Neuroradiolog y / Head and Neck Imaging • Original Research

Hoxworth et al.
Digital Subtraction Myelography of Spinal CSF Leaks

Neuroradiology/Head and Neck Imaging

Original Research

The Role of Digital Subtraction

Downloaded from www.ajronline.org by 129.67.118.203 on 05/12/19 from IP address 129.67.118.203. Copyright ARRS. For personal use only; all rights reserved

Myelography in the Diagnosis

and Localization of Spontaneous
Spinal CSF Leaks
Joseph M. Hoxworth1 OBJECTIVE. The objective of our study was to review the clinical utility of digital sub-
Terrence L. Trentman2 traction myelography for the diagnosis of spinal CSF leaks in patients with spontaneous in-
Amy L. Kotsenas 3 tracranial hypotension (SIH) and those with superficial siderosis.
Kent R. Thielen 3 MATERIALS AND METHODS. Procedure logs from 2007 to 2011 were reviewed to
Kent D. Nelson1 identify cases in which digital subtraction myelography was performed to diagnose spinal
CSF leaks. Electronic medical records were reviewed to obtain information regarding diag-
David W. Dodick 4
nosis and outcome. For patients to be included in the study, preprocedural spinal MRI had to
Hoxworth JM, Trentman TL, Kotsenas AL, Thielen show an extradural fluid collection spanning more than one vertebral level and postmyelo-
KR, Nelson KD, Dodick DW graphic CT had to confirm the presence of an active CSF leak. If digital subtraction myelog-
raphy successfully showed the site of the CSF leak, the location was documented.
RESULTS. Eleven patients (seven men and four women; mean age, 49.0 years) under-
went digital subtraction myelography during the study period. Six patients had SIH and five
patients had superficial siderosis. The extradural fluid collection on spinal MRI averaged a
length of 15.5 vertebral levels. Digital subtraction myelography successfully showed the site
of the CSF leak in nine of the 11 patients, and all of the dural tears were located in the tho-
racic spine between T3 and T11.
CONCLUSION. Digital subtraction myelography is a valuable diagnostic tool for the
localization of rapid spinal CSF leaks and should be considered in patients who are clinically
suspected to have a dural tear that is accompanied by a longitudinally extensive extradural
fluid collection on spinal MRI.

Keywords: digital subtraction myelography, dural tear,
spinal CSF leak, spontaneous intracranial hypotension,
superficial siderosis
DOI:10.2214/AJR.11.8238
Received November 16, 2011; accepted after revision
January 2, 2012.
1 CSF leak [3–6]. SIH prototypically manifests
Department of Radiology, Mayo Clinic, 5777 E Mayo
Blvd, Phoenix, AZ 85054. Address correspondence to
J. M. Hoxworth (Hoxworth.Joseph@mayo.edu).
2
Department of Anesthesiology, Mayo Clinic, Phoenix, AZ.
3
Department of Radiology, Mayo Clinic, Rochester, MN.
4
Department of Neurology, Mayo Clinic, Phoenix, AZ.
AJR 2012; 199:649–653
0361–803X/12/1993–649
© American Roentgen Ray Society
C
SF leaks along the spinal axis are
now recognized as a cause of both
spontaneous intracranial hypoten-
sion (SIH) and superficial siderosis
of the CNS [1, 2]. Although sharing spinal dural
defects as a cause, these entities have different
clinical presentations. The clinical syndrome
commonly referred to as SIH has a variety of
associated spine imaging findings, one of which
is an extradural fluid collection secondary to the
as an orthostatic headache and is potentially ac-
companied by a wide range of other neurologic
symptoms. In patients with SIH, the spontane-
ous spinal CSF leak culminates in decreased
buoyancy of the brain, which leads to down-
ward traction on pain-sensitive structures. In
contrast, superficial siderosis is caused by re-
current bleeding into the subarachnoid space
and results in subpial deposition of hemosid-
erin within the brain and spinal cord [7, 8].
Patients with superficial siderosis character-
istically come to clinical attention because of
adult-onset, slowly progressive sensorineural
hearing loss, gait ataxia, and cerebellar dysar-
thria. Similar to patients with SIH, a subset of
superficial siderosis patients also have a longi-
tudinally extensive extradural fluid collection
in the spinal canal that communicates with the
subarachnoid space through a dural defect, the
closure of which has been associated with
symptom stabilization or improvement [7, 9].
Although the precise mechanism of chronic
bleeding remains unclear, the link between
spinal CSF leak and superficial siderosis is
now well established.
Definitive treatment of a spinal CSF leak re-
quires closure of the dural tear. Closure is per-
formed surgically in most patients with superfi-
cial siderosis [7, 9]. When conservative medical
therapy or empirical lumbar epidural blood
patches fail to achieve durable symptom reso-
lution, a smaller number of SIH patients require
a localized epidural blood patch, a targeted fi-
brin glue injection, or surgery [1]. In these cas-
es, treatment requires that the precise site of the

AJR:199, September 2012 649

 Hoxworth et al.

Fig. 1—Gray lines show longitudinal extent of extradural fluid collections identified on preprocedural MRI of
C2 11 study patients who underwent digital subtraction myelography. Location of leak () is shown for patients in
whom site of CSF leak was ultimately identified. This was accomplished with digital subtraction myelography
C5 in 9 of 11 patients. Digital subtraction myelography was unsuccessful in localizing the CSF leak in two patients.
Patient 7 was unable to undergo additional evaluation because of declining health, so the site of CSF leak is
T1
not known. The CSF leak in patient 8 was localized with dynamic CT myelography after digital subtraction
myelography was unsuccessful.
Vertebral Level

T4
Downloaded from www.ajronline.org by 129.67.118.203 on 05/12/19 from IP address 129.67.118.203. Copyright ARRS. For personal use only; all rights reserved

T7

T10

L1

L4

S2
1 2 3 4 5 6 7 8 9 10 11
Patient No.

A B C
Fig. 2—54-year-old man with superficial siderosis and CSF leak at T3 from remote fracture.
A, Sagittal T2-weighted image of thoracic spine shows ventral extradural fluid collection (arrowheads) dorsally displacing dura. Because fat-suppressed imaging was not
available for review, this collection was confirmed to represent fluid through comparison with T1-weighted sequences (not shown).
B, Axial CT myelogram through T3 level confirms that ventral extradural fluid collection (asterisk) freely communicates with subarachnoid space. Dorsally displaced dura
(arrowheads) is visible as thin line.
C, Lateral digital subtraction myelographic image is annotated to designate normal flow of contrast material within subarachnoid space (black arrows), site of CSF leak
at T3 (white arrow), and contrast material spreading within extradural fluid collection (arrowheads). For reference, please note that left side and top of digital subtraction
myelographic image correspond to anterior and cephalad aspects of patient, respectively.

CSF leak be known. Although noninvasive im- techniques is typically needed to conclusively material injected into the subarachnoid space
aging of the spine with MRI may show features document the exact site of the dural defect. via lumbar puncture may egress from the the-
that indicate the presence of a CSF leak in both In patients with a dural tear accompanied by cal sac too rapidly to allow identification of the
cohorts, additional testing with myelographic a sizeable extradural fluid collection, contrast site of the dural tear. Although routine postmy-

650 AJR:199, September 2012

 Digital Subtraction Myelography of Spinal CSF Leaks
Downloaded from www.ajronline.org by 129.67.118.203 on 05/12/19 from IP address 129.67.118.203. Copyright ARRS. For personal use only; all rights reserved
Fig. 3—31-year-old woman with spontaneous
intracranial hypotension and CSF leak at T8–T9 from
transdural disk extrusion. Lateral digital subtraction
myelographic image is annotated to designate normal
flow of contrast material within subarachnoid space
(black arrows), site of CSF leak at T8–T9 (white
arrow), and contrast spreading within extradural
fluid collection (arrowheads). For reference, please
note that left side and top of digital subtraction
myelographic image correspond to anterior and
cephalad aspects of patient, respectively.
elographic CT usually confirms the presence
of a CSF leak in these cases, localization of
the dural tear may not be possible because the
extravasated contrast material in the epidural
space spans multiple vertebral levels. Dynamic
CT myelography was introduced in an attempt
to overcome this latter limitation, but the tem-
poral resolution of this technique remains in-
adequate in some cases [10]. Digital subtrac-
tion myelography was subsequently reported
as an alternative means of localizing these rap-
id spinal CSF leaks [11]. The goal of the cur-
rent study was to review the diagnostic utility
of digital subtraction myelography in a series
of patients with documented spinal CSF leaks.
Materials and Methods
This retrospective study was approved by our in-
stitutional review board. The procedure logs at our
institution for the period from November 1, 2007
(beginning with the index case for which digital sub-
traction myelography was developed), through July
31, 2011, were reviewed to identify cases in which
digital subtraction myelography was performed. All
of the patients underwent evaluation by a board-cer-
tified neurologist, and clinical notes were reviewed
AJR:199, September 2012 651
in the electronic medical record to obtain informa-
tion regarding diagnosis and follow-up. The patients
had undergone preprocedural MRI of the entire
spine that included sagittal and axial T2-weighted
sequences. A board-certified neuroradiologist re-
viewed the spinal MRI examinations to confirm the
presence of and determine the length of the extradu-
ral fluid collection. In addition to the presence of an
abnormal extradural fluid collection within the spi-
nal canal spanning more than one vertebral level on
preprocedural MRI, extradural contrast material had
to be seen on postmyelographic CT to confirm the
presence of an active CSF leak. Patients with con-
traindications for lumbar puncture or iodinated con-
trast administration were not permitted to undergo
digital subtraction myelography.
The digital subtraction myelographic procedure
has previously been described in detail [11]. As is
standard for myelography at our institution, nei-
ther IV conscious sedation nor general anesthe-
sia was necessary. In brief, the patient was placed
in the prone position on the fluoroscopy table. Ei-
ther through table tilt or elevation of the pelvis with
foam padding, the lumbar spinal canal was elevated
to a higher gravitational position relative to the cer-
vical and thoracic spine. Given that the entire spine
cannot be visualized simultaneously because of the
limited size of the image intensifier, the region of
greatest suspicion based on prior spinal MRI and
other cross-sectional imaging, if available, was tar-
geted. Specifically, most of the extradural fluid col-
lection was included in the imaging field with partic-
ular attention directed toward known predisposing
factors for dural tear such as transdural disk extru-
sion, a focally prominent osteophyte, a remote ver-
tebral compression fracture, and so on. Radiopaque
skin markers were placed to allow reliable postpro-
cedure vertebral numbering. Lumbar puncture was
performed in the usual fashion with a 20-gauge spi-
nal needle. While patient respiration was suspend-
ed and during the steady intrathecal infusion of ap-
proximately 12–16 mL of iopamidol (Isovue M 200,
Bracco Diagnostics) or iohexol (Omnipaque 300,
GE Healthcare), a digital subtraction acquisition at a
film rate of 3–6 frames per second was performed in
the lateral plane; the anteroposterior plane was add-
ed when biplane equipment was available. Local-
ization of the CSF leak with digital subtraction my-
elography was defined as identification of contrast
material clearly diverging from the normal contrast
stream within the subarachnoid space.
Descriptive statistics were calculated using Mi-
crosoft Excel 2011 for Macintosh (Apple).
Results
A total of 11 patients underwent digital sub-
traction myelography during the study period
and were included in the analysis. None of the
patients experienced procedure-related com-
plications. The cohort was composed of sev-
en men and four women with a mean age of
49.0 years (range, 28–75 years). Six of the pa-
tients met the criteria established in the sec-
ond edition of the International Classification
of Headache Disorders [12] for headache at-
tributed to spontaneous or idiopathic low CSF
pressure (i.e., SIH), whereas superficial sider-
osis had been diagnosed in the other five pa-
tients. The duration of symptoms ranged from
3 months to 7 years (mean ± SD, 31.5 ± 25.0
months). All patients had an abnormal ventral-
predominant extradural fluid collection on spi-
nal MRI spanning an average of 15.5 vertebral
levels (SD, 5.1 vertebral levels), and the site of
the CSF leak relative to each extradural fluid
collection is illustrated in Figure 1. Routine
postmyelographic CT uniformly confirmed
the presence of an active spinal CSF leak in
all patients through the clear communication
of contrast medium between the subarachnoid
space and extradural fluid (Fig. 2B).
The site of the CSF leak was visualized
on digital subtraction myelography as active
extravasation of contrast medium from the
thecal sac in nine of the 11 patients (Figs.
Fig. 4—57-year-old man with superficial siderosis
and CSF leak at T6–T7 from focal disk osteophyte.
Lateral digital subtraction myelographic image is
annotated to designate normal flow of contrast
material within subarachnoid space (black arrows),
site of CSF leak at T6–T7 (white arrow), and contrast
material spreading within extradural fluid collection
(arrowheads). For reference, please note that left
side and top of digital subtraction myelographic
image correspond to anterior and cephalad aspects
of patient, respectively.

Downloaded from www.ajronline.org by 129.67.118.203 on 05/12/19 from IP address 129.67.118.203. Copyright ARRS. For personal use only; all rights reserved
A B
2–5). Because of the ventral location of the
dural tears, the CSF leaks were typical-
ly best visualized in the lateral projection,
but the anteroposterior plane was useful for
defining laterality in some instances (Fig.
5B). Of the nine cases confirmed with digi-
tal subtraction myelography, the most com-
mon cause for the dural tear was a focally
prominent disk osteophyte (five patients).
An additional patient had a noncalcified
transdural disk extrusion. Two patients had
a history of remote trauma with associated
mild vertebral compression fractures at the
site of CSF leak. One patient had no clear
structural cause for the dural tear but had
a presumed genetic predisposition to dural
fragility given his history of multiple pre-
vious spontaneous CSF leaks involving dif-
ferent spinal levels.
Of the two patients in whom digital sub-
traction myelography was nondiagnostic,
the CSF leak is presumed to have been a
slow process that could not readily be vis-
ualized during the brief 20- to 30-second
digital subtraction myelographic acquisi-
tion. The clinical condition of one patient
with superficial siderosis deteriorated sec-
ondary to multiple medical comorbidities,
and no further workup was undertaken for
his CSF leak. The second patient with SIH
had a partially calcified disk extrusion at
T11–T12, which was confirmed as the site
of the CSF leak on subsequently performed
dynamic CT myelography.
Six patients (two of six with SIH, four of
five with superficial siderosis) underwent
primary surgical closure of the dural tear.
The remaining four patients with SIH have
652 AJR:199, September 2012
Hoxworth et al.
been managing their symptoms with con-
servative treatment. As we noted earlier,
the poor medical condition of one patient
with superficial siderosis precluded further
workup and treatment.
Discussion
Spontaneous spinal CSF leaks have been
firmly established as a cause of SIH, and spinal
dural tears are also an increasingly recognized
association with superficial siderosis [1, 2].
In both of these instances, routine MRI of the
spine and conventional CT myelography are
frequently sufficient to suggest the presence of
a CSF leak but may not offer conclusive local-
ization. Targeted minimally invasive therapies
and surgical approaches, when necessary, re-
quire precise localization of the site of the CSF
leak, and additional myelographic techniques
have been developed and used successfully for
this purpose. Dynamic CT myelography with
intrathecal administration of contrast material
performed during CT acquisition improves the
likelihood of visualizing the site of a CSF leak
[10]. Although offering superior anatomic de-
tail, dynamic CT myelography may still not be
able to localize very rapid CSF leaks, so digi-
tal subtraction myelography was developed as
an adjunct test that provides improved tempo-
ral resolution [11].
The results of the current retrospective
study confirm that digital subtraction my-
elography is a useful test for the localization
of spinal dural tears in select patients. In our
experience, the rapidity of a spinal CSF leak
can be estimated through an analysis of pre-
procedural spinal MRI. Elongated extradural
fluid collections within the spinal canal span-
Fig. 5—38-year-old man with spontaneous intracranial
hypotension and CSF leak to right of midline at T11
from unknown cause. Genetic predisposition to dural
fragility was presumed because of his history of
multiple previous spontaneous CSF leaks involving
different spinal levels.
A, Lateral digital subtraction myelographic image is
annotated to designate normal flow of contrast material
within subarachnoid space (black arrows), site of CSF
leak at T11 (white arrow), and contrast material spreading
within extradural fluid collection (arrowheads). For
reference, please note that left side and top of digital
subtraction myelographic image correspond to anterior
and cephalad aspects of patient, respectively.
B, In frontal projection, site of CSF leak is visualized
as double-density sign (white arrows) superimposed
on contrast material within subarachnoid space
(black arrows). Please note that left side and top of
digital subtraction myelographic image correspond
to left and cephalad aspects of patient, respectively.
(Because patient was in prone position, normal
radiographic convention was not followed.)
ning multiple vertebral levels tend to behave
analogous to pseudomeningoceles in which
the fluid freely communicates with the sub-
arachnoid space. This equilibration permits
the rapid extravasation of contrast material,
often in a matter of seconds, from the thecal
Fig. 6—Lateral digital subtraction myelography
of cervical spine with left side and top of digital
subtraction myelographic image corresponding
to anterior and cephalad aspects of 42-year-old
man with spontaneous intracranial hypotension,
respectively. Visualization of myelographic
contrast material near cervicothoracic junction
is impaired because of radiation attenuation from
shoulders (arrows). In addition, ventral projection of
contrast material within normal nerve root sleeves
(arrowheads) could potentially obscure contrast
material extravasating into ventral epidural fluid
collection.
 Digital Subtraction Myelography of Spinal CSF Leaks
sac into the extradural collection. As a result,
the injection of intrathecal contrast material
in the fluoroscopy suite using a standard my-
elographic technique followed by transfer of
the patient to CT for cross-sectional imaging
allows too much time to elapse. The extra-
Downloaded from www.ajronline.org by 129.67.118.203 on 05/12/19 from IP address 129.67.118.203. Copyright ARRS. For personal use only; all rights reserved
dural contrast material has frequently already
disseminated over multiple vertebral levels
so that localization of the dural tear is not
successful. Dynamic CT myelography offers
greatly improved temporal resolution com-
pared with conventional postmyelograph-
ic CT but is unable to achieve the multiple
frames per second acquired with digital sub-
traction myelography. Consequently, when
a patient with a suspected rapid spinal CSF
leak is referred for myelography, intrathe-
cal injection of contrast material using digi-
tal subtraction myelography is now frequent-
ly performed at our institution to overcome
the temporal limitations of other techniques.
In spite of the 82% success rate in local-
izing the site of the dural tear in this series,
digital subtraction myelography has sever-
al potential limitations that must be consid-
ered when determining whether this imag-
ing technique is appropriate for a given case.
First, the digital subtraction acquisition is ex-
tremely sensitive to motion degradation, so
patients must be able to comfortably suspend
respiration in the prone position. With ap-
propriate patient counseling, we have found
general anesthesia with suspended respira-
tion is not necessary for diagnostic-quality
imaging. Second, the detection of cervical
and upper thoracic CSF leaks may be lim-
ited on digital subtraction myelography for
technical reasons. The added density of the
patient’s shoulders can make the stream of
myelographic contrast agent appear attenu-
ated, and the normal anterolateral projection
of cervical nerve root sleeves could poten-
tially mask a ventral CSF leak (Fig. 6). Last,
the duration of the digital subtraction acqui-
sition is only 20–30 seconds, so some CSF
leaks may not be rapid enough to be visual-
ized. Slow leaks presumably occurred in the
two cases in this series for which the sites of
the CSF leaks were not identified with digital
subtraction myelography.
If digital subtraction myelography fails to
identify the site of a CSF leak but longitudi-
nally extensive extradural fluid is present on
spinal MRI, then we typically examine the
region of interest using dynamic CT myelog-
raphy on a different day, to allow the previ-
ously injected contrast medium to clear, as
AJR:199, September 2012 653
the next step. Again, limiting the digital sub-
traction myelographic technique to patients
with multilevel extradural fluid collections is
an important part of preprocedural screen-
ing. For example, routine CT myelography
usually works well in evaluating patients
who have small, localized extradural fluid
collections. Other patients with CSF leaks
that are suspected to be slow or intermit-
tent, as is often the case with CSF leaks from
meningeal diverticula, require techniques
that afford delayed imaging. Although de-
layed imaging can be accomplished with CT,
the added radiation exposure is undesirable
and the gravitational dependence of iodin-
ated contrast material requires the patient
to remain in the recumbent position. Radio-
nuclide cisternography has the advantage of
delayed imaging and has been used success-
fully but often must rely on indirect signs of
CSF leak for diagnosis because the lack of
spatial detail precludes direct localization
of the CSF leak in many cases [13–15]. As
a result, the intrathecal injection of gadolin-
ium-based contrast agents for MR myelogra-
phy has shown increased promise in patients
who require delayed imaging [16–18]. Spe-
cifically, the gadolinium contrast agent re-
mains better dispersed within the subarach-
noid space rather than moving to the most
gravitationally dependent region. Gadolini-
um permits patients to be active and perform
provocative maneuvers while waiting to un-
dergo delayed imaging hours after injection.
Because intrathecal injection of gadolini-
um-based contrast material is not approved
by the U.S. Food and Drug Administration,
MR myelography represents an off-label use.
However, recent data are increasingly sup-
portive of its safety profile [16].
In conclusion, digital subtraction myelog-
raphy is a valuable diagnostic tool for the
localization of rapid spinal CSF leaks and
should be considered in patients who are
clinically suspected to have a dural tear that
is accompanied by a longitudinally extensive
extradural fluid collection on spinal MRI.
References
1. Schievink WI. Spontaneous spinal cerebrospinal
fluid leaks. Cephalalgia 2008; 28:1345–1356
2. Kumar N. Neuroimaging in superficial siderosis:
an in-depth look. AJNR 2010; 31:5–14
3. Medina JH, Abrams K, Falcone S, Bhatia RG.
Spinal imaging findings in spontaneous intracra-
nial hypotension. AJR 2010; 195:459–464
4. Chiapparini L, Farina L, D’Incerti L, et al. Spinal
radiological findings in nine patients with sponta-
neous intracranial hypotension. Neuroradiology
2002; 44:143–150; discussion, 151–152
5. Watanabe A, Horikoshi T, Uchida M, Koizumi H,
Yagishita T, Kinouchi H. Diagnostic value of spi-
nal MR imaging in spontaneous intracranial hy-
potension syndrome. AJNR 2009; 30:147–151
6. Rabin BM, Roychowdhury S, Meyer JR, Cohen
BA, LaPat KD, Russell EJ. Spontaneous intracra-
nial hypotension: spinal MR findings. AJNR
1998; 19:1034–1039
7. Kumar N, Cohen-Gadol AA, Wright RA, Miller
GM, Piepgras DG, Ahlskog JE. Superficial sider-
osis. Neurology 2006; 66:1144–1152
8. Koeppen AH, Michael SC, Li D, et al. The pathol-
ogy of superficial siderosis of the central nervous
system. Acta Neuropathol 2008; 116:371–382
9. Kumar N. Superficial siderosis: associations and
therapeutic implications. Arch Neurol 2007;
64:491–496
10. Luetmer PH, Mokri B. Dynamic CT myelography:
a technique for localizing high-flow spinal cere-
brospinal fluid leaks. AJNR 2003; 24:1711–1714
11. Hoxworth JM, Patel AC, Bosch EP, Nelson KD.
Localization of a rapid CSF leak with digital sub-
traction myelography. AJNR 2009; 30:516–519
12. Headache Classification Subcommittee of the In-
ternational Headache Society. The international
classification of headache disorders: 2nd edition.
Cephalalgia 2004; 24(suppl 1):9–160
13. Hyun SH, Lee KH, Lee SJ, et al. Potential value of
radionuclide cisternography in diagnosis and
management planning of spontaneous intracrani-
al hypotension. Clin Neurol Neurosurg 2008;
110:657–661
14. Hashizume K, Watanabe K, Kawaguchi M, Taoka
T, Shinkai T, Furuya H. Comparison of computed
tomography myelography and radioisotope cister-
nography to detect cerebrospinal fluid leakage in
spontaneous intracranial hypotension. Spine (Ph-
ila Pa 1976) 2012; 37:E237–E242
15. Morioka T, Aoki T, Tomoda Y, et al. Cerebrospi-
nal fluid leakage in intracranial hypotension syn-
drome: usefulness of indirect findings in radionu-
clide cisternography for detection and treatment
monitoring. Clin Nucl Med 2008; 33:181–185
16. Albayram S, Kilic F, Ozer H, Baghaki S, Kocer N,
Islak C. Gadolinium-enhanced MR cisternogra-
phy to evaluate dural leaks in intracranial hypo-
tension syndrome. AJNR 2008; 29:116–121
17. Dillon WP. Intrathecal gadolinium: its time has
come? AJNR 2008; 29:3–4
18. Vanopdenbosch LJ, Dedeken P, Casselman JW,
Vlaminck SA. MRI with intrathecal gadolinium
to detect a CSF leak: a prospective open-label co-
hort study. J Neurol Neurosurg Psychiatry 2011;
82:456–458