaebel. German research network on schizophrenia — bridging the gap between research and care. Eur. Arch. Psychiatry Clin. Neurosci.. 2003;253:321-329
  • Wölwer et al., 2008 W. Wölwer, J. Brinkmeyer, M. Riesbeck, L. Freimüller, A. Klimke, M. Wagner, H.J. Möller, S. Klingberg, W. Gaebel, German Study Group on First Episode Schizophrenia. Neuropsychological impairments predict the clinical course in schizophrenia. Eur. Arch. Psychiatry Clin. Neurosci.. 2008;258(Suppl. 5):28-34
  • Zubin and Spring, 1977 J. Zubin, B. Spring. Vulnerability: a new view of schizophrenia. J. Abnorm. Psychol.. 1977;86:103-126 Crossref

Footnotes

Department of Psychiatry and Psychotherapy, Medical Faculty, Heinrich Heine University, Duesseldorf, Germany

lowast Corresponding author at: Department of Psychiatry and Psychotherapy Medical Faculty, Heinrich Heine University, LVR-Klinikum Düsseldorf, Bergische Landstrasse 2, 40629 Düsseldorf, Germany. Tel.: + 49 211 922 2000; fax: + 49 211 922 2020.