Drug Safety 7 (Suppl. I): 17-25, 1992 0114-5916/92/0001-0017/$4.50/0 © Adis International Limited. All rights reserved.

DSSUP3292

Mechanisms of Clozapine t -Induced Agranulocytosis

Stanton L. Gerson I and Herbert M eltzer2

Department of Medicine and the Ireland Cancer Center, University Hospitals of Cleveland, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA

2 Department of Psychiatry, University Hospitals of Cleveland, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA

Summary The aetiology of c1ozapine-induced agranulocytosis remains unknown. Leading candidates in-clude an immune mechanism that is possibly complement- or drug-dependent and a toxic mech­anism. We analysed these mechanisms by culturing the granulocyte precursor stem cell from the bone marrow in the presence of patients' serum, c10zapine or c10zapine metabolites. Studies with patients' serum failed to identify an immune mechanism. On the basis of our preliminary data, it appears that a toxic mechanism may be responsible, and this is more likely to be due to a metabolite than to c10zapine itself. Further studies are required to determine the sensitivity of bone marrow precursors to these c10zapine derivatives. For instance, prospective collection of serum will make it possible to evaluate whether high metabolite concentrations develop in sen­sitive individuals and whether they are responsible for agranulocytosis. If such elevated levels occur, further studies will be required to determine whether prospective monitoring will effec­tively identify patients at risk and ultimately prevent the onset of agranulocytosis by early dis­continuation of the drug.

Clozapine is an effective therapy for treatment­resistant schizophrenia. Its widespread use has been hampered by a high incidence of agranulocytosis. In the US, the annual rate approaches 1.6% (Bal­dessarini & Frankenburg 1991), which is higher than that with any other drug causing idiosyncratic agranulocytosis (Coulter & Edwards 1990; Pis­ciotta 1973). As with other drugs causing blood dyscrasias, the aetiology of the disorder and the presence of risk factors have been difficult to es­tablish. Currently, there are no defined mechan­isms of clozapine-induced agranulocytosis. To pro­vide a framework for our recent studies into this disorder, we will review mechanisms of agranulo­cytosis induced by other drugs and the character-

tClozaril® /Leponex®

IStlCS of clozapine-induced agranulocytosis, since these may provide clues to the mechanism asso­ciated with clozapine use. We will then discuss our recent studies into the mechanism of clozapine­induced agranulocytosis.

1. Mechanisms of Agranulocytosis

There are many mechanisms potentially re­sponsible for drug-induced agranulocytosis. These include the presence of either antigranulocyte or anti myeloid stem cell antibodies, the presence of a toxic metabolite that accumulates in the blood­stream of certain individuals, or a selective idio­syncratic sensitivity of the bone marrow stem cells

18

Toxic suppression of myelopoiesis

Drug Safety 7 (Suppl. 1) 1992

Drug ~ Metabolite

!/ @-

CFU-GM stem cell

Myelocyte Granulocyte blood cell

Immune-mediated suppression of myelopoiesis by drugs

r - Drug

+ _ Complement

~ --.... 6-lymphocyte

Antibody

,/' +

~ @-CFU-GM Myelocyte Granulocyte stem cell blood cell

Fig. 1. Possible toxic and immunological mechanisms involved in drug-induced agranulocytosis. CFU-GM = colony forming unit-granulocyte-macrophage.

to the drug itself (fig. I). For instance, penicillins (Murphy et at. 1983), aprindine (Pisciotta & Cronkite 1983), ibuprofen (Mamus et at. 1986) and quinidine (Kelton et at. 1979) have all been asso­ciated with antibody-mediated agranulocytosis, in which antibodies, either in the absence or presence of the parent compound, are toxic to myeloid pre­cursors in the bone marrow. Chloramphenicol (Jiminez et at. 1987; Yunis 1980), phenylbutazone (Smith et at. 1977) and penicillin (Neftel et at. 1983) are 3 drugs whose toxic metabolites have been shown to suppress haematopoiesis, as indicated by

inhibition of the growth of haematopoietic stem cells of the myeloid lineage in vitro. With these 3 compounds, the metabolites are much more toxic to the bone marrow than the parent compound. However, blood levels of these metabolites in patients affected and unaffected by agranulocytosis have not been shown conclusively to reach the lev­els at which toxicity is identified in vitro. Less com­monly, drugs can elicit a T cell-mediated response in which a cytotoxic T lymphocyte is responsible for agranulocytosis in the absence of the drug or metabolite (Gualde & Malinvaud 1982). In this

C1ozapine-Associated Agranulocytosis

case, the cytotoxic T lymphocyte is thought to di­rectly kill the haematopoietic precursor in the bone marrow and thus suppress haematopoiesis. Chlor­promazine is thought to be directly toxic to the bone marrow and has been identified as being re­sponsible for metabolic suppression of haemato­poietic growth in susceptible individuals (Pisciotta 1971). Carbamazepine causes a dose-dependent suppression of myelopoiesis, and this is thought to be responsible for the relatively high incidence of neutropenia associated with its use (Gallicchio & Hulette 1989). It is interesting that the combina­tion of carbamazepine and lithium, which stimu­lates haematopoiesis through the production of colony stimulating factors such as granulocyte­macrophage colony stimulating factor (GM-CSF), offsets the neutropenia caused by carbamazepine (Kramlinger & Post 1990). However, as we will discuss in this paper, the combination of lithium and clozapine is not generally indicated, and has been reported to be harmful in at least one case (Gerson et al. 1991).

2. Overview of Haematopoiesis

In order to understand more fully the nature of clozapine-induced agranulocytosis, we will review those aspects of haematopoiesis that appear to be disrupted in patients with this disorder. Peripheral blood granulocytes, red blood cells and platelets are all derived from precursors in the bone marrow. The myeloid precursor is known as the colony forming unit-granulocyte-macrophage (CFV-GM), the red cell is derived from the burst forming unit­erythroid (BFU-E) and the platelet is derived from the CFU-megakaryocyte (CFV-Meg). Each of these precursor haematopoietic stem cells is derived from a common stem cell known as the CFU-granulo­cyte-erythrocyte-monocyte-megakaryocyte (CFV­GEMM), which is itself derived from a pluripo­tential haematopoietic stem cell that produces both myeloid and lymphocytic progeny. Mean transit time of peripheral blood granulocytes is approxi­mately 8 to 9 hours, whereas red cells survive ap­proximately 120 days, and platelets approximately 7 to 10 days. Thus, bone marrow suppression of

19

any multi-lineage cell such as the CFU-GEMM or the pi uri potential stem cell will manifest itself in­itially by a decrease in the granulocyte count, fol­lowed by a decline in the platelet count, whereas a much longer period of time is required before a decline in the haematocrit is observed. In addition, granulocyte maturation requires 4 to 7 days from the appearance of the earliest stem cell to that of the mature peripheral blood granulocyte. Thus, suppression of the myeloid stem cell becomes manifest 4 to 7 days later by a precipitous drop in the granulocyte count, leading to agranulocytosis.

3. Characteristics of Clozapine-Induced Agranulocytosis

Dr Lieberman has reviewed the epidemiology of clozapine-induced agranulocytosis in this issue and elsewhere (Lieberman et al. 1989a). Here, we will review certain characteristics of the onset and clinical course of clozapine-induced agranulocyto­sis that may be helpful in the understanding of its aetiology. Firstly, clozapine-induced agranulocy­tosis has a delayed onset: approximately 76% of cases occur between the fourth and eighteenth week of therapy (Anderman & Griffith 1977; Idanpaan­Heikkila et al. 1977; Krupp & Barnes 1989). Since the dose of clozapine is usually increased from a starting dose of 25 to 50 mg/day up to 300 to 600 mg/day over the first 4 to 6 weeks of treatment, this lag time does not allow one to distinguish be­tween a toxic or an immune mechanism. For in­stance, if a toxic mechanism is involved, build-up of the drug or metabolite may be delayed, whereas if an immune response takes place, this may be both time and dose dependent. Secondly, once the granulocyte count begins to fall, agranulocytosis develops rapidly, usually over 2 to 5 days, which is consistent with a sudden halt in production of granulocytes within the bone marrow. This sug­gests that the target cell is an early myeloid pre­cursor. In some patients, a gradual decline in gran­ulocyte count is observed before the onset of agranulocytosis, whereas in others only mild to moderate granulocytopenia is observed, with an absolute neutrophil count (ANC) of approximately

20

500 to 1500/~1 and failure to progress to agranu­locytosis. In those patients who only develop gran­ulocytopenia, a compensatory increase in granu­lopoiesis may take place or, alternatively, the aetiology may be quite different from that ob­served with patients who go on to develop agran­ulocytosis.

The third striking feature of clozapine-induced agranulocytosis is the fact that even when the medication is stopped because the ANC falls below 1000/~1 patients go on to develop agranulocytosis lasting between 14 and 24 days. Since early drug suspension does not prevent the complication, the mechanism of inhibition must be set in place by the time it is first recognised. It also suggests that whatever the impediment to myelopoiesis, it is in­fluencing or acting on an early stem cell and having a uniformly myelosuppressive action in affected individuals. Otherwise, there would be more var­iability in the course and onset of the disease. Fourthly, the bone marrow in affected individuals almost always shows an absence of myeloid pre­cursors and the presence of only occasional pro­myelocytes and myeloblasts. Marked relative erythroid hyperplasia is present, with an inversion of the characteristic myeloid to erythroid ratio from a normal value of 3 : I to approximately I : 5.

Taken together, these factors suggest that an early myeloid precursor is affected by clozapine and that simple peripheral destruction of granulocytes is not responsible for the observed agranulocytosis. In fact, it is remarkable that most cases of cloza­pine-associated agranulocytosis show an isolated defect in myeloid precursors in the bone marrow and a predominance of erythroid maturation. Mild anaemia is noted in some patients receiving clo­zapine, irrespective of whether or not they develop agranulocytosis.

Furthermore, we have recently reported a case of hypoplastic anaemia in a patient who died of clozapine-associated agranulocytosis; however, this patient had been taking a number of other medi­cations, including carbamazepine, clonazepam and had recently stopped taking lithium (Gerson et al. 1991). Cases of mild thrombocytopenia have also occurred, raising the possibility that an early hae-

Drug Safety 7 (Suppl. 1) 1992

matopoietic stem cell could also be affected in this disorder (P. Barnes; S.L. Gerson, unpublished ob­servations).

4. Risk of Clozapine-Induced Agranulocytosis

In the United States, clozapine appears to cause a higher incidence of agranulocytosis than any other chronically prescribed medication, with the excep­tion ofthose used for the treatment of malignancy. This is especially important considering the chronic outpatient use of this medication in patients with schizophrenia and other mental disorders. The in­cidence of clozapine-induced agranulocytosis var­ies from less than 0.2% in some European coun­tries to 1.6% in the United States (Anderman & Griffith 1977; Baldessarini & Frankenburg 1991; Idanpaan-Heikkila et al. 1977; Krupp & Barnes 1989; Lieberman et al. 1989b). The incidence of agranulocytosis may be higher in some ethnically uniform populations such as the Finnish popula­tion or the Ashkenazi Jewish population in New York (Idanpaan-Heikkila et al. 1977; Lieberman et al. I 989b). The risk of death from clozapine-as­sociated agranulocytosis has declined sharply since its introduction in 1974 (Baldessarini & Franken­burg 1991; Krupp & Barnes 1989). In the United States, there has been only one death in approxi­mately 80 cases of agranulocytosis in the 16-month period between February 1990 and June 1991 (Gerson et al. 1991). The improved survival rate in patients with this disorder probably emanates from improved recognition, better patient man­agement and supportive care during periods of sev­ere agranulocytosis, and may have been positively influenced in the United States by the use of pro­spective weekly blood count monitoring, which was designed to detect cases at an early stage and per­haps prevent the sequela of unrecognised neutro­penic fever.

S. Studies into the Mechanism of Clozapine-Induced Agranulocytosis

We have evaluated both immune and toxic mechanisms of clozapine-induced agranulocytosis. In the studies into an immune mechanism, acute­phase serum was collected within 48 hours of the onset of agranulocytosis and was evaluated for the

Clozapine-Associated Agranulocytosis

Table I. Effect of acute-phase serum on stem cell growthB

Patient donor CFU-GM + serum (% control growth)

1

2 3 4

5

89 119 95 98

104

Healthy donor 114

a Acute-phase serum was collected from 5 patients with clozapine-induced agranulocytosis within 48 hours of discontinuing clozapine. Normal nonadherent mononuclear bone marrow cells were cultured at 1 x 105 cells/ml, in the absence or presence of 10% v/v serum, in 0.8% methylcellulose supplemented with 200 units granulocyte­macrophage colony stimulating factor, as previously described (Lazarus et al. 1991).

Abbreviation: CFU-GM = colony forming unit-granulocyte­macrophage.

presence of anti myeloid precursor antibodies using the bone marrow of both healthy donors and patients as the source of haematopoietic precursor stem cells. In the studies into a toxic mechanism, clozapine and a variety of clozapine metabolites and derivatives were evaluated for their toxicity by measuring the effects of these compounds on haematopoietic stem cell growth in vitro, using haematopoietic precursors from healthy donors as targets.

6. Studies of Possible Immune Mechanisms of Clozapine-Induced Agranulocytosis

When acute-phase serum from 5 patients with clozapine-induced agranulocytosis was added to haematopoietic stem cell colony assays, no inhi­bition ofCFU-GM stem cell growth occurred (table I), suggesting that an antibody directly toxic to these stem cells is absent. Further studies evaluating the need to combine complement and/or clozapine or a clozapine metabolite with acute-phase serum in order to reveal a potential immune-mediated pro­cess are currently underway and may yield im­portant data, particularly since a combination of

21

drug, complement and patient's serum has been re­quired to detect an immune mechanism for some drugs (Kelton et al. 1979; Pisciotta & Cronkite 1983). On the other hand, sera from patients with clozapine-induced agranulocytosis occasionally caused growth stimulation, as might be expected if a cytotoxic antibody was absent and if circulating levels of haematopoietic growth factors such as granulocyte-CSF (G-CSF) were elevated, as they often are during neutropenia. In very preliminary studies, we have observed inhibitory activity in the serum of 2 other patients with clozapine-induced agranulocytosis (data not shown), and we are in the process of evaluating whether this is an immune­or toxin-mediated process. However, for the most part, there does not appear to be clear evidence for immune-mediated bone marrow stem suppression in patients with clozapine-induced agranulocytosis.

7. Studies of Possible Toxic Mechanisms of C1ozapine-Induced Agranulocytosis

We are currently analysing data from an exten­sive study into possible toxic mechanisms respon­sible for clozapine-induced agranulocytosis (Ger­son et aI., unpublished data). This study is intended to test the hypothesis that the parent compound, clozapine, or a derivative, is directly toxic to hae­matopoietic stem cells in the bone marrow, and that this toxicity can be detected by in vitro stem cell assays. Steady-state plasma levels of N-des­methyl clozapine and N-oxide clozapine, the major metabolites of clozapine in humans, have been measured by a number of investigators (Ackenheil 1989; Bondesson & Lindstrom 1988; Haring et al. 1989; Perry et al. 1991; Simpson & Cooper 1978) and are outlined in table II. We postulated that, if I or more than I clozapine metabolite is toxic to the bone marrow at concentrations 2 to 5 times higher than those normally achieved in the serum, then the idiosyncratic nature of clozapine-associ­ated agranulocytosis might be due to the accu­mulation of I or more than I metabolite in the blood or bone marrow. Alternatively, patients with agranulocytosis may be particularly sensitive to the drug or a metabolite. Table III shows the mean

22 Drug Safety 7 (Suppl. 1) 1992

Table II. Steady-state plasma concentrations of clozapine and its metabolites

Reference Mean plasma concentrations (mg/L) [range] Days on treatment

clozapine N-desmethyl N-oxide clozapine

clozapine

Ackenheil (1889) 0.31 0.11 0.1 10-30

Bondesson & 0.5 [0.06-0.82] 0.4 [0.03-0.83] > 30 Lindstrom (1988)

Perry et al. (1991) 0.41 [0.08-1.09] 0.123 [0.03-0.27] 28

Simpson & Cooper [1.3-2.2]

(1978)

Haring et al. (1989) 0.16 [0.01-0.53] >8

Table III. Haematopoietic stem cell toxicity of clozapine and its derivatives8

Compound

Clozapine N-Desmethyl clozapine

N-Oxide clozapine

N-Nitroso clozapine

6-Hydroxy clozapine

N-Desmethyl-N-hydroxy clozapine

IC50 for CFU-GM

(mg/L)

32 3.2

> 50 24

35 16

a Healthy human bone marrow cells were cultured in

methylcellulose as previously described (Lazarus et al. 1991),

in the absence or presence of increasing concentrations of

the compounds listed. Briefly, nonadherent mononuclear cells

from the bone marrow of healthy donors were suspended at 1 x 105 cells/ml in 0.8% methylcellulose in the presence

of 30% fetal calf serum, 50 units/ml interleukin-3 and 200

units/ml granulocyte-macrophage colony stimulating factor

(kindly provided by Drs D. Oette and E. Liehl, Sandoz Research Institute). Cells were cultured for 14 days at 37°C

and scored for CFU-GM. The maximum drug concentration used was 50 mg/L. Stem cell toxicity is expressed in terms

of the drug concentration producing 50% inhibition of cell

growth (IC50). Abbreviation: CFU-GM = colony forming unit-granulocyte­

macrophage.

concentrations of clozapine and its derivatives that inhibit growth of the myeloid CFU, CFU-GM, by 50% (ICso). In most instances, higher concentra­tions than those achieved in the serum are required for a toxic effect to be observed. This makes it un-

likely that these compounds are responsible for mediating clozapine-associated agranulocytosis through a direct toxic mechanism. In this regard, our results are similar to those of Pisciotta et al. (1985). In contrast, one compound, N-desmethyl clozapine, the most common metabolite found in the serum of patients receiving clozapine, was much more toxic to the CFU-GM, with an ICso value of approximately 3 mgfL, only 3- to 5-fold higher than the concentrations normally found in the serum.

This suggests 2 possible mechanisms through which N-desmethyl clozapine may be a cause of clozapine-associated agranulocytosis. Firstly, it is possible that patients with agranulocytosis are hy­persensitive to clozapine or N-desmethyl cloza­pine. Alternatively, it is possible that affected patients metabolise clozapine differently and have increased serum levels of N-desmethyl clozapine. For this reason, we are evaluating the sensitivity of CFU-GM to N-desmethyl clozapine in healthy donors and in patients who have recovered from clozapine-induced agranulocytosis. As it is difficult to obtain bone marrow from these patients, we have developed an assay that enables us to culture CFU­GM from the peripheral blood. This will allow us to establish whether the affected patients are more sensitive to N-desmethyl clozapine than healthy donors or patients without clozapine-associated agranulocytosis. Evaluation of the second option requires prospective collection of serum to deter­mine whether N-desmethyl clozapine accumulates in the blood of patients and whether there is a re-

Clozapine-Associated Agranulocytosis

lationship between N-desmethyl clozapine blood levels and the onset of either neutropenia or agran­ulocytosis. If an association can be identified, it may be possible to devise a predictive blood test that measures N-desmethyl clozapine levels and that may be valuable for identifying cases early and/ or preventing the development offull-blown agran­ulocytosis.

8. Clozapine-Associated Agranulocytosis in Cases of Polypharmacy

We recently reported the first death from clo­zapine-associated agranulocytosis in the United States (Gerson et al. 1991). This patient developed a severe hypoplastic bone marrow and a moderate anaemia and moderate thrombocytopenia associ­ated with profound agranulocytosis of rapid onset. The bone marrow was hypoplastic, with decreased erythroid and megakaryocytic precursors, and myeloid precursors were absent. The acute-phase bone marrow from this patient grew very small numbers of CFU-GM (10/105 bone marrow cells) and these were not increased by the use of hydro­cortisone, which is known to block the effect of cytotoxic T lymphocytes (Bagby et al. 1983) [table

23

IV]. Acute-phase serum from the patient was not cytotoxic to normal haematopoietic precursors. Since this patient was receiving carbamazepine and clonazepam and had recently been withdrawn from lithium, all of which have themselves been re­ported to cause bone marrow suppression (Barn­hart 1991; Gallicchio & Hulette 1989; Jefferson 1975), it seemed reasonable to conclude that the hypoplastic anaemia and moderate pancytopenia were not due to clozapine itself, but may have been caused by the combination of clozapine with the other medications. To evaluate this further, we studied the effect of carbamazepine alone and in combination with clozapine on the growth ofCFU­GM. From these studies (table IV) it is clear that carbamazepine is more toxic to bone marrow than clozapine, but less toxic than N-desmethyl cloza­pine. The combination is at least additive in its cytotoxic effect and this suggests that the combined use of these compounds may increase the potential for bone marrow suppression. For this reason, it seems prudent to advise against the combination of carbamazepine and clozapine in the treatment of psychiatric disorders. Furthermore, since this case was associated with withdrawal from lithium, it is unclear whether coadministration of lithium and clozapine can be advocated.

Table IV. Polypharmacy in drug-induced bone marrow suppressiona

Medium

10% patient serum Normal serum 10 I'mol/L hydrocortisone Clozapine 10 mg/L Clozapine 10 mg/L + 10% patient serum Carbamazepine 10 mg/L Carbamazepine 10 mg/L + 10% patient serum N-Desmethyl clozapine 5 mg/L N-Desmethyl clozapine 5 mg/L + carbamazepine 10 mg/L

CFU-GM/l05 bone marrow cells (% of control)

donor patients

90 110 80

healthy donors (n = 3)

104 ± 9 103

85 ± 6 107 ± 8 108 ± 10 123 ± 18 36 ± 8 23 ± 4

a Bone marrow cells were prepared as outlined in table III. Clozapine, carbamazepine and N-desmethyl clozapine were freshly prepared and added directly to cell suspensions in 0.8% methylcellulose. Colonies were counted after 14 days of incubation at 37°C.

Abbreviation: CFU-GM = colony forming unit-granulocyte-macrophage.

24

9. Conclusions

The aetiology of c1ozapine-induced agranulo­cytosis remains unclear. The delayed onset, pro­tracted duration, idiosyncratic clustering, and ab­sence of myeloid precursors in the bone marrow suggest that a haematopoietic stem cell is the target for c1ozapine's effect. We screened acute-phase serum from 5 patients with acute onset c1ozapine­induced agranulocytosis. No cytotoxic antibody di­rected against the myeloid haematopoietic precur­sor in the bone marrow was identified. Screening for c1ozapine- and c10zapine metabolite-related toxicity found that most compounds are relatively nontoxic to the bone marrow, especially at the con­centrations achieved systemically. In contrast, N­desmethyl c10zapine is much more toxic to mye­loid precursors, exhibiting toxicity at concentra­tions 3- to 5-fold higher than those normally seen in patients taking c1ozapine. Thus, it is possible that a c10zapine metabolite is responsible for c1ozapine­induced agranulocytosis. If so, this metabolite might accumulate at high levels in susceptible patients. Prospective screening of serum samples from patients receiving c1ozapine, undertaken to iden­tify patients who develop elevated levels of c10-zapine metabolites, might serve as a predictive test that could forestall the onset of c1ozapine-induced agranulocytosis.

Acknowledgement

Supported in part by grants MH47440, MH41684, P3OCA43703, arid GCRC MOIRR00080 from the Na­tional Institutes of Health and funds from Sandoz Re­search Institute and the Elisabeth Severance Prentiss and John Pascal Sawyer Foundations

The authors thank Cheryl Arce, Karin Johnson and Suzzanne Clavey for expert technical assistance and Lynne Lucas for preparation ofthe manuscript. Dr Peter Barnes, Sandoz Pharma Ltd, Basle, is thanked for his early in­terest in this work and Dr Paul Herrling of Sandoz Re­search Institute, Berne, Switzerland, is thanked for his kind donation of the c10zapine and c10zapine metabolites used in our studies.

References Ackenheil M. Clozapine-pharmacokinetic investigations and bio­

chemical effects in man. Psychopharmacology 99: S32-837, 1989 Anderman B. Griffith RW. Clozapine-induced agranulocytosis: a

Drug Safety 7 (Suppl. 1) 1992

situation report up to August 1976. European Journal of Clinical Pharmacology II: 199-20 I. 1977

Bagby Jr GC, Lawrence HJ. Neerhout Re. T-Iymphocyte-me­diated granulocyte failure. New England Journal of Medicine 309: 1073-1078, 1983

Baldessarini RJ. Frankenburg FR. Clozapine: a novel antipsy­chotic agent. New England Jounal of Medicine 324: 746-754, 1991

Barnhart ER. Physicians Desk Reference. 45th Ed., p. 1838, Medical Economics Co. Ltd .• Oradell, NJ, 1991

Bondesson U, Lindstrom LH. Determination ofclozapine and its N-demethylated metabolite in plasma by use of gas chroma­tography-mass spectrometry with single ion detection. Psycho­pharmacology 95: 472-475. 1988

Coulter DM, Edwards IR. Mianserin and agranulocytosis in New Zealand. Lancet 336: 785-787. 1990

Ereshefsky L, Watanabe MD. Tran-Johnson TK. Clozapine: an atypical antipsychotic agent. Clinical Pharmacy 8: 691-709, 1989

Gallicchio VS, Hulette Be. In vitro effect of lithium on carba­mazepine-induced inhibition of murine and human bone mar­row-derived granulocyte-macrophage. erythroid, and mega­karyocyte progenitor stem cells. Proceedings of the Society for Experimental Biology and Medicine 190: 109-116, 1989

Gerson SL, Lieberman JA. Friedenberg WR, Lee D, Marx Jr JJ, et al. Polypharmacy in fatal clozapine-associated agranUlocy­tosis. Lancet 338: 262, 1991

Gualde N, Malinvaud G. In vitro generation of cytotoxic lympho­cytes during noramidopyrine-induced agranulocytosis. Clinical Immunology and Immunopathology 24: 220-226. 1982

Haring C, Barnas C, Humpel C, Saria A, Aeischhacker WW. Dose­related plasma levels of clozapine. Journal of Clinical Psycho­pharmacology 9: 71-72, 1989

Idiinpiiiin-Heikkilii J, Alhava E, Olkinuora M, Paiva IP. Agran­ulocytosis during treatment with clozapine. European Journal of Clinical Pharmacology II: 193-198, 1977

Jefferson JW. Aplastic anemia associated with lithium. Lancet I: 413, 1975

Jiminez JJ, Arimura GK, Abou-Khalil WHo Isildar M. Yunis AA. Chloramphenicol-induced bone marrow injury: possible role of bacterial metabolites of chloramphenicol. Blood 70: 1180-1185, 1987

Kelton JG, Huang AT, Mold N, Logue G, Rosse WF. The use of in vitro techniques to study drug-induced pancytopenia. New England Journal of Medicine 301: 621-624,1979

Kramlinger KG, Post RM. Addition oflithium carbonate to carb­amazepine: hematological and thyroid effects. American Jour­nal of Psychiatry 147: 615-620, 1990

Krupp P, Barnes P. Leponex-associated granulocytopenia: a re­view of the situation. Psychopharmacology 99: S 118-S 121. 1989

Lazarus HM, Andersen J, Chen MG, Variakojis D, Mansour EG. et al. Recombinant granulocyte-macrophage colony stimulat­ing factor after autologous bone marrow transplantation for relapsed non-Hodgkin's lymphoma: blood and bone marrow progenitor growth studies. A phase II Eastern Cooperative On­cology Group Trial. Blood 78: 830-837, 1991

Lieberman JA, Kane JM, Johns CA. Clozapine: guidelines for clinical management. Journal of Clinical Psychiatry 50: 9-13. 1989a

Lieberman JA, Yunis J, Egea E, Canoso RT. Kane JM. et al. HLA-B38, DR4, DQw3 and clozapine-induced agranulocytosis in Jewish patients with schizophrenia. Archives of General Psychiatry 47: 945-948, 1989b

Mamus SW, Buron JK, Groat JD, Schulte DA. Lobell M, et al. Ibuprofen-associated pure white-cell aplasia. New England Journal of Medicine 314: 623-625, 1986

Murphy JF, Riordan T, Minchinton RM, Chapman JF. Amess JAL, et al. Demonstration of an immune-mediated mechan­ism of penicillin-induced neutropenia and thrombocytopenia. British Journal of Haematology 55: 155-160, 1983

Clozapine-Associated Agranulocytosis

Neftel KA, Muller MR, Walti M, Erni J, Gugler M, et al. Peni­cillin G degradation products inhibit in vitro granulopoiesis. British Journal of Haematology 54: 255-260, 1983

Perry PJ, Miller DO, Arndt SY, Cadoret RJ. Clozapine and nor­c10zapine plasma concentrations and clinical response of treat­ment-refractory schizophrenic patients. American Journal of Psychiatry 148: 231-235, 1991

Pisciotta A Y. Studies on agranulocytosis: a biochemical defect in chlorpromazine sensitive marrow cells. Journal of Laboratory and Clinical Medicine 78: 435-443, 1971

Pisciotta A Y. Immune and toxic mechanisms in drug-induced agranulocytosis. Seminars in Hematology 10: 279-310, 1973

Pisciotta A Y, Cronkite C. Aprindine-induced agranulocytosis. Evidence for immunology mechanism. Archives of Internal Medicine 143: 241-243, 1983

Pisciotta A Y, Miller RC, Ban S, Cronkite C. Inhibition of fibro­blast proliferation by myelotoxic drugs. Annals of the New York Academy of Sciences 459: 198-210, 1985

25

Simpson GM, Cooper TA. Clozapine plasma levels and convul­sions. American Journal of Psychiatry 135: 99-100, 1978

Smith CS, Chinn S, Watts RWE. The sensitivity of human bone marrow granulocyte/monocyte precursor cells to phenylbuta­zone, oxyphenbutazone and gamma-hydroxyphenylbutazone in vitro. with observations on the bone marrow colony for­mation in phenylbutazone-induced granulocytopenia. Bio­chemical Pharmacology 26: 847-852, 1977

Yunis AA. Comparative toxicity of chloramphenicol and thiam­phenicol with particular reference to aplastic anemia. Chem­ioterapie and Antimicrobials 4: 52-61, 1980

Correspondence and reprints: Dr Stanton L. Gerson. Division of Hematology/Oncology, 2074 Abbington Road, Cleveland, OH 44106, USA.