Blood Dyscrasias Induced by Psychotropic Drugs Original Paper S70
S. Stübner1 R. Grohmann1 R. Engel1 B. Bandelow2 W.-D. Ludwig3 G. Wagner1 B. Müller-Oerlinghausen4 H.-J. Möller1 H. Hippius1 E. Rüther2
Drugs can cause a variety of blood dyscrasias, e. g., by interfering with hematopoiesis in the bone marrow or damaging mature blood cells by antibodies. Although numerous reports on the risks of adverse hematological effects associated with psychotropic drugs have led to stringent monitoring requirements for some compounds, particularly neuroleptics, it is still difficult to estimate the true prevalence of such risks. Sixteen episodes of thrombocytopenia, 63 of neutropenia, 22 of agranulocytosis, 4 episodes of severe neutro- and thrombocytopenia, and 2 of pancytopenia were documented by the drug safety program in psychiatry AMSP (Arzneimittelsicherheit in der Psychiatrie) in a population of 122,562 patients between 1993 and 2000. All cases were related to the epidemiological data provided for this population and systematically analyzed as regards history of medication, co-medication, and the clinical course. Putative risk rates for the main groups of medications and a number of drugs could be estimated with this database.
Most changes in the white blood cell counts, which were rated as probably or definitely drug-induced, were attributed to clozapine (0.18 % of patients exposed), carbamazepine (0.14 %) and perazine (0.09 %). In patients on newer atypical neuroleptics, we documented neutropenia assumed to be probably or definitely drug-related in five patients during treatment with olanzapine and in one case with risperidone. In all five olanzapine-related cases, the drugs were the sole cause of the adverse drug reactions. All surveyed patients who received clozapine showed no difference in age and gender distribution from those who developed hematological changes. Incidences of hematological changes for antidepressants were much lower (about 0.01 %). Although the methodological accuracy of these findings has to be critically discussed these data could be of considerable clinical relevance and should be helpful in making clinical treatment decisions.
Introduction
An estimate of the background incidence of hematopoietic disorders in the general population revealed that agranulocytosis accounts for 7.2 cases/million patients/year and neutropenia for 35.7 cases/million patients/year [39].
Drug-induced blood dyscrasias may be due to toxicity of the drug, inborn errors of metabolism or immunologic mechanisms. Depending on the site of the hematopoietic damage and its underlying pathomechanisms, one, two or all three major cell lines of hematopoiesis (i. e., erythrocytic, myeloid, and megakaryocytic) can be involved.
Drug-induced agranulocytosis and aplastic anemia are extremely rare, with an incidence of only a few cases per million population per year [23], whereas drug-induced thrombocytopenia and especially drug-induced neutropenia occur more frequently
Affiliation 1 Department of Psychiatry, Ludwig-Maximilians University, Munich, Germany 2 Department of Psychiatry and Psychotherapy, Georg-August University, Göttingen, Germany 3 HELIOS Klinikum Berlin, Department of Hematology, Oncology, and Tumor Immunology, CharitØ (Campus Berlin-Buch), Humboldt-University of Berlin, Germany 4 Drug Commission of the German Medical Association, Berlin/Cologne, Germany Correspondence Dr. med. Susanne Stübner ´ Psychiatrische Klinik und Poliklinik ´ Abteilung für Forensische Psychiatrie ´ Nussbaumstraûe 7 ´ 80336 München ´ Germany ´ Phone: +49-89-5160-2736 ´ Fax: +49-89-5160-3389 ´ E-Mail:
[email protected] Bibliography Pharmacopsychiatry 2004; 37 Suppl 1: S70±S78 ´ Georg Thieme Verlag Stuttgart ´ New York ´ ISSN 0936-9528 ´ DOI 10.1055/s-2004-815513
[20, 35]. In contrast to acute agranulocytosis, which is mostly due to drugs (62 ± 90 % of all cases), drug exposure is a rather rare cause of thrombocytopenia (5 ± 20 % of all cases) and hemolytic anemia (< 10 % of all cases) [20]. Bone marrow suppression is probably the most frequent mechanism of drug-induced blood dyscrasias [20]. Known or suspected causes include antibacterial or anticonvulsive drugs, antihistamines, antimalarial, antirheumatic, antithyroid, cardiotropic, or cytostatic drugs, diuretics, psychotropic drugs, and sulphonylureas [22]. In the case of combination therapies, synergisms must be considered as a potential cause of adverse drug reactions (ADR).
Drug-induced blood dyscrasia is of great interest in clinical psychiatry. Newer and ªatypicalº neuroleptic drugs have been developed mainly to minimize other adverse drug reactions (ADRs) such as extrapyramidal motoric symptoms (EPMS). However, clozapine, the first atypical drug, is known to cause an annual incidence of 8000 cases of agranulocytosis/million patients [2]. Thus, during treatment with clozapine, blood cell counts controls must be strictly monitored, and the patient's informed written consent is required before starting treatment. AMÜP (Arzneimittelüberwachung in der Psychiatrie) [17] conducted an intensive drug surveillance program in psychiatric inpatients that documented all adverse reactions to all psychotropic drugs, even if they were very common or discrete, during a limited period of time. To our knowledge this investigation is the only published prospective study in this area. Observations of 13.000 inpatients treated with antipsychotics between 1979 and 1988 were reported. Seven cases of agranulocytosis were detected, one while on clozapine and six while on perazine (three in combinations with other drugs) [19]. So far no risk rates for blood dyscrasia with the newer atypical neuroleptics have been published. Much less observations exist on the hematological risk profile of other groups of psychotropic changes, such as antidepressants [18]. In this paper we report on the hematological changes observed with psychotropic drugs in psychiatric inpatients surveyed by the AMSP study.
Currently 35 psychiatric institutions participate in the drug safety program in psychiatry AMSP for assessing severe ADRs [18]. Details of the methods used in AMSP and its history are given by [36]. In brief, in the naturalistic setting of routine psychiatric inpatient treatment all marketed psychotropic drugs are under surveillance. However, only clinically severe ADRs are assessed according to the criteria of the project [36], and comprise psychiatric, neurological or other medical conditions occurring as ADRs. AMSP is supported by many German pharmaceutical manufacturers involved in CNS research. However, the participating hospitals also contribute largely as the drug monitors serve this function in addition to their usual routine work. The individual events are rated and classified according to the probability of a causal relationship between the application of a drug and the observed adverse event. There are five degrees: no (0), possible (1), probable (2) or definite (3) relationship assumed, or not assessable (4). A possible relation is assumed if an ADR has not yet been described with the drug, the time course does not show a clear correlation between onset or increase of the dosage of a drug and the development of the adverse event, or if an alternative explanation is more likely. A relation is assessed as probable if an ADR is known for the drug(s) in question, the time sequence agrees with previous knowledge of the drug, and there is no plausible alternative explanation. A relationship is considered definite, if a rechallenge leads to the reappearance of the ADR. Polypharmacy is common in psychiatric inpatients, and combinations of drugs with known potential to cause hematologic ADR are often administered. In such cases, the drugs used in combination therapy are considered as common offenders if the time sequence does not allow a decision on the causative agent or if additive effects are assumed. Dosage, time sequence, and potential risk are taken into account for each drug separately, which may lead to different degrees of probability for the drugs in question in each case. Data on the drug utilization in the participating centers is derived from reference days (twice per year and hospital). The participating centers also provide information on the number of inpatients monitored per year as well as the mean duration of inpatient treatment for all monitored patients. Both are broken down according to diagnostic groups. The following definitions of severe drug-induced hematologic reactions are based on common hematological standards: thrombocytopenia: platelet counts < 100.00/ml; neutropenia: neutrophils < 1.500/ml, (agranulocytosis: < 0.500/ml); anemia: erythrocytes < 4 *106 /ml and/or hemoglobin < 12 g/dl. Pancytopenia is defined as neutropenia and thrombocytopenia in combination with anemia.
Stübner S et al. Blood Dyscrasias Induced ¼ Pharmacopsychiatry 2004; 37 Suppl 1: S70 ± S78
Original Paper
Different pathomechanisms may play a role in drug-induced selective neutropenia [34], including the sudden destruction of neutrophils by the formation of drug-dependent antineutrophil antibodies, development of a lupus-like syndrome, and toxic suppression of hematopoietic precursors by prolonged administration of a drug. The latter has often been observed after the use of phenothiazines [33]. As the effect of phenothiazines is cumulative, bone marrow depression is likely to occur in psychiatric patients on maintenance therapy. Agranulocytosis due to treatment with phenothiazines was reported already in the fifties [21]. Clinical symptoms can be observed 20 to 40 days after the onset of treatment, and bone marrow investigations disclose marked marrow hypocellularity.
Methods
S71
Leukocytes, neutrophils, and platelets are counted in newly admitted patients by all participating hospitals. Additionally, controls of the white blood cell counts are performed at least monthly during psychopharmacological treatment in about 75 % of the participating hospitals, and more frequently ± every 2 weeks ± in about 25 %. In a few hospitals the controls are even more frequent. Controls are made weekly during treatment with clozapine.
Original Paper
Data are presented as absolute numbers of ADRs and as relative frequencies in the patients exposed. The number of patients exposed to a drug (or drug group) is predicted by data collected at two reference days per year from every hospital under surveillance and by the number of patients per year. ADR rates are presented together with their 95 % confidence intervals. In view of the very low rate of serious side effects and the large number of patients exposed, confidence intervals are calculated according to the exact (asymmetric) method first described by Clopper and Pearson [7], avoiding the bias of the commonly used approximate methods (for a discussion of methods see e. g. [1]. Statistical interference is used as an exploratory technique, because the focus of the AMSP project is to provide descriptive ADR data, not to prove hypotheses.
Results A total of 122,562 patients, treated with psychotropic drugs between 1993 and 2000 were surveyed by the AMSP program. 107 cases of hematological changes were documented resulting in an incidence of 0.0873 %. S72
Only four of these cases were rated as definitely drug-related on the basis of a positive rechallenge (two cases of neutropenia, one of agranulocytosis, and one of pancytopenia), 75 were rated as probably drug-related, and 28 cases were rated as only ªpossiblyº drug-related. Symptomatology of the observed severe hematological ADRs During this period of surveillance, 16 episodes of thrombocytopenia, 63 episodes of neutropenia, 22 episodes of agranulocytosis, 4 episodes of severe neutro- and thrombocytopenia, 2 episodes of pancytopenia, and no haemolytic anemia were documented. Associated clinical signs were observed in 18 cases (see Table 1). All other events were detected by routine blood counts. Clinical symptoms were present in three of the patients with thrombocytopenia (hemorrhage); six of those with neutropenia and seven of those with agranulocytosis had fever and/or infections. Eight of the patients with agranulocytosis were transferred to a department of internal medicine and seven were treated with granulocyte-colony stimulating factor (G-CSF). The imputed drugs were discontinued in 104 cases, in the remaining three their dosage was reduced. The causal relationship between psychotropic drug treatment and ADR was rated as probable or definite in 9 episodes of thrombocytopenia, 42 of neutropenia, 21 episodes of agranulocytosis and one of pancytopenia.
Table 1
Hematological changes due to psychotropic drugs and presence of clinical signs. A causal relationship between treatment and ADR was assessed as possible, probable or definite. G-CSF: granulocyte-colony stimulating factor. ADR: adverse drug reaction
Type of ADR
Number of documented episodes
Clinical signs
(all probabilities)
per type of ADR
Thrombocytopenia
16
3
Neutropenia
63
6
Agranulocytosis
22
7 - internal hospitalization in 8 cases; administration of G-CSF in 7 cases
Pancytopenia
2
1
Neutropenia and thrombopenia
4
1
107
18
All ADR
Severe ADRs of the white blood cells (WBC) Table 2 gives an overview of the 91 documented episodes of severe disorders affecting the WBCs including agranulocytosis, neutropenia, neutro- and thrombocytopenia, and pancytopenia. In the following, all these are summarized as ªsevere neutropeniaº. The ADRs are related to different groups and subgroups of psychotropic drugs. Sometimes only one pathological WBC count was documented; sometimes the pathological value was the first one after admission. In all these cases (n = 27; 30 % of all) only a possible relationship between ADR and treatment was assumed, because of a potential error in the hematological laboratory or the possibility of another explanation such as a physical illness. The highest rates of ªsevere neutropeniasº were observed during treatment with anticonvulsants. The incidence with anticonvulsants alone was 0.08 %, and higher than with all (typical and atypical) neuroleptics (0.05 %). In comparison, the incidence of hematologic ADR in patients treated with antidepressants was low (0.01 %). Among the typical neuroleptic drugs phenothiazines caused the highest incidence (0.02 %). Fig. 1 shows the ADR rates for the different drug groups. The incidences for single drugs are given in Table 3, which lists only cases for which a causal relationship between treatment and ADR was rated as probable or definite. For rarely used drugs, incidence rates may be an artifact and are not really comparable with that of drugs used more often. Therefore, we listed only those substances that were administered in more than 4.000 patients. Clozapine rated highest among the single substances (0.16 %, only clozapine imputed); it was followed by carbamazepine (0.11 %), perazine (0.07 %) and olanzapine (0.05 %). All other drugs caused severe hematological ADRs in less than 0.04 % (all rates for drugs imputed alone). For cases rated as probably or definitely drug-induced and in which a single substance was imputed, there was s significant statistical difference on the 5 %-niveau (Fisher's exact test) between clozapine and olanzapine (p = 0.022).
In all cases there was full recovery. Stübner S et al. Blood Dyscrasias Induced ¼ Pharmacopsychiatry 2004; 37 Suppl 1: S70 ± S78
Table 2
Frequency of ªsevere neutropeniaº for different psychotropic drug groups and subgroups; all cases, and only cases rated as probably or definitely drug-related. All: drug or drug group was imputed alone or in combination with other drugs as probable or definite; imputed alone: only this drug or drug group was imputed with a probable or definite rating (that is, other drugs may be imputed as well but only with a possible rating). *estimated from data on reference days. ** Mirtazapine, Mianserine, Maprotiline, Venlafaxine. Exposed patients*
Episodes with the drug all probabilities
Incidence
probably or definitely drug related
All
All
Imputed alone
All
Imputed alone
N
N
n
n
(%)
(%)
122,562
91
72
Anticonvulsants
18,857
21
21
16
0.11
0.08
Antidepressants
53,041
19
8
5
0.02
0.01
Neuroleptics
86,439
66
43
39
0.05
0.05
Butyrophenones
30,537
6
1
1
0.003
0.003
Phenothiazines
28,350
16
8
7
0.03
0.02
Atypical Neuroleptics without Clozapine
21,728
12
7
7
0.03
0.03
Clozapine
15,414
29
27
25
0.18
0.16
Thioxanthenes
15,020
7
1
0
0.01
0.00
Tricyclic Antidepressants
25,764
12
6
4
0.02
0.02
Selective Serotonine Reuptake Inhibitors
16,698
4
0
0
0.00
0.00
Other Antidepressants**
11,121
4
2
1
0.02
0.01
Benzodiazepines
33,826
1
1
1
0.003
0.003
Patients with psychopharmacological treatment
Frequency of ªsevere neutropeniaº associated with different single drugs; a relation between the drug application and the adverse event was rated as probable or definite. *estimated from reference day data.
Clozapine
Fig. 2 gives the incidences of ªsevere neutropeniaº with their exact confidence intervals for all drugs applied to more than 4000 patients. There was a clear peak of ªsevere neutropeniasº between the 20th and 30th day after onset of treatment (Fig. 3). The time course of the clozapine-induced changes, however, showed another pattern: the main peak appeared later, between the 40th and 50th day of treatment (Fig. 4).
Exposed episodes with the patients* drug, imputed at all alone n n
incidence imputed at all alone (%) (%)
15,414
27
25
0.18
0.16
Olanzapine
9,231
5
5
0.05
0.05
Risperidone
7,495
1
1
0.01
0.01
16,293
1
1
0.01
0.01
Levomepromazine
5,714
0
0
0.00
0.00
Perazine
9,391
8
7
0.09
0.07
Promethazine
6,257
0
0
0.00
0.00
Chlorprothixene
6,291
1
0
0.02
0.00
Amitriptyline
7,044
1
1
0.01
0.01
Doxepine
7,324
3
2
0.04
0.03
Imipramine
7,324
1
0
0.01
0.00
Trimipramine
5,943
1
1
0.02
0.02
Lorazepam
18,885
1
1
0.01
0.01
Carbamazepine
13,525
19
15
0.14
0.11
4,333
2
1
0.05
0.02
Haloloperidol
Fig. 1 Incidences of ªsevere neutropeniaº (i. e. agranulocytosis, neutropenia, neutropenia, thrombocytopenia, and pancytopenia) after administration of psychotropic drug groups, imputed alone or in combination; a causal relationship was assumed as probable or definite.
Original Paper
Table 3
0.06
Valproate
As regards potential risk factors, the distribution of gender was nearly the same in the group of all surveyed patients
Stübner S et al. Blood Dyscrasias Induced ¼ Pharmacopsychiatry 2004; 37 Suppl 1: S70 ± S78
S73
Fig. 2 Incidences of ªsevere neutropeniaº due to single drugs (imputed alone) with 95 % confidence intervals.
Original Paper Fig. 3 Time (in days) of treatment (without clozapine) until onset of ADR within 3 months (57 cases; 9 additional cases were documented at a time beyond this period).
S74
(n = 122,562) compared to those with drug-induced ªsevere neutropeniasº (n = 72), as well as the distribution among all clozapine treated patients (n = 15,414) compared to those with ADR due to clozapine treatment (n = 27) (Fig. 5). The age of the affected patients on clozapine (n = 27) showed a similar distribution to that of all patients having received clozapine (n = 15,414) (Fig. 6). Case report ± Neutropenia (see Fig. 7) A 62-year-old woman was diagnosed to have a moderate depressive episode in a recurrent major depressive disorder. She also had a chronic sinusitis, sleep apnoe syndrome, and vertebral arthrosis. She had been treated with budesonide, zolpidem and pipamperone for several weeks. After her hospitalization, venlafaxine and (briefly) indomethacine had been added. About 8 weeks after onset of venlafaxine, the dosage of which had reached 225 mg, neutrophil counts of 1.400/ml were meas-
ured. The bone marrow biopsy disclosed hypocellularity of hematopoiesis, which was interpreted as drug-induced toxic damage. Since the discontinuation of indomethacine did not lead to any improvement, venlafaxine was suspected for having possibly caused the hematologic changes. After its discontinuation, leucocyte and neutrophil counts completely recovered. To our knowledge venlafaxine has not yet been reported to cause hematological changes. However, in this case the time course of the events strongly suggests a causal role for venlafaxine. Case report 2 ± Neutropenia (see Fig. 8) A 40-year-old patient with a history of paranoid schizophrenia was hospitalized and treated with a combination of perazine and lorazepam. As the patient also had an anemia due to iron deficiency, she also had to take ferrosulfate. After approximately 1 week of treatment, neutrophil counts were slightly reduced compared to those before treatment (2,900/ml versus 3,300/ml). At that time,
Stübner S et al. Blood Dyscrasias Induced ¼ Pharmacopsychiatry 2004; 37 Suppl 1: S70 ± S78
Fig. 4 Time (in days) of treatment with clozapine until onset of ADR within 3 months (27 cases; 7 additional cases were documented at a time beyond this period).
Original Paper
haloperidol was added. Another week later cotrimoxazol was given in order to treat cystitis. Seven days later, blood cell counts showed a neutropenia of 1,400/ml. Perazine was promptly discontinued, and 3 days later, the antibiotic treatment course was finished. Neutrophil cell counts completely recovered. In this case, perazine was imputed to be the probable causative agent, an additional effect of antibiotics on the severity of the neutropenia was rated as only possible, because neutrophils had begun to increase again the very first day after the last dose of cotrimoxazol. This case report characterizes a typical clinical situation. Patients often are treated with combinations of psychopharmacological drugs and in addition a medication for accompanying internal diseases. Thus, polypharmacy arises which may lead to additive effects, interactions, altered metabolism, higher blood concentrations of the single substances, and increased ADR risk.
Fig. 5 Distribution of gender in the group of all surveyed patients, patients with drug-induced neutropenia or agranulocytosis, clozapinetreated patients, and patients with ADR after clozapine treatment ± only cases rated as probably or definitely drug-induced are included.
Pancytopenia Pancytopenia was documented in two cases. In one case both carbamazepine and lovastatine were rated as probable causative agents in combination, because the time course did not allow to distinguish between both drugs as being potentially responsible for pancytopenia. In the second case, pancytopenia was observed during a treatment with trimipramine. Thrombocytopenia Thrombocytopenia was observed in 20 patients (in four of these neutropenia was present simultaneously) and considered to be probably caused by psychotropic drugs in ten cases. In five of these episodes carbamazepine as combined treatment was involved (incidence 0.03 %); other drugs were imputed as well. In one case only valproate was judged to be the cause; in the other four cases, perazine (twice), dibenzepine and risperidone were involved. Case report 3 ± Thrombocytopenia (see Fig. 9) A 56-year-old woman with a schizoaffective psychosis was treated with atenolol and chlorthalidone for hypertension and clomi-
Fig. 6 Age distribution of the group of patients with drug-induced neutropenia or agranulocytosis after clozapine treatment and all clozapine-treated patients ± only cases rated as probably or definitely druginduced are included.
Stübner S et al. Blood Dyscrasias Induced ¼ Pharmacopsychiatry 2004; 37 Suppl 1: S70 ± S78
S75
Original Paper
Fig. 7 Course of hematological changes in a 62-year-old patient receiving psychotropic and medical treatment for internal diseases. Drugs not having caused the ADR are represented in light gray, those assumed to possibly cause the ADR in a darker gray.
pramine because of a depressive syndrome over several weeks. Fourteen days after carbamazepine was added, she developed vertigo, nausea, and an exanthema with little spots. Platelet counts decreased to a nadir of 42.000/ml. An internal consultation was made, and the drugs thought to cause the hematological changes were discontinued: carbamazepine probably, and atenolol, chlorthalidone, and clomipramine possibly caused the thrombocytopenia. As a consequence, lormetazepam had to be added, and lithium was applied in order to stabilize the mood. This treatment led to a rapid improvement of platelet counts and the clinical signs. S76
This case report emphasizes that it is important to monitor not only the WBC but also the platelet counts for potential ADRs.
Fig. 9 Course of hematological changes in a 56-year-old patient receiving psychotropic and medical treatment for internal diseases. Drugs not assumed to cause the ADR are represented in light gray: those assumed to possibly cause the ADR in a darker gray; and those assumed to probably cause the ADR in black.
107 severe hematological changes occurred during psychopharmacological treatment and mostly affected the white blood cells (in 85 %). The findings suggest that severe hematologic ADRs are generally rare events. No fatal ADR was documented. Furthermore, it could be shown that discontinuation of the offending drugs generally leads to fast improvement of blood cell counts. There are differences between drug groups and single substances. ªSevere neutropeniaº occurred with an incidence of 0.1 % during anticonvulsive treatment and 0.05 % with neuroleptic treatment. Clozapine rated highest among the single substances (0.18 %), followed by carbamazepine (0.14 %). In contrast, severe neutropenia during treatment with antidepressants was very rare (0.02 %).
Discussion No single case due to treatment with SSRI was documented. Our current investigation has extended over 8 years, and more than 120,000 patients have been under surveillance. A total of
To determine the probability of single substances for inducing hematological changes, the frequencies of administration have to be taken into account. As some substances are not given very often, the extrapolations of events per 100,000 people may lead to artifacts and false high-risk assessments. Clozapine is commonly assumed to have a cumulative incidence of 0.8 %/year for agranulocytosis [2]. During the average treatment in AMSP from 1993 to 2000 of about 50 days for schizophrenic patients and about 35 days for all patients, only one segment of the treatment period was monitored. This was not necessarily the high-risk treatment period (week 4 ± 8). Thus, it is to be expected that a lower rate than Alvir's cumulative incidence is assessed with a system like AMSP.
Fig. 8 Course of hematological changes in a 40-year-old patient receiving psychotropic treatment. Drugs not assumed to cause the ADR are represented in light gray, those assumed to possibly cause the ADR in a darker gray; and those assumed to probably cause the ADR in black.
It has to be considered that our data may include systematic errors due to the fact that changes in blood cell counts often do not lead to any clinical signs, and due to the recording methods. Some cases may fail to be identified, because they are only transitory, and not all cases observed by the treating clinicians may
Stübner S et al. Blood Dyscrasias Induced ¼ Pharmacopsychiatry 2004; 37 Suppl 1: S70 ± S78
be communicated to the drug monitors. As the changes do not lead to clinical symptoms in each case, some alterations are only noticed in routine blood cell counts. If they are not performed, changes will probably not attract attention, as long as they do not progress to very severe ADRs.
No other comprehensive drug surveillance studies with a methodology comparable to that of AMSP have been performed in psychiatry. Although hematological ADRs are of great clinical relevance, there are still few epidemiological data available, especially those comparing several drugs. King and Wager reported on hematological disorders caused by 16 antipsychotic drugs under surveillance by the UK monitoring system (Medicines Control Agency) [24]. They provided collections of spontaneous ADR reports between 1963 and 1996 and corrected those for relative use based on prescription rates in N. Ireland in 1995. A ranking based on the prescription rates indicated that the highest risk rates occurred with clozapine. They concluded that higher risks of hemopoietic reactions seemed to be associated with the aliphatic phenothiazine derivates thioridazine and chlorpromazine, and that there was no evidence of any increased risk with high-potency drugs such as haloperidol or pimozide or with newer drugs like sulpiride or risperidone. It has been suggested that women may be at slightly higher risk of developing agranulocytosis due to clozapine than men [2, 25]. In our data the distribution of gender was nearly the same in the group of all surveyed patients compared to those with severe neutropenias as well as the distribution among all clozapine treated patients compared to those with severe neutropenia due to clozapine treatment; only a slight tendency could be observed. In the meantime several reports on hematological disorders due to treatment with newer atypical drugs have been published. For example, the mechanisms of atypical antipsychotic-induced neutropenia have been investigated in dogs. DMP 406, a clozapine-like drug, was observed to induce changes that seemed to result from both immune-mediated and direct drug-induced bone marrow suppression [27]. Many case reports on hematological changes due to atypical drugs have been published in the last decade. Some focused on risperidone: in one case agranulocytosis was observed following risperidone treatment in a 40-year old woman, who had already developed agranulocytosis during classical neuroleptic treatment [13]. Thrombopenia and lymphopenia during treatment with risperidone and persistent after a switch to clozapine were
The hematological safety of olanzapine was questioned after some cases of granulocytopenia due to clozapine though persisting after a switch to olanzapine were reported [4,15, 26]. By now, several cases of olanzapine-induced neutropenia or leukopenia have been reported by AMSP (38) as well as by other authors [6, 9, 29, 31, 41, 42]. Some authors claimed the hematological safety of olanzapine based on case reports about successful treatment [14, 40] even in patients who had developed blood dyscrasias previously on clozapine and risperidone [11] or on clozapine alone [32]. From Canada six cases of neutropenia associated with olanzapine were reported up to April, 1998; olanzapine has been approved and marketed in Canada in July 1996. The authors claimed that this compound is not associated with a particular risk of agranulocytosis [43]. The findings of the present study suggest that olanzapine possesses a certain hematological risk, but that it is significantly lower than that of clozapine. Recently, interesting results on the plasma levels of granulocyte colony-stimulating factor (G-CSF) during clozapine- and olanzapine-induced granulocytopenia were published. They showed a significant positive correlation: G-CSF was not detected when granulocytopenia occurred. The authors concluded that both clozapine and olanzapine are able to induce transient granulocytopenia via a similar or common mechanism that is not associated with a compensatory increase in G-CSF levels [37]. HLA-genotyping in patients with clozapine-induced agranulocytosis revealed a significant association between HLA haplotype DQB*0502 and agranulocytosis [8,10]. Interestingly, one patient who had developed neutropenia during perazine and later on during olanzapine treatment [9] was also positive for HLADQB*05. Though such an association could not be demonstrated in the patient with olanzapine-induced agranulocytosis described by the group from Dresden [31], this case and the two cases published by Dettling et al. [9] converge in that all of them are carriers of haplotype DQB1*06. The two of these three cases with confirmed agranulocytosis additionally shared haplotype DQB1*0302 [12]. As many groups of drugs with different psychiatric or somatic indications are known to cause hematological changes, combined treatment, which is very common in clinical daily life, may realistically pose an increased risk. Some of the risk-promoting mechanisms in patients on multiple drugs are as follows: (1) multimorbid or elderly patients often
Stübner S et al. Blood Dyscrasias Induced ¼ Pharmacopsychiatry 2004; 37 Suppl 1: S70 ± S78
Original Paper
However, compared with other data collected as a kind of reference, our results seem quite exact. The previous drug safety project, AMÜP (Arzneimittelüberwachung in der Psychiatrie), which documented all adverse reactions, yielded an incidence of 0.13 % for severe neutropenia associated with perazine and of 0.10 % with clozapine [17]. WBC counts were performed weekly. From the AMSP data an incidence of 0.18 % for clozapine and of 0.09 % for perazine was calculated. The small number of patients on clozapine within the context of AMÜP (n = 967) probably led to a less reliable rate in AMÜP. This correspondence of the actual data to the reference data evaluated previously speaks against a systematic error; in fact, the data may reflect adequately the clinical situation.
reported once [3]. There was also a report of agranulocytosis after adding risperidone to clozapine treatment [16]. A case of reversible neutropenia during a cold was reported to possibly involve risperidone [30], but also a negative rechallenge in this patient was documented [5]. Two cases of treatment with risperidone were reported to have a subsequent good therapeutic response after blood dyscrasias due to classic antipsychotics, suggesting risperidone to be a safe alternative [28]. The review of the relevant literature shows that so far an alternative cause had to be taken into account in almost all cases of hematological ADR connected with risperidone. Our data documented only one case connected with risperidone. Thus, risperidone seems to be quite safe as regards the development of severe hematological changes.
S77
have a lower metabolism rate, (2) drugs may inhibit enzymatic processes and thus metabolic substances may accumulate easier, (3) distressed organisms are in general more sensitive and susceptible.
Original Paper
In conclusion, it seems wise to regularly monitor blood cell counts during treatment with neuroleptics and anticonvulsants and to be aware of the potential hazard of combinations of several drugs with known risk of hematotoxicity. In addition to leukocyte counts the proportion of neutrophils should be checked. Special caution is advisable when compounds introduced only recently to the market are prescribed. Any adverse hematological reaction, no matter whether it occurs with a traditional or new drug, should be reported to either AMSP or directly to national voluntary reporting systems as to the Drug Commission of the German Medical Association. Furthermore, when ever possible, plasma or serum of such patients should be stored in order to facilitate future assessments of metabolites or antibodies against the imputed drugs.
References 1
S78
Agresti A, Coull BA. Approximate is better than ªexactº for interval estimation of binominal proportions. The American Statistician 1998; 52: 119 ± 126 2 Alvir JMJ, Liebermann JA, Saffermann AZ, Schwimmer JL, Schaaf JA. Clozapine-induced agranulocytosis: increased and risk factors in the United States. New Engl J Med 1993; 329: 162 ± 167 3 Assion HJ, Kolbinger HM, Rao ML, Laux G. Lymphocytopenia and thrombocytopenia during treatment with risperidone or clozapine. Pharmacopsychiatry 1996; 29: 227 ± 228 4 Benedetti F, Cavallaro R, Smeraldi E. Olanzapine-induced neutropenia after clozapine-induced neutropenia. Lancet 1999; 354: 567 5 Bondolfi G, Morena P, Dascal D-R, Bertschy G. Risperidone and reversible neutropenia: a negative rechallenge. Europ Neuropsychopharmacol 1996; 6: 257 6 Buchman N, Strous RD, Ulman A-M, Lerner M, Kotler M. Olanzapineinduced leukopenia with human leukocyte antigen profiling. Int Clin Psychopharmacol 2001; 16: 55 ± 57 7 Clopper CJ, Pearson ES. The use of confidence of fiducial limits illustrated in the case of the binominal. Biometrika 1934; 26: 404 ± 413 8 Dettling M, Cascorbi I, Roots I, Mueller-Oerlinghausen B. Genetic Determinants of Clozapine-Induced Agranulocytosis: Recent Results of HLA Subtyping in a Non-Jewish Caucasian Sample. Arch Gen Psychiatry 2001; 58: 93 9 Dettling M, Cascorbi R, Hellweg R, Deicke U, Weise L, Müller-Oerlinghausen B. Genetic Determinants of Drug-Induced Agranulocytosis: Potential Risk of Olanzapine?.Pharmacopsychiatry 1999; 32: 110 ± 112 10 Dettling M, Schaub RT, Mueller-Oerlinghausen B, Roots I, Cascorbi I. Further evidence of human leukocyte antigen-encoded susceptibility to clozapine-induced agranulocytosis independent of ancestry. Pharmacogenetics 2000; 10: 1 ± 7 11 Dvernovsek M, Tavcar R. Olanzapine appears haematologically safe in patients who developed blood dyscrasia on clozapine and risperidone. Int Clin Psychopharmacol 2000; 15: 237 ± 238 12 Felber W, Naumann R, Schuler Uet al. Are There Genetic Determinants of Olanzapine-Induced Agranulocytosis?. Pharmacopsychiatry 2000; 33: 197 ± 198 13 Finkel B, Lerner AG, Oyfee I, Sigal M. Risperidone-associated agranulocytosis. Am J Psychiatry 1998; 155(6): 855 ± 856 14 Finkel B, Lerner AG, Oyffe I, Rudinski D, Sigal M, Weizman A. Olanzapine treatment in patients with typical and atypical neuroleptic-associated agranulocytosis. Int Clin Psychopharmacol 1998; 13: 133 ± 135 15 Flynn S, Altmann S, Macewan GW, Black LL, Greenidge LL, Horner WG. Prolongation of clozapine.induced graulocytopenia associated with olanzapine. J Clin Psychopharmacol 1997; 17: 494 ± 495
16
Godleski LI, Sernyak M. Agranulocytosis afterAddition of Risperidone to Clozapine Treatment. Am J Psychiatry 1996; 153(5): 735 ± 736 17 Grohmann R, Hippius H, Helmchen H, Rüther E, Schmidt L. The AMÜP study for drug surveillance in psychiatry - a summary of inpatient data. Pharmacopsychiatry 2004; 37 Suppl 1: S16 ± 26 18 Grohmann R, Rüther E, Engel R, Hippius H. Assessment of adverse reactions with the AMSP drug safety program: Methods and first results on tricyclic antidepressants and SSRI. Pharmacopsychiatry 1999; 31: 1 ± 8 19 Grohmann R, Schmidt L, Spiess-Kiefer C, Rüther E. Agranulocytosis and significant luecopenia with neuroleptic drugs: results from the AMÜP program. Psychopharmacology 1989; 99: (109 ± 112) 20 Heimpel H. When should the clinician suspect a drug-induced blood dyscrasia, and how should he proceed?. Eur J Haematol 1996; 60(suppl): 11 ± 15 21 Hippius H, Kanig K. Die Agranulocytose unter neuro-psychiatrischer Phenothiazin-Therapie. ¾rztliche Wochenschrift 1958; 13(23): 501 ± 507 22 Horler AR. Blood disorders. In: Davies DM, ed. Textbook of adverse Drug Reactions,; 3. ed Oxford: Oxford University Press, 1985: 578± 594 23 Kaufman DW, Kelly JP, Jurgelon J et al. Drugs in the aetiology of agranulocytosis and aplastic anaemia. Eur-J-Haematol 1996; 60 Suppl: 23± 30 24 King DJ, Wager E. Haematological safety of antipsychotic drugs. J Psychopharmacol 1998; 12(3): 283 ± 288 25 Krupp P, Barnes P. Clozapine-induced agranulocytosis: risk and aetiology. British Journal of Psychiatry 1992; 160: 38 ± 40 26 Lambert T. Olanzapine after clozapine: the rare case of prolongation if graulocytopaenia. Aust NZ J Psychiatry 1998; 32: 591 ± 592 27 Lorenz M, Evering WE, Provencher A et al. Atypical antipsychotic-induced neutropenia in dogs. Toxicology and Applied Pharmacology 1998; 155: 227 ± 236 28 Mahmood T, Silverstone T, Spittle B. Risperidone appears safe in patients with antipsychotic-induced blood dyscrasias. Int Clin Psychopharmacol 1996; 11: 53 ± 54 29 Meissner W, Schmidt T, Kupsch A, Trottenberg T, Lempert T. Reversible Leucopenia related to Olanzapine. Movement Disorders 1999; 14: 872 30 Meylan C, Bondolfi G, Aubert A-C, Baumann P. Reversible neutropenia during a cold: possible involvement of risperidone?.Europ Neuropsychopharmacol 1995; 5: 1 ± 2 31 Naumann R, Felber W, Heilemann H, Reuster T. Olanzapine-induced agranulocytosis. Lancet 1999; 354: 566 ± 567 32 Oyewumi LK, Al-Seman Y. Olanzapine: Safe during Clozapine-Induced Agranulocytosis. J Clin Psychopharmacol 2000; 20(2): 279 ± 281 33 Pisciotta AV. Agranulocytosis induced by certain phenothiazine derivates. JAMA 1969; 208: 1862 34 Pisciotta AV. Drug-induced agranulocytosis. Drugs 1978; 15: 132 35 Rizvi MA, Shah SR, Raskob GE, George JN. Drug-induced thrombocytopenia. Curr-Opin-Hematol 1999; 6(5): 349 ± 53 36 Grohmann R, Engel R, Rüther E, Hippius H. The AMSP Drug Safety Program: Methods and Global Results. Pharmacopsychiatry 2004; 37 Suppl 1: S4 ± 11 37 Schuld A, Kraus T, Hinze-Selch D, Haack M, Pollmächer T. Granulocyte colony-stimulating factor plasma levels during clozapine- and olanzapine-induced granulocatopenia. Acta Psychiatr Scand 2000; 102: 153 ± 155 38 Steinwachs A, Grohmann R, Pedrosa F, Rüther E, Schwerdtner I. Two Cases of Olanzapine-Induced Reversible Neutropenia. Pharmacopsychiatry 1999; 32: 154 ± 156 39 Strom BL, Carson JL, Schinnar R, Snyder ES, Shaw M. Descriptive epidemiology of agranulocytosis. Arch Intern Med 1992; 152: 1475 ± 1480 40 Swartz JR, Ananth J, Smith MW, Burgyone KS, Gadasally R, Arai Y. Olanzapine Treatment After Clozapine-Induced Granulocytopenia in 3 Patients. J Clin Psychiatry 1999; 60(2): 119 ± 121 41 Teter CJ, Early JJ, Frachtling RJ. Olanzapine-Induced Neutropenia in Patients With History of Clozapine Treatment; Two Case Reports From a State Psychiatric Institution. J Clin Psychiatry 2000; 61(11): 872 ± 873 42 Tolosa C, Ruiz A, Mari B, Naval E. Olanzapine-Induced Agranulocytosis. Int J Neuropsychopharmacol 2000; suppl 1: S93 43 Wray C-M, Sztuke-Fournier A. Olanzapine: hematological reactions. CMAJ 1998; 159(1): 81 ± 82
Stübner S et al. Blood Dyscrasias Induced ¼ Pharmacopsychiatry 2004; 37 Suppl 1: S70 ± S78
