Test-retest reliability of central [11C]raclopride binding at high D2 receptor occupancy. A PET study in haloperidol-treated patients

PSYCHIATRY RESEARCH NEUROIMAGING ELSEVIER

Psychiatry Research: Neuroimaging Section 67 (1996) 163-171

Test-retest reliability of central [alC]raclopride binding at high D 2 receptor occupancy. A PET study in haloperidol-treated patients Svante Nyberg*, Lars Farde, Christer Halldin Karolinska Institute, Department of Clinical Neuroscience, Psychiatry Section, Karolinska Hospital, S-171 76 Stockholm, Sweden

Received 27 September 1995; revised 29 April 1996; accepted 5 May 1996

Abstract

Central D 2 dopamine receptor occupancy may be a useful measure to establish clinical guidelines for optimal antipsychotic drug treatment. The use of positron emission tomography (PET) to explore quantitative relationships among D 2 receptor occupancy and clinical effects depends on the reliability of such measurements. The calculation of D 2 receptor occupancy using [HC]raclopride is routinely based on a ratio-equilibrium analysis, in which the ratio of radioactivity concentration in the striatum to that in the cerebellum is determined. To examine the reliability of such ratios, a test-retest analysis was performed in four schizophrenic patients treated with haloperidol decanoate. PET experiments with [~C]raclopride were repeated in each subject during the same day. The putamen to cerebellum ratio ( P / C ratio) ranged from 1.44 to 1.07 among the four patients, corresponding to a D 2 receptor occupancy of 62 to 71%. In each subject, the P / C ratios remained highly similar, with quotients 0.98, 1.01, 1.04 and 1.06 between the two experiments. The high test-retest reproducibility of the P / C ratios indicates that measurements of D 2 receptor occupancy with the present methods are highly reliable, and support the further use of PET to optimize the drug treatment of schizophrenia. Keywords: Positron emission tomography; Human; Schizophrenia; Haloperidol

1. I n t r o d u c t i o n

Positron emission tomography (PET) is a brain imaging technique that permits quantitative stud-

*Corresponding author, Tel: +46 8 729 29 01; fax: +46 8 34 65 63; e-mail: [email protected].

ies of central neuroreceptors, and the degree by which psychotropic drugs occupy such receptors (Wagner et al., 1983; Sedvall et al., 1986). The d o p a m i n e hypothesis of antipsychotic drug action has been firmly supported by the consistent findings of a uniformly high D 2 d o p a m i n e receptor occupancy ( 7 0 - 8 9 % ) in patients treated with antipsy~:hotic drugs (Farde et al., 1986; Baron et al.,

0925-4927/96/$15.00 © 1996 Elsevier Science Ireland Ltd. All rights reserved. PIIS0925-4927(96)02921-6

164

S. Nyberg et al. / Psychiatry Research: Neuroimaging 6 7 (1996) 163-171

1989). We have used the selective radioligand [11C]raclopride for clinical studies on D 2 receptor occupancy in antipsychotic drug-treated patients (Farde et al., 1986, 1988b). Significant relations have been demonstrated between striatal D 2 receptor occupancy and clinical effects (Farde et al., 1992; Nordstr/~m et al., 1993). Based on these findings, we have proposed that 70-80% may be the optimal therapeutic interval for the treatment of acute schizophrenic psychoses (Nyberg et al., 1995b). The wider use of PET to establish clinical guidelines for antipsychotic drug treatment is highly dependent on the reliability of the occupancy measurements. Quantitative PET analyses of receptor binding are based on the time curves for radioactivity in different brain regions. The ratio of radioactivity concentration in the striatum to that in the cerebellum is routinely used as an index of receptor density and to calculate the D E receptor occupancy induced by antipsychotic drugs (Farde et al., 1986, 1989). A high test-retest reliability of PET measurements of [ 11C]raclopride binding has previously been reported (NordstrSm et al., 1992; Volkow et al., 1993). Those evaluations were performed in healthy subjects and unmedicated patients, in whom the ratio between the time curves for the striatum and the cerebellum is favorably high (approximately 5:1). In antipsychotic drug-treated patients, this ratio is lower (below 2:1) due to receptor occupancy. The evaluation of relationships between D 2 receptor occupancy and clinical effects in applied studies thus depends on the reliability of such low ratios. The objective of this study was to examine the test-retest reliability of the putamen-to-cerebellum ratio ( P / C ratio) as determined by PET and [l~C]raclopride in antipsychotic drug-treated patients. 2. Subjects and methods

2.1. Study design The study was approved by the Ethics and the Radiation Safety Committees of the Karolinska Hospital, and was performed at the Department

of Clinical Neuroscience at the Karolinska Hospital. Four patients (A, B, C and D) treated with haloperidol decanoate in monotherapy were recruited after review of their medication records and gave their informed consent. To ensure stable plasma concentrations of haloperidol, PET examinations were performed 6 or 14 days after a depot injection. In each patient two PET experiments were performed 2.5 h apart (Table 1). The patients were removed from the PET camera system but remained resting between experiments. 2.2. Patients and treatment The patients fulfilled DSM-III-R criteria for chronic schizophrenia, and did not have any other mental or physical illness. They had not been hospitalized for at least 5 months before the study. The patients were rated as 'borderline' (2) to 'moderately' (4) mentally ill in the Clinical Global Impression (CGI) (range 1-7) (Guy, 1976) (Table 1). The patients had been treated in monotherapy with haloperidol decanoate (Haldol Depot s, 50 or 100 mg/ml, Janssen Pharmaceutica, Belgium) 30-80 mg every 4 weeks for at least 6 months (Table 1). Patients A and C received their injections 6 days before, and patients B and D 14 days before the PET experiments. No concomitant medication was administered. Blood samples (20 ml) for the determination of plasma concentrations of haloperidol were collected at the start of the first PET experiment. The samples were drawn into heparin-treated glass tubes and centrifuged. Plasma was frozen at -20°C until analyzed. Haloperidol plasma concentrations were analyzed by HPLC (limit of determination of 1 riM) (Llerena et al., 1992). 2. 3. PET experimental procedure A plaster helmet was made for each subject. The helmet was used with a head fixation system to allow transfer of positioning from CT to PET, and to allow repeated PET experiments with the same positioning (Bergstrfm et al., 1981). The

4,2 4,3 2,4 3,5

144 336 144 336

2:25 2:25 2:30 2:30

Time between starts of PET experiments (hours: rain.)

296/292 297/307 316/313 310/311

6.0 6.3 5.5 6.0

25.9 20.2 22.1 18.1

1.251 1.082 1.437 1.155

1.275 1.068 1.384 1.091

Injected Putamen Cerebellum P / C ratio, P / C ratio, second radioactivity VOI (mL) VOI (mL) first experiment experiment (MBq) first/second experiment

Note: The plasma concentration of haloperidol was sampled at the start of the first experiment.

3 4 4 1

50 80 50 30

6 4 8 25

A B C D

32 45 32 42

Haloperiodol Haloperiodol Time since decanoate p-conc, last depot dose (nM) injection (mg/4w.) (hours)

Patient Age Duration CGI (years) of score illness (years)

Table 1 Patient and treatment characteristics, experimental conditions and results

0.98 1.01 1.04 1.06

Relative difference first/ second P / C ratio

I

t...a

.,....

] 66

S. Nyberg et al. / Psychiat O, Research: Neuroimaging 67 (1996) 163-171 4O0

I ~"

Putamen Cerebellum First experiment Second experiment

........

Patient

A Patient B

300 >

~200-

2o0

"o

e"

~ 100-

"~ 100.

1'0

2"0

30

40

0

5()

0

1'0

2"0

time (min.)

3b

go

go

time (rain.) 400.

400.

Patient D Patient C E

300.

c v

~ 200.

15 .~ 200.

--°~1o0.

0

lb

~0

3'o

,~

50

o

o

tb

~

time (rain.)

3b

i.o

go

time (rain.)

Fig. 1. Regional radioactivity, as a function of time, in the putamen (upper curves) and the cerebellum (lower curves) in two consecutive PET experiments in each of four drug-treated schizophrenic patients (A-D). First experiments: unbroken line, second experiments: dotted lines.

foramen of Monro was identified by CT. The helmet was connected by four metal plates to a metal ring, which was rigidly fastened to a metal holder on the bed frame. The bed position was adjusted in steps of 0.5 mm in the sagittal (y) and coronal (z) directions to a standardized position relative to the foramen of Monro (Farde et al., 1988a). [ltC]raclopride was prepared as previously described (Halldin et al., 1991). The specific radioactivity at the time of injection was higher than 500 C i / m m o l and the injected radioactivity was 292-316 MBq (Table 1), which gives a total mass of injected radioligand of about 5/xg. PET experiments were conducted as previously described (Farde et al., 1989). In each experiment, radioactivity in brain tissue was measured

for 51 rain with the P E T system Scanditronix PC2048-15B, which has a resolution of 4.5 mm (FWHM) in the reconstructed image plane (Litton et al., 1990). The distance between the center of each section is 6.5 mm. Data were acquired in 19 consecutive time frames. Frames 1-6 (0-3 min) had a duration of 20 s, frames 7-11 (4-15 min) had a duration of 90 s, and frames 12 to 19 (16-51 min) had a duration of 6 min. Regional radioactivity was corrected for decay and was plotted versus time, each point representing the midtime of each scan (Fig. 1). 2.4. Regions of interest

Regions of interest (ROI) were drawn on the reconstructed PET images from the first experi-

S. Nyberg et aL / Psychiat~ Research: Neuroimaging 6 7 (1996) 163-171

167

ment after visual inspection of all 15 sections. The putamen was drawn bilaterally in two adjacent sections, covering the level of the foramen of Monro. The cerebellum was drawn in two sections, with the highest section showing the pars petrosa of the temporal bone as the upper limitation. Data from adjacent sections of the same ROI was pooled before calculation of regional radioactivity, so that the average radioactivity concentration for the whole volume of interest (VOI) was obtained. The pooled data represents a VOI which was about 6 ml for the putamen and about 20 ml for the cerebellum (Table 1). In each plane where regions were drawn, an extra ROI was drawn for the perimeter of the brain. At the reconstructed images from the second experiments, perimeter ROIs were drawn in the same sections. Thereafter, the regions drawn at images from the first experiment were imported to the images from the second experiment. The fit of the perimeter ROIs were evaluated visually. If necessary, all regions in the same section could be moved simultaneously for optimal fit of the perimeters.

The integration period was chosen to include the time of equilibrium and to reduce the effect noise. D 2 receptor occupancy was defined as the percent reduction in R during drug treatment as compared to the R of 3.77; 0.57 (mean; + 1 S.D.) previously obtained in 34 healthy subjects (age 18-50 years), using the same experimental and analytical methods as described above (Farde et al., 1990; Nordstr/Sm et al., 1992; Nyberg et al., 1995a).

2.5. Calculation of D 2 receptor occupancy

Thus, r estimates the uncertainty (1 S.D.) of an observed difference between two measurements in the same subject. This measure of uncertainty takes all sources of variation into account. With four repeated measurements (4 degrees of freedom), the t-quantile is 2.78 when P < 0.05. A difference between two measured R in the same subject greater than 2.78 r should thus represent a significant difference at the 5% level.

The theory for analysis of D 2 receptor occupancy using [ ~ C]raclopride and the P / C ratio has been presented in detail earlier (Farde et al., 1988b, 1989). The total radioactivity in the cerebellum, Cf(t), was used as an estimate of the free radioligand concentration in the brain. Radioactivity (nCi/ml) representing ligand bound specifically to D 2 receptors, Cb(t), was defined as

2.6. Statistical ana~sis The P / C ratio R (equation 2) was determined twice in each of four subjects, and the difference between the first and second measurement was calculated. The variance of the difference between two measurements, 20-2, was estimated according to the equation:

(Rli-R2i)

r2 = 2s 2 = i= 1

2

4

(3)

3. Results

Ch(t) = Cp,,,(t) - Cf(t)

(1)

3.1. Experimental conditions where Cp,,(t) is the regional radioactivity in the putamen. The curves for Cb(t) and Cf(t) were integrated from 9 to 45 min (scans 12-18) after radioligand injection and the P / C ratio R was obtained according to the equation 45

45

R= f Cb(t)/ f ce(t) 9

9

(2)

PET experiments were conducted according to schedule and procedures. The injected radioactivity was about 300 MBq in all experiments, with no greater difference than 3% between two experiments (Table 1). The plasma levels of haloperidol (range 2.4 to 4.3) were within the range found in previous studies (Reyntjens et al., 1982; Nyberg et al., 1995a) (Table 1).

168

S. Nyberg et al. / Psvchiat~ Research: Neuroimaging 6 7 (1996) 163-171

3.2. PET experimental data Images were reconstructed based on radioactivity summed for the whole experiment. After importation of ROIs to images from the second experiment, the positioning of the ROIs was assessed visually. No moving of the regions was required. Regional radioactivity was calculated for each scan, corrected for decay and plotted versus time (Fig. 1). The curves from both experiments in each patient were superimposed. Both the putamen and the cerebellum curves were very close in the two experiments. The P / C ratios used for the calculation of receptor occupancy were highly similar (Table 1). The quotients were 0.98, 1.01, 1.04 and 1.06 between the two experiments in the same subject. The standard deviation for the difference between two P / C ratio measurements in the same subject was estimated to r = 0.0436 (equation 3). Thus, a difference between two P / C ratios greater than 0.121 in the same subject should be significant at the 5% level. This corresponds to a difference in D e receptor occupancy of 3.2%. 4. Discussion

The test-retest reproducibility of the P / C ratio was very high in the present four patients. These observations indicate that PET measurements of [ix C]raclopride binding are highly reliable also in patients with high D 2 receptor occupancy. The test-retest reliability of the striatum to cerebellum ratio for [11C]raclopride binding has previously been examined. A mean quotient of 1.01 + 0.07 (mean + S.D., n = 10) was reported between repeated measurements of the P / C ratio in the same individual, even when experiments were conducted several years apart (NordstriSm et al., 1992). In a similar study (n = 5), the difference in the striatum-to-cerebellum ratio ranged from - 7 to 8% between experiments performed 24 h apart (Volkow et al., 1993). Both studies examined subjects without treatment, and thus with a favorably high striatum-to-cerebellum ratio. In the present study at clinical conditions with low ratios, the difference between ratios ranged

from - 2 to 6%. The statistical analyses indicate that a change in D 2 receptor occupancy greater than 3.2% measured in the same subject should be regarded as probably significant. 4.1. Calculation o l D 2 receptor occupancy In this study, D 2 receptor occupancy was calculated using an average binding ratio determined in a control sample. However, the density (Bma×) of D 2 receptors varies among individuals. In a recent PET study, we reported an about 2.5-fold range in Bmax among 20 healthy men and women (Farde et al., 1995). An age-related decline in Bmax has been reported in healthy subjects (NordstriSm et al., 1992; Rinne et al., 1993). The error thus introduced using an average binding ratio has been discussed in detail elsewhere (Farde et al., 1992). The measured radioactivity represents the amount of ligand bound to available receptors. Consequently, the magnitude of the error introduced by using the standard baseline P / C ratio is dependent on receptor occupancy, and decreases linearly to 0% when the true occupancy is 100%. Thus, if 70% D 2 receptor occupancy is calculated based on the healthy controls' average ratio 3.77, then + 1 S.D. of the average ratios (3.20 and 4.34) gives a difference in occupancy of - 5 . 3 % and + 3.9%. Moreover, the use of a binding ratio from healthy subjects is based on the assumption that the number of receptors is the same for drug-treated patients, unmedicated schizophrenic patients and healthy controls. Previous PET studies from our group found no difference in D 2 receptor density among healthy subjects and drug-naive schizophrenic patients (Farde et al., 1990; NordstriSm et al., 1995). However, experimental studies in animals show that the administration of high doses of antipsychotic drugs increases D 2 receptor density (Burt et al., 1977; Clow et al., 1980; Owen et al., 1980). A high D 2 receptor density, with a binding ratio 50% higher than the control group average, was demonstrated with PET in a patient 2 weeks after withdrawal from high doses of sulpiride (1600 mg/day) (Farde et al., 1990). Seven weeks later the ratio was only 2% above the control average. These findings suggest an increase of the receptor

S. Nyberg et al. / Psychiatry Research: Neuroimaging 67 (1996) 163-171

density during clinical treatment with neuroleptic drugs. A higher than average receptor density would introduce a systematic underestimation of D 2 receptor occupancy. If a 70% occupancy is calculated using the 3.77 average ratio, but the true 'baseline' is increased by 50% to 5.66, the occupancy would be underestimated by about 10%. We have conducted several studies exploring relations among D 2 receptor occupancy and antipsychotic drug effects. We have shown statistically significant relations between the degree of D 2 receptor occupancy and antipsychotic effect (Nordstr6m et al., 1993). Patients with EPS had a significantly higher D 2 receptor occupancy, above 80% (Farde et al., 1992). Thus, a tentative optimal interval of 70-80% D 2 receptor occupancy may be defined for the acute treatment of schizophrenia (Nyberg et al., 1995b). The future application of PET in larger samples and with controlled study designs depends on the reliability of D 2 receptor occupancy within a few percent. The present results support the use of PET and [J' C]raclopride as accurate tools for such detailed evaluations to further optimize the drug treatment of schizophrenia.

4.2. Haloperidol plasma concentrations The aim of this study was to examine the test-retest reproducibility of measured D 2 receptor occupancy. Therefore, it was important to choose subjects in whom the actual receptor occupancy could be expected to remain stable during the time between experiments. The pharmacokinetics of orally administered antipsychotic drugs indicate that plasma concentrations may change during the course of a few hours. By contrast, patients treated with injectable decanoate formulations have very stable drug plasma concentrations. Pharmacokinetic studies on haloperidol decanoate indicate that the time to peak plasma concentration is about 6 days after injection (Nayak et al., 1987), and that the apparent plasma elimination half-life is about 3 weeks (Reyntjens et al., 1982). Thus, 6 days or more after injection there should be a negligible difference in haloperidol plasma concentration dur-

169

ing the 2.5 h between the two PET experiments. The plasma levels were measured only at the time of the first experiment, since the test-retest reliability of haloperidol levels was not the focus of this investigation.

4.3. Positioning and regions of interest The head fixation system is used to reproduce the position between repeated experiments and to avoid movement artifacts (Bergstr6m et al., 1981). A shift of position may occur for the bed relative to the camera, and for the head relative to the fixation system. When an older PET camera system was used, the reproducibility of the position was reported to be better than 3 mm in each dimension (Kingsley et al., 1980). In the present study, the realignment in the plane was very good (within _ 1 mm), as indicated by the good fit of perimeter ROIs. In this study, we examined patients where the P / C ratio was below 2:1, and found a high testretest reliability. In a recent study (Nyberg et al., submitted), patients treated with very high doses of fluphenazine were examined. The P / C ratios ranged from 0.13 to 0.41, and the calculated D 2 receptor occupancy was 89-97% (mean 93%). This indicates, that [llC]raclopride binding approaches the theoretical maximal occupancy of 100% with sufficiently high doses of competing ligand. These observations further support the view that [~lC]raclopride is well suited for extensive exploration of both low and high degrees of D 2 receptor occupancy. However, caution is warranted in the interpretation of such extreme receptor occupancy based on very low P / C ratios.

Acknowledgements The members of the PET group at Karolinska Institutet are gratefully acknowledged. The members of the Division of Clinical Pharmacology at Huddinge Hospital are thanked for analyses of plasma haloperidol. Mr. Fredrik Granath, Institute of Actuarial Mathematics and Statistics, Stockholm University, is thanked for advice on statistics. The study was supported by grants from the National Institute of Mental Health, USA (MH

170

S. Nyberg et aL / Psychiatry Research: Neuroimaging 67 (1996) 163-171

41205-05) and the Swedish Medical Research Council (9114). References Baron, J.C., Martinot, J.L., Cambon, H., Boulenger, J.P. Poirier, M.F., Caillard, V., Blin, J., Huret, J.D., Loc'h, C. and Mazi~re, B. (1989) Striatal dopamine receptor occupancy during and following withdrawal from neuroleptic treatment: correlative evaluation by positron emission tomography and plasma prolactin levels. Psychophannacologv 99, 463-472. Bergstr6m, M., Bo/~thius, J., Eriksson, L., Greitz, T., Ribbe, T. and Wid6n, L. (1981) Head fixation device for reproducible position alignment in transmission CT and positron emission tomography. J Comput Assist Tomogr 5, 136-141. Burr, D., Creese, I. and Snyder, S. (1977) Antischizophrenic drugs: Chronic treatment elevated dopamine receptor binding in brain. Science 196, 326-328. Clow, A., Theodorou, A. and Jenner, P. (1980) Changes in rat striatal dopamine turnover and receptor activity during one year's neuroleptic administration. Eur J Pharmacol 63, 135-144. Farde, L., Eriksson, L., Blomquist, G. and Halldin, C. (1989) Kinetic analysis of central ['C]raclopride binding to D2dopamine receptors studied by PET - - A comparison to the equilibrium analysis. J Cereb Blood Flow Metab 9, 696-708. Farde, L., Hall, H., Ehrin, E. and Sedvall, G. (1986) Quantitative analysis of D2 dopamine receptor binding in the living human brain by PET. Sc&nce 231,258-261. Farde, L., Hall, H., Pauli, S. and Halldin C. (1995) Variability in D2-dopamine receptor density and affinity - - a PETstudy with [HC]raclopride in man. Synapse 20, 200-208. Farde, L., Nordstr6m, A.-L., Wiesel, F.-A., Pauli, S., Halldin, C. and Sedvall, G. (1992) Positron emission tomographic analysis of central D l- and D:-dopamine receptor occupancy in patients treated with classical neuroleptics and clozapine - - relation to extrapyramidal side effects. Arch Gen Psychiatry 49, 538-544. Farde, L., Pauli, S., Hall, H., Eriksson, L., Halldin, C., H6gberg, T., Nilsson, L., Sj6gren, I. and Stone-Elander, S. (1988a) Stereoselective binding of ~C-raclopride in living human brain - - a search for extrastriatal central D2-dopamine receptors by PET. Psychopharrnacology 94, 471-478. Farde, L., Wiesel, F.-A., Halldin, C. and Sedvall, G. (1988b) Central D2-dopamine receptor occupancy in schizophrenic patients treated with antipsychotic drugs. Arch Gen Psychiatry 45, 71-76. Farde, L., Wiesel, F.-A., Stone-Elander, S., Halldin, C., Nordstr6m, A.-L., Hall, H. and Sedvall, G. (1990) D 2 dopamine receptors in neuroleptic-naive schizophrenic patients. Arch Gen Psychiatry 47, 213-219. Guy, W. (1976) Clinical Global Impressions. In: ECDEU Assessment Manual for Psychopharrnacology, revised (DHEW Publication No. [ADM] 76-338). Superintendent of Documents, U.S. Government Printing Office, Washington, DC.

Halldin, C., Farde, L., H6gberg, T., Hall, H., Str6m, P., Ohlberger, A. and Solin, O. (1991) A comparative PETstudy of five carbon-ll or fluorine-18 labelled salicylamides. Preparation and in vitro dopamine D-2 binding. lnt J Rad Appl lnstrum [B] 18, 871-881. Kingsley, D., Bergstr6m, M. and Berggren, B.-M. (1980) A critical evaluation of two methods of head fixation. Neuroradiologv 19, 7-12. Litton, J.E., Holte, S. and Eriksson, L. (1990) Evaluation of the Karolinska new positron camera system; the Scanditronix PC2048-15B. IEEE Trans Nucl Sci 37, 743-748. Llerena, A., Aim, C., Dahl M-L, Ekqvist, B. and Bertilsson, L. (1992) Haloperidol disposition is dependent on debrisoquine hydroxylation phenotype. Ther Dnlg Monit 14, 92-97. Nayak, R.K., Doose, D.R. and Nair, N.P.V. (1987) The bioavailability and pharmacokinetics of oral and depot intramuscular haloperidol in schizophrenic patients. J Clin Pharmacol 27, 144-150. Nordstr6m, A.-L., Farde, L., Eriksson, L. and Halldin, C. (1995) No elevated D z dopamine receptors in neurolepticnaive schizophrenic patients revealed by positron emission tomography and [" C]N-methylspiperone. Psychiatry Res Neuroimagmg 61, 67-83. Nordstr6m, A.-L., Farde, L., Pauli, S., Litton, J.-E. and Halldin, C. (1992) PET analysis of central [It C]raclopride binding in healthy young adults and schizophrenic patients - - reliability and age effects. Hum Psychophatvnacol 7, 157-165. Nordstr6m, A.-L., Farde, L., Wiesel, F.-A., Forslund, K., Pauli, S., Halldin, C. and Uppfeldt, G. (1993) Central D2-dopamine receptor occupancy in relation to antipsychotic drug effects - - a double blind PET study of schizophrenic patients. Biol Psychiatry 33, 227-235. Nyberg, S., Farde, L., Halldin, C., Dahl, M.-L. and Bertilsson, L. (1995a) D 2 dopamine receptor occupancy during lowdose treatment with haloperidol decanoate. A m J Psychiatry 152, 173-178. Nyberg, S., Nordstr6m, A.-L., Halldin, C. and Farde, L. (1995b) PET studies on D2 dopamine receptor occupancy and antipsychotic drug plasma levels in man. lnt Clin Psychopharmacology 10 (Suppl. 3), 81-85. Owen, F., Cross, A.J. and Waddington, J.L. (1980) Dopaminemediated behaviour and 3H-spiperone binding to striatal membranes in rats after nine months' haloperidol administration. Life Sci 26, 55-59. Reyntjens, A.J.M., Heykants, J.J.P., Woestenborghs, R.J.H., Gelders, Y.G. and Aerts, T.J.L. (1982) Pharmacokinetics of haloperidol decanoate. A 2-year follow-up, lnt Pharmacopsychiatry 17, 238-246. Rinne, J.O., Hietala, J., Ruotsalainen, U., Sako, E., Laihinen, A., Nagren, K., Lehikoinen, P., Oikonen, V. and Syvalahti, E. (1993) Decrease in human striatal dopamine D2 receptor density with age: a PET study with [llC]raclopride. J Cereb Blood Flow Metab 13, 310-314. Sedvall, G., Farde, L., Persson, A. and Wiesel, F.-A. (1986) Imaging of neurotransmitter receptors in the living human brain. Arch Gen Psychiatry 43, 995-1005.

S. Nyberg et al. / Psvchiat9, Research: Neuroimaging 6 7 (1996) 163-171

Volkow, N.D., Fowler, J.S., Wang, G.J., Dewey, S.L., Schlyer, D., MacGregor, R., Logan, J., Alexoff, D., Shea, C., Hitzemann, R., Angrist, B. and Wolf, A. (1993) Reproducibility of repeated measures of carbon-ll-ractopride binding in the human brain. J Nucl Med 34, 609-13. Wagner, H.N., Jr, Burns, H.D., Dannals, R.F., Wong, D.F.,

171

L;]ngstr6m, B., Duelfer, T., Frost, J.J., Ravert, H.T., Links. J.M.. Rosenbloom, S.B., Lukas, S.E., Kramer, A.V. and Kuhar, M.J. (1983) Imaging dopamine receptors in the human brain by positron tomography. Science 221, 1264-1266.