Psychiatry Research, 37:321-331
321
Elsevier
Plasma Catecholamines Obsessive-Compulsive
and Their Disorder
Metabolites
in
Chawki Benkelfat, Ivan N. Mefford, Carolyn F. Masters, Thomas A. Catherine King, Robert M. Cohen, and Dennis L. Murphy Received
Fehruarr,
14, 1991; accepted
E. Nordahl,
Ma-r I, 1991.
Abstract. Plasma catecholamines and their metabolites were sampled in 13 medication-free patients with obsessive-compulsive disorder (OCD) and 29 normal controls. In addition to severe OCD symptoms, the patients had significantly higher anxiety, tension, and resting pulse rates than the controls. Nonetheless, mean plasma concentrations of norepinephrine (NE) and epinephrine (E), the catecholamine metabolites 3-methoxy-4-hydroxyphenylglycol (M H PC) and homovanillic acid (HVA), and the stress-related hormone cortisol did not differ between OCD patients and normal controls. When the patients and control populations were combined and average plasma NE and E levels calculated over 35 min. subjects with a higher mean NE output (> 1.1 pm/ ml) had higher Profile of Mood States depression scores than subjects with a low NE output (< I. I pm/ ml). Altogether, these results indicate that elevated plasma catecholamine measuresare not likely to be associated with the pathophysiology of OCD. Key Words. Epinephrine, norepinephrine. homovanillic hydroxyphenylglycol, cortisol, sympathoadrenal activity.
acid,
3-methoxy-4-
Obsessive-compulsive disorder (OCD) is a chronic anxiety disorder characterized by dysphoric obsessional thoughts and repeated ritualistic behaviors or compulsions. Recent findings implicate central serotonergic dysfunction in the neurobiology of this disorder (Murphy et al., 1989). Some have suggested that other central neurotransmitters, such as norepinephrine (NE), dopamine, and opioid peptides, are involved in the chemical mediation of the OCD symptom complex. Recently, Rasmussen et al. (1987) summarized several studies that supported a noradrenergic involvement in the pathophysiology of the OCD symptom complex. In particular, these studies emphasized the role of central norepinephrine (NE) in the pathogenesis of anxiety in OCD. In a study of a small number of patients, OCD patients had a blunted growth hormone (GH) response following intravenous (i.v.) challenge with the cY,-agonist clonidine (Siever et al., 1983). Furthermore, OCD Chawki Benkelfat, M.D. is Assistant Professor, Neurobiological Psychiatry Unit, Department of Psychiatry, McGill University, Montreal. Quebec, Canada. Ivan N. Mefford, Ph.D., is Chief. and Carolvn F. Masters. B.Sc.. is a research assistant. Unit of Clinical Neurochemistry, Experimental Therapeutic Branch, National Institute of Mental Health (NIMH). Bethesda, MD. Dennis L. Murphv. M.D.. is Chief. Laboratory of Clinical Science, NIMH, Bethesda, MD. Thomas E. Nordahl, M.D., Ph.D., is Awstant Professor, Department of Psychiatry, University of California at Davis. A. Catherine King is a research assistant, Child Psychiatry Branch, NIMH, Bethesda, MD. Robert M. Cohen, M.D., Ph.D., is Chief, Section on Clinical Brain Imaging, Laboratory of Cerebral Metabolism, NI M H, Bethesda, MD. (Reprint requests to Dr. C. Benkelfat. Research & Training Bldg., Rm. 208, Dept. of Psychiatry, McCiill University, 1033 Pine
Ave. W.. Montreal. Que. H3A-IAI, Canada.) 0165-1781/91/$03.50 @ 1991 Elsevier Scientific Publishers Ireland Ltd.
322 patients had higher preinfusion plasma levels of NE and its main metabolite, 3methoxy-4-hydroxyphenylglycol (M HPG), than normal controls did (Siever et al., 1983; Rasmussen et al., 1987). These studies led to the hypothesis that OCD is associated with both increased presynaptic noradrenergic activity and decreased postsynaptic noradrenergic receptor responsiveness. This hypothesis, however, was not supported by a study that measured M H PC and homovanillic acid (HVA) in the cerebrospinal fluid (CSF) of nonmedicated adult OCD patients (Thor&r et al., 1980), or by a study that measured plasma NE and epinephrine (E) in children with OCD (Flament et al., 1987). Furthermore, other OCD studies that used clonidine (Lee et al., 1990) and yohimbine challenges also failed to support this hypothesis. In the latter study, single-dose administration of yohimbine produced no differences in the behavioral responses or MHPG concentrations of OCD patients as compared with those of healthy controls (Rasmussen et al., 1987). In the present study, we measured the plasma catecholamines NE and E and the catecholamine metabolites MHPG and HVA in OCD patients and normal controls during a nonstressful procedure to compare these catecholamine measures in the two groups.
Plasma
catecholamine
and metabolite
concentrations
were sampled
in fore-
venous blood, as skeletal muscle contributions to plasma NE concentrations can be a significant variable (Esler et al., 1984). Because plasma half-lives of NE and E are known to be very short (l-2 min) (Christensen et al., 1984), we collected serial blood samples every 5 min for about 35 min to evaluate any group differences in pulsatile release of catecholamines.
arm
arterial blood, increases in venous
rather
than
Methods 7hirteen OCD patients, with an average duration of illness of 16.8 (SD = 8.7) years, who were drug free for at least 3 weeks (with 8: I3 drug-free for more than a year) (Table I). and 29 normal controls, with similar age and sex distribution (Table 2). were included in the study. All subjects were recruited as participants in an ongoing fluorodeoxyglucose (FDG) positron emission tomography (PET) research project that was part of the National Institute of. Mental Health’s Intramural PET program in psychiatry (Nordahl et al., 1989; Benkelfat et al.. 1990). All OCD patients met the following criteria: (I) primary DSM-III-R diagnosis of OCD (American Psychiatric Association, 1987); (2) National Institute of Mental Health (NIMH) Global OC Scale value > 7 (OC symptoms interfered mildly to moderately with daily activities); (3) good physical health (as determined by physical examination, routine laboratory tests, electrocardiogram (EKG), and chest X-ray); and (4) medication free for at least 3 weeks. To avoid the possible confounding effect of severe depression on plasma catecholamine levels. we excluded all but one patient who had an Axis I USM-III- R diagnosis of current primary major depression. Normal controls were devoid of any significant medical problems (as determined by physical examination, routine laboratory tests, EKG, and chest X-ray) and of a personal or family (first-degree relative) history of psychiatric illness. Subjects.
Behavioral Evaluation. Behavioral observations, which were carried out at the time of blood sampling, included a careful evaluation of anxiety, OC symptoms, and mood disturbances. Anxiety levels were evaluated using the Spielberger State Anxiety Scale (SSAS; Spielberger et al,. 1970) and the Profile of Mood States (POMS; McNair et al., 1971). The following rating scales were used to measure OC symptomatology: (I) the OC subscale of the Comprehensive Psychiatric Rating Scale (CPRS-OC) developed by Asberg et al. (1978); (2) the National Institute of Mental Health Global OC Scale (Zohar et al., 1988); and (3) the Maudsley Obsessional Compulsive Inventory (Hodgson and Rachman, 1965). Mood was evaluated with
323 the Beck Depression Inventory (BDI; Beck et al., 1961). the 17-item Hamilton Rating Scale for Depression (HRSD-17; Hamilton, 1960), and the Profile of Mood States-Depression (POMSD). Procedure. An indwelling arterial catheter was placed in the subject’s right radial artery 90 min before PET scan. After catheter placement and before blood collection, subjects rested for 30 min in the supine position. During the scanning procedure, both normal controls and OCD patients remained supine with their heads within the PET scanner ring and with eyes patched. While in the scanner, they performed a simple nonstressful auditory task (FDG uptake period), which consisted of pushing a button when listening to a soft tone through headphones. The patients did not differ from the normal controls on their performance of the auditory task (number of false hits in successive 5-min blocks for 30 min) (data not shown). Blood samples were collected in ethylenediamine tetraacetic acid (EDTA) tubes, Before obtaining the first blood sample, we systematically recorded heart rate in all subjects. The first sample (baseline) was obtained before iv. injection of a traceramount of [lXF]FDCi. Ina subset of normal controls and OCD patients, blood was also obtained every 5 min for 35 min (7 time points) after the injection to measure serial levels of plasma catecholamines. A total blood volume of 42 ml of blood was collected from each subject during each procedure. All blood samples were put on ice immediately and then spun in a refrigerated centrifuge ( IO min. 6OOOg,5 “C). Plasma was separated and then stored in a freezer at -70 “C. Cortisol was assayed using a standardized radioimmunoassay (Mueller et al., 19X5). Laboratory Methods. Catecholamines were extracted using a method described by Masters et al. (1988). Samples were run with the following chromatographic parameters. Pulseless solvent delivery (200 ~1: mn) was accomplished using a model 400-02 reciprocating pump (Applied Chromatography Systems, Inc.). Sample injection was accomplished with a 50-~11fixed volume loop injector (Gilson autosampler, model 23 I). Chromatograms were recorded on a strip chart recorder (LKB Instruments). Catecholamines were detected amperometrically using a glassy carbon electrode (TL-8A) at a potential of +0.6 volts vs. Ag, AgCl reference using a model LC-4B amperometric detector (Bioanalytical Systems, Inc.). Chromatographic separation was performed on a 10 cm length x I mm i.d. ODS 3 pm Ultrasphere column (Beckman Instruments). The column was packed in the laboratory with a high pressure slurry packer (Chemco Model 124A). The use of ion pair reverse phase chromatography with Igepon (Mefford et al., 1987) as the surfactant enabled us to measure E (> 0.1 pm:ml) and NE (> 0.3 pm/ml) concentrations simultaneously. All plasma NE and E samples for both patients and controls were extracted together and measured in one chromatographic run, which enabled us to eliminate interassay variability. Intra-assay varrability vvas 39; for NE and E. of Plasma Concentrations of HVA and MHPG. Extraction of M H PC; and HVA was performed using a modified technique of Scheinin et al. ( 1983) (MH PC) and Chang et al. (1983) (HVA). The internal standard was 100 ~1 of IO-” M 3-ethoxy-4-hydroxyphenylglycol (EHPG) and vortexed with 600 ~1 of plasma. Protein w’as removed with Amicon Centrifree filters, and then a single biphasic extraction was performed. Fifty microliters of phosphate buffer (supersaturated, pH q 6.9) 300 mg NaCI. and 4 ml ethyl acetate were added to each sample. The samples were vortexed and centrifuged, and the ethyl acetate was transferred and then evaporated under NJ in a 37 “C water bath. For injection, the metabolites were resuspended with 200 ~1 sodium metabisulfate IO-5 M. Pulseless solvent delivery ( I.5 ml; min) w’as accomplished using a model 400-02 reciprocating pump (Applied Chromatography Systems, Inc.). A Gilson autosample (model 231) with a 50-~1 fixed volume loop injector was again used to inject the samples. The metabolites were detected amperometrically at a potential of 0.75 volts vs. Ag: AgCl reference using a model LC-4B amperometric detector (Bioanalytical Systems, Inc.) and recorded on an LKB Instruments strip chart recorder. The plasma samples were again assayed in one chromatographic run, limiting variability. lntra-assay variability was 12.3c/o for MHPG and 10% for HVA. Determination
324 Statistics. The results were analyzed in the following four ways: (I) Comparisons of the means for behavioral, physiological, and biochemical variables between normal controls and OCD patients were made using Student’s f test (Z-tailed). (2) A search for a difference in pulse release of catecholamines (pulse release differences that might occur at regular intervals, when plasma is sampled every 5 min for 35 min in subjects at rest) between patients and controls was made using a two-way (time/group) analysis of variance (ANOVA) for repeated measures. (3) Behavioral measures in the low vs. high mean plasma levels of NE groups were compared using a factorial ANOVA. (4) Regression analysis was performed on clinical (measures of OC. mood, and anxiety symptoms, continuous performance test accuracy), physiological, and biochemical measures of interest (pulse, cortisol. NE, E, M H PC, and HVA) in the group of normal controls and the group of the OCD patients separately and, whenever possible, with both groups combined. Not all measures were available in all subjects. Table 2 presents the sample size for
each determination. Results Table I presents clinical data describing the OCD patients and normal controls. Self-rating scores of tension, apprehension, and anxiety obtained immediately before blood collection were significantly higher in the OCD patients (SSAS, p < 0.0001; POMS-Tension (T), p < 0.001). Pulse (u < 0.01) was also higher in OCD patients. There were no significant baseline differences between OCD patients and normal controls in plasma E, NE, and MHPG (Table 2). In the subset of seven normal controls and seven OCD patients who had plasma E and NE samples collected every 5 min for 35 min, no significant differences emerged between patients and controls for time or group effect for pulse release of plasma catecholamines. As indicated in Fig. I, Table
Pt.
1. Demographic Ma (yr)
profile
Sex
OCD history (yr)
1
29
M
7
2
35
F
18
3
24
M
13
4
26
M
8
5
40
M
30
6
42
F
7
29
of the 13 OCD Past Rx
Maudsley total
patients
CPRS
NIMHoc
HRSD
Subtype
26
13
30
5
12
12
31
14
23
IO
20
8
checker
-
21
13
29
10
checker
-
23
IO
27
7
washer
30
al.2
22
10
27
M
16
-
14
9
23
5
checker mixed
alz+imi
mixed mixed
washer
8
34
F
22
imi
25
12
29
5
9
27
F
17
-
26
14
30
10
mixed
10
35
M
29
-
12
12
27
11
washer
11
44
M
11
-
26
14
35
13
mixed
12
43
F
12
ami
22
18
49
24
washer
13
32
M
5
flx
14
13
43
25
atypical
Mean
33.8
16.8
20.5
12.3
30.6
11.1
8.7
5.5
2.4
7.8
6.6
SD
6.8
5F/8M
Note. Past Rx (treatment]: alz = alprazolam; am, = amltrlptyllne; urn = Imlpramlne; fix = fluoxetlne. CpRS = Comprehenswe Psychlatrlc Rating Scale. NIMH-OC = National lnstltute of Mental Health ObsessweCompulsive Scale. HRSD = 17.item HamIlton Rating Scale for Depression.
325 Table 2. Plasma levels of E, NE, MHPG, HVA, and cortisol in untreated OCD Datients and controls
Age iyr)
Normal controls
OCD patients
(n = 29)
(n = 13)
34.8 210
Sex SSAS POMS Anxiety
6.8
13F/16M
5F/8M
32.6 + 9.6
51.4 ?I 51
9.2 i: 8.6 (r/=24)
Pulse (resting)
33.82
64.9 i
8.8
17.9 + 6.12 jn=12) 75.8 k11.82
Plasma concentrations E (pm/ml)
0.44? 0.16 1rJ=lO)
0.56+ 0.37 jn=lO)
NE (pm/ml)
1.16k 0.50 (n=12)
1.32? 0.52
MHPG
(pm/ml)
HVA (pm/ml)
Cortisol (ng/ml
J
(n=lO)
17.3 + 7.4 in=261
16.6 + 3.7 (n=lO)
19.4 k 8.6
18.6 t 6.3 (n=9)
8.7 + 3.6
9.7 * 9.5
Note. E = epinephrine. NE = norepinephrine. HVA= homovanillicacid. MHPG = 3-methoxy-4-hydroxyphenylglycol. POMS= Profileof Mood States. SSAS= Spielberger State Anxiety Scale. OCD = obsessive-compulsive disorder. Numbers are indicated in parentheses when fewer than 29 controls or 13 patients were samples. 1. p -c0.001 2. p < 0.01.
were quite stable over the multiple sampling time period. The plasma catecholamine levels of OCD patients appeared to be higher than those of normal controls, but these differences were not significant. Mean plasma cortisol levels did not differ between the two groups (Table 2). Since the suggestion has been made that plasma catecholamines may reflect a state variable (“tension,” “ arousal,” and “sensitivity to external stressors”), cutting across nosological categories, we elected to combine NE plasma levels in OCD and healthy subjects, and correlate them with mood states. For that purpose, a mean plasma NE output was calculated for 21 subjects (9 OCD patients and 12 normal controls) by averaging the NE plasma levels obtained at each time point. By splitting the mean plasma NE levels (around the median distribution of NE values of 1.1 pmiml), we formed two subgroups: subjects with high NE concentrations (> 1.1 pm/ml; n = 11) and subjects with low NE concentrations (< 1.1 pm/ ml; n = 10). The subgroup with high plasma NE concentrations had significantly higher scores on the POMS-D subscale(9.5?7.0,n= 11,vs.4.1+5.2,n~10;t~2.0,df~19,p=0.06)butnotonthe POMS-T Scale, SSAS scale, pulse, or cortisol.
values
326 Fig. 1. Lack of difference in serial plasma levels of epinephrine (E) and norepinephrine (NE) between OCD patients and normal controls
2.0 -
15-
1 o-
0.5 -
0.0 I
I
1
2
4
SAMPLING OCD patients, n : 7. normal represent SD.
controls,
TIME
(every
6
6
5 minutes)
n mm 7. Blood samples were obtalned
every 5 mln for 35 min. Error bars
In normal controls, but not in OCD patients, plasma MHPG concentrations correlated modestly with scores on one measure of anxiety, the SSAS (r = 0.48, p < 0.05). In this group, plasma MHPG was also correlated with plasma cortisol concentrations (r = 0.38, p < 0.05). In OCD patients, plasma E concentrations at baseline correlated positively with pulse (r= 0.66,p< O.OS)and with NE plasma concentrations (v = 0.62, p < 0.05) (Table 3). In OCD patients, scores on the CPRS-OC subscale also correlated negatively with plasma HVA (r = -0.80, p = 0.01) NE (r = -0.62,p< 0.05) and E (r = -0.9 1, p < 0.00 1) concentrations, but not with the concentrations of plasma MGPG. Discussion In the present study, OCD patients had significantly higher self-rated anxiety and heart rate than did normal controls. These findings are consistent with previous reports of higher anxiety ratings (Barlow et al., 1985; Rasmussen et al., 1987; Charney et al., 1988) and reports of mild increases in physiological arousal in OCD patients, as measured by both cardiovascular and electrodermal indices (Kelly, 1980; Insel et al., 1984). We also measured plasma concentrations of the catecholamines, NE and E, the catecholamine metabolites, MHPG and HVA, and the stress-related hormone, cortisol. Despite the fact that OCD patients were significantly more anxious than normal controls, no group differences were found in the plasma levels of NE, E, MHPG, HVA, or cortisol. In OCD patients, the mean plasma levels for E and NE, although not significantly different from those in controls, were about 13% higher than the normal control values. This finding suggests that studies in a larger group, using a different test
0.16
0.48
0.11
0.08
0.55
0.47
0.18
0.36
0.73
13
13
10
10
10
9
13
12
NIMH-OC
Pulse
NE
E
MHPG
HVA
Cortisol
CPT
0.42
0.08
0.15
0.05
0.29
0.592
0.26
0.16 0.31 0.47 0.13 -0.752 0.05
0.2
0.05
0.13 0.11 0.21 0.38
0.1
0.911
0.21
0.761
0.12
BDI
0.20
0.37
POMS-T
0.11
0.08
-0.83
-0.46
-0.911
-0.622
0.16
0.801
CPRS-8
0.31 0.25
-0.722 0.01
0.08
0.16
0.22
0.1
0.05 0.662
-0.50
Pulse
-0.44
0.07
NIMH-OC
0.64
0.37
0.40
0.54
0.622
NE
and behavior in obsessive-compulsive
0.74
0.38
0.13 0.40
0.752
MHPG
0.3
0.51
E
disorder
-0.2
0.07
HVA
-0.524
Cortisol
1. 2. 3. 4.
p p p p
< < <
321
Elsevier
Plasma Catecholamines Obsessive-Compulsive
and Their Disorder
Metabolites
in
Chawki Benkelfat, Ivan N. Mefford, Carolyn F. Masters, Thomas A. Catherine King, Robert M. Cohen, and Dennis L. Murphy Received
Fehruarr,
14, 1991; accepted
E. Nordahl,
Ma-r I, 1991.
Abstract. Plasma catecholamines and their metabolites were sampled in 13 medication-free patients with obsessive-compulsive disorder (OCD) and 29 normal controls. In addition to severe OCD symptoms, the patients had significantly higher anxiety, tension, and resting pulse rates than the controls. Nonetheless, mean plasma concentrations of norepinephrine (NE) and epinephrine (E), the catecholamine metabolites 3-methoxy-4-hydroxyphenylglycol (M H PC) and homovanillic acid (HVA), and the stress-related hormone cortisol did not differ between OCD patients and normal controls. When the patients and control populations were combined and average plasma NE and E levels calculated over 35 min. subjects with a higher mean NE output (> 1.1 pm/ ml) had higher Profile of Mood States depression scores than subjects with a low NE output (< I. I pm/ ml). Altogether, these results indicate that elevated plasma catecholamine measuresare not likely to be associated with the pathophysiology of OCD. Key Words. Epinephrine, norepinephrine. homovanillic hydroxyphenylglycol, cortisol, sympathoadrenal activity.
acid,
3-methoxy-4-
Obsessive-compulsive disorder (OCD) is a chronic anxiety disorder characterized by dysphoric obsessional thoughts and repeated ritualistic behaviors or compulsions. Recent findings implicate central serotonergic dysfunction in the neurobiology of this disorder (Murphy et al., 1989). Some have suggested that other central neurotransmitters, such as norepinephrine (NE), dopamine, and opioid peptides, are involved in the chemical mediation of the OCD symptom complex. Recently, Rasmussen et al. (1987) summarized several studies that supported a noradrenergic involvement in the pathophysiology of the OCD symptom complex. In particular, these studies emphasized the role of central norepinephrine (NE) in the pathogenesis of anxiety in OCD. In a study of a small number of patients, OCD patients had a blunted growth hormone (GH) response following intravenous (i.v.) challenge with the cY,-agonist clonidine (Siever et al., 1983). Furthermore, OCD Chawki Benkelfat, M.D. is Assistant Professor, Neurobiological Psychiatry Unit, Department of Psychiatry, McGill University, Montreal. Quebec, Canada. Ivan N. Mefford, Ph.D., is Chief. and Carolvn F. Masters. B.Sc.. is a research assistant. Unit of Clinical Neurochemistry, Experimental Therapeutic Branch, National Institute of Mental Health (NIMH). Bethesda, MD. Dennis L. Murphv. M.D.. is Chief. Laboratory of Clinical Science, NIMH, Bethesda, MD. Thomas E. Nordahl, M.D., Ph.D., is Awstant Professor, Department of Psychiatry, University of California at Davis. A. Catherine King is a research assistant, Child Psychiatry Branch, NIMH, Bethesda, MD. Robert M. Cohen, M.D., Ph.D., is Chief, Section on Clinical Brain Imaging, Laboratory of Cerebral Metabolism, NI M H, Bethesda, MD. (Reprint requests to Dr. C. Benkelfat. Research & Training Bldg., Rm. 208, Dept. of Psychiatry, McCiill University, 1033 Pine
Ave. W.. Montreal. Que. H3A-IAI, Canada.) 0165-1781/91/$03.50 @ 1991 Elsevier Scientific Publishers Ireland Ltd.
322 patients had higher preinfusion plasma levels of NE and its main metabolite, 3methoxy-4-hydroxyphenylglycol (M HPG), than normal controls did (Siever et al., 1983; Rasmussen et al., 1987). These studies led to the hypothesis that OCD is associated with both increased presynaptic noradrenergic activity and decreased postsynaptic noradrenergic receptor responsiveness. This hypothesis, however, was not supported by a study that measured M H PC and homovanillic acid (HVA) in the cerebrospinal fluid (CSF) of nonmedicated adult OCD patients (Thor&r et al., 1980), or by a study that measured plasma NE and epinephrine (E) in children with OCD (Flament et al., 1987). Furthermore, other OCD studies that used clonidine (Lee et al., 1990) and yohimbine challenges also failed to support this hypothesis. In the latter study, single-dose administration of yohimbine produced no differences in the behavioral responses or MHPG concentrations of OCD patients as compared with those of healthy controls (Rasmussen et al., 1987). In the present study, we measured the plasma catecholamines NE and E and the catecholamine metabolites MHPG and HVA in OCD patients and normal controls during a nonstressful procedure to compare these catecholamine measures in the two groups.
Plasma
catecholamine
and metabolite
concentrations
were sampled
in fore-
venous blood, as skeletal muscle contributions to plasma NE concentrations can be a significant variable (Esler et al., 1984). Because plasma half-lives of NE and E are known to be very short (l-2 min) (Christensen et al., 1984), we collected serial blood samples every 5 min for about 35 min to evaluate any group differences in pulsatile release of catecholamines.
arm
arterial blood, increases in venous
rather
than
Methods 7hirteen OCD patients, with an average duration of illness of 16.8 (SD = 8.7) years, who were drug free for at least 3 weeks (with 8: I3 drug-free for more than a year) (Table I). and 29 normal controls, with similar age and sex distribution (Table 2). were included in the study. All subjects were recruited as participants in an ongoing fluorodeoxyglucose (FDG) positron emission tomography (PET) research project that was part of the National Institute of. Mental Health’s Intramural PET program in psychiatry (Nordahl et al., 1989; Benkelfat et al.. 1990). All OCD patients met the following criteria: (I) primary DSM-III-R diagnosis of OCD (American Psychiatric Association, 1987); (2) National Institute of Mental Health (NIMH) Global OC Scale value > 7 (OC symptoms interfered mildly to moderately with daily activities); (3) good physical health (as determined by physical examination, routine laboratory tests, electrocardiogram (EKG), and chest X-ray); and (4) medication free for at least 3 weeks. To avoid the possible confounding effect of severe depression on plasma catecholamine levels. we excluded all but one patient who had an Axis I USM-III- R diagnosis of current primary major depression. Normal controls were devoid of any significant medical problems (as determined by physical examination, routine laboratory tests, EKG, and chest X-ray) and of a personal or family (first-degree relative) history of psychiatric illness. Subjects.
Behavioral Evaluation. Behavioral observations, which were carried out at the time of blood sampling, included a careful evaluation of anxiety, OC symptoms, and mood disturbances. Anxiety levels were evaluated using the Spielberger State Anxiety Scale (SSAS; Spielberger et al,. 1970) and the Profile of Mood States (POMS; McNair et al., 1971). The following rating scales were used to measure OC symptomatology: (I) the OC subscale of the Comprehensive Psychiatric Rating Scale (CPRS-OC) developed by Asberg et al. (1978); (2) the National Institute of Mental Health Global OC Scale (Zohar et al., 1988); and (3) the Maudsley Obsessional Compulsive Inventory (Hodgson and Rachman, 1965). Mood was evaluated with
323 the Beck Depression Inventory (BDI; Beck et al., 1961). the 17-item Hamilton Rating Scale for Depression (HRSD-17; Hamilton, 1960), and the Profile of Mood States-Depression (POMSD). Procedure. An indwelling arterial catheter was placed in the subject’s right radial artery 90 min before PET scan. After catheter placement and before blood collection, subjects rested for 30 min in the supine position. During the scanning procedure, both normal controls and OCD patients remained supine with their heads within the PET scanner ring and with eyes patched. While in the scanner, they performed a simple nonstressful auditory task (FDG uptake period), which consisted of pushing a button when listening to a soft tone through headphones. The patients did not differ from the normal controls on their performance of the auditory task (number of false hits in successive 5-min blocks for 30 min) (data not shown). Blood samples were collected in ethylenediamine tetraacetic acid (EDTA) tubes, Before obtaining the first blood sample, we systematically recorded heart rate in all subjects. The first sample (baseline) was obtained before iv. injection of a traceramount of [lXF]FDCi. Ina subset of normal controls and OCD patients, blood was also obtained every 5 min for 35 min (7 time points) after the injection to measure serial levels of plasma catecholamines. A total blood volume of 42 ml of blood was collected from each subject during each procedure. All blood samples were put on ice immediately and then spun in a refrigerated centrifuge ( IO min. 6OOOg,5 “C). Plasma was separated and then stored in a freezer at -70 “C. Cortisol was assayed using a standardized radioimmunoassay (Mueller et al., 19X5). Laboratory Methods. Catecholamines were extracted using a method described by Masters et al. (1988). Samples were run with the following chromatographic parameters. Pulseless solvent delivery (200 ~1: mn) was accomplished using a model 400-02 reciprocating pump (Applied Chromatography Systems, Inc.). Sample injection was accomplished with a 50-~11fixed volume loop injector (Gilson autosampler, model 23 I). Chromatograms were recorded on a strip chart recorder (LKB Instruments). Catecholamines were detected amperometrically using a glassy carbon electrode (TL-8A) at a potential of +0.6 volts vs. Ag, AgCl reference using a model LC-4B amperometric detector (Bioanalytical Systems, Inc.). Chromatographic separation was performed on a 10 cm length x I mm i.d. ODS 3 pm Ultrasphere column (Beckman Instruments). The column was packed in the laboratory with a high pressure slurry packer (Chemco Model 124A). The use of ion pair reverse phase chromatography with Igepon (Mefford et al., 1987) as the surfactant enabled us to measure E (> 0.1 pm:ml) and NE (> 0.3 pm/ml) concentrations simultaneously. All plasma NE and E samples for both patients and controls were extracted together and measured in one chromatographic run, which enabled us to eliminate interassay variability. Intra-assay varrability vvas 39; for NE and E. of Plasma Concentrations of HVA and MHPG. Extraction of M H PC; and HVA was performed using a modified technique of Scheinin et al. ( 1983) (MH PC) and Chang et al. (1983) (HVA). The internal standard was 100 ~1 of IO-” M 3-ethoxy-4-hydroxyphenylglycol (EHPG) and vortexed with 600 ~1 of plasma. Protein w’as removed with Amicon Centrifree filters, and then a single biphasic extraction was performed. Fifty microliters of phosphate buffer (supersaturated, pH q 6.9) 300 mg NaCI. and 4 ml ethyl acetate were added to each sample. The samples were vortexed and centrifuged, and the ethyl acetate was transferred and then evaporated under NJ in a 37 “C water bath. For injection, the metabolites were resuspended with 200 ~1 sodium metabisulfate IO-5 M. Pulseless solvent delivery ( I.5 ml; min) w’as accomplished using a model 400-02 reciprocating pump (Applied Chromatography Systems, Inc.). A Gilson autosample (model 231) with a 50-~1 fixed volume loop injector was again used to inject the samples. The metabolites were detected amperometrically at a potential of 0.75 volts vs. Ag: AgCl reference using a model LC-4B amperometric detector (Bioanalytical Systems, Inc.) and recorded on an LKB Instruments strip chart recorder. The plasma samples were again assayed in one chromatographic run, limiting variability. lntra-assay variability was 12.3c/o for MHPG and 10% for HVA. Determination
324 Statistics. The results were analyzed in the following four ways: (I) Comparisons of the means for behavioral, physiological, and biochemical variables between normal controls and OCD patients were made using Student’s f test (Z-tailed). (2) A search for a difference in pulse release of catecholamines (pulse release differences that might occur at regular intervals, when plasma is sampled every 5 min for 35 min in subjects at rest) between patients and controls was made using a two-way (time/group) analysis of variance (ANOVA) for repeated measures. (3) Behavioral measures in the low vs. high mean plasma levels of NE groups were compared using a factorial ANOVA. (4) Regression analysis was performed on clinical (measures of OC. mood, and anxiety symptoms, continuous performance test accuracy), physiological, and biochemical measures of interest (pulse, cortisol. NE, E, M H PC, and HVA) in the group of normal controls and the group of the OCD patients separately and, whenever possible, with both groups combined. Not all measures were available in all subjects. Table 2 presents the sample size for
each determination. Results Table I presents clinical data describing the OCD patients and normal controls. Self-rating scores of tension, apprehension, and anxiety obtained immediately before blood collection were significantly higher in the OCD patients (SSAS, p < 0.0001; POMS-Tension (T), p < 0.001). Pulse (u < 0.01) was also higher in OCD patients. There were no significant baseline differences between OCD patients and normal controls in plasma E, NE, and MHPG (Table 2). In the subset of seven normal controls and seven OCD patients who had plasma E and NE samples collected every 5 min for 35 min, no significant differences emerged between patients and controls for time or group effect for pulse release of plasma catecholamines. As indicated in Fig. I, Table
Pt.
1. Demographic Ma (yr)
profile
Sex
OCD history (yr)
1
29
M
7
2
35
F
18
3
24
M
13
4
26
M
8
5
40
M
30
6
42
F
7
29
of the 13 OCD Past Rx
Maudsley total
patients
CPRS
NIMHoc
HRSD
Subtype
26
13
30
5
12
12
31
14
23
IO
20
8
checker
-
21
13
29
10
checker
-
23
IO
27
7
washer
30
al.2
22
10
27
M
16
-
14
9
23
5
checker mixed
alz+imi
mixed mixed
washer
8
34
F
22
imi
25
12
29
5
9
27
F
17
-
26
14
30
10
mixed
10
35
M
29
-
12
12
27
11
washer
11
44
M
11
-
26
14
35
13
mixed
12
43
F
12
ami
22
18
49
24
washer
13
32
M
5
flx
14
13
43
25
atypical
Mean
33.8
16.8
20.5
12.3
30.6
11.1
8.7
5.5
2.4
7.8
6.6
SD
6.8
5F/8M
Note. Past Rx (treatment]: alz = alprazolam; am, = amltrlptyllne; urn = Imlpramlne; fix = fluoxetlne. CpRS = Comprehenswe Psychlatrlc Rating Scale. NIMH-OC = National lnstltute of Mental Health ObsessweCompulsive Scale. HRSD = 17.item HamIlton Rating Scale for Depression.
325 Table 2. Plasma levels of E, NE, MHPG, HVA, and cortisol in untreated OCD Datients and controls
Age iyr)
Normal controls
OCD patients
(n = 29)
(n = 13)
34.8 210
Sex SSAS POMS Anxiety
6.8
13F/16M
5F/8M
32.6 + 9.6
51.4 ?I 51
9.2 i: 8.6 (r/=24)
Pulse (resting)
33.82
64.9 i
8.8
17.9 + 6.12 jn=12) 75.8 k11.82
Plasma concentrations E (pm/ml)
0.44? 0.16 1rJ=lO)
0.56+ 0.37 jn=lO)
NE (pm/ml)
1.16k 0.50 (n=12)
1.32? 0.52
MHPG
(pm/ml)
HVA (pm/ml)
Cortisol (ng/ml
J
(n=lO)
17.3 + 7.4 in=261
16.6 + 3.7 (n=lO)
19.4 k 8.6
18.6 t 6.3 (n=9)
8.7 + 3.6
9.7 * 9.5
Note. E = epinephrine. NE = norepinephrine. HVA= homovanillicacid. MHPG = 3-methoxy-4-hydroxyphenylglycol. POMS= Profileof Mood States. SSAS= Spielberger State Anxiety Scale. OCD = obsessive-compulsive disorder. Numbers are indicated in parentheses when fewer than 29 controls or 13 patients were samples. 1. p -c0.001 2. p < 0.01.
were quite stable over the multiple sampling time period. The plasma catecholamine levels of OCD patients appeared to be higher than those of normal controls, but these differences were not significant. Mean plasma cortisol levels did not differ between the two groups (Table 2). Since the suggestion has been made that plasma catecholamines may reflect a state variable (“tension,” “ arousal,” and “sensitivity to external stressors”), cutting across nosological categories, we elected to combine NE plasma levels in OCD and healthy subjects, and correlate them with mood states. For that purpose, a mean plasma NE output was calculated for 21 subjects (9 OCD patients and 12 normal controls) by averaging the NE plasma levels obtained at each time point. By splitting the mean plasma NE levels (around the median distribution of NE values of 1.1 pmiml), we formed two subgroups: subjects with high NE concentrations (> 1.1 pm/ml; n = 11) and subjects with low NE concentrations (< 1.1 pm/ ml; n = 10). The subgroup with high plasma NE concentrations had significantly higher scores on the POMS-D subscale(9.5?7.0,n= 11,vs.4.1+5.2,n~10;t~2.0,df~19,p=0.06)butnotonthe POMS-T Scale, SSAS scale, pulse, or cortisol.
values
326 Fig. 1. Lack of difference in serial plasma levels of epinephrine (E) and norepinephrine (NE) between OCD patients and normal controls
2.0 -
15-
1 o-
0.5 -
0.0 I
I
1
2
4
SAMPLING OCD patients, n : 7. normal represent SD.
controls,
TIME
(every
6
6
5 minutes)
n mm 7. Blood samples were obtalned
every 5 mln for 35 min. Error bars
In normal controls, but not in OCD patients, plasma MHPG concentrations correlated modestly with scores on one measure of anxiety, the SSAS (r = 0.48, p < 0.05). In this group, plasma MHPG was also correlated with plasma cortisol concentrations (r = 0.38, p < 0.05). In OCD patients, plasma E concentrations at baseline correlated positively with pulse (r= 0.66,p< O.OS)and with NE plasma concentrations (v = 0.62, p < 0.05) (Table 3). In OCD patients, scores on the CPRS-OC subscale also correlated negatively with plasma HVA (r = -0.80, p = 0.01) NE (r = -0.62,p< 0.05) and E (r = -0.9 1, p < 0.00 1) concentrations, but not with the concentrations of plasma MGPG. Discussion In the present study, OCD patients had significantly higher self-rated anxiety and heart rate than did normal controls. These findings are consistent with previous reports of higher anxiety ratings (Barlow et al., 1985; Rasmussen et al., 1987; Charney et al., 1988) and reports of mild increases in physiological arousal in OCD patients, as measured by both cardiovascular and electrodermal indices (Kelly, 1980; Insel et al., 1984). We also measured plasma concentrations of the catecholamines, NE and E, the catecholamine metabolites, MHPG and HVA, and the stress-related hormone, cortisol. Despite the fact that OCD patients were significantly more anxious than normal controls, no group differences were found in the plasma levels of NE, E, MHPG, HVA, or cortisol. In OCD patients, the mean plasma levels for E and NE, although not significantly different from those in controls, were about 13% higher than the normal control values. This finding suggests that studies in a larger group, using a different test
0.16
0.48
0.11
0.08
0.55
0.47
0.18
0.36
0.73
13
13
10
10
10
9
13
12
NIMH-OC
Pulse
NE
E
MHPG
HVA
Cortisol
CPT
0.42
0.08
0.15
0.05
0.29
0.592
0.26
0.16 0.31 0.47 0.13 -0.752 0.05
0.2
0.05
0.13 0.11 0.21 0.38
0.1
0.911
0.21
0.761
0.12
BDI
0.20
0.37
POMS-T
0.11
0.08
-0.83
-0.46
-0.911
-0.622
0.16
0.801
CPRS-8
0.31 0.25
-0.722 0.01
0.08
0.16
0.22
0.1
0.05 0.662
-0.50
Pulse
-0.44
0.07
NIMH-OC
0.64
0.37
0.40
0.54
0.622
NE
and behavior in obsessive-compulsive
0.74
0.38
0.13 0.40
0.752
MHPG
0.3
0.51
E
disorder
-0.2
0.07
HVA
-0.524
Cortisol
1. 2. 3. 4.
p p p p
< < <
