Psychological Medicine, 1976, 6, 407-415
Clinical significance of plasma chlorpromazine levels I. Plasma levels of the drug, some of its metabolites and prolactin during acute treatment D. H. WILES, T. KOLAKOWSKA, A. S. McNEILLY, B. M. MANDELBROTE AND M. G. GELDER 1 From the University of Oxford, Department of Psychiatry Research Unit, Littlemore Hospital, Oxford, and Department of Chemical Pathology, St Bartholomew''s Hospital, London
Seventeen acute psychotic patients were studied in the course of chlorpromazine (CPZ) treatment. Blood samples were taken weekly both before and two hours after the morning CPZ dose. Plasma levels of CPZ, CPZ sulphoxide (CPZ SO) monodesmethylated CPZ (NORXCPZ) and 7-hydroxy CPZ (7OH CPZ) were estimated by gas chromatography. Plasma prolactin, luteinizing hormone, testosterone and oestrogens were measured by radioimmunoassay. Six of the seven patients who showed no clinical improvement had plasma CPZ levels equal to or higher than those of patients who improved. 'Non-responders' had a greater proportion of CPZ SO in pre-dosage samples. The occurrence of parkinsonian side effects was associated with a mean plasma CPZ of > 50 ng/ml and a mean plasma prolactin of > 30 ng/ml two hours after dosage. The elevation of prolactin preceded the onset of parkinsonian symptoms by 1-2 weeks. There was a significant positive correlation between mean plasma prolactin and mean plasma CPZ levels. The prolactin response may prove a useful index of the central antidopaminergic effect of neuroleptic drugs.
SYNOPSIS
INTRODUCTION
Patients who take the same dose of chlorpromazine (CPZ) do not all respond in the same way. For many years attempts have been made to relate therapeutic response and the appearance of adverse reactions to such factors as 'target symptoms', the clinical course of the illness and such general characteristics of the patient as age, sex, general health and family history. As a result, the indications for selection of patients for treatment have become rather more precise, but the problem of the relationship between the dose of phenothiazines and their therapeutic and unwanted effects remains unsolved, even for chlorpromazine, the best-known representative of this group of drugs. This lack of information is reflected in the large differences in the dosage schedules used by psychiatrists in Britain and America. 1 Address for correspondence: Professor M. G. Gelder, University of Oxford Department of Psychiatry, Warneford Hospital, Oxford 0X3 7JX.
27
407
Variations in plasma level of the phenothiazine might explain part of this variability in the response to the drug. Thus, Curry et al. (1970) have shown that patients receiving the same dose of CPZ differ in their plasma CPZ level, although they observed consistency between an alteration of the plasma level and a change of clinical response in the individual subject. The relationship between dose administered and plasma level is further complicated by the observation that acute patients treated with a constant dose show considerable variations in CPZ levels from week to week (Sakalis et al. 1972). The relationship of plasma drug level to clinical effect is also uncertain. Curry and his collaborators have shown a wide range of therapeutic plasma levels of CPZ during both acute and long-term treatment (Curry et al. 1970) and other authors confirm this (Mackay et al. 1974; Rivera-Calimlin et al. 1974). With chronic patients, poor therapeutic effect was found in association with high, low and/or unstable
408
D, H. Wiles and others
levels (Curry & Marshall, 1968). In a detailed study of patients undergoing treatment for acute illness, improvement during the first fortnight was related to plasma CPZ and so were some peripheral side effects, but no overall relationship was found between CPZ level and therapeutic effect (Sakalis et al. 1972). The recent reports of Sakalis et al. (1973) and of Mackay et al. (1974) present evidence that plasma levels of CPZ SO and 7OH CPZ may be related to the therapeutic effect. However, discrepant data on levels of CPZ metabolites and their relation to the levels of the parent drugs have been reported by different groups of workers (Curry & Marshall, 1968; March et al. 1972). SUBJECTS
Seventeen subjects were studied in the course of their CPZ treatment. They were recently admitted acute psychotic patients for whom phenothiazine treatment was indicated and who were free from physical illness. Five had never received psychiatric treatment, another six had received no neuroleptic drugs for at least three previous months, while four had been receiving phenothiazines and two patients antidepressants for various periods before the investigation. It was not considered ethical to withhold treatment to detect any spontaneous remitters. CPZ in syrup was prescribed. A dose between 150 and 600 mg/day was decided on clinical judgement by the patient's psychiatrist of the patient's response in the early days. This dose was kept constant throughout the period of study for all but three patients where there were clinical reasons for changing. Benzhexol was added only if parkinsonian side effects appeared and occasional nitrazepam was the only other drug prescribed. Three patients were discharged as improved within the first 17 days; the remaining fourteen received CPZ from 3 to 8 weeks. Clinical data about the seventeen subjects are shown in Table 1. PROCEDURE
Blood samples for the estimation of plasma levels of CPZ and hormones were collected weekly from all patients (and in six twice a week during the first fortnight). Samples were taken both immediately before and two hours after the morning dose of CPZ (these are referred to
subsequently as the zero and two-hour samples respectively). Pulse and blood pressure were measured in the standing and reclining positions before taking each sample. Clinical assessment was repeated on the days of blood collection. METHODS (a) CHLORPROMAZINE AND METABOLITES
Plasma drug level was determined by Curry's gas-chromatographic method (Curry, 1968) in a modification described by Mackay et al. (1974). This modified method allows measurement of CPZ and three of its metabolites (CPZ SO, 7OH CPZ and N O R ^ P Z ) by using diethyl ether instead of heptane for extraction, and corrects for the variability of extraction by introducing 2,4-dichlorpromazine as internal standard. Precision and accuracy Ten samples containing a 50 ng/ml mixture in plasma when assayed together gave mean levels ±SD of 96 + 4, 99±9, 104± 14 and 105± 11 % of the actual concentration for CPZ, NORjCPZ, 7OH CPZ and CPZ SO. Similar results were obtained over the range 10-100 ng/ml for CPZ, NORjCPZ and CPZ SO; and between 25 and 100 ng/ml for 7OH CPZ. This data was confirmed in a multi-laboratory collaborative qualitycontrol study organized by Dr W. J. Turner of the Central Islip Psychiatric Research Center, New York (Turner et al. 1975). Detector sensitivity Using a Pye Unicam Series 104 E.C. Detector, this was 0-5—1 ng in 5 ft\ injection for CPZ, NORjCPZ and CPZ SO: 1-2 ng for 7OH CPZ. Averages from individual weekly measurements throughout treatment were used to characterize plasma CPZ of each subject both immediately before (time 'zero') and two hours after the morning dose. (In the few patients whose dose was changed, the level on the more prolonged dose was taken.) (b)
HORMONES
Plasma prolactin, luteinizing hormone (LH), testosterone and oestrogens were estimated by
409
Clinical significance of plasma chlorpromazine levels TABLE 1 CLINICAL DETAILS OF SEVENTEEN PATIENTS
Patient
Age Sex
L.W. B.G. J.C. G.S. J.M. D.M. M.M. H.S. A.C. L.Ph. A.A. F.M.
47 26 45 23 27 60 32 26 23 23 33 37
F F M M M F M M M M M F
Paranoid schizophrenia Paranoid-depressive syndrome (puerperal) Paranoid-depressive syndrome Acute paranoid psychosis Paranoid schizophrenia Paranoid reaction Paranoid schizophrenia Paranoid schizophrenia Paranoid schizophrenia Schizophrenia Paranoid schizophrenia Schizophrenia
300/450 400 150-300 300 600-800 400 400 400 500 500 500 300
5-4-8-2 70 2-3-4-7 4-3 10-0-12-7 6-9 4-6 5-2 80 61 8-3 41
51 32 36 7 42 21 56 42 15 27 23 28
M.T.M.
21
F
Paranoid schizophrenia
400
5-5
43
J.G.
21
F
Paranoid schizophrenia
400
7-8
40
J.A.
24
F
450
71
35
V.G. R.A.
25 38
M M
Acute paranoid psychosis (intoxication with antidepressants) Acute paranoid psychosis Catatonic schizophrenia
150 300
2-5 4-3
8 36
Diagnosis
radioimmunoassays. Plasma prolactin was measured by the specific double antibody radioimmunoassay described by McNeilly & Hagen (1974). LH was measured by the method of Midgley (1966) using the second international reference preparation of human menopausal gonadotrophin(2nd IRP-HMG) and the Medical Research Council Pituitary LH 68/40 as standards. Testosterone was estimated as described by Furuyama et al. (1970). Immunoreactive oestrogens were measured by the technique of Hotchkiss et al. (1971). The antiserum was relatively specific for oestradiol 17/? and crossreaction to oestradiol 17a, oestrone and oestriol at 40, 35 and 8 % respectively.
mg/day
No No No No No Stopped over 1 year ago Stopped over 3 months ago Stopped over 3 months ago Stopped over 3 months ago Stopped over 6 months ago Stopped over 3 months ago Tricyclic antidepressants in low doses Trifluoperazine + orphenadzine (4 weeks) CPZ + trifluoperazine (irregular, 2 weeks) Amitriptyline + Parstelin (4 weeks) Thioridazine (6 days) Fluphenazine decanoate + CPZ
scale.2 In addition, changes in sleep, habits, activity and insight were rated on a 4-point scale. A check-list for subjective autonomic symptoms (ratings 0-3) and a rating scale for parkinsonian symptoms (ratings 0-3) were used to assess CPZ side-effects. All the clinical ratings were carried out during the initial interview and repeated on the days of blood sampling. At the end of CPZ treatment a global assessment of change in mental state was made on a 4-point scale (0 = worse/unchanged; 1 = some symptomatic improvement only; 2 = satisfactory but partial improvement, residual symptoms; 3 = good improvement, patient asymptomatic).
RESULTS
(c) CLINICAL RATINGS
A full psychiatric interview was carried out at the first contact with the patient, after obtaining his consent to participate in the study, and the Brief Psychiatric Rating Scale (BPRS, Overall & Gorham, 1962) was used to assess the mental state. The prognostic index (Stephens et al. 1966) was completed at the end of the first interview. Two items were added to the BPRS to assess elation and psychomotor excitation. A detailed glossary has been developed to rate the severity of each symptom on a 7-point
Psychotropic drugs prior to admission
mg/kg/day Days
(a) PLASMA LEVELS OF CPZ AND METABOLITES
A wide range of plasma levels of CPZ and its metabolites was found in patients taking therapeutic doses (Table 2). This variation was not reduced by expressing CPZ levels per daily dose of CPZ in mg or mg/kg body weight. The overall relation between CPZ dose and plasma level is not simple. The data suggest that plasma CPZ levels tend to increase with dosage 2
The glossary is available on request. 27-2
410
D. H. Wiles and others TABLE 2 PLASMA CPZ, CPZ SO, NOR,CPZ AND 7OH CPZ LEVELS IN SEVENTEEN SUBJECTS TREATED WITH A CONSTANT DAILY DOSE AT 150 TO 600 mg CPZ
CPZ SO
CPZ
7OH CPZ
NOR ,CPZ
Oh
2h
Oh
2h
Oh
2h
Oh
2h
All samples
Mean SD Range N
24 24 0-125 82
56 42 0-196 78
11 21 0-126 76
23 30 0-97 72
1 1 0-6 76
2 4 0-26 72
3 7 0-26 76
7 10 0-36 72
Averages for individuals
Mean SD Range N
21 22 0-92 17
50 33 0-110 17
11 20 0-85 16
28 19 3-77 16
1 1 0-4 16
2 4 0-10 16
3 6 0-23 16
6 8 0-27 16
80
r
60
N a. U 5 40 a c
20
200 400 Dose CPZ (mg/day)
600
FIG. 1. Mean plasma CPZ level in ng per ml measured immediately before the morning dose of the drug, plotted against the daily dose.
up to a dose of 400 mg/day; however, patients who received larger doses than this had a lower average level of CPZ than those on 400 mg/day (Fig. 1). If plasma CPZ at two hours is expressed per dose of 1 mg/kg body weight and plotted against the full dose in mg/kg body weight, negative correlation is found between dose and plasma level (r = -0-57; P < 0-05). Indeed the six patients who received CPZ in doses above 7 mg/kg/day had among the lowest ratios of plasma CPZ to daily dose. The clinical features of these six were those of a florid paranoid syndrome with some elation and/or psychomotor excitation, while only two of the remaining 11 patients had a similar clinical picture. The plasma CPZ levels of individual patients
also varied considerably from week to week even though dosage was unchanged: thus up to a sevenfold difference was recorded at zero time. Among 14 patients for whom we have repeated samples over at least three weeks, 9 had their highest CPZ levels in the first fortnight, 3 had this later and in two levels did not change. In the 3 patients whose dosage was altered, plasma CPZ changed in the corresponding direction and fluctuated around a new mean. Benzhexol was introduced during the fourth or fifth week in four subjects. This was not followed by any consistent change in plasma CPZ levels. With only one exception, CPZ SO was detected both before and after the CPZ dose in all subjects when CPZ was present. Its level was on average about half that of CPZ but the mean ratio of CPZ SO to CPZ from individual patients ranged from 0-14 to 0-89 at zero time and 0-33 to 0-93 at two hours. There was, however, in the group of patients a strong correlation between plasma CPZ and CPZ SO levels both at zero time (r = 0-92; P < 0001) and at two hours (r = 0-87; P < 0001). (b) PLASMA CPZ RESPONSE
LEVEL
AND
CLINICAL
Of the seventeen patients, ten improved and had been discharged at the end of the investigation. Their global rating was 2 or 3 and their BPRS score decreased by 58-94% of initial value. For seven patients the therapeutic effect was poor (global rating 0 or 1; change in BPRS score from - 4 4 to +30 %) and for six patients other methods of treatment had to be introduced.
411
Clinical significance of plasma chlorpromazine levels TABLE 3
AVERAGE PLASMA LEVELS OF CPZ AND CPZ SO AND THEIR RATIO IN 'RESPONDERS' AND 'NON-RESPONDERS'
CPZ (ng/ml)
'Responders'
CPZ SO (ng/ml)
SO/CPZ
Daily dose (mg)
Oh
2h
Oh
2h
Oh
2h
370 106 10
19 13 10
45 29 10
6 5 9
25 13 9
0-29
0-54 018 9
34 28 7
P < 002 0-53 0-21 5
Mean SD N
on 9
P if significant 'Non-responders'
453 155 7
Mean SD N
25 29 7
15 27 7
62 42 7
0-52 014 6
TABLE 4 AVERAGE PLASMA LEVELS OF CPZ, CPZ SO AND PROLACTIN IN EIGHT SUBJECTS SHOWING THE EXTRAPYRAMIDAL SYNDROME (EPS) AND SEVEN SUBJECTS WITH NO EPS. ALL HAD RECEIVED CPZ TREATMENT FOR AT LEAST FIFTEEN DAYS
Plasma CPZ (ng/ml)
EPS
Plasma prolactin (ng/ml)
Daily dose (mg)
Oh
2h
Oh
2h
Oh
2h
406 126 8
30 26 8
75 29 8
16 24 8
40 21 8
37 16 8
41 14 8
—
—
P < 001
—
P < 005
—
—
454 112 7
15 12 7
31 23 7
5 4 6
17 12 6
25 13 7
28 21 7
Mean SD N
P if significant Mean SD N
No EPS
Plasma CPZ SO (ng/ml)
100 r
80 -
0
60 ft.
5. 40
o o
o
O
* 20 "
o
o i
10
i
o
;
i
i
20 30 40 50 Mean plasma prolactin (ng/ml)
i
i
60
70
FIG. 2. Mean plasma prolactin and CPZ two hours after dosage in eight subjects showing the extrapyramidal syndrome (EPS) (•) and seven subjects showing no EPS (O).
412
D. H. Wiles and others
Six of these unimproved patients had plasma CPZ levels equal to or higher than those of the improved group (Table 3); only one had very low or undetectable plasma drug levels on nearly all occasions. A significant (P < 0-02) feature of these non-responders was a higher ratio of CPZ SO to CPZ at zero time (0-53) than that of the responders (0-29). Although the groups did not differ in the initial BPRS score or prognostic index, they showed some clinical differences. All 'responders' had an episode of acute onset and either no previous admission (5) or relatively good remissions in the past (5). Among 'nonresponders', three had a long history of schizophrenic illness with poor remissions, three presented a paranoid episode with marked elation and one suffered from puerperal psychosis with paranoid syndrome coloured by depression and anxiety. More 'non-responders' than 'responders' developed parkinsonian side effects (5 of 7 and 3 of 10 respectively).
(c) PLASMA DRUG LEVELS AND EXTRAPYRAMIDAL SIDE EFFECTS (EPS)
Among 15 patients treated with CPZ for fifteen days or more, 8 developed a parkinsonian syndrome (all within the first three weeks of CPZ administration) and 7 did not. The difference in average CPZ dosage between the groups was insignificant (406 and 454 mg/day respectively) but they differed considerably in plasma CPZ level (Table 4). Seven out of eight patients with EPS had mean levels at two hours > 50 ng/ ml, while only one out of seven patients with no EPS had as high a plasma CPZ (Fig. 2). Mean levels of both CPZ and CPZ SO were higher in the EPS group, significantly so at two hours (P < 001 and P < 005). There was no difference in levels of 7OH and NORjCPZ. Changes in pulse and blood pressure were rather small and unrelated to plasma CPZ levels. There was no overall correlation between subjective autonomic side effects and plasma CPZ but in individual patients the maximal severity of these symptoms were either soon before or at the same time as the highest CPZ levels. Only 1 patient out of the 6 whose mean
plasma prolactin at two hours remained below 30ng/ml developed EPS, while 7 out of 9 patients with a higher two hours prolactin developed parkinsonism.
{d) PROLACTIN LEVELS
Plasma prolactin was measured throughout treatment in 17 patients (7 women and 10 men). In all 7 women and in 5 of the 10 men prolactin levels were elevated (i.e. above 20ng/ml) from the end of the first week of CPZ administration and throughout the treatment period. Prolactin fluctuated erratically with no consistent further increase after the first week. (The sole exception was a women who developed mild galactorrhoea in the third week of treatment (with 300 mg/day) and whose prolactin levels rose progressively until the third week.) Only in two male patients was there a rise in plasma prolactin after every morning dose of CPZ. In the remaining subjects the changes of prolactin levels following the morning dose of CPZ were inconsistent. The mean plasma prolactin level of individual patients was unrelated to the daily dose of CPZ but there was a significant positive correlation between mean plasma prolactin and mean CPZ concentration. Fig. 3 shows this at zero time (r = 0-77; P < 001); the correlation coefficient between corresponding levels at two hours was 0-52 (P < 005). Means of weekly estimations were used here because prolactin levels fluctuate from hour to hour and CPZ levels also vary. Thus the average or sustained level of prolactin is more likely to be biologically related to the sustained level of CPZ. However, correlations were also calculated using every pair of values and these were also significant: for' zero' samples -r = 0-51; P < 0001, for all twohour samples -r = 0-34; P < 0005. There were insufficient readings from some subjects to allow calculation of r for each subject separately or for variance analysis. Mean plasma prolactin also correlated with mean CPZ SO levels, significantly in 'zero' samples (r = 0-58; P < 002) and both in zero and two-hour samples when all paired observations were included (P < 0001). Plasma prolactin levels tended to be higher
413
Clinical significance of plasma chlorpromazine levels 80 r
I60 c
I
I 40 CO
V)
"B. c
± 20
20
40
60
80
100
Mean plasma CPZ (ng/ml) FIG. 3. Correlation between plasma levels of CPZ and prolactin in seventeen subjects. Mean levels for individuals from samples taken before the morning dose of the drug, r = 0-67; P < 0 0 1 .
in subjects who developed extrapyramidal symptoms (EPS). Thus 7 out of 8 such patients had a mean two-hours prolactin > 30 ng/ml, while in only 2 of 7 patients without these symptoms was prolactin so elevated (Fig. 2). Prolactin elevation preceded the appearance of EPS by 1-2 weeks. (e) SEX HORMONES
Measurements of LH, testosterone, oestrogens and progesterone throughout treatment gave values within the normal ranges. No correlation was found between sex hormone levels and plasma CPZ. DISCUSSION
We used the method described by Mackay et al. (1974) in preference to Curry's method because we have found it to be more reliable and because it measures the derivatives we have described. Although an improvement on Curry's method, it is not without potential disadvantages. Thus ether would extract chlorpromazine Noxide if this were present in significant amounts in plasma; and this substance would in turn decompose to CPZ and desmethyl-CPZ on the column, thus falsifying the values obtained for these substances. However, although CPZ-N-
oxide has been found in urine, it has not been demonstrated convincingly in significant amounts in plasma. The use of 2,4-dichlorpromazine as a standard improves the reliability, even though this substance will not, of course, behave exactly in the same way as CPZ in the extraction procedure. The range of plasma CPZ levels found in these patients is comparable to those described by Sakalis et al. (1972) in acute patients, and by Curry et al. (1970) and Mackay et al. (1974) in chronic ones. The relationship between CPZ dose and plasma level was complicated: at doses up to 400 mg/day or 7 mg/kg/day, plasma levels were related to the dose but with higher doses plasma levels of CPZ were actually lower. Mackay et al. (1974) reported a similar finding in chronic patients, namely lower levels in subjects on 600 mg/day than in those receiving 400 mg/day; and Martensson & Roos (1973) describe a comparable result with thioridazine. In our group a low ratio of plasma level of CPZ to dose was often found among patients who were on doses exceeding 400 mg/day and who also presented a florid paranoid syndrome. CPZ absorption and metabolism may depend on the dose administered, but it might also be affected by physiological changes related to a particular psychotic syndrome. Like other
414
D. H. Wiles and others
investigators, we found considerable variation of CPZ levels in the same patient on different occasions, but we did not observe the regular decline after the first week reported by Sakalis et al. (1972). We also failed to repeat the observation of Rivera-Calamlin et al. (1973) that addition of benzhexol is accompanied consistently by lower CPZ levels. Although we used the method of assaying metabolites that Mackay et al. (1974) described, our findings were different. Thus we did not detect 7OH CPZ regularly, nor could we relate the ratio of either 7OH CPZ to CPZ or of 7OH CPZ to CPZ SO, to outcome as these workers, and also Sakalis et al. (1973), had done. This may be because our assay was less sensitive to 7OH CPZ, or may be because these authors studied chronic patients who had received CPZ for 9-6 years on average, while ours had received the drug for a much shorter time. The ratio of sulphoxide to CPZ was consistently higher two hours after the morning dose than twelve hours after the evening dose (i.e. immediately before the morning dose). As there was no significant overall correlation between this ratio and CPZ level, the difference may have been due to the shorter plasma halflife of the sulphoxide (Davidson et al. 1957). We were also unable to confirm the finding of Mackay et al. (1974) that response is related to the ratio of 7OH CPZ to CPZ; we did, however, find a higher ratio of sulphoxide to CPZ in unresponsive patients. In any case there are clinical differences between responders and nonresponders that must obscure relationships between responsiveness and drug levels. We have reported in a preliminary account that prolactin level was related both to plasma CPZ level and to the appearance of a parkinsonian syndrome (Kolakowska et al. 1975). This rise in prolactin preceded the extrapyramidal syndrome (EPS) by 1-2 weeks. The parkinsonian syndrome was also related to plasma CPZ levels. The effect on prolactin release and on EPS are both thought to depend on dopaminergic mechanisms, though of different location - one nigrostriatal and the other hypothalamic. The findings suggest that the two dopaminergic systems have a similar threshold of response to CPZ, and
that prolactin level can be used as an index of the antidopaminergic central effect of CPZ. One of the theories put forward to explain the antipsychotic effects of CPZ is that it depends on antidopamine properties of the drug. However, these were not related to plasma CPZ level in our patients. It has been suggested that the dopaminergic system in the limbic area is of particular importance for pathophysiology of schizophrenia (Stevens, 1973). It has also been shown in rats that effects of neuroleptics on dopamine metabolism differ among these three brain areas (Anden, 1972; Bowers & Rozitis, 1974). These regional differences might be of importance for the dissociation between antipsychotic effects of the drug on one hand and, parkinsonism and prolactineemia on the other. It is also possible that subjects' response to changes of dopamine metabolism varies more for psychological than for extrapyramidal and hormonal effects. The changes in the mental state which occur during CPZ administration are not only more complex and more difficult to measure but they depend on factors other than the pharmacological action of the drug. Finally, the relationship between CPZ plasma level and therapeutic effect is obscured by the nature of the psychosis which tends to remit spontaneously in some patients, while in others it takes a progressive course, irrespective of any treatment. Perhaps it is only within a more uniform group of potential 'responders' that plasma levels of the drug could be related to therapeutic effects. Such a group would presumably consist of patients who are likely neither to have an early spontaneous remission nor to have a poor prognosis irrespective of treatment. However, neither clinical judgement nor prognostic scales permit an early and reliable separation of such groups of patients. We are grateful to Miss A. Garfield for technical assistance. Drug standards were kindly supplied by Dr A. A. Manian, United States National Institute of Mental Health. We are grateful to Drs S. H. Curry and A. F. Healey for advice on methodology, to Mrs E. Green and Mrs E. Kuht for nursing assistance and to the consultants of Littlemore Hospital for permission to study patients under their care.
Clinical significance of plasma chlorpromazine levels REFERENCES Anden, N. E. (1972). Dopamine turnover in the corpus striatum and the limbic system after treatment with neuroleptics and anti-acetylocholine drugs. Journal of Pharmacy and Pharmacology 29, 340-348. Beumont, P. S. V., Corker, C. S., Friesen, H. G., Kolakowska, T , Mandelbrote, B. M., Marshall, J., Murray, M. A. F. & Wiles, D. H. (1974). The effects of phenothiazines on endocrine function: II. British Journal of Psychiatry 124, 420-430. Bowers, M. B., Jr & Rozitis, A. (1974). The regional differences in homovanilic acid concentrations after acute and chronic administration of antipsychotic drugs. Journal of Pharmacy and Pharmacology 26, 743-745. Cooper, S. F., Albert, J. M., Hillel, i. & Caille, G. (1973). Plasma level studies of chlorpromazine following the administration of chlorpromazine hydrochloride and chlorpromazine embonate in chronic schizophrenics. Current Therapeutic Research 15, 73-77. Curry, S. H. (1968). Determination of nanogram quantities of chlorpromazine and some of its metabolites using gas liquid chromatography with an electron capture detector. Analytical Chemistry 40, 1251-1255. Curry, S. H. & Marshall, J. H. L. (1968). Plasma levels of chlorpromazine and some of its relatively non-polar metabolites. Life Sciences 7, 9-17. Curry, S. H., Marshall, J. H. L., Davis, J. M. & Janovsky, D. S. (1970). Chlorpromazine plasma levels and effects. Archives of General Psychiatry 22, 289-296. Davidson, J. D., Terry, L. L. & Sjoerdsma, A. S. (1957). Action and metabolism of chlorpromazine sulfoxide in man. Journal of Pharmacy and Experimental Therapeutics 121,8-12. Furuyama, S., Mayes, D. M. & Nugent, C. A. (1970). A radioimmunoassay for plasma testosterone. Steroids 16, 415-428. Hotchkiss, J., Atkinson, L. E. & Knobil, E. (1971). Time course of serum oestrogen and luteinizing hormone concentrations during the menstrual cycle of the rhesus monkey. Endocrinology 89, 177-183. Kolakowska, T., Wiles, D. H., McNeilly, A. S. & Gelder, M. (1975). Correlation between plasma levels of prolactin and chlorpromazine. Psychological Medicine 5, 214-216. Mackay, A. V. P., Healey, A. F. & Baker, J. (1974). The relationship of plasma chlorpromazine to its 7-hydroxy and sulphoxide metabolites in a large population of chronic
415
schizophrenics. British Journal of Clinical Pharmacology 1, McNeilly, A. S. & Hagen, C. (1974). Prolactin, TSH, LH and FSH responses to a combined LHRH/TRH test at different stages of the menstrual cycle. Clinical Endocrinology 3, 427-435. March, J. E., Donato, D.,Turano, P. & Turner, W. J. (1972). Interpatient variation and significance of plasma levels of chlorpromazine in psychotic patients. Journal of Medicine (Basel) 3, 146-162. Martensson, E. & Roos, B. E. (1973). Serum levels of thioridazine in psychiatric patients and healthy volunteers. European Journal of Clinical Pharmacology 6, 181-186. Midgley, A. R. (1966). Radioimmunoassay: a method for human chorionic gonadotrophin and human luteinizing hormone. Endocrinology 79, 10-18. Overall, G. E. & Gorham, D. R. (1962). The brief psychiatric rating scale. Psychological Reports 10, 799. Rivera-Calimlin, L., Castaneda, L. & Lasagna, L. (1973). Effects of mode of management on plasma chlorpromazine in psychiatric patients. Clinical Pharmacology and Therapeutics 14, 979-985. Sakalis, G., Curry, S. H., Mould, G. P. & Lader, M. H. (1972). Physiologic and clinical effects of chlorpromazine and their relationship to plasma level. Clinical Pharmacology and Therapeutics 13, 931-946. Sakalis, G., Chan, T. L., Gershon, S. & Park, S. (1973). The possible role of metabolites in therapeutic response to chlorpromazine treatment. Psychopharmacologia (Berlin) 32, 279-284. Stephens, J. H., Astrup, Ch. & Mangrum, S. C. (1966). Prognostic factors in recovered and deteriorated schizophrenics. American Journal of Psychiatry 120, 1116-1121. Stevens, J. R. (1973). An anatomy of schizophrenia. Archives of General Psychiatry 29, 177-189. Turano, P., March, J. E., Turner, W. J. & Merlis, S. (1972). Qualitative and quantitative report on chlorpromazine and metabolites in plasma, erythocytes and erythrocyte washings from chronically medicated schizophrenic patients. Journal of Medicine (Basel) 3, 109-120. Turner, W. J., Turano, P., Badzinski, S., Breyer, U., Curry, S., Kaul, P. & Wiles, D. (1975). An attempt to establish quality control in determination of plasma chlorpromazine by a multi-laboratory collaboration. In Pharmacokinetics, Blood Levels and Clinical Response of Psychoactive Drugs (ed. L. Gottschalk & S. Merlis). Spectrum: New York.
Clinical significance of plasma chlorpromazine levels I. Plasma levels of the drug, some of its metabolites and prolactin during acute treatment D. H. WILES, T. KOLAKOWSKA, A. S. McNEILLY, B. M. MANDELBROTE AND M. G. GELDER 1 From the University of Oxford, Department of Psychiatry Research Unit, Littlemore Hospital, Oxford, and Department of Chemical Pathology, St Bartholomew''s Hospital, London
Seventeen acute psychotic patients were studied in the course of chlorpromazine (CPZ) treatment. Blood samples were taken weekly both before and two hours after the morning CPZ dose. Plasma levels of CPZ, CPZ sulphoxide (CPZ SO) monodesmethylated CPZ (NORXCPZ) and 7-hydroxy CPZ (7OH CPZ) were estimated by gas chromatography. Plasma prolactin, luteinizing hormone, testosterone and oestrogens were measured by radioimmunoassay. Six of the seven patients who showed no clinical improvement had plasma CPZ levels equal to or higher than those of patients who improved. 'Non-responders' had a greater proportion of CPZ SO in pre-dosage samples. The occurrence of parkinsonian side effects was associated with a mean plasma CPZ of > 50 ng/ml and a mean plasma prolactin of > 30 ng/ml two hours after dosage. The elevation of prolactin preceded the onset of parkinsonian symptoms by 1-2 weeks. There was a significant positive correlation between mean plasma prolactin and mean plasma CPZ levels. The prolactin response may prove a useful index of the central antidopaminergic effect of neuroleptic drugs.
SYNOPSIS
INTRODUCTION
Patients who take the same dose of chlorpromazine (CPZ) do not all respond in the same way. For many years attempts have been made to relate therapeutic response and the appearance of adverse reactions to such factors as 'target symptoms', the clinical course of the illness and such general characteristics of the patient as age, sex, general health and family history. As a result, the indications for selection of patients for treatment have become rather more precise, but the problem of the relationship between the dose of phenothiazines and their therapeutic and unwanted effects remains unsolved, even for chlorpromazine, the best-known representative of this group of drugs. This lack of information is reflected in the large differences in the dosage schedules used by psychiatrists in Britain and America. 1 Address for correspondence: Professor M. G. Gelder, University of Oxford Department of Psychiatry, Warneford Hospital, Oxford 0X3 7JX.
27
407
Variations in plasma level of the phenothiazine might explain part of this variability in the response to the drug. Thus, Curry et al. (1970) have shown that patients receiving the same dose of CPZ differ in their plasma CPZ level, although they observed consistency between an alteration of the plasma level and a change of clinical response in the individual subject. The relationship between dose administered and plasma level is further complicated by the observation that acute patients treated with a constant dose show considerable variations in CPZ levels from week to week (Sakalis et al. 1972). The relationship of plasma drug level to clinical effect is also uncertain. Curry and his collaborators have shown a wide range of therapeutic plasma levels of CPZ during both acute and long-term treatment (Curry et al. 1970) and other authors confirm this (Mackay et al. 1974; Rivera-Calimlin et al. 1974). With chronic patients, poor therapeutic effect was found in association with high, low and/or unstable
408
D, H. Wiles and others
levels (Curry & Marshall, 1968). In a detailed study of patients undergoing treatment for acute illness, improvement during the first fortnight was related to plasma CPZ and so were some peripheral side effects, but no overall relationship was found between CPZ level and therapeutic effect (Sakalis et al. 1972). The recent reports of Sakalis et al. (1973) and of Mackay et al. (1974) present evidence that plasma levels of CPZ SO and 7OH CPZ may be related to the therapeutic effect. However, discrepant data on levels of CPZ metabolites and their relation to the levels of the parent drugs have been reported by different groups of workers (Curry & Marshall, 1968; March et al. 1972). SUBJECTS
Seventeen subjects were studied in the course of their CPZ treatment. They were recently admitted acute psychotic patients for whom phenothiazine treatment was indicated and who were free from physical illness. Five had never received psychiatric treatment, another six had received no neuroleptic drugs for at least three previous months, while four had been receiving phenothiazines and two patients antidepressants for various periods before the investigation. It was not considered ethical to withhold treatment to detect any spontaneous remitters. CPZ in syrup was prescribed. A dose between 150 and 600 mg/day was decided on clinical judgement by the patient's psychiatrist of the patient's response in the early days. This dose was kept constant throughout the period of study for all but three patients where there were clinical reasons for changing. Benzhexol was added only if parkinsonian side effects appeared and occasional nitrazepam was the only other drug prescribed. Three patients were discharged as improved within the first 17 days; the remaining fourteen received CPZ from 3 to 8 weeks. Clinical data about the seventeen subjects are shown in Table 1. PROCEDURE
Blood samples for the estimation of plasma levels of CPZ and hormones were collected weekly from all patients (and in six twice a week during the first fortnight). Samples were taken both immediately before and two hours after the morning dose of CPZ (these are referred to
subsequently as the zero and two-hour samples respectively). Pulse and blood pressure were measured in the standing and reclining positions before taking each sample. Clinical assessment was repeated on the days of blood collection. METHODS (a) CHLORPROMAZINE AND METABOLITES
Plasma drug level was determined by Curry's gas-chromatographic method (Curry, 1968) in a modification described by Mackay et al. (1974). This modified method allows measurement of CPZ and three of its metabolites (CPZ SO, 7OH CPZ and N O R ^ P Z ) by using diethyl ether instead of heptane for extraction, and corrects for the variability of extraction by introducing 2,4-dichlorpromazine as internal standard. Precision and accuracy Ten samples containing a 50 ng/ml mixture in plasma when assayed together gave mean levels ±SD of 96 + 4, 99±9, 104± 14 and 105± 11 % of the actual concentration for CPZ, NORjCPZ, 7OH CPZ and CPZ SO. Similar results were obtained over the range 10-100 ng/ml for CPZ, NORjCPZ and CPZ SO; and between 25 and 100 ng/ml for 7OH CPZ. This data was confirmed in a multi-laboratory collaborative qualitycontrol study organized by Dr W. J. Turner of the Central Islip Psychiatric Research Center, New York (Turner et al. 1975). Detector sensitivity Using a Pye Unicam Series 104 E.C. Detector, this was 0-5—1 ng in 5 ft\ injection for CPZ, NORjCPZ and CPZ SO: 1-2 ng for 7OH CPZ. Averages from individual weekly measurements throughout treatment were used to characterize plasma CPZ of each subject both immediately before (time 'zero') and two hours after the morning dose. (In the few patients whose dose was changed, the level on the more prolonged dose was taken.) (b)
HORMONES
Plasma prolactin, luteinizing hormone (LH), testosterone and oestrogens were estimated by
409
Clinical significance of plasma chlorpromazine levels TABLE 1 CLINICAL DETAILS OF SEVENTEEN PATIENTS
Patient
Age Sex
L.W. B.G. J.C. G.S. J.M. D.M. M.M. H.S. A.C. L.Ph. A.A. F.M.
47 26 45 23 27 60 32 26 23 23 33 37
F F M M M F M M M M M F
Paranoid schizophrenia Paranoid-depressive syndrome (puerperal) Paranoid-depressive syndrome Acute paranoid psychosis Paranoid schizophrenia Paranoid reaction Paranoid schizophrenia Paranoid schizophrenia Paranoid schizophrenia Schizophrenia Paranoid schizophrenia Schizophrenia
300/450 400 150-300 300 600-800 400 400 400 500 500 500 300
5-4-8-2 70 2-3-4-7 4-3 10-0-12-7 6-9 4-6 5-2 80 61 8-3 41
51 32 36 7 42 21 56 42 15 27 23 28
M.T.M.
21
F
Paranoid schizophrenia
400
5-5
43
J.G.
21
F
Paranoid schizophrenia
400
7-8
40
J.A.
24
F
450
71
35
V.G. R.A.
25 38
M M
Acute paranoid psychosis (intoxication with antidepressants) Acute paranoid psychosis Catatonic schizophrenia
150 300
2-5 4-3
8 36
Diagnosis
radioimmunoassays. Plasma prolactin was measured by the specific double antibody radioimmunoassay described by McNeilly & Hagen (1974). LH was measured by the method of Midgley (1966) using the second international reference preparation of human menopausal gonadotrophin(2nd IRP-HMG) and the Medical Research Council Pituitary LH 68/40 as standards. Testosterone was estimated as described by Furuyama et al. (1970). Immunoreactive oestrogens were measured by the technique of Hotchkiss et al. (1971). The antiserum was relatively specific for oestradiol 17/? and crossreaction to oestradiol 17a, oestrone and oestriol at 40, 35 and 8 % respectively.
mg/day
No No No No No Stopped over 1 year ago Stopped over 3 months ago Stopped over 3 months ago Stopped over 3 months ago Stopped over 6 months ago Stopped over 3 months ago Tricyclic antidepressants in low doses Trifluoperazine + orphenadzine (4 weeks) CPZ + trifluoperazine (irregular, 2 weeks) Amitriptyline + Parstelin (4 weeks) Thioridazine (6 days) Fluphenazine decanoate + CPZ
scale.2 In addition, changes in sleep, habits, activity and insight were rated on a 4-point scale. A check-list for subjective autonomic symptoms (ratings 0-3) and a rating scale for parkinsonian symptoms (ratings 0-3) were used to assess CPZ side-effects. All the clinical ratings were carried out during the initial interview and repeated on the days of blood sampling. At the end of CPZ treatment a global assessment of change in mental state was made on a 4-point scale (0 = worse/unchanged; 1 = some symptomatic improvement only; 2 = satisfactory but partial improvement, residual symptoms; 3 = good improvement, patient asymptomatic).
RESULTS
(c) CLINICAL RATINGS
A full psychiatric interview was carried out at the first contact with the patient, after obtaining his consent to participate in the study, and the Brief Psychiatric Rating Scale (BPRS, Overall & Gorham, 1962) was used to assess the mental state. The prognostic index (Stephens et al. 1966) was completed at the end of the first interview. Two items were added to the BPRS to assess elation and psychomotor excitation. A detailed glossary has been developed to rate the severity of each symptom on a 7-point
Psychotropic drugs prior to admission
mg/kg/day Days
(a) PLASMA LEVELS OF CPZ AND METABOLITES
A wide range of plasma levels of CPZ and its metabolites was found in patients taking therapeutic doses (Table 2). This variation was not reduced by expressing CPZ levels per daily dose of CPZ in mg or mg/kg body weight. The overall relation between CPZ dose and plasma level is not simple. The data suggest that plasma CPZ levels tend to increase with dosage 2
The glossary is available on request. 27-2
410
D. H. Wiles and others TABLE 2 PLASMA CPZ, CPZ SO, NOR,CPZ AND 7OH CPZ LEVELS IN SEVENTEEN SUBJECTS TREATED WITH A CONSTANT DAILY DOSE AT 150 TO 600 mg CPZ
CPZ SO
CPZ
7OH CPZ
NOR ,CPZ
Oh
2h
Oh
2h
Oh
2h
Oh
2h
All samples
Mean SD Range N
24 24 0-125 82
56 42 0-196 78
11 21 0-126 76
23 30 0-97 72
1 1 0-6 76
2 4 0-26 72
3 7 0-26 76
7 10 0-36 72
Averages for individuals
Mean SD Range N
21 22 0-92 17
50 33 0-110 17
11 20 0-85 16
28 19 3-77 16
1 1 0-4 16
2 4 0-10 16
3 6 0-23 16
6 8 0-27 16
80
r
60
N a. U 5 40 a c
20
200 400 Dose CPZ (mg/day)
600
FIG. 1. Mean plasma CPZ level in ng per ml measured immediately before the morning dose of the drug, plotted against the daily dose.
up to a dose of 400 mg/day; however, patients who received larger doses than this had a lower average level of CPZ than those on 400 mg/day (Fig. 1). If plasma CPZ at two hours is expressed per dose of 1 mg/kg body weight and plotted against the full dose in mg/kg body weight, negative correlation is found between dose and plasma level (r = -0-57; P < 0-05). Indeed the six patients who received CPZ in doses above 7 mg/kg/day had among the lowest ratios of plasma CPZ to daily dose. The clinical features of these six were those of a florid paranoid syndrome with some elation and/or psychomotor excitation, while only two of the remaining 11 patients had a similar clinical picture. The plasma CPZ levels of individual patients
also varied considerably from week to week even though dosage was unchanged: thus up to a sevenfold difference was recorded at zero time. Among 14 patients for whom we have repeated samples over at least three weeks, 9 had their highest CPZ levels in the first fortnight, 3 had this later and in two levels did not change. In the 3 patients whose dosage was altered, plasma CPZ changed in the corresponding direction and fluctuated around a new mean. Benzhexol was introduced during the fourth or fifth week in four subjects. This was not followed by any consistent change in plasma CPZ levels. With only one exception, CPZ SO was detected both before and after the CPZ dose in all subjects when CPZ was present. Its level was on average about half that of CPZ but the mean ratio of CPZ SO to CPZ from individual patients ranged from 0-14 to 0-89 at zero time and 0-33 to 0-93 at two hours. There was, however, in the group of patients a strong correlation between plasma CPZ and CPZ SO levels both at zero time (r = 0-92; P < 0001) and at two hours (r = 0-87; P < 0001). (b) PLASMA CPZ RESPONSE
LEVEL
AND
CLINICAL
Of the seventeen patients, ten improved and had been discharged at the end of the investigation. Their global rating was 2 or 3 and their BPRS score decreased by 58-94% of initial value. For seven patients the therapeutic effect was poor (global rating 0 or 1; change in BPRS score from - 4 4 to +30 %) and for six patients other methods of treatment had to be introduced.
411
Clinical significance of plasma chlorpromazine levels TABLE 3
AVERAGE PLASMA LEVELS OF CPZ AND CPZ SO AND THEIR RATIO IN 'RESPONDERS' AND 'NON-RESPONDERS'
CPZ (ng/ml)
'Responders'
CPZ SO (ng/ml)
SO/CPZ
Daily dose (mg)
Oh
2h
Oh
2h
Oh
2h
370 106 10
19 13 10
45 29 10
6 5 9
25 13 9
0-29
0-54 018 9
34 28 7
P < 002 0-53 0-21 5
Mean SD N
on 9
P if significant 'Non-responders'
453 155 7
Mean SD N
25 29 7
15 27 7
62 42 7
0-52 014 6
TABLE 4 AVERAGE PLASMA LEVELS OF CPZ, CPZ SO AND PROLACTIN IN EIGHT SUBJECTS SHOWING THE EXTRAPYRAMIDAL SYNDROME (EPS) AND SEVEN SUBJECTS WITH NO EPS. ALL HAD RECEIVED CPZ TREATMENT FOR AT LEAST FIFTEEN DAYS
Plasma CPZ (ng/ml)
EPS
Plasma prolactin (ng/ml)
Daily dose (mg)
Oh
2h
Oh
2h
Oh
2h
406 126 8
30 26 8
75 29 8
16 24 8
40 21 8
37 16 8
41 14 8
—
—
P < 001
—
P < 005
—
—
454 112 7
15 12 7
31 23 7
5 4 6
17 12 6
25 13 7
28 21 7
Mean SD N
P if significant Mean SD N
No EPS
Plasma CPZ SO (ng/ml)
100 r
80 -
0
60 ft.
5. 40
o o
o
O
* 20 "
o
o i
10
i
o
;
i
i
20 30 40 50 Mean plasma prolactin (ng/ml)
i
i
60
70
FIG. 2. Mean plasma prolactin and CPZ two hours after dosage in eight subjects showing the extrapyramidal syndrome (EPS) (•) and seven subjects showing no EPS (O).
412
D. H. Wiles and others
Six of these unimproved patients had plasma CPZ levels equal to or higher than those of the improved group (Table 3); only one had very low or undetectable plasma drug levels on nearly all occasions. A significant (P < 0-02) feature of these non-responders was a higher ratio of CPZ SO to CPZ at zero time (0-53) than that of the responders (0-29). Although the groups did not differ in the initial BPRS score or prognostic index, they showed some clinical differences. All 'responders' had an episode of acute onset and either no previous admission (5) or relatively good remissions in the past (5). Among 'nonresponders', three had a long history of schizophrenic illness with poor remissions, three presented a paranoid episode with marked elation and one suffered from puerperal psychosis with paranoid syndrome coloured by depression and anxiety. More 'non-responders' than 'responders' developed parkinsonian side effects (5 of 7 and 3 of 10 respectively).
(c) PLASMA DRUG LEVELS AND EXTRAPYRAMIDAL SIDE EFFECTS (EPS)
Among 15 patients treated with CPZ for fifteen days or more, 8 developed a parkinsonian syndrome (all within the first three weeks of CPZ administration) and 7 did not. The difference in average CPZ dosage between the groups was insignificant (406 and 454 mg/day respectively) but they differed considerably in plasma CPZ level (Table 4). Seven out of eight patients with EPS had mean levels at two hours > 50 ng/ ml, while only one out of seven patients with no EPS had as high a plasma CPZ (Fig. 2). Mean levels of both CPZ and CPZ SO were higher in the EPS group, significantly so at two hours (P < 001 and P < 005). There was no difference in levels of 7OH and NORjCPZ. Changes in pulse and blood pressure were rather small and unrelated to plasma CPZ levels. There was no overall correlation between subjective autonomic side effects and plasma CPZ but in individual patients the maximal severity of these symptoms were either soon before or at the same time as the highest CPZ levels. Only 1 patient out of the 6 whose mean
plasma prolactin at two hours remained below 30ng/ml developed EPS, while 7 out of 9 patients with a higher two hours prolactin developed parkinsonism.
{d) PROLACTIN LEVELS
Plasma prolactin was measured throughout treatment in 17 patients (7 women and 10 men). In all 7 women and in 5 of the 10 men prolactin levels were elevated (i.e. above 20ng/ml) from the end of the first week of CPZ administration and throughout the treatment period. Prolactin fluctuated erratically with no consistent further increase after the first week. (The sole exception was a women who developed mild galactorrhoea in the third week of treatment (with 300 mg/day) and whose prolactin levels rose progressively until the third week.) Only in two male patients was there a rise in plasma prolactin after every morning dose of CPZ. In the remaining subjects the changes of prolactin levels following the morning dose of CPZ were inconsistent. The mean plasma prolactin level of individual patients was unrelated to the daily dose of CPZ but there was a significant positive correlation between mean plasma prolactin and mean CPZ concentration. Fig. 3 shows this at zero time (r = 0-77; P < 001); the correlation coefficient between corresponding levels at two hours was 0-52 (P < 005). Means of weekly estimations were used here because prolactin levels fluctuate from hour to hour and CPZ levels also vary. Thus the average or sustained level of prolactin is more likely to be biologically related to the sustained level of CPZ. However, correlations were also calculated using every pair of values and these were also significant: for' zero' samples -r = 0-51; P < 0001, for all twohour samples -r = 0-34; P < 0005. There were insufficient readings from some subjects to allow calculation of r for each subject separately or for variance analysis. Mean plasma prolactin also correlated with mean CPZ SO levels, significantly in 'zero' samples (r = 0-58; P < 002) and both in zero and two-hour samples when all paired observations were included (P < 0001). Plasma prolactin levels tended to be higher
413
Clinical significance of plasma chlorpromazine levels 80 r
I60 c
I
I 40 CO
V)
"B. c
± 20
20
40
60
80
100
Mean plasma CPZ (ng/ml) FIG. 3. Correlation between plasma levels of CPZ and prolactin in seventeen subjects. Mean levels for individuals from samples taken before the morning dose of the drug, r = 0-67; P < 0 0 1 .
in subjects who developed extrapyramidal symptoms (EPS). Thus 7 out of 8 such patients had a mean two-hours prolactin > 30 ng/ml, while in only 2 of 7 patients without these symptoms was prolactin so elevated (Fig. 2). Prolactin elevation preceded the appearance of EPS by 1-2 weeks. (e) SEX HORMONES
Measurements of LH, testosterone, oestrogens and progesterone throughout treatment gave values within the normal ranges. No correlation was found between sex hormone levels and plasma CPZ. DISCUSSION
We used the method described by Mackay et al. (1974) in preference to Curry's method because we have found it to be more reliable and because it measures the derivatives we have described. Although an improvement on Curry's method, it is not without potential disadvantages. Thus ether would extract chlorpromazine Noxide if this were present in significant amounts in plasma; and this substance would in turn decompose to CPZ and desmethyl-CPZ on the column, thus falsifying the values obtained for these substances. However, although CPZ-N-
oxide has been found in urine, it has not been demonstrated convincingly in significant amounts in plasma. The use of 2,4-dichlorpromazine as a standard improves the reliability, even though this substance will not, of course, behave exactly in the same way as CPZ in the extraction procedure. The range of plasma CPZ levels found in these patients is comparable to those described by Sakalis et al. (1972) in acute patients, and by Curry et al. (1970) and Mackay et al. (1974) in chronic ones. The relationship between CPZ dose and plasma level was complicated: at doses up to 400 mg/day or 7 mg/kg/day, plasma levels were related to the dose but with higher doses plasma levels of CPZ were actually lower. Mackay et al. (1974) reported a similar finding in chronic patients, namely lower levels in subjects on 600 mg/day than in those receiving 400 mg/day; and Martensson & Roos (1973) describe a comparable result with thioridazine. In our group a low ratio of plasma level of CPZ to dose was often found among patients who were on doses exceeding 400 mg/day and who also presented a florid paranoid syndrome. CPZ absorption and metabolism may depend on the dose administered, but it might also be affected by physiological changes related to a particular psychotic syndrome. Like other
414
D. H. Wiles and others
investigators, we found considerable variation of CPZ levels in the same patient on different occasions, but we did not observe the regular decline after the first week reported by Sakalis et al. (1972). We also failed to repeat the observation of Rivera-Calamlin et al. (1973) that addition of benzhexol is accompanied consistently by lower CPZ levels. Although we used the method of assaying metabolites that Mackay et al. (1974) described, our findings were different. Thus we did not detect 7OH CPZ regularly, nor could we relate the ratio of either 7OH CPZ to CPZ or of 7OH CPZ to CPZ SO, to outcome as these workers, and also Sakalis et al. (1973), had done. This may be because our assay was less sensitive to 7OH CPZ, or may be because these authors studied chronic patients who had received CPZ for 9-6 years on average, while ours had received the drug for a much shorter time. The ratio of sulphoxide to CPZ was consistently higher two hours after the morning dose than twelve hours after the evening dose (i.e. immediately before the morning dose). As there was no significant overall correlation between this ratio and CPZ level, the difference may have been due to the shorter plasma halflife of the sulphoxide (Davidson et al. 1957). We were also unable to confirm the finding of Mackay et al. (1974) that response is related to the ratio of 7OH CPZ to CPZ; we did, however, find a higher ratio of sulphoxide to CPZ in unresponsive patients. In any case there are clinical differences between responders and nonresponders that must obscure relationships between responsiveness and drug levels. We have reported in a preliminary account that prolactin level was related both to plasma CPZ level and to the appearance of a parkinsonian syndrome (Kolakowska et al. 1975). This rise in prolactin preceded the extrapyramidal syndrome (EPS) by 1-2 weeks. The parkinsonian syndrome was also related to plasma CPZ levels. The effect on prolactin release and on EPS are both thought to depend on dopaminergic mechanisms, though of different location - one nigrostriatal and the other hypothalamic. The findings suggest that the two dopaminergic systems have a similar threshold of response to CPZ, and
that prolactin level can be used as an index of the antidopaminergic central effect of CPZ. One of the theories put forward to explain the antipsychotic effects of CPZ is that it depends on antidopamine properties of the drug. However, these were not related to plasma CPZ level in our patients. It has been suggested that the dopaminergic system in the limbic area is of particular importance for pathophysiology of schizophrenia (Stevens, 1973). It has also been shown in rats that effects of neuroleptics on dopamine metabolism differ among these three brain areas (Anden, 1972; Bowers & Rozitis, 1974). These regional differences might be of importance for the dissociation between antipsychotic effects of the drug on one hand and, parkinsonism and prolactineemia on the other. It is also possible that subjects' response to changes of dopamine metabolism varies more for psychological than for extrapyramidal and hormonal effects. The changes in the mental state which occur during CPZ administration are not only more complex and more difficult to measure but they depend on factors other than the pharmacological action of the drug. Finally, the relationship between CPZ plasma level and therapeutic effect is obscured by the nature of the psychosis which tends to remit spontaneously in some patients, while in others it takes a progressive course, irrespective of any treatment. Perhaps it is only within a more uniform group of potential 'responders' that plasma levels of the drug could be related to therapeutic effects. Such a group would presumably consist of patients who are likely neither to have an early spontaneous remission nor to have a poor prognosis irrespective of treatment. However, neither clinical judgement nor prognostic scales permit an early and reliable separation of such groups of patients. We are grateful to Miss A. Garfield for technical assistance. Drug standards were kindly supplied by Dr A. A. Manian, United States National Institute of Mental Health. We are grateful to Drs S. H. Curry and A. F. Healey for advice on methodology, to Mrs E. Green and Mrs E. Kuht for nursing assistance and to the consultants of Littlemore Hospital for permission to study patients under their care.
Clinical significance of plasma chlorpromazine levels REFERENCES Anden, N. E. (1972). Dopamine turnover in the corpus striatum and the limbic system after treatment with neuroleptics and anti-acetylocholine drugs. Journal of Pharmacy and Pharmacology 29, 340-348. Beumont, P. S. V., Corker, C. S., Friesen, H. G., Kolakowska, T , Mandelbrote, B. M., Marshall, J., Murray, M. A. F. & Wiles, D. H. (1974). The effects of phenothiazines on endocrine function: II. British Journal of Psychiatry 124, 420-430. Bowers, M. B., Jr & Rozitis, A. (1974). The regional differences in homovanilic acid concentrations after acute and chronic administration of antipsychotic drugs. Journal of Pharmacy and Pharmacology 26, 743-745. Cooper, S. F., Albert, J. M., Hillel, i. & Caille, G. (1973). Plasma level studies of chlorpromazine following the administration of chlorpromazine hydrochloride and chlorpromazine embonate in chronic schizophrenics. Current Therapeutic Research 15, 73-77. Curry, S. H. (1968). Determination of nanogram quantities of chlorpromazine and some of its metabolites using gas liquid chromatography with an electron capture detector. Analytical Chemistry 40, 1251-1255. Curry, S. H. & Marshall, J. H. L. (1968). Plasma levels of chlorpromazine and some of its relatively non-polar metabolites. Life Sciences 7, 9-17. Curry, S. H., Marshall, J. H. L., Davis, J. M. & Janovsky, D. S. (1970). Chlorpromazine plasma levels and effects. Archives of General Psychiatry 22, 289-296. Davidson, J. D., Terry, L. L. & Sjoerdsma, A. S. (1957). Action and metabolism of chlorpromazine sulfoxide in man. Journal of Pharmacy and Experimental Therapeutics 121,8-12. Furuyama, S., Mayes, D. M. & Nugent, C. A. (1970). A radioimmunoassay for plasma testosterone. Steroids 16, 415-428. Hotchkiss, J., Atkinson, L. E. & Knobil, E. (1971). Time course of serum oestrogen and luteinizing hormone concentrations during the menstrual cycle of the rhesus monkey. Endocrinology 89, 177-183. Kolakowska, T., Wiles, D. H., McNeilly, A. S. & Gelder, M. (1975). Correlation between plasma levels of prolactin and chlorpromazine. Psychological Medicine 5, 214-216. Mackay, A. V. P., Healey, A. F. & Baker, J. (1974). The relationship of plasma chlorpromazine to its 7-hydroxy and sulphoxide metabolites in a large population of chronic
415
schizophrenics. British Journal of Clinical Pharmacology 1, McNeilly, A. S. & Hagen, C. (1974). Prolactin, TSH, LH and FSH responses to a combined LHRH/TRH test at different stages of the menstrual cycle. Clinical Endocrinology 3, 427-435. March, J. E., Donato, D.,Turano, P. & Turner, W. J. (1972). Interpatient variation and significance of plasma levels of chlorpromazine in psychotic patients. Journal of Medicine (Basel) 3, 146-162. Martensson, E. & Roos, B. E. (1973). Serum levels of thioridazine in psychiatric patients and healthy volunteers. European Journal of Clinical Pharmacology 6, 181-186. Midgley, A. R. (1966). Radioimmunoassay: a method for human chorionic gonadotrophin and human luteinizing hormone. Endocrinology 79, 10-18. Overall, G. E. & Gorham, D. R. (1962). The brief psychiatric rating scale. Psychological Reports 10, 799. Rivera-Calimlin, L., Castaneda, L. & Lasagna, L. (1973). Effects of mode of management on plasma chlorpromazine in psychiatric patients. Clinical Pharmacology and Therapeutics 14, 979-985. Sakalis, G., Curry, S. H., Mould, G. P. & Lader, M. H. (1972). Physiologic and clinical effects of chlorpromazine and their relationship to plasma level. Clinical Pharmacology and Therapeutics 13, 931-946. Sakalis, G., Chan, T. L., Gershon, S. & Park, S. (1973). The possible role of metabolites in therapeutic response to chlorpromazine treatment. Psychopharmacologia (Berlin) 32, 279-284. Stephens, J. H., Astrup, Ch. & Mangrum, S. C. (1966). Prognostic factors in recovered and deteriorated schizophrenics. American Journal of Psychiatry 120, 1116-1121. Stevens, J. R. (1973). An anatomy of schizophrenia. Archives of General Psychiatry 29, 177-189. Turano, P., March, J. E., Turner, W. J. & Merlis, S. (1972). Qualitative and quantitative report on chlorpromazine and metabolites in plasma, erythocytes and erythrocyte washings from chronically medicated schizophrenic patients. Journal of Medicine (Basel) 3, 109-120. Turner, W. J., Turano, P., Badzinski, S., Breyer, U., Curry, S., Kaul, P. & Wiles, D. (1975). An attempt to establish quality control in determination of plasma chlorpromazine by a multi-laboratory collaboration. In Pharmacokinetics, Blood Levels and Clinical Response of Psychoactive Drugs (ed. L. Gottschalk & S. Merlis). Spectrum: New York.
