<<

ANNALS OF CLINICAL AND LABORATORY SCIENCE, Vol. 19, No. 2 Copyright © 1989, Institute for Clinical Science, Inc.

Hematologic Side Effects of Drugs

MICHAEL M. LUBRAN, M.D., P h .D.

Harbor-UCLA Medical Center, Department of Pathology, Torrance, CA 90509

ABSTRACT

Bone marrow and peripheral blood cells may be adversely affected by drugs. Although the risk from most drugs is very small, many cases are reported because of the millions of doses of drugs taken each year by the population. , , hemolytic , , and are the commonest effects, in that order. Aplastic anemia is rare, but very serious when it does occur. Adverse effects may be produced by a direct toxic action of the drug or its metabo­ lites on the or, less often, on circulating cells. Antineoplastic drugs and are examples. Most drugs produce their adverse effects through an immunological mechanism. The drug may act as a hapten or may affect the leading to the production of antidrug and sometimes autoantibodies. may result. may behave in this manner. Some drugs act on erythrocytes with enzyme defects, e.g. glucose-6-phosphate dehydroge­ nase (G-6-PD) abnormalities, to produce . In many cases, the mechanism underlying the adverse effect is unknown. The paper lists the drugs reported to have caused some hematological adverse effect and describes the mechanisms in those cases where they are known.

Introduction related, and the affected cells return to normal after the drug is discontinued; Many drugs may adversely affect the however, in some cases the adverse hematopoietic system and peripheral effect is irreversible, and the patient may blood cells. The principal effects, in die. Serious adverse effects are uncom­ order of frequency, are neutropenia, mon, when viewed against the back­ thrombocytopenia, hemolytic anemia, ground of the millions of doses of drugs aplastic anemia (and ), and taken daily in the total population. Tran­ macrocytic (or megaloblastic) anemia. In sient effects are commoner and pass addition, other types of anemia may unnoticed for the most part unless dis­ occur and also and covered during some clinical examina­ coagulation defects. A drug may cause tion. The frequency of adverse reactions more than one adverse reaction and may to drugs is thus difficult to determine. act by two or more different mecha­ Much of our knowledge is derived from nisms. Often, the effect is dose and time reports in the literature describing one

114 0091-7370/89/0300-01X4 $01.20 © Institute for Clinical Science, Inc. HEMATOLOGIC SIDE EFFECTS OF DRUGS 115 or a small number of cases. Many of side-effect of a drug. In the former case, these reports are anecdotal, and it is the toxic effect is dose and time depen­ sometimes difficult to be sure that the dent; in the latter case, it is often not effect described was caused by the drug dose dependent. Immunological mecha­ and not by the illness for which it was nisms require prior exposure to the given, or by some other drug. However, drug. Subsequent administration of the a few large scale studies have been pub­ drug results in formation, lead­ lished,5,816,17 and the modes of action of ing to damage of blood cells. Enzyme some drugs on the blood system have inhibition may be a property of the drug been studied. Many countries now col­ and occur in all subjects receiving it (as, lect systematically adverse drug reac­ for example, with many chemotherapeu­ tions of all types reported to them by tic agents), or the drug may act on cells physicians, and there are several books genetically deficient in an essential detailing adverse reactions.9,11,15,22 This enzyme (e.g., glucose-6-phosphate paper will describe the adverse hemato­ dehydrogenase). A few drugs decrease logical reactions (excluding coagulation) absorption of or B12, causing a attributed to drugs, gleaned from the lit­ macrocytic anemia or megaloblastosis. erature and, whenever possible, the Although the adverse effect can be mechanisms underlying the reactions. explained in some cases, the cause is Only drugs used in the United States unknown for the majority of drugs. will be considered. These idiosyncratic effects are unpredic­ table, cannot be deduced from the Adverse Effects of Drugs known pharmacological properties of the drug and may or may not be dose Drugs may produce their adverse related. They may be related to some effects through toxicity, by an immuno­ genetic feature of the patient, e.g., the logical process, by inhibiting the action patient may be a slow acetylator, but in of important enzymes, by decreasing the most cases no cause can be found. Idio­ absorption of substances essential for syncratic drug reactions affecting blood normal hematopoiesis, or by as yet cells are uncommon; nevertheless, unknown mechanisms. A well-studied because of the enormous number of drug, now not often used, is chloram­ doses of drugs given to the population, phenicol.23 This drugs binds to mito­ they will occur almost inevitably, even chondria, inhibiting the formation of with the safest drugs. Most of the inner membrane enzymes by inhibiting reported adverse drug effects relate to peptidyltransferase, necessary for the the idiosyncratic behavior of the drug in formation of peptide bonds. The toxic one or a very few subjects. However, effects of chloramphenicol are caused by there are some drugs which produce its action on mitochondria. Bone marrow adverse effects in an appreciable number aplasia occurs in about 1:19,000 to of subjects. The following sections will 1:200,000 courses of therapy. Most describe the various adverse hematologi­ patients develop a dose-related revers­ cal effects and the drugs causing them. ible erythroid suppression, caused by inhibition of ferrochelatase, an inner Aplastic Anemia membrane enzyme. The toxic effect of drugs may be delib­ Aplastic anemia is an occasional com­ erately sought to suppress abnormal cell plication of the of cancer. activity, as in the use of chemotherapeu­ Antineoplastic drugs are usually used in tic agents, or it may be an unwanted combination to treat cancers and hema­ 116 LUBRAN tological malignancies by interfering Tamoxifen, an antagonist, may with and division. In the cause or thrombocytopenia. doses used, rapidly dividing normal cells Vinca alkaloids (vinblastine, vincristine, are also affected, including those in the vindesine) and podophyllin derivatives bone marrow. A selective anemia, neu­ (etoposide, teniposide) arrest mitosis in tropenia, thrombocytopenia, or combi­ metaphase and stop cell division. The nation of these occur regularly depend­ podophyllin drugs may severely depress ing on the dose, duration of treatment, the bone marrow; the vinca drugs are and susceptibility of the patient. Less less toxic, causing usually a transient leu­ commonly, aplastic anemia or pancyto­ kopenia and thrombocytopenia. Asparag­ penia may develop. inase may cause a leukopenia. Alkylating agents act on DNA, cross- Aplastic anemia is an uncommon linking its strands or breaking them. All result of other drugs. Various estimates phases of the are affected and exist for the risk of developing it follow­ all bone marrow cells may be involved. ing use of certain classes of drugs. The Alkylating agents include busulfan, car- incidence of aplastic anemia among sub­ mustine, chlorambucil, cisplatin, cyclo­ jects taking therapeutic drugs is about phosphamide, dacarbazine, lomustine, 2.2:1,000,000. More than half of these mechlorethamine, melphalan, pipobro- cases result from drugs and the mortality man, thiotepa, and uracil mustard. is high. For indomethacin the risk of Antimetabolites affect DNA synthesis aplastic anemia is about 1:100,000, for by enzyme inhibition or by blocking the about 1:300,000 and synthesis of or com­ much lower for other analgesics. The risk ponents. , , is also very low for some antithyroid thioguanine are purine antagonists; drugs. cytarahine, floxuridine, fluorouracil are Drugs with a relatively high risk for pyrimidine antagonists; aplastic anemia include , binds to and , chloramphenicol, gold salts, hydroxyurea inhibits ribonucleotide indomethacin, meclofenamate, mefe- reductase. Antimetabolites may act dur­ namic acid, mephenytoin, oxyphenbuta- ing the entire cell cycle but are most zone, paramethadione, , effective in the S phase (DNA synthesis). phenylbutazone, , quinacrine, Methotrexate inhibits folic acid synthesis . Phenytoin toxicity may and may cause megaloblastosis. be related to a transient antibody forma­ antineoplastic agents affect tion and production of abnormal T-sup- DNA and RNA synthesis. Dactinomycin pressor cells.10 Phenytoin and carbamaz- and daunorubicin depress all cell types epine give rise to toxic arene-oxide of the bone marrow; dexorubicin may intermediate metabolites, which bind cause ; mithramycin and covalently to macromolecules of stem m itom ycin produce thrombocytopenia; cells in the marrow and lymphocytes.13 bleom ycin does not depress the bone About 70 percent of patients on long­ marrow. term phenytoin therapy have mild to Other antineoplastic agents: Procar­ severe abnormalities of the immune sys­ bazine inhibits DNA, RNA and protein tem. synthesis and may cause leukopenia and Low risk drugs, reported in the litera­ an em ia. Aminoglutethimide, w hich ture once or a few times include acet­ inhibits the synthesis of some steroid aminophen, acetazolamide, acetohexa- hormones, causes a transient leukopenia mide, aspirin, bendroflumethiazide, but may give a severe pancytopenia. benthiazide, captopril, , HEMATOLOGIC SIDE EFFECTS OF DRUGS 117

carbimazole, chlordiazepoxide, chloro- prothixene, , , desi- quine, chlorothiazide, , pramine, , ethacrynate, fluphen- chlorpropamide, chlorthalidone, chlor- azine, furazolidone , , , cimetidine, , co- ibuprofen, , levamisole, trimoxazole, cyclothiazide, dichlorphen- , mercaptomerin, mesorida- amide, ethosuximide, ethotoin, zine, methotrimeprazine, , ethoxzolamide, hydrochlorothiazide, metronidazole, , , hydroxychloroquine, , mepro­ novobiocin, para-aminosalicylic acid, bamate, methamazole, metharbital, penicillamine, , piperaceta- methazolamide, methimazole, methsuxi- zine, procainamide, , quini- mide, methyclothiazide, metolazone, dine, , thiethylperazine, thiorid­ oxytetracycline, , phenacemide, azine, thiothixene, tocainide, trifluoper­ phensuximide, polythiazide, prima­ azine, , trimeprazine, quine, , , , tybamate. Drugs usually , propylthiouracil, pyrimeth­ producing a selective neutropenia amine, quinethazone, salicylate, strepto­ include , ampicillin, bacampi- mycin, sulfacytine, sulfadiazine, sulfa- cillin, capreomycin, , cefa­ merazine, sulf amethizole , clor, , , , sulfamethoxazole, sulfapyridine, sulfasa­ , , cephalexin, cepha- lazine, sulfisoxazole, sulindac, tolaza­ loglycin, , cephalothin, mide, tolbutamide, trichlormethiazide, cephapirin, cephradine, chlorphenira­ trimethadione, , val­ mine, , clofibrate, cyclacil- proate. lin, dantrolene, demeclocycline, diaze­ pam, diazoxide, doxycycline, , , , levodopa, linco- Agranulocytosis and Neutropenia mycin, , methacycline, methyl- dopa, , , moxalac- Agranulocytosis may arise through the tam, nafcillin, , oxy tetracycline, toxic action of the drug on bone marrow penicillin G, penicillin V, pyrimeth­ precursors6,7 or by destruction of cells by amine, rifampin, tetracycline, , circulating antibodies. Some , , vidarabine, drugs depress suppressor T-cells and allow the production of autoimmune antibodies by the B-cells.25 Anti Hemolytic Anemia drugs may behave in this way. In most cases, the mechanism is unknown and Drug-induced hemolytic anemia may presumed to be idiosyncratic. occur as a result of a direct toxic effect on Neutropenia and agranulocytosis may normal circulating erythrocytes, on cells be produced by the drugs listed pre­ with certain enzyme defects, on cells viously causing aplastic anemia. Anti­ with some unstable , or neoplastic drugs are responsible for through an immune mechanism. many of the cases of neutropenia;1 in Direct toxic effects are dose related, fact, a selective neutropenia is much have a slow onset, and occur in all sub­ commoner than aplastic anemia or jects. Few drugs are known to cause agranulocytosis. Thrombocytopenia may hemolytic anemia in this way. The best or may not occur as well. Other drugs known are phenacetin (which also acts on producing neutropenia, which may prog­ enzyme-defective cells) and the sulfones ress to agranulocytosis, include aceto- used in the treatment of leprosy, espe­ phenazine, , , cially dapsone. This drug causes hemo­ , carphenazine, chlor- lysis in all subjects, methemoglobinemia 118 LUBRAN in many and hemolytic anemia in a few duce lysis may be formed by two differ­ on prolonged treatment. Drugs which ent mechanisms. In the hapten-cell may be toxic to erythrocytes, but may mechanism, certain drugs given in large also act by other mechanisms, include doses bind to proteins of the red cell cephalothin, chlorpromazine, clemas­ membrane and act as haptens. Cephalo­ tine, ethoxzolamide, hydrochlorothia­ sporins and penicillins may produce zide, indomethacin, mefanamic acid, these antibodies. Penicillin is present on metaxolone, methazolamide, methyl- the surface of red cells of all patients dopa, oxyphenbutazone, phenazopyri- receiving the drug. About three percent dine, phenylbutazone, streptomycin, of patients on high doses of penicillin sulfamethazine. have IgG antibodies, and a small number Enzyme defects: The principal of these patients go on to develop hemo­ enzyme involved in hemolytic anemia is lytic anemia. g lu co se-6 -phosphate dehydrogenase In the immune complex mechanism, (G-6 -PD). These drugs give rise to acute the drugs bind to a serum protein and episodes of hemolytic anemia which may combine with the resulting antibody to be dose related, when administered to form an immune complex. This adsorbs G-6 -PD deficient patients. In some to the surface of the red cell and to other cases, the condition is self-limiting and cells, e.g., thrombocytes and leukocytes, disappears in about seven days, even binds complement and causes cell though drug administration is continued. destruction. The antibody is usually IgM This occurs when existing G-6 -PD defi­ in type. Quinidine and rifampin produce cient cells have been destroyed and the immune complexes. newly formed cells have become resis­ An auto-immune hemolytic anemia tant to the drug. The erythrocytes in may be drug-induced. It is believed that drug-induced hemolytic anemia caused the drug binds to one of the blood group by enzyme deficiency usually show baso­ antigens on the red cell surface, usually philic stippling and may contain Heinz an Rh antigen. The resulting antibody is bodies. Drugs include chloramphenicol, directed against the blood group antigen; , dapsone, , fur­ occasionally, an anti-drug antibody coex­ azolidone, isoniazid, menadione, nali­ ists. 14 About 15 percent of patients on dixic acid, nitrofurantoin, phenacetin, methyldopa and about nine percent on phytonadione, primaquine, probenecid, levodopa have demonstrable autoim­ quinidine, quinine, sulfapyridine, sulfi- mune antibodies in their sera after treat­ soxazole, sulfoxone. ment for three months to a year; about Hemolytic anemia is seldom caused by one in ten thousand develop hemolytic drugs in the case of other red cell anemia. An alternative to the hapten enzyme deficiencies, but existing anemia mechanism is the suppression of T-cell may be exacerbated. Sulfoxone may function through an increase of c-AMP in cause hemolytic anemia in glutathione lymphocytes, leading to an unregulated reductase deficiency. production of autoantibodies by B-cells. Unstable hemoglobins: Exacerbation The autoantibodies are usually of the Rh of existing hemolytic anemia may occur system . 18 The drug does not act as a hap­ in subjects with some unstable hemoglo­ ten, but produces its effects via the bins, e.g., Hb-Zurich, when given the immune system. The immune system drugs causing hemolysis in G-6 -PD defi­ plays an important role in regulating ciency. normal hematopoiesis. 21 Drug-induced immune hemolytic ane­ Drugs giving rise to antibodies include mia: Antibodies which react with red cefamandole, cefotaxine, cefoxitine, cells to activate complement and pro­ cephalothin, co-trimoxazole, isoniazid, HEMATOLOGIC SIDE EFFECTS OF DRUGS 119 levodopa, mefanamic acid, , crynate, ethclorvynol, ethinamate, eth­ methyldopa, moxalactam, nalidixic acid, ionamide, flucytosine, , oxacillin, para-aminosalicylic acid, peni­ furosemide, isoniazid, lincomycin, cillin G, penicillin V, quinidine, quinine, mephobarbital, meprobamate, methar- rifampin, sulfacytine, sulfadiazine, sul- bital, methimazole, para-aminosalicylic famerazine, sulfamethizole, sulfapyri- acid, penicillamine, penicillins, phenace- dine, , ticarcillin, tolmetin. tin, phenobarbitone, phenylbutazone, Autoantibodies are found with levodopa, , phenytoin, procaina­ methyldopa and mefanamic acid. mide, propylthiouracil, quinidine, qui­ nine, rifampin, salicylates, streptomy­ Thrombocytopenia cin, sulfonamides, ticarcillin, tocainide, trichlormethiazide, trimethoprim, val­ This may result from a direct toxic proate, vidarabine. effect of the drug on precursors in the bone marrow or on the circulating . Drugs associated with aplastic anemia also cause thrombocytopenia, This may result from administration of through action on megakaryocyte pre­ drugs interfering with the absorption of cursors in the marrow. Low risk drugs folic acid or B12 or the metabolic may cause a selective thrombocytopenia, functions in which they are involved. which is often the only indication of sup­ Frank megaloblastic anemia is uncom­ pression of marrow activity. Ristocetin mon. The usual effect is , acts directly on circulating platelets. unless the patient is severely ill or Lithium carbonate given for one or more undernourished and deficient in folic years lowers the platelet count, but acid or . In many cases, there rarely causes purpura.2 is no clinical evidence of folate or B12 Thrombocytopenia may also result deficiency, but their concentrations in from an immune mechanism. Drug and the blood are low. platelet combine and produce an anti­ Folic acid: Low serum folate concen­ body against the drug. An immune com­ trations caused by poor absorption of plex is formed which is adsorbed non- folate may result from administration of specifically on to platelets and other , ethotoin, mephobarbital, cells. This is referred to as “the innocent metharbital, nitrofurantoin, phensuxi- bystander” hypothesis. As an example, mide, phenytoin, primidone. Megaloblas­ valproate depresses the platelet count in tic anemia may occur in a few cases. about one third of the cases. 3,24 How­ About 50 percent of patients on pheny­ ever, the immune complex may combine toin have low serum folate concentra­ specifically with platelets. The antibody tions and about 30 percent have red cell to quinidine combines with the major macrocytosis and early megaloblastic glycoprotein complex on the platelet changes in the bone marrow. About one surface and the antibody to ticarcillin percent develop frank megaloblastic ane­ combines with a polymorphic platelet m ia. 4 It may take many years of drug antigen . 14 Drugs acting by an immune administration for this to occur. It is not mechanism include antimony potassium clear why some patients should develop tartrate, , brompheniramine, this condition, but dietary intake of capreomycin, captopril, chlorothiazide, folate may be a contributory factor. chlorpromazine, chlorpropamide, cle­ Dihydrofolate reductase is inhibited mastine, clindamycin, clonazepam, by methotrexate, triamterene, pyrimeth­ deferoxamine, diazoxide, digitoxin, etha- amine, trimethoprim. This is a key 120 LUBRAN enzyme in folate metabolism and DNA associated with aplastic anemia, may synthesis. Megaloblastic anemia will produce anemia by depression of the occur if the drug is given for a sufficient bone marrow or by other means. Isonia- length of time, in high enough doses, zid increases the severity of the anemia but in practice it is unusual. in patients suffering from pyridine- , floxuridine, fluoruracil responsive . Rarely, (antimetabolite antineoplastic agents) it may produce a pure red cell hypopla­ inhibit pyrimidine synthesis; hydroxy­ sia . 19 This condition may also be pro­ urea inhibits ribonucleotide reductase; duced by chlorpropamide, co-trimoxa- azathioprene, mercaptopurine, thiogua- zole, penicillin, phenylbutazone, nine inhibit purine synthesis; sulfon­ phenytoin, tolbutamide. Indomethacin amides block the synthesis of folic acid. produces an iron-deficiency anemia and All of these drugs cause a low serum amphotericin B a normochromic, nor- folate and eventually megaloblastic ane­ mocytic anemia. Other drugs which may mia. produce anemia, usually hypochromic, Vitamin B,,: Low serum concentra­ include amoxicillin, ampicillin, bacampi- tions may result from administration of cillin, cyclacillin, cefadroxil, cefazolin, colchicine, neomycin, para-aminosalicy- cephalexin, cephaloglycin, cephalori- lic acid, slow release potassium chloride, dine, cephapirin, novobiocin, oleovita­ which impair absorption of vitamin B12. min A, vidarabine.

Drug-induced Methemoglobinemia References

1. Ar n e b o r n , P. and Pa lm b la d , J.: Drug-induced The commonest cause is self-medica­ neutropenia-a survey for Stockholm 1973-1978. tion with pain-relieving drug combina­ Acta Med. Scand. 212:289-292, 1982. tions containing phenacetin.20 O th er 2. B a l o n , R., B e r c h o u , R., Ly c a k i, H ., and P o h l , R.: The effect of lithium on platelet drugs are amyl nitrite, dapsone, prilo- count. Acta Psychiatr. Scand. 74:474-478, 1988. caine, primaquine, sulfoxone, trimetho­ 3. B arr, R. D ., C o p e l a n d , S. A ., St o c k w e l l , prim. Primaquine acts on red cells defi- M . L ., M o r ris, N., andKELTON, J. C.: Valproic acid and immune thrombocytopenia. Arch. Dis. cient in NADH methemoglobin C hild h o o d 57:681-684, 1982. reductase. Methemoglobinemia is pro­ 4. B e g h i , E ., M a s c io , R., and T o g n o n i, G.: duced by drugs which are redox com­ Adverse effects of drugs-a critical review. Adv. Drug React. Ac. Pois. Rev. pounds or form a complex with heme. 5:63-80, 1986. Hydrogen peroxide or peroxide radicals 5. Bo tt ig er , L. E ., F u r h o f f , A. K ., and H o l m - are formed, leading to the formation of BERG, H .: Drug-induced blood dyscrasias, Acta Med. Scand. 205:457-461, 1979. methemoglobin . 12 In some cases, the 6 . C h a p l in , S.: Bone marrow depression due to reaction proceeds further, resulting in manserin, phenylbutazone, oxyphenbutazone, Heinz bodies and hemolytic anemia. and chloramphenicol. Part 1. Adv. Drug React. Ac. Pois. Rev. 5:97-136, 1986. Most drugs producing hemolytic anemia 7. C h a p l in , S.: Bone marrow depression due to by direct toxic action on the red cells will manserin, phenylbutazone, oxyphenbutazone, also produce methemoglobin in the early and chloramphenicol. Part 2. Adv. Drug React. Ac. Pois. Rev. 5:181-196, 1986. stages. Some drugs act by producing 8. D a n ie l s o n , D . A., D o u g la s, S. W. I l l, H er ­ toxic metabolites in the . z o g , P., H e r s h e l , J., an d P o r ter, J. B .: D rug- in d u ced blood disorders. J. Amer. Med. Assoc. 252:3257-3260, 1984. Anemia 9. D e G ruchy, G. C.: Drug-induced Blood Dis­ orders. Oxford, Blackwell, 1975. Drugs which may cause aplastic ane­ 10. D o s c h , H-M., J aso n , J., and G e l f a n d , E. W.: Transient antibody deficiency and abnormal T mia may also have lesser effects, of suppressor cells induced by phenytoin. New which anemia is one. Other drugs, not Engl. J. Med. 306:406-409, 1982. HEMATOLOGIC SIDE EFFECTS OF DRUGS 121

11. D u k e s , M. N. G. and B ee le y , L .: Side effects anaemia in relation to use of antithyroid drugs. of drugs. Annual 11. Amsterdam, Elsevier, Brit. Med. J. 297:262-265, 1988. 1987. 18. K ir t l a n d , H . H . Ill, M o h l e r , D. N., and 12. G abdu za , T. G .: Drug- interactions. H o r w it z, D. A.: Methyldopa inhibition of sup- Drugs and Hematologic Reactions. Dimitrov, pressor-lymphocyte function. New Engl. J. N. V. and Nodine, J. H., eds. New York, Grune Med. 302:825-832, 1980. and Stratton, 1974, pp. 49-64. 19. L e w is , C. R. an d M anoharan , A.: Pure red cell 13. G e r s o n , W. T ., F in e , D . G ., Sp ie l b e r g , S. P., hypoplasia secondary to isoniazid. Postgrad. a n d S ensenbrenner , L. L.: Anticonvulsant- M ed. J. 63:309-310, 1987. induced aplastic anemia: increased susceptibility 20. N a g e l , R . L. and R a n n e y , H. M.: Drug- to toxic drug metabolites in vitro. B lood 61:889- induced oxidative denaturation of hemoglobin. 893, 1983. Semin. Hematol. 10:269-278, 1973. 14. H a b ib i, B.: Drug induced autoan­ 21. Parkm an, R.: The immunopathology of marrow tibodies co-developed with drug specific anti­ failure. Clin. Haematol. 7:475-486, 1978. bodies causing haemolytic anaemias. Brit. J. 22. Rid d e l l , R. H.: Pathology of Drug-induced and Haematol. 61:139-143, 1983. Toxic . New York, Churchill Living­ 15. H a m il t o n , H . K.: Professional Guide to Drugs. stone, 1982. Pennsylvania, International Communications 23. Sm it h , A. L. and W e b e r , A. Pharmacology of Inc., 2nd ed ., 1983. chloramphenicol. Pediatr. Clin. N. Amer. 16. International Agranulocytosis and Aplastic Ane­ 30:209-235, 1983. mia Study. Risks of agranulocytosis and aplastic 24. T urnbu Ll , D. M.: Adverse effects of valproate. anemia. A first report of their relation to drug Adv. Drug React. Ac. Pois Rev. 2:191-216, use with special reference to analgesics. J. Amer. 1983. Med. Assoc. 256:1749-1757, 1986. 25. W e t z m a n , S. A., St o s s e l , T. P., and D e s ­ 17. International Agranulocytosis and Aplastic Anae­ m o n d , M.: Drug-induced immunological neu­ mia Study. Risk of agranulocytosis and aplastic tropenia. Lancet 1:1068-1072, 1978.