Schizophrenia Research 252 (2023) 110–117

Contents lists available at ScienceDirect

Schizophrenia Research
journal homepage: www.elsevier.com/locate/schres

Schizophrenia or schizoaffective disorder? A 50-year assessment of

A R T I C L E I N F O A B S T R A C T

Keywords: Background: Schizoaffective disorder (SAD) remains a controversial diagnosis in terms of necessity and reliability.
Schizophrenia Objectives: We assessed diagnostic patterns of SAD and schizophrenia (SZ) among hospitalized psychiatric pa­
Schizoaffective disorder tients over a fifty-year period.
Diagnostic stability
Method: Data from the Israeli National Psychiatric Registry on 16,341 adults diagnosed with SZ or SAD, hospi­
Reliability
DSM-5
talized at least twice in 1963–2017, were analyzed. Stability between most-frequent, first and last diagnosis, and
diagnostic-constancy (the same diagnosis in >75 % of a person's hospitalizations) were calculated. Three groups
were compared: People with both SAD and SZ diagnoses over the years (SZ-SAD), and people with only one of
these diagnoses (SZ-only; SAD-only). The incidence of SAD and SZ before and after DSM-5 publication was
compared.
Results: Reliability between last and first diagnosis was 60 % for SAD and 94 % for SZ. Agreement between first
and most-frequent diagnosis was 86 % for SAD and 92 % for SZ. Diagnostic shifts differ between persons with
SAD and with SZ. Diagnostic-constancy was observed for 50 % of SAD-only patients. In the SZ-SAD group, 9 %
had a constant SAD diagnosis. Compared to the other groups, the SZ-SAD group exhibited a higher substance use
prevalence, younger age at first-hospitalization, and more hospitalizations/person (p < 0.0001). The incidence of
a first-hospitalization SAD diagnosis increased by 2.2 % in the 4-years after vs. prior to DSM-5.
Conclusions: A SAD diagnosis is less stable than SZ. The incidence of a SAD diagnosis increased after DSM-5,
despite stricter diagnostic criteria. The SZ-SAD group exhibited the poorest outcomes. SAD may evolve over
time necessitating periodic re-evaluation.

1. Introduction Since its inception, SAD diagnosis has been controversial in terms of
its diagnostic necessity, validity and reliability (Maj et al., 2000;
Schizoaffective disorder was first recognized after a period in which Vollmer-Larsen et al., 2006; Abrams et al., 2008; Bromet et al., 2011;
Kraepelin's diagnostic dichotomy dominated, differentiating between Murru et al., 2012; Santelmann et al., 2015; Murru et al., 2016; Hung
psychotic mood disorders (‘manic-depression’) and schizophrenia (‘de­ et al., 2018). SAD was first introduced as a separate diagnostic entity in
mentia praecox’) (Kraepelin, 1883). However, clinical evidence had DSM-III (1980) although without operational diagnostic criteria
indicated that some people with a dominant psychotic disorder, also (American Psychiatric Association, 1980). In DSM-III-R, SAD appeared
experienced affective symptoms or affective episodes, leading Kasanin, as a separate diagnosis, requiring, in addition to affective symptoms
in 1933, to coin the term “acute schizoaffective psychosis” or schizo­ present for a “substantial” period of the illness, a period of at least two
affective disorder (SAD) (Kasanin, 1993). weeks with psychotic symptoms without affective symptoms (American

* Corresponding author at: Department of Psychiatry, Hadassah Medical Center, Jerusalem 9103401, Israel.
E-mail addresses: [email protected] (S. Florentin), [email protected] (I. Reuveni), [email protected] (P. Rosca), shlomo80@hadassah.
org.il (S.R. Zwi-Ran), [email protected] (Y. Neumark).

https://doi.org/10.1016/j.schres.2023.01.007
Received 1 May 2022; Received in revised form 30 December 2022; Accepted 3 January 2023
Available online 12 January 2023
0920-9964/© 2023 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
S. Florentin et al. Schizophrenia Research 252 (2023) 110–117

Psychiatric Association, 1994). This contrasts with ICD-10, which does Registry (NPCR) maintained by the Ministry of Health of Israel. The
not require a period of psychotic symptoms without affective symptoms NPCR is the official registry of all psychiatric admissions and discharges
(World Health Organization, 1994). Based on the DSM-IV (1994) countrywide since 1950 (Lichtenberg et al., 1999). The NPCR is based
criteria, the SAD diagnosis was expected to be relatively rare, yet it has on ICD-10 codes, although DSM-5 criteria are widely used in Israel by
become quite prevalent (Perälä et al., 2007; Boland et al., 2021). psychiatrists in clinical practice and for teaching. We identified all
Moreover, SAD's diagnostic reliability has been shown to be low psychiatric inpatients aged 18–65 years who were hospitalized during
compared to its differential diagnoses, including bipolar disorder, the years 2010–2015 with an ICD-10 diagnosis of either schizophrenia
schizophrenia and major depression (Santelmann et al., 2015; Fusar-Poli (SZ) or schizoaffective disorder (SAD) on their last discharge. For each
et al., 2016; Hung et al., 2018; Malhi and Bell, 2019a). Despite the person, all psychiatric diagnoses were recorded from all of his/her
controversy over the necessity of the use of schizoaffective diagnosis as a hospitalizations from 1963 through April 30, 2017. A substance use
stand-alone nosological category (Kempf et al., 2005; Maier, 2006; Lake disorder (SUD) diagnosis was made based on a recorded ICD-10 diag­
and Hurwitz, 2007; Marneros, 2007; Cheniaux et al., 2008; Heckers, nosis (excluding F17 - nicotine dependence) and/or a psychiatrist-
2009; Jäger et al., 2011; Kotov et al., 2013; Malhi, 2013; Malhi and Bell, documented indication of alcohol and/or drug abuse. Twenty-nine pa­
2019b), SAD diagnosis was retained as such in DSM-5 (2013) (American tients with anomalous numbers of hospitalizations (≥80) were excluded.
Psychiatric Association, 2013), in the absence of sufficient clinical in­ A total of 18,684 patients, with 168,377 hospitalizations, were identi­
formation supporting its re-conceptualization (Cosgrove and Suppes, fied, of whom 16,341 were hospitalized at least twice and were included
2013; Malaspina et al., 2013; Murru et al., 2016). In an attempt to in the analyses.
enhance its diagnostic consistency and reduce over-diagnosis, more
stringent criteria were introduced, requiring affective symptoms to be 2.2. Statistical analysis
present throughout most of the course of the disease.
Overall, the clinical and demographic characteristics of SAD on We compared the following characteristics among three diagnostic
average are in-between those of bipolar disorder and of schizophrenia groups (SZ-only, SAD-only, SZ-SAD): age at time of establishment of the
but more similar to schizophrenia (Pagel et al., 2013). The proportion of database (April 30, 2017), sex, ethnicity (Jew or Arab), SUD, average
women with SAD is greater than that of SZ (American Psychiatric As­ LOS per hospitalization and total number of hospitalizations. First-order
sociation, 2013; Pagel et al., 2013). SAD is often expected to have a descriptive statistics examined distributions; chi-square and t-tests were
better prognosis than schizophrenia, due to the affective symptoms that used to assess associations for continuous and categorical variables.
constitute a positive prognostic factor (American Psychiatric Associa­ Cohen's kappa coefficient was used to measure reliability between the
tion, 2013), as well as a better outcome (Harrow et al., 2000). first vs. last (i.e., most recent) diagnosis, and between the most-frequent
To date, most assessments regarding the reliability and temporal diagnosis vs. first/last diagnosis. ‘First diagnosis’ refers to the first of
stability of SAD diagnosis have investigated persons with SAD at two either SZ or SAD, regardless of whether the patient had previously
different time points (Maj et al., 2000; Bromet et al., 2011; Santelmann received a different diagnosis (e.g., acute psychotic disorder). We
et al., 2015). However, many psychiatric patients are hospitalized measured prospective consistency (the proportion of patients with first
multiple times during their lifetime, and receive different diagnoses over diagnosis of SZ or SAD who retained the same diagnosis at the last
time (Hung et al., 2018). Causes for changes in diagnosis may include discharge) and retrospective consistency (the proportion of patients
low inter-observer reliability, diagnostic shift (e.g., temporal changes in whose diagnosis at last discharge was consistent with the first SZ or SAD
diagnostic standards or diagnostic substitution), and/or temporal diagnosis). Also “diagnostic constancy” was assessed, defined as the
changes in the dominant clinical characteristics of the disease (Murru presence of a SAD diagnosis or a SZ diagnosis in >75 % of a person's
et al., 2012). Comparing the diagnosis at two time-points may, there­ diagnoses, a measure that has been used in previous studies (Baca-
fore, not adequately reflect reliability or temporal stability of SAD Garcia et al., 2007; Cegla-Schvartzman et al., 2021).
diagnosis. Logistic regression modeling was performed to predict the stability
The current study set three objectives: 1. Describe differences in between first and most-frequent diagnoses adjusting for age, sex, num­
demographic and hospitalization characteristics across three diagnostic ber of hospitalizations, LOS, SUD, age at first hospitalization, and last
groups: people diagnosed with SZ who were never diagnosed with SAD diagnosis (SAD or SZ). Statistical significance level was set at p < 0.05.
(SZ-only), people diagnosed with SAD who were never diagnosed with The data were analyzed using IBM® SPSS® Statistics 24.0.
SZ (SAD-only), people with both SAD and SZ diagnoses (SZ-SAD). Ac­ The study was conducted in accordance with the Declaration of
cording to our hypothesis, the SZ-SAD group's demographic and clinical Helsinki and was approved by the IRB of the Israel Ministry of Health.
characteristics would be in-between the SZ-only and SAD-only groups. All datasets were anonymized prior to being released to the researchers.
For example, we anticipate that the proportion of men, the prevalence of
SUD, age at first hospitalization, number of hospitalizations and length 3. Results
of stay (LOS) in the SZ-SAD group will be lower than in SZ-only group
and greater than in the SAD-only group. 2. Compare, for the first time, Among the 16,341 people who had at least two recorded hospitali­
the stability of SAD diagnosis between the most-frequent diagnosis and zations and were diagnosed with SZ and/or SAD, 64.6 % were in the SZ-
the first and last diagnosis for each person, and determine factors related only group, 11.5 % were in the SAD-only group, and 23.9 % in the mixed
to diagnostic instability. We hypothesized greater instability for the SAD SZ-SAD group.
diagnosis than the SZ diagnosis. As the number of hospitalizations in­ Among persons first diagnosed with SAD, 38 % subsequently
creases, people first diagnosed with SAD will have a greater probability received a diagnosis of SZ; 21 % of persons first diagnosed with SZ,
of being subsequently diagnosed with SZ, than those with an initial subsequently received an SAD diagnosis. Among those hospitalized five
diagnosis of SZ being subsequently diagnosed with SAD. 3. Assess or more times, 51 % of those with a first diagnosis of SAD subsequently
whether the incidence of SAD declined following changes in the SAD received a SZ diagnosis, and 28 % of those first diagnosed with SZ were
diagnostic criteria in DSM-5. subsequently diagnosed with SAD.
As seen in Table 1, compared with the SZ-only and SAD-only groups,
2. Methods the SZ-SAD group was older, on average, while their mean age at first
hospitalization was younger. The lifetime prevalence of SUD was higher
2.1. Study design and participants in the SZ-SAD group (42 %) than in the other two groups. The average
number of hospitalizations per person was also higher in the SZ-SAD
Data for this study were extracted from the National Psychiatric Case group – in total and across three age groups, as was the mean number

111
S. Florentin et al. Schizophrenia Research 252 (2023) 110–117

Table 1 87.9 % for 10 or more hospitalizations. The agreement for SAD de­
Demographic and hospitalization-related characteristics by diagnostic group, creases more steeply - from 94.4 % for patients with 2–5 hospitaliza­
1963–2016, Israel. tions, 82.6 % for 6–9, and 74.6 % for 10 or more hospitalizations.
SZ-only SAD- SZ-SAD F-test/ Eta Number of hospitalizations was also seen to be associated with the
(N = only (N (N = χ2 squared relative size of the diagnostic groups. Among individuals hospitalized
10,551) = 1882) 3908) 2–5 times (n = 7107), 10.9 % were SZ-SAD, compared with 24.7 %
43.3 42.4 46.7 among those hospitalized 6–9 times (n = 3542) and 39.7 % among those
Mean age (SD) 146.62 0.02
(11.4) (11.6) (11.0) with 10 or more hospitalizations (n = 5639) (p < 0.0001). The pro­
Male 72.3 % 48.0 % 57.9 % 572.86 0.03
portion of the SAD-only group, on the other hand, declined from 16.2 %
Population group
Jews (n = 15,143) 85.0 % 86.5 % 89.1 % 38.37 0.00 among those with 2–5 hospitalizations to 6.0 % among those hospital­
Arabs (n = 2556) 15.0 % 13.5 % 10.9 % ized 10 or more times (n = 339).
Lifetime SUD 37.4 % 34.5 % 42.3 % 42.16 0.01 Little change was noted in the proportion of patients diagnosed with
Mean age at first SAD at their first hospitalization from the pre-DSM-5 period (July 1,
26.4 29.6 24.8
hospitalization 171.70 0.03
(SD)
(9.2) (10.6) (8.3) 2009–May 31, 2013) to the post-DSM-5 period (June 1, 2013–April 30,
Mean LOS (SD)/ 2017). Of those first hospitalized in the former period, 8.5 % were
113.7 50.1 72.6
Person (number of 91.73 0.01 diagnosed with SAD at their first hospitalization, compared with 10.7 %
(273.8) (63.6) (101.0)
days) of those first hospitalized in the latter period (χ2 = 5.0, p = 0.026).
Mean number (SD) of
6.5 15.1 Whereas the proportion of those diagnosed with SZ at first hospitaliza­
hospitalizations/ 8.8 (9.0) 730.21 0.11
person
(6.6) (12.4) tion did not change between the periods - 41.8 % in the 2009–2013
4.2 period and 42.2 % in the 2013–2017 period (χ2 = 0.1, p = 0.795).
Age < 30 years 5.1 (4.5) 8.7 (7.2) 0.10
(3.1) As seen in Fig. 1, among people hospitalized ≥5 times, the proba­
6.0 12.6 bility of additional diagnoses is much greater in the SAD-only group and
Age 30–44 years 7.9 (7.7) 0.08
(5.9) (10.5)
the SZ-SAD-group than in the SZ-group. Among those never diagnosed
10.9 7.9 17.5
Age ≥ 45 years 0.10 with SAD (SZ-only group; Fig. 1A), 51 % received a SZ diagnosis at first
(10.6) (8.0) (13.5)
Mean number (SD) of hospitalization and this increased to 91 % by the fourth hospitalization.
2.0
different 1.6 (0.8)
(0.9)
2.7 (1.2) 1880.76 0.21 Among those in the SAD-only group (i.e., never diagnosed with SZ;
diagnoses/person
Fig. 1B), 19 % were diagnosed with SAD at first hospitalization and this
Note: p < 0.0001 for all cross-group comparisons. increased monotonically to 73 % at the fifth hospitalization. Bipolar
Acronyms: LOS, length of stay; SUD, substance use disorder. disorder was the most common diagnosis given in the first hospitaliza­
tion in this group (25 %) and accounted for 16 % at the fifth hospitali­
of diagnoses. The proportion of males was highest in the SZ-only group. zation. About one-third (35 %) of the SZ-SAD group were diagnosed
Overall, a SAD diagnosis was the first diagnosis (between SZ and with SZ at first hospitalization and this increased to about 50 % by the
SAD) for 18.5 % of patients, and the last diagnosis for 23.7 %. Among second hospitalization (Fig. 1C); 10 % of the SZ-SAD group were diag­
those with an initial diagnosis of SAD, 23.3 % were last diagnosed with nosed with SAD at first hospitalization, and 34 % at the fifth
SZ, compared with 11.6 % of those who were first diagnosed with SZ and hospitalization.
were last diagnosed with SAD. Half (50.6 %) of those in the SZ-SAD Finally, we divided the population into those who were first diag­
group had a SAD diagnosis as their last diagnosis. nosed with SZ and those who were first diagnosed with SAD (regardless
Diagnostic constancy (a diagnosis obtained in >75 % of the person's of a subsequent SAD or SZ diagnosis, or a previous other diagnosis). As
diagnoses) was noted for 72 % of patients in the SZ-only group and 50 % seen in Fig. 2, among those hospitalized ≥5 times and first diagnosed
of those in the SAD-only group. In the SZ-SAD group, 23 % were cate­ with SZ, just over half (52 %) received a SZ diagnosis at first hospitali­
gorized as having a constant diagnosis of SZ, and 9 % were categorized zation and about 80 % by the third hospitalization. Among those first
as having a constant SAD diagnosis. diagnosed with SAD, 23 % were diagnosed with SZ at the fifth hospi­
The prospective and retrospective consistency between first and last talization, and the proportion of those with a SAD diagnosis increased
diagnosis of SZ was 88.4 % and 94.4 %, respectively, and 76.7 % and from 29 % at first hospitalization to 57 % at the fifth hospitalization
59.9 % between first and last diagnosis of SAD (kappa between first and (Fig. 2).
last diagnosis = 0.59). The agreement (prospective consistency) be­
tween the first and the most-frequent diagnosis was 91.5 % for SZ (i.e., 4. Discussion
for 91.5 % of those who had a first diagnosis of SZ, the most-frequent
diagnosis was also SZ), and 85.7 % for SAD. Overall, kappa between Data from the National Psychiatric Case Registry of Israel, among
first and most-frequent diagnosis was 0.71. High agreement was also >16,000 people hospitalized in the years 1963–2016 with a diagnosis of
noted between last diagnosis of SZ and most-frequent diagnosis (96.2 SZ and/or SAD shows that SZ was 2.5 times more frequently diagnosed
%), while the agreement was lower for a last diagnosis of SAD (83.8 %). than SAD, similar to previous findings (Perälä et al., 2007; American
Kappa between first and most-frequent diagnosis = 0.81. Psychiatric Association, 2013; Boland et al., 2021). Overall, the SZ-only
Results of logistic-regression modeling revealed that disagreement group was nearly 6 times larger than the SAD-only group among those
between the first and the most-frequent diagnosis was independently with a greater number of hospitalizations. In line with our hypothesis,
associated with older age (OR = 1.04, 95 % CI 1.03–1.04, p < 0.001), the relative size of the SAD-only group diminished with increasing
female sex (OR = 1.18, 95 % CI 1.06–1.31, p = 0.002), increasing number of hospitalizations.
number of hospitalizations (OR = 1.04, 95 % CI 1.03–1.04, p < 0.001)
lifetime SUD (OR = 1.35, 95 % CI 1.13–1.62, p = 0.001), younger age at 4.1. Diagnostic stability
first hospitalization (OR = 1.06, 95 % CI 1.05–1.06, p < 0.001), and last
diagnosis of SAD (OR = 1.34, 95 % CI 1.20–1.50, p < 0.0001). No as­ Nearly 40 % of people first diagnosed with SAD subsequently
sociation was found with LOS. received a diagnosis of SZ. The SZ-SAD group was characterized by a
As the number of hospitalizations increases, the agreement between higher degree of diagnostic instability, which is not surprising, by mere
a first diagnosis of SZ and the most-frequent diagnosis decreases from fact that they are in a mixed diagnostic group. However, the average
94.5 % for patients with 2–5 hospitalizations, to 91.7 % for 6–9, and number of diagnoses per person was less different between the SZ-SAD
and SAD-only groups than we expected. The prospective consistency

112
S. Florentin et al. Schizophrenia Research 252 (2023) 110–117

Fig. 1. Diagnosis distribution for patients hospitalized ≥5 times (n = 10,538) by diagnosis group: a. SZ-only, b. SAD-only, c. SZ-SAD
NOTE: Diagnoses: Schizophrenia (SZ); Schizoaffective disorder (SAD); Bipolar disorder (BP); Major Depressive disorder (MDD). Unspecified psychosis includes also
other nonorganic psychotic disorders.

was lower for SAD than for SZ. As we hypothesized, and as reported by diagnosed with SAD and who subsequently received a diagnosis of
others, the consistency of a diagnosis of SAD between two time points schizophrenia was nearly two-fold higher than patients with an initial SZ
(most often the follow-up was after the first psychotic episode) ranged diagnosis and who were subsequently diagnosed with SAD. One of the
from 55 % to 76 %, and 68 %–100 % for SZ (Schwartz et al., 2000; Hung factors that may contribute to the diagnostic instability is substance use,
et al., 2018; Wood et al., 2021). Differential follow-up time may explain which was found to be significantly more prevalent in the SZ-SAD group
the higher degree of stability reported by Haahr et al. (2008) who re- than in the other diagnostic groups. Substance use may cause psychosis
assessed patients after one or two years, compared with the lower sta­ and affective symptoms to appear or to worsen (Campbell et al., 2021),
bility reported by Lenz et al. (1991), in which re-assessment took place hampering the ability of making a reliable diagnosis. However, this is
after seven years. This raises the question of the clinically appropriate likely to be only a partial explanation, as the other groups also include
length of time during which repeated clinical assessments should be substance-using individuals.
made to confirm diagnosis and during which diagnostic stability should It is also possible, as previously suggested, that SAD is a milder
be measured. version of SZ (Harrow et al., 2000; Pagel et al., 2013), that tends to
develop more slowly into SZ, with fewer affective features, more nega­
tive symptoms and decreased functioning over the years. Therefore, it
4.2. Understanding the diagnostic instability of the SZ-SAD group
may take time for clinicians to identify the gradual development to SZ,
which is associated with worsening prognosis and functional decline.
In the SZ-SAD group, the proportion of patients who were initially

113
S. Florentin et al. Schizophrenia Research 252 (2023) 110–117

Fig. 2. Diagnoses by hospitalizations for patients hospitalized ≥5 times whose first diagnosis between SZ and SAD was: a. SZ; b. SAD (n = 10,538)
NOTE: Diagnoses: Schizophrenia (SZ); Schizoaffective disorder (SAD); Bipolar disorder (BP); Major Depressive disorder (MDD). Unspecified psychosis includes also
other nonorganic psychotic disorders.

Since SAD is thought to have a better prognosis than SZ (American degree of adherence. Alternatively, people in the SZ-SAD group may
Psychiatric Association, 2013; Mondragón-Maya et al., 2017; Boland have been misdiagnosed at an early hospitalization and should have
et al., 2021), these individuals may receive insufficient therapy and been diagnosed with SZ from the outset (Pai and Vella, 2019; Ayano
rehabilitation in the early stages, and may exhibit more negative out­ et al., 2021). It is also possible that the SZ-SAD group constitutes the
comes as their disorder evolves. “true” population of people with SAD, whereas people in the SAD-only
In addition, we noted that persons in the SZ-SAD group were younger group, which diminishes in size as the number of hospitalizations in­
at first hospitalization, on average, than the other groups. Also, contrary creases, are more likely to have bipolar disorder.
to our prior hypothesis, we found that the mixed SZ-SAD group was According to the distribution of diagnoses across hospitalizations
characterized by more hospitalizations - about twice as many as the SZ- (Fig. 1), diagnostic shifts differ between persons with SAD and with SZ.
only group and three times more than the SA-only group. A younger age In the SAD-only group, the most-frequent diagnosis at first hospitaliza­
of onset, multiple hospitalizations and substance use are negative tion is bipolar disorder (BP), and as the number of hospitalizations in­
prognostic factors (Boland et al., 2021), and were found to be statisti­ creases, the prevalence of BP decreases and SAD increases. On the other
cally significant, albeit not strong predictors of diagnostic instability in hand, in the SZ-only group BP barely appeared, and the rates of SZ
the regression model. This raises the question whether the SZ-SAD group diagnosis in each hospitalization (53–91 %) were much higher than the
is a distinct diagnostic subgroup with a worse prognosis than that of the rates of SAD in the SAD-only group (19–73 %). These trends may
SZ-only group. It is possible that the greater number of hospitalizations highlight the need to recognize schizoaffective disorder as a time-
in the SZ-SAD group indicates a lower adherence to treatment. Also, varying disorder that can begin as BP, progress to SA, and then to SZ.
SUD, which was found to be more prevalent in the SZ-SAD group, may The process of development and transition from BP to SAD to SZ is
increase the frequency of exacerbations and in turn, lessen treatment also consistent with the findings of the Genome-Wide Association
adherence. Unfortunately, our data does not allow us to assess the Studies (GWAS) which show a significant genetic overlap between SZ

114
S. Florentin et al. Schizophrenia Research 252 (2023) 110–117

and BP, and therefore do not support the dichotomous separation be­ Psychiatric Case Registry of Israel that captures virtually all psychiatric
tween diagnoses (Schizophrenia Psychiatric Genome-Wide Association hospitalizations in the country over a fifty-year period. Our analyses
Study (GWAS) Consortium, 2011; Visscher et al., 2012). addressed, we believe for the first time, diagnostic stability also with
reference to the most-frequent diagnosis, and compared three diagnostic
4.3. Changes in the incidence of SAD since DSM5 groups (SZ-only, SAD-only, SZ-SAD) to deepen and refine our under­
standing about SAD. The main limitation of this study is that the study
Due to the change in the definition of SAD in DSM-5, especially in population was limited to people whose last diagnosis was either SZ or
criterion C (“symptoms that meet criteria for a major mood episode are SAD, and excluded those whose last diagnosis was BP. Although most
present for the majority of the total duration of the active and residual studies show that schizophrenia is more similar to SAD than BP, there is
portions of the illness”), we anticipated to observe a decrease in the a diagnostic sequence between BP, SAD, and SZ (Fusar-Poli et al., 2016;
prevalence of SAD and a concomitant increase in the SZ/SAD ratio Santelmann et al., 2016b; Baethge, 2020). In addition, the comorbidity
following the publication of DSM-5 in 2013. Yet, little change was noted with SUD, which was more prevalent in the SZ-SAD group, may create
from the pre-DSM-5 period. It is possible that four years (2013–2017) some ambiguity in terms of diagnosis, accounting for more diagnostic
since the launch of DSM-5 are insufficient for a new diagnosis to be fully changes. Similarly, due to the study's retrolective design, it was only
adopted into clinical practice. However, SAD's diagnostic criteria (at possible to track the patients' diagnoses according to number of hospi­
least two weeks of psychosis without affective symptoms; mood episodes talizations rather than at defined time points common to all patients.
last at least half the time of the disorder) that were designed to prevent Another shortcoming is the absence from the database of some clinical
overuse of the diagnosis, are not always easy or possible to fulfil. It is and prognostic factors (e.g., severity of the psychotic disorder, level of
therefore not surprising that the changes in the diagnostic criteria did functioning, negative symptoms, motivation for rehabilitation, adher­
not lead to a reduction in the prevalence of SAD. This might explain the ence to treatment, and level of family support).
tendency of psychiatrists to diagnose SAD according to prototypes and a
general clinical picture (Santelmann et al., 2016a), despite formal DSM- 5. Conclusions
5 criteria.
In light of the ongoing controversy regarding the diagnostic utility of
4.4. Clinical and demographic characteristics of the study population SAD, the current study shows that diagnostic shifts, as well as de­
mographic and hospitalization characteristics differ between persons
SAD is conceptualized as a hybrid of both SZ and BP, and thus its with SAD and with SZ. Yet, the reliability of SAD is relatively low, and
demographic characteristics tend to be in-between the two disorders, many people with SAD subsequently received a SZ diagnosis. The inci­
and usually more similar to SZ (American Psychiatric Association, 2013; dence of SAD has not decreased since the publication of DSM-5, despite
Pagel et al., 2013). We therefore expected the characteristics of the stricter diagnosis criteria.
mixed SZ-SAD group to be between SZ and SAD, however this was true Further research is needed to illuminate the differences between the
for some characteristics (e.g., LOS, gender), but not others (e.g., age at diagnostic groups particularly the mixed SZ-SAD group, and whether
first hospitalization). through early detection and appropriate and intensive treatment in the
Overall, males accounted for two-thirds of the study population. early stages of the disorder it will be possible to halt progression at the
They comprised 72 % of the SZ-only group and 58 % of the SZ-SAD stage of BP or SAD and reduce the likelihood of transition to schizo­
group, while in the SAD-only group the proportion of men and women phrenia. Persons with a SAD-only diagnosis are, overall, more clinically
was similar. The higher proportion of males with SZ is consistent with stable and exhibit better prognosis, and may not require as much
findings from another Israeli study (Lurie and Fleischman, 2018), yet intensive follow-up.
higher than other studies which report no difference between women A diagnosis based on domains/dimensions (Peralta et al., 2020;
and men (Abel et al., 2010; Boland et al., 2021; Sommer et al., 2021). Cuthbert and Morris, 2021) may be more appropriate for psychotic
The greater representation of men in our study may be due, in part, to disorders, in particular for the evolving nature of SAD. Whether there
the inclusion of hospitalized patients only, and does not represent people will be a formal transition to diagnosis by domains in the future, or we
treated in the community. As the prognosis of women with schizo­ will continue to rely on dichotomous diagnoses as presented in DSM
phrenia is known to be better, with lower rates of SUD (Florentin et al., (American Psychiatric Association, 2022), we should keep in mind that
2019, 2021), violence and completed suicide (Boland et al., 2021), there at least for most SAD patients, the clinical course of SAD changes and
is likely to be a greater proportion of women treated in the community develops over time and may eventually transform to SZ. It may be worth
without being hospitalized. Results of the logistic regression model show considering defining SAD in current diagnostic systems as a provisional
that diagnostic stability was lower among women. This may be due to diagnosis, prompting recurrent clinical and diagnostic assessments.
the fact that women are more prone to develop affective symptoms
(Boland et al., 2021) and thus might be misdiagnosed with major CRediT authorship contribution statement
depression with psychotic features or bipolar disorder instead of a psy­
chotic disorder. Also, women with SZ have more self-destructive and All the authors designed the study. SF wrote the first draft of the
suicidal behaviors and are misdiagnosed in some cases with borderline manuscript. SF and YN designed the methods for data analysis. SF per­
personality disorder (Sommer et al., 2021). formed the statistical analysis. YN, IR, PR and SR proofed and critically
We noted a slightly higher proportion of Arabs in the SZ-only group revised the manuscript. All authors contributed to the article and
than in the SAD-only or SZ-SAD groups. This might be due to clinicians, approved the submitted version.
the majority of whom are Jewish, being less able to perceive affective
features among the Arab minority patients (Youngmann et al., 1999;
Funding
Katz et al., 2012; Boland et al., 2021). As reported for Afro-Americans
and Hispanic minorities in the United States (Boland et al., 2021), this
This research received no grant from any funding agency, commer­
diagnostic challenge may be culturally influenced (Youngmann et al.,
cial or not-for-profit sectors.
1999).

4.5. Strengths and limitations Declaration of competing interest

This longitudinal study of >16,000 individuals utilized the National The authors declare that they have no competing interests.

115
S. Florentin et al. Schizophrenia Research 252 (2023) 110–117

Data availability Haahr, U., Friis, S., Larsen, T.K., Melle, I., Johannessen, J.O., Opjordsmoen, S.,
Simonsen, E., Rund, B.R., Vaglum, P., McGlashan, T., 2008. First-episode psychosis:
diagnostic stability over one and two years. Psychopathology 41 (5), 322–329.
The datasets analyzed in this study were obtained from the Ministry https://doi.org/10.1159/000146070.
of Health of Israel. Access to the data is highly restricted due to the Harrow, M., Grossman, L.S., Herbener, E.S., Davies, E.W., 2000. Ten-year outcome:
sensitive nature of psychiatric patient data. Requests to access these patients with schizoaffective disorders, schizophrenia, affective disorders and mood-
incongruent psychotic symptoms. Br. J. Psychiatry 177, 421–426. https://doi.org/
datasets should be directed to Mr. Reuven Eliyahu (reuven.eliahu@moh. 10.1192/bjp.177.5.421.
gov.il) Data Security Office, Ministry of Health. Heckers, S., 2009. Is schizoaffective disorder a useful diagnosis? Curr. Psychiatry Rep. 11
(4), 332–337. https://doi.org/10.1007/s11920-009-0048-3.
Hung, Y.N., Yang, S.Y., Kuo, C.J., Lin, S.K., 2018. Diagnostic consistency and
Acknowledgments interchangeability of schizophrenic disorders and bipolar disorders: a 7-year follow-
up study [published correction appears in Psychiatry Clin Neurosci. 72(8):627.
The authors thank Ms. Tali Bdolah-Abram for her invaluable statis­ Psychiatry Clin. Neurosci. 72 (3), 180–188. https://doi.org/10.1111/pcn.12629.
Jäger, M., Haack, S., Becker, T., Frasch, K., 2011. Schizoaffective disorder–an ongoing
tical advice. challenge for psychiatric nosology. Eur. Psychiatry. 26 (3), 159–165. https://doi.
org/10.1016/j.eurpsy.2010.03.010.
References Kasanin, J., 1993. The acute schizoaffective psychosis. Am. J. Psychiatry 90, 97–126.
Katz, G., Grunhaus, L., Deeb, S., Shufman, E., Bar-Hamburger, R., Durst, R., 2012.
A comparative study of Arab and jewish patients admitted for psychiatric
Abel, K.M., Drake, R., Goldstein, J.M., 2010. Sex differences in schizophrenia. Int. Rev.
hospitalization in Jerusalem: the demographic, psychopathologic aspects, and the
Psychiatry 22 (5), 417–428. https://doi.org/10.3109/09540261.2010.515205.
drug abuse comorbidity. Compr. Psychiatry 53 (6), 850–853. https://doi.org/
Abrams, D.J., Rojas, D.C., Arciniegas, D.B., 2008. Is schizoaffective disorder a distinct
10.1016/j.comppsych.2011.11.005.
categorical diagnosis? A critical review of the literature. Neuropsychiatr. Dis. Treat.
Kempf, L., Hussain, N., Potash, J.B., 2005. Mood disorder with psychotic features,
4 (6), 1089–1109. https://doi.org/10.2147/ndt.s4120.
schizoaffective disorder, and schizophrenia with mood features: trouble at the
American Psychiatric Association, 1980. Diagnostic and Statistical Manual of Mental
borders. Int. Rev. Psychiatry 17 (1), 9–19. https://doi.org/10.1080/
Disorders: DSM-3, 3th edition.
09540260500064959.
American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental
Kotov, R., Leong, S.H., Mojtabai, R., Erlanger, A.C., Fochtmann, L.J., Constantino, E.,
Disorders: DSM-4, 4th edition.
Carlson, G.A., Bromet, E.J., 2013. Boundaries of schizoaffective disorder: revisiting
American Psychiatric Association, 2013. Diagnostic and Statistical Manual of Mental
kraepelin. JAMA Psychiatry 70 (12), 1276–1286. https://doi.org/10.1001/
Disorders: DSM-5, 5th edition.
jamapsychiatry.2013.2350.
American Psychiatric Association, 2022. Diagnostic and Statistical Manual of Mental
Kraepelin, E., 1883. Psychiatrie. Ein Lehrbuch für Studirende und Aerzte.
Disorders: DSM-5-TR, 5th edition. TR.
Lake, C.R., Hurwitz, N., 2007. Schizoaffective disorder merges schizophrenia and bipolar
Ayano, G., Demelash, S., Yohannes, Z., Haile, K., Tulu, M., Assefa, D., Tesfaye, A.,
disorders as one disease–there is no schizoaffective disorder. Curr. Opin. Psychiatry.
Haile, K., Solomon, M., Chaka, A., Tsegay, L., 2021. Misdiagnosis, detection rate, and
20 (4), 365–379. https://doi.org/10.1097/YCO.0b013e3281a305ab.
associated factors of severe psychiatric disorders in specialized psychiatry centers in
Lenz, G., Simhandl, C., Thau, K., Berner, P., Gabriel, E., 1991. Temporal stability of
Ethiopia. Ann. General Psychiatry 20 (1), 10. https://doi.org/10.1186/s12991-021-
diagnostic criteria for functional psychoses. Results from the Vienna follow-up study.
00333-7.
Psychopathology 24 (5), 328–335. https://doi.org/10.1159/000284733.
Baca-Garcia, E., Perez-Rodriguez, M.M., Basurte-Villamor, I., López-Castromán, J.,
Lichtenberg, P., Kaplan, Z., Grinshpoon, A., Feldman, D., Nahon, D., 1999. The goals and
Fernandez del Moral, A.L., Jimenez-Arriero, M.A., Gronzalez de Rivera, J.L., Saiz-
limitations of Israel's psychiatric case register. Psychiatr. Serv. 50, 1043–1048.
Ruiz, J., Leiva-Murillo, J.M., de Prado-Cumplido, M., Santiago-Mozos, R., Artés-
https://doi.org/10.1176/ps.50.8.1043.
Rodríguez, A., Oquendo, M.A., de Leon, J., 2007. Diagnostic stability and evolution
Lurie, I., Fleischman, A., 2018. Psychiatric hospitalizations among the Arab population in
of bipolar disorder in clinical practice: a prospective cohort study. Acta Psychiatr.
Israel: a historic cohort study. Retrieved from Isr. J. Psychiatry 55 (3), 12–22.
Scand. 115 (6), 473–480. https://doi.org/10.1111/j.1600-0447.2006.00984.x.
https://www.proquest.com/scholarly-journals/psychiatric-hospitalizations-amon
Baethge, C., 2020. How reasonable is it to uphold the diagnostic concept of
g-arab/docview/2248965886/se-2?accountid=14546.
schizoaffective disorder? Bipolar Disord. 22 (4), 328–330. https://doi.org/10.1111/
Maier, W., 2006. Do schizoaffective disorders exist at all? Acta Psychiatr. Scand. 113 (5),
bdi.12920.
369–371. https://doi.org/10.1111/j.1600-0447.2006.00763.x.
Boland, R., Verduin, M., Ruiz, P., 2021. Kaplan & Sadock's Synopsis of Psychiatry:
Maj, M., Pirozzi, R., Formicola, A.M., Bartoli, L., Bucci, P., 2000. Reliability and validity
Behavioral Sciences/Clinical Psychiatry, 12th edition. Wolters Kluwer, Philadelphia.
of the DSM-IV diagnostic category of schizoaffective disorder: preliminary data.
Bromet, E.J., Kotov, R., Fochtmann, L.J., Carlson, G.A., Tanenberg-Karant, M.,
J. Affect. Disord. 57 (1–3), 95–98. https://doi.org/10.1016/s0165-0327(99)00059-
Ruggero, C., Chang, S.W., 2011. Diagnostic shifts during the decade following first
2.
admission for psychosis. Am. J. Psychiatry 168 (11), 1186–1194. https://doi.org/
Malaspina, D., Owen, M.J., Heckers, S., Tandon, R., Bustillo, J., Schultz, S., Barch, D.M.,
10.1176/appi.ajp.2011.11010048.
Gaebel, W., Gur, R.E., Tsuang, M., Van Os, J., Carpenter, W., 2013. Schizoaffective
Campbell, E.C., Caroff, S.N., Mann, S.C., 2021. Co-occurring schizophrenia and substance
disorder in the DSM-5. Schizophr. Res. 150 (1), 21–25. https://doi.org/10.1016/j.
use disorder: Epidemiology, pathogenesis, clinical manifestations, course,
schres.2013.04.026.
assessment and diagnosis. UptoDate. Available from. https://www.uptodate.com/co
Malhi, G.S., 2013. Making up schizoaffective disorder: cosmetic changes to a sad
ntents/co-occurring-schizophrenia-and-substance-use-disorder-epidemiology-patho
creation? Aust. N. Z. J. Psychiatry 47 (10), 891–894. https://doi.org/10.1177/
genesis-clinical-manifestations-course-assessment-and-diagnosis.
0004867413505522.
Cegla-Schvartzman, F., Ovejero, S., López-Castroma, J., Palomar-Ciria, N., Migoya-
Malhi, G.S., Bell, E., 2019. Fake views: schizoaffective disorder is not 'SAD', just bad.
Borja, M., Bello, H., Martínez-Alés, G., Baca-García, E., 2021. Diagnostic stability in
Aust. N. Z. J. Psychiatry 53 (5), 481–484. https://doi.org/10.1177/
bipolar disorder: a follow-up study in 130,000 patient-years. J. Clin. Psychiatry 82
0004867419846396.
(6), 20m13764. https://doi.org/10.4088/JCP.20m13764.
Malhi, G.S., Bell, E., 2019. Schizoaffective disorder: a "dysfunctional" proposal? Bipolar
Cheniaux, E., Landeira-Fernandez, J., Lessa Telles, L., Lessa, J.L., Dias, A., Duncan, T.,
Disord. 21 (4), 298–301. https://doi.org/10.1111/bdi.12806.
Versiani, M., 2008. Does schizoaffective disorder really exist? A systematic review of
Marneros, A., 2007. Do schizoaffective disorders exist at all? Acta Psychiatr. Scand. 115
the studies that compared schizoaffective disorder with schizophrenia or mood
(2), 162–163. https://doi.org/10.1111/j.1600-0447.2006.00945.x.
disorders. J. Affect. Disord. 106 (3), 209–217. https://doi.org/10.1016/j.
Mondragón-Maya, A., Ramos-Mastache, D., Yáñez-Téllez, G., Saracco-Álvarez, R., 2017.
jad.2007.07.009.
SA32. Cognitive functions in schizophrenia and schizoaffective disorder patients: a
Cosgrove, V.E., Suppes, T., 2013. Informing DSM-5: biological boundaries between
comparative study. Schizophr. Bull. 43 (Suppl. 1), S124–S125. https://doi.org/
bipolar I disorder, schizoaffective disorder, and schizophrenia. BMC Med. 11, 127.
10.1093/schbul/sbx023.031.
https://doi.org/10.1186/1741-7015-11-127.
Murru, A., Pacchiarotti, I., Nivoli, A.M., Colom, F., Vieta, E., 2012. Is schizoaffective
Cuthbert, B.N., Morris, S.E., 2021. Evolving concepts of the schizophrenia spectrum: a
disorder still a neglected condition in the scientific literature? Psychother.
research domain criteria perspective. Front. Psychiatry 12, 641319. https://doi.org/
Psychosom. 81, 389–390.
10.3389/fpsyt.2021.641319.
Murru, A., Manchia, M., Tusconi, M., Carpiniello, B., Pacchiarotti, I., Colom, F., Vieta, E.,
Florentin, S., Rosca, P., Raskin, S., Bdolah-Abram, T., Neumark, Y., 2019. Psychiatric
2016. Diagnostic reliability in schizoaffective disorder. Bipolar Disord. 18 (1),
hospitalizations of chronic psychotic disorder patients with and without dual
78–80. https://doi.org/10.1111/bdi.12366.
diagnosis, Israel, 1963–2016. J. Dual. Diagn. 15 (3), 130–139. https://doi.org/
Pagel, T., Baldessarini, R.J., Franklin, J., Baethge, C., 2013. Characteristics of patients
10.1080/15504263.2019.1609149.
diagnosed with schizoaffective disorder compared with schizophrenia and bipolar
Florentin, S., Neumark, Y., Raskin, S., Bdolah-Abram, T., Rosca, P., 2021. Differential
disorder. Bipolar Disord. 15 (3), 229–239. https://doi.org/10.1111/bdi.12057.
effect of community rehabilitation reform on hospitalizations of patients with
Pai, N., Vella, S.L., 2019. The misdiagnosis of schizoaffective disorder. Aust. N. Z. J.
chronic psychotic disorders with and without substance use disorder, Israel,
Psychiatry 53 (11), 1124–1125. https://doi.org/10.1177/0004867419872246.
1991–2016. Admin. Pol. Ment. Health 48 (2), 354–362. https://doi.org/10.1007/
Perälä, J., Suvisaari, J., Saarni, S.I., Kuoppasalmi, K., Isometsä, E., Pirkola, S.,
s10488-020-01077-4.
Partonen, T., Tuulio-Henriksson, A., Hintikka, J., Kieseppä, T., Härkänen, T.,
Fusar-Poli, P., Cappucciati, M., Rutigliano, G., Heslin, M., Stahl, D., Brittenden, Z.,
Koskinen, S., Lönnqvist, J., 2007. Lifetime prevalence of psychotic and bipolar I
Caverzasi, E., McGuire, P., Carpenter, W.T., 2016. Diagnostic stability of ICD/DSM
disorders in a general population. Arch. Gen. Psychiatry 64 (1), 19–28. https://doi.
first episode psychosis diagnoses: meta-analysis. Schizophr. Bull. 42 (6), 1395–1406.
org/10.1001/archpsyc.64.1.19.
https://doi.org/10.1093/schbul/sbw020.

116
S. Florentin et al. Schizophrenia Research 252 (2023) 110–117

Peralta, V., Gil-Berrozpe, G.J., Librero, J., Sánchez-Torres, A., Cuesta, M.J., 2020. The admission diagnosis of psychosis. Arch. Gen. Psychiatry 57 (6), 593–600. https://
symptom and domain structure of psychotic disorders: a network analysis approach, doi.org/10.1001/archpsyc.57.6.593.
schizophrenia. Bulletin. Open. Volume 1, Issue 1 https://doi.org/10.1093/ Sommer, I.E., Tiihonen, J., van Mourik, A., Tanskanen, A., Taipale, H., 2021. The clinical
schizbullopen/sgaa008. course of schizophrenia in women and men-a nation-wide cohort study. NPJ
Santelmann, H., Franklin, J., Bußhoff, J., Baethge, C., 2015. Test–retest reliability of Schizophr. 6 (1), 12. https://doi.org/10.1038/s41537-020-0102-z.
schizoaffective disorder compared with schizophrenia, bipolar disorder, and Visscher, P.M., Goddard, M.E., Derks, E.M., Wray, N.R., 2012. Evidence-based
unipolar depression–a systematic review and meta-analysis. Bipolar Disord. 17, psychiatric genetics, AKA the false dichotomy between common and rare variant
753–768. hypotheses. Mol. Psychiatry 17 (5), 474–485. https://doi.org/10.1038/mp.2011.65.
Santelmann, H., Franklin, J., Bußhoff, J., Baethge, C., 2016. Interrater reliability of Vollmer-Larsen, A., Jacobsen, T.B., Hemmingsen, R., Parnas, J., 2006. Schizoaffective
schizoaffective disorder compared with schizophrenia, bipolar disorder, and disorder– the reliability of its clinical diagnostic use. Acta Psychiatr. Scand. 113 (5),
unipolar depression - a systematic review and meta-analysis. Schizophr. Res. 176 402–407. https://doi.org/10.1111/j.1600-0447.2005.00744.x. PMID: 16603031.
(2–3), 357–363. https://doi.org/10.1016/j.schres.2016.07.012. Wood, A.J., Carroll, A.R., Shinn, A.K., Ongur, D., Lewandowski, K.E., 2021. Diagnostic
Santelmann, H., Franklin, J., Bußhoff, J., Baethge, C., 2016. Diagnostic shift in patients stability of primary psychotic disorders in a research sample. Front. Psychiatry 12,
diagnosed with schizoaffective disorder: a systematic review and meta-analysis of 734272. https://doi.org/10.3389/fpsyt.2021.734272.
rediagnosis studies. Bipolar Disord. 18 (3), 233–246. https://doi.org/10.1111/ World Health Organization, 1994. International classification of diseases for mortality
bdi.12388. and morbidity statistics (10th Revision). Retrieved from: https://icd.who.int/browse
Schizophrenia Psychiatric Genome-Wide Association Study (GWAS) Consortium, 2011. 11/l-m/en.
Genome-wide association study identifies five new schizophrenia loci. Nat. Genet. 43 Youngmann, R., Minuchin-Itzigsohn, S., Barasch, M., 1999. Manifestations of emotional
(10), 969–976. https://doi.org/10.1038/ng.940. distress among ethiopian immigrants in Israel: patient and clinician perspectives.
Schwartz, J.E., Fennig, S., Tanenberg-Karant, M., Carlson, G., Craig, T., Galambos, N., Transcult.. Psychiatry 36 (1), 45–63. https://doi.org/10.1177/
Lavelle, J., Bromet, E.J., 2000. Congruence of diagnoses 2 years after a first- 136346159903600103.

117