- Email: [email protected]
Brain distribution of selected antipsychotics in schizophrenia
Forensic Science International 157 (2006) 121–130 www.elsevier.com/locate/forsciint
Brain distribution of selected antipsychotics in schizophrenia Kabrena E. Rodda a, Brian Dean b,c,d,e, Iain M. McIntyre a,1, Olaf H. Drummer a,* a
Victorian Institute of Forensic Medicine, Department of Forensic Medicine, Monash University, 57-83 Kavanagh Street, Southbank, Vic. 3006, Australia b The Mental Health Research Institute of Victoria, Vic., Australia c The University of Melbourne Department of Psychiatry, Parkville, Vic., Australia d The University of Melbourne Department of Pathology, Parkville, Vic., Australia e Department of Psychological Medicine, Monash University, Clayton, Vic., Australia Received 23 July 2004; accepted 17 March 2005 Available online 4 June 2005
Abstract The brain distribution of phenothiazine antipsychotics in 22 confirmed schizophrenic and 11 control subjects were collected at autopsy. Specimens were homogenized, extracted with n-butyl chloride, and analyzed via liquid chromatography–mass spectrometry, using atmospheric pressure electrospray ionization operating in the positive mode. Drug concentrations normalized for those observed in cerebellum showed three distinct patterns of distribution corresponding to different structural features of each type of phenothiazine. Those drugs with high affinity for dopamine receptors were detected in the highest concentrations in regions with high concentrations of such receptors. However, those associated with relatively lower dopaminergic activity were found in the highest concentration in the occipital cortex, a region with a relatively low concentration of dopamine receptors. The regional brain distribution of thioridazine and its metabolites was concentration dependent. These results have implications for determining the role of these drugs in the sudden and unexpected deaths of schizophrenics. # 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Post-mortem; LC–MS; Antipsychotics; Brain distribution; Schizophrenia
1. Introduction There is a lack of studies examining how different antipsychotic drugs distribute in the brain at death. Due to differences in drug measurement between studies, comparison of results within and between drugs is difficult. One study of plasma concentrations of thioridazine and metabolites in humans found that although drug concentrations varied widely between patients, in all patients the relative concentration of thioridazine to mesoridazine was about one * Corresponding author. Tel.: +61 3 9684 4334; fax: +61 3 9682 7353. E-mail address: [email protected] (O.H. Drummer). 1 Present address: County of San Diego Medical Examiner’s Office, 5555 Overland Ave., Bldg. 14, San Diego, CA 92123, USA.
half and thioridazine to sulforidazine was two-fold [1]. In a study of antipsychotic drug distribution between blood and brain in rats [2], the observed ratio of brain-to-serum concentration varied widely, but the high potency drugs fluphenazine and haloperidol had more favorable brain-to-blood distribution than the low potency drug thioridazine. Svendsen et al published the only study that reports postmortem concentrations of thioridazine and its metabolites and chlorpromazine in schizophrenic brains [3]. A review of the clinical histories revealed that five subjects had received thioridazine within 8 h of death. In only one subject who had been withdrawn from medication 4 days before death after chronic treatment, thioridazine and its metabolites selectively accumulated in particular regions of the brain. The most remarkable difference was between cortical and sub-cortical structures, in which the former
0379-0738/$ – see front matter # 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2005.03.017
122
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
had high thioridazine levels and lower mesoridazine levels, whereas the latter had high mesoridazine and lower thioridazine concentrations. Although these data are limited, there is evidence to suggest that most anti-psychotics distribute preferentially into the frontal cortex, midbrain, and caudateputamen [3–6], and a degree of differential distribution between the left and right brain hemispheres may occur [5,7]. How a drug distributes in brain would not only seem to be important for post-mortem studies but could provide insight as to how different drugs may distribute in the living brain. Differential distribution may account for why drugs such as clozapine and olanzapine, that have similar receptor pharmacology [8], can have differing therapeutic outcomes [9]. The purpose of this study was to examine whether selected antipsychotic drugs distribute preferentially to particular brain regions in schizophrenic subjects, and to investigate the extent of correlation of blood concentrations with those concentrations in different brain regions. This will allow a better understanding of the interpretation and causes of sudden and expected death in this population.
entiation between grey and white matter from the frontal cortex was made to determine whether the antipsychotic drugs accumulate preferentially in grey matter, where the dopamine neurons are located. Similarly, the occipital cortex was included as a form of negative control (it is the cortical region which is not thought to be involved in schizophrenia). Finally, cerebellum was included to investigate whether antipsychotic drug distribution was related to receptor location. Although GABAA receptors and a1A-adrenoceptors are found in cerebellum, there are very few dopamine or serotonin receptors. 2.3. Determination of postmortem interval (PMI)
2. Materials and methods
Where death did not involve suicide, tissue was collected from subjects whose death was witnessed and the postmortem interval (PMI) was the time of death to autopsy. With suicides, tissue was only taken from individuals who had been seen alive up to 5 h before being found deceased. In those cases, the PMI was estimated at half way between the donor being found dead and being last seen alive. In all cases, the cadavers were refrigerated within 5 h of being found.
2.1. Materials
2.4. Diagnosis of schizophrenia
Mesoridazine was obtained from the Division of Analytical Laboratories of the New South Wales Health Department (Lidcombe, New South Wales). Sulforidazine and fluphenazine sulfoxide were obtained from the Australian Government Analytical Laboratory (South Melbourne, Victoria). Thioridazine, haloperidol, chlorpromazine, and trazodone, (internal standard), were purchased from Sigma Aldrich Pty Ltd. (Castle Hill, New South Wales). All other reagents were HPLC grade and were obtained from Sigma Aldrich Pty Ltd.
The provisional diagnosis of schizophrenia was confirmed by a senior psychologist and senior psychiatrist after an extensive case history review [5]. These reviews were carried out using a structure instrument called Diagnostic Instrument for Brain Studies (see Keks et al. for a review [11]). In this study all diagnoses were confirmed using DSMIV criteria (APA). In addition, information on the type and amount of antipsychotic drugs prescribed close to death was obtained from the case history and the final recorded dose of antipsychotic drug was converted to chlorpromazine dose equivalents [12].
2.2. Tissue collection 2.5. LC–MS conditions Caudate-putamen, grey and white matter from the frontal cortex, occipital cortex, and cerebellum were collected at autopsy from 22 subjects with a provisional diagnosis of schizophrenia (police report of death to the Coroner), and from 11 subjects with no history of psychiatric illness (controls). Collection of tissue for the purpose of this study was approved by the VIFM ethics approval process. The tissue was rapidly frozen to 70 8C and stored at this temperature until required. Sub-cortical structures were identified from an atlas of the human brain. The cortical regions were based on a cytoarchitectural system devised by Brodmann in the early 1900s [10]. Caudate-putamen was chosen because of the high concentration of dopamine receptors in this region and its resulting involvement in schizophrenia. The frontal cortex is also implicated in schizophrenia, since it is regarded as an area of the brain involved in cognitive function. Differ-
LC–MS analysis was performed on a 1100 Series HPLC (Agilent Technologies, Forest Hill, Vic., Australia) configured with both a G1946A mass selective detector (MSD). It was operated in APCI mode as described previously [13]. Briefly, a Zorbax Extend C18 column (2.1 mm 150 mm, 5 mm particle size, Agilent Technologies) was used with a mobile phase of 0.05 M ammonia/methanol/tetrahydrofuran (32.5:67:0.5) at pH 10.5. For quantitation purposes, mass spectral detection was conducted in SIM mode with three ions monitored per drug. 2.6. Sample preparation The extraction technique used was performed based on a procedure described by McIntyre et al. [14] and modified for use with LC–MS. Approximately 2.5 g of each brain
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
123
2.9. Statistical analyses
sample collected at autopsy was added to an equal weight of water and homogenized using an Ultra-turrax T25 ultrasonic probe (IKA Labortechnik, Staufen, Germany). One milliliter of the resulting homogenate, 10 ng trazodone (internal standard), and 0.2 M sodium carbonate buffer (1 mL) were mixed and extracted with butyl chloride (6 mL) after centrifuging at 3000 rpm for 5 min. The organic fraction was back-extracted with 0.2% phosphoric acid (100 mL), which was then basified with 28% ammonia solution and saturated ammonium carbonate and reextracted with chloroform (100 mL). The chloroform fraction was evaporated to dryness and reconstituted in mobile phase. A volume of 30 mL of this solution was injected for LC–MS analysis.
3. Results
2.7. Routine toxicology testing
3.1. Subject demographics
Routine toxicological testing was conducted on urine (or blood if urine was unavailable) for drugs of abuse (EMIT, Dade Behring, Australia), blood for drugs of abuse and any basic/neutral drugs using established routine Victorian Institute of Forensic Medicine procedures. Specimens were also tested for the presence of alcohol by GC. Where screening tests showed the presence of other drugs (including benzodiazepines, opiates, antidepressants, and/ or antipsychotics), standard confirmation and quantitation analyses were conducted using GC-MS SIM or HPLC photodiode techniques.
Of the 22 schizophrenic subjects included in this study, there were approximately twice as many males as females. The average age was 45 4.0 years. Table 1 lists demographic and toxicologic data in blood of all subjects. Data includes the last recorded antipsychotic drug(s) and, where available, amount of time between the last dose and death. Unfortunately, the estimated time between dose and death was unavailable in 17 of the 22 cases. In 3 of the 17 cases, no information on drug histories was available. In the remaining cases, it was not possible to be certain if the subjects had received medication close to death. However, a low blood concentration of antipsychotic drugs relative to the accepted therapeutic range suggests that most of the subjects had been non-compliant in the days leading up to their deaths. Antipsychotic drug concentrations in blood and brains from several of the schizophrenic subjects (subjects 5, 6, 9, 14, 16, 18–22) were not detected. This includes one trifluoperazine-positive subject (subject 6), thioridazine-positive subject (subject 9), one chlorpromazine-positive subject (subject 16), and all of the haloperidol- and flupenthixolpositive subjects (subjects 5, 14, 18–22). In some of these cases the lack of detected neuroleptic suggests the drug had not been taken for some time before death, as the half-lives of the oral forms of fluphenazine, trifluoperazine, chlorpromazine, and thioridazine are 1–2 days [15]. In the case of the higher potency haloperidol and the depot drug flupenthixol, the lower brain concentrations detected are more likely to be the result of therapeutic use rather than non-compliance, as therapeutic blood concentrations of these drugs (0.5–250 ng/ mL and >2 ng/mL, respectively) are generally much lower than for thioridazine (50–3900 ng/mL) and chlorpromazine (1–750 ng/mL). The detection limits for these drugs were 5– 10 times lower for brain tissue assays using LC–MS than the conventional HPLC UV or photodiode array assays used to obtain drug concentrations in blood. However, for the higher potency drugs the subjects in this study would have had to have taken doses far above those used for therapeutic benefit in order to be detected in blood or brain. Because of this, such data was not included for comparison between subjects.
2.8. Comparison of analytical results This method was validated by analyzing a series of five replicate brain specimens spiked with 100 ng/mL standards of the target antipsychotics. Results were evaluated on the basis of drug recovery, intra- and inter-assay precision (measured as coefficients of variation, or CVs), and accuracy. Recovery was calculated by comparing the peak area of an extracted brain specimen spiked with a known drug concentration to that of unextracted standards of known concentrations and expressed as a percentage. CVs were calculated using the standard deviation (s) and mean (x) of replicate analyses of specimens spiked with known concentrations using the formula CV = (s/x) 100. Accuracy was calculated by dividing the measured concentration by the calculated concentration and multiplying the result by 100. The LOQ for mass spectral detection was defined as the concentration that consistently resulted in CVs <20% in SIM mode for target drugs. The data from each brain region was normalized for each subject by expressing concentration as a percentage of the cerebellum concentration. Brain distribution of chlorpromazine, fluphenazine sulfoxide, and trifluoperazine was compared for inter-subject variability. Data from each brain region in thioridazine-positive subjects was separated and the parent drug data compared to that of the two principal metabolites, mesoridazine and sulforidazine.
Statistical evaluation of these data was performed using SPSS V9.0.1 for Windows software on an IBM personal computer. Parametric Pearson correlation tests at the 95% confidence interval were used to determine correlation of concentrations measured in each brain region to those in blood for thioridazine and its metabolites and chlorpromazine. Pearson correlation tests were also used to determine the extent of correlation between regional drug concentrations and Vd, extent of protein binding, and lipophilicity.
124
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
Table 1 Demographic data, antipsychotic drug history, and tissue concentrations in schizophrenic subjects Sex
Age (year)
Cause of death
Last recorded dose of antipsychotic drug
CPZ eqa
Drugs detected in blood
Concentrationb
PMI (h)
1
M
55
Coronary arterial thrombosis
Thioridazine
400
Detected
25
2
F
81
Aspiration of food
Trifluoperazine
100
Benzodiazepines (urine) Thioridazine Mesoridazine Sulforidazine Temazepam Thioridazine Mesoridazine Sulforidazine Diazepam Nordiazepam No blood toxicology
0.2 0.2 0.1 0.1 1.2 0.5 0.2 0.2 0.2 –
610
Thioridazine Mesoridazine Sulforidazine No drugs detected
6.8 0.3 0.13 –
700
Chlorpromazine
0.75
Lithium Paracetamol Chlorpromazine Ethanol Amitriptyline Nortriptyline Clonazepam 7-Aminoclonazepam 7-Aminoflunitrazepam
1.9 5.0 0.10 0.11 g/dL 2.0 1.6 0.06 0.2 0.14 0.1 18 urine 0.3 urine 4.0 urine 13 3.0 0.05 urine 2.2 10 0.1 0.1
ID
3
M
71
Aspiration of food
Thioridazine
150
4
M
47
CAD
530
5
M
44
CAD
Fluphenazine decanoatethioridazine Thioridazine
6
M
42
Drowning
7
F
48
Pulmonary thromboembolism
Flupenthixol decanoate chlorpromazine Fluphenazine decanoate chlorpromazine
8
F
48
Mixed drug toxicity
History refused
9
M
29
Heroin toxicity
No drug history available
10
F
43
Pneumonia
No drug history available
11 12
F M
68 65
CAD Bronchopneumonia
Trifluoperazine Trifluoperazine Haloperidol decanoate
13
14
M
M
53
41
Coronary arterial thrombosis
Mixed drug toxicity
Trifluoperazine chlorpromazine
Fluphenazine decanoate trifluoperazine
600
400 460
Chlorpromazine Morphine 6-Acetylmorphine Codeine Carbamazepine Paracetamol Cannabinoids Chlorpromazine Moclobemide Diazepam Nordiazepam No blood toxicology
25
48
33
32 35
53
65
0.0
51
42
300
No drugs detected Diazepam Nordiazepam Paracetamol
0.1 0.1 5.0
42
500
Fluoxetine Norfluoxetine Trifluoperazine Temazepam Diazepam
0.1 0.1 0.11 4.1 2.0
Nordiazepam Dothiepin
0.3 6.5
6.2
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
125
Table 1 (Continued ) ID
Sex
Age (year)
Cause of death
Last recorded dose of antipsychotic drug
15
M
22
Pericarditis
Flupenthixol, 8-days before death; suppl.: trifluoperazine
16
M
35
17
F
30
18
F
35
19
M
20 21
22
a b
Drugs detected in blood
Concentrationb
PMI (h)
450
Diazepam Nordiazepam
Trace Trace
37
Cannabinoids
Trace
Fluphenazine decanoate
400
No drugs detected
600
No blood toxicology
–
48
300
Cannabinoids
0.075
15
38
Coronary arterial thrombosis Mediastinitis
Flupenthixol, 4 days before death Haloperidol Haloperidol decanoate
160
19
Unascertained
Haloperidol, time of death
750
0.2 <0.1 <0.1 3 5 Detected 0.1 0.1
40
M
Morphine Diazepam Nordiazepam Paracetamol Frusemide Metronidazole Diazepam Nordiazepam
43
M
23
Multiple injuries
Haloperidol, 1 year, 3 weeks before death
300
Diazepam
<0.05
78
500
Nordiazepam Carboxyhemoglobin Desipramine 7-Aminoclonazepam
0.1 90% sat 0.5 0.07
52
Benztropine
Trace
M
26
Perforated gastric ulcer Hanging
CPZ eqa
Carbon monoxide poisoning
Haloperidol decanoate, 2 days before death
47
CPZ eq: chlorpromazine dose equivalents. Concentrations in mg/L, unless otherwise specified. CAD: coronary artery disease; PMI: post-mortem interval.
3.2. Case types and other drugs detected in blood from antipsychotic-positive subjects In antipsychotic positive cases one or more commonly abused drugs were detected. In particular, benzodiazepines were detected in 11 cases, and opioids and cannabinoids were detected in two cases each. Antidepressants of varying types (SSRIs, MAOIs, and TCAs) were present in blood of five subjects. The manner of death in the studied cases spanned the four major categories. There were 12 deaths ascribed to natural causes, comprising the largest number of cases (subjects 1, 3, 6–8, 11–15, 18, and 19). There were three accidents (subjects 9, 10, and 21), three suicides not involving drugs (subjects 2, 17 and 22), and three cases of mixed drug toxicities (subjects 4, 5, and 16). The cause of death of subject 20 could not be determined. 3.3. Regional brain distribution of antipsychotic drugs The thioridazine-positive cases revealed the most information out of all antipsychotic drugs studied, since the metabolites mesoridazine and sulforidazine were measured in
addition to parent drug in all five cases. Table 2 shows substantial inter-subject variability in the brain distribution of thioridazine concentrations. In subjects 3 and 4, the highest concentration of thioridazine, mesoridazine, and sulforidazine were detected in the caudate putamen. This was true for only the metabolites in case 2, the highest parent drug concentration being detected in the white matter of the frontal cortex. This region had the lowest concentrations of each compound in subjects 2 and 5. Because the concentrations of all three compounds were not detected in the remaining regions for subject 1, this case was not included in further analyses. Subject 5 had the highest parent and metabolite concentrations in the occipital cortex, the region with the second highest concentration in the remaining subjects. Brain region concentrations of chlorpromazine detected in subjects 1 and 6–10 are shown in Table 2. Because chlorpromazine was not detected in any brain regions for subject 9, this case was not included in further data analysis. These data show that like the thioridazine cases, a great deal of inter-subject variability in brain distribution of chlorpromazine was observed. In all subjects positive for the drug, the lowest concentration was found in the cerebellum, and in four of the five remaining subjects the highest concentration
126
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
Table 2 Summary of brain neuroleptic concentrationsa,b Subject
CPu
FCx (grey)
FCx (white)
OCx
Cerebellum
10 (M:10, S: 10) 240 (M: 10, S: 10) 3100 (M: 430, S: 20) 870 (M: 260, S: 10) 73000 (M: 21000, S: 3800)
10 (M:10, S: 10) 270 (M: 10, S: 10) 1000 (M: 320, S: 10) 10 (M: 260, S: 10) 12000 (M: 5300, S: 770)
20 (M:10, S: 10) 260 (M: 20, S: 110) 2000 (M: 330, S: 10) 10 (M: 260, S: 10) 91000 (M: 26000, S: 4600)
10 (M:10, S: 10) 210 (M: 10, S: 10) 1600 (M: 230, S: 10) 10 (M: 150, S: 10) 51000 (M: 18000, S: 2900)
Chlorpromazine 1 440 6 70 7 6800 8 1700 9 11 10 14000
580 70 1500 10 <10 14000
630 20 1500 20 <10 13000
590 20 1600 10 <10 14000
310 10 1100 10 <10 9900
Trifluoperazine 2 10 11 10 12 10 13 10 14 9100
100 1100 140 60 9100
100 580 90 60 5400
150 1200 190 60 9300
100 570 150 50 8100
490 600
400 470
470 570
410 400
Thioridazine 1 90 (M: 2 40 (M: 3 20000 (M: 4 1200 (M: 5 9500 (M: S: 560)
30, S: 20) 530, S: 250) 2100, S: 610) 430, S: 10) 3500,
Fluphenazine sulfoxide 15 360 16 10 a b
All concentrations in ng/mg; thioridazine data includes data for two metabolites mesoridazine (M) and sulforidazine (S) in parentheses. CPu: caudate-putamen; FCx: frontal cortex; OCx: occipital cortex.
was in the caudate-putamen. However, the relative concentrations in occipital cortex and grey and white matter from frontal cortex differed between the subjects. Table 2 summarizes the regional brain concentrations detected for each trifluoperazine-positive case. Trifluoperazine was found at highest concentration in occipital cortex and was lowest in caudate-putamen, but again there was some degree of inter-subject variability in results obtained from grey and white matter from the frontal cortex and the cerebellum. Since fluphenazine was not available from any Australian supplier, its presence was determined by measuring the concentration of its major metabolite, fluphenazine sulfoxide. This was deemed to be adequate as a study of fluphenazine and metabolite concentrations in rat brain showed fluphenazine sulfoxide concentrations ranged 24–96% of parent drug concentrations [4]. As shown in Table 2, tissue concentrations in subjects 15 and 16 ranged (from highest to lowest) in the following manner: grey matter from the frontal cortex, occipital cortex, white matter from the frontal cortex/ cerebellum, and caudate-putamen. 3.4. Correlation of blood antipsychotic concentrations to those of brain tissue There were seven cases in which target antipsychotic drugs were detected in both blood and brain (subjects 2, 3, 5,
7, 8, 10, and 14). Of these cases, subjects 7, 8, and 10 were chlorpromazine-positive, subjects 2, 3, and 5 were thioridazine-positive, and trifluoperazine was detected in subject 14. Because of the range of drugs detected, there was insufficient data to test for correlation between blood and brain concentrations for individual drugs. Therefore, the seven cases for which both types of specimens had been analyzed were combined and the extent of correlation between the two specimen types calculated. Significant correlations with blood were observed for concentrations in grey matter of the frontal cortex (r2 = 0.944, P < 0.001), occipital cortex (r2 = 0.937, P < 0.001), and cerebellum (r2 = 0.927, p < 0.001). If combined regional concentrations were expressed in terms of those in cerebellum, much lower correlations were observed, the only significant one being with occipital cortex (r2 = 0.783, p < 0.01). Mean absolute drug concentrations (Table 3) in caudateputamen exhibited a positive, significant correlation with lipophilicity (measured as log p) (r2 = 0.949, p < 0.05). Additionally, when corrected for those in cerebellum, mean drug concentrations in occipital cortex were positively correlated with Vd (r2 = 0.994, p < 0.01). However, this correlation is likely skewed by the high Vd of fluphenazine (Table 4) and therefore was not reflective of their actual relationship. No particular correlation between regional drug concentrations and any other parameters were observed.
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
127
Table 3 Mean absolute regional drug concentrations and pharmacokinetic parameters of psychiatric drugsa Drug
CPu
FCx (grey)
FCx (white)
OCx
Cere
pKa
log p
Vd
Fb
Chlorpromazine Fluphenazine Thioridazine Trifluoperazine
3800 5600 190 250 1800 4100 6200 8700
2700 5600 550 78 2100 3900 15000 32000
5200 2500 440 49 6200 2300 2700 5200
2700 690 520 71 2200 4000 19000 40000
1900 3900 410 7.0 1800 3500 11000 23000
9.30 8.10 9.50 8.10
4.55 3.46 5.13 4.26
23 220 18 13
0.98 0.99 0.90 0.95
All concentrations expressed in ng/g. pKa’s and log p’s calculated using CompuDrug Software [16]. CPu: caudate-putamen; FCx: frontal cortex; OCx: occipital cortex; Cere: cerebellum; Fb: fraction protein bound. a Mean S.D.
Table 4 Patterns of regional brain distribution for thioridazine Pattern
Cases exhibiting pattern
Relative sites of concentration (highest to lowest)
A
2
B
3, 4
C
5
FCx (white) > OCx>FCx (grey) > cerebellum > CPu CPu > FCx (grey) > OCx/FCx (white)/cerebellum OCx > FCx(grey) > cerebellum > FCx (white) > CPu
CPu: caudate-putamen; FCx: frontal cortex; OCx: occipital cortex; Cere: cerebellum.
3.5. Blood versus brain concentrations In our study no fluphenazine was detected in the blood of subjects whose histories had suggested this drug had been administered (Table 1), and hence no comparison of brain and blood concentrations could be made. The same is true for the trifluoperazine-positive subjects (n = 4), among whom trifluoperazine was only detected in blood from subject 6. However, as long as these subjects had been taking these drugs therapeutically it is not likely they would have been detected in their blood, as the detection limits for these drugs using conventional assays were about two times the highest expected therapeutic concentration for either of these drugs. It is of interest that the last recorded antipsychotic taken by this subject were haloperidol and flupenthixol decanoate and not trifluoperazine, although only trifluoperazine was detected in both blood and brain tissue from the subject. However, one study of fluphenazine and metabolites in rat brain and other tissues showed that concentrations of fluphenazine in the brain and plasma correlated more closely than those of its metabolites [4]. The correlation of plasma concentrations was strongest for caudate-putamen, cortex, and hippocampus (r2 = 0.83, 0.68, and 0.70, respectively). By contrast, they found no significant correlation between plasma and brain concentrations of fluphenazine sulfoxide for any brain region. Among the four subjects in whom chlorpromazine was detected in brain, it was also detected in the blood in all but subject 12. Our chlorpromazine data indicates the drug must also accumulate more in brain than blood. The chlorproma-
zine concentrations in subjects 13 and 15 were higher in all brain regions analyzed than in blood (750 and 2200 mg/L in blood, respectively), but in subject 15 only the caudate putamen concentration was higher (1700 mg/kg compared to 100 mg/L). The brain region in which measured concentrations correlated most closely with blood levels for chlorpromazine was caudate-putamen (r2 = 0.99), while the lowest correlation region was found with both grey and white matter from frontal cortex (r2 = 0.96 for both brain regions). Consistent with our chlorpromazine and trifluoperazine data, thioridazine also seems to accumulate at higher concentrations in brain than in blood. As with chlorpromazine, thioridazine and metabolites were detected in blood as well as brain in subjects 3, 8, and 11 (Table 1). In subject 11, the concentrations of all three compounds were higher in brain than in blood. In subject 8, only the caudate putamen concentration was higher than that of blood for all three compounds. The caudate putamen concentration in subject 3 was higher than in blood for the two metabolites, but not the parent compound. The caudate-putamen concentrations showed the lowest correlation to blood data for all three compounds (r2 = 0.01, 0.85, and 0.26, for thioridazine, mesoridazine, and sulforidazine, respectively). White matter from the frontal cortex showed the highest correlation with blood data for both thioridazine and mesoridazine (r2 = 0.99 for both compounds), whereas sulforidazine blood data correlated most with those found in the cerebellum (r2 = 0.99).
4. Discussion 4.1. Regional brain distribution The regional brain distribution for each compound using mean absolute concentrations is shown in Fig. 1. Fig. 2 shows this distribution with concentrations normalized for those in cerebellum. As shown in these figures, three distinct patterns of distribution were observed in this set of experiments (see also Table 4). Chlorpromazine, a low-potency phenothiazine with an aliphatic side chain, was found at highest concentrations in caudate-putamen and lowest in cerebellum (group 1). Thioridazine and its metabolites
128
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
Fig. 1. Mean absolute regional brain distribution of selected antipsychotic, grouped according to structural features. Group 1: aliphatic side chain, group 2: piperidine ring in side chain, and group 3: piperazine ring in side chain.
(group 2) were also detected in caudate-putamen at highest concentrations, while the lowest mean concentrations of these compounds were detected in white matter of the frontal cortex. Phenothiazines in group 2 possess a piperidine ring in the side chain and are considered atypical antipsychotics because they are associated with fewer extrapyramidal side effects. It would appear, therefore, that antipsychotics belonging to groups 1 and 2 accumulate in regions of high dopamine receptor concentration. The highest and lowest concentrations of both fluphenazine and trifluoperazine (group 3) were found in occipital cortex and caudate-putamen, respectively. Both of these compounds have a piperazine ring in the side chain and are known to possess anticholinergic activity, thus carrying an increased risk of extrapyramidal side effects in comparison to other antipsychotics. The comparatively lower relative concentrations of both fluphenazine sulfoxide and trifluoperazine may be a result of their comparatively lower affinity for dopamine receptors or their higher affinity for peripheral cholinergic receptors. These data partially agree with the findings of Aravagiri et al. in rat brain, in which the lowest fluphenazine concentrations were found in caudate-putamen [4]. Like their study in rats, the brain fluphenazine sulfoxide concentrations detected in subjects 15 and 16 showed little regional variation in comparison to the other targeted anti-
psychotics. However, as was observed in the presented thioridazine-positive cases, it may be that fluphenazine and its metabolites exhibit different regional brain distribution. In the presented data, it appears that the regional distribution of thioridazine depends on overall drug concentrations. For example, subject 5, who had the highest drug concentrations out of all thioridazine-positive cases, had concentrations in occipital cortex (an area of the brain not thought to be involved in schizophrenia) that were higher than those in other regions. By contrast, subjects 3 and 4, who had moderate overall thioridazine concentrations relative to the other cases, had higher regional concentrations in caudate-putamen (an area of high dopamine receptor concentration). This suggests that antipsychotics first distribute into regions where dopamine receptors and their neurons are concentrated, until active sites are saturated. After reaching this point, the drugs may then begin to distribute more into regions possessing fewer receptors or neurons or bind to less specific sites. It should also be noted that mesoridazine and sulforidazine followed the same pattern of distribution as thioridazine in cases 3–5, but their pattern of distribution differed in subject 9, in which concentrations of all three compounds were significantly lower. This may indicate that although thioridazine and its metabolites appear to distribute in the same way when concentrations are high, the distribution of
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
129
Fig. 2. Mean regional brain distribution of selected antipsychotics relative to cerebellum concentrations and grouped according to structural features. Group 1: aliphatic side chain, group 2: piperidine ring in side chain, and group 3: piperazine ring in side chain.
each compound begins to differentiate upon cessation of drug use, reflecting a possible differential rate of drug elimination between brain regions. It is not possible to say whether the variations observed in regional thioridazine distribution in this study are reflective of differences in time between last dose and death, since this information was not available for any thioridazine-positive subjects. Svendsen et al. have found that no one area of the brain selectively accumulated thioridazine or its metabolites in subjects known to have taken their medication within 8 h or less before death [3]. However, they did find that in one subject who had been taken off medication (800 mg/day) 4 days before death, there was a more specific distribution of thioridazine and its metabolites in different brain regions. Most significant was the difference between cortical and subcortical structures in which the former had high thioridazine and lower mesoridazine concentrations, whereas the latter had high mesoridazine and lower thioridazine concentrations. 4.2. Tissue-to-femoral blood correlations In all seven cases where both blood and brain tissue specimens were positive for antipsychotic drugs, they were detected at concentrations which were 70% higher on average in brain than in blood. This is in accord with findings
from earlier studies [3,4]. Hence, published data on blood therapeutic and toxic concentrations can generally be expected to be lower than those in brain. However, it was interesting to note that several brain regions (including grey matter from frontal cortex, occipital cortex, and cerebellum) exhibited positive significant correlations with blood (r2 > 0.92, p < 0.001). This suggests antipsychotic drug concentrations in these brain regions under some circumstances show some proportion to femoral blood concentrations, and vice-versa. It appears from these results that caudate-putamen may be the best region to sample for determination of fluphenazine or trifluoperazine in terms of likelihood of the drug’s presence at measurable concentrations. In the case of either chlorpromazine or thioridazine, however, more data is needed to understand what appears to be a complex distribution profile in postmortem schizophrenic brains. These data allow two important conclusions to be made. First, low or undetectable concentrations of antipsychotic drug in blood generally correspond to low concentrations in selected regions of the CNS. Second, residual antipsychotic drugs are absent in a large proportion of the brains using detection limits of 10 ng/g (45%). It is notable that in many of the cases there was a difference between those drugs detected in blood and their
130
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
detection in brain. Specifically, although case histories indicated a particular antipsychotic drug should have been present, they were often not detected in blood but often were detected in at least one brain region. This is most likely due to two main factors. Firstly, the brain data was obtained using an LC–MS method that specifically targets the antipsychotic drugs in this study with a low detection limit (10 ng/g), whereas the blood toxicology data was collected using older, routine GC screening methods with higher detection limits (50–100 ng/mL). Neither the decanoates of flupenthixol and fluphenazine nor haloperidol is detected routinely with these routine GC techniques. Unfortunately, due to time limits imposed on specimen holding, the blood was unable to be re-analyzed using the LC–MS method. Secondly, in cases where the antipsychotic drug was found in the blood as well as brain, brain concentrations of drug were higher than those found in blood. Therefore, in cases where the antipsychotic was detected in brain at low concentrations but not detected in blood, its blood concentration may have been below the assay LOD. In conclusion, these data indicate different antipsychotic drugs preferentially distribute to different regions of the brain. Importantly, the distinct patterns of distribution for each of the three types of phenothiazines appear to correspond with their relative affinity for dopamine receptors and the regional brain concentration of such receptors. In addition, it appears from the presented data that regional distribution of antipsychotics may be time-dependent, such that different patterns of distribution are observed at differing points in therapy (commencement of treatment, maintenance, and withdrawal of medication). More data are needed to confirm these findings. Future experiments that focus on the regional distribution of antipsychotics not represented in this study would be especially useful. Additionally, more information about the possibility of differential clearance rates of parent drugs and metabolites from brain tissue would be of value.
Acknowledgments This work was supported in part by the NH&MRC Schizophrenia Unit, the State Government of Victoria, the United States Air Force, The Brockhoff Foundation and the Rebecca L. Cooper Medical Research Foundation. Brian Dean is the recipient of a NARSAD Young Investigator award. The authors wish to thank Professor Nicholas Keks and Ms. Christine Hill for carrying out case-history reviews and diagnostic confirmation, and the Victorian Institute of Forensic Medicine for access to tissue and to staff for assistance with tissue collection. Finally, we thank Agilent Australia for loan of the LC–MS.
References [1] B.M. Cohen, J.F. Lipinski, C. Waternaux, A fixed dose study of the plasma concentration and clinical effects of thioridazine and its major metabolites, Psychopharmacology (Berl) 97 (4) (1989) 481–488. [2] T. Tsuneizumi, S.M. Babb, B.M. Cohen, Drug distribution between blood and brain as a determinant of antipsychotic drug effects, Biol. Psychiatry 32 (9) (1992) 817–824. [3] C.N. Svendsen, C.C. Hrbek, M. Casendino, R.D. Nichols, E.D. Bird, Concentration and distribution of thioridazine and metabolites in schizophrenic post-mortem brain tissue, Psychiatry Res. 23 (1) (1988) 1–10. [4] M. Aravagiri, S.R. Marder, A. Yuwiler, K.K. Midha, N.S. Kula, R.J. Baldessarini, Distribution of fluphenazine and its metabolites in brain regions and other tissues of the rat, Neuropsychopharmacology 13 (3) (1995) 235–247. [5] T.C. Merrick, A. Felo, A.J. Jenkins, Tissue distribution of olanzapine in a postmortem case, Am. J. Forensic Med. Pathol. 22 (2001) 270–274. [6] L.E. van Beijsterveldt, R.J. Geerts, J.E. Leysen, A.A. Megens, H.M. Van den Eynde, W.E. Meuldermans, et al. Regional brain distribution of risperidone and its active metabolite 9-hydroxyrisperidone in the rat, Psychopharmacology (Berl) 114 (1) (1994) 53–62. [7] T.C. Merrick, J.A. Felo, A.J. Jenkins, Tissue distribution of olanzapine in a postmortem case, Am. J. Forensic Med. Pathol. 22 (3) (2001) 270–274. [8] F.P. Bymaster, D.O. Calligaro, J.F. Falcone, R.D. Marsh, N.A. Moore, N.C. Tye, et al. Radioreceptor binding profile of the atypical antipsychotic olanzapine, Neuropsychopharmacology 14 (2) (1996) 87–96. [9] R.R. Conley, C.A. Tamminga, D.L. Kelly, C.M. Richardson, Treatment-resistant schizophrenic patients respond to clozapine after olanzapine non-response, Biol. Psychiatry 46 (1) (1999) 73–77. [10] L.J. Garey (Ed.), Brodman’s Localisation in the Cerebral Cortex, Smith-Gordon, London, 1994. [11] N.A. Keks, K. Hill, K. Opeskin, et al. Psychiatric diagnosis after death: the problems of accurate diagnosis, in: B. Dean, T.M. Hyde, J. Kleinman (Eds.), The Use of CNS Autopsy Tissue in Psychiatric Death, Gordon & Breach Science Publishers, Sydney, 1999, pp. 19–37. [12] P. Foster, Neuroleptic equivalence, Pharmacol. J. 243 (1989) 431–432. [13] K.E. Goeringer, M. McIntyre, O.H. Drummer, LC–MS analysis of serotonergic drugs, J. Anal. Toxicol. 27 (1) (2003) 30– 35. [14] I.M. McIntyre, M.L. Syrjanen, K. Crump, S. Horomidis, A.W. Peace, O.H. Drummer, Simultaneous HPLC gradient analysis of 15 benzodiazepines and selected metabolites in postmortem blood, J. Anal. Toxicol. 17 (4) (1993) 202–207. [15] R.C. Baselt, Disposition of Toxic Drugs and Chemicals in Man, 5th ed., Year Book Medical Publishers, 2000. [16] C. PALLAS Expert System for Today’s Chemists, CompDrug Internationsl Inc., San Francisco, 1999.
Brain distribution of selected antipsychotics in schizophrenia Kabrena E. Rodda a, Brian Dean b,c,d,e, Iain M. McIntyre a,1, Olaf H. Drummer a,* a
Victorian Institute of Forensic Medicine, Department of Forensic Medicine, Monash University, 57-83 Kavanagh Street, Southbank, Vic. 3006, Australia b The Mental Health Research Institute of Victoria, Vic., Australia c The University of Melbourne Department of Psychiatry, Parkville, Vic., Australia d The University of Melbourne Department of Pathology, Parkville, Vic., Australia e Department of Psychological Medicine, Monash University, Clayton, Vic., Australia Received 23 July 2004; accepted 17 March 2005 Available online 4 June 2005
Abstract The brain distribution of phenothiazine antipsychotics in 22 confirmed schizophrenic and 11 control subjects were collected at autopsy. Specimens were homogenized, extracted with n-butyl chloride, and analyzed via liquid chromatography–mass spectrometry, using atmospheric pressure electrospray ionization operating in the positive mode. Drug concentrations normalized for those observed in cerebellum showed three distinct patterns of distribution corresponding to different structural features of each type of phenothiazine. Those drugs with high affinity for dopamine receptors were detected in the highest concentrations in regions with high concentrations of such receptors. However, those associated with relatively lower dopaminergic activity were found in the highest concentration in the occipital cortex, a region with a relatively low concentration of dopamine receptors. The regional brain distribution of thioridazine and its metabolites was concentration dependent. These results have implications for determining the role of these drugs in the sudden and unexpected deaths of schizophrenics. # 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Post-mortem; LC–MS; Antipsychotics; Brain distribution; Schizophrenia
1. Introduction There is a lack of studies examining how different antipsychotic drugs distribute in the brain at death. Due to differences in drug measurement between studies, comparison of results within and between drugs is difficult. One study of plasma concentrations of thioridazine and metabolites in humans found that although drug concentrations varied widely between patients, in all patients the relative concentration of thioridazine to mesoridazine was about one * Corresponding author. Tel.: +61 3 9684 4334; fax: +61 3 9682 7353. E-mail address: [email protected] (O.H. Drummer). 1 Present address: County of San Diego Medical Examiner’s Office, 5555 Overland Ave., Bldg. 14, San Diego, CA 92123, USA.
half and thioridazine to sulforidazine was two-fold [1]. In a study of antipsychotic drug distribution between blood and brain in rats [2], the observed ratio of brain-to-serum concentration varied widely, but the high potency drugs fluphenazine and haloperidol had more favorable brain-to-blood distribution than the low potency drug thioridazine. Svendsen et al published the only study that reports postmortem concentrations of thioridazine and its metabolites and chlorpromazine in schizophrenic brains [3]. A review of the clinical histories revealed that five subjects had received thioridazine within 8 h of death. In only one subject who had been withdrawn from medication 4 days before death after chronic treatment, thioridazine and its metabolites selectively accumulated in particular regions of the brain. The most remarkable difference was between cortical and sub-cortical structures, in which the former
0379-0738/$ – see front matter # 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2005.03.017
122
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
had high thioridazine levels and lower mesoridazine levels, whereas the latter had high mesoridazine and lower thioridazine concentrations. Although these data are limited, there is evidence to suggest that most anti-psychotics distribute preferentially into the frontal cortex, midbrain, and caudateputamen [3–6], and a degree of differential distribution between the left and right brain hemispheres may occur [5,7]. How a drug distributes in brain would not only seem to be important for post-mortem studies but could provide insight as to how different drugs may distribute in the living brain. Differential distribution may account for why drugs such as clozapine and olanzapine, that have similar receptor pharmacology [8], can have differing therapeutic outcomes [9]. The purpose of this study was to examine whether selected antipsychotic drugs distribute preferentially to particular brain regions in schizophrenic subjects, and to investigate the extent of correlation of blood concentrations with those concentrations in different brain regions. This will allow a better understanding of the interpretation and causes of sudden and expected death in this population.
entiation between grey and white matter from the frontal cortex was made to determine whether the antipsychotic drugs accumulate preferentially in grey matter, where the dopamine neurons are located. Similarly, the occipital cortex was included as a form of negative control (it is the cortical region which is not thought to be involved in schizophrenia). Finally, cerebellum was included to investigate whether antipsychotic drug distribution was related to receptor location. Although GABAA receptors and a1A-adrenoceptors are found in cerebellum, there are very few dopamine or serotonin receptors. 2.3. Determination of postmortem interval (PMI)
2. Materials and methods
Where death did not involve suicide, tissue was collected from subjects whose death was witnessed and the postmortem interval (PMI) was the time of death to autopsy. With suicides, tissue was only taken from individuals who had been seen alive up to 5 h before being found deceased. In those cases, the PMI was estimated at half way between the donor being found dead and being last seen alive. In all cases, the cadavers were refrigerated within 5 h of being found.
2.1. Materials
2.4. Diagnosis of schizophrenia
Mesoridazine was obtained from the Division of Analytical Laboratories of the New South Wales Health Department (Lidcombe, New South Wales). Sulforidazine and fluphenazine sulfoxide were obtained from the Australian Government Analytical Laboratory (South Melbourne, Victoria). Thioridazine, haloperidol, chlorpromazine, and trazodone, (internal standard), were purchased from Sigma Aldrich Pty Ltd. (Castle Hill, New South Wales). All other reagents were HPLC grade and were obtained from Sigma Aldrich Pty Ltd.
The provisional diagnosis of schizophrenia was confirmed by a senior psychologist and senior psychiatrist after an extensive case history review [5]. These reviews were carried out using a structure instrument called Diagnostic Instrument for Brain Studies (see Keks et al. for a review [11]). In this study all diagnoses were confirmed using DSMIV criteria (APA). In addition, information on the type and amount of antipsychotic drugs prescribed close to death was obtained from the case history and the final recorded dose of antipsychotic drug was converted to chlorpromazine dose equivalents [12].
2.2. Tissue collection 2.5. LC–MS conditions Caudate-putamen, grey and white matter from the frontal cortex, occipital cortex, and cerebellum were collected at autopsy from 22 subjects with a provisional diagnosis of schizophrenia (police report of death to the Coroner), and from 11 subjects with no history of psychiatric illness (controls). Collection of tissue for the purpose of this study was approved by the VIFM ethics approval process. The tissue was rapidly frozen to 70 8C and stored at this temperature until required. Sub-cortical structures were identified from an atlas of the human brain. The cortical regions were based on a cytoarchitectural system devised by Brodmann in the early 1900s [10]. Caudate-putamen was chosen because of the high concentration of dopamine receptors in this region and its resulting involvement in schizophrenia. The frontal cortex is also implicated in schizophrenia, since it is regarded as an area of the brain involved in cognitive function. Differ-
LC–MS analysis was performed on a 1100 Series HPLC (Agilent Technologies, Forest Hill, Vic., Australia) configured with both a G1946A mass selective detector (MSD). It was operated in APCI mode as described previously [13]. Briefly, a Zorbax Extend C18 column (2.1 mm 150 mm, 5 mm particle size, Agilent Technologies) was used with a mobile phase of 0.05 M ammonia/methanol/tetrahydrofuran (32.5:67:0.5) at pH 10.5. For quantitation purposes, mass spectral detection was conducted in SIM mode with three ions monitored per drug. 2.6. Sample preparation The extraction technique used was performed based on a procedure described by McIntyre et al. [14] and modified for use with LC–MS. Approximately 2.5 g of each brain
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
123
2.9. Statistical analyses
sample collected at autopsy was added to an equal weight of water and homogenized using an Ultra-turrax T25 ultrasonic probe (IKA Labortechnik, Staufen, Germany). One milliliter of the resulting homogenate, 10 ng trazodone (internal standard), and 0.2 M sodium carbonate buffer (1 mL) were mixed and extracted with butyl chloride (6 mL) after centrifuging at 3000 rpm for 5 min. The organic fraction was back-extracted with 0.2% phosphoric acid (100 mL), which was then basified with 28% ammonia solution and saturated ammonium carbonate and reextracted with chloroform (100 mL). The chloroform fraction was evaporated to dryness and reconstituted in mobile phase. A volume of 30 mL of this solution was injected for LC–MS analysis.
3. Results
2.7. Routine toxicology testing
3.1. Subject demographics
Routine toxicological testing was conducted on urine (or blood if urine was unavailable) for drugs of abuse (EMIT, Dade Behring, Australia), blood for drugs of abuse and any basic/neutral drugs using established routine Victorian Institute of Forensic Medicine procedures. Specimens were also tested for the presence of alcohol by GC. Where screening tests showed the presence of other drugs (including benzodiazepines, opiates, antidepressants, and/ or antipsychotics), standard confirmation and quantitation analyses were conducted using GC-MS SIM or HPLC photodiode techniques.
Of the 22 schizophrenic subjects included in this study, there were approximately twice as many males as females. The average age was 45 4.0 years. Table 1 lists demographic and toxicologic data in blood of all subjects. Data includes the last recorded antipsychotic drug(s) and, where available, amount of time between the last dose and death. Unfortunately, the estimated time between dose and death was unavailable in 17 of the 22 cases. In 3 of the 17 cases, no information on drug histories was available. In the remaining cases, it was not possible to be certain if the subjects had received medication close to death. However, a low blood concentration of antipsychotic drugs relative to the accepted therapeutic range suggests that most of the subjects had been non-compliant in the days leading up to their deaths. Antipsychotic drug concentrations in blood and brains from several of the schizophrenic subjects (subjects 5, 6, 9, 14, 16, 18–22) were not detected. This includes one trifluoperazine-positive subject (subject 6), thioridazine-positive subject (subject 9), one chlorpromazine-positive subject (subject 16), and all of the haloperidol- and flupenthixolpositive subjects (subjects 5, 14, 18–22). In some of these cases the lack of detected neuroleptic suggests the drug had not been taken for some time before death, as the half-lives of the oral forms of fluphenazine, trifluoperazine, chlorpromazine, and thioridazine are 1–2 days [15]. In the case of the higher potency haloperidol and the depot drug flupenthixol, the lower brain concentrations detected are more likely to be the result of therapeutic use rather than non-compliance, as therapeutic blood concentrations of these drugs (0.5–250 ng/ mL and >2 ng/mL, respectively) are generally much lower than for thioridazine (50–3900 ng/mL) and chlorpromazine (1–750 ng/mL). The detection limits for these drugs were 5– 10 times lower for brain tissue assays using LC–MS than the conventional HPLC UV or photodiode array assays used to obtain drug concentrations in blood. However, for the higher potency drugs the subjects in this study would have had to have taken doses far above those used for therapeutic benefit in order to be detected in blood or brain. Because of this, such data was not included for comparison between subjects.
2.8. Comparison of analytical results This method was validated by analyzing a series of five replicate brain specimens spiked with 100 ng/mL standards of the target antipsychotics. Results were evaluated on the basis of drug recovery, intra- and inter-assay precision (measured as coefficients of variation, or CVs), and accuracy. Recovery was calculated by comparing the peak area of an extracted brain specimen spiked with a known drug concentration to that of unextracted standards of known concentrations and expressed as a percentage. CVs were calculated using the standard deviation (s) and mean (x) of replicate analyses of specimens spiked with known concentrations using the formula CV = (s/x) 100. Accuracy was calculated by dividing the measured concentration by the calculated concentration and multiplying the result by 100. The LOQ for mass spectral detection was defined as the concentration that consistently resulted in CVs <20% in SIM mode for target drugs. The data from each brain region was normalized for each subject by expressing concentration as a percentage of the cerebellum concentration. Brain distribution of chlorpromazine, fluphenazine sulfoxide, and trifluoperazine was compared for inter-subject variability. Data from each brain region in thioridazine-positive subjects was separated and the parent drug data compared to that of the two principal metabolites, mesoridazine and sulforidazine.
Statistical evaluation of these data was performed using SPSS V9.0.1 for Windows software on an IBM personal computer. Parametric Pearson correlation tests at the 95% confidence interval were used to determine correlation of concentrations measured in each brain region to those in blood for thioridazine and its metabolites and chlorpromazine. Pearson correlation tests were also used to determine the extent of correlation between regional drug concentrations and Vd, extent of protein binding, and lipophilicity.
124
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
Table 1 Demographic data, antipsychotic drug history, and tissue concentrations in schizophrenic subjects Sex
Age (year)
Cause of death
Last recorded dose of antipsychotic drug
CPZ eqa
Drugs detected in blood
Concentrationb
PMI (h)
1
M
55
Coronary arterial thrombosis
Thioridazine
400
Detected
25
2
F
81
Aspiration of food
Trifluoperazine
100
Benzodiazepines (urine) Thioridazine Mesoridazine Sulforidazine Temazepam Thioridazine Mesoridazine Sulforidazine Diazepam Nordiazepam No blood toxicology
0.2 0.2 0.1 0.1 1.2 0.5 0.2 0.2 0.2 –
610
Thioridazine Mesoridazine Sulforidazine No drugs detected
6.8 0.3 0.13 –
700
Chlorpromazine
0.75
Lithium Paracetamol Chlorpromazine Ethanol Amitriptyline Nortriptyline Clonazepam 7-Aminoclonazepam 7-Aminoflunitrazepam
1.9 5.0 0.10 0.11 g/dL 2.0 1.6 0.06 0.2 0.14 0.1 18 urine 0.3 urine 4.0 urine 13 3.0 0.05 urine 2.2 10 0.1 0.1
ID
3
M
71
Aspiration of food
Thioridazine
150
4
M
47
CAD
530
5
M
44
CAD
Fluphenazine decanoatethioridazine Thioridazine
6
M
42
Drowning
7
F
48
Pulmonary thromboembolism
Flupenthixol decanoate chlorpromazine Fluphenazine decanoate chlorpromazine
8
F
48
Mixed drug toxicity
History refused
9
M
29
Heroin toxicity
No drug history available
10
F
43
Pneumonia
No drug history available
11 12
F M
68 65
CAD Bronchopneumonia
Trifluoperazine Trifluoperazine Haloperidol decanoate
13
14
M
M
53
41
Coronary arterial thrombosis
Mixed drug toxicity
Trifluoperazine chlorpromazine
Fluphenazine decanoate trifluoperazine
600
400 460
Chlorpromazine Morphine 6-Acetylmorphine Codeine Carbamazepine Paracetamol Cannabinoids Chlorpromazine Moclobemide Diazepam Nordiazepam No blood toxicology
25
48
33
32 35
53
65
0.0
51
42
300
No drugs detected Diazepam Nordiazepam Paracetamol
0.1 0.1 5.0
42
500
Fluoxetine Norfluoxetine Trifluoperazine Temazepam Diazepam
0.1 0.1 0.11 4.1 2.0
Nordiazepam Dothiepin
0.3 6.5
6.2
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
125
Table 1 (Continued ) ID
Sex
Age (year)
Cause of death
Last recorded dose of antipsychotic drug
15
M
22
Pericarditis
Flupenthixol, 8-days before death; suppl.: trifluoperazine
16
M
35
17
F
30
18
F
35
19
M
20 21
22
a b
Drugs detected in blood
Concentrationb
PMI (h)
450
Diazepam Nordiazepam
Trace Trace
37
Cannabinoids
Trace
Fluphenazine decanoate
400
No drugs detected
600
No blood toxicology
–
48
300
Cannabinoids
0.075
15
38
Coronary arterial thrombosis Mediastinitis
Flupenthixol, 4 days before death Haloperidol Haloperidol decanoate
160
19
Unascertained
Haloperidol, time of death
750
0.2 <0.1 <0.1 3 5 Detected 0.1 0.1
40
M
Morphine Diazepam Nordiazepam Paracetamol Frusemide Metronidazole Diazepam Nordiazepam
43
M
23
Multiple injuries
Haloperidol, 1 year, 3 weeks before death
300
Diazepam
<0.05
78
500
Nordiazepam Carboxyhemoglobin Desipramine 7-Aminoclonazepam
0.1 90% sat 0.5 0.07
52
Benztropine
Trace
M
26
Perforated gastric ulcer Hanging
CPZ eqa
Carbon monoxide poisoning
Haloperidol decanoate, 2 days before death
47
CPZ eq: chlorpromazine dose equivalents. Concentrations in mg/L, unless otherwise specified. CAD: coronary artery disease; PMI: post-mortem interval.
3.2. Case types and other drugs detected in blood from antipsychotic-positive subjects In antipsychotic positive cases one or more commonly abused drugs were detected. In particular, benzodiazepines were detected in 11 cases, and opioids and cannabinoids were detected in two cases each. Antidepressants of varying types (SSRIs, MAOIs, and TCAs) were present in blood of five subjects. The manner of death in the studied cases spanned the four major categories. There were 12 deaths ascribed to natural causes, comprising the largest number of cases (subjects 1, 3, 6–8, 11–15, 18, and 19). There were three accidents (subjects 9, 10, and 21), three suicides not involving drugs (subjects 2, 17 and 22), and three cases of mixed drug toxicities (subjects 4, 5, and 16). The cause of death of subject 20 could not be determined. 3.3. Regional brain distribution of antipsychotic drugs The thioridazine-positive cases revealed the most information out of all antipsychotic drugs studied, since the metabolites mesoridazine and sulforidazine were measured in
addition to parent drug in all five cases. Table 2 shows substantial inter-subject variability in the brain distribution of thioridazine concentrations. In subjects 3 and 4, the highest concentration of thioridazine, mesoridazine, and sulforidazine were detected in the caudate putamen. This was true for only the metabolites in case 2, the highest parent drug concentration being detected in the white matter of the frontal cortex. This region had the lowest concentrations of each compound in subjects 2 and 5. Because the concentrations of all three compounds were not detected in the remaining regions for subject 1, this case was not included in further analyses. Subject 5 had the highest parent and metabolite concentrations in the occipital cortex, the region with the second highest concentration in the remaining subjects. Brain region concentrations of chlorpromazine detected in subjects 1 and 6–10 are shown in Table 2. Because chlorpromazine was not detected in any brain regions for subject 9, this case was not included in further data analysis. These data show that like the thioridazine cases, a great deal of inter-subject variability in brain distribution of chlorpromazine was observed. In all subjects positive for the drug, the lowest concentration was found in the cerebellum, and in four of the five remaining subjects the highest concentration
126
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
Table 2 Summary of brain neuroleptic concentrationsa,b Subject
CPu
FCx (grey)
FCx (white)
OCx
Cerebellum
10 (M:10, S: 10) 240 (M: 10, S: 10) 3100 (M: 430, S: 20) 870 (M: 260, S: 10) 73000 (M: 21000, S: 3800)
10 (M:10, S: 10) 270 (M: 10, S: 10) 1000 (M: 320, S: 10) 10 (M: 260, S: 10) 12000 (M: 5300, S: 770)
20 (M:10, S: 10) 260 (M: 20, S: 110) 2000 (M: 330, S: 10) 10 (M: 260, S: 10) 91000 (M: 26000, S: 4600)
10 (M:10, S: 10) 210 (M: 10, S: 10) 1600 (M: 230, S: 10) 10 (M: 150, S: 10) 51000 (M: 18000, S: 2900)
Chlorpromazine 1 440 6 70 7 6800 8 1700 9 11 10 14000
580 70 1500 10 <10 14000
630 20 1500 20 <10 13000
590 20 1600 10 <10 14000
310 10 1100 10 <10 9900
Trifluoperazine 2 10 11 10 12 10 13 10 14 9100
100 1100 140 60 9100
100 580 90 60 5400
150 1200 190 60 9300
100 570 150 50 8100
490 600
400 470
470 570
410 400
Thioridazine 1 90 (M: 2 40 (M: 3 20000 (M: 4 1200 (M: 5 9500 (M: S: 560)
30, S: 20) 530, S: 250) 2100, S: 610) 430, S: 10) 3500,
Fluphenazine sulfoxide 15 360 16 10 a b
All concentrations in ng/mg; thioridazine data includes data for two metabolites mesoridazine (M) and sulforidazine (S) in parentheses. CPu: caudate-putamen; FCx: frontal cortex; OCx: occipital cortex.
was in the caudate-putamen. However, the relative concentrations in occipital cortex and grey and white matter from frontal cortex differed between the subjects. Table 2 summarizes the regional brain concentrations detected for each trifluoperazine-positive case. Trifluoperazine was found at highest concentration in occipital cortex and was lowest in caudate-putamen, but again there was some degree of inter-subject variability in results obtained from grey and white matter from the frontal cortex and the cerebellum. Since fluphenazine was not available from any Australian supplier, its presence was determined by measuring the concentration of its major metabolite, fluphenazine sulfoxide. This was deemed to be adequate as a study of fluphenazine and metabolite concentrations in rat brain showed fluphenazine sulfoxide concentrations ranged 24–96% of parent drug concentrations [4]. As shown in Table 2, tissue concentrations in subjects 15 and 16 ranged (from highest to lowest) in the following manner: grey matter from the frontal cortex, occipital cortex, white matter from the frontal cortex/ cerebellum, and caudate-putamen. 3.4. Correlation of blood antipsychotic concentrations to those of brain tissue There were seven cases in which target antipsychotic drugs were detected in both blood and brain (subjects 2, 3, 5,
7, 8, 10, and 14). Of these cases, subjects 7, 8, and 10 were chlorpromazine-positive, subjects 2, 3, and 5 were thioridazine-positive, and trifluoperazine was detected in subject 14. Because of the range of drugs detected, there was insufficient data to test for correlation between blood and brain concentrations for individual drugs. Therefore, the seven cases for which both types of specimens had been analyzed were combined and the extent of correlation between the two specimen types calculated. Significant correlations with blood were observed for concentrations in grey matter of the frontal cortex (r2 = 0.944, P < 0.001), occipital cortex (r2 = 0.937, P < 0.001), and cerebellum (r2 = 0.927, p < 0.001). If combined regional concentrations were expressed in terms of those in cerebellum, much lower correlations were observed, the only significant one being with occipital cortex (r2 = 0.783, p < 0.01). Mean absolute drug concentrations (Table 3) in caudateputamen exhibited a positive, significant correlation with lipophilicity (measured as log p) (r2 = 0.949, p < 0.05). Additionally, when corrected for those in cerebellum, mean drug concentrations in occipital cortex were positively correlated with Vd (r2 = 0.994, p < 0.01). However, this correlation is likely skewed by the high Vd of fluphenazine (Table 4) and therefore was not reflective of their actual relationship. No particular correlation between regional drug concentrations and any other parameters were observed.
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
127
Table 3 Mean absolute regional drug concentrations and pharmacokinetic parameters of psychiatric drugsa Drug
CPu
FCx (grey)
FCx (white)
OCx
Cere
pKa
log p
Vd
Fb
Chlorpromazine Fluphenazine Thioridazine Trifluoperazine
3800 5600 190 250 1800 4100 6200 8700
2700 5600 550 78 2100 3900 15000 32000
5200 2500 440 49 6200 2300 2700 5200
2700 690 520 71 2200 4000 19000 40000
1900 3900 410 7.0 1800 3500 11000 23000
9.30 8.10 9.50 8.10
4.55 3.46 5.13 4.26
23 220 18 13
0.98 0.99 0.90 0.95
All concentrations expressed in ng/g. pKa’s and log p’s calculated using CompuDrug Software [16]. CPu: caudate-putamen; FCx: frontal cortex; OCx: occipital cortex; Cere: cerebellum; Fb: fraction protein bound. a Mean S.D.
Table 4 Patterns of regional brain distribution for thioridazine Pattern
Cases exhibiting pattern
Relative sites of concentration (highest to lowest)
A
2
B
3, 4
C
5
FCx (white) > OCx>FCx (grey) > cerebellum > CPu CPu > FCx (grey) > OCx/FCx (white)/cerebellum OCx > FCx(grey) > cerebellum > FCx (white) > CPu
CPu: caudate-putamen; FCx: frontal cortex; OCx: occipital cortex; Cere: cerebellum.
3.5. Blood versus brain concentrations In our study no fluphenazine was detected in the blood of subjects whose histories had suggested this drug had been administered (Table 1), and hence no comparison of brain and blood concentrations could be made. The same is true for the trifluoperazine-positive subjects (n = 4), among whom trifluoperazine was only detected in blood from subject 6. However, as long as these subjects had been taking these drugs therapeutically it is not likely they would have been detected in their blood, as the detection limits for these drugs using conventional assays were about two times the highest expected therapeutic concentration for either of these drugs. It is of interest that the last recorded antipsychotic taken by this subject were haloperidol and flupenthixol decanoate and not trifluoperazine, although only trifluoperazine was detected in both blood and brain tissue from the subject. However, one study of fluphenazine and metabolites in rat brain and other tissues showed that concentrations of fluphenazine in the brain and plasma correlated more closely than those of its metabolites [4]. The correlation of plasma concentrations was strongest for caudate-putamen, cortex, and hippocampus (r2 = 0.83, 0.68, and 0.70, respectively). By contrast, they found no significant correlation between plasma and brain concentrations of fluphenazine sulfoxide for any brain region. Among the four subjects in whom chlorpromazine was detected in brain, it was also detected in the blood in all but subject 12. Our chlorpromazine data indicates the drug must also accumulate more in brain than blood. The chlorproma-
zine concentrations in subjects 13 and 15 were higher in all brain regions analyzed than in blood (750 and 2200 mg/L in blood, respectively), but in subject 15 only the caudate putamen concentration was higher (1700 mg/kg compared to 100 mg/L). The brain region in which measured concentrations correlated most closely with blood levels for chlorpromazine was caudate-putamen (r2 = 0.99), while the lowest correlation region was found with both grey and white matter from frontal cortex (r2 = 0.96 for both brain regions). Consistent with our chlorpromazine and trifluoperazine data, thioridazine also seems to accumulate at higher concentrations in brain than in blood. As with chlorpromazine, thioridazine and metabolites were detected in blood as well as brain in subjects 3, 8, and 11 (Table 1). In subject 11, the concentrations of all three compounds were higher in brain than in blood. In subject 8, only the caudate putamen concentration was higher than that of blood for all three compounds. The caudate putamen concentration in subject 3 was higher than in blood for the two metabolites, but not the parent compound. The caudate-putamen concentrations showed the lowest correlation to blood data for all three compounds (r2 = 0.01, 0.85, and 0.26, for thioridazine, mesoridazine, and sulforidazine, respectively). White matter from the frontal cortex showed the highest correlation with blood data for both thioridazine and mesoridazine (r2 = 0.99 for both compounds), whereas sulforidazine blood data correlated most with those found in the cerebellum (r2 = 0.99).
4. Discussion 4.1. Regional brain distribution The regional brain distribution for each compound using mean absolute concentrations is shown in Fig. 1. Fig. 2 shows this distribution with concentrations normalized for those in cerebellum. As shown in these figures, three distinct patterns of distribution were observed in this set of experiments (see also Table 4). Chlorpromazine, a low-potency phenothiazine with an aliphatic side chain, was found at highest concentrations in caudate-putamen and lowest in cerebellum (group 1). Thioridazine and its metabolites
128
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
Fig. 1. Mean absolute regional brain distribution of selected antipsychotic, grouped according to structural features. Group 1: aliphatic side chain, group 2: piperidine ring in side chain, and group 3: piperazine ring in side chain.
(group 2) were also detected in caudate-putamen at highest concentrations, while the lowest mean concentrations of these compounds were detected in white matter of the frontal cortex. Phenothiazines in group 2 possess a piperidine ring in the side chain and are considered atypical antipsychotics because they are associated with fewer extrapyramidal side effects. It would appear, therefore, that antipsychotics belonging to groups 1 and 2 accumulate in regions of high dopamine receptor concentration. The highest and lowest concentrations of both fluphenazine and trifluoperazine (group 3) were found in occipital cortex and caudate-putamen, respectively. Both of these compounds have a piperazine ring in the side chain and are known to possess anticholinergic activity, thus carrying an increased risk of extrapyramidal side effects in comparison to other antipsychotics. The comparatively lower relative concentrations of both fluphenazine sulfoxide and trifluoperazine may be a result of their comparatively lower affinity for dopamine receptors or their higher affinity for peripheral cholinergic receptors. These data partially agree with the findings of Aravagiri et al. in rat brain, in which the lowest fluphenazine concentrations were found in caudate-putamen [4]. Like their study in rats, the brain fluphenazine sulfoxide concentrations detected in subjects 15 and 16 showed little regional variation in comparison to the other targeted anti-
psychotics. However, as was observed in the presented thioridazine-positive cases, it may be that fluphenazine and its metabolites exhibit different regional brain distribution. In the presented data, it appears that the regional distribution of thioridazine depends on overall drug concentrations. For example, subject 5, who had the highest drug concentrations out of all thioridazine-positive cases, had concentrations in occipital cortex (an area of the brain not thought to be involved in schizophrenia) that were higher than those in other regions. By contrast, subjects 3 and 4, who had moderate overall thioridazine concentrations relative to the other cases, had higher regional concentrations in caudate-putamen (an area of high dopamine receptor concentration). This suggests that antipsychotics first distribute into regions where dopamine receptors and their neurons are concentrated, until active sites are saturated. After reaching this point, the drugs may then begin to distribute more into regions possessing fewer receptors or neurons or bind to less specific sites. It should also be noted that mesoridazine and sulforidazine followed the same pattern of distribution as thioridazine in cases 3–5, but their pattern of distribution differed in subject 9, in which concentrations of all three compounds were significantly lower. This may indicate that although thioridazine and its metabolites appear to distribute in the same way when concentrations are high, the distribution of
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
129
Fig. 2. Mean regional brain distribution of selected antipsychotics relative to cerebellum concentrations and grouped according to structural features. Group 1: aliphatic side chain, group 2: piperidine ring in side chain, and group 3: piperazine ring in side chain.
each compound begins to differentiate upon cessation of drug use, reflecting a possible differential rate of drug elimination between brain regions. It is not possible to say whether the variations observed in regional thioridazine distribution in this study are reflective of differences in time between last dose and death, since this information was not available for any thioridazine-positive subjects. Svendsen et al. have found that no one area of the brain selectively accumulated thioridazine or its metabolites in subjects known to have taken their medication within 8 h or less before death [3]. However, they did find that in one subject who had been taken off medication (800 mg/day) 4 days before death, there was a more specific distribution of thioridazine and its metabolites in different brain regions. Most significant was the difference between cortical and subcortical structures in which the former had high thioridazine and lower mesoridazine concentrations, whereas the latter had high mesoridazine and lower thioridazine concentrations. 4.2. Tissue-to-femoral blood correlations In all seven cases where both blood and brain tissue specimens were positive for antipsychotic drugs, they were detected at concentrations which were 70% higher on average in brain than in blood. This is in accord with findings
from earlier studies [3,4]. Hence, published data on blood therapeutic and toxic concentrations can generally be expected to be lower than those in brain. However, it was interesting to note that several brain regions (including grey matter from frontal cortex, occipital cortex, and cerebellum) exhibited positive significant correlations with blood (r2 > 0.92, p < 0.001). This suggests antipsychotic drug concentrations in these brain regions under some circumstances show some proportion to femoral blood concentrations, and vice-versa. It appears from these results that caudate-putamen may be the best region to sample for determination of fluphenazine or trifluoperazine in terms of likelihood of the drug’s presence at measurable concentrations. In the case of either chlorpromazine or thioridazine, however, more data is needed to understand what appears to be a complex distribution profile in postmortem schizophrenic brains. These data allow two important conclusions to be made. First, low or undetectable concentrations of antipsychotic drug in blood generally correspond to low concentrations in selected regions of the CNS. Second, residual antipsychotic drugs are absent in a large proportion of the brains using detection limits of 10 ng/g (45%). It is notable that in many of the cases there was a difference between those drugs detected in blood and their
130
K.E. Rodda et al. / Forensic Science International 157 (2006) 121–130
detection in brain. Specifically, although case histories indicated a particular antipsychotic drug should have been present, they were often not detected in blood but often were detected in at least one brain region. This is most likely due to two main factors. Firstly, the brain data was obtained using an LC–MS method that specifically targets the antipsychotic drugs in this study with a low detection limit (10 ng/g), whereas the blood toxicology data was collected using older, routine GC screening methods with higher detection limits (50–100 ng/mL). Neither the decanoates of flupenthixol and fluphenazine nor haloperidol is detected routinely with these routine GC techniques. Unfortunately, due to time limits imposed on specimen holding, the blood was unable to be re-analyzed using the LC–MS method. Secondly, in cases where the antipsychotic drug was found in the blood as well as brain, brain concentrations of drug were higher than those found in blood. Therefore, in cases where the antipsychotic was detected in brain at low concentrations but not detected in blood, its blood concentration may have been below the assay LOD. In conclusion, these data indicate different antipsychotic drugs preferentially distribute to different regions of the brain. Importantly, the distinct patterns of distribution for each of the three types of phenothiazines appear to correspond with their relative affinity for dopamine receptors and the regional brain concentration of such receptors. In addition, it appears from the presented data that regional distribution of antipsychotics may be time-dependent, such that different patterns of distribution are observed at differing points in therapy (commencement of treatment, maintenance, and withdrawal of medication). More data are needed to confirm these findings. Future experiments that focus on the regional distribution of antipsychotics not represented in this study would be especially useful. Additionally, more information about the possibility of differential clearance rates of parent drugs and metabolites from brain tissue would be of value.
Acknowledgments This work was supported in part by the NH&MRC Schizophrenia Unit, the State Government of Victoria, the United States Air Force, The Brockhoff Foundation and the Rebecca L. Cooper Medical Research Foundation. Brian Dean is the recipient of a NARSAD Young Investigator award. The authors wish to thank Professor Nicholas Keks and Ms. Christine Hill for carrying out case-history reviews and diagnostic confirmation, and the Victorian Institute of Forensic Medicine for access to tissue and to staff for assistance with tissue collection. Finally, we thank Agilent Australia for loan of the LC–MS.
References [1] B.M. Cohen, J.F. Lipinski, C. Waternaux, A fixed dose study of the plasma concentration and clinical effects of thioridazine and its major metabolites, Psychopharmacology (Berl) 97 (4) (1989) 481–488. [2] T. Tsuneizumi, S.M. Babb, B.M. Cohen, Drug distribution between blood and brain as a determinant of antipsychotic drug effects, Biol. Psychiatry 32 (9) (1992) 817–824. [3] C.N. Svendsen, C.C. Hrbek, M. Casendino, R.D. Nichols, E.D. Bird, Concentration and distribution of thioridazine and metabolites in schizophrenic post-mortem brain tissue, Psychiatry Res. 23 (1) (1988) 1–10. [4] M. Aravagiri, S.R. Marder, A. Yuwiler, K.K. Midha, N.S. Kula, R.J. Baldessarini, Distribution of fluphenazine and its metabolites in brain regions and other tissues of the rat, Neuropsychopharmacology 13 (3) (1995) 235–247. [5] T.C. Merrick, A. Felo, A.J. Jenkins, Tissue distribution of olanzapine in a postmortem case, Am. J. Forensic Med. Pathol. 22 (2001) 270–274. [6] L.E. van Beijsterveldt, R.J. Geerts, J.E. Leysen, A.A. Megens, H.M. Van den Eynde, W.E. Meuldermans, et al. Regional brain distribution of risperidone and its active metabolite 9-hydroxyrisperidone in the rat, Psychopharmacology (Berl) 114 (1) (1994) 53–62. [7] T.C. Merrick, J.A. Felo, A.J. Jenkins, Tissue distribution of olanzapine in a postmortem case, Am. J. Forensic Med. Pathol. 22 (3) (2001) 270–274. [8] F.P. Bymaster, D.O. Calligaro, J.F. Falcone, R.D. Marsh, N.A. Moore, N.C. Tye, et al. Radioreceptor binding profile of the atypical antipsychotic olanzapine, Neuropsychopharmacology 14 (2) (1996) 87–96. [9] R.R. Conley, C.A. Tamminga, D.L. Kelly, C.M. Richardson, Treatment-resistant schizophrenic patients respond to clozapine after olanzapine non-response, Biol. Psychiatry 46 (1) (1999) 73–77. [10] L.J. Garey (Ed.), Brodman’s Localisation in the Cerebral Cortex, Smith-Gordon, London, 1994. [11] N.A. Keks, K. Hill, K. Opeskin, et al. Psychiatric diagnosis after death: the problems of accurate diagnosis, in: B. Dean, T.M. Hyde, J. Kleinman (Eds.), The Use of CNS Autopsy Tissue in Psychiatric Death, Gordon & Breach Science Publishers, Sydney, 1999, pp. 19–37. [12] P. Foster, Neuroleptic equivalence, Pharmacol. J. 243 (1989) 431–432. [13] K.E. Goeringer, M. McIntyre, O.H. Drummer, LC–MS analysis of serotonergic drugs, J. Anal. Toxicol. 27 (1) (2003) 30– 35. [14] I.M. McIntyre, M.L. Syrjanen, K. Crump, S. Horomidis, A.W. Peace, O.H. Drummer, Simultaneous HPLC gradient analysis of 15 benzodiazepines and selected metabolites in postmortem blood, J. Anal. Toxicol. 17 (4) (1993) 202–207. [15] R.C. Baselt, Disposition of Toxic Drugs and Chemicals in Man, 5th ed., Year Book Medical Publishers, 2000. [16] C. PALLAS Expert System for Today’s Chemists, CompDrug Internationsl Inc., San Francisco, 1999.