Introduction
Despite the large body of research that has been carried out on psychotic illness, fundamental questions remain unanswered (Tandon et al. Reference Tandon, Keshavan and Nasrallah2008; MacDonald & Schulz, Reference MacDonald and Schulz2009). Among these challenges, we are still seeking to define the boundaries of psychosis and the extent to which those aspects of psychotic illness that we resolve into separate diagnostic categories are in fact distinct in any fundamental way. This is exemplified by (i) classical debate regarding the relationship between ‘non-affective psychosis’ (typified by schizophrenia) and ‘affective psychosis’ [typified by manic-depressive psychosis (now bipolar disorder) and psychotic depression] (Jablensky, Reference Jablensky1999; Demjaha et al. Reference Demjaha, Morgan, Morgan, Landau Dean, Reichenberg, Sham, Fearon, Hutchinson, Jones, Murray and Dazzan2009; Fischer & Carpenter, Reference Fischer and Carpenter2009; Craddock & Owen, Reference Craddock and Owen2010) and (ii) the broader, contemporary debate regarding a dimensional rather than a categorical concept of psychosis; current psychotic diagnoses may reflect not discrete entities but, rather, domains defined by certain psychopathological and functional characteristics, the boundaries of which are probably notional and in continuity with other domains of mental illness, through to the limits of ‘normal’ human behavior (Tandon et al. Reference Tandon, Nasrallah and Keshavan2009; van Os & Kapur, Reference van Os and Kapur2009; Waddington et al. Reference Waddington, Hennessy, O'Tuathaigh, Owoeye, Russell, Brown and Patterson2012).
DSM-IV (APA, 1994) defines 12 diagnostic categories that involve psychotic symptoms: schizophrenia (SZ); schizo-affective disorder (SA); schizophreniform disorder (SF); delusional disorder (DD); brief psychotic disorder (BrP); bipolar I disorder (BD); major depressive disorder, with psychotic features (MDDP); substance-induced psychotic disorder (SIP); substance-induced mood disorder, with manic features (SIM); psychotic disorder due to a general medical condition (PGMC); mood disorder due to a general medical condition, with manic features (MGMC); and psychotic disorder not otherwise specified (PNOS). However, although these diagnoses differ operationally in terms of etiological assumptions and characteristics such as duration of illness and extent of affective symptoms, they all share the presence of psychotic symptoms at some point during the course of illness. Early in the course of psychotic illness, the maelstrom of symptoms can lead to considerable diagnostic fluidity (Malla & Payne, Reference Malla and Payne2005). Thereafter, long-term (in)stability of diagnosis may yield important information as to their import: do initial diagnoses remain stable, diverge, or converge to a smaller number of more highly populated categories that might be considered fundamental diagnostic nodes?
One essential strategy for addressing the challenges raised by this categorical approach has yet to be applied in the numerous follow-up studies of first-episode cohorts to date (Bromet et al. Reference Bromet, Naz, Fochtmann, Carlson and Tanenberg-Karant2005; Menezes et al. Reference Menezes, Arenovich and Zipursky2006), primarily due to attendant methodological difficulties: to study prospectively the ‘totality’ of first-episode psychosis (FEP), ascertained on an epidemiological basis across the whole adult lifespan and through all routes to care, in the absence of a priori diagnostic restriction; longitudinal application of contemporary diagnostic algorithms as post-hoc assessment, rather than as a criterion for inclusion/exclusion, would then allow the epidemiological, clinical and pathobiological characteristics of resultant diagnoses to be compared systematically. Described here are (i) procedures that attain 97% follow-up, at a mean interval of 6 years, of 202 subjects constituting an epidemiological cohort of FEP, and (ii) the diagnostic trajectory, interplay and extent of convergence/divergence across each of the 12 DSM-IV psychotic diagnoses over this interval.
Method
Study setting
Subjects were participants in the Cavan-Monaghan First Episode Psychosis Study (CAMFEPS). This is a prospective study, operating since 1995, of ‘all’ incident cases presenting with a first episode of any psychotic disorder in two rural counties in Ireland, Cavan and Monaghan, as described previously in detail (Scully et al. Reference Scully, Quinn, Morgan, Kinsella, O'Callaghan, Owens and Waddington2002; Baldwin et al. Reference Baldwin, Browne, Scully, Quinn, Morgan, Kinsella, Owens, Russell, O'Callaghan and Waddington2005). Cavan and Monaghan are two contiguous rural counties with a population of 109139 (55821 males and 53318 females) at the 2002 census; the region consists of remote areas, villages and towns, in the absence of any major urban areas, and is of substantial ethnic and social homogeneity, with the vast majority of the population being white Irish (Central Statistics Office, 2003).
The study is based within the Cavan-Monaghan Mental Health Service, which operates a community-based service model with a focus on home treatment, general practice liaison and services based in small local clinics. It involves two community mental health teams, a specialist service for the elderly and a community rehabilitation team; central to the delivery of health services in this model is the use of home-based treatment as an alternative to hospital admission (McCauley et al. Reference McCauley, Rooney, Clarke, Carey and Owens2003). All cases from this catchment area who present to services in other parts of the country are returned to the Cavan-Monaghan Mental Health Service as soon as is practicable.
Study cohort
In outline, CAMFEPS involves the following ascertainment procedures (Baldwin et al. Reference Baldwin, Browne, Scully, Quinn, Morgan, Kinsella, Owens, Russell, O'Callaghan and Waddington2005): cases resident in the Cavan-Monaghan Mental Health Service catchment area are identified from (a) all treatment teams in the catchment areas, (b) cases from the catchment areas who choose to avail of private mental health care in St Patrick's Hospital, Dublin, or St John of God Hospital, Co. Dublin, which together account for >98% of all national private psychiatric admissions (Daly et al. Reference Daly, Walsh, Moran and O'Doherty2004), and (c) cases from the catchment areas having admission to the national forensic service at the Central Mental Hospital, Dublin; the small number of cases missed initially and identified subsequently are incorporated retrospectively. The primary criterion for entry to the study is a first lifetime episode of any psychotic illness, to include a first manic episode, at age ⩾16 years, with no upper age limit. DSM-IV diagnosis is made at first presentation, together with psychopathological and cognitive assessments that are outlined elsewhere (Owoeye et al. Reference Owoeye, Kingston, Hennessy, Baldwin, Browne, Scully, Kinsella, Russell, O'Callaghan, Waddington, Gattaz and Busatto2010). At 6 months thereafter, as described previously (Baldwin et al. Reference Baldwin, Browne, Scully, Quinn, Morgan, Kinsella, Owens, Russell, O'Callaghan and Waddington2005), all clinical information, that is case-notes and discussions with the treating teams, is reviewed to confirm or update, by DSM-IV criteria, the initial diagnosis. There are no exclusion criteria other than a previously treated psychotic episode. Thus, all psychotic diagnoses included in DSM-IV are incepted into the study (i.e. SZ, SA, SF, DD, BrP, BD, MDDP, SIP, SIM, PGMC, MGMC and PNOS). We additionally encountered simple deteriorative disorder (SDD) (see DSM-IV Appendix B: Criteria Sets and Axes Provided for Further Study), which encompasses schizophrenia-like deterioration in functioning with primarily negative symptoms and transient psychotic symptoms that do not reach the threshold for diagnosis of a psychotic disorder. This disorder, having a long history as ‘simple schizophrenia’ and retained in ICD-9, was included here on the basis of it conforming in considerable part to classical concepts of the ‘clinical core’ of schizophrenia (Black & Boffeli, Reference Black and Boffeli1989; Parnas, Reference Parnas2011).
Follow-up
This study is a follow-up of the 202 cases having their first presentation over the first 8 years of CAMFEPS, between May 1995 and April 2003 (Baldwin et al. Reference Baldwin, Browne, Scully, Quinn, Morgan, Kinsella, Owens, Russell, O'Callaghan and Waddington2005). Follow-up was undertaken between July 2005 and June 2007. Study protocols were approved by the Research Ethics Committees of the North Eastern Health Board (and, following restructuring, of the Health Service Executive Dublin North East Area), St Patrick's Hospital, Dublin, St John of God Hospital, Co. Dublin, and the Central Mental Hospital, Dublin, to include (a) subjects giving written, informed consent to formal assessment and (b) obtaining diagnostic, clinical and demographic information from case-notes and treating health professionals for subjects declining formal assessment. Given that, for some subjects, more than a decade had elapsed since onset of what might have been a single episode of psychosis at a vulnerable time of their lives, it was felt that initial contact should be made in a sensitive manner; thus, one of the key tenets determining the mode of re-establishing contact with patients to seek their participation was that this should be made initially by a health professional with whom they had previous clinical contact.
Once name and date of birth had been identified for each subject, a search of the electronic records of Cavan-Monaghan Mental Health Service was performed to identify those still in contact with the service and their point of contact. For those who were no longer in contact with the service, a search of their medical records was performed to gather information on address, general (family) practitioner (GP) and past key workers. A notice was placed on the front page of each subject's medical records, identifying them as a potential candidate for the study and requesting mental health professionals to (i) ask if they would be willing to participate, and then (ii) contact the investigator with the outcome of the request. Where medical records were missing, as many details as possible were gleaned from the CAMFEPS dataset to facilitate tracing.
Contact was then made with individual clinical teams to establish, for those subjects still in contact with the service, their next appointment; community psychiatric nurses (CPNs) were extremely important in this process, as the quality of the relationships established by CPNs with subjects greatly facilitated the process of tracing and obtaining consent to participate. For those who declined to participate, information was obtained on the basis of an interview with their mental health professional and this process pertained to any other key worker identified (GP or psychiatrist outside of the Cavan-Monaghan Mental Health Service), as long as they had had regular contact with the subject within the past year.
For those subjects who were no longer under the care of a mental health service (approximately 50% of participants), telephone contact was made with their GP to establish whether they still attended and, if not, who was their new GP. The relevant GPs were then asked to inform the subjects about the study and ask if they would be willing to participate. If the subjects were willing to participate, they were informed that the investigator would make contact with them to arrange an appointment for assessment.
A small number of subjects had been incepted through private psychiatric hospitals or the national forensic service; follow-up involved contacting their treating consultant psychiatrist, who was requested to ask the subject if they were willing to participate and be contacted by the investigator.
Where the subject had moved out of the catchment area and was now being treated by another mental health service, similar procedures were followed as for primary care, substituting consultant psychiatrist for GP.
In instances where there was difficulty in identifying a current address for a subject, other community agencies were used, such as Medical Card Offices, Community Welfare Officers and police. Once an address was identified, all GPs in the area were contacted to establish the GP currently providing care; this GP then initiated contact with the subject on behalf of the investigator.
The Births, Marriages and Deaths Offices for Cavan and Monaghan were contacted with a list of names of all subjects who were thought by health-care professionals to be deceased, or who had not been traced successfully through the above means. The Officer provided death certificates for all subjects who were deceased; from these, date and cause of death were noted and included in the dataset.
Follow-up assessment
For subjects giving informed consent to reassessment at follow-up, each was interviewed, either in their own home, in a community clinic, in hospital or elsewhere at their convenience. The Structured Clinical Interview for DSM-IV (SCID-IV; First et al. Reference First, Spitzer, Gibbon and Williams2002) was conducted by the first author (T.K.), who was blind to diagnosis at 6 months, to establish current DSM-IV diagnosis; additional assessments included psychopathology, functioning, quality of life and service engagement, to be described elsewhere. For subjects declining reassessment, DSM-IV diagnosis was made on the basis of all available information gathered from case-notes and their current key worker, whether a mental health professional or a GP.
Additional demographic information ascertained at follow-up included whether deceased, with cause of death, marital status, living status and any co-morbidity. The investigator was a Clinical Research fellow who conducted all follow-up assessments without reference to assessments at onset and 6 months.
Diagnostic consistency was determined as prospective consistency (PC; the percentage of subjects with each diagnosis at 6 months who retained that diagnosis at 6 years) and retrospective consistency (RC; the percentage of subjects with each diagnosis at 6 years who had received that diagnosis at 6 months). Data were analyzed using SPSS version 15 (SPSS Inc., USA).
Results
Demographics
We sought to follow up all 202 cases incepted into CAMFEPS over its first 8 years [mean age at first presentation 36.0 (s.d. = 18.1) years; 121 male, mean age 32.4 (s.d. = 15.7) years; 81 female, mean age 41.5 (s.d. = 20.2) years]. Of these, 196 [97%; mean age at follow-up 42.6 (s.d. = 18.6) years; 115 male, mean age 38.9 (s.d. = 16.3) years; 81 female, mean age 48.0 (s.d. = 20.3) years] were followed up at a mean of 6.4 (s.d. = 2.3, range 2.6–11.9) years after first presentation, for which ‘6-year follow-up’ is used hereafter as convenient shorthand. It was not possible to obtain follow-up data for six cases [3%, all male; mean age at first presentation 29.3 (s.d. = 9.2) years]: for two, no record of the case was identifiable; two had moved out of the country and were untraceable; and for two, their records were in a private hospital and could not be accessed. Their diagnoses at first presentation were: one each of SZ, SA, BD, MDDP, DD and PNOS.
Of the 196 cases for whom follow-up data and DSM-IV diagnoses were obtained [179 (91%) through the SCID; 17 (9%) through all available sources of clinical information, that is case-notes and discussions with the treating teams (Baldwin et al. Reference Baldwin, Browne, Scully, Quinn, Morgan, Kinsella, Owens, Russell, O'Callaghan and Waddington2005)], 13 (seven male, six female) were found to be deceased: for four cases death was by suicide [all male, mean age at death 58.5 (s.d. = 16.8) years; diagnoses at first presentation: two SZ; two MDDP); for one case death was the result of a road traffic accident (female, age at death 27 years; diagnosis at first presentation: MDDP); and for eight cases deaths were the result of typical natural causes [five cardiovascular, three pneumonia; three male, five female; mean age at death 74.0 (s.d. = 11.4) years; diagnoses at first presentation: one SF, one BD, four MDDP and one SIP].
Diagnostic interplay from 6 months to 6-year follow-up
Data on diagnostic shifts between first presentation and 6-month assessments, together with the incidence of individual psychotic diagnoses at 6 months, have been presented previously in detail (Baldwin et al. Reference Baldwin, Browne, Scully, Quinn, Morgan, Kinsella, Owens, Russell, O'Callaghan and Waddington2005). Demographics for each psychotic diagnosis at the 6-month assessment are given in Table 1. For these 196 cases, details of stability of, or transitions from, diagnoses at 6 months to other categories at 6 years are summarized in Table 3. Diagnosis of SF at both 6 months and 6 years refers to no recurrence of psychosis since the first episode. One case with a diagnosis of SF at 6 months died subsequently of natural causes. Diagnosis of BrP at 6 years refers to no recurrence of psychosis since the first episode. One case with a diagnosis of MDDP at 6 months subsequently experienced a hypomanic episode and then a further depressive episode with psychotic features; this resulted in a diagnosis at 6 years of bipolar II disorder, depressed, with psychotic features. Seven cases with a diagnosis of MDDP at 6 months died subsequently (four natural causes, one road traffic accident, two suicide); an 80-year-old female with a diagnosis of PNOS at 6 months received a diagnosis of Alzheimer's disease (ALZ) at 6 years.
Table 1. Demographics for each psychotic diagnosis at 6 months following first presentation
SZ, Schizophrenia; SA, schizo-affective disorder; SF, schizophreniform disorder; DD, delusional disorder; BrP, brief psychotic disorder; BD, bipolar I disorder; MDDP, major depressive disorder, with psychotic features; SIP, substance-induced psychotic disorder; SIM, substance-induced mood disorder, with manic features; PGMC, psychotic disorder due to a general medical condition; MGMC, mood disorder due to a general medical condition, with manic features; PNOS, psychotic disorder not otherwise specified; SDD, simple deteriorative disorder; PC, prospective consistency to 6 years; s.d., standard deviation.
Trajectories to ‘convergence’ at 6-year follow-up
Demographics for each psychotic diagnosis at the 6-year follow-up are given in Table 2. For these 196 cases, specifics of stability of, or transitions to, diagnoses at 6 years from other categories at 6 months are summarized in Table 3. Two cases with a diagnosis of SF at 6 years involve no recurrence of psychosis between the first episode and 6 months but re-emergence of SF at 6 years. One case with a diagnosis of DD at 6 years involved a first episode of BrP with no recurrence of psychosis at 6 months but emergence of DD at 6 years. One case with a diagnosis of BrP at 6 years involved a first episode of substance-induced psychosis with no recurrence of psychosis at 6 months until the emergence of BrP at 6 years. One case involved a first episode of substance-induced mood disorder with psychotic features, with no recurrence of psychosis after 6 months until the emergence of BrP at 6 years.
Table 2. Demographics for each psychotic diagnosis at 6-year follow-up
SZ, Schizophrenia; SA, schizo-affective disorder; SF, schizophreniform disorder; DD, delusional disorder; BrP, brief psychotic disorder; BD, bipolar I disorder; BD2DP, bipolar II disorder, depressed, with psychotic features; MDDP, major depressive disorder, with psychotic features; SIP, substance-induced psychotic disorder; SIM, substance-induced mood disorder, with manic features; PGMC, psychotic disorder due to a general medical condition; MGMC, mood disorder due to a general medical condition, with manic features; ALZ, Alzheimer's disease; PNOS, psychotic disorder not otherwise specified; SDD, simple deteriorative disorder; RIP, deceased other than suicide; RC, retrospective consistency from 6 months; s.d., standard deviation.
Table 3. Summary of diagnostic transitions over illness trajectory from 6-month assessment to 6-year follow-up
SZ, Schizophrenia; SA, schizo-affective disorder; SF, schizophreniform disorder; DD, delusional disorder; BrP, brief psychotic disorder; BD, bipolar I disorder; BD2DP, bipolar II disorder, depressed, with psychotic features; MDDP, major depressive disorder, with psychotic features; SIP, substance-induced psychotic disorder; SIM, substance-induced mood disorder, with manic features; PGMC, psychotic disorder due to a general medical condition; MGMC, mood disorder due to a general medical condition, with manic features; ALZ, Alzheimer's disease; PNOS, psychotic disorder not otherwise specified; SDD, simple deteriorative disorder; RIP, deceased other than suicide.
Discussion
CAMFEPS was initiated in 1995 (Baldwin et al. Reference Baldwin, Browne, Scully, Quinn, Morgan, Kinsella, Owens, Russell, O'Callaghan and Waddington2005) using methodology that seeks the closest approximation to epidemiological completeness for FEP. To our knowledge, no other study has included each of the defined catchment areas, all routes to care (i.e. public, private or forensic), all modes of care (i.e. in-patient, out-patient or home-based), full diagnostic scope (i.e. all 12 DSM-IV psychotic diagnoses), no arbitrary upper-age cut-off (i.e. cases incepted throughout the adult lifespan) and Research Ethics Committee approvals to include demographic and diagnostic information on those cases declining formal assessment.
In the present study, we have continued this methodological approach to seek the closest approximation to epidemiological completeness of follow-up at 6 years following the first psychotic episode. The purpose of the study was to determine, for the 202 cases incepted over the first 8 years of CAMFEPS, the extent to which each of the 12 DSM-IV psychotic diagnoses, made following the first 6 months of illness, remain stable, converge to a smaller number of accumulating diagnoses or diverge. These methods for case ascertainment and assessment at long-term follow-up allowed us to obtain diagnoses for 97% of the initial cohort, with 91% of these being through a face-to-face SCID and 9% through case records and/or current key worker information, whether a mental health professional or GP. We consider each of the 12 DSM-IV psychotic diagnoses in turn, beginning with the most populous categories. It should be emphasized that interpretation vis-à-vis previous studies is rendered problematic by none having adopted the current methodological approach in a comparable (i.e. rural) setting.
At follow-up, SZ was a generally stable and accumulating diagnosis (PC > RC), primarily from SF, DD and PNOS; this is in accordance with both the classical literature (van Os & Kapur, 2009) and prospective first-episode studies (Bromet et al. Reference Bromet, Naz, Fochtmann, Carlson and Tanenberg-Karant2005, Reference Bromet, Kotov, Fochtmann, Carlson, Tanenberg-Karant, Ruggero and Chang2011; Addington et al. Reference Addington, Chaves and Addington2006; Chang et al. Reference Chang, Pang, Chung and Chan2009). In a ‘sister’ cohort, conducted in an urban population in Ireland, the incidence of SZ was higher than in the present rural cohort but showed similar long-term diagnostic stability; thus, importantly, the excess of SZ associated with urbanicity does not seem to be attributable to variations in diagnostic stability (Whitty et al. Reference Whitty, Clarke, McTigue, Browne, Kamali, Larkin and O'Callaghan2005; Kelly et al. Reference Kelly, O'Callaghan, Waddington, Feeney, Browne, Scully, Clarke, Quinn, McTigue, Morgan, Kinsella and Larkin2010). The small number of cases evolving to BD or MDDP over time indicates that vigilance should be maintained in monitoring for and treating affective symptoms over the long-term course of an illness that seems essentially psychotic at onset.
Despite SA having generated so much controversy historically (Maier, Reference Maier2006; Cheniaux et al. Reference Cheniaux, Landeira-Fernandez, Telles, Lessa, Dias, Duncan and Versiani2008) and being a ‘longstanding problem that plagues psychiatric nosology’ (Mahli et al. Reference Mahli, Green, Fagiolini, Peselow and Kumari2008), we have not been able to identify any contemporary, prospective studies of FEP that have ascertained and followed up meaningful numbers of cases of SA, these being more commonly pooled pragmatically into a category of ‘schizophrenia spectrum’ psychoses (Addington et al. Reference Addington, Chaves and Addington2006; Chang et al. Reference Chang, Pang, Chung and Chan2009; Bromet et al. Reference Bromet, Kotov, Fochtmann, Carlson, Tanenberg-Karant, Ruggero and Chang2011). Here, we find SA to also be a generally stable and accumulating diagnosis (PC > RC), but with a profile of accumulation, that is from BD/MDDP, distinct from that seen in SZ. This differential accumulation from ‘feeder’ diagnoses would indicate that SA might be more correctly characterized as a psychotic mood disorder (Lake & Hurwitz, Reference Lake and Hurwitz2006).
Over the past several years BD has been increasingly included in prospective studies of FEP, with affective psychosis having previously been more commonly a reason for exclusion (Menezes et al. Reference Menezes, Arenovich and Zipursky2006). We found BD to be a generally stable diagnosis, in accordance with both the classical literature (Müller-Oerlinghausen et al. Reference Müller-Oerlinghausen, Berghöfer and Bauer2002) and more recent prospective first-episode studies (Schimmelmann et al. Reference Schimmelmann, Conus, Edwards, McGorry and Lambert2005; Chang et al. Reference Chang, Pang, Chung and Chan2009). However, transitions to BD (PC = RC) were less frequent than for SZ and SA (see also Ruggero et al. Reference Ruggero, Carlson, Kotov and Bromet2010) and were commonly from MDDP due to the emergence of at least one manic episode (Ketter et al. Reference Ketter, Wang, Becker, Nowarkowska and Yang2004). Transitions from BD were more frequent than for SZ and SA, and were most commonly to SA (Conus et al. Reference Conus, Abdel-Baki, Harrigan, Lambert, McGorry and Berk2010).
Among those psychotic diagnoses occurring at the first episode, MDDP has received the least systematic study; even more so than for BD, MDDP has been a common reason for exclusion from FEP studies (Menezes et al. Reference Menezes, Arenovich and Zipursky2006). We find MDDP to be a less stable and more complex diagnosis (PC < RC) than SZ, SA or BD (Ruggero et al. Reference Ruggero, Kotov, Carlson, Tanenberg-Karant, Gonzáles and Bromet2011; Tohen et al. Reference Tohen, Khalsa, Salvatore, Vieta, Ravichandran and Baldessarini2012). Emergence of psychosis in the context of MDD reflects most robustly the intersection of affective and psychotic psychopathological dimensions (Demjaha et al. Reference Demjaha, Morgan, Morgan, Landau Dean, Reichenberg, Sham, Fearon, Hutchinson, Jones, Murray and Dazzan2009; Linscott & van Os, Reference Linscott and van Os2010; Waddington et al. Reference Waddington, Hennessy, O'Tuathaigh, Owoeye, Russell, Brown and Patterson2012). Here, MDDP was characterized in the long term by three primary trajectories: (i) for the majority, an enduring diagnosis of MDDP; thus, these affective and psychotic dimensions continued to intersect rather than converge on an affective psychotic diagnosis; (ii) for a minority, enduring psychosis, with transition to more portentous psychotic diagnoses, particularly those having an affective dimension (BD > SA > SZ); however, these transitions occurred less often than reported recently in an FEP study that adopted a different methodological approach (Bromet et al. Reference Bromet, Kotov, Fochtmann, Carlson, Tanenberg-Karant, Ruggero and Chang2011); and (iii) for a minority, death due to suicide, accidental death or natural causes. Although explanations for this high rate of mortality over follow-up are likely to include the older age and increased risk for suicide among cases of MDDP, it suggests scope for further attention to both mental and physical health in this group to reduce such long-term adversities.
Long-term transition of SF primarily to SZ (PC < RC) was in accordance with both classical (Strakowski, Reference Strakowski1994) and contemporary (Marchesi et al. Reference Marchesi, Paini, Ruju, Rosi, Turrini and Maggini2007) literature, due in the main to enduring psychosis that ultimately attains the DSM-IV duration-of-illness criterion for SZ; however, we did not encounter previously described transitions to BD or MDDP (Marchesi et al. Reference Marchesi, Paini, Ruju, Rosi, Turrini and Maggini2007). For a smaller number of cases, follow-up of SF clarified the role of substance abuse and led to subsequent diagnosis of SIP; for only a small minority of cases was SF a benign initial diagnosis without long-term import.
The literature regarding BrP is confounded by studies of ICD-10 F23, ‘acute and transient psychotic disorders’, which encompasses a wide variety of clinical concepts that overlap with, but extend far beyond, BrP and indeed SF (Castagnini & Berrios, Reference Castagnini and Berrios2009; Chang et al. Reference Chang, Pang, Chung and Chan2009). We have been able to identify only one contemporary, prospective study of FEP that has ascertained and followed up subjects with DSM-IV BrP; this involved five cases assessed over 1 year, one of whom evolved to schizophrenia (Addington et al. Reference Addington, Chaves and Addington2006). In the present study, all cases of BrP at presentation, occurring primarily among women, were by definition characterized initially by the rapid remission of psychosis; thereafter, the large majority remained well over the subsequent 6 months, with a small minority experiencing transition to BD or MDDP during this period (Baldwin et al. Reference Baldwin, Browne, Scully, Quinn, Morgan, Kinsella, Owens, Russell, O'Callaghan and Waddington2005). However, among those cases of BrP remaining well at 6 months, by 6 years the large majority had experienced deterioration to SZ, SA, DD, BD or MDDP. The case of SDD showed a similar course, as noted previously (Serra-Mestres et al. Reference Serra-Mestres, Gregory, Tandon, Stansfield, Kemp and McKenna2000).
Thus, over the duration of CAMFEPS to date, 85% of cases of BrP have experienced rapid remission, followed only in the longer term by the emergence of a more portentous psychotic diagnosis. BrP may be a far from benign occurrence that is, in reality, the harbinger of long-term evolution to serious psychotic illness, the diagnostic diversity of which may further challenge categorical nosology and favor a dimensional model. The emergence of BrP, with its defining rapid remission, may be a transient exacerbation over some arbitrary threshold of what is currently conceptualized subclinically as ‘attenuated psychotic symptoms’ or ‘brief, limited, intermittent psychotic symptoms’, which can confer ‘ultra high risk’ for subsequent psychotic illness (Nelson et al. Reference Nelson, Yuen and Yung2011). This arbitrary threshold may be a position along a dimension that extends from the breadth of psychotic ideation in the ‘normal’ population through to clinical psychosis (van Os et al. Reference van Os, Linscott, Myin-Germeys, Delespaul and Krabbendam2009; Demjaha et al. Reference Demjaha, Morgan, Morgan, Landau Dean, Reichenberg, Sham, Fearon, Hutchinson, Jones, Murray and Dazzan2009; Linscott & van Os, Reference Linscott and van Os2010; Waddington et al. Reference Waddington, Hennessy, O'Tuathaigh, Owoeye, Russell, Brown and Patterson2012); BrP may straddle the upper boundary along a dimension of ‘ultra high risk’ for psychotic illness that then transitions to variable diagnostic categories with overlapping dimensions of psychopathology. Irrespective of these theoretical considerations, BrP may require more vigorous, sustained intervention than is currently appreciated, so as to reduce risk for such long-term sequelae.
The trajectory of DD was similar to that of SF in most commonly evolving to SZ, as noted previously (Whitty et al. Reference Whitty, Clarke, McTigue, Browne, Kamali, Larkin and O'Callaghan2005; Chang et al. Reference Chang, Pang, Chung and Chan2009); one case of BrP at first presentation evolved to DD over follow-up. As with BrP and SF, for only a minority of cases was DD a benign initial diagnosis without long-term import. This trajectory suggests that treatment at first presentation and thereafter should take into account the likelihood of DD being the harbinger of long-term evolution to more serious psychotic illness and possibly benefiting from more vigorous, sustained intervention.
Contemporary studies have reported varying findings as to the stability of SIP versus evolution to a primary psychotic diagnosis over varying periods of follow-up (Whitty et al. Reference Whitty, Clarke, McTigue, Browne, Kamali, Larkin and O'Callaghan2005; Addington et al. Reference Addington, Chaves and Addington2006; Caton et al. Reference Caton, Hasin, Shrout, Drake, Dominguez, First, Samet and Schanzer2007; Bromet et al. Reference Bromet, Kotov, Fochtmann, Carlson, Tanenberg-Karant, Ruggero and Chang2011). Here, SIP occurred only in men and was due to abuse of a wide range of individual or combined substances. Over the 6-year follow-up, similar proportions of SIP retained that diagnosis, or evolved to SZ/SA. Thus, it is important in the early treatment of SIP to note that simple abstinence from the putative causative agent may not be adequate for achieving freedom from future psychotic episodes. We have been unable to identify any other study that has specifically ascertained and systematically followed up subjects with DSM-IV SIM. Here, SIM occurred in both men and women due to substances of either abuse or pharmacotherapy [primarily antidepressant-induced mania (Goldberg & Truman, Reference Goldberg and Truman2003; Daray et al. Reference Daray, Thommi and Ghaemi2010) and corticosteroid-induced mania (Brown & Suppes, Reference Brown and Suppes1998; Kenna et al. Reference Kenna, Poon, de los Angeles and Koran2011)]. Over follow-up, similar proportions of SIM retained that diagnosis or evolved to SA/BD.
We have been unable to identify any other study that has ascertained and followed up subjects with DSM-IV PGMC and MGMC. Here, similar proportions of PGMC retained that diagnosis or evolved to SZ/SA. The single case of MGMC retained that diagnosis at 6 years.
The primary limitation of the present study is that, despite the methodologies used, an unknown number of cases still may have been missed and information on cases declining formal assessment is less complete. Additionally, the total number of cases identified results in the least common diagnostic categories having small numbers. Furthermore, contemporary studies have reported varying findings regarding the long-term diagnostic outcome of PNOS over differing periods of follow-up, with evolution to SZ being most common (Whitty et al. Reference Whitty, Clarke, McTigue, Browne, Kamali, Larkin and O'Callaghan2005; Addington et al. Reference Addington, Chaves and Addington2006; Chang et al. Reference Chang, Pang, Chung and Chan2009). Here, over 6 years the most common diagnostic transition from PNOS was to SZ/SA as the long-term course of psychotic illness became clearer. For one elderly female, diagnosis at 6 years was Alzheimer's disease; her atypical psychosis at 6 months may have been an early manifestation of Alzheimer's neuropathology, attesting the challenges of distinguishing in such circumstances between DSM-IV diagnoses of PNOS versus dementia of the Alzheimer type, with delusions (Ropacki & Jeste, Reference Ropacki and Jeste2005). However, even with all the assessments available at the first episode, 6 months and 6 years, supplemented by all available information from case records and health professionals, almost one-third of cases diagnosed initially as PNOS continued to defy the SCID/DSM-IV algorithm and endured as PNOS.
Conclusions
CAMFEPS methodology reveals limited diagnostic convergence, primarily to SZ/SA (from 27% to 44% of the cohort), and elaborates the diversity of stabilities in, and transitions between, all 12 DSM-IV psychotic diagnoses over 6 years; thus, the present findings indicate that psychosis is manifested within a dimensional space, with diagnostic outcomes that can show longitudinal disrespect to current categories. The extent to which psychosis might be better described dimensionally (van Os & Kapur, Reference van Os and Kapur2009; Waddington et al. Reference Waddington, Hennessy, O'Tuathaigh, Owoeye, Russell, Brown and Patterson2012), as typified by MDDP, may be further clarified by systematic, comparative evaluation of pre-morbid characteristics and, prospectively, of psychopathology, neuropsychology and functional outcome in relation to treatments across these 12 DSM-IV psychotic diagnoses. In particular, the findings suggest that first-episode diagnoses of BrP, MDDP and DD may benefit from more vigorous, sustained interventions to reduce risk for potentially adverse outcomes in the long term.
Acknowledgments
The CAMFEPS was supported by the Stanley Medical Research Institute and Cavan-Monaghan Mental Health Service. We thank J. Quinn and M. Morgan for their contributions to data collection and the staff of Cavan-Monaghan Mental Health Service, particularly M. Maguire and T. McEnteggart, its associated GPs, St Patrick's Hospital, Dublin, St John of God Hospital, Co. Dublin, and the Central Mental Hospital, Dublin, for their kind assistance.
Declaration of Interest
None.
