Académique Documents
Professionnel Documents
Culture Documents
in Obsessive-Compulsive Disorder
A Comparison With Rates in Unipolar Depression and in Normal Controls
Lewis R. Baxter, Jr, MD; Michael E. Phelps, PhD; John C. Mazziotta, MD, PhD;
Barry H. Guze, MD; Jeffrey M. Schwartz, MD; Carl E. Selin, MS
studied 14 patients with obsessive-compulsive disor-
\s=b\ We to the anxiety disorders, psychoses, Gilles de la Tourette's
der (OCD) by positron emission tomography and the fluo- disorder ("Tourette's disorder") and particularly depres¬
rodeoxyglucose method, looking for abnormalities in local sion, has been questioned.14,6"12 The inability of the patient's
cerebral metabolic rates for glucose in brain structures that "will" to control this unwanted mental activity suggested a
have been hypothesized to function abnormally in OCD.These deficit in mental energy ("psychasthenia") to Janet,1"3,13·14
patients were compared with 14 normal controls and 14 pa- while the apparently symbolic nature of obsessions and
tients with unipolar depression. The patients with unipolar compulsions led Freud to view the problem as being related
depression and OCD did not differ in levels of anxiety, tension, to complex defenses invoked to deal with potent psychosex-
or depression. In OCD, metabolic rates were significantly ual conflicts.1"4 Some have viewed both obsessions and
increased in the left orbital gyrus and bilaterally in the caudate compulsions as the product of conditioning.1"4 To others, the
nuclei. This was apparent on all statistical comparisons with association of OCD with Tourette's disorder15"17 and with
both controls and unipolar depression. The right orbital gyrus cases of von Enconomo's encephalitis13,18"20 have suggested
showed at least a trend to an increased metabolic rate in all neurologic dysfunction.21 Studies of cerebrospinal fluid
comparisons. The metabolic rate in the left orbital gyrus, amines,22 neuropsychological tests,2325 electroen-
relative to that in the ipsilateral hemisphere (orbital gyrus/ cephalographic measures,26,26 sleep studies,9 and response
hemisphere ratio), was significantly elevated compared to to pharmacologie treatments3,27"29 have likewise suggested
controls and subjects with unipolar depression, and stayed biologic abnormalities in this disorder.30,31
high even with successful drug treatment. Though it was in the We studied global and local cerebral metabolic rates for
normal range in the morbid state, with improvement in OCD glucose (LCMRGlc) in patients with OCD using positron
symptoms after drug treatment, the caudate/hemisphere meta- emission tomography and fluoro-2-deoxyglucose labeled
bolic ratio increased uniformly and significantly bilaterally. with fluorine 18.32,33 The results were compared with those
This ratio did not increase in patients who did not respond to obtained from normal controls and patients with major
treatment. Thus, OCD showed cerebral glucose metabolic depressive disorder, unipolar type, who were studied under
patterns that differed from controls in both the symptomatic similar conditions. A subgroup of the patients with OCD
and recovered states. was also examined after treatment with medication and the
(Arch Gen Psychiatry 1987;44:211-218) results were compared with pretreatment data and the data
from the normal group.
disorder
(OCD)1"4 characterizedis
Obsessi
by vpersistent,
e-compulsirecurrent,
ve against
that invade conscious awareness
and
repugnant thoughts
an individual's will
SUBJECTS AND METHODS
Subjects
This study was carried out under guidelines established by the
("obsessions"). These are usually accompanied by ego- UCLA Human Subjects' Protection Committee. Characteristics of
dystonic, ritualistic behaviors ("compulsions") that the indi¬ subjects and their scores on various behavior rating scales are
vidual feels he must perform to prevent overwhelming shown in Table 1. Values are given as the mean ± SD.
anxiety. Insight into the perverse and senseless nature of Patients With OCD.—All 14 patients met both DSM-IIP* and
these frequently bizarre rituals is usually preserved. A Research Diagnostic Criteria35 for OCD, as diagnosed using the
serious disorder, the psychosocial morbidity from OCD is Schedule for Affective Disorders and Schizophrenia, Lifetime
thought to be among the highest seen in any major psychi¬ Version.36 All patients had compulsive rituals as well as obsessions.
atric disorder." Though formerly thought to be rare, the Rituals were further categorized into washing or checking types1
recent National Institute of Mental Health (NIMH) catch¬ (Table 1). Nine of these individuals also met DSM-III criteria for
ment area epidemiologie study indicated a six-month point major depression (Research Diagnostic Criteria secondary depres¬
sion subtype) at the time of scanning, but had histories of OCD
prevalence of 1% to 2% and a lifetime prevalence of 2% to 3% when not depressed and showed residual OCD symptoms after
for OCD.5 their index depressive episode resolved with treatment. Nine of
Obsessive-compulsive disorder has been the subject of the patients with OCD were studied when they had been free of
both diagnostic and etiologic controversy. Its relationship psychotropic drugs for at least two weeks (mean, 6.2 ±3.7; range,
two to 12 weeks). Five patients were not able or willing to stop
Accepted for publication Oct 14, 1986. taking medications for clinical reasons and they were studied while
From the Division of Biophysics and Nuclear Medicine, Department of they received the following medications (one patient each):
Radiological Sciences (Drs Baxter, Phelps, Mazziotta, and Guze and Mr alprazolam, 3 mg/d; alprazolam, 3 mg/d, and desipramine hydro¬
Selin), Department of Psychiatry and Biobehavioral Sciences (Drs Baxter, chloride, 150 mg/d; diazepam, 5 mg/d, phenelzine sulfate, 60 mg/d,
Guze, and Schwartz), and Department of Neurology (Dr Mazziotta), UCLA and perphenazine, 8 mg/d; haloperidol, 5 mg/d; and trazodone
School of Medicine.
Presented in part at the Annual Meeting of the American College of hydrochloride, 50 mg/d. Ten of the patients were restudied after
treatment with trazodone, with or without a monoamine oxidase
Neuropsychopharmacology, Kaanapali, Maui, Hawaii, Dec 12, 1985.
Reprint requests to UCLA Neuropsychiatric Institute, 760 Westwood inhibitor.37 Female subjects with OCD underwent scanning with¬
Plaza, Los Angeles, CA 90024 (Dr Baxter). out regard for time in the menstrual cycle. Thirteen subjects were
scanned without regard for time in the menstrual cycle. Thirteen 3 Mean 4.66±0.76 4.47 ±0.58 5.26 ±0.80
subjects were right-handed and one (woman) was ambidextrous, ±SD (N 14) =
(N 14)
=
(N 14)
=
O
-
2 7'
DC
Í8
Ir 6
S c
119>
t
CD ';=
Cl co
=
¬
c
o
w
o
a
g
E 5
4
f 5
S "5CD O
(ANOVA) with post hoc t tests between pairings of subject groups, Absolute Glucose Metabolic Rates
with Bonferroni's correction for multiple tests.50,51 This was done
for the following neuroanatomic regions that had been postulated Cerebral Hemispheres.—Both left and right hemispheres had
to show pathology in OCD, based on the literature cited in the significantly (P<.05) higher LCMRGlc values in OCD than in
introduction: (1) whole cerebral hemispheres, (2) prefrontal cortex unipolar depression. This was seen on all statistical comparisons.
(precentrai gyrus, rectal gyrus, orbital gyrus, lateral superior Though patients with OCD had a significantly higher LCMRGlc on
frontal gyrus, medial superior frontal gyrus, middle frontal gyrus, the total-subject group and the drug-free comparisons, rates were
inferior frontal gyrus, and medial prefrontal cortex—all examined not significantly greater than normals on the analysis of the second
individually), (3) temporal cortex (lateral superior temporal gyrus, seven patients with OCD. Figure 1 depicts these results in the
transverse superior temporal gyrus, middle temporal gyrus, and right hemisphere for the total group of subjects.
inferior temporal gyrus—all examined individually), (4) hippocam- Heads of the Caudate Nuclei.—All comparisons performed for
pal-parahippocampal complex, (5) anterior cingulate gyrus, (6) both left and right caudate nuclei showed the LCMRGlc for OCD to
heads of the caudate nuclei, (7) putamen, and (8) thalamus. These be significantly (P<.05) higher than that for both normal controls
structures were examined (1) in terms of absolute LCMRGlc, (2) and patients with unipolar depression. Figure 2 shows these
relative to the rate of the ipsilateral hemisphere by means of a ratio results for the total group of subjects for the right head of the
of the two (LCMRGlc-structure/LCMRGlc-hemisphere), and (3) caudate nucleus.
for left-right asymmetries ([LCMRGlc-left-LCMRGlc-right]/ Orbital Gyri.—Both orbital gyri showed a significantly (P<.05)
[LCMRGlc-left + LCMRGlc-right]/2). higher LCMRGlc in patients with OCD than in normal subjects in
Structures showing significant (P<.05, two-tailed) differences all comparisons. The left orbital gyrus had a significantly higher
from both control groups were then examined in the second group LCMRGlc in patients with OCD compared with patients with
of seven patients with OCD (against the same controls) employing unipolar depression on all statistical comparisons. The right or¬
planned orthagonal comparisons of group means using a t statistic, bital gyrus, however, showed only a trend (P<.10) to be higher in
with one-tailed probability limits.52 In addition, a third analysis patients with OCD than in those with unipolar depression in the
was undertaken that examined only those patients with OCD who second group of seven subjects with OCD and in the drug-free
were drug free (ANOVA and post hoc t tests with Bonferroni's group. Figure 3 shows these results in the right orbital gyrus for
correction). Student's t test, with two-tailed values, was used for the total group of subjects. (Technical difficulties made it impossi-
Obsessive-Compulsive
Normal Controls (N) Unipolar Depressed ( ) Disorder ( ) Significant
-'-' '-*-' -*-· Differences
L Hemisphere Hemisphere L Hemisphere R Hemisphere L Hemisphere R Hemisphere (P .05)t
Cerebral hemispheres
All subjects 4.73 + 0.76(14) 4.66±0.76 (14) 4.54±0.58 (14) 4.47±0.58 (14) 5.29±0.88 (14) 5.26±0.80 (14) OCD>UD,NI;
_Land R
Drug-free subjects 5.30 ±0.65 (9) 5.25 ±0.61 (9) OCD>UD,NI;
... ... ... ...
Land R
Head of caudate
All subjects 5.70±0.90 (14) 5.81 ±0.89 (14) 5.34±0.62 (14) 5.52±0.69 (14) 6.55±1.31 (14) 6.79±1.32 (14) OCD>UD,NI;
Land R
Drug-free subjects 6.52 ±0.96 (9) 6.73±1.00(9) OCD>UD,NI;
... ... ... ...
Land R
Orbital gyrit
All subjects 5.08±0.91 (13) 5.09±0.92 (13) 5.16±0.71 (14) 5.23±0.75 (14) 6.20±1.32 (13) 6.12 + 1.18 (13) OCD>UD,NI;
L and R
Drug-free subjects 6.02 ±0.90 (8) 5.92 ±0.84 (8) L:OCD>UD,NI;
... ... ... ... R:OCD>NI
*Glucose metabolic rates are given as milligrams of glucose per 100 g of brain per minute. All values are the mean ± SD.
tOCD indicates obsessive-compulsive disorder; UD, unipolar depressed; and NI, normal.
¿Structurenot visible on scans of all subjects.
Table 3.—Glucose Metabolic Rates in Orbital Gyri, Relative to Hemispheric Mean (Orbit-Hemisphere Ratio)*
Obsessive-Compulsive Significant
Normal Controls (N) Unipolar Depressed (N) Disorder (N) Differences
-'-
-'- -'-
(P<.05)
_L_R_L_R_L_R_(See Text)
All subjects 1.09±0.06 (13) 1.11 ±0.08 (13) 1.14±0.05 (14) 1.17±0.06 (14) 1.17±0.08 (13) 1.17±0.08 (13) OCD>NI;
L only
Drug-free subjects ... ... 1.15±0.06 (8) 1.14 ±0.07 (8) OCD>NI;
... ...
L only
*OCD indicates obsessive-compulsive disorder; NI, normal.
ble to calculate the LCMRGlc for the orbital gyri of one normal 15, respectively, on the HAM-D scale. All ten patients underwent
subject and one patient with OCD.) scanning before and after drug treatment, with scans performed
Table 2 gives the LCMRGlc for the cerebral hemisphere, heads two to four months apart.
of the caudate nuclei, and orbital gyri, both for the total group of The LCMRGlc for both hemispheres, caudate nuclei, and orbital
subjects and for those who were drug free. gyri did not show statistically significant changes after treatment
in the group of eight patients who responded to treatment (P>.25
Relative Glucose Metabolic Rates for all groups), although in all six structures the group means for
these values were lower after treatment than before. These same
Statistical analysis of metabolic ratio values from the first seven
values also dropped in the two patients whose conditions did not
subjects with OCD, compared with the control groups, showed improve.
significant results only for the left LCMRGlc-orbital gyrus/ Figure 5 shows that there were significant changes in the
LCMRGlc-hemisphere ratio (Table 3). In this structure, ratio LCMRGlc-caudate/LCMRGlc-hemisphere ratio in the group that
values in subjects with OCD were significantly higher than in
normal controls, but were not higher than those seen in subjects responded to drug treatment, but not in the patients who did not
with unipolar depression on all statistical comparisons. (Even respond. In the right hemisphere of responders the ratio changed
+ 6.7%, while in nonresponders it changed -5.3%. These were
when the six patients with unipolar depression who had obsessions
were excluded from the analysis, the value of this ratio was not
significant differences (P<.005). Values in responders and nonre¬
significantly higher in subjects with OCD compared with subjects sponders changed + 5.9% and -11.4%, respectively (P<.03), in the
with unipolar depression.) Figure 4 depicts results in the left left hemisphere. Mean pretreatment values for this ratio in sub¬
orbital gyrus for the total group of subjects with OCD and normal jects with OCD were not significantly different from normal
controls. values, either in the total patient group (right: normals, 1.25 ± 0.09;
OCD, 1.29 + 0.06; left: normals, 1.21 + 0.07; OCD, 1.23 + 0.05) or in
Glucose Metabolic Changes After Drug Treatment the group of eight patients with OCD who responded to drug
treatment (OCD, right, 1.26 + 0.06; left, 1.22 + 0.04) (P>.20). With
Ten subjects with OCD were treated with trazodone hydro¬ successful treatment, however, this ratio increased (right,
chloride, with or without the addition of a monoamine oxidase 1.34 + 0.07; left, 1.29 + 0.08). These posttreatment values were
inhibitor, in an open-study design.37 (Full details of treatment significantly higher than both the pretreatment values for these
strategy and results are reported elsewhere.53) The pretreatment patients with OCD (right, P<.025; left, P<.05) and those values
mean NIMH-OCD scale score was 32.9 ±7.0 and the HAM-D score obtained from the normal controls (right and left, P<.02). In
was 18.1±6.3. Eight subjects had a clear response and,
in these contrast, the LCMRGlc-orbital gyrus/LCMRGlc-hemisphere ratio
patients, the NIMH-OCD scale score went from 34.4±6.7 to did not change after treatment in either the patients who were
12.6±8.3 and the HAM-D score went from 20.0±5.5 to 6.5±2.7. treated successfully (1.17 + 0.06 before treatment vs 1.17 + 0.09
The scores of two patients who did not respond (one treated with after treatment) or unsuccessfully (the mean value went from 1.20
trazodone and a monoamine oxidase inhibitor and the other treated to 1.17).
with trazodone alone) went from 30 and 23 to 31 and 27, respec¬ Pretreatment and posttreatment percentage changes in the
tively, on the NIMH-OCD scale, and went from 12 and 6 to 12 and LCMRGlc-caudate/LCMRGlc-hemisphere ratio ([pretreatment -
1.20
•
___
Œ
1.10t d»
" "
1.00
Mean 1.09±0.06 1.17±0.08
±SD
(N 13)
=
(N 13)
=
Normal Obsessive-
Compulsive
Fig 4.—Glucose metabolic rates for left orbital gyrus divided by rate Fig 5.—Glucose metabolic rate of right head of caudate nucleus
for ipsilateral hemisphere as whole. Ratio was significantly higher in divided by that of ipsilateral hemisphere for patients with obsessive-
subjects with obsessive-compulsive disorder than in normals. compulsive disorder studied before and after drug treatment. A
Values for the right hemisphere did not show significant differences. indicates posttreatment values for two patients who did not respond
Circles indicate drug-free subjects; squares, obsessive-compulsive to drug treatment. B, values on pretreatment scans; C, values for
disorder. one subject scanned after partial response to treatment with
trazodone hydrochloride, after which tranylcypromine was added,
with additional response and another scan performed, as indicated
at point D; and D, values for subjects scanned after optimal
posttreatmentMpretreatment + posttreatment]/2) were corre¬ response. All subjects who responded to drug showed increase in
lated with both pretreatment HAM-D and NIMH-OCD scores and ratio (P-C.025, paired t test), while the two who did not respond did
pretreatment to posttreatment percentage changes in these not increase ratio. Percentage change In ratio was significantly
two scales (percent change in HAM-D score; percent change in different between group of patients who responded to drug treat¬
NIMH-OCD score). Correlations were stronger with the HAM-D ment vs those who did not (P<.005). Values for left hemisphere
than the NIMH-OCD score (right hemisphere: pretreatment
score were essentially same.
HAM-D score, r, .85, P<.01; pretreatment NIMH-OCD score,
=
Fig 6.—Positron emission tomographic scans of patients with obsessive-compulsive disorder. Top, Location of heads
of caudate nuclei and orbital gyri in scans shown below. Center, Scan of patient vs that of age- and sex-matched normal
control. Colors of scans correspond to glucose metabolic rates, which are indicated in color bar to right (in micromoles
of glucose per minute per 100 g of brain tissue; 1 mg of glucose equals 5.56 µ ). Glucose metabolic rates in
obsessive-compulsive disorder are especially high in caudate nuclei and orbital gyri (particularly on left) compared
with normal controls. Bottom, Scan of patient with obsessive-compulsive disorder before and after successful drug
treatment. Scan has been coded to reflect metabolic rate of each area of brain, divided by that of brain as whole. As in
center illustration, red indicates highest values and blue, lowest. Note how value of ratio for head of caudate nuclei
Increased with successful treatment. Patients who did not respond did not show increase in ratio. (See Fig 5.) Notice
how ratio value does not change in left orbital gyrus; despite symptomatic Improvement, it remains high when
compared with normal controls.
ness and awareness.55,56 The orbital gyri in particular, aside on other aspects of personality.58"65 However, procedures
from their probable involvement in the regulation of mood that isolate the anterior cingulate gyrus may produce
states, seem important in maintaining the integrated per¬ equally good results.21,66 We are not aware of any attempts to
ception of complex temporal relationships through the remove only the left orbital gyrus as a treatment for OCD.
inhibition of external interfering stimuli.55"57 The difficulty Abnormal metabolic activity in the caudate nuclei can
patients with OCD have in diverting their attention from also be understood in light of available literature. It has
obsessions might be due to the overactivity of such a been suggested that the basal ganglia system acts as a
system. Furthermore, certain destructive lesions of the sensory-information gating station or processor and, when
orbital cortex result in disinhibition characterized by face- it is dysfunctional, unmodified afferent information leads to
tiousness, sexual and personal hedonism, irritability, and a inappropriate behavioral responses.61 Impairment of the
lack of socially appropriate concern for others.55 Such basal ganglia may even produce perseverative behaviors.68
syndromes are the behavioral opposite of the symptoms and Dysfunction in this region might relate to the perseverative
typical personality manifestations seen in individuals with nature of obsessions and compulsions.
OCD.lJ< Though consisting mostly of anecdotal reports, It has been postulated that the orbital cortex and the
there is a large body of literature that suggests that ventral caudate nucleus determine if a behavioral response
neurosurgical procedures that obliterate or isolate the will occur when an animal must redirect its behavior in re¬
orbital gyri from the rest of the brain may alleviate some of sponse to the environment.69 Lesions in either the orbital
the symptoms of OCD, while having a relatively small effect area or areas of the caudate nuclei to which the orbital area