- Email: [email protected]
Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial
Schizophrenia Research xxx (xxxx) xxx
Contents lists available at ScienceDirect
Schizophrenia Research journal homepage: www.elsevier.com/locate/schres
Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial Yuta Saito a, b, 1, Hitoshi Sakurai a, c, John M. Kane d, e, f, Nina R. Schooler d, g, Takefumi Suzuki h, Masaru Mimura a, Hiroyuki Uchida a, i, * a
Department of Neuropsychiatry, Keio University School of Medicine, Tokyo, Japan Department of Neuropsychiatry, Kawasaki Municipal Kawasaki Hospital, Kanagawa, Japan Department of Psychiatry, Massachusetts General Hospital, Boston, MA, USA d Department of Psychiatry, The Zucker Hillside Hospital, Northwell Health, New York, USA e Department of Psychiatry and Molecular Medicine, The Donald and Barbara Zucker School of Medicine at Hofstra/Northwell, Hempstead, N.Y, USA f Center of Psychiatric Neuroscience, The Feinstein Institute for Medical Research, New York, USA g Department of Psychiatry and Behavioral Sciences, SUNY Downstate Medical Center, New York, USA h Department of Neuropsychiatry and Clinical Ethics, University of Yamanashi, Yamanashi, Japan i Geriatric Mental Health Program, Centre for Addiction and Mental Health, Toronto, Canada b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 11 January 2019 Received in revised form 4 October 2019 Accepted 12 October 2019 Available online xxx
Introduction: Little attention has been paid to the contribution of individual residual symptom to predict relapse in patients with schizophrenia receiving oral or long-acting injectable (LAI) antipsychotics. Method: We used the data from the Preventing Relapse on Oral Antipsychotics Compared to Injectables Evaluating Efficacy (PROACTIVE) study, in which 305 outpatients with schizophrenia were randomly allocated to either biweekly LAI-risperidone (LAI-R) or daily oral second-generation antipsychotics (SGA) and assessed for up to 30 months. Baseline individual symptoms that could predict subsequent relapse were identified, using a Cox proportional hazards model. Moreover, among those who relapsed during the study (n ¼ 73), individual symptoms were compared between baseline and biweekly ratings 8 to 2 weeks before relapse, using the linear mixed model. Results: A greater score in grandiosity at baseline was significantly associated with subsequent relapse (adjusted HR ¼ 1.24, p ¼ 0.006). When the two treatment groups were separately analyzed, more severe grandiosity (adjusted HR ¼ 1.43, p ¼ 0.003) and less severe hallucinatory behavior (adjusted HR ¼ 0.70, p ¼ 0.013) at baseline were significantly associated with relapse in the oral SGA group, but none was identified in the LAI-R group. Emotional withdrawal was significantly worse 8 and 2 weeks before relapse compared to the baseline (p ¼ 0.032 and p ¼ 0.043, respectively). Discussion: More severe grandiosity and less hallucination may have led to more frequent relapses in patients with schizophrenia receiving oral antipsychotics, which was not a case in those receiving LAI-R. The exploratory analysis indicates an increase in emotional withdrawal before relapse may be a useful marker for earlier interventions to possibly avert relapse. © 2019 Elsevier B.V. All rights reserved.
Keywords: Antipsychotics Early warning signs Relapse Residual symptoms Schizophrenia
1. Introduction Schizophrenia is a chronic, frequently recurrent illness as a high proportion of patients experience relapse after improvement, remission or even recovery (Goff et al., 2017). Relapse results in serious consequences, including deterioration in psychosocial and
* Corresponding author. Department of Neuropsychiatry, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo, 160-8582, Japan. E-mail address: [email protected] (H. Uchida). 1 Both authors equally contributed to this work.
vocational functioning (Pillai et al., 2018), as well as social disruption (Tandon et al., 2010), and puts substantial burden on the patients, families and caregivers (Genduso and Haley, 1997; Kane, 2007). Additionally, relapse increases healthcare costs and potential work force loss (Genduso and Haley, 1997). Therefore, identifying those at high risk for relapse and detecting signs of relapse at the earliest opportunity is of critical importance for the management of this chronic illness. Prediction of relapse based on clinical characteristics, including symptomatology, is clinically relevant in providing individuallytailored medicine (Gaebel and Riesbeck, 2014). There have been
https://doi.org/10.1016/j.schres.2019.10.037 0920-9964/© 2019 Elsevier B.V. All rights reserved.
Please cite this article as: Saito, Y et al., Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.10.037
2
Y. Saito et al. / Schizophrenia Research xxx (xxxx) xxx
some reports suggesting residual positive and negative symptoms as predictors for relapse (Gaebel and Riesbeck, 2014; Schennach et al., 2012). However, these reports focused solely on total scores or symptom domains in the rating scales and did not evaluate the relevance of each individual symptom. In light of the contribution of each residual symptom to subsequent relapse, remission, as well as response reported in other psychiatric illnesses such as major depressive disorder (Madhoo et al., 2015; Sakurai et al., 2013, 2017), further investigations on this topic is warranted in schizophrenia. On the other hand, a majority of patients with schizophrenia experience cognitive, perceptual, and affective change a few weeks before relapse (Birchwood and Spencer, 2001). These early signs which are thought to be inherent symptoms of psychosis or a part of a relapsing process (Birchwood and Spencer, 2001) include nonspecific symptoms such as anxiety, dysphoria, insomnia, poor concentration as well as attenuated psychotic symptoms (Eisner et al., 2013; Gaebel and Riesbeck, 2014). Similarly, several objectively assessed symptoms, including depression and unusual thought content, have been reported to predict subsequent relapse (Marder et al., 1994; Norman and Malla, 1995; Subotnik and Nuechterlein, 1988). However, most of the studies were published in the 1990s and their sample sizes were as small as 50 to 80, which underscores the need of examining these early signs by using a more recent, larger dataset. The Preventing Relapse on Oral Antipsychotics Compared to Injectables - Evaluating Efficacy (PROACTIVE) study provides an ideal dataset derived from generalizable, large-scale samples to address these important issues. Utilizing this dataset, we examined the relevance of individual residual symptoms to the prediction of subsequent relapse. We also tried to evaluate the timing and degree of individual objective symptom worsening before actual relapse. 2. Materials and methods 2.1. Study design The PROACTIVE study was initiated in 2006 and conducted at eight academic centers in the United States to compare the effectiveness of daily oral second-generation antipsychotics (SGAs) and biweekly long-acting injectable risperidone (LAI-R). It was intended to inform clinical decision making by conducting an up-to-date relapse prevention study that included both efficacy/explanatory and effectiveness/pragmatic features; details have been described elsewhere (Buckley et al., 2015). Briefly, 305 outpatients with schizophrenia who were randomized either to oral SGA group or LAI-R group were followed for up to 30 months. Following a complete description of the study, participants provided written informed consent. The original study had been approved by the relevant regulatory bodies, and because of the completely anonymous nature of this analysis and absence of direct human involvement, no ethical approval was sought for the present analysis. 2.2. Study population Inclusion criteria were as follows: (a) outpatients who met the criteria for schizophrenia or schizoaffective disorder according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSM-IV-TR), (b) symptom exacerbation within 12 months and community dwelling for at least 4 weeks, (c) a score of ¼<4 on the Clinical Global Impressions - Severity of Illness (CGI-S), (d) age between 18 and 65 years, and (e) being capable of providing informed consent. Exclusion criteria were: (a) a current episode as their first episode of psychosis, (b) allergy to risperidone, (c) inadequate prior response to risperidone, (d) treatment-refractoriness and/or lack of response to clozapine, and
(e) pregnant or lactating women. 2.3. Treatment Enrolled participants were randomized to an oral SGA or LAI-R. In the oral SGA group, participants continued the ongoing oral SGA or started to take one of the following oral SGAs: aripiprazole, olanzapine, quetiapine, risperidone, paliperidone, asenapine, and iloperidone. In the LAI-R group, LAI-R was initiated with a 25-mg injection. The previous oral antipsychotics or LAIs were tapered in a cross-titration manner. Oral antipsychotics could be titrated or changed based on lack of efficacy, side effects, or patient preference every two weeks. LAI-R was continuously injected biweekly in the range of 12.5 mg to 50 mg and could be supplemented by oral antipsychotic. 2.4. Assessment measures The following assessments were performed by a rater who was blinded to the randomization status via secure, live, two-way video connection: the Brief Psychiatric Rating Scale (BPRS); CGI-S at baseline and quarterly visits; and Clinical Global Impression Global Improvement (CGI-GI) at quarterly visits. In addition, the following assessments were rated biweekly by unblinded raters at the sites: CGI-S; CGI-GI; and an abbreviated version of the BPRS that was composed of eight items (grandiosity, suspiciousness, unusual thought content, hallucinatory behavior, conceptual disorganization, mannerisms and posturing, blunted affect, and emotional withdrawal). Relapse was defined with a CGI-GI score of 6 (much worse) or 7 (very much worse) for the present analysis. It was defined with (1) a CGI-GI score of 6 or 7, (2) psychotic hospitalization except for social one, (3) increase in level of psychotic care, (4) continuous increase in psychotic symptoms that was judged by raters, or (5) self-injury, suicidal or homicidal ideation, and violent behavior that was clinically significant in the original study (Buckley et al., 2015). However, since some of the information on (2)e(5) could not be located for some subjects in the dataset, we solely focused on the CGI. 2.5. Statistical analysis Baseline sociodemographic and clinical characteristics were compared between the oral SGA group and the LAI-R group. Spearman correlation coefficients were calculated between the individual scores on the BPRS at baseline in order to examine multicollinearity. To identify which individual symptoms in the BPRS at baseline could predict subsequent relapse, a Cox hazards model, which is commonly used for investigating the association between survival time and predictor variables, was employed. We ran one Cox proportional hazard model with several individual symptoms included as explanatory variables which was stratified on age, sex, illness duration, and treatment; the Bonferroni correction was not necessary in this analysis. To evaluate the timing and degree of individual symptom worsening before relapse, the individual symptom scores between baseline and 8, 6, 4, and 2 weeks before relapse were compared among those who relapsed (relapsers), using the linear mixed model that contained week as a factor with autoregressive AR (1) correlation matrix among time points. The Bonferroni correction was used to compare the main effects. Those analyses conducted for the whole sample were also repeated separately for the oral SGA group, the LAI-R group, patients with schizophrenia, and those with schizoaffective disorder. To deal with missing values, per-protocol (PP) analysis was employed. A p-value of <0.05 was considered statistically significant. All of the data were analyzed using the Statistical Package for
Please cite this article as: Saito, Y et al., Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.10.037
Y. Saito et al. / Schizophrenia Research xxx (xxxx) xxx
3
Table 1 Sociodemographic and clinical characteristics of subjects. Characteristics Sex Male, n (%) Race Caucasian, n (%) African American, n (%) Hispanic, n (%) Others, n (%) Age, years, mean ± SD Diagnoses Schizophrenia, Paranoid Type, n (%) Schizophrenia Disorganized Type, n (%) Schizophrenia, Residual Type, n (%) Schizophrenia, Undifferentiated Type, n (%) Schizophrenia, NOS, n (%) Schizoaffective Disorder, n (%) Duration of illness, years, mean ± SD Time since last psychiatric hospitalization, months, mean ± SD
All (N ¼ 305)
LAI (n ¼ 153)
Oral SGA (n ¼ 152)
218 (71%)
108 (71%)
110 (72%)
156 (51%) 85 (28%) 58 (19%) 6 (2%) 38.3 ± 12.1
78 (51%) 45 (29%) 27 (18%) 3 (2%) 38.1 ± 11.8
78 (51%) 40 (26%) 31 (20%) 3 (2%) 38.3 ± 12.3
123 (40%) 23 (8%) 1 (0.3%) 54 (18%) 5 (2%) 99 (32%) 17.4 ± 11.6 31.4 ± 64.2
66 (43%) 11 (7%) 0 (0%) 26 (17%) 4 (3%) 46 (30%) 16.7 ± 11.3 22.4 ± 35.9
57 (38%) 12 (8%) 1 (1%) 28 (18%) 1 (1%) 53 (35%) 18.0 ± 11.9 40.4 ± 82.4
Abbreviations: LAI, long-acting injectable; SGA, second-generation antipsychotic; SD, standard deviation.
Social Science (SPSS) version 24 for Windows (IBM Corporation, Armonk, NY).
Table 2 BPRS scores at baseline.
3. Results 3.1. Subject characteristics A total of 305 participants were randomly assigned to the oral SGA group (n ¼ 152) and the LAI-R group (n ¼ 153). The mean ± SD score in the total BPRS score was 38.0 ± 9.4 and the mean score in the CGI-S was 4.0 ± 0.8 at baseline, without significant group differences. Of these participants, an overall 23.9% of patients (n ¼ 73) (26.8% in the LAI-R group and 21.1% in the SGA group) experienced at least one relapse, and the first relapse occurred at a mean ± SD of 42.1 ± 31.9 weeks (41.9 ± 30.5 weeks in the LAI-R group and 42.4 ± 34.6 weeks in the oral SGA group) after the enrollment to the study. There were 129 patients (42.3%) who completed the study without relapse (42.5% in LAI-R group and 42.1% in SGA group; n.s.); 33.8% (n ¼ 103) (30.7% in LAI-R group and 36.8% in SGA group; n.s.) dropped out for reasons other than relapse during the follow-up period. Table 1 summarizes sociodemographic and clinical characteristics of the study sample.
Somatic Concern Anxiety Depressive Mood Guilt Feelings Hostility Suspiciousness Unusual Thought Content Grandiosity Hallucinatory behavior Disorientation Conceptual Disorganization Excitement Motor Retardation Blunted Affect Tension Mannerisms and Posturing Uncooperativeness Emotional Withdrawal Total Score
All (N ¼ 305)
LAI (n ¼ 153)
Oral SGA (n ¼ 152)
2.32 ± 1.25 3.64 ± 1.35 2.90 ± 1.44 2.23 ± 1.30 2.32 ± 1.19 3.48 ± 1.94 3.19 ± 1.73 1.77 ± 1.36 3.14 ± 1.87 1.23 ± 0.51 1.68 ± 1.18 1.17 ± 0.61 1.35 ± 0.76 2.10 ± 1.15 1.75 ± 0.96 1.33 ± 0.79 1.12 ± 0.46 1.32 ± 0.74 38.0 ± 9.36
2.31 ± 1.30 3.75 ± 1.37 2.90 ± 1.50 2.19 ± 1.36 2.25 ± 1.16 3.46 ± 1.96 3.14 ± 1.81 1.67 ± 1.32 3.31 ± 1.94 1.25 ± 0.54 1.66 ± 1.18 1.14 ± 0.54 1.33 ± 0.69 2.13 ± 1.19 1.79 ± 0.99 1.30 ± 0.77 1.14 ± 0.47 1.38 ± 0.76 38.0 ± 9.60
2.34 ± 1.20 3.53 ± 1.33 2.91 ± 1.37 2.27 ± 1.23 2.39 ± 1.23 3.50 ± 1.92 3.25 ± 1.65 1.87 ± 1.40 2.97 ± 1.79 1.21 ± 0.48 1.70 ± 1.19 1.19 ± 0.67 1.36 ± 0.83 2.07 ± 1.12 1.71 ± 0.93 1.36 ± 0.80 1.11 ± 0.46 1.27 ± 0.72 38.0 ± 9.14
Values are shown as mean ± SD. Abbreviations: BPRS, Brief Psychiatric Rating Scale; LAI, long-acting injectable; SGA, second-generation antipsychotic.
3.2. Prediction of relapse by individual symptom severities The mean values of each individual symptom in the BPRS at baseline are shown in Table 2. In all participants, Spearman correlation coefficients among these symptom scores were less than 0.5 except for anxiety and depressive mood, depressive mood and guilt feeling, suspiciousness and unusual thought content, as well as unusual thought content and hallucinatory behavior. In the Cox proportional hazards model, from which depressive mood and unusual thought content were removed in the light of multicollinearity (c2(20) ¼ 24.87), only grandiosity in the BPRS at baseline was significantly associated with subsequent relapse (adjusted HR ¼ 1.24, 95% CI ¼ 1.06e1.45, p ¼ 0.006) (Table 3). Similarly, in the Cox proportional hazards model with all individual symptoms (c2(22) ¼ 26.36), only grandiosity was also significantly associated with subsequent relapse (adjusted HR ¼ 1.26, 95% CI ¼ 1.07e1.50, p ¼ 0.007) (Supplementary Table 1). In the LAI-R group, Spearman correlation coefficients between individual symptom scores at baseline were less than 0.5 except for anxiety and depressive mood, anxiety and guilt feelings, anxiety and hallucinatory behavior, depressive mood and guilt feelings, suspiciousness and unusual thought content, suspiciousness and
Table 3 Associations between individual symptom scores in the BPRS and relapse in all participants. Adjusted Hazard Ratio (95% CI)a Somatic Concern Anxiety Guilt Feelings Hostility Suspiciousness Grandiosity Hallucinatory behavior Disorientation Conceptual Disorganization Excitement Motor Retardation Blunted Affect Tension Mannerisms and Posturing Uncooperativeness Emotional Withdrawal
0.98 1.01 0.98 1.06 1.14 1.24 0.93 1.09 0.99 0.98 0.95 1.15 0.92 0.93 1.01 1.00
(0.79e1.23) (0.79e1.29) (0.78e1.22) (0.83e1.35) (0.98e1.33) (1.06e1.45)b (0.79e1.09) (0.69e1.70) (0.77e1.27) (0.65e1.50) (0.66e1.36) (0.89e1.49) (0.67e1.27) (0.66e1.32) (0.63e1.61) (0.69e1.44)
Abbreviation: BPRS, Brief Psychiatric Rating Scale. a Adjustedafor age, gender, illness duration, treatment, and other individual symptoms. b p < 0.05.
Please cite this article as: Saito, Y et al., Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.10.037
4
Y. Saito et al. / Schizophrenia Research xxx (xxxx) xxx
hallucinatory behavior, unusual thought content and hallucinatory behavior, as well as motor retardation and blunted affect. According to the Cox proportional hazards model, from which anxiety, depressive mood, suspiciousness, and unusual thought content were removed in the light of multicollinearity (c2(17) ¼ 13.78), there were no baseline symptoms that were significantly associated with relapse (Table 4). Also in the Cox proportional hazards model with all individual symptoms (c2(21) ¼ 16.57), there were no baseline symptoms that were significantly associated with relapse, either (Supplementary Table 2). In the oral SGA group, Spearmen correlation coefficients were less than 0.5 except for anxiety and depressive mood, suspiciousness and unusual thought content, as well as unusual thought content and hallucinatory behavior. In contrast to the LAI-R group, according to the Cox proportional hazards model, from which anxiety and unusual thought content were removed in the light of multicollinearity (c2(19) ¼ 19.55), the following two symptoms in the BPRS at baseline were significantly associated with subsequent relapse in the oral SGA group; grandiosity (adjusted HR ¼ 1.43, 95% CI ¼ 1.13e1.80, p ¼ 0.003) and less hallucinatory behavior (adjusted HR ¼ 0.70, 95% CI ¼ 0.53e0.93, p ¼ 0.013) (Table 5). Similarly, according to the Cox proportional hazards model with all individual symptoms (c2(21) ¼ 19.71), greater grandiosity (adjusted HR ¼ 1.42, 95% CI ¼ 1.09e1.86, p ¼ 0.01) and depressive mood (adjusted HR ¼ 1.55, 95% CI ¼ 1.00e2.41, p ¼ 0.049), and less hallucinatory behavior (adjusted HR ¼ 0.70, 95% CI ¼ 0.52e0.95, p ¼ 0.02) were significantly associated with subsequent relapse (Supplementary Table 3). In the schizophrenia group, Spearmen correlation coefficients were less than 0.5 except for anxiety and depressive mood, anxiety and suspiciousness, suspiciousness and unusual thought content, as well as unusual thought content and hallucinatory behavior. According to the Cox proportional hazards model from which anxiety and unusual thought content were removed in the light of multicollinearity (c2(20) ¼ 42.47), the following four symptoms in the BPRS at baseline were significantly associated with subsequent relapse; blunted affect (adjusted HR ¼ 1.47, 95% CI ¼ 1.05e2.06, p ¼ 0.026), hostility (adjusted HR ¼ 1.46, 95% CI ¼ 1.06e2.00, p ¼ 0.019), suspiciousness (adjusted HR ¼ 1.23, 95% CI ¼ 1.01e1.50, p ¼ 0.042), and less hallucinatory behavior (adjusted HR ¼ 0.80, 95% CI ¼ 0.64e0.99, p ¼ 0.041) (Supplementary Table 4). Similarly, according to the Cox proportional hazards model with all individual symptoms (c2(22) ¼ 42.33), blunted affect (adjusted HR ¼ 1.45, 95% CI ¼ 1.04e2.03, p ¼ 0.029), and hostility (adjusted HR ¼ 1.49, 95% CI ¼ 1.07e2.08, p ¼ 0.018) were significantly associated with
Table 4 Associations between individual symptom scores in BPRS and relapse in LAI-R group. Adjusted Hazard Ratio (95% CI)a Somatic Cancern Guilt Feelings Hostility Grandiosity Hallucinatory behavior Disorientation Conceptual Disorganization Excitement Motor Retardation Blunted Affect Tension Mannerisms and Posturing Uncooperativeness Emotional Withdrawal
0.91 0.95 1.03 1.16 1.12 1.28 1.05 0.97 1.05 1.14 0.92 0.99 1.10 1.02
(0.65e1.26) (0.70e1.28) (0.72e1.47) (0.93e1.44) (0.92e1.36) (0.74e2.22) (0.74e1.50) (0.48e1.96) (0.58e1.90) (0.79e1.65) (0.57e1.48) (0.60e1.64) (0.57e2.10) (0.62e1.69)
Abbreviation: BPRS, Brief Psychiatric Rating Scale; LAI-R, long-acting injectable risperidone. a Adjusted for age, gender, illness duration, and other individual symptoms..
Table 5 Associations between individual symptom scores in BPRS and relapse in oral SGA group. Adjusted Hazard Ratio (95% CI)a a
Somatic C ncern Depressive Mood Guilt Feelings Hostility Suspiciousness Grandiosity Hallucinatory behavior Disorientation Conceptual Disorganization Excitement Motor Retardation Blunted Affect Tension Mannerisms and Posturing Uncooperativeness Emotional Withdrawal
1.02 1.42 0.97 1.04 1.27 1.43 0.70 0.79 1.01 1.01 0.87 1.08 0.90 0.70 1.09 1.04
(0.72e1.43) (0.98e2.06) (0.67e1.39) (0.71e1.53) (0.98e1.63) (1.13e1.80)b (0.53e0.93)b (0.33e1.89) (0.68e1.50) (0.57e1.79) (0.52e1.45) (0.71e1.65) (0.54e1.51) (0.40e1.20) (0.47e2.52) (0.54e2.04)
Abbreviation: BPRS, Brief Psychiatric Rating Scale; SGA, second-generation antipsychotic. a Adjusted for age, gender, illness duration, and other individual symptoms. b . p < 0.05.
subsequent relapse (Supplementary Table 5). In the schizoaffective disorder group, Spearmen correlation coefficients were less than 0.5 except for anxiety and depressive mood, depressive mood and hallucinatory behavior, suspiciousness and unusual thought content, and conceptual disorganization and excitement. According to the Cox proportional hazards model, from which depressive mood, unusual thought content, and conceptual disorganization were removed in the light of multicollinearity (c2(19) ¼ 22.97), the following two symptoms in the BPRS at baseline were significantly associated with subsequent relapse; uncooperativeness (adjusted HR ¼ 3.11, 95% CI ¼ 1.22e7.98, p ¼ 0.018) and less somatic concern (adjusted HR ¼ 0.66, 95% CI ¼ 0.44e0.98, p ¼ 0.041) (Supplementary Table 6). According to the Cox proportional hazards model in this group with all individual symptoms (c2(22) ¼ 32.12), uncooperativeness (adjusted HR ¼ 3.61, 95% CI ¼ 1.37e9.53, p ¼ 0.01), motor retardation (adjusted HR ¼ 3.46, 95% CI ¼ 1.47e8.13, p ¼ 0.004), grandiosity (adjusted HR ¼ 2.19, 95% CI ¼ 1.31e3.69, p ¼ 0.003), suspiciousness (adjusted HR ¼ 1.53, 95% CI ¼ 1.00e2.32, p ¼ 0.048), less disorientation (adjusted HR ¼ 0.06, 95% CI ¼ 0.00e0.85, p ¼ 0.04), less hostility (adjusted HR ¼ 0.50, 95% CI ¼ 0.28e0.87, p ¼ 0.01), less blunted affect (adjusted HR ¼ 0.52, 95% CI ¼ 0.28e0.98, p ¼ 0.04), less unusual thought content (adjusted HR ¼ 0.53, 95% CI ¼ 0.32e0.87, p ¼ 0.01), and less somatic concern (adjusted HR ¼ 0.55, 95% CI ¼ 0.35e0.86, p ¼ 0.01) were significantly associated with subsequent relapse (Supplementary Table 7).
3.3. Comparison of individual symptom severities between baseline and pre-relapse among relapsers Scores of unusual thought content, hallucinatory behavior, conceptual disorganization, and emotional withdrawal in the abbreviated BPRS at the time of relapse were significantly higher than those at baseline (Fig. 1). Emotional withdrawal scores were significantly higher 8 and 2 weeks before relapse compared to the baseline value (p ¼ 0.032 and p ¼ 0.043, respectively). No symptom scores before relapse were significantly higher compared to the baseline values in the LAI-R group and the oral SGA group, respectively. While there were no symptoms whose scores before relapse were significantly higher than their baseline values in the schizophrenia group, an emotional withdrawal score was significantly higher 8 weeks before relapse compared to the baseline value in the schizoaffective disorder group (p ¼ 0.018).
Please cite this article as: Saito, Y et al., Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.10.037
Y. Saito et al. / Schizophrenia Research xxx (xxxx) xxx
5
Fig. 1. Symptom scores in the abbreviated BPRS before relapse.
4. Discussion The present exploratory data analysis indicated that more severe grandiosity and less severe hallucinatory behavior at baseline
were associated with subsequent relapse in patients treated with oral SGAs, but no symptoms emerged as predictors in those receiving LAI-R. Additionally, we also found that the score of emotional withdrawal had significantly increased already 8 weeks
Please cite this article as: Saito, Y et al., Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.10.037
6
Y. Saito et al. / Schizophrenia Research xxx (xxxx) xxx
before relapse compared to the baseline score. While it is unclear whether this symptom worsening is a predictor of relapse or a part of a relapsing process, this timeline may offer a window for possible interventions in an effort to avert relapses. The advantage of this study is that not only psychotic symptoms but also other BPRS symptoms were analyzed. These findings underscore the relevance of individual symptoms as well as overall illness severity in predicting relapse in the maintenance treatment of schizophrenia. Among the residual BPRS symptoms, grandiosity was the only symptom that was associated with subsequent relapse in the whole sample. Grandiosity may well have a negative effect on patients’ social function. For example, Rocca et al. examined the associations between positive, negative, and disorganization symptoms, and real-world functioning among 880 patients with schizophrenia, using the Specific Levels of Functioning Scale. In the analysis using structural equation modeling, only grandiosity was identified to be related to poor work skills (Rocca et al., 2018). While grandiosity may remain in a relatively small proportion of patients (Schennach et al., 2015), it was frequently observed in patients with chronic schizophrenia compared to those with first episode psychosis (Del Bello et al., 2016), thus indicating a possibility to serve as a sign of chronicity. Taken together, these findings suggest that residual grandiosity may negatively influence social function and selfmanagement skill, which in turn might lead to further worsening to the point where relapse results, among patients with schizophrenia. Interestingly, in the analysis of the oral SGA group, residual grandiosity and less severity in hallucinatory behavior at baseline emerged as factors associated with subsequent relapse, whereas no residual symptoms including grandiosity were identified in the LAIR group. This difference could be explained by differences in medication adherence between the two groups. In a cross-sectional study of 719 patients in emergency psychiatric settings, grandiosity in the BPRS was found to be associated with a lack of motivation for treatment, suggesting that grandiosity could lead to low adherence (Mulder et al., 2005). Expert consensus guidelines also refer to persistent grandiosity as potential causes of adherence problems in patients with schizophrenia (Velligan et al., 2009). On the other hand, the finding of the association between less hallucinatory behaviors at baseline and subsequent relapse may appear paradoxical. According to a survey by Moritz and colleagues, approximately one quarter of patients with schizophrenia reported at least one positive aspect of the illness as a reason for poor medication adherence (Moritz et al., 2014). Some patients with schizophrenia feel that psychosis makes their life meaningful, that psychosis provides them a feeling of importance and power, and that psychosis helps them to get the attention they deserved; these patients may be at greater risk for poor adherence once psychosis is under control (Moritz et al., 2009, 2013, 2014). However, this somewhat complicated explanation only has any real value if it is confirmed by prospective analysis of a new sample. Thus, these findings might indicate that a greater degree of grandiosity may render adherence to oral antipsychotics worse in this specific group of patients, resulting in subsequent relapse. In this context, measuring antipsychotic concentrations in blood may be useful to identify patients with poor adherence who have an increased risk of relapse (McCutcheon et al., 2018). Neither of subgroup analyses of patients with schizophrenia nor those with schizoaffective disorder did not identify grandiosity as a predictor of subsequent relapse. Moreover, a variety of different symptoms at baseline were found to be associated with subsequent relapse in these patient groups: hostility, suspiciousness, blunted affect, and less severity in hallucinatory behavior in patients with schizophrenia; uncooperativeness and less severity in somatic concern in patients with schizoaffective disorder. These differences may indicate differences in the nature of the illness between these
two disorders or possibilities of type-I and type-II errors. Therefore, the preliminary findings in the present analysis have to be confirmed in larger, prospective clinical trials that are specifically designed to address this issue. In the present study, the biweekly, clinical structured assessment identified emotional withdrawal as an early warning sign as early as eight weeks before relapse in the whole sample. The findings on early signs of relapse in schizophrenia have not always been consistent in the literature. In an early interview study, 145 patients with schizophrenia and 80 of their family members were retrospectively asked about their experiences of early signs of relapse (Herz and Melville, 1980). The symptoms that were frequently reported before relapse were dysphoric in nature: troubled sleep, troubled concentration, loss of appetite, and depressive mood. Previous reviewed articles examined prospective studies that performed weekly to monthly assessments of symptomatology in order to characterize early signs before relapse in patients with schizophrenia (Birchwood and Spencer, 2001; Eisner et al., 2013). They found that some psychotic symptoms (e.g. BPRS thought disturbance subscale) (Subotnik and Nuechterlein, 1988), non-psychotic symptoms (e.g. BPRS anxiety-depression subscale) (Marder et al., 1991), and their combination (e.g. combined Early Signs Scale and Positive and Negative Syndrome Scale general subscale) (Jørgensen, 1998) showed high sensitivity and specificity in predicting subsequent relapse. More recently, Wang et al. analyzed data from three long-term studies and examined symptom changes among 907 patients with schizophrenia who received assessment with the PANSS every 1e4 weeks. 267 patients who relapsed presented with delusions, suspiciousness, hallucinations, anxiety, tension, excitement, and conceptual disorganization as early warning signs (Wang et al., 2018). Notably, this finding are inconsistent with ours probably due to the difference in patients’ characteristics and study design: the participants in the present study had longer duration of illness (17.4 vs 10e12 years [mean]), a greater number of past hospitalizations (10.9 vs 1e3 times [mean]), and worse baseline CGI-S (4.0 vs 2.7e3.5 [mean]), which needs to be taken into account when interpreting the results. Since early warning signs may vary depending on the patients’ background and treatment, further research is clearly needed to confirm our findings. There are several limitations to be noted in the present study. First, this is a reanalysis of the PROACTIVE study, which was not designed to evaluate the issues addressed herein. Second, the generalizability of our findings may be limited, since the participants were outpatients enrolled at eight US academic centers in the PROACTIVE study. Additionally, 71% of the participants were male. Finally, although the dropout rate of 33.8% during 30-month follow-up was not always high, a lower rate might have resulted in slightly different findings. In conclusion, residual grandiosity may serve as a risk factor of subsequent relapse in the long-term treatment of schizophrenia in the whole sample. More notably, severe grandiosity and less hallucination at baseline were associated with subsequent relapse in patients receiving oral SGAs, which was not observed in those treated with LAI-R. These symptoms may lower medication adherence, which in turn may lead to more frequent relapses. Though we did not observe more frequent relapses among patients receiving oral antipsychotics in the present study, we did see significantly less psychopathology in the LAI-R group. Additionally, detection of exacerbation in emotional withdrawal at the earliest opportunity may help to identify those who are likely to relapse in the near future and who could benefit from additional intervention/attention to potentially prevent such a negative event. Since these findings are based on post-hoc analyses they should be replicated in other populations. However, these results point to the value of thorough evaluation of individual symptoms as well as
Please cite this article as: Saito, Y et al., Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.10.037
Y. Saito et al. / Schizophrenia Research xxx (xxxx) xxx
overall symptomatology for relapse prevention among patients with schizophrenia.
7
Funding body agreements and policies No funding support for the present analysis of the data.
Declaration of competing interest Dr. Saito has received speaker’s fee from Dainippon Sumitomo. Dr. Sakurai has received manuscript or speaker’s honoraria from Dainippon Sumitomo, Eli Lilly, Meiji-Seika Pharma, Otsuka Pharmaceutical, Tanabe Mitsubishi Pharma, and Yoshitomi Yakuhin within the past three years. Dr, Kane has received honoraria or consulting fees from Alkermes, Genentech, Intracellular Therapeutics, Janssen, Lundbeck, Merck, Neurocrine, Neuron, Otsuka, Pierre Fabre, Reviva, Roche, Sunovion, Teva. He is a shareholder in LB Pharma and The Vangurad Research Group. Dr Schooler has received honoraria or consulting fees from Alkermes, Allergan, EnVivo/Forum, Lundbeck, Otsuka, Roche and Sunovion in the past three years. Dr. Suzuki has received manuscript or speaker’s fees from Astellas, Dainippon Sumitomo Pharma, Eli Lilly, Elsevier Japan, Janssen Pharmaceuticals, Meiji Seika Pharma, Novartis, Otsuka Pharmaceutical, Wiley Japan, and Yoshitomi Yakuhin, and research grants from Eisai, Mochida Pharmaceutical, and Meiji Seika Pharma. Dr. Mimura has received speaker’s honoraria from Daiichi Sankyo, Dainippon-Sumitomo Pharma, Eisai, Eli Lilly, Fuji Film RI Pharma, Janssen Pharmaceutical, Mochida Pharmaceutical, MSD, Nippon Chemiphar, Novartis Pharma, Ono Yakuhin, Otsuka Pharmaceutical, Pfizer, Takeda Yakuhin, Tsumura, and Yoshitomi Yakuhin within the past three years. Also, he received grants from Daiichi Sankyo, Eisai, Pfizer, Shionogi, Takeda, Tanabe Mitsubishi and Tsumura within the past three years. Dr. Uchida has received grants from Eisai, Otsuka Pharmaceutical, Dainippon-Sumitomo Pharma, Mochida Pharmaceutical, Meiji-Seika Pharmaceutical, and Novartis; speaker’s honoraria from Otsuka Pharmaceutical, Eli Lilly, Shionogi, Pfizer, Yoshitomi Yakuhin, Dainippon-Sumitomo Pharma, Meiji-Seika Pharma, MSD, and Janssen Pharmaceutical; and advisory panel payments from DainipponSumitomo Pharma within the past three years. Acknowledgments Data and/or research tools used in the preparation of this manuscript were obtained and analyzed from the controlled access datasets distributed from the NIMH-supported National Database for Clinical Trials (NDCT). NDCT is a collaborative informatics system created by the National Institute of Mental Health to provide a national resource to support and accelerate discovery related to clinical trial research in mental health. Dataset identifier: NIMH Data Repositories Study identification number #2154. This manuscript reflects the views of the authors and may not reflect the opinions or views of the NIMH or of the submitters submitting original data to NDCT. Data used in the preparation of this article were obtained from the limited access datasets from the Preventing Relapse Oral Antipsychotics Compared to Injectables - Evaluating Efficacy (PROACTIVE) study. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.schres.2019.10.037. Contributors HU designed the study and wrote the protocol. YS and HS managed the data analyses and drafted the manuscript. JMK, NRS, TS, MM, and HU revised the manuscript critically for important intellectual content and contributed substantially to data analysis. All authors have approved the final manuscript and are accountable for all aspects of the work.
References Birchwood, M., Spencer, E., 2001. Early intervention in psychotic relapse. Clin. Psychol. Rev. 21 (8), 1211e1226. Buckley, P.F., Schooler, N.R., Goff, D.C., Hsiao, J., Kopelowicz, A., Lauriello, J., Manschreck, T., Mendelowitz, A.J., Miller del, D., Severe, J.B., Wilson, D.R., Ames, D., Bustillo, J., Mintz, J., Kane, J.M., Study, P., 2015. Comparison of SGA oral medications and a long-acting injectable SGA: the PROACTIVE study. Schizophr. Bull. 41 (2), 449e459. Del Bello, V., Menculini, G., Moretti, P., Tortorella, A., 2016. Psychopathological characteristics of patients with first-episode psychosis and chronic schizophrenia: a descriptive comparison. Psychiatr. Danub. 28 (Suppl. 1), 39e44. Eisner, E., Drake, R., Barrowclough, C., 2013. Assessing early signs of relapse in psychosis: review and future directions. Clin. Psychol. Rev. 33 (5), 637e653. Gaebel, W., Riesbeck, M., 2014. Are there clinically useful predictors and early warning signs for pending relapse? Schizophr. Res. 152 (2e3), 469e477. Genduso, L.A., Haley, J.C., 1997. Cost of illness studies for schizophrenia: components, benefits, results, and implications. Am. J. Manag. Care 3 (6), 873e877. Goff, D.C., Falkai, P., Fleischhacker, W.W., Girgis, R.R., Kahn, R.M., Uchida, H., Zhao, J., Lieberman, J.A., 2017. The long-term effects of antipsychotic medication on clinical course in schizophrenia. Am. J. Psychiatry 174 (9), 840e849. Herz, M.I., Melville, C., 1980. Relapse in schizophrenia. Am. J. Psychiatry 137 (7), 801e805. Jørgensen, P., 1998. Early signs of psychotic relapse in schizophrenia. Br. J. Psychiatry 172, 327e330. Kane, J.M., 2007. Treatment strategies to prevent relapse and encourage remission. J. Clin. Psychiatry 68 (Suppl. 14), 27e30. Madhoo, M., Levine, S.Z., 2015. Initial severity effects on residual symptoms in response and remission: a STAR*D study during and after failed citalopram treatment. J. Clin. Psychopharmacol. 35 (4), 450e453. Marder, S.R., Mintz, J., Van Putten, T., Lebell, M., Wirshing, W.C., Johnston-Cronk, K., 1991. Early prediction of relapse in schizophrenia: an application of receiver operating characteristic (ROC) methods. Psychopharmacol. Bull. 27 (1), 79e82. Marder, S.R., Wirshing, W.C., Van Putten, T., Mintz, J., McKenzie, J., JohnstonCronk, K., Lebell, M., Liberman, R.P., 1994. Fluphenazine vs placebo supplementation for prodromal signs of relapse in schizophrenia. Arch. Gen. Psychiatr. 51 (4), 280e287. McCutcheon, R., Beck, K., D’Ambrosio, E., Donocik, J., Gobjila, C., Jauhar, S., Kaar, S., Pillinger, T., Reis Marques, T., Rogdaki, M., Howes, O.D., 2018. Antipsychotic plasma levels in the assessment of poor treatment response in schizophrenia. Acta Psychiatr. Scand. 137 (1), 39e46. Moritz, S., Peters, M.J., Karow, A., Deljkovic, A., Tonn, P., Naber, D., 2009. Cure or curse? Ambivalent attitudes towards neuroleptic medication in schizophrenia and non-schizophrenia patients. Ment. Illn. 1 (1), e2. Moritz, S., Favrod, J., Andreou, C., Morrison, A.P., Bohn, F., Veckenstedt, R., Tonn, P., Karow, A., 2013. Beyond the usual suspects: positive attitudes towards positive symptoms is associated with medication noncompliance in psychosis. Schizophr. Bull. 39 (4), 917e922. Moritz, S., Hünsche, A., Lincoln, T.M., 2014. Nonadherence to antipsychotics: the role of positive attitudes towards positive symptoms. Eur. Neuropsychopharmacol. 24 (11), 1745e1752. Mulder, C.L., Koopmans, G.T., Hengeveld, M.W., 2005. Lack of motivation for treatment in emergency psychiatry patients. Soc. Psychiatry Psychiatr. Epidemiol. 40 (6), 484e488. Norman, R.M., Malla, A.K., 1995. Prodromal symptoms of relapse in schizophrenia: a review. Schizophr. Bull. 21 (4), 527e539. Pillai, A., Schooler, N.R., Peter, D., Looney, S.W., Goff, D.C., Kopelowicz, A., Lauriello, J., Manschreck, T., Mendelowitz, A., Miller, D.D., Severe, J.B., Wilson, D.R., Ames, D., Bustillo, J., Kane, J.M., Buckley, P.F., 2018. Predicting relapse in schizophrenia: is BDNF a plausible biological marker? Schizophr. Res. 193, 263e268. Rocca, P., Galderisi, S., Rossi, A., Bertolino, A., Rucci, P., Gibertoni, D., Montemagni, C., Bellino, S., Aguglia, E., Amore, M., Bellomo, A., Biondi, M., Carpiniello, B., Cuomo, A., D’Ambrosio, E., dell’Osso, L., Girardi, P., Marchesi, C., Monteleone, P., Montemitro, C., Oldani, L., Pacitti, F., Roncone, R., Siracusano, A., Tenconi, E., Vita, A., Zeppegno, P., Steardo Jr., L., Vignapiano, A., Maj, M., 2018. Disorganization and real-world functioning in schizophrenia: results from the multicenter study of the Italian Network for Research on Psychoses. Schizophr. Res. 201, 105e112. Sakurai, H., Uchida, H., Abe, T., Nakajima, S., Suzuki, T., Pollock, B.G., Sato, Y., Mimura, M., 2013. Trajectories of individual symptoms in remitters versus nonremitters with depression. J. Affect. Disord. 151 (2), 506e513. Sakurai, H., Suzuki, T., Yoshimura, K., Mimura, M., Uchida, H., 2017. Predicting relapse with individual residual symptoms in major depressive disorder: a reanalysis of the STAR*D data. Psychopharmacology 234 (16), 2453e2461. Schennach, R., Obermeier, M., Meyer, S., Jager, M., Schmauss, M., Laux, G., Pfeiffer, H., Naber, D., Schmidt, L.G., Gaebel, W., Klosterkotter, J., Heuser, I., Maier, W., Lemke, M.R., Ruther, E., Klingberg, S., Gastpar, M., Seemuller, F., Moller, H.J., Riedel, M., 2012. Predictors of relapse in the year after hospital discharge among patients with schizophrenia. Psychiatr. Serv. 63 (1), 87e90. Schennach, R., Riedel, M., Obermeier, M., Spellmann, I., Musil, R., Jager, M.,
Please cite this article as: Saito, Y et al., Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.10.037
8
Y. Saito et al. / Schizophrenia Research xxx (xxxx) xxx
Schmauss, M., Laux, G., Pfeiffer, H., Naber, D., Schmidt, L.G., Gaebel, W., Klosterkotter, J., Heuser, I., Maier, W., Lemke, M.R., Ruther, E., Klingberg, S., Gastpar, M., Moller, H.J., 2015. What are residual symptoms in schizophrenia spectrum disorder? Clinical description and 1-year persistence within a naturalistic trial. Eur. Arch. Psychiatry Clin. Neurosci. 265 (2), 107e116. Subotnik, K.L., Nuechterlein, K.H., 1988. Prodromal signs and symptoms of schizophrenic relapse. J. Abnorm. Psychol. 97 (4), 405e412. Tandon, R., Nasrallah, H.A., Keshavan, M.S., 2010. Schizophrenia, "just the facts" 5. Treatment and prevention. Past, present, and future. Schizophr. Res. 122 (1e3),
1e23. Velligan, D.I., Weiden, P.J., Sajatovic, M., Scott, J., Carpenter, D., Ross, R., Docherty, J.P., Expert Consensus Panel on Adherence Problems in Serious and Persistent Mental Illness, 2009. The expert consensus guideline series: adherence problems in patients with serious and persistent mental illness. J. Clin. Psychiatry 70 (Suppl. 4), 1e48. Wang, D., Gopal, S., Baker, S., Narayan, V.A., 2018. Trajectories and changes in individual items of positive and negative syndrome scale among schizophrenia patients prior to impending relapse. NPJ Schizophr 4 (1), 10.
Please cite this article as: Saito, Y et al., Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.10.037
Contents lists available at ScienceDirect
Schizophrenia Research journal homepage: www.elsevier.com/locate/schres
Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial Yuta Saito a, b, 1, Hitoshi Sakurai a, c, John M. Kane d, e, f, Nina R. Schooler d, g, Takefumi Suzuki h, Masaru Mimura a, Hiroyuki Uchida a, i, * a
Department of Neuropsychiatry, Keio University School of Medicine, Tokyo, Japan Department of Neuropsychiatry, Kawasaki Municipal Kawasaki Hospital, Kanagawa, Japan Department of Psychiatry, Massachusetts General Hospital, Boston, MA, USA d Department of Psychiatry, The Zucker Hillside Hospital, Northwell Health, New York, USA e Department of Psychiatry and Molecular Medicine, The Donald and Barbara Zucker School of Medicine at Hofstra/Northwell, Hempstead, N.Y, USA f Center of Psychiatric Neuroscience, The Feinstein Institute for Medical Research, New York, USA g Department of Psychiatry and Behavioral Sciences, SUNY Downstate Medical Center, New York, USA h Department of Neuropsychiatry and Clinical Ethics, University of Yamanashi, Yamanashi, Japan i Geriatric Mental Health Program, Centre for Addiction and Mental Health, Toronto, Canada b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 11 January 2019 Received in revised form 4 October 2019 Accepted 12 October 2019 Available online xxx
Introduction: Little attention has been paid to the contribution of individual residual symptom to predict relapse in patients with schizophrenia receiving oral or long-acting injectable (LAI) antipsychotics. Method: We used the data from the Preventing Relapse on Oral Antipsychotics Compared to Injectables Evaluating Efficacy (PROACTIVE) study, in which 305 outpatients with schizophrenia were randomly allocated to either biweekly LAI-risperidone (LAI-R) or daily oral second-generation antipsychotics (SGA) and assessed for up to 30 months. Baseline individual symptoms that could predict subsequent relapse were identified, using a Cox proportional hazards model. Moreover, among those who relapsed during the study (n ¼ 73), individual symptoms were compared between baseline and biweekly ratings 8 to 2 weeks before relapse, using the linear mixed model. Results: A greater score in grandiosity at baseline was significantly associated with subsequent relapse (adjusted HR ¼ 1.24, p ¼ 0.006). When the two treatment groups were separately analyzed, more severe grandiosity (adjusted HR ¼ 1.43, p ¼ 0.003) and less severe hallucinatory behavior (adjusted HR ¼ 0.70, p ¼ 0.013) at baseline were significantly associated with relapse in the oral SGA group, but none was identified in the LAI-R group. Emotional withdrawal was significantly worse 8 and 2 weeks before relapse compared to the baseline (p ¼ 0.032 and p ¼ 0.043, respectively). Discussion: More severe grandiosity and less hallucination may have led to more frequent relapses in patients with schizophrenia receiving oral antipsychotics, which was not a case in those receiving LAI-R. The exploratory analysis indicates an increase in emotional withdrawal before relapse may be a useful marker for earlier interventions to possibly avert relapse. © 2019 Elsevier B.V. All rights reserved.
Keywords: Antipsychotics Early warning signs Relapse Residual symptoms Schizophrenia
1. Introduction Schizophrenia is a chronic, frequently recurrent illness as a high proportion of patients experience relapse after improvement, remission or even recovery (Goff et al., 2017). Relapse results in serious consequences, including deterioration in psychosocial and
* Corresponding author. Department of Neuropsychiatry, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo, 160-8582, Japan. E-mail address: [email protected] (H. Uchida). 1 Both authors equally contributed to this work.
vocational functioning (Pillai et al., 2018), as well as social disruption (Tandon et al., 2010), and puts substantial burden on the patients, families and caregivers (Genduso and Haley, 1997; Kane, 2007). Additionally, relapse increases healthcare costs and potential work force loss (Genduso and Haley, 1997). Therefore, identifying those at high risk for relapse and detecting signs of relapse at the earliest opportunity is of critical importance for the management of this chronic illness. Prediction of relapse based on clinical characteristics, including symptomatology, is clinically relevant in providing individuallytailored medicine (Gaebel and Riesbeck, 2014). There have been
https://doi.org/10.1016/j.schres.2019.10.037 0920-9964/© 2019 Elsevier B.V. All rights reserved.
Please cite this article as: Saito, Y et al., Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.10.037
2
Y. Saito et al. / Schizophrenia Research xxx (xxxx) xxx
some reports suggesting residual positive and negative symptoms as predictors for relapse (Gaebel and Riesbeck, 2014; Schennach et al., 2012). However, these reports focused solely on total scores or symptom domains in the rating scales and did not evaluate the relevance of each individual symptom. In light of the contribution of each residual symptom to subsequent relapse, remission, as well as response reported in other psychiatric illnesses such as major depressive disorder (Madhoo et al., 2015; Sakurai et al., 2013, 2017), further investigations on this topic is warranted in schizophrenia. On the other hand, a majority of patients with schizophrenia experience cognitive, perceptual, and affective change a few weeks before relapse (Birchwood and Spencer, 2001). These early signs which are thought to be inherent symptoms of psychosis or a part of a relapsing process (Birchwood and Spencer, 2001) include nonspecific symptoms such as anxiety, dysphoria, insomnia, poor concentration as well as attenuated psychotic symptoms (Eisner et al., 2013; Gaebel and Riesbeck, 2014). Similarly, several objectively assessed symptoms, including depression and unusual thought content, have been reported to predict subsequent relapse (Marder et al., 1994; Norman and Malla, 1995; Subotnik and Nuechterlein, 1988). However, most of the studies were published in the 1990s and their sample sizes were as small as 50 to 80, which underscores the need of examining these early signs by using a more recent, larger dataset. The Preventing Relapse on Oral Antipsychotics Compared to Injectables - Evaluating Efficacy (PROACTIVE) study provides an ideal dataset derived from generalizable, large-scale samples to address these important issues. Utilizing this dataset, we examined the relevance of individual residual symptoms to the prediction of subsequent relapse. We also tried to evaluate the timing and degree of individual objective symptom worsening before actual relapse. 2. Materials and methods 2.1. Study design The PROACTIVE study was initiated in 2006 and conducted at eight academic centers in the United States to compare the effectiveness of daily oral second-generation antipsychotics (SGAs) and biweekly long-acting injectable risperidone (LAI-R). It was intended to inform clinical decision making by conducting an up-to-date relapse prevention study that included both efficacy/explanatory and effectiveness/pragmatic features; details have been described elsewhere (Buckley et al., 2015). Briefly, 305 outpatients with schizophrenia who were randomized either to oral SGA group or LAI-R group were followed for up to 30 months. Following a complete description of the study, participants provided written informed consent. The original study had been approved by the relevant regulatory bodies, and because of the completely anonymous nature of this analysis and absence of direct human involvement, no ethical approval was sought for the present analysis. 2.2. Study population Inclusion criteria were as follows: (a) outpatients who met the criteria for schizophrenia or schizoaffective disorder according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSM-IV-TR), (b) symptom exacerbation within 12 months and community dwelling for at least 4 weeks, (c) a score of ¼<4 on the Clinical Global Impressions - Severity of Illness (CGI-S), (d) age between 18 and 65 years, and (e) being capable of providing informed consent. Exclusion criteria were: (a) a current episode as their first episode of psychosis, (b) allergy to risperidone, (c) inadequate prior response to risperidone, (d) treatment-refractoriness and/or lack of response to clozapine, and
(e) pregnant or lactating women. 2.3. Treatment Enrolled participants were randomized to an oral SGA or LAI-R. In the oral SGA group, participants continued the ongoing oral SGA or started to take one of the following oral SGAs: aripiprazole, olanzapine, quetiapine, risperidone, paliperidone, asenapine, and iloperidone. In the LAI-R group, LAI-R was initiated with a 25-mg injection. The previous oral antipsychotics or LAIs were tapered in a cross-titration manner. Oral antipsychotics could be titrated or changed based on lack of efficacy, side effects, or patient preference every two weeks. LAI-R was continuously injected biweekly in the range of 12.5 mg to 50 mg and could be supplemented by oral antipsychotic. 2.4. Assessment measures The following assessments were performed by a rater who was blinded to the randomization status via secure, live, two-way video connection: the Brief Psychiatric Rating Scale (BPRS); CGI-S at baseline and quarterly visits; and Clinical Global Impression Global Improvement (CGI-GI) at quarterly visits. In addition, the following assessments were rated biweekly by unblinded raters at the sites: CGI-S; CGI-GI; and an abbreviated version of the BPRS that was composed of eight items (grandiosity, suspiciousness, unusual thought content, hallucinatory behavior, conceptual disorganization, mannerisms and posturing, blunted affect, and emotional withdrawal). Relapse was defined with a CGI-GI score of 6 (much worse) or 7 (very much worse) for the present analysis. It was defined with (1) a CGI-GI score of 6 or 7, (2) psychotic hospitalization except for social one, (3) increase in level of psychotic care, (4) continuous increase in psychotic symptoms that was judged by raters, or (5) self-injury, suicidal or homicidal ideation, and violent behavior that was clinically significant in the original study (Buckley et al., 2015). However, since some of the information on (2)e(5) could not be located for some subjects in the dataset, we solely focused on the CGI. 2.5. Statistical analysis Baseline sociodemographic and clinical characteristics were compared between the oral SGA group and the LAI-R group. Spearman correlation coefficients were calculated between the individual scores on the BPRS at baseline in order to examine multicollinearity. To identify which individual symptoms in the BPRS at baseline could predict subsequent relapse, a Cox hazards model, which is commonly used for investigating the association between survival time and predictor variables, was employed. We ran one Cox proportional hazard model with several individual symptoms included as explanatory variables which was stratified on age, sex, illness duration, and treatment; the Bonferroni correction was not necessary in this analysis. To evaluate the timing and degree of individual symptom worsening before relapse, the individual symptom scores between baseline and 8, 6, 4, and 2 weeks before relapse were compared among those who relapsed (relapsers), using the linear mixed model that contained week as a factor with autoregressive AR (1) correlation matrix among time points. The Bonferroni correction was used to compare the main effects. Those analyses conducted for the whole sample were also repeated separately for the oral SGA group, the LAI-R group, patients with schizophrenia, and those with schizoaffective disorder. To deal with missing values, per-protocol (PP) analysis was employed. A p-value of <0.05 was considered statistically significant. All of the data were analyzed using the Statistical Package for
Please cite this article as: Saito, Y et al., Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.10.037
Y. Saito et al. / Schizophrenia Research xxx (xxxx) xxx
3
Table 1 Sociodemographic and clinical characteristics of subjects. Characteristics Sex Male, n (%) Race Caucasian, n (%) African American, n (%) Hispanic, n (%) Others, n (%) Age, years, mean ± SD Diagnoses Schizophrenia, Paranoid Type, n (%) Schizophrenia Disorganized Type, n (%) Schizophrenia, Residual Type, n (%) Schizophrenia, Undifferentiated Type, n (%) Schizophrenia, NOS, n (%) Schizoaffective Disorder, n (%) Duration of illness, years, mean ± SD Time since last psychiatric hospitalization, months, mean ± SD
All (N ¼ 305)
LAI (n ¼ 153)
Oral SGA (n ¼ 152)
218 (71%)
108 (71%)
110 (72%)
156 (51%) 85 (28%) 58 (19%) 6 (2%) 38.3 ± 12.1
78 (51%) 45 (29%) 27 (18%) 3 (2%) 38.1 ± 11.8
78 (51%) 40 (26%) 31 (20%) 3 (2%) 38.3 ± 12.3
123 (40%) 23 (8%) 1 (0.3%) 54 (18%) 5 (2%) 99 (32%) 17.4 ± 11.6 31.4 ± 64.2
66 (43%) 11 (7%) 0 (0%) 26 (17%) 4 (3%) 46 (30%) 16.7 ± 11.3 22.4 ± 35.9
57 (38%) 12 (8%) 1 (1%) 28 (18%) 1 (1%) 53 (35%) 18.0 ± 11.9 40.4 ± 82.4
Abbreviations: LAI, long-acting injectable; SGA, second-generation antipsychotic; SD, standard deviation.
Social Science (SPSS) version 24 for Windows (IBM Corporation, Armonk, NY).
Table 2 BPRS scores at baseline.
3. Results 3.1. Subject characteristics A total of 305 participants were randomly assigned to the oral SGA group (n ¼ 152) and the LAI-R group (n ¼ 153). The mean ± SD score in the total BPRS score was 38.0 ± 9.4 and the mean score in the CGI-S was 4.0 ± 0.8 at baseline, without significant group differences. Of these participants, an overall 23.9% of patients (n ¼ 73) (26.8% in the LAI-R group and 21.1% in the SGA group) experienced at least one relapse, and the first relapse occurred at a mean ± SD of 42.1 ± 31.9 weeks (41.9 ± 30.5 weeks in the LAI-R group and 42.4 ± 34.6 weeks in the oral SGA group) after the enrollment to the study. There were 129 patients (42.3%) who completed the study without relapse (42.5% in LAI-R group and 42.1% in SGA group; n.s.); 33.8% (n ¼ 103) (30.7% in LAI-R group and 36.8% in SGA group; n.s.) dropped out for reasons other than relapse during the follow-up period. Table 1 summarizes sociodemographic and clinical characteristics of the study sample.
Somatic Concern Anxiety Depressive Mood Guilt Feelings Hostility Suspiciousness Unusual Thought Content Grandiosity Hallucinatory behavior Disorientation Conceptual Disorganization Excitement Motor Retardation Blunted Affect Tension Mannerisms and Posturing Uncooperativeness Emotional Withdrawal Total Score
All (N ¼ 305)
LAI (n ¼ 153)
Oral SGA (n ¼ 152)
2.32 ± 1.25 3.64 ± 1.35 2.90 ± 1.44 2.23 ± 1.30 2.32 ± 1.19 3.48 ± 1.94 3.19 ± 1.73 1.77 ± 1.36 3.14 ± 1.87 1.23 ± 0.51 1.68 ± 1.18 1.17 ± 0.61 1.35 ± 0.76 2.10 ± 1.15 1.75 ± 0.96 1.33 ± 0.79 1.12 ± 0.46 1.32 ± 0.74 38.0 ± 9.36
2.31 ± 1.30 3.75 ± 1.37 2.90 ± 1.50 2.19 ± 1.36 2.25 ± 1.16 3.46 ± 1.96 3.14 ± 1.81 1.67 ± 1.32 3.31 ± 1.94 1.25 ± 0.54 1.66 ± 1.18 1.14 ± 0.54 1.33 ± 0.69 2.13 ± 1.19 1.79 ± 0.99 1.30 ± 0.77 1.14 ± 0.47 1.38 ± 0.76 38.0 ± 9.60
2.34 ± 1.20 3.53 ± 1.33 2.91 ± 1.37 2.27 ± 1.23 2.39 ± 1.23 3.50 ± 1.92 3.25 ± 1.65 1.87 ± 1.40 2.97 ± 1.79 1.21 ± 0.48 1.70 ± 1.19 1.19 ± 0.67 1.36 ± 0.83 2.07 ± 1.12 1.71 ± 0.93 1.36 ± 0.80 1.11 ± 0.46 1.27 ± 0.72 38.0 ± 9.14
Values are shown as mean ± SD. Abbreviations: BPRS, Brief Psychiatric Rating Scale; LAI, long-acting injectable; SGA, second-generation antipsychotic.
3.2. Prediction of relapse by individual symptom severities The mean values of each individual symptom in the BPRS at baseline are shown in Table 2. In all participants, Spearman correlation coefficients among these symptom scores were less than 0.5 except for anxiety and depressive mood, depressive mood and guilt feeling, suspiciousness and unusual thought content, as well as unusual thought content and hallucinatory behavior. In the Cox proportional hazards model, from which depressive mood and unusual thought content were removed in the light of multicollinearity (c2(20) ¼ 24.87), only grandiosity in the BPRS at baseline was significantly associated with subsequent relapse (adjusted HR ¼ 1.24, 95% CI ¼ 1.06e1.45, p ¼ 0.006) (Table 3). Similarly, in the Cox proportional hazards model with all individual symptoms (c2(22) ¼ 26.36), only grandiosity was also significantly associated with subsequent relapse (adjusted HR ¼ 1.26, 95% CI ¼ 1.07e1.50, p ¼ 0.007) (Supplementary Table 1). In the LAI-R group, Spearman correlation coefficients between individual symptom scores at baseline were less than 0.5 except for anxiety and depressive mood, anxiety and guilt feelings, anxiety and hallucinatory behavior, depressive mood and guilt feelings, suspiciousness and unusual thought content, suspiciousness and
Table 3 Associations between individual symptom scores in the BPRS and relapse in all participants. Adjusted Hazard Ratio (95% CI)a Somatic Concern Anxiety Guilt Feelings Hostility Suspiciousness Grandiosity Hallucinatory behavior Disorientation Conceptual Disorganization Excitement Motor Retardation Blunted Affect Tension Mannerisms and Posturing Uncooperativeness Emotional Withdrawal
0.98 1.01 0.98 1.06 1.14 1.24 0.93 1.09 0.99 0.98 0.95 1.15 0.92 0.93 1.01 1.00
(0.79e1.23) (0.79e1.29) (0.78e1.22) (0.83e1.35) (0.98e1.33) (1.06e1.45)b (0.79e1.09) (0.69e1.70) (0.77e1.27) (0.65e1.50) (0.66e1.36) (0.89e1.49) (0.67e1.27) (0.66e1.32) (0.63e1.61) (0.69e1.44)
Abbreviation: BPRS, Brief Psychiatric Rating Scale. a Adjustedafor age, gender, illness duration, treatment, and other individual symptoms. b p < 0.05.
Please cite this article as: Saito, Y et al., Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.10.037
4
Y. Saito et al. / Schizophrenia Research xxx (xxxx) xxx
hallucinatory behavior, unusual thought content and hallucinatory behavior, as well as motor retardation and blunted affect. According to the Cox proportional hazards model, from which anxiety, depressive mood, suspiciousness, and unusual thought content were removed in the light of multicollinearity (c2(17) ¼ 13.78), there were no baseline symptoms that were significantly associated with relapse (Table 4). Also in the Cox proportional hazards model with all individual symptoms (c2(21) ¼ 16.57), there were no baseline symptoms that were significantly associated with relapse, either (Supplementary Table 2). In the oral SGA group, Spearmen correlation coefficients were less than 0.5 except for anxiety and depressive mood, suspiciousness and unusual thought content, as well as unusual thought content and hallucinatory behavior. In contrast to the LAI-R group, according to the Cox proportional hazards model, from which anxiety and unusual thought content were removed in the light of multicollinearity (c2(19) ¼ 19.55), the following two symptoms in the BPRS at baseline were significantly associated with subsequent relapse in the oral SGA group; grandiosity (adjusted HR ¼ 1.43, 95% CI ¼ 1.13e1.80, p ¼ 0.003) and less hallucinatory behavior (adjusted HR ¼ 0.70, 95% CI ¼ 0.53e0.93, p ¼ 0.013) (Table 5). Similarly, according to the Cox proportional hazards model with all individual symptoms (c2(21) ¼ 19.71), greater grandiosity (adjusted HR ¼ 1.42, 95% CI ¼ 1.09e1.86, p ¼ 0.01) and depressive mood (adjusted HR ¼ 1.55, 95% CI ¼ 1.00e2.41, p ¼ 0.049), and less hallucinatory behavior (adjusted HR ¼ 0.70, 95% CI ¼ 0.52e0.95, p ¼ 0.02) were significantly associated with subsequent relapse (Supplementary Table 3). In the schizophrenia group, Spearmen correlation coefficients were less than 0.5 except for anxiety and depressive mood, anxiety and suspiciousness, suspiciousness and unusual thought content, as well as unusual thought content and hallucinatory behavior. According to the Cox proportional hazards model from which anxiety and unusual thought content were removed in the light of multicollinearity (c2(20) ¼ 42.47), the following four symptoms in the BPRS at baseline were significantly associated with subsequent relapse; blunted affect (adjusted HR ¼ 1.47, 95% CI ¼ 1.05e2.06, p ¼ 0.026), hostility (adjusted HR ¼ 1.46, 95% CI ¼ 1.06e2.00, p ¼ 0.019), suspiciousness (adjusted HR ¼ 1.23, 95% CI ¼ 1.01e1.50, p ¼ 0.042), and less hallucinatory behavior (adjusted HR ¼ 0.80, 95% CI ¼ 0.64e0.99, p ¼ 0.041) (Supplementary Table 4). Similarly, according to the Cox proportional hazards model with all individual symptoms (c2(22) ¼ 42.33), blunted affect (adjusted HR ¼ 1.45, 95% CI ¼ 1.04e2.03, p ¼ 0.029), and hostility (adjusted HR ¼ 1.49, 95% CI ¼ 1.07e2.08, p ¼ 0.018) were significantly associated with
Table 4 Associations between individual symptom scores in BPRS and relapse in LAI-R group. Adjusted Hazard Ratio (95% CI)a Somatic Cancern Guilt Feelings Hostility Grandiosity Hallucinatory behavior Disorientation Conceptual Disorganization Excitement Motor Retardation Blunted Affect Tension Mannerisms and Posturing Uncooperativeness Emotional Withdrawal
0.91 0.95 1.03 1.16 1.12 1.28 1.05 0.97 1.05 1.14 0.92 0.99 1.10 1.02
(0.65e1.26) (0.70e1.28) (0.72e1.47) (0.93e1.44) (0.92e1.36) (0.74e2.22) (0.74e1.50) (0.48e1.96) (0.58e1.90) (0.79e1.65) (0.57e1.48) (0.60e1.64) (0.57e2.10) (0.62e1.69)
Abbreviation: BPRS, Brief Psychiatric Rating Scale; LAI-R, long-acting injectable risperidone. a Adjusted for age, gender, illness duration, and other individual symptoms..
Table 5 Associations between individual symptom scores in BPRS and relapse in oral SGA group. Adjusted Hazard Ratio (95% CI)a a
Somatic C ncern Depressive Mood Guilt Feelings Hostility Suspiciousness Grandiosity Hallucinatory behavior Disorientation Conceptual Disorganization Excitement Motor Retardation Blunted Affect Tension Mannerisms and Posturing Uncooperativeness Emotional Withdrawal
1.02 1.42 0.97 1.04 1.27 1.43 0.70 0.79 1.01 1.01 0.87 1.08 0.90 0.70 1.09 1.04
(0.72e1.43) (0.98e2.06) (0.67e1.39) (0.71e1.53) (0.98e1.63) (1.13e1.80)b (0.53e0.93)b (0.33e1.89) (0.68e1.50) (0.57e1.79) (0.52e1.45) (0.71e1.65) (0.54e1.51) (0.40e1.20) (0.47e2.52) (0.54e2.04)
Abbreviation: BPRS, Brief Psychiatric Rating Scale; SGA, second-generation antipsychotic. a Adjusted for age, gender, illness duration, and other individual symptoms. b . p < 0.05.
subsequent relapse (Supplementary Table 5). In the schizoaffective disorder group, Spearmen correlation coefficients were less than 0.5 except for anxiety and depressive mood, depressive mood and hallucinatory behavior, suspiciousness and unusual thought content, and conceptual disorganization and excitement. According to the Cox proportional hazards model, from which depressive mood, unusual thought content, and conceptual disorganization were removed in the light of multicollinearity (c2(19) ¼ 22.97), the following two symptoms in the BPRS at baseline were significantly associated with subsequent relapse; uncooperativeness (adjusted HR ¼ 3.11, 95% CI ¼ 1.22e7.98, p ¼ 0.018) and less somatic concern (adjusted HR ¼ 0.66, 95% CI ¼ 0.44e0.98, p ¼ 0.041) (Supplementary Table 6). According to the Cox proportional hazards model in this group with all individual symptoms (c2(22) ¼ 32.12), uncooperativeness (adjusted HR ¼ 3.61, 95% CI ¼ 1.37e9.53, p ¼ 0.01), motor retardation (adjusted HR ¼ 3.46, 95% CI ¼ 1.47e8.13, p ¼ 0.004), grandiosity (adjusted HR ¼ 2.19, 95% CI ¼ 1.31e3.69, p ¼ 0.003), suspiciousness (adjusted HR ¼ 1.53, 95% CI ¼ 1.00e2.32, p ¼ 0.048), less disorientation (adjusted HR ¼ 0.06, 95% CI ¼ 0.00e0.85, p ¼ 0.04), less hostility (adjusted HR ¼ 0.50, 95% CI ¼ 0.28e0.87, p ¼ 0.01), less blunted affect (adjusted HR ¼ 0.52, 95% CI ¼ 0.28e0.98, p ¼ 0.04), less unusual thought content (adjusted HR ¼ 0.53, 95% CI ¼ 0.32e0.87, p ¼ 0.01), and less somatic concern (adjusted HR ¼ 0.55, 95% CI ¼ 0.35e0.86, p ¼ 0.01) were significantly associated with subsequent relapse (Supplementary Table 7).
3.3. Comparison of individual symptom severities between baseline and pre-relapse among relapsers Scores of unusual thought content, hallucinatory behavior, conceptual disorganization, and emotional withdrawal in the abbreviated BPRS at the time of relapse were significantly higher than those at baseline (Fig. 1). Emotional withdrawal scores were significantly higher 8 and 2 weeks before relapse compared to the baseline value (p ¼ 0.032 and p ¼ 0.043, respectively). No symptom scores before relapse were significantly higher compared to the baseline values in the LAI-R group and the oral SGA group, respectively. While there were no symptoms whose scores before relapse were significantly higher than their baseline values in the schizophrenia group, an emotional withdrawal score was significantly higher 8 weeks before relapse compared to the baseline value in the schizoaffective disorder group (p ¼ 0.018).
Please cite this article as: Saito, Y et al., Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.10.037
Y. Saito et al. / Schizophrenia Research xxx (xxxx) xxx
5
Fig. 1. Symptom scores in the abbreviated BPRS before relapse.
4. Discussion The present exploratory data analysis indicated that more severe grandiosity and less severe hallucinatory behavior at baseline
were associated with subsequent relapse in patients treated with oral SGAs, but no symptoms emerged as predictors in those receiving LAI-R. Additionally, we also found that the score of emotional withdrawal had significantly increased already 8 weeks
Please cite this article as: Saito, Y et al., Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.10.037
6
Y. Saito et al. / Schizophrenia Research xxx (xxxx) xxx
before relapse compared to the baseline score. While it is unclear whether this symptom worsening is a predictor of relapse or a part of a relapsing process, this timeline may offer a window for possible interventions in an effort to avert relapses. The advantage of this study is that not only psychotic symptoms but also other BPRS symptoms were analyzed. These findings underscore the relevance of individual symptoms as well as overall illness severity in predicting relapse in the maintenance treatment of schizophrenia. Among the residual BPRS symptoms, grandiosity was the only symptom that was associated with subsequent relapse in the whole sample. Grandiosity may well have a negative effect on patients’ social function. For example, Rocca et al. examined the associations between positive, negative, and disorganization symptoms, and real-world functioning among 880 patients with schizophrenia, using the Specific Levels of Functioning Scale. In the analysis using structural equation modeling, only grandiosity was identified to be related to poor work skills (Rocca et al., 2018). While grandiosity may remain in a relatively small proportion of patients (Schennach et al., 2015), it was frequently observed in patients with chronic schizophrenia compared to those with first episode psychosis (Del Bello et al., 2016), thus indicating a possibility to serve as a sign of chronicity. Taken together, these findings suggest that residual grandiosity may negatively influence social function and selfmanagement skill, which in turn might lead to further worsening to the point where relapse results, among patients with schizophrenia. Interestingly, in the analysis of the oral SGA group, residual grandiosity and less severity in hallucinatory behavior at baseline emerged as factors associated with subsequent relapse, whereas no residual symptoms including grandiosity were identified in the LAIR group. This difference could be explained by differences in medication adherence between the two groups. In a cross-sectional study of 719 patients in emergency psychiatric settings, grandiosity in the BPRS was found to be associated with a lack of motivation for treatment, suggesting that grandiosity could lead to low adherence (Mulder et al., 2005). Expert consensus guidelines also refer to persistent grandiosity as potential causes of adherence problems in patients with schizophrenia (Velligan et al., 2009). On the other hand, the finding of the association between less hallucinatory behaviors at baseline and subsequent relapse may appear paradoxical. According to a survey by Moritz and colleagues, approximately one quarter of patients with schizophrenia reported at least one positive aspect of the illness as a reason for poor medication adherence (Moritz et al., 2014). Some patients with schizophrenia feel that psychosis makes their life meaningful, that psychosis provides them a feeling of importance and power, and that psychosis helps them to get the attention they deserved; these patients may be at greater risk for poor adherence once psychosis is under control (Moritz et al., 2009, 2013, 2014). However, this somewhat complicated explanation only has any real value if it is confirmed by prospective analysis of a new sample. Thus, these findings might indicate that a greater degree of grandiosity may render adherence to oral antipsychotics worse in this specific group of patients, resulting in subsequent relapse. In this context, measuring antipsychotic concentrations in blood may be useful to identify patients with poor adherence who have an increased risk of relapse (McCutcheon et al., 2018). Neither of subgroup analyses of patients with schizophrenia nor those with schizoaffective disorder did not identify grandiosity as a predictor of subsequent relapse. Moreover, a variety of different symptoms at baseline were found to be associated with subsequent relapse in these patient groups: hostility, suspiciousness, blunted affect, and less severity in hallucinatory behavior in patients with schizophrenia; uncooperativeness and less severity in somatic concern in patients with schizoaffective disorder. These differences may indicate differences in the nature of the illness between these
two disorders or possibilities of type-I and type-II errors. Therefore, the preliminary findings in the present analysis have to be confirmed in larger, prospective clinical trials that are specifically designed to address this issue. In the present study, the biweekly, clinical structured assessment identified emotional withdrawal as an early warning sign as early as eight weeks before relapse in the whole sample. The findings on early signs of relapse in schizophrenia have not always been consistent in the literature. In an early interview study, 145 patients with schizophrenia and 80 of their family members were retrospectively asked about their experiences of early signs of relapse (Herz and Melville, 1980). The symptoms that were frequently reported before relapse were dysphoric in nature: troubled sleep, troubled concentration, loss of appetite, and depressive mood. Previous reviewed articles examined prospective studies that performed weekly to monthly assessments of symptomatology in order to characterize early signs before relapse in patients with schizophrenia (Birchwood and Spencer, 2001; Eisner et al., 2013). They found that some psychotic symptoms (e.g. BPRS thought disturbance subscale) (Subotnik and Nuechterlein, 1988), non-psychotic symptoms (e.g. BPRS anxiety-depression subscale) (Marder et al., 1991), and their combination (e.g. combined Early Signs Scale and Positive and Negative Syndrome Scale general subscale) (Jørgensen, 1998) showed high sensitivity and specificity in predicting subsequent relapse. More recently, Wang et al. analyzed data from three long-term studies and examined symptom changes among 907 patients with schizophrenia who received assessment with the PANSS every 1e4 weeks. 267 patients who relapsed presented with delusions, suspiciousness, hallucinations, anxiety, tension, excitement, and conceptual disorganization as early warning signs (Wang et al., 2018). Notably, this finding are inconsistent with ours probably due to the difference in patients’ characteristics and study design: the participants in the present study had longer duration of illness (17.4 vs 10e12 years [mean]), a greater number of past hospitalizations (10.9 vs 1e3 times [mean]), and worse baseline CGI-S (4.0 vs 2.7e3.5 [mean]), which needs to be taken into account when interpreting the results. Since early warning signs may vary depending on the patients’ background and treatment, further research is clearly needed to confirm our findings. There are several limitations to be noted in the present study. First, this is a reanalysis of the PROACTIVE study, which was not designed to evaluate the issues addressed herein. Second, the generalizability of our findings may be limited, since the participants were outpatients enrolled at eight US academic centers in the PROACTIVE study. Additionally, 71% of the participants were male. Finally, although the dropout rate of 33.8% during 30-month follow-up was not always high, a lower rate might have resulted in slightly different findings. In conclusion, residual grandiosity may serve as a risk factor of subsequent relapse in the long-term treatment of schizophrenia in the whole sample. More notably, severe grandiosity and less hallucination at baseline were associated with subsequent relapse in patients receiving oral SGAs, which was not observed in those treated with LAI-R. These symptoms may lower medication adherence, which in turn may lead to more frequent relapses. Though we did not observe more frequent relapses among patients receiving oral antipsychotics in the present study, we did see significantly less psychopathology in the LAI-R group. Additionally, detection of exacerbation in emotional withdrawal at the earliest opportunity may help to identify those who are likely to relapse in the near future and who could benefit from additional intervention/attention to potentially prevent such a negative event. Since these findings are based on post-hoc analyses they should be replicated in other populations. However, these results point to the value of thorough evaluation of individual symptoms as well as
Please cite this article as: Saito, Y et al., Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.10.037
Y. Saito et al. / Schizophrenia Research xxx (xxxx) xxx
overall symptomatology for relapse prevention among patients with schizophrenia.
7
Funding body agreements and policies No funding support for the present analysis of the data.
Declaration of competing interest Dr. Saito has received speaker’s fee from Dainippon Sumitomo. Dr. Sakurai has received manuscript or speaker’s honoraria from Dainippon Sumitomo, Eli Lilly, Meiji-Seika Pharma, Otsuka Pharmaceutical, Tanabe Mitsubishi Pharma, and Yoshitomi Yakuhin within the past three years. Dr, Kane has received honoraria or consulting fees from Alkermes, Genentech, Intracellular Therapeutics, Janssen, Lundbeck, Merck, Neurocrine, Neuron, Otsuka, Pierre Fabre, Reviva, Roche, Sunovion, Teva. He is a shareholder in LB Pharma and The Vangurad Research Group. Dr Schooler has received honoraria or consulting fees from Alkermes, Allergan, EnVivo/Forum, Lundbeck, Otsuka, Roche and Sunovion in the past three years. Dr. Suzuki has received manuscript or speaker’s fees from Astellas, Dainippon Sumitomo Pharma, Eli Lilly, Elsevier Japan, Janssen Pharmaceuticals, Meiji Seika Pharma, Novartis, Otsuka Pharmaceutical, Wiley Japan, and Yoshitomi Yakuhin, and research grants from Eisai, Mochida Pharmaceutical, and Meiji Seika Pharma. Dr. Mimura has received speaker’s honoraria from Daiichi Sankyo, Dainippon-Sumitomo Pharma, Eisai, Eli Lilly, Fuji Film RI Pharma, Janssen Pharmaceutical, Mochida Pharmaceutical, MSD, Nippon Chemiphar, Novartis Pharma, Ono Yakuhin, Otsuka Pharmaceutical, Pfizer, Takeda Yakuhin, Tsumura, and Yoshitomi Yakuhin within the past three years. Also, he received grants from Daiichi Sankyo, Eisai, Pfizer, Shionogi, Takeda, Tanabe Mitsubishi and Tsumura within the past three years. Dr. Uchida has received grants from Eisai, Otsuka Pharmaceutical, Dainippon-Sumitomo Pharma, Mochida Pharmaceutical, Meiji-Seika Pharmaceutical, and Novartis; speaker’s honoraria from Otsuka Pharmaceutical, Eli Lilly, Shionogi, Pfizer, Yoshitomi Yakuhin, Dainippon-Sumitomo Pharma, Meiji-Seika Pharma, MSD, and Janssen Pharmaceutical; and advisory panel payments from DainipponSumitomo Pharma within the past three years. Acknowledgments Data and/or research tools used in the preparation of this manuscript were obtained and analyzed from the controlled access datasets distributed from the NIMH-supported National Database for Clinical Trials (NDCT). NDCT is a collaborative informatics system created by the National Institute of Mental Health to provide a national resource to support and accelerate discovery related to clinical trial research in mental health. Dataset identifier: NIMH Data Repositories Study identification number #2154. This manuscript reflects the views of the authors and may not reflect the opinions or views of the NIMH or of the submitters submitting original data to NDCT. Data used in the preparation of this article were obtained from the limited access datasets from the Preventing Relapse Oral Antipsychotics Compared to Injectables - Evaluating Efficacy (PROACTIVE) study. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.schres.2019.10.037. Contributors HU designed the study and wrote the protocol. YS and HS managed the data analyses and drafted the manuscript. JMK, NRS, TS, MM, and HU revised the manuscript critically for important intellectual content and contributed substantially to data analysis. All authors have approved the final manuscript and are accountable for all aspects of the work.
References Birchwood, M., Spencer, E., 2001. Early intervention in psychotic relapse. Clin. Psychol. Rev. 21 (8), 1211e1226. Buckley, P.F., Schooler, N.R., Goff, D.C., Hsiao, J., Kopelowicz, A., Lauriello, J., Manschreck, T., Mendelowitz, A.J., Miller del, D., Severe, J.B., Wilson, D.R., Ames, D., Bustillo, J., Mintz, J., Kane, J.M., Study, P., 2015. Comparison of SGA oral medications and a long-acting injectable SGA: the PROACTIVE study. Schizophr. Bull. 41 (2), 449e459. Del Bello, V., Menculini, G., Moretti, P., Tortorella, A., 2016. Psychopathological characteristics of patients with first-episode psychosis and chronic schizophrenia: a descriptive comparison. Psychiatr. Danub. 28 (Suppl. 1), 39e44. Eisner, E., Drake, R., Barrowclough, C., 2013. Assessing early signs of relapse in psychosis: review and future directions. Clin. Psychol. Rev. 33 (5), 637e653. Gaebel, W., Riesbeck, M., 2014. Are there clinically useful predictors and early warning signs for pending relapse? Schizophr. Res. 152 (2e3), 469e477. Genduso, L.A., Haley, J.C., 1997. Cost of illness studies for schizophrenia: components, benefits, results, and implications. Am. J. Manag. Care 3 (6), 873e877. Goff, D.C., Falkai, P., Fleischhacker, W.W., Girgis, R.R., Kahn, R.M., Uchida, H., Zhao, J., Lieberman, J.A., 2017. The long-term effects of antipsychotic medication on clinical course in schizophrenia. Am. J. Psychiatry 174 (9), 840e849. Herz, M.I., Melville, C., 1980. Relapse in schizophrenia. Am. J. Psychiatry 137 (7), 801e805. Jørgensen, P., 1998. Early signs of psychotic relapse in schizophrenia. Br. J. Psychiatry 172, 327e330. Kane, J.M., 2007. Treatment strategies to prevent relapse and encourage remission. J. Clin. Psychiatry 68 (Suppl. 14), 27e30. Madhoo, M., Levine, S.Z., 2015. Initial severity effects on residual symptoms in response and remission: a STAR*D study during and after failed citalopram treatment. J. Clin. Psychopharmacol. 35 (4), 450e453. Marder, S.R., Mintz, J., Van Putten, T., Lebell, M., Wirshing, W.C., Johnston-Cronk, K., 1991. Early prediction of relapse in schizophrenia: an application of receiver operating characteristic (ROC) methods. Psychopharmacol. Bull. 27 (1), 79e82. Marder, S.R., Wirshing, W.C., Van Putten, T., Mintz, J., McKenzie, J., JohnstonCronk, K., Lebell, M., Liberman, R.P., 1994. Fluphenazine vs placebo supplementation for prodromal signs of relapse in schizophrenia. Arch. Gen. Psychiatr. 51 (4), 280e287. McCutcheon, R., Beck, K., D’Ambrosio, E., Donocik, J., Gobjila, C., Jauhar, S., Kaar, S., Pillinger, T., Reis Marques, T., Rogdaki, M., Howes, O.D., 2018. Antipsychotic plasma levels in the assessment of poor treatment response in schizophrenia. Acta Psychiatr. Scand. 137 (1), 39e46. Moritz, S., Peters, M.J., Karow, A., Deljkovic, A., Tonn, P., Naber, D., 2009. Cure or curse? Ambivalent attitudes towards neuroleptic medication in schizophrenia and non-schizophrenia patients. Ment. Illn. 1 (1), e2. Moritz, S., Favrod, J., Andreou, C., Morrison, A.P., Bohn, F., Veckenstedt, R., Tonn, P., Karow, A., 2013. Beyond the usual suspects: positive attitudes towards positive symptoms is associated with medication noncompliance in psychosis. Schizophr. Bull. 39 (4), 917e922. Moritz, S., Hünsche, A., Lincoln, T.M., 2014. Nonadherence to antipsychotics: the role of positive attitudes towards positive symptoms. Eur. Neuropsychopharmacol. 24 (11), 1745e1752. Mulder, C.L., Koopmans, G.T., Hengeveld, M.W., 2005. Lack of motivation for treatment in emergency psychiatry patients. Soc. Psychiatry Psychiatr. Epidemiol. 40 (6), 484e488. Norman, R.M., Malla, A.K., 1995. Prodromal symptoms of relapse in schizophrenia: a review. Schizophr. Bull. 21 (4), 527e539. Pillai, A., Schooler, N.R., Peter, D., Looney, S.W., Goff, D.C., Kopelowicz, A., Lauriello, J., Manschreck, T., Mendelowitz, A., Miller, D.D., Severe, J.B., Wilson, D.R., Ames, D., Bustillo, J., Kane, J.M., Buckley, P.F., 2018. Predicting relapse in schizophrenia: is BDNF a plausible biological marker? Schizophr. Res. 193, 263e268. Rocca, P., Galderisi, S., Rossi, A., Bertolino, A., Rucci, P., Gibertoni, D., Montemagni, C., Bellino, S., Aguglia, E., Amore, M., Bellomo, A., Biondi, M., Carpiniello, B., Cuomo, A., D’Ambrosio, E., dell’Osso, L., Girardi, P., Marchesi, C., Monteleone, P., Montemitro, C., Oldani, L., Pacitti, F., Roncone, R., Siracusano, A., Tenconi, E., Vita, A., Zeppegno, P., Steardo Jr., L., Vignapiano, A., Maj, M., 2018. Disorganization and real-world functioning in schizophrenia: results from the multicenter study of the Italian Network for Research on Psychoses. Schizophr. Res. 201, 105e112. Sakurai, H., Uchida, H., Abe, T., Nakajima, S., Suzuki, T., Pollock, B.G., Sato, Y., Mimura, M., 2013. Trajectories of individual symptoms in remitters versus nonremitters with depression. J. Affect. Disord. 151 (2), 506e513. Sakurai, H., Suzuki, T., Yoshimura, K., Mimura, M., Uchida, H., 2017. Predicting relapse with individual residual symptoms in major depressive disorder: a reanalysis of the STAR*D data. Psychopharmacology 234 (16), 2453e2461. Schennach, R., Obermeier, M., Meyer, S., Jager, M., Schmauss, M., Laux, G., Pfeiffer, H., Naber, D., Schmidt, L.G., Gaebel, W., Klosterkotter, J., Heuser, I., Maier, W., Lemke, M.R., Ruther, E., Klingberg, S., Gastpar, M., Seemuller, F., Moller, H.J., Riedel, M., 2012. Predictors of relapse in the year after hospital discharge among patients with schizophrenia. Psychiatr. Serv. 63 (1), 87e90. Schennach, R., Riedel, M., Obermeier, M., Spellmann, I., Musil, R., Jager, M.,
Please cite this article as: Saito, Y et al., Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.10.037
8
Y. Saito et al. / Schizophrenia Research xxx (xxxx) xxx
Schmauss, M., Laux, G., Pfeiffer, H., Naber, D., Schmidt, L.G., Gaebel, W., Klosterkotter, J., Heuser, I., Maier, W., Lemke, M.R., Ruther, E., Klingberg, S., Gastpar, M., Moller, H.J., 2015. What are residual symptoms in schizophrenia spectrum disorder? Clinical description and 1-year persistence within a naturalistic trial. Eur. Arch. Psychiatry Clin. Neurosci. 265 (2), 107e116. Subotnik, K.L., Nuechterlein, K.H., 1988. Prodromal signs and symptoms of schizophrenic relapse. J. Abnorm. Psychol. 97 (4), 405e412. Tandon, R., Nasrallah, H.A., Keshavan, M.S., 2010. Schizophrenia, "just the facts" 5. Treatment and prevention. Past, present, and future. Schizophr. Res. 122 (1e3),
1e23. Velligan, D.I., Weiden, P.J., Sajatovic, M., Scott, J., Carpenter, D., Ross, R., Docherty, J.P., Expert Consensus Panel on Adherence Problems in Serious and Persistent Mental Illness, 2009. The expert consensus guideline series: adherence problems in patients with serious and persistent mental illness. J. Clin. Psychiatry 70 (Suppl. 4), 1e48. Wang, D., Gopal, S., Baker, S., Narayan, V.A., 2018. Trajectories and changes in individual items of positive and negative syndrome scale among schizophrenia patients prior to impending relapse. NPJ Schizophr 4 (1), 10.
Please cite this article as: Saito, Y et al., Predicting relapse with residual symptoms in schizophrenia: A secondary analysis of the PROACTIVE trial, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.10.037