Secondary Schizophrenia (76 page)

stress being contributory, and the lack of generalizabil-A similar case-control study by Fujii and Ahmed
ity from a special population.

[16]
included 25 cases of PDTBI and 25 controls, and
A study from Belgium
[19]
retrospectively studied
reported that those with psychosis had history of prior
530 brain-injured patients who had been assessed at
head injury or evidence of neurological disorder, sug-a neuropsychological center and followed up for 1–10

gesting brain damage as a source of vulnerability. The
years. Eighteen were diagnosed with a major psychi-small size of the study and the broad and heteroge-atric disorder, mostly (83%) within 6 months of the
neous nature of brain damage make the results incon-injury, of which 12 (2.3%) had DSM-III schizophrenia
clusive. Moreover, the authors excluded patients who
or schizophreniform disorder. The authors reported
had a family history of psychosis, thereby making it
high rates of premorbid psychopathology and sug-impossible to examine an interaction with genetic vul-gested that the injury aggravated or precipitated a pre-nerability.

existing condition rather than caused it.

Some of the findings of these studies in relation to
Roberts
[20]
attempted to follow up 479 patients
the nature of the psychosis will be discussed later.

with severe head injury (post-traumatic amnesia > 1

week) admitted to the Radcliffe Infirmary, Oxford,
after 10–24 years. Of the 291 survivors with sufficient
Cohort studies

clinical data, 7 had a diagnosis of “paranoid demen-The earliest attempt to review the cohort studies was
tia” and 2 of SLP (9/291
=
3.09%). In the 2 with SLP,
by Davison and Bagley
[17]
who examined 8 stud-the onset was 9 and 17 years postinjury, and affective
ies published between 1917 and 1964, 7 of which
symptoms were prominent.

were in war veterans, with follow-up periods of 10 to
There have been some other noteworthy studies.

20 years (except for 2 studies with 2 years or less).

Ota
[21]
studied 1,168 adults who were admitted to
The cumulative incidence of schizophrenia in the TBI
hospital following TBI and reported psychosis, which
subjects ranged from 0.07% to 9.8%, with the median
could be either “functional” or “organic”, in 2.7%.

figure of 1.35%. Using a comparison figure of 0.8%

Miller and Stern
[22]
, in a long-term follow-up of 100

189

incidence over 25 years in the general population, the
TBI subjects, found 10 with psychosis, all of whom
Organic Syndromes of Schizophrenia – Section 3

Table 13.1
Association between traumatic brain injury (TBI) and schizophrenia-like psychosis (SLP): epidemiological evidence
Duration of

Authors

N

follow-up

Risk

Comment

Case Reports

Fujii & Ahmed, 2002 [5]

69

–

–

Anecdotal reports

Cross-sectional Studies

Silver et al., 2001 [6]

361

–

1.8

3.4% vs. 1.9%

Case control studies

Wilcox & Nasrallah, 1987 [9]

200 (controls
=
137)
–
OR 16.6
∗∗
(95% CI 2.6–689)
TBI < 10 years

Nielsen et al., 2002 [11]

8,288

–

OR 0.94/0.89 (0.71
∗
, corrected)
N.S. (corrected for

fractures)

Harrison et al., 2006 [12]

748 (204 controls)

–

OR1.10 (CI 0.82–1.47) [OR1.37

Nonschizophrenic,

(CI1.14–1.66)]

nonaffective psychosis

Corcoran & Malaospina, 2001 [13]

581

–

OR 4.27 (CI 1.4–3.0)

Malik et al., 2003 [14]

23 families

OR 2.34 (CI 1.03–5.03)

TBI < 10 years

suffered from dementia. A study by Lishman
[23],

small, cannot be dismissed, the clinical characteristics
using contemporary diagnostic criteria, identified only
of putative post-traumatic SLP to determine if it is dis-5 patients with schizophrenia-like illness in 670

tinct from schizophrenia in its phenomenology, clini-patients with penetrating head injury.

cal associations, and prognosis will now be examined.

A recent historical cohort study is of interest
[24].

It included 939 adults diagnosed with TBI in a emer-

Clinical presentation

gency department during 1993 and searched their
This description is based primarily on the review by
medical records for the period of 1 year before and 3

Davison and Bagley
[17]
and 2 recent reports of SLP-years following the injury for psychiatric diagnoses or
TBI
[5, 25],
1 of which
[5]
is an analysis of 69 published
treatment. An unexpected finding of the study was that
cases.

patients with TBI had a higher rate of pre-existing psy-The majority of the subjects are young, with a male
chosis (OR 10.0 for moderate to severe TBI, and 1.7

preponderance in the case reports. The mean age of
for mild TBI). In those with no pre-existing psychi-onset of psychosis was 26.3 (SD 10.2) years and 33.4

atric disorder (n
=
85), the risk of psychosis was sig-

(SD 15.4) years in the 2 studies, with 80% and 90%

nificantly elevated by moderate to severe head injury
being men. Because TBI is more common in young
in the second year (OR 5.9, CI 1.6 to 22.1) and third
men, it is uncertain whether the SLP rates merely
year (OR 3.6, CI 1.0 to 1.3) of follow-up. The limita-reflect this. It has been suggested
[5]
that men may
tions of this study include the reliance on case notes
be over-represented after accounting for the base rates
for diagnosis, the short preinjury period covered, and
for TBI, but this is far from established. In the large
the relatively small sample size.

Danish study, the authors examined the effect of sex
The diversity of the evidence discussed earlier
and found that the mean interval between the TBI and
makes for a difficult synthesis, but the conservative
the development of psychosis was 4.6
[15]
and 4.1
[5]

conclusion is that population-based studies suggest
years in the 2 reports, with a wide range (0–34 years).

that civilian TBI leads to only a small increase in the
Although most civilian TBI is due to motor vehicle
risk of SLP if at all, and the evidence for this is not con-accidents; assaults, gunshot injuries, and falls are all
sistent. The association may be stronger in those with
represented.

a genetic vulnerability to schizophrenia or the presence of preinjury psychopathology. War-related brain
injury may be a special case that cannot be generalized
Characteristics of head injury

to civilian populations, and the reported increase of
The TBI is more likely to be closed, but an open
risk may well be due to psychological rather than neu-injury is not necessarily protective against future SLP,
190

rological factors. Because the increase in risk, albeit
as has sometimes been suggested
[18].
The severity
Chapter 13 – Schizophrenia-like psychosis and traumatic brain injury

of the injury is usually moderate to severe intensity.

predominantly a positive syndrome, with only 22%

The severity of the head injury, by neuroimaging and
and 15% of patients in the two studies demonstrating
neuropsychological criteria, was greater in the SLP-negative symptoms such as flattening of affect, avoli-TBI group than the TBI group without psychosis
[15].

tion, or asociality. Agitated and aggressive behavior
Although many SLP-TBI occur after head injury in
are common.

childhood, early head injury was not over-represented
in one large study
[15].
In about 40% cases in this
Diagnostic criteria

study
[15]
, the head injury was followed by personAs is apparent from the earlier description, the clin-ality or behavioral change, the main characteristics of
ical picture resembles schizophrenia, with the differ-which were impulsivity, aggressiveness, loss of social
ence being that like other secondary schizophrenias,
graces, moodiness, and, less commonly, apathy. There
the emphasis is on a paranoid-hallucinatory psychosis
was evidence for brain damage in the temporal, pari-with few negative symptoms. To meet the DSM-IV cri-etal, and frontal lobes, more often unilateral than bilat-teria for a Psychotic Disorder due to Traumatic Brain
eral, on the basis of neuroimaging, clinical, and neu-Injury, it is crucial to establish that i) the psychosis is
ropsychological data. The SLP-TBI subjects had more
a direct physiological consequence of a brain disor-widespread neuropsychological deficits than the conder, and ii) the psychosis is not better accounted for by
trol group
[15].

a primary psychiatric disorder such as schizophrenia.

The pathophysiology of SLP is discussed later, but the
Psychopathology

association between TBI and SLP cannot be considered proof of causality even though it is a recognized
Prodromal symptoms are common and include bizarre
association, psychosis as a consequence of brain injury
or antisocial behavior, social withdrawal, affective
is plausible, and many of the patients are not otherwise
instability, and deterioration in work, often lasting
vulnerable to schizophrenia.

for months. Depressive symptoms are often present at
the time of presentation but confusional symptoms at
onset are unusual.

Course

The psychosis is delusional-hallucinatory in
The long-term course of SLP-TBI is poorly studied.

nature. A range of delusional symptoms, similar to
Fujii and Ahmed
[5]
found some follow-up informa-that seen in schizophrenia, is present and includes
tion on 39 out of the 69 cases reviewed: of these, 25

first rank Schneiderian symptoms. In the study by
patients were reported to have improved, 11 did not
Sachdev and colleagues
[15]
, one or more delusions
improve, and 3 worsened. Sachdev and colleagues
[15]

were present in all subjects, with persecutory (55.5%),
did not follow patients systematically; but it is not
referential (22.2%), control (22.2%), grandiose (20%),
unusual for clinicians to encounter TBI related psy-and religious (15.4%) delusion being the most com-chosis that, despite its prominence of positive symp-mon types. Delusions of thought alienation, thought
toms, responds poorly to neuroleptic medication.

insertion, withdrawal, or broadcast were present in
six (13.3%) and somatic passivity in three (6.7%).

Risk factors and pathophysiology

The review by Fujii and Ahmed
[5]
emphasized the
presence of persecutory delusions. Organic themes,
Are there specific neuroanatomical substrates of SLP-TBI?

described by Cutting
[26],
were absent in the Sachdev
Psychosis following focal brain injury is a relatively
and colleagues
[15]
study. However, the review of
rare event, but several consistent observations relat-published cases
[5]
did find 5 cases with the Capgras
ing right brain dysfunction to delusional disorders
delusion and 3 cases each with reduplicative amne-have emerged. Anatomically, lesions of the temporo-sia, erotomania, and stealing. Delusions relating to
parietal region are associated with the highest fre-misidentification, stealing, or hiding that are prone
quency of lesion-related psychosis
[26, 27].
The data
to occur in dementia patients with psychosis, are not
on anatomical localization in relation to TBI have
generally seen in SLP-TBI. Hallucinations were pre-not been consistent, and no convincing theoretical
dominantly auditory or visual, with the former being
framework has emerged. In the study by Sachdev and
more likely. Formal thought disorder and catatonic
colleagues
[15],
the comparison of the two groups
191

features are usually absent. The psychosis is therefore
on neuroimaging data suggested greater damage in
Organic Syndromes of Schizophrenia – Section 3

the SLP group in the left temporal and right parietal
Malaspina
[13],
there was no increase in risk for
regions, but the differences were not significant after
schizophrenia after TBI in the members of bipolar
Bonferroni correction. The Fujii and Ahmed
[5]