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

A five year diagnostic follow-up of 1840 patients after a first episode non-schizophrenia and non-affective psychosis

Schizophrenia Research, 1, 150, pages 205 - 210



It is not clear which patients with a first psychotic episode will develop schizophrenia. We performed a diagnostic follow-up of patients treated for a first time non-affective, non-schizophrenia psychosis and explored potential predictors of a subsequent schizophrenia or schizoaffective diagnosis.


This register-based cohort study comprises individuals born between 1973 and 1978 in Sweden, with a first hospital-treated psychosis excluding schizophrenia, schizoaffective disorder, bipolar disorder and depressive disorder with psychotic symptoms (n = 1840). The patients were followed for five years regarding subsequent diagnoses. Psychiatric, social, family history of psychiatric illness, premorbid intellectual level, head injuries and obstetrical complications were investigated by logistic regression as predictors of schizophrenia or schizoaffective diagnosis.


During the follow-up, 18% were diagnosed with schizophrenia or schizoaffective disorder, 5% were diagnosed with bipolar disorder, whereas 29% were not re-admitted to a psychiatric clinic. Patients with a first-degree relative hospitalized for schizophrenia and/or bipolar disorder had an increased risk of subsequent diagnosis for schizophrenia or schizoaffective disorder (odds ratio 1.9 and 95% confidence interval 1.1 to 3.0)), whereas previous severe criminality was associated with a decreased risk (odds ratio 0.5, 95% confidence interval 0.3–0.8).


Diagnostic outcome was diverse after a first non-schizophrenia and non-affective psychosis. Family history of severe mental illness and no previous conviction for severe criminality were the strongest risk factors for a future schizophrenia or schizoaffective diagnosis.

Keywords: Schizophrenia, Diagnostic follow-up, Psychosis, Predictors, Cohort study, Population based.

1. Introduction

Schizophrenia is a diagnosis that is stable over time and with notorious poor prognosis (Robinson et al, 2004, Bromet et al, 2005, and Bromet et al, 2011) However, for patients with a first-episode psychosis (FEP), diagnostic shifts are relatively common during follow-up (Ramirez et al, 2010, Castro-Fornieles et al, 2011, and Kim et al, 2011). Studies of the diagnostic stability in FEP have suggested that other psychosis diagnoses are less stable than schizophrenia spectrum diagnoses (Baldwin et al, 2005, Addington et al, 2006, and Rahm and Cullberg, 2007). Affective psychosis and psychosis not otherwise specified have a low stability (Rahm and Cullberg, 2007 and Bromet et al, 2011).

A significant proportion of patients with schizophrenia have initially been diagnosed with other psychotic disorder many years before a definite schizophrenia diagnosis has been established (Castagnini et al, 2008 and Bromet et al, 2011). On the other hand, patients with certain psychosis diagnoses such as acute transient psychosis may not develop schizophrenia and about 20% are not readmitted ( Castagnini et al., 2008 ). A number of risk factors for developing schizophrenia in a population perspective have been identified: familial factors, poor premorbid adjustment ( Ramirez et al., 2010 ), prodromal symptoms ( Moukas et al., 2010 ), high paternal age, longer duration of untreated psychosis, obstetrical complications ( Maki et al., 2005 ), head injuries, and substance abuse ( Maki et al., 2005 ). Nevertheless, there is a lack of knowledge regarding which of the patients with a first psychotic episode will develop a chronic psychotic disorder ( Murphy, 2010 ). Available studies have suggested poor function, more negative and positive symptoms ( Bromet et al., 2011 ), and having first-degree relatives with schizophrenia ( Das et al., 1999 ), as predictors of a subsequent schizophrenia diagnosis after a first non-affective psychotic break, but these studies are hampered by small and selected samples with high loss to follow-up.

Predicting the course of a psychotic disorder by analyzing diagnostic shifts over time is of great importance not only to give reliable information about the clinical prognosis to both patients and their families, but also to implement adequate therapeutic and psychosocial interventions. With the aim of describing shifts of diagnoses over time, we performed a register based diagnostic follow-up of patients treated for a first episode psychosis that was neither schizophrenia nor an affective psychosis. A secondary aim was to assess potential predictors for a future diagnosis of schizophrenia or schizoaffective disorder.

2. Method

2.1. Study population

Swedish national registers make it possible to study the entire Swedish population and to perform linkage of data between different registers on an individual level. In the present study, the unique personal identity number assigned to each permanent resident in Sweden was used to link information from nine population-based registers ( Ludvigsson et al., 2009 ).

The Medical Birth Register, established in 1973, includes information on almost all births in Sweden ( Cnattingius et al., 1990 ). The National Patient Register includes all individuals admitted to psychiatric or general hospitals, with nearly complete coverage for psychiatric care since 1973 and for somatic care since 1987 ( Ludvigsson et al., 2011 ). Through the Multi-Generation Register one is able to link children and parents (biological and adoptive) together. The Causes of Death Register comprises information on all deaths of Swedish residents since 1952 ( National Board of Health and Welfare, 2011 ). The Register of Court Convictions contains information on all court convictions in Sweden for persons 15 years of age or older ( National Council for Crime Prevention, 2011 ). We used the Swedish Register of Children and Young Persons subjected to child welfare measures to obtain records on out-of-home care foster family and residential care. The Total Enumeration Income Survey contains data on the income of and governmental benefits provided to all Swedish residents. The Total Population Register, established in 1968, includes information on age, sex, place of residence etc. ( Statistics Sweden, 2009 ). Finally, the National School Register contains information on school grades.

The selection of the study population is illustrated in Fig. 1 . We only included individuals with at least one biological parent born in Sweden and for whom we could obtain the personal identity number of the biological mother. We selected all individuals with a first hospital-treated psychosis (as defined by the International Classification of Disease, ICD) from age 15 (1988–1993) until year 2003 from the National Patient Register. We chose this age as cut off because it is uncommon to be diagnosed before this age and diagnoses among younger individuals may therefore be unreliable. A total of six individuals had been diagnosed before age 15. We further excluded all individuals who had a hospitalization with any of the following primary diagnoses: schizophrenia, schizoaffective disorder or bipolar disorder. The reason for this is that these diagnoses are considered stable over time and were defined by us as outcome variables. Moreover, depressive disorder with psychotic symptoms was excluded because it is difficult to translate this diagnosis between different versions of ICD. Our final cohort comprised 1840 individuals. These patients were classified according to the following index diagnoses: persistent delusional disorder, psychotic disorder due to substance use, acute and transient psychotic disorder, and other psychoses.


Fig. 1 Flow chart for the study population.

2.2. Variables

2.2.1. Predictors

We assessed a number of potential predictors occurring before the first diagnosis that we categorized into six main categories: psychiatric factors, social factors, premorbid intellectual functional level, familial factors, and other predictors (including injuries to the head, injuries sustained during delivery, and obstetrical complications). Psychiatric factors

We created three dichotomous variables using the National Patient Register. Only hospitalizations occurring before the first diagnosis for psychosis were considered.

  • - Former in-patient care with any psychiatric diagnosis
  • - Former in-patient care with a diagnosis for substance abuse
  • - Former in-patient care for intentional self-harm Social factors

As indicator for severe criminality we selected convictions that led to severe sentences, i.e. imprisonment or probation. This information was obtained from the Register of Court Convictions.

Experience of interventions by the social services before age 12 is a well-known risk factor for mental health problems (Hjern et al, 2004 and Vinnerljung et al, 2006). Child welfare intervention, retrieved from the Swedish Register of Children and Young Persons subjected to child welfare measures, was defined as out-of-home care or provision of a respite care. Parental social assistance dependency was measured when the individual was between ages 12 and 17. To fulfill this criterion, at least one parent had to be receiving social assistance during at least one year where more than 50% of the yearly income constituted social assistance. The information was obtained from the Total Enumeration Income Survey.

Place of residence at the time of first diagnosis, obtained from the Total Population Register was categorized as: big city (the three largest city areas in Sweden), smaller city (> 90 000 inhabitants) or rural municipalities. Premorbid intellectual functional level

In an attempt to capture academic performance pre diagnosis we created the following three variables:

  • - No, incomplete or low grades from the nine year compulsory school, defined as less than one standard deviation (SD) below average
  • - Grade point average from the nine year compulsory school
  • - No graduation from high school at age 20. Familial factors

Using the National Patient Register, we created dichotomous variables for familial mental disorders (i.e. biological parents or siblings) occurring any time between 1973 and the individuals' index diagnosis, for the following diagnoses: schizophrenia or bipolar disorder, substance abuse or other mental disorder. Other predictors

The National Patient Register was used to retrieve information on hospitalization for intracranial injuries, or fractures of skull and facial bones to form the predictor “head injuries”. We also created a variable called “any obstetrical complications” comprising the following: preeclampsia, neonatal jaundice, being born small for gestational age, preterm birth (before 33 weeks of gestation), microcephaly (i.e. head circumference < 32 cm) and an Apgar score of < 7 at 5 min.

2.2.2. Outcome

The individuals of the study population were followed for five years from the time of the first diagnosis. Four outcome groups were created according to a hierarchy of diagnoses recorded in the National Patient Register: i) schizophrenia or schizoaffective disorder (ICD-9: 295; ICD-10: F20, F25), ii) bipolar disorder (ICD-9: 296A, 296C-296E, 296W; ICD-10: F30-31), iii) other diagnoses of mental disorder except the ones above, and iv) no diagnoses for mental disorder during the follow up period. The primary outcome was a diagnosis of schizophrenia or schizoaffective disorder. We further studied a set of diagnosis groups that were not mutually exclusive: diagnoses for substance abuse, depression, mental retardation or autism, and other diagnoses for psychoses.

2.3. Statistical analysis

In the first descriptive part of the study we compared background factors and follow-up diagnoses between the different exposure groups. A bivariate descriptive analysis with a frequency distribution including proportions was performed. In the second analytical part, patients who died during the follow-up period were censored (for the outcome schizophrenia and schizoaffective disorder: n = 79, for bipolar disorder: n = 97). The outcome groups were defined as binary variables. Unconditional logistic regression analyses were used to statistically evaluate the association between the predictors and the primary outcome. Each predictor was studied in separate regression analyses.

Odds ratios (OR) with 95% confidence intervals (CI) were calculated using a SAS v. 9.2 (SAS Institute Inc. Cary, NC, USA) in order to estimate risk ratios.

2.4. Ethics statement

The study population was based on linkage of several public national registers. Ethical vetting is always required when using register data for research purposes in Sweden. The ethical vetting is performed by a regional ethical review board and the risk appraisal associated with the Law on Public Disclosure and Secrecy is done by data owners. The ethical review boards can however waive the requirement to consult the data subjects (or in case of minors/children the next of kin, careers or guardians) directly to obtain their informed consent, and will often do so if the research is supported by the ethical review board and the data has already been collected in some other context. This project was evaluated and approved according to these standards by the Regional Ethical Review Board in Stockholm, Sweden.

4. Results

Cohort characteristics for the 1840 patients (728 women and 1 112 men) are presented in Table 1 . Nearly 50% of the patients had a baseline diagnosis of acute and transient psychotic disorder, followed by psychotic disorders due to substance use (24%). Almost 25% were diagnosed with other psychoses (mostly unspecified nonorganic psychosis). The remaining 81 patients were initially diagnosed with persistent delusional disorder. More than 30% had a history of in-patient care with a psychiatric diagnosis and 14% had previously been hospitalized with a diagnosis for substance abuse. Approximately one third of the patients had incomplete or no grades from the nine-year compulsory school and almost half of them had not finished secondary school at age 20. Nearly half of the patients had a first-degree biological relative hospitalized for mental disorder.

Table 1 Cohort characteristics.

  Persistent delusional disorders Psychotic disorder due to substance use Acute and transient psychotic disorders Other psychoses Total
N 81 (4%) 445 (24%) 868 (47%) 446 (25%) 1840 (100%)
Mean age at index diagnosis (sd) 24.0 (3.5) 23.2 (3.3) 21.3 (3.6) 24.2 (2.6) 22.6 (3.6)
Psychiatric factors
Former inpatient care with a psychiatric diagnosis 27 (33%) 199 (45%) 225 (26%) 135 (30%) 586 (32%)
Former inpatient care with a diagnosis for substance abuse 9 (11%) 154 (35%) 56 (6%) 42 (9%) 261 (14%)
Former inpatient care for intentional self-harm 7 (9%) 58 (13%) 50 (6%) 35 (8%) 150 (8%)
Social factors
First child welfare intervention before age 12 6 (7%) 72 (16%) 78 (9%) 38 (9%) 194 (11%)
Severe criminality 6 (7%) 97 (22%) 37 (4%) 24 (5%) 164 (9%)
Parental social assistance recipiency 17 (21%) 137 (31%) 142 (16%) 78 (17%) 374 (20%)
Place of residence at index diagnosis          
 Small (rural municipalities) 23 (28%) 115 (26%) 234 (27%) 110 (25%) 482 (26%)
 Intermediate (> 90 000 inhabitants) 30 (37%) 173 (39%) 326 (38%) 151 (34%) 680 (37%)
 Big city area (the three large city areas) 28 (35%) 157 (35%) 308 (35%) 185 (41%) 678 (37%)
Premorbid intellectual functional level
No/incomplete grades from compulsory school 29 (36%) 250 (56%) 242 (28%) 122 (27%) 643 (35%)
Grade point average from compulsory school 2.8 (0.8) 2.4 (0.7) 3.1 (0.8) 3.0 (0.8) 2.9 (0.8)
No graduation from high school at age 20 34 (42%) 273 (61%) 376 (43%) 171 (38%) 854 (46%)
Familial factors
First degree relative hospitalized for schizophrenia and/or bipolar disorder 2 (2%) 10 (2%) 48 (6%) 25 (6%) 85 (5%)
First degree relative hospitalized for substance abuse disorder 6 (7%) 86 (19%) 63 (7%) 43 (10%) 198 (11%)
First degree relative hospitalized with a diagnosis for mental disorder 26 (32%) 229 (51%) 366 (42%) 182 (41%) 803 (44%)
Other predictors
Head injuries (ICD-10: S061-S069, S02) 1 (1%) 22 (5%) 20 (2%) 10 (2%) 53 (3%)
Any obstetrical complications 20 (25%) 73 (16%) 133 (15%) 59 (13%) 285 (15%)

Table 2 presents the follow-up diagnoses for the patients. Almost 20% were diagnosed with schizophrenia or schizoaffective disorder during the five years following the index diagnosis whereas 5% developed bipolar disorder. Another 5% died during the follow-up period. Around one third were not re-admitted to a psychiatric clinic.

Table 2 Follow-up diagnoses.

  Persistent delusional disorders Psychotic disorder due to substance use Acute and transient psychotic disorders Other Total
Same as index diagnosis 17 (21%) 71 (8%) 88 (5%)
Schizophrenia or schizoaffective disorder 14 (17%) 35 (8%) 192 (22%) 96 (22%) 337 (18%)
Bipolar disorder 4 (5%) 6 (1%) 69 (8%) 18 (4%) 97 (5%)
Other diagnoses for mental disorder (except the ones above) 18 (22%) 248 (56%) 250 (29%) 177 (40%) 693 (38%)
No diagnosis for mental disorder during the follow-up period 28 (35%) 156 (35%) 286 (33%) 155 (35%) 625 (33%)
Subgroups (not mutually exclusive)          
 Diagnosis for substance abuse 5 (6%) 204 (46%) 52 (6%) 39 (9%) 300 (16%)
 Diagnosis for depression 4 (5%) 39 (9%) 69 (8%) 33 (7%) 145 (8%)
 Diagnosis for mental retardation or autism 3 (4%) 3 (1%) 18 (2%) 20 (4%) 44 (2%)
 Other diagnoses for psychoses 23 (28%) 31 (7%) 183 (21%) 110 (25%) 347 (19%)

Odds ratios for a subsequent diagnosis of schizophrenia or schizoaffective disorder by the different predictors are presented in Table 3 . Patients with a first-degree relative hospitalized for schizophrenia and/or bipolar disorder had an increased risk of subsequent diagnosis of schizophrenia or schizoaffective disorder (OR: 1.9 [95% CI 1.1 to 3.0]) when adjusting for sex and age. A slight risk increase for schizophrenia (however not significant) was found for patients with no graduation from secondary school at age 20. Previous severe criminality was associated with a decreased risk of a subsequent diagnosis of schizophrenia or schizoaffective disorder.

Table 3 Odds ratios a for predictors of having a subsequent schizophrenia or schizoaffective disorder diagnosis.

Predictors Schizophrenia or schizoaffective disorder
n = 337 OR (95% CI)
Former inpatient care with a psychiatric diagnosis 111 1.1 (0.9–1.4)
Former inpatient care with a diagnosis for substance abuse 42 0.9 (0.6–1.3)
Former inpatient care for intentional self-harm 17 0.6 (0.4–1.0)
First child welfare intervention before age 12 37 1.0 (0.7–1.5)
Severe criminality 18 0.5 (0.3–0.8)
Parental social assistance recipiency 66 0.9 (0.7–1.3)
No/incomplete grades from the compulsory school 141 0.9 (0.7–1.2)
No graduation from high school at age 20 174 1.3 (1.0–1.7)
First degree relative hospitalized for schizophrenia and/or bipolar disorder 24 1.9 (1.1–3.0)
First degree relative hospitalized for substance abuse disorder 23 0.6 (0.4–0.9)
First degree relative hospitalized with a diagnosis for other mental disorder 137 0.9 (0.7–1.1)
Any obstetrical complication 49 0.9 (0.7–1.3)
Head injuries 8 0.8 (0.4–1.7)

a Adjusted for sex and birth year.

5. Discussion

5.1. Key results

Our study of more than 1800 patients hospitalized with a first non-schizophrenia and non-affective psychosis showed that approximately one fifth of the patients in the cohort shifted diagnosis to schizophrenia or schizoaffective disorder during the first five years. A third of the patients were not readmitted during the first five years. A family history of schizophrenia or bipolar disorder was the strongest predictor for a subsequent schizophrenia or schizoaffective diagnosis, whereas previous severe criminality was associated with a decreased risk.

5.2. Interpretation & comparison with other studies

Our results show that there is a wide range in the outcome after a first non-schizophrenia and non-affective psychosis diagnoses; a third of the patients in our cohort were not readmitted during the first five years, whereas one in every five patients received a subsequent schizophrenia or schizoaffective disorder diagnosis. Results from earlier studies have been inconsistent, for instance a study evaluating the stability of ICD-10 diagnoses among first-episode psychosis patients showed that 15% shifted to schizophrenia during a five-year follow-up ( Chang et al., 2009 ) whereas another study reported 32% ( Castagnini et al., 2008 ), though with different settings of psychosis patients. A Dutch study reported an even higher proportion ( Veen et al., 2004 ). A shift to schizophrenia spectrum disorder has often been the most frequent diagnostic change (Veen et al, 2004 and Subramaniam et al, 2007), suggesting that schizophrenic disorders are underdiagnosed rather than over-diagnosed at first contact. Another Swedish study following 146 first-episode psychosis that also included schizophrenia as first diagnosis, showed that, whereas very few patients leave the schizophrenic spectrum group, many with other first diagnoses change to specific diagnoses or end up as recovered over a 3-year follow-up period ( Rahm and Cullberg, 2007 ). Compared to one third in our study, over 50% did not receive a psychosis diagnosis at follow-up. A consistent finding in studies that have included all first-episode patients is the high diagnostic stability over time for schizophrenia, followed by affective psychoses (Veen et al, 2004, Whitty et al, 2005, Rahm and Cullberg, 2007, Subramaniam et al, 2007, and Chang et al, 2009). The follow-up time has typically spanned from two to five years. The lowest stability has been shown in categories of drug-induced psychosis, and “other” psychosis (Whitty et al, 2005 and Chang et al, 2009).

The strongest predictor for a subsequent schizophrenia or schizoaffective disorder diagnosis in our cohort of patients with a first-episode psychosis was having a first-degree relative with schizophrenia or bipolar disorder. The finding is concordant with population based studies on genetic risk factors for schizophrenia in the general population (Lichtenstein et al, 2009 and Murphy, 2010), but one of largest previous clinical follow-up studies did not identify psychiatric family history as a risk factor for subsequent schizophrenia diagnosis ( Schwartz et al., 2000 ). However, our cohort was more than three times larger and did not rely on interview or case-note data regarding family history, which may have introduced report bias in the study by Schwartz et al. (2000) . Several other clinical follow-up studies have not investigated psychiatric family history as a risk factor (Schimmelmann et al, 2005 and Salvatore et al, 2009). Earlier studies have often been based on small samples with rather short follow-up periods (Addington et al, 2006, Chang et al, 2009, and Castro-Fornieles et al, 2011). With the use of national registers, we were able to study a large cohort with a long follow-up time.

We also observed that severe criminality was associated with a decreased risk of a subsequent schizophrenia diagnosis. To our knowledge, this is a novel finding. This decreased risk should be interpreted in the light of the other binary outcome, not having a subsequent schizophrenia diagnosis, which encompasses not having another admission at all or being admitted for other diagnoses. Maybe these patients are at higher risk of drug induced psychoses or of being imprisoned and therefore not having further diagnoses. Having a history of self-harm was also associated with decreased risk of schizophrenia or schizoaffective disorder, even though of borderline statistical significance. Both these predictors might be grouped as factors associated with impulsivity and could therefore be more associated to other diagnoses such as personality disorders, affective disorders and substance use disorders, than to schizophrenia. This result can be compared with a previous study of the stability of first episode psychosis diagnoses where aggression was associated with subsequent shift to substance induced psychosis and suicidality to a shift to major depressive disorder ( Bromet et al., 2011 ).

According to our results, poor school performance did not increase the risk of subsequent diagnosis for schizophrenia in first-episode psychosis patients. Studies of the association of school performance and subsequent schizophrenia have been inconclusive. There are studies which have shown associations between excellent school performance preceding schizophrenia (Isohanni et al, 1999 and Karlsson, 2004), poor school performance (van Oel et al, 2002 and MacCabe et al, 2008), as well as no association between school performance and schizophrenia ( Isohanni et al., 1998 ).

In our study, 458 patients diagnosed with schizophrenia or schizoaffective disorder at first hospitalization were excluded. Nearly 20% of the remaining patients received a subsequent schizophrenia or schizoaffective diagnoses. These observations can be compared with the results of a previous Danish population based study which found that almost every second patient with a first admission for schizophrenia have had previous admissions for other psychotic disorders ( Mortensen and Eaton, 1994 ). This is an important observation as studies only including patients with a first hospitalization for schizophrenia and not having previous exposure to antipsychotics (Tiihonen et al, 2011 and Tiihonen et al, 2012), thus are studying selected schizophrenia populations as they do not include patients that first have another psychosis diagnosis and later develop schizophrenia.

Although there was as low mean age of our cohort as 22.6 years, 5% died during five years of follow-up. This can be compared with 3% in an Australian study including all first-episode psychosis diagnoses ( Robinson et al., 2010 ). Excess mortality has been repeatedly reported in patients with schizophrenia (Bromet et al, 2011 and Bjorkenstam et al, 2012), and suicide is one of the contributors to the mortality of schizophrenia and especially first-episode psychosis patients (Robinson et al, 2004, Reutfors et al, 2009, and Pompili et al, 2011). Thus, mortality in psychotic disorders may be underestimated in studies only including first admission schizophrenia as some patients may initially have been diagnosed with other provisional psychosis diagnoses and then die before receiving a definite schizophrenia diagnosis.

5.3. Clinical implication

The identified predictors of the risk of a subsequent schizophrenia or schizoaffective diagnosis may be used clinically to identify first diagnosed non-affective non-schizophrenia psychoses patients at high risk of subsequent schizophrenia to better communicate prognosis to patients and relatives. Moreover, early identification of patients who are at higher risk of developing a chronic psychotic disorder might inform a rational allocation of health care resources to patients with most needs.

5.4. Strengths & limitations

The strength of the study includes the population-based design, using national registers with high completeness and validity (Dalman et al, 2002, Hjern et al, 2004, Ekholm et al, 2005, and Ludvigsson et al, 2011). The large cohort including all first psychosis patients in Sweden with a complete follow-up yields high generalizability. The richness of risk factors ranging from family history to early environmental factors such as obstetrical complications and clinical factors such as comorbid psychiatric diagnoses allowed a multivariate assessment of these risk factors which has been called for ( Murphy, 2010 ).

However, relying solely on register data, there are important predictors that we were not able to study, which is a limitation. Such predictors not available in our registers are detailed baseline clinical variables such as symptoms, cognitive assessments and medication, duration of untreated psychosis, symptoms' evolution, and age of onset. Another limitation is the lack of validation studies of delusional disorder, as well as acute and unspecified psychoses diagnoses. On the other hand the clinical schizophrenia spectrum diagnoses and bipolar disorder are considered to have high validity (Ludvigsson et al, 2011 and Sellgren et al, 2011).

5.5. Conclusion

There is a wide range of diagnostic outcomes after a first non-schizophrenia and non-affective psychosis, but few predictors of subsequent schizophrenia or schizoaffective diagnoses have been identified. Further studies including clinical variables such as symptoms and medication response and biological predictors are warranted.

Role of funding

This study was supported by unrestricted grants from Märta and Nicke Nasvell Foundation.


RB and JR originated the idea. EB and CB analyzed the data in consultation with RB. EB and RB wrote the manuscript draft. All authors contributed in designing the study and to the interpretation of the results and to the writing and approval of the final article.

Conflict of interest





  • Addington et al., 2006 J. Addington, A. Chaves, D. Addington. Diagnostic stability over one year in first-episode psychosis. Schizophr. Res.. 2006;86(1–3):71-75
  • Baldwin et al., 2005 P. Baldwin, D. Browne, P.J. Scully, J.F. Quinn, M.G. Morgan, A. Kinsella, J.M. Owens, V. Russell, E. O'Callaghan, J.L. Waddington. Epidemiology of first-episode psychosis: illustrating the challenges across diagnostic boundaries through the Cavan–Monaghan study at 8 years. Schizophr. Bull.. 2005;31(3):624-638
  • Bjorkenstam et al., 2012 E. Bjorkenstam, R. Ljung, B. Burstrom, E. Mittendorfer-Rutz, J. Hallqvist, G.R. Weitoft. Quality of medical care and excess mortality in psychiatric patients—a nationwide register-based study in Sweden. BMJ open. 2012;2:e000778
  • Bromet et al., 2005 E.J. Bromet, B. Naz, L.J. Fochtmann, G.A. Carlson, M. Tanenberg-Karant. Long-term diagnostic stability and outcome in recent first-episode cohort studies of schizophrenia. Schizophr. Bull.. 2005;31(3):639-649
  • Bromet et al., 2011 E.J. Bromet, R. Kotov, L.J. Fochtmann, G.A. Carlson, M. Tanenberg-Karant, C. Ruggero, S.W. Chang. Diagnostic shifts during the decade following first admission for psychosis. Am. J. Psychiatry. 2011;168(11):1186-1194
  • Castagnini et al., 2008 A. Castagnini, A. Bertelsen, G.E. Berrios. Incidence and diagnostic stability of ICD-10 acute and transient psychotic disorders. Compr. Psychiatry. 2008;49(3):255-261
  • Castro-Fornieles et al., 2011 J. Castro-Fornieles, I. Baeza, E. de la Serna, A. Gonzalez-Pinto, M. Parellada, M. Graell, D. Moreno, S. Otero, C. Arango. Two-year diagnostic stability in early-onset first-episode psychosis. J. Child Psychol. Psychiatry. 2011;52(10):1089-1098
  • Chang et al., 2009 W.C. Chang, S.L. Pang, D.W. Chung, S.S. Chan. Five-year stability of ICD-10 diagnoses among Chinese patients presented with first-episode psychosis in Hong Kong. Schizophr. Res.. 2009;115(2–3):351-357
  • Cnattingius et al., 1990 S. Cnattingius, A. Ericson, J. Gunnarskog, B. Kallen. A quality study of a medical birth registry. Scand. J. Soc. Med.. 1990;18(2):143-148
  • Dalman et al., 2002 C. Dalman, J. Broms, J. Cullberg, P. Allebeck. Young cases of schizophrenia identified in a national inpatient register—are the diagnoses valid?. Soc. Psychiatry Psychiatr. Epidemiol.. 2002;37(11):527-531
  • Das et al., 1999 S.K. Das, S. Malhotra, D. Basu. Family study of acute and transient psychotic disorders: comparison with schizophrenia. Soc. Psychiatry Psychiatr. Epidemiol.. 1999;34(6):328-332
  • Ekholm et al., 2005 B. Ekholm, A. Ekholm, R. Adolfsson, M. Vares, U. Osby, G.C. Sedvall, E.G. Jonsson. Evaluation of diagnostic procedures in Swedish patients with schizophrenia and related psychoses. Nord. J. Psychiatry. 2005;59(6):457-464
  • Hjern et al., 2004 A. Hjern, B. Vinnerljung, F. Lindblad. Avoidable mortality among child welfare recipients and intercountry adoptees: a national cohort study. J. Epidemiol. Community Health. 2004;58(5):412-417
  • Isohanni et al., 1998 I. Isohanni, M.R. Jarvelin, P. Nieminen, P. Jones, P. Rantakallio, J. Jokelainen, M. Isohanni. School performance as a predictor of psychiatric hospitalization in adult life. A 28-year follow-up in the Northern Finland 1966 Birth Cohort. Psychol. Med.. 1998;28(4):967-974
  • Isohanni et al., 1999 I. Isohanni, M.R. Jarvelin, P. Jones, J. Jokelainen, M. Isohanni. Can excellent school performance be a precursor of schizophrenia? A 28-year follow-up in the Northern Finland 1966 birth cohort. Acta Psychiatr. Scand.. 1999;100(1):17-26
  • Karlsson, 2004 J.L. Karlsson. Psychosis and academic performance. Br. J. Psychiatry J. Ment. Sci.. 2004;184:327-329
  • Kim et al., 2011 J.S. Kim, J.H. Baek, J.S. Choi, D. Lee, J.S. Kwon, K.S. Hong. Diagnostic stability of first-episode psychosis and predictors of diagnostic shift from non-affective psychosis to bipolar disorder: a retrospective evaluation after recurrence. Psychiatry Res.. 2011;188(1):29-33
  • Lichtenstein et al., 2009 P. Lichtenstein, B.H. Yip, C. Björk, Y. Pawitan, T.D. Cannon, P.F. Sullivan, C.M. Hultman. Common genetic determinants of schizophrenia and bipolar disorder in Swedish families: a population-based study. Lancet. 2009;373(9659):234-239
  • Ludvigsson et al., 2009 J.F. Ludvigsson, P. Otterblad-Olausson, B.U. Pettersson, A. Ekbom. The Swedish personal identity number: possibilities and pitfalls in healthcare and medical research. Eur. J. Epidemiol.. 2009;24(11):659-667
  • Ludvigsson et al., 2011 J.F. Ludvigsson, E. Andersson, A. Ekbom, M. Feychting, J.L. Kim, C. Reuterwall, M. Heurgren, P.O. Olausson. External review and validation of the Swedish national inpatient register. BMC Publ. Health. 2011;11:450
  • MacCabe et al., 2008 J.H. MacCabe, M.P. Lambe, S. Cnattingius, A. Torrang, C. Bjork, P.C. Sham, A.S. David, R.M. Murray, C.M. Hultman. Scholastic achievement at age 16 and risk of schizophrenia and other psychoses: a national cohort study. Psychol. Med.. 2008;38(8):1133-1140
  • Maki et al., 2005 P. Maki, J. Veijola, P.B. Jones, G.K. Murray, H. Koponen, P. Tienari, J. Miettunen, P. Tanskanen, K.E. Wahlberg, J. Koskinen, E. Lauronen, M. Isohanni. Predictors of schizophrenia—a review. Br. Med. Bull.. 2005;73–74:1-15
  • Mortensen and Eaton, 1994 P.B. Mortensen, W.W. Eaton. Predictors for readmission risk in schizophrenia. Psychol. Med.. 1994;24(1):223-232
  • Moukas et al., 2010 G. Moukas, A. Stathopoulou, P. Gourzis, I.N. Beratis, S. Beratis. Relationship of “prodromal” symptoms with severity and type of psychopathology in the active phase of schizophrenia. Compr. Psychiatry. 2010;51(1):1-7
  • Murphy, 2010 B.P. Murphy. Beyond the first episode: candidate factors for a risk prediction model of schizophrenia. Int. Rev. Psychiatry. 2010;22(2):202-223
  • National Board of Health and Welfare, 2011 National Board of Health and Welfare. Causes of Death 2010. National Board of Health and Welfare (Stockholm, Sweden, 2011)
  • National Council for Crime Prevention, 2011 National Council for Crime Prevention. Kriminalvårdsstatistik 2011 [Criminal statistics 2011]. (Edita Västra Aros, Västerås, Sweden, 2011)
  • Pompili et al., 2011 M. Pompili, G. Serafini, M. Innamorati, D. Lester, A. Shrivastava, P. Girardi, M. Nordentoft. Suicide risk in first episode psychosis: a selective review of the current literature. Schizophr. Res.. 2011;129(1):1-11
  • Rahm and Cullberg, 2007 C. Rahm, J. Cullberg. Diagnostic stability over 3 years in a total group of first-episode psychosis patients. Nord. J. Psychiatry. 2007;61(3):189-193
  • Ramirez et al., 2010 N. Ramirez, B. Arranz, J. Salavert, E. Alvarez, I. Corripio, R.M. Duenas, V. Perez, L. San. Predictors of schizophrenia in patients with a first episode of psychosis. Psychiatry Res.. 2010;175(1–2):11-14
  • Reutfors et al., 2009 J. Reutfors, L. Brandt, E.G. Jonsson, A. Ekbom, P. Sparen, U. Osby. Risk factors for suicide in schizophrenia: findings from a Swedish population-based case–control study. Schizophr. Res.. 2009;108(1–3):231-237
  • Robinson et al., 2004 D.G. Robinson, M.G. Woerner, M. McMeniman, A. Mendelowitz, R.M. Bilder. Symptomatic and functional recovery from a first episode of schizophrenia or schizoaffective disorder. Am. J. Psychiatry. 2004;161(3):473-479
  • Robinson et al., 2010 J. Robinson, M. Harris, S. Cotton, A. Hughes, P. Conus, M. Lambert, B.G. Schimmelmann, P. McGorry. Sudden death among young people with first-episode psychosis: an 8–10 year follow-up study. Psychiatry Res.. 2010;177(3):305-308
  • Salvatore et al., 2009 P. Salvatore, R.J. Baldessarini, M. Tohen, H.M. Khalsa, J.P. Sanchez-Toledo, C.A. Zarate Jr., E. Vieta, C. Maggini. McLean-Harvard International First-Episode Project: two-year stability of DSM-IV diagnoses in 500 first-episode psychotic disorder patients. J. Clin.Psychiatry. 2009;70(4):458-466
  • Schimmelmann et al., 2005 B.G. Schimmelmann, P. Conus, J. Edwards, P.D. McGorry, M. Lambert. Diagnostic stability 18 months after treatment initiation for first-episode psychosis. J. Clin. Psychiatry. 2005;66(10):1239-1246
  • Schwartz et al., 2000 J.E. Schwartz, S. Fennig, M. Tanenberg-Karant, G. Carlson, T. Craig, N. Galambos, J. Lavelle, E.J. Bromet. Congruence of diagnoses 2 years after a first-admission diagnosis of psychosis. Arch. Gen. Psychiatry. 2000;57(6):593-600
  • Sellgren et al., 2011 C. Sellgren, M. Landen, P. Lichtenstein, C.M. Hultman, N. Langstrom. Validity of bipolar disorder hospital discharge diagnoses: file review and multiple register linkage in Sweden. Acta Psychiatr. Scand.. 2011;124(6):447-453
  • Statistics Sweden, 2009 Statistics Sweden. SCB-data för forskning 2009. Innehållsbeskrivning av olika register (SCB-data for research. Contents of different registers) (, 2009)
  • Subramaniam et al., 2007 M. Subramaniam, E. Pek, S. Verma, Y.H. Chan, S.A. Chong. Diagnostic stability 2 years after treatment initiation in the early psychosis intervention programme in Singapore. Aust. N. Z.J. Psychiatry. 2007;41(6):495-500
  • Tiihonen et al., 2011 J. Tiihonen, J. Haukka, M. Taylor, P.M. Haddad, M.X. Patel, P. Korhonen. A nationwide cohort study of oral and depot antipsychotics after first hospitalization for schizophrenia. Am. J. Psychiatry. 2011;168(6):603-609
  • Tiihonen et al., 2012 J. Tiihonen, J.T. Suokas, J.M. Suvisaari, J. Haukka, P. Korhonen. Polypharmacy with antipsychotics, antidepressants, or benzodiazepines and mortality in schizophrenia. Arch. Gen. Psychiatry. 2012;69(5):476-483
  • van Oel et al., 2002 C.J. van Oel, M.M. Sitskoorn, M.P. Cremer, R.S. Kahn. School performance as a premorbid marker for schizophrenia: a twin study. Schizophr. Bull.. 2002;28(3):401-414
  • Veen et al., 2004 N.D. Veen, J.P. Selten, D. Schols, W. Laan, H.W. Hoek, I. van der Tweel, R.S. Kahn. Diagnostic stability in a Dutch psychosis incidence cohort. Br. J. Psychiatry J. Ment. Sci.. 2004;185:460-464
  • Vinnerljung et al., 2006 B. Vinnerljung, A. Hjern, F. Lindblad. Suicide attempts and severe psychiatric morbidity among former child welfare clients—a national cohort study. J. Child Psychol. Psychiatry. 2006;47(7):723-733
  • Whitty et al., 2005 P. Whitty, M. Clarke, O. McTigue, S. Browne, M. Kamali, C. Larkin, E. O'Callaghan. Diagnostic stability four years after a first episode of psychosis. Psychiatr. Serv.. 2005;56(9):1084-1088


a Department of Public Health Sciences, Division of Social Medicine, Karolinska Institutet, Stockholm, Sweden

b Department of Statistics, Monitoring and Evaluation, National Board of Health and Welfare, Stockholm, Sweden

c Centre for Health Equity Studies (CHESS), Stockholm University/Karolinska Institutet, Stockholm, Sweden

d Department of Medicine Solna, Clinical Epidemiology Unit, Karolinska Institutet, Stockholm, Sweden

e Department of Neuroscience, Psychiatry, Uppsala University, Uppsala, Sweden

lowast Corresponding author at: Department of Neuroscience, Psychiatry, Uppsala University, SE-751 85 Uppsala, Sweden.