predicting spectrums of adult mania, psychosis and depression by ...376212/uq376212_oa.pdf ·...
TRANSCRIPT
Accepted Manuscript
Predicting spectrums of adult mania, psychosis and depression by prospectivelyascertained childhood neurodevelopment
Dr Kim S. Betts, PhD, MPH, BEd., Prof Gail M. Williams, PhD, MSc(Epi), MSc(ApplStat), BSc(Hons)., Prof Jacob M. Najman, PhD, BA(Hons), Prof Rosa Alati, PhD,MApplSc(Health Sc), GradDip(Aboriginal Studies), BA(Hons).
PII: S0022-3956(15)00302-7
DOI: 10.1016/j.jpsychires.2015.10.013
Reference: PIAT 2749
To appear in: Journal of Psychiatric Research
Received Date: 16 June 2015
Revised Date: 3 September 2015
Accepted Date: 16 October 2015
Please cite this article as: Betts KS, Williams GM, Najman JM, Alati R, Predicting spectrums of adultmania, psychosis and depression by prospectively ascertained childhood neurodevelopment, Journal ofPsychiatric Research (2015), doi: 10.1016/j.jpsychires.2015.10.013.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service toour customers we are providing this early version of the manuscript. The manuscript will undergocopyediting, typesetting, and review of the resulting proof before it is published in its final form. Pleasenote that during the production process errors may be discovered which could affect the content, and alllegal disclaimers that apply to the journal pertain.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
1
Predictingspectrums of adult mania, psychosis and depressionby
prospectively ascertained childhood neurodevelopment
Running head: Neurodevelopment, mania, psychosis and depression
Abstract word count is 247 words.
Main text word count is 3,988 words.
There are 4 tables and 1 figure.
There are 1 supplementary tables.
Dr Kim S. Betts (corresponding author), PhD, MPH, BEd.
School of Population Health, University of Queensland, Brisbane, Australia
Contact details of Kim Betts:
(Care of) Rosa Alati
School of Population Health
The University of Queensland
4th
floor, Public Health Building
Herston Rd, Herston QLD 4006
Australia
phone: +617 33655509
Prof Gail M. Williams, PhD, MSc(Epi), MSc(Appl Stat), BSc(Hons).
School of Population Health, University of Queensland, Brisbane, Australia
Prof Jacob M. Najman, PhD, BA(Hons)
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
2
School of Social Science and Population Health, University of Queensland, Brisbane,
Australia
Prof Rosa Alati, PhD, MApplSc(Health Sc), GradDip(Aboriginal Studies), BA(Hons).
School of Public Health and Centre for Youth Substance Abuse Research, University of
Queensland, Brisbane, Australia
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
3
Background: We used a novel approach to investigateearlyneurodevelopmental factors of
later adult spectrums of mania, depression and psychosis as a means to identify etiological
similarities and differences among the three constructs.
Methods:Participants were from the Mater University Study of Pregnancy (MUSP), a pre-
birth cohort study started in Brisbane, Australia in 1981.A range of neurodevelopmental
variables were ascertained at age 5, including measures of cognitive ability, developmental
delay and behaviour problems. At age 21 offspring were assessed using a semi-structured
psychiatric interview. We used structural equation modelling to establish three latent factors
of mania, depression and psychotic symptoms. We thenregressedthese factors on the
neurodevelopmental variables and covariates.
Results: In both univariate and multivariate analysis premorbid cognitive ability predicted
only psychotic symptoms, developmental delay predicted only manic symptoms, while
behaviour problems predicted both depressive and psychotic symptoms. In a supplementary
analysis the three factors were also found to have unique relationships with a number of
outcomes also measured at age 21, including anxiety and substance use.
Conclusion:By assessing the impact of early childhood neurodevelopmenton the continuous
spectrums which underlie three serious adult psychiatric disorders in a general population
sample, we provide unique evidence regarding potential etiological similarities and
differences. Perhaps of most interest is that our findings suggestthat the manic and depressive
symptoms in bipolar depression, despite often overlapping in clinical presentations, may in
fact be somewhat separate entities with origins that are at least partly unique to either
disorder.
Key words: Mania, depression, psychosis, neurodevelopment, structural equation modelling
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
4
Introduction
Determining the etiologies of serious mental health problems including major depression,
bipolar disorders and schizophrenia remains a priority in efforts to reduce the considerable
morbidity and mortality associated with these disorders(Saha et al., 2005, Slade et al., 2009).
One area of research has focused on pinpointing thechildhood neurodevelopmental
abnormalities which predict such psychiatric disorders inlater life, in the hope of identifying
at risk individuals for early intervention. Such research hashighlighted a number of
neurodevelopmental similarities and differences among the disorders.Impairments in
cognitive ability, social functioning, attention and motor functioninghave been found to
predict schizophrenia (Dickson et al., 2012, Erlenmeyer-Kimling et al., 2000).While some
studies suggest similar impairments are not associated with later onset bipolar disorders
(Cannon et al., 2002, Murray et al., 2004),declines in early academic adjustment have been
noted in individuals who later develop bipolar disorders (Payá et al., 2013). Lastly, cognitive
ability deficits are also associated with unipolar depression (Zammit et al., 2004).
Such research also plays a central role in establishing the nosological boundaries among these
disorders by elucidating their impairedpremorbid profiles(Zammit et al., 2004).However,
conclusions about which neurodevelopmental factors predict which disorders and what this
suggests about their etiologies are confounded by the high levels of comorbidity among the
disorders. For example, a recent study discussed the conflicting findings with regard to the
levels of premorbid adjustment observed in bipolar disorders, for which various previous
studies support as being either better, the same or worse than controls (Payá et al., 2013).
The authors suggest the differences may be due to subgroups within bipolar individuals who
vary by the severity of psychotic symptoms. The importance of such comorbidity was again
highlighted in a recent population register study which found that higher intelligence did
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
5
predict bipolar disorder, but only in a minority of cases who had ‘pure bipolar’(Gale et al.,
2013), and by an earlier study which found a number of neurodevelopmental abnormalities
predicted schizophrenia but not non-psychotic bipolar disorder(Reichenberg et al.,
2002).Importantly, even studies which take comorbidity into accountare likely to include
participantswith considerable sub-threshold levels of comorbid disorders, and therefore
symptoms of another disorder which are unaccounted for in the analyses (Merikangas et al.,
2011).
The comorbidity which complicates etiological enquiry is adefining feature of the prevailing
diagnostic systems, which categorise and class disorders according to their primary features
built uponexpert determined thresholds of psychopathology(Goldberg et al., 2009, Linscott
and van Os, 2010). Thus, approaches aimed at elucidating the unique or shared origins of
related psychiatric disorders may benefit from employing alternative ways to define the
disorders under study. One approach suggested in the literature, but yet to be implemented in
a truly effective manner,is to incorporate the related psychiatric disorders into a single
analysis, and to operationalise the disorders as spectrums thereby accounting for comorbidity
even at sub-clinical levels(Cederlöf et al., 2013, Goldberg et al., 2009, Hickie,
2013).Accumulating research supports the dimensional structure of psychiatric disorders,
whereby the distribution of symptoms in the general population forms a continuous spectrum
(with perhaps more than a single spectrum underlying the symptoms)(Linscott and van Os,
2010, van Os et al., 2009), and where it may be reasonable to expect the increasing symptom
load to linearly relate to distress, impairment and to the etiological factors found to
precipitate the related clinical diagnoses(Blanchard et al., 2010, Kelleher et al., 2012).
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
6
In this study we used structural equation modelling to construct separate but correlated
spectrums (continuous factors) of mania, depression and psychosis measured in early
adulthood in a general population birth cohort. We then examined the relationships among a
range of neurodevelopmental factors ascertained during early childhood (age 5) with these
spectrums. We hypothesised that this unique methodology, which betteraccounts for
comorbidity among the disorder spectrums (even at sub-clinical levels), may more explicitly
demonstrate that the different disorders have partly unique underlying etiological factors.
Byproviding a clearer picture of how specific disorders are linked with specific
neurodevelopmental abnormalities wemay have the potential to inform early intervention
strategies.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
8
Methods
Participants
Participants are from the Mater University Study of Pregnancy (MUSP), a prospective pre-
birth cohort study following mothers and their children for over 20 years. A total of 7,223
mothers attending their first clinic visit at Brisbane’s Mater Misericordiae Hospital were
recruited between 1981 and 1984, with subsequent follow-ups at birth, and child age 6
months, and 5, 14 and 21 years, further information found elsewhere (Najman et al., 2005).
At 21 years 2,566 offspring completed the Composite International Diagnostic Interview
(CIDI-Auto 2.1) (World Health Organization, 1997), forming the sample for the
measurement model.The CIDI is a fully structured and comprehensive diagnostic interview
for the assessment of mental disorders and provides diagnosis by computerised algorithms,
with the questions eliciting symptoms and behaviours from respondents and mapping these to
diagnostic criteria. Regression analyses were limited to participants with values on all risk
factors of interest and covariates (n = 1,934).
Symptoms of mania, depression and psychosis
At the 21 year follow-up the lifetime version of the CIDI-Auto (World Health Organization,
1997) was administered by lay trained interviewers, including 11 symptoms of mania, 11
symptoms of depression and 12 symptoms of psychosis (see table one for symptom
descriptions and prevalence). Due to the format of the interview, symptoms of mania were
restricted to participants who had experienced at least four days of either (i) being so
happy/excited it caused problems with friends, family or a doctor told them that they were
manic, or (ii) being so irritable that they complained, started arguments, shouted at or hit
people. Also due to the format of the interview, symptoms of depression were restricted to
participants who for at least two weeks either (i) felt sad, empty or depressed for most of the
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
9
day nearly every day, or (ii) had lost interest in things they usually enjoyed. The psychotic
symptoms included delusions and hallucinations, for which positive responses were probed to
be surer the experience was psychotic. Due to the low prevalence of a number of the
symptoms itemsand the need to satisfy the requirements of the covariance matrix for the
subsequent structural equation model, a number of the items were either combined or
dropped as done in a previous study(Betts et al., 2014) (see supplementary text 1).
Childhood neurodevelopmental factors
At 5 years mothers completed a shortened version of the Achenbach Child Behaviour
Checklist (CBCL)(Achenbach, 1991),which included the most commonly occurring
behaviours, assessing 10 items each from the internalising and externalising scales, and 10
items from the social/attention/thought sub-scales. These items were combined into a total
behaviour problems scale (Alpha = 0.897), and dichotomised defining ‘cases’ using a cut-off
consistent with the percentage of cases identified by Achenbach in a community sample (Bor
et al., 1997). The correlation between the full form (ascertained on a subsample) and short
form for total behaviour problems was found to be very high (r = 0.98) (Bor et al., 1997).
Children completed the Peabody Picture Vocabulary Test – Revised (PPVT-R), requiring
them to indicate which one of four illustrations best represented a word expressed verbally by
the examiner, resulting in a score measuring the subjects verbal ability (Jongsma, 1982). The
PPVT-R has been validated against other standardised intelligence tests used on children
(Childers et al., 1994, Dunn, 1981, Johnson et al., 1993).Children were alsoadministered the
Denver Developmental Screening Test (DDST) (Frankenburg and Dodds, 1967), which
assesses developmental delays in the four key sectors of gross motor, fine motor-adaptive,
language and personal-social development. The DDST was administered by trained
researchers and a standard algorithm determined if the child’s performance on the relevant
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
10
task was normal, questionable or abnormal (Frankenburg et al., 1973). In this study we
classified any child with a rating of abnormal or questionable in any of the four areas as
developmentally delayed.
Covariates and concurrent factors at 21
As with similar studies using the MUSP cohort we included offspring gender, age (ages
ranged from 18-24 at the 21 year follow-up), and education (maternal education at baseline as
the child’s education at 21 may be influenced by the psychiatric symptoms) as covariates, and
only included significant risk factors and covariates in the multivariate analysis (Scott et al.,
2009, Welham et al., 2010). Lastly, concurrent factors measured at 21 were used in a number
of supplementary analyses, including the PPVT-R which was readministered at age 21, DSM-
IV anxiety disorders from the CIDI-Auto, and levels of alcohol and tobacco use. In addition,
the Attachment Style Questionnaire (ASQ) was used to assess five aspects of participant
attachment styles, consisting of 40 items scored on a 6-point Likert scale (Feeney et al.,
1994), and has been used frequently in studies assessing the relationship between various
mental health problems and adult relationships in population samples (Chotai et al., 2005,
Twaite and Rodriguez-Srednicki, 2004), including in the MUSP (Varghese et al., 2013). This
was done as one means of assessing the validity of the latent factors of depression, mania and
psychosis.
Statistical analysis
We used confirmatory factor analysis to create three separate but correlated latent factors of
mania, depression and psychosis using geomin rotation with the WLSMV estimator
appropriate for categorical data available in Mplus version 6(Muthén, 1998-2010). The a
priori measurement model was assessed for suitability using the standard model fit indices
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
11
including the Root Mean Square Error of Approximation (RMSEA), the Comparative Fit
Index (CFI), and the Tucker-Lewis Index (TLI), for which adequate fit is indicated by
RMSEA<0.06, CFI≥0.95 and TLI≥0.95 (Hu and Bentler, 1998).A series of univariate
regression analyses then examined associations between the risk factors and covariates on the
three latent factors, with statistically significant variables (p < 0.05) retained for the
multivariate analysis (see supplementary section for greater detail regarding model
specifications). We ran a series of descriptive (i.e., univariate) supplementary analysis using
the three latent factors to predict concurrent outcomes at age 21, to examine if our latent
constructs associate with other negative psychiatric and psychosocial outcomes (i.e., testing
construct validity). Lastly, to address loss to follow-up, we used multivariate logistic
regression to compare those who had been lost to follow-up with those used in the final
analysis by the neurodevelopmental factors at age 5. We then used the results from this
analysis to compute Inverse Probability Weights (IPW)representing the inverse probability of
each participant being included in the study, and replicated the final analysis using the IPWs
as in a previous study (Betts et al.).
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
12
Results
Symptom and disorder prevalence at age 21
The prevalence of life time mania, depression and psychosissymptoms are shown in table 1
revealing that some symptoms were rarely endorsed, particularly delusions and some mania
symptoms, while a number of depressive symptoms were highly prevalent. Among the 2,566
participants who undertook the CIDI-Auto,1.7% had a DSM-IV life time bipolar disorder,
19.7% had a life time DSM-IV Major Depressive Disorder, and 1.4% had a life time DSM-IV
any psychotic disorder.
Measurement model of symptoms
Table 1 also contains the factor loadings for the three latent factors (i.e., spectrums) of mania,
depression and psychosis, showing that all symptoms items loaded strongly onto their
respective factors. Importantly, the fit indices revealed the model fit the data very well (CFI =
0.99; TLI = 0.99, RMSEA = 0.018; chi-2 = 941.22; free parameters = 71), and the three
factors were strongly correlated.
Regression analyses
In both univariate and multivariate analyses (table 2 and figure 1) premorbid cognitive ability
predicted only the psychotic spectrum, female gender predicted only the depression spectrum,
developmental delay predicted only the maniaspectrum, while behaviour problems predicted
both the depression and psychotic spectrums. The effect sizes are standardised probit
regression estimates and represent the increase in standard deviation of the outcome variable
(i.e., continuous latent depression, mania or psychosis factor) for a one unit increase in the
dichotomous predictor variables (i.e., behaviour problems and developmental delay), or a one
standard deviation increase in the continuous predictor variables (i.e., cognitive ability).
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
13
Supplementary analyses
Supplementary table 1 shows the analyses of the univariate association among the three
spectrums of mania, depression and psychosis and a number of outcomes measured at the
same time point (i.e., age 21). After adjusting for cognitive ability at age 5, only the
depressive spectrum predicted lower cognitive ability at age 21. In addition, the mania
spectrum most strongly predicted alcohol and tobacco use, the depression spectrum predicted
all three DSM-IV anxiety disorders, and the psychosis spectrum only was associated with
feeling relationships were secondary and not feeling the need for approval. The attrition
analysis showed that participants who did not attend the 21 year psychiatric assessment were
more likely to be male and hadon average a small but significant 1 point increase on the
PPVT-R, but did not differ by developmental delays or behaviour problems (table 3). The
results of the final model after adjusting for the IPW were not substantively different
(supplementary table 2).
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
14
Discussion
We employed a novel approach and examined the impact of a range of early
neurodevelopmental abnormalitieson spectrums of later severe psychiatric disorders. As such,
our findings offer a unique insight into the early neurodevelopmental similarities and
differencesunderlying disorders characterised by mania, depression and psychosis (Zammit et
al., 2004). By avoiding the threshold and sub-thresholdcomorbid confoundinginherent in the
clinical diagnoses, and by operationalising the disorders as the corresponding symptomsmay
more naturally occur in the general population, we provide novel evidence of which types of
abnormalities precede the symptoms of which types of disorder. Child developmental
delaywas strongly predictive of later mania butdid not predict depression or
psychosis,whereaschild behaviour problems were strongly predictive of later depression and
psychosis but not mania, and cognitive ability in childhood predicted only psychosis.
Recent evidence suggests that mania and depression can be viewed as two separate
spectrums, rather than opposite ends of the same spectrum, and thus supports our use of
independent factors to represent mania and depression in efforts to determine aetiology
(Hickie, 2013, Johnson et al., 2011, Merikangas et al., 2012). While mania and bipolar
depression have been found to have separate biological, environmental and personality risk
factors, few studies have separately examined the mania and depression which occurs in
bipolar disorders (Cuellar et al., 2005). Our unique methodology has allowed us to assess that
the developmental delays strongly associated with mania are not related to the depressive
symptoms with which mania strongly correlates. This suggests that the manic and depressive
symptoms in bipolar depression, despite often overlapping in clinical presentations, may in
fact be somewhat separate entities with origins that are at least partly unique to either
disorder.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
15
The relationship we identified between lower cognitive ability and psychosis is in keeping
with previous findings, which have identified lower cognitive ability as predictive of not only
psychotic disorders such as schizophrenia as outlined above but also positive symptoms of
psychosis (Barnett et al., 2012, Polanczyk et al., 2010). As expected we found no relationship
between lower cognitive ability and mania once accounting for depression and psychosis
(Zammit et al., 2004), but did not find an increase in cognitive ability either (MacCabe et al.,
2010). Interestingly, in our supplementary analysis, in which we used the three spectrums of
mania, depression and psychosis to predict cognitive ability measured concurrently at age 21
and adjusted for cognitive ability at age 5, depression but not psychosis predicted lower
concurrent cognitive ability. With regards to psychosis, our findings largely confirm those
from a recent investigation using the Dunedin cohort (Meier et al., 2014). The authors found
that while a range of cognitive and mental functions were found to decline from childhood to
adulthood in schizophrenics, verbal IQ deficits emerged early and did not worsen thereafter.
Our study add to this evidence by showing that after accounting for mania and psychosis, this
relationship was reversed in depression, with no apparent deficits in childhood but worsened
cognitive ability having emerged by early adulthood.
While behaviour problems have been linked with depression and psychosis in many previous
studies including analyses based on the MUSP cohort (Erlenmeyer-Kimling et al., 2000,
Gooding et al., 2013, Matheson et al., 2013, Scott et al., 2009), the similar predictive strength
on depression and psychosis and the absence of an association with mania were surprising. It
seems that broad-band behaviour problems (which in this study included 10 items each of (i)
internalising, (ii) externalising and (iii) social/attention/thought problems) are a common
source of later depression and psychosis but not mania. If mania does not manifest in
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
16
measureable behavioural abnormalities in early childhood this may limit efforts to formulate
effective intervention strategies. On the other handit may simply be that the indicators of
behaviour problems we collected at age 5 are not specific to the paediatric temperamental
markers predictive of later disorders characterised by mania (Luby and Navsaria, 2010), and
further research is needed into the specific early behavioural abnormalities which predict
mania as opposed to bipolar depression (Cuellar et al., 2005).
The exclusive relationship between developmental delay and mania diverges from the
previous literature. In this and similar research developmental delay is understood to be an
early prodromal marker for serious psychiatric disorders in later life, rather than the result of
an environmental exposure such as family poverty. As such, the manifestation of a prodromal
marker would be unlikely to greatly distinguish between two later psychiatric disorders which
share similar genetic and biological origins. It has been demonstrated that the variance in
bipolar and psychotic disorders contributed by SNPs is highly correlated (Consortium, 2013a)
and both disorders share similar underlying biological pathways from gene to phenotype
(Consortium, 2013b). Despite this, as with our findings most previous research shows that
early neurodevelopmental factors do in fact differently predict these disorders, albeit in a
manner opposite to ours.The majority of previous studies show thatlanguage, motor function
and social delaysare linked with psychotic disorders and not bipolar disorders(Arango et al.,
2013, Dickson et al., 2012, Murray et al., 2004). While previous evidencehas highlighted an
association between these early developmental delays and paediatric bipolar disorders
(Sigurdsson et al., 1999), the study was relatively small-scale in comparison to other studies.
We ran three simple post-hoc univariate logistic regression analyses, using developmental
delays to predict the diagnoses of DSM-IV major depression, bipolar disorders and psychotic
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
17
disorders, to see if it corroborated the findings from our SEM. The respective effect sizes and
95% confidence intervals were 1.01 (0.64, 1.58), 4.24 (1.89, 9.50) and 1.29 (0.30, 5.54);
aligning with our main findings. Thus, the relationship with developmental delay is robust
against the way the manic symptoms are operationalised and the statistical methods
employed, and in both approaches developmental delay is strongly associated with mania/
bipolar disorders and not with depression or psychosis.We understand the need for caution
when disseminating our finding which is of a contradictory nature when compared with the
overwhelming body of existing literature. Nonetheless, the association was robust and not
without precedent (Sigurdsson et al., 1999), and can serve as an important reminder that we
efforts to assign specific neurodevelopmental pathways to specific adult psychiatric outcomes
remains to be settled.
With regards to the non-significant relationship between developmental abnormalities and
later psychosis,it is possible that the developmental delays may not be predictive ofthe
positive symptoms of psychosis, which were also largely subthreshold symptoms in our
sample with few participants receiving a clinical diagnosis. A diagnosis of schizophrenia
includes negative psychotic symptoms which we did not include in our study and which may
be more closely aligned with developmental delays(Collip et al., 2013).Positive symptoms
and particularly hallucinations may be more strongly related to environmental exposures such
as childhood trauma (Daalman et al., 2012).The inability to account for the negative
symptoms, which mark the chronic profile and are also the major causes of disability and
impairment in schizophrenia, is a limitation of our methodology. However, the positive
symptoms are the predominate characteristic of clinical presentations(Salloum and Mezzich,
2009), and thus comprise an important characteristic of psychotic disorders to understand in
the general population.A recent finding also suggests that while psychotic like experiences
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
18
are a valid predictor of a range of mental health problems, their use as a proxy to investigate
the etiology of schizophrenia may be less valid than previously though (Kounali et al., 2014).
Lastly, the age of the participants may have played a role in the null-finding, with a later age
of psychiatric assessment needed to identify an association between developmental delay and
psychosis. Our team will soon be in a position to test whether this is the case using the newly-
collected MUSP measures at offspring age 30 years.
The main strength of this study was the unique methodological approach adopted. A number
of recent studies support the importance of using spectrums to investigate etiological factors
among these related disorders, in addition to the need to separate mania from depression
(Cuellar et al., 2005, Hickie, 2013, Johnson et al., 2011, Merikangas et al., 2012).Our
structural equation model approach allowed us to make a number of unique observations
concerning the relationship among neurodevelopmental abnormalities and mania, depression
and psychosis, and we encourage further research of this type. We also found some
interesting relationships among the spectrums and outcomes at age 21. Depression predicted
all three anxiety disorders, while mania predicted just panic disorder and psychosis just social
phobias. Mania was unique in predicting alcohol use and the strongest predictor of tobacco,
while with regard to attachment styles psychosis uniquely associated with relationships as of
secondary importance, and not feeling the need for approval.
There has been growing interest within psychiatric research to utilise symptoms-level data to
construct what may be a more accurate representation of the phenotype of psychopathology
in efforts to both understand comorbidity and identify etiological factors(Kramer et al., 2008,
Krueger and Markon, 2006, Kushner et al., 2013, Markon, 2010, Simms et al., 2012).
However, with this approach comes the disadvantage of moving away from the diagnostic
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
19
categories to which clinically meaningful impairment and distress are inherent. So although
we believe research such as ours has the potential to contribute to psychiatric nosology and
etiology in a manner previously outlined, our findings should not be understood as directly
generlisable to clinical diagnoses. In addition to this consideration,our study has also some
limitations. Firstly, although validated against other standard intelligence tests for children
and considered a strong indicator of verbal intelligence, the PPVT-R is a less comprehensive
measure of intelligence than tests such as the Wechsler Intelligence Scale for Children-
Revised (WISC-R). However, verbal intelligence does appear to precipitate psychotic
disorders as opposed to other measures of IQ, some of which are found to worsen along with
the development of psychotic symptoms(Zammit et al., 2004). So while future studies with a
greater capacity to measure cognitive ability are needed, verbal ability appears to be a good
predictor of future psychotic disorders and experiences. Secondly, we were unable to explore
associations between more specific and more abnormal developmental profiles of the DDST
and the outcomes due to low numbers. This was due to using item level data for the
psychiatric outcomesand the needto satisfy the requirements of the covariance matrix in the
resulting structural equation model (i.e., no zero cells in bivariate relationships). Although
from a public health perspective our approach had the benefit of relating a general
developmentally delayed profile in early childhood with adult mania, the use of more detailed
delayed developmental profiles may have identified more specific relationships with the
outcomes.
Thirdly, in our analyses we made no correction for multiple comparisons. Importantly, it has
been persuasively argued that adjustment for multiple comparisons may not be necessary in
exploratory epidemiological research (Rothman, 1990), particularly when investigating
relationships which are logical, produce strong effect estimates and have been linked in
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
20
previous research (Feise, 2002). Indeed, our analysis was exploratory and the relationships
we tested have all been found in previous studies described above, as the novelty in our study
was not the establishment of new risk factors but employing a new conceptualisation of
psychopathology. However, it is important to acknowledge that some of the associations we
identified, particularly those with small effect sizes and borderline p-values, could have
resulted from chance.Lastly our cohort was subject to considerable attrition. Our attrition
analysis indicated that those lost to follow-up were no more or less likely to have a
developmental delay or behaviour problem, and had on average only a one point higher
cognitive ability score.In addition, the inverse probability weighting analysis showedvery
similar results to the main analysis, thus our findings are unlikely to besubstantively biased
by attrition.
Our study employed a novel strategy to provide a uniquepicture of the neurodevelopmental
similarities and differences underlying disorders characterised by depression, mania and
psychosis. In particular, our findings suggest that manic symptoms may be associated with
neurodevelopmental factors that are not associated with the depressive symptoms they often
accompany. This has important implications for research into bipolar disorders, as it may
suggest that depressive and manic symptoms are not simply the opposite end of the same
pole. We also found a robust relationship between developmental delays in childhood and
later symptoms of mania, which we presented cautiously as it contradicts existing studies
which place such delays as a specific prodrome of psychotic disorders. Being the first of its
kind, we encourage future studies to employ similar spectrum-based analysis to confirm our
findings and to further develop and improve upon a methodology which will in the future will
likely take a greater role in identifying the etiological factors of psychiatric disorders (Owen,
2014).
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
22
Funding and finance
Conflict of interest: All authors declare no conflicts of interest
Funding/Support: This study was funded by the National Health and Medical Research
Council (NHMRC). R.A. is funded by a NHMRC Career Development Award Level 2 in
Population Health (APP1012485).
Contributors: Author one (see order of authors below under affiliations) has been designated
principal author and was responsible for the bulk of the literature review, drafting, statistical
analysis and discussion. Author 2 assisted with the design of the analysis, provided statistical
consultancy and provided support in the interpretation of statistical findings. Author 3 is the
principal investigator of the MUSP cohort and helped with numerous revisions of the
manuscript. Author 4 contributed substantially to the drafting and revision of the manuscript
and assisted in the literature review and methodology.
Acknowledgements
Additional Contributions: The authors thank the MUSP team, MUSP participants, the Mater
Misericordiae Hospital, and the Schools of Social Science, Population Health and Medicine
(University of Queensland).
Ethical Standards: Informed consent from all participants was gained, all data was coded
for confidentiality and ethics was approved for the cohort by the institution and funding body.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
23
References
Achenbach, T. M. (1991). Manual for the Child Behavior Checklist 4/18 and 1991 Profile. University of Vermont Department of Psychiatry: Burlington, VT. Arango, C., Fraguas, D. & Parellada, M. (2013). Differential Neurodevelopmental Trajectories in Patients With Early-onset Bipolar and Schizophrenia Disorders. Schizophrenia bulletin, sbt198. Barnett, J. H., McDougall, F., Xu, M. K., Croudace, T. J., Richards, M. & Jones, P. B. (2012). Childhood cognitive function and adult psychopathology: associations with psychotic and non-psychotic symptoms in the general population. The British Journal of Psychiatry201, 124-130. Betts, K. S., Williams, G. M., Najman, J. M., Scott, J. & Alati, R. Maternal prenatal infection, early susceptibility to illness and adult psychotic experiences: A birth cohort study. Schizophrenia Research. Betts, K. S., Williams, G. M., Najman, J. M., Scott, J. & Alati, R. (2014). Maternal prenatal infection, early susceptibility to illness and adult psychotic experiences: A birth cohort study. Schizophrenia research. Blanchard, M. M., Jacobson, S., Clarke, M. C., Connor, D., Kelleher, I., Garavan, H., Harley, M. & Cannon, M. (2010). Language, motor and speed of processing deficits in adolescents with subclinical psychotic symptoms. Schizophrenia research123, 71-6. Bor, W., Najman, J. M., Andersen, M. J., O'Callaghan, M., Williams, G. M. & Behrens, B. C. (1997). The relationship between low family income and psychological disturbance in young children: an Australian longitudinal study. Aust N Z J Psychiatry31, 664-75. Cannon, M., Caspi, A., Moffitt, T. E., Harrington, H., Taylor, A., Murray, R. M. & Poulton, R. (2002). Evidence for early-childhood, pan-developmental impairment specific to schizophreniform disorder: results from a longitudinal birth cohort. Archives of general psychiatry59, 449-56. Cederlöf, M., Östberg, P., Pettersson, E., Anckarsäter, H., Gumpert, C., Lundström, S. & Lichtenstein, P. (2013). Language and mathematical problems as precursors of psychotic-like experiences and juvenile mania symptoms. Psychological medicine, 1-10. Childers, J. S., Durham, T. W. & Wilson, S. (1994). Relation of performance on the Kaufman Brief Intelligence Test with the Peabody Picture Vocabulary Test--Revised among preschool children. Percept Mot Skills79, 1195-9. Chotai, J., Jonasson, M., Hägglöf, B. & Adolfsson, R. (2005). Adolescent attachment styles and their relation to the temperament and character traits of personality in a general population. European Psychiatry20, 251-259. Collip, D., Wigman, J. T., Lin, A., Nelson, B., Oorschot, M., Vollebergh, W. A., Ryan, J., Baksheev, G., Wichers, M. & van Os, J. (2013). Dynamic association between interpersonal functioning and positive symptom dimensions of psychosis over time: a longitudinal study of healthy adolescents. Schizophrenia bulletin39, 179-185. Consortium, C.-D. G. o. t. P. G. (2013a). Genetic relationship between five psychiatric disorders estimated from genome-wide SNPs. Nature genetics45, 984-994. Consortium, C.-D. G. o. t. P. G. (2013b). Identification of risk loci with shared effects on five major psychiatric disorders: a genome-wide analysis. The Lancet381, 1371-1379. Cuellar, A. K., Johnson, S. L. & Winters, R. (2005). Distinctions between bipolar and unipolar depression. Clinical psychology review25, 307-339. Daalman, K., Diederen, K. M. J., Derks, E. M., van Lutterveld, R., Kahn, R. S. & Sommer, I. E. (2012). Childhood trauma and auditory verbal hallucinations. Psychological medicine42, 2475-2484. Dickson, H., Laurens, K., Cullen, A. & Hodgins, S. (2012). Meta-analyses of cognitive and motor function in youth aged 16 years and younger who subsequently develop schizophrenia. Psychol Med42, 743-755.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
24
Dunn, L. (1981). Peabody Picture Vocabulary Test-Revised: Circle Lines. American Guidance Service: MN. Erlenmeyer-Kimling, L., Rock, D., Roberts, S. A., Janal, M., Kestenbaum, C., Cornblatt, B., Adamo, U. H. & Gottesman, II (2000). Attention, memory, and motor skills as childhood predictors of schizophrenia-related psychoses: the New York High-Risk Project. The American journal of psychiatry157, 1416-22. Feeney, J. A., Noller, P. & Hanrahan, M. (1994). Assessing adult attachment. Feise, R. J. (2002). Do multiple outcome measures require p-value adjustment? BMC Medical Research Methodology2, 8. Frankenburg, W. K. & Dodds, J. B. (1967). The Denver developmental screening test. The Journal of pediatrics71, 181-191. Frankenburg, W. K., Dodds, J. B. & Fandal, A. W. (1973). Denver Developmental Screening Test: Manual/workbook for Nursing & Paramedical Personnel. University of Colorado Medical Center. Gale, C. R., Batty, G. D., McIntosh, A. M., Porteous, D. J., Deary, I. J. & Rasmussen, F. (2013). Is bipolar disorder more common in highly intelligent people? A cohort study of a million men. Molecular psychiatry18, 190-194. Goldberg, D., Andrews, G. & Hobbs, M. (2009). Where should bipolar disorder appear in the meta-structure? Psychological medicine39, 2071. Gooding, D., Ott, S., Roberts, S. & Erlenmeyer-Kimling, L. (2013). Thought disorder in mid-childhood as a predictor of adulthood diagnostic outcome: findings from the New York High-Risk Project. Psychological medicine43, 1003-1012. Hickie, I. (2013). Evidence for separate inheritance of mania and depression challenges current concepts of bipolar mood disorder. Molecular psychiatry. Hu, L. T. & Bentler, P. M. (1998). Fit indices in covariance structure modeling: Sensitivity to underparameterized model misspecification. Psychological Methods3, 424-453. Johnson, D. L., Howie, V. M., Owen, M., Baldwin, C. D. & Luttman, D. (1993). Assessment of three-year-olds with the Stanford-Binet Fourth Edition. Psychol Rep73, 51-7. Johnson, S., Morriss, R., Scott, J., Paykel, E., Kinderman, P., Kolamunnage‐Dona, R. & Bentall, R. (2011). Depressive and manic symptoms are not opposite poles in bipolar disorder. Acta Psychiatrica Scandinavica123, 206-210. Jongsma, E. A. (1982). Peabody Picture-Vocabulary Test - Revised (Ppvt-R) - Dunn,Lm, Dunn,Lm. Journal of Reading25, 360-364. Kelleher, I., Keeley, H., Corcoran, P., Lynch, F., Fitzpatrick, C., Devlin, N., Molloy, C., Roddy, S., Clarke, M. C., Harley, M., Arseneault, L., Wasserman, C., Carli, V., Sarchiapone, M., Hoven, C., Wasserman, D. & Cannon, M. (2012). Clinicopathological significance of psychotic experiences in non-psychotic young people: evidence from four population-based studies. The British journal of psychiatry : the journal of mental science201, 26-32. Kounali, D., Zammit, S., Wiles, N., Sullivan, S., Cannon, M., Stochl, J., Jones, P., Mahedy, L., Gage, S. & Heron, J. (2014). Common versus psychopathology-specific risk factors for psychotic experiences and depression during adolescence. Psychological medicine44, 2557-2566. Kramer, M. D., Krueger, R. F. & Hicks, B. M. (2008). The role of internalizing and externalizing liability factors in accounting for gender differences in the prevalence of common psychopathological syndromes. Psychological Medicine38, 51-61. Krueger, R. F. & Markon, K. E. (2006). Reinterpreting comorbidity: A model-based approach to understanding and classifying psychopathology. Annual review of clinical psychology2, 111. Kushner, M., Krueger, R., Wall, M., Maurer, E., Menk, J. & Menary, K. (2013). Modeling and treating internalizing psychopathology in a clinical trial: a latent variable structural equation modeling approach. Psychological medicine43, 1611-1623. Linscott, R. J. & van Os, J. (2010). Systematic reviews of categorical versus continuum models in psychosis: evidence for discontinuous subpopulations underlying a psychometric continuum. Implications for DSM-V, DSM-VI, and DSM-VII. Annual review of clinical psychology6, 391-419.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
25
Luby, J. L. & Navsaria, N. (2010). Pediatric bipolar disorder: evidence for prodromal states and early markers. Journal of Child Psychology and Psychiatry51, 459-471. MacCabe, J. H., Lambe, M. P., Cnattingius, S., Sham, P. C., David, A. S., Reichenberg, A., Murray, R. M. & Hultman, C. M. (2010). Excellent school performance at age 16 and risk of adult bipolar disorder: national cohort study. The British Journal of Psychiatry196, 109-115. Markon, K. (2010). Modeling psychopathology structure: a symptom-level analysis of Axis I and II disorders. Psychological Medicine40, 273-288. Matheson, S. L., Vijayan, H., Dickson, H., Shepherd, A. M., Carr, V. J. & Laurens, K. R. (2013). Systematic meta-analysis of childhood social withdrawal in schizophrenia, and comparison with data from at-risk children aged 9–14 years. Journal of psychiatric research47, 1061-1068. Meier, M. H., Caspi, A., Reichenberg, A., Keefe, R. S., Fisher, H. L., Harrington, H., Houts, R., Poulton, R. & Moffitt, T. E. (2014). Neuropsychological decline in schizophrenia from the premorbid to the postonset period: evidence from a population-representative longitudinal study. American Journal of Psychiatry171, 91-101. Merikangas, K. R., Cui, L., Kattan, G., Carlson, G. A., Youngstrom, E. A. & Angst, J. (2012). Mania with and without depression in a community sample of US adolescents. Archives of general psychiatry69, 943-951. Merikangas, K. R., Jin, R., He, J.-P., Kessler, R. C., Lee, S., Sampson, N. A., Viana, M. C., Andrade, L. H., Hu, C. & Karam, E. G. (2011). Prevalence and correlates of bipolar spectrum disorder in the world mental health survey initiative. Archives of general psychiatry68, 241-251. Murray, R. M., Sham, P., Van Os, J., Zanelli, J., Cannon, M. & McDonald, C. (2004). A developmental model for similarities and dissimilarities between schizophrenia and bipolar disorder. Schizophrenia research71, 405-416. Muthén, L. K., & Muthén, B.O. (1998-2010). Mplus user's guide. Muthén & Muthén: Los Angeles. Najman, J. M., Bor, W., O'Callaghan, M., Williams, G. M., Aird, R. & Shuttlewood, G. (2005). Cohort Profile: The Mater-University of Queensland Study of Pregnancy (MUSP). Int J Epidemiol34, 992-7. Owen, M. J. (2014). New Approaches to Psychiatric Diagnostic Classification. Neuron84, 564-571. Payá, B., Rodríguez-Sánchez, J. M., Otero, S., Muñoz, P., Castro-Fornieles, J., Parellada, M., Gonzalez-Pinto, A., Soutullo, C., Baeza, I. & Rapado-Castro, M. (2013). Premorbid impairments in early-onset psychosis: Differences between patients with schizophrenia and bipolar disorder. Schizophrenia research146, 103-110. Polanczyk, G., Moffitt, T. E., Arseneault, L., Cannon, M., Ambler, A., Keefe, R. S., Houts, R., Odgers, C. L. & Caspi, A. (2010). Etiological and clinical features of childhood psychotic symptoms: results from a birth cohort. Archives of general psychiatry67, 328-338. Reichenberg, A., Weiser, M., Rabinowitz, J., Caspi, A., Schmeidler, J., Mark, M., Kaplan, Z. & Davidson, M. (2002). A population-based cohort study of premorbid intellectual, language, and behavioral functioning in patients with schizophrenia, schizoaffective disorder, and nonpsychotic bipolar disorder. American Journal of Psychiatry159, 2027-2035. Rothman, K. J. (1990). No adjustments are needed for multiple comparisons. Epidemiology1, 43-46. Saha, S., Chant, D., Welham, J. & McGrath, J. (2005). A systematic review of the prevalence of schizophrenia. PLoS medicine2, e141. Salloum, I. M. & Mezzich, J. E. (2009). Psychiatric diagnosis: challenges and prospects. Wiley. com. Scott, J., Martin, G., Welham, J., Bor, W., Najman, J., O’Callaghan, M., Williams, G., Aird, R. & McGrath, J. (2009). Psychopathology during childhood and adolescence predicts delusional-like experiences in adults: a 21-year birth cohort study. American Journal of Psychiatry166, 567-574. Sigurdsson, E., Fombonne, E., Sayal, K. & Checkley, S. (1999). Neurodevelopmental antecedents of early-onset bipolar affective disorder. The British Journal of Psychiatry174, 121-127. Simms, L., Prisciandaro, J., Krueger, R. & Goldberg, D. (2012). The structure of depression, anxiety and somatic symptoms in primary care. Psychological medicine42, 15-28. Slade, T., Johnston, A., Oakley Browne, M. A., Andrews, G. & Whiteford, H. (2009). 2007 National Survey of Mental Health and Wellbeing: methods and key findings. Aust N Z J Psychiatry43, 594-605.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
26
Twaite, J. A. & Rodriguez-Srednicki, O. (2004). Childhood sexual and physical abuse and adult vulnerability to PTSD: The mediating effects of attachment and dissociation. Journal of Child Sexual Abuse13, 17-38. van Os, J., Linscott, R. J., Myin-Germeys, I., Delespaul, P. & Krabbendam, L. (2009). A systematic review and meta-analysis of the psychosis continuum: evidence for a psychosis proneness-persistence-impairment model of psychotic disorder. Psychological medicine39, 179-95. Varghese, D., Scott, J. G., Bor, W., Williams, G. M., Najman, J. M. & McGrath, J. J. (2013). The association between adult attachment style and delusional-like experiences in a community sample of women. The Journal of nervous and mental disease201, 525-529. Welham, J., Scott, J., Williams, G., Najman, J., Bor, W., O’Callaghan, M. & McGrath, J. (2010). The antecedents of non‐affective psychosis in a birth‐cohort, with a focus on measures related to cognitive ability, attentional dysfunction and speech problems. Acta Psychiatrica Scandinavica121, 273-279. World Health Organization (1997). CIDI-Auto version 2.1: administrator's guide and reference. Training and reference centre for WHO CIDI: Sydney. Zammit, S., Allebeck, P., David, A. S., Dalman, C., Hemmingsson, T., Lundberg, I. & Lewis, G. (2004). A longitudinal study of premorbid IQ Score and risk of developing schizophrenia, bipolar disorder, severe depression, and other nonaffective psychoses. Arch Gen Psychiatry61, 354-60.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
27
Table 1: Prevalence and standardised factor loadings (with standard errors) and fit
indices for the 3 factor (correlated) models of lifetime symptoms of mania, depression
and psychosis at 21 years (n = 2,566)
Symptoms Items from CIDI-Auto 2.1 Prevalence 3 Factor CFA (correlated factors)
% (n) Mania Depression Psychosis
Mania
much more active 2.8% (72) 0.84 (0.03)
unable to sit still 6.2% (160) 0.94 (0.01)
spend too much money 2.9% (75) 0.81 (0.03)
increased interest in sex 2.2% (57) 0.83 (0.03)
less careful about sexual activities 1.6% (40) 0.81 (0.03)
very talkative 4.2% (108) 0.88 (0.02)
thoughts raced through your head 7.3% (186) 0.95 (0.01)
did something normally would be ashamed of 5.7% (145) 0.90 (0.02)
overly friendly 4.0% (102) 0.87 (0.02)
hardly slept but didn’t feel tired 5.1% (131) 0.91 (0.02)
easily distracted 8.6% (220) 0.95 (0.01)
Depression
lacking energy 32.3% (825) 0.95 (0.01)
gain or loss in appetite or weight 33.3% (853) 0.95 (0.01)
insomnia or hypersomnia 35.9% (921) 0.97 (0.00)
psychomotor retardation or agitation 16.4% (420) 0.84 (0.01)
felt worthless or guilty 24.4% (626) 0.92 (0.01)
lacked confidence 29.2% (749) 0.93 (0.01)
trouble thinking 39.1% (1002) 0.99 (0.00)
thoughts of death and suicide 17.0% (436) 0.82 (0.01)
Psychosis
spied on 1.8% (46) 0.77 (0.04)
followed 2.0% (51) 0.76 (0.04)
people discussing you 1.4% (37) 0.57 (0.07)
tested on/plotted against 1.1% (28) 0.73 (0.05)
sent messages via media 1.1% (28) 0.70 (0.06)
experienced mind reading 1.6% (40) 0.67 (0.06)
hear others' thought 3.4% (86) 0.73 (0.04)
others' hear your thoughts 1.1% (28) 0.78 (0.04)
manipulated by external force 1.3% (32) 0.75 (0.06)
visual hallucinations 7.7% (196) 0.74 (0.04)
auditory hallucinations 4.3% (111) 0.76 (0.03)
voice hearing 2.2% (55) 0.87 (0.03)
olfactory hallucinations 4.1% (104) 0.64 (0.05)
gustatory hallucinations 4.7% (120) 0.60 (0.05)
tactile hallucination 8.8% (225) 0.76 (0.04)
Note:Parameters were derived using the WLSMV estimator and all factor loadings and factor variances were
significant(CFI = 0.99; TLI = 0.99, RMSEA = 0.018; chi-2 = 941.22; free parameters = 71).
Factor correlates: mania with depression = 0.49 (p <0.001); mania with psychosis = 0.51 (p<0.001); depression
with psychosis = 0.50 (p<0.001).
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
28
Table 2: Univariate and multivariate associations of the risk factors and confounders with spectrums of mania, depression and psychosis
at 21 years (n = 1,934)
Mania Depression Psychosis
Univariate
USPE SE p-Value SPE USPE SE p-Value SPE USPE SE p-Value SPE
PPVT-R cognitive ability 0.00 0.00 0.673 -0.02 0.00 0.00 0.441 -0.02 -0.01 0.00 0.029 -0.08
DDST delayed development 0.31 0.11 0.007 0.37 0.12 0.11 0.235 0.13 0.06 0.12 0.631 0.07
CBCL behaviour problems -0.03 0.11 0.780 -0.01 0.27 0.08 0.001 0.28 0.25 0.09 0.007 0.31
offspring gender 0.03 0.06 0.689 0.02 0.03 0.05 <0.001 0.37 0.07 0.06 0.196 0.09
offspring age 0.02 0.04 0.647 0.02 0.04 0.03 0.235 0.03 0.00 0.04 0.961 0.00
maternal education 0.08 0.06 0.156 0.06 0.07 0.04 0.127 0.04 0.09 0.05 0.062 0.06
Multivariate
PPVT-R cognitive ability 0.00 0.00 0.956 0.00 0.00 0.00 0.520 -0.02 -0.01 0.00 0.038 -0.08
DDST delayed development 0.32 0.12 0.006 0.38 0.17 0.12 0.114 0.17 0.01 0.12 0.929 0.01
CBCL behaviour problems -0.04 0.11 0.678 -0.05 0.28 0.08 <0.001 0.29 0.25 0.10 0.011 0.30
offspring gender 0.05 0.07 0.486 0.06 0.37 0.05 <0.001 0.38 0.09 0.06 0.129 0.11
Note: Estimates presented as Unstandardised Parameter Estimates (USPE), Standard Errors (SE) and p-Values, and Standardised Parameter Estimates (SPE).
All factor loadings and factor variances were significant in the multivariate model (CFI = 0.99; TLI = 0.99, RMSEA = 0.018; chi-2 = 1032.30; free parameters = 83).
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
29
Figure 1: Structural equation model showing the associations among the risk factors and gender with the latent
factors of mania, depression and psychosis (curved arrows show the latent factors are correlated). Effect sizes
are standardised paramter estimates (with p-values and standard errors included in table 2).
Mania (21 yrs)
Depression
(21 yrs)
Psychosis (21 yrs)
Delayed development (5 yrs)
Offspring gender
Cognitive ability (5 yrs)
Behaviour problems (5 yrs)
0.38
0.38
0.29
0.30
-0.08
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
30
Table3: Multivariate attrition analysis comparing those included in the analysis (n =
1,934) versus those lost to follow-up but with values on childhood
neurodevelopmental variables (n = 2,006) [expressed in OR with 95%
Confidence Intervals (CI)] (total n = 3,940)
Effect OR (95% CI) P-value
Offspring gender ref: male
Female 0.78 (0.69, 0.89) <0.001
PPVT-R cognitive ability 0.98 (0.98, 0.99) <0.001
DDST delayed development 0.94 (0.71, 1.22) 0.632
CBCL behaviour problems 1.12 (0.92, 1.36) 0.274
Note: Showing the odds of not being included in the study by childhood neurodevelopmental factors.
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Highlights
Attention Editors
Re: paper ‘Predicting spectrums of adult mania, psychosis and depression by prospectively
ascertained childhood neurodevelopment’.
We used SEM to specify spectrums of mania, depression and psychosis.
Mania and depression were predicted by different neurodevelopmental factors
A robust relationship between developmental delay and mania was identified
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Acknowledgment
Attention Editors
Re: paper ‘Predicting spectrums of adult mania, psychosis and depression by prospectively
ascertained childhood neurodevelopment’.
Additional Contributions: The authors thank the MUSP team, MUSP participants, the Mater
Misericordiae Hospital, and the Schools of Social Science, Population Health and Medicine
(University of Queensland).
Dr. Kim Betts (Corresponding author)
School of Population Health
The University of Queensland
4th
floor, Public Health Building
Herston Rd, Herston QLD 4006
Australia
phone: +617 33655509
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Conflict of Interest
Attention Editors
Re: Predicting spectrums of adult mania, psychosis and depression by prospectively
ascertained childhood neurodevelopment’.
All authors declare no conflicts of interest exist.
Dr. Kim Betts (Corresponding author)
School of Population Health
The University of Queensland
4th
floor, Public Health Building
Herston Rd, Herston QLD 4006
Australia
phone: +617 33655509
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Contributors
Attention Editors
Re: paper ‘Predicting spectrums of adult mania, psychosis and depression by prospectively
ascertained childhood neurodevelopment’.
Author one (see order of authors below under affiliations) has been designated principal
author and was responsible for the bulk of the literature review, drafting, statistical analysis
and discussion. Author 2 assisted with the design of the analysis, provided statistical
consultancy and provided support in the interpretation of statistical findings. Author 3 is the
principal investigator of the MUSP cohort and helped with numerous revisions of the
manuscript. Author 4 contributed substantially to the drafting and revision of the manuscript
and assisted in the literature review and methodology.
Dr. Kim Betts (Corresponding author)
School of Population Health
The University of Queensland
4th
floor, Public Health Building
Herston Rd, Herston QLD 4006
Australia
phone: +617 33655509
MANUSCRIP
T
ACCEPTED
ACCEPTED MANUSCRIPT
Funding Source
Attention Editors
Re: paper ‘Predicting spectrums of adult mania, psychosis and depression by prospectively
ascertained childhood neurodevelopment’.
This study was funded by the National Health and Medical Research Council (NHMRC).
R.A. is funded by a NHMRC Career Development Award in Population Health (ID 519721).
We acknowledge the support of the Australian Health Management (AHM) in the review and
preparation of this manuscript. The funding sources did not have any influence over the
analyses or interpretation presented in our paper.
Dr. Kim Betts (Corresponding author)
School of Population Health
The University of Queensland
4th
floor, Public Health Building
Herston Rd, Herston QLD 4006
Australia
phone: +617 33655509