Type
Characteristic
Mechanism
Examples
Type A
Dose-dependent
Chemotherapy-induced myelosuppression
Predictable
Type B
Dose-independent
Nonimmune-mediated
Primaquine-induced hemolytic anemia, metformin-induced megaloblastic anemia, methotrexate-induced megaloblastic anemia, clozapine-induced neutropenia
Unpredictable
Immune-mediated
IgG-mediated cytotoxicity (type II Gell and Coombs)
Penicillin-induced hemolytic anemia, cephalosporin-induced hemolytic anemia, methyldopa-induced hemolytic anemia, aminopyrine-induced neutropenia
Immune-complex deposition (type III Gell and Coombs)
Quinine-induced hemolytic anemia, heparin-induced thrombocytopenia, bevacizumab-related thrombocytopenia
From a clinical point of view, drug-induced hematological disorders are quite difficult to diagnose for at least three reasons: they occur rarely, clinical presentation may be indistinguishable from that induced by other causes of dyscrasias, and often the available literature is not exhaustive [6]. Moreover, they may have multifaceted clinical presentations, either in terms of the blood cell lines which are involved or of the time of onset after drug exposure, this spanning from few hours to several weeks.
From this analysis it follows that hematological ADRs represent challenging threats for clinicians, as phenotypic unpredictability is common place.
The aim of this chapter is to focus on the epidemiological, pathogenetic, clinical, and therapeutical aspects of drug-induced blood dyscrasias. Each blood cell line is considered separately for ease of consultation.
8.2 Drug-Induced Thrombocytopenia
8.2.1 Epidemiology
Drug–induced thrombocytopenia (DIT) was first noted in 1865, when Vipan reported the occurrence of purpura in patients treated with quinine. Nowadays, several epidemiological studies, performed both in the USA and in Europe, estimated the minimum incidence of DIT at about 10 cases per million population per year in the general population, but higher estimates were shown in some settings, such as among hospitalized patients and among the elderly [7].
Among ICU patients, thrombocytopenia is a particularly frequent occurrence, as it was reported in 15–58 % of admitted patients, and DIT is reported in up to 25 % of acutely ill patients [8].
8.2.2 Causative Drugs
More than 200 drugs have been reported as possible causes of DIT and the list is expected to become longer in the ensuing years in the light of the increasing number of new biotherapeutic drugs which are becoming clinically available [9]. Among the various attempts at classifying levels of evidence of DIT, George et al. proposed in 1998 some specific criteria for the assessment of probability of causal relationship for DIT, on the basis of which four different levels of evidence were defined [10]. Essentially, four different criteria were identified: (1) Drug administration came before thrombocytopenia, and complete and sustained recovery was obtained after drug withdrawal; (2) the suspected drug was the only one administered before the onset of thrombocytopenia, or the other drugs were continued or reintroduced without any persistence of thrombocytopenia; (3) other etiologies of thrombocytopenia were excluded; (4) reexposure to the suspected drug led to recurrence of thrombocytopenia. On the basis of these criteria, four different levels of evidence were defined: Definitive: 4 criteria met; probable: the first 3 criteria are met; possible: criteria 1 is met; unlikely: criteria 1 is not met [10].
For comprehensive lists of drugs causing DIT, on the basis of these and of other criteria, the readers are referred to some recent reviews [11–13] and to the following URL: http://www.ouhsc.edu/platelets.
Some of the most relevant drugs with definite or probable evidence for DIT are listed in Table 8.2. Among drugs that are frequently used in the ICU setting, those deserving major attention are heparin, sodium valproate, some antibiotics, and the GPIIb/IIIa inhibitors (tirofiban, eptifibatide, and abciximab).
Table 8.2
Drugs with definite or probable evidence for causality of drug-induced thrombocytopenia
Drug category | Level of evidence: definitive | Level of evidence: probable |
---|---|---|
Analgesics | Acetaminophen, diclofenac, meclofenamate, tolmetin, | Ibuprofen, naproxen, oxyphenbutazone, sulindac |
Anti-infective drugs | Amphotericin B, cephalotin, ethambutol, interferon-alfa, isoniazid, methicillin, nalidixic acid, novobiocin, quinine, piperacillin, rifampicin, sulphisoxazole, TMT-SMX, vancomycin | Ampicillin, fluconazole, oxytetracycline |
Antiepileptics | Carbamazepine, oxcarbazepin, phenytoin | |
Antineoplastics | Aminoglutethimide, rituximab, tamoxifen, trastuzumab | |
Antirheumatics | Levamisole, sulfasalazine | Gold salts, infliximab |
Cardiovascular drugs | Abciximab, alprenolol, amiodarone, amrinone, atorvastatin, heparin, diazoxide, digoxin, eptifibatide, methyldopa, minoxidil, nitroglycerine, oxprenolol, quinidine, tirofiban | Captopril, clopidogrel, hydrochlorothiazide, procainamide, rosuvastatin, simvastatin, ticlopidine |
Gastrointestinal drugs | Aminosalicylic acid, cimetidine | Ranitidine |
Psychotropic drugs | Chlorpromazine, diazepam, haloperidol, lithium, thiothixene | |
Other drugs | Danazol, deferoxamine, diatrizoate, diethylstilbestrol, difluoromethylornithine, iopanoic acid, naphazoline, meglumine | Glibenclamide |
It has been estimated that approximately 1–5 % of patients treated with heparin may experience heparin–induced thrombocytopenia (HIT). It occurs ten times more frequently with unfractioned heparin than with low-molecular-weight heparins, and with the highest frequency among orthopedic patients [14].
Thrombocytopenia is the most common hematologic abnormality associated with sodium valproate. Its incidence has been reported between 5 and 40 %, regardless of absolute values of plasma drug concentrations [15].
As far as antimicrobials are concerned, a Dutch retrospective case-controlled study in hospitalized patients showed that beta-lactams were associated with an increased risk of DIT, with an odds ratio of 7.4 [16]. Among the various beta-lactams, piperacillin is considered to be one of the most frequently associated with DIT [17]. Also vancomycin is shown to be an important cause of DIT [18]. Trimethoprim–sulfamethoxazole-induced thrombocytopenia was estimated at 1 in 25,000 patients [8]. Recently, linezolid was shown to be frequently associated with dose-dependent DIT in patients on long-term treatment [19].
The prevalence of DIT caused by GPIIb/IIIa inhibitors is estimated between 0.1 and 2 % of cases [20–22].
Among other potential causes of thrombocytopenia to take into account for differential diagnosis versus DIT, it should not be overlooked that iodinated contrast media, some common beverages (tonic water and bitter lemon), some herbal remedies (Jui herbal tea), and foods (tahini sesame seeds and Lupinus termis beans) may be involved [23].
8.2.3 Pathogenesis
DIT may occur through two major pathogenetic mechanisms [23].
The first is a nonimmune-mediated mechanism, which in most cases leads, in a dose-dependent fashion, to bone-marrow suppression with pancytopenia. This is typically related to several antineoplastic agents that may block cell proliferation. Additionally, other drugs that may cause nonimmune-mediated DIT may be linezolid, thiazide diuretics, colchicine, and amrinone.
The second mechanism, which is more common and challenging, is immune-mediated and yields to a specific lysis of circulating platelets. Indeed, six different mechanisms are currently advocated for immune-mediated DITs (Table 8.3) [21, 24]. Among these, the hapten–induced antibody theory was originally advocated to explain hemolytic anemia caused by high-dose penicillin, and it is less certain whether this mechanism may be applied also to DIT. This process involves the covalent binding of the penicilloyl group of an opened beta-lactam ring to the amino groups of glycoproteins in the platelet membrane with consequent perturbation of the antigen processing of protein that ends up with a new antigenic structure, or neo-epitope, toward which a specific immune response is elicited.
Table 8.3
Mechanisms of immune-mediated drug-induced thrombocytopenia
Type | Mechanism | Clinical consequence | Examples |
---|---|---|---|
Hapten-dependent antibody | Drugs covalently bind to platelet glycoproteins producing a neoantigen that elicits an immune response | Hemorrhage | Penicillin, cephalosporin |
Drug-dependent antibody | Antibodies bind a complex formed by drugs noncovalently linked to platelet glycoprotein | Hemorrhage | Quinine, quinidine, sulfonamides, NSAIDs |
Ligand-induced binding site | Drugs bind to platelet GPIIb/IIIa producing a neoantigen that elicits an immune response | Hemorrhage | Eptifibatide, Tirofiban |
Drug-specific antibody | Antibodies react to murine component of the drug | Hemorrhage | Abciximab |
Autoantibody induction | Drugs induce antibodies that reacts with platelet glycoprotein in absence of drug | Hemorrhage | Gold salts, procainamide, rituximab, infliximab |
Immune-complex | Antibodies bind drug-PF4 immune-complex and cross-link between platelets, leading to aggregation | Thrombosis | Heparin |
Drug–dependent antibody theory has been recently revisited and an integrated view of previous theories is now well-accepted. According to this mechanism, antibodies causing DIT may react weakly with epitopes or glycoproteins on the surface of platelet membranes, increasing their avidity to their targets when a specific drug, presenting structural polar affinities, interacts noncovalently both with platelets and with antibodies. Consequently, drug is trapped between antibodies and platelets, and this strengthens the interaction. Whether the production of antibodies is promoted by the drug itself or whether it derives from a pool of naturally produced immunoglobulins is not well-understood. Drugs recognized to belong to this category are quinine, quinidine, nonsteroidal anti-inflammatory drugs, and some sulfonamide antibiotics.
Tirofiban and eptifibatide are synthetic small molecules that react with the arginine-glycine-aspartic acid (RGD) recognition site on platelet glycoprotein IIb/IIIa, preventing this receptor from binding fibrinogen, thus blocking platelet aggregation and thrombus formation. These drugs, by binding to the GPIIb/IIIa complex on the platelet surface, may induce conformational changes leading to the emergence of cryptic domains with resultant antibody formation and platelet destruction (ligand–induced binding site).
Another antiplatelet agent causing thrombocytopenia is abciximab, a chimeric monoclonal antibody specific for GPIIIa, that may induce the production of specific antibodies which recognize the murine component of the molecule itself, and platelets are consequently destroyed as bystanders when coated with this drug (drug–specific antibody).
Some drugs may promote the production of autoantibodies (autoantibody induction), which may last for a long period of time even after drug discontinuation, leading to chronic autoimmune thrombocytopenic purpura. Typical drugs in this category are gold salts, once used in the treatment of rheumatoid arthritis, and procainamide. In recent years, numerous reports have suggested that DIT arising after treatment with the monoclonal antibodies rituximab, infliximab, etanercept and efalizumab should be referred to this mechanism.
The last type of immune-mediated mechanism is due to immune–complex formation, and is typical of heparin and heparin-like drugs. Essentially, heparin links to platelet factor 4 (PF-4) on platelet surface. This aggregate, in turns, binds to heparin-induced immunoglobulin G (HIT-IgG), whose Fc portion cross-links with FcγRII on the same platelet or on adjacent platelets, thus promoting further activation of platelets, release of PF-4, and platelet aggregation.
8.2.4 Clinical Presentation
Whereas the time of onset of pancytopenia due to cytotoxic myelosuppression may take several weeks to become of clinical concern, conversely the onset of symptoms due to immune-mediated mechanisms is generally more rapid. Indeed, the onset depends on the time to mount the immunologic response. In case of primary immunization, frequently it takes 2–3 weeks before clinical appearance. However, the onset of thrombocytopenia may occur in less than 24 h if significant drug-induced antibody titer due to recent exposure exists, or in 3–10 days if the antibody titer has fallen after previous exposure [25].
A notable exception to this rule is represented by platelet inhibitors tirofiban, eptifibatide, and abciximab, whose naturally occurring antibodies may lead to acute thrombocytopenia within hours of the first exposure [22].
Most patients experiencing DIT usually have moderate-to-severe thrombocytopenia (platelet count less than 55,000 platelets/μL), and platelet counts lower than 20,000/μL are relatively frequent. Clinical presentations of DIT may be petechial hemorrhages, bruising, and epixtasis. Systemic symptoms such as chills, fever, nausea, and vomiting may often anticipate bleeding signs. Major bleeding was reported in 9 % of patients with DIT, whereas minor bleeding occurred in 28 % [8]. Life-threatening conditions due to platelet drop to 1,000 platelets/μL could also be expected. In these cases, patients could experience gastrointestinal, genitourinary, intracranial, or pulmonary hemorrhage [25].
Differently from other DITs, heparin may be associated with two different types of thrombocytopenia. The first, known as HIT–1, is a spontaneously resolving benign form characterized by a mild decrease in platelet count, rarely below 100,000/μL. The second type, described as HIT–2 and considered the typical HIT-form, occurs 5–10 days after the initiation of therapy and is associated with a more than 30–50 % drop in platelet count [26]. About one-third of HIT occurs within few days since initiation of therapy, and is usually associated with reexposure to heparin within 100 days since the first episode [27].
Interestingly, in contrast to the usual course of DIT, platelet nadir in HIT is less severe (~55,000 platelets/μL), since counts below 20,000/μL occur in less than 10 % of cases. Of note, HIT has a different clinical evolution with respect to other DITs. The major clinical complication of HIT is thrombosis as a consequence of the immune-complex mechanism with platelet activation and consumption. The most frequent clinical scenario is represented by venous thrombosis, with pulmonary embolism being the most common fatal complication. Conversely, spontaneous hemorrhage is uncommon. Interestingly, some authors have postulated the “iceberg” model of HIT [28, 29]. This model postulates that whereas a significant proportion of patients may produce HIT–IgG while on treatment with heparin, indeed only a smaller proportion become thrombocytopenic and an even smaller proportion may develop thrombotic complications.
8.2.5 Management
The unexpected occurrence of thrombocytopenia in a patient with a recent history of drug exposure should always give rise to the suspicion of DIT. In most cases the identification of the putative drug is quite challenging, as therapeutic schedules often include polytherapy, especially in the ICU. For this reason, a stepwise approach in the diagnosis of DIT is recommended [25, 30].
After excluding other causes of DIT, such as pseudothrombocytopenia (i.e., spurious in vitro causes of low platelet count), disseminated intravascular coagulation, or thrombotic thrombocytopenic purpura/hemolytic uremic syndrome, a causal relationship should be evaluated according to the criteria proposed by George et al. [10]. Although the diagnosis of DIT is usually based mainly on clinical criteria, definitive confirmation of the suspicion of DIT may come through the identification of drug–induced antiplatelet antibodies. Unfortunately, testing is technically challenging and, with the notable exception of heparin, few laboratories may provide with these antibody assays. Two major categories of test are available, namely immunoassays and functional tests [23, 26]. Immunoassays measure immunoglobulins associated with or bound to platelets. Among the different available methods, flow cytometry is the most rapid and sensitive for the detection of antibodies induced by the most common causative drugs.
Functional assays, as for example the 14 C–serotonin release assay and the heparin–induce platelet aggregation test, measure antibody-induced changes in platelet activation. These tests are widely applied for the diagnosis of HIT, but rarely for other DIT.
Overall, it should also be taken into account that in patients with a highly suggestive clinical history of DIT, laboratory tests may sometimes result falsely negative. This could be due to some technical problems (i.e., insolubility of many drugs in in vitro testing) [31].
As far as the treatment of DIT is concerned, discontinuation of the causal medication is of paramount importance. After drug discontinuation, the prognosis of DIT is generally excellent. Platelet count recovers to more than 100,000/μL generally in 1–10 days, with a median of about 7 days [25].
Patients presenting with severe thrombocytopenia should be aggressively treated with platelet transfusions in order to avoid fatal intracranial or intrapulmonary haemorrhage. The role of high-dose corticosteroids and of intravenous immunoglobulins in DIT remains controversial [20].
Platelet transfusions should not be applied to HIT, as they could increase the rate of thrombotic complications [32]. In patients with HIT, clinicians may face the need to continue the anticoagulative treatment. The choice of low-molecular-weight heparins and of warfarin is not recommended, the former because of the high cross-reactivity with HIT-IgG and the latter for reports of necrotic complications in acute DIT [33]. Conversely, direct thrombin inhibitors (lepirudin, argatroban, bivalirudin) or heparinoids (danaparoid) should be used. Oral anticoagulation shift toward warfarin should be applied only after platelet count recover to more than 150,000/μL and after an overlap with direct thrombin inhibitors for at least 5 days [33].
8.3 Drug-Induced Anemia
8.3.1 Epidemiology
There are several types of drug–induced anemias, namely immune–mediated hemolytic anemia (IHA), nonimmune-mediated hemolytic anemia, sideroblastic anemia, megaloblastic anemia, and methemoglobinemia.
Unfortunately, exhaustive epidemiological data on drug-induced anemia are available only for some of these forms. The estimated incidence of drug–induced immune hemolytic anemia (DIIHA) is about 1 case per million individuals. However, the real incidence of DIIHA is probably underestimated for some reasons: firstly, because of misdiagnosis with classical autoimmune hemolytic anemia (AIHA), which has been reported to occur in 1 in 80,000 of the population; and secondly, because serological confirmation of DIIHA is usually performed only in the presence of severe hemolysis, this leaving less relevant causative drugs potentially undiscovered [34].
Epidemiological data on the other types of drug-induced anemias are still sparse or totally absent. Megaloblastic anemia is reported to occur in 9 % of patients on continuous treatment with biguanides (phenformin or metformin), and between 3 and 9 % of patients on treatment with methotrexate [35]. Non immune hemolytic anemia is a major side effect of the antiviral ribavirin and it occurs in more than two-thirds of patients with HCV hepatitis treated with the dual combination ribavirin–interferon [36]. Sideroblastic anemia is well-known for some antimicrobials, such as isoniazid and chloramphenicol, the former requiring regular pyridoxine prophylaxis in every antitubercular regimen. Some reports have suggested that linezolid could induce anemia in long-term treatments, even if less frequently than thrombocytopenia [37].
8.3.2 Causative Drugs
Several drugs that are commonly prescribed can induce anemia as ADR [38]. The most notable examples, with the underlying mechanism and with the type of anemia, are reported in Table 8.4.
Table 8.4
General scheme of drug-induced anemia, with relative mechanism of action and main examples
Phenotype of anemia | Mechanism of anemia | Common medications |
---|---|---|
Hemolytic anemia | Immune (DIIHA) | |
DDAB | Cephalosporins | |
Cefotetan | ||
Ceftriaxone | ||
Penicillin | ||
NSAIDs | ||
Quinine/Quinidine | ||
DIAB | Fludarabine | |
Methyldopa | ||
Levodopa | ||
Beta-lactamase inhibitors | ||
Nonimmune | ||
G6PD deficiency | Primaquine | |
Nitrofurantoin | ||
Sulfamethoxazole | ||
Nalidixic Acid | ||
Other | Ribavirin | |
Sideroblastic anemia | Pyridoxine deficiency | Isoniazid |
Pyrazinamide | ||
Copper chelation | Penicillamine | |
Myelosuppression | Chloramphenicol | |
Linezolid | ||
Busulfan | ||
Tetracycline | ||
Megaloblastic anemia | Folic acid deficiency | Methotrexate |
Trimethoprim/Sulfamethoxazole | ||
Pyrimethamine | ||
Sulfasalazine | ||
Triamterene | ||
Anticonvulsants | ||
Phenobarbital | ||
Phenytoin | ||
Primidone | ||
Cobalamin deficiency | Metformin | |
Omeprazole | ||
Variable: sideroblastic/megaloblastic/aplastic anemia | Myelosuppression | Antiviral |
Zidovudine | ||
Didanosine | ||
Stavudine | ||
Lamivudine | ||
Methemoglobinemia | Indirect oxidizers | Dapsone |
Primaquine | ||
Direct oxidizers | Metoclopramide | |
Benzocaine | ||
Prilocaine | ||
Phenazopyridine |
DIIHA has been widely investigated since the first report of suspected hemolysis due to the antiepileptic drug mephenytoin in 1953 [39]. In the 1980s, the most common drugs reported to cause DIIHA were methyldopa (67 %) and penicillin (23 %). Subsequently, the spectrum of implicated drugs has progressively changed and second- and third-generation cephalosporins, in particular cefotetan and ceftriaxone, were reported as responsible for more than 80 % of DIIHA [34].
Differently from what occurred with drug-induced thrombocytopenia and with drug-induced neutropenia, for which well-defined criteria of eligibility for causative drugs have been established, conversely well-defined criteria are still lacking for DIIHA. The most comprehensive update on DIIHA listed 125 drugs on the basis of the number of reports available in the literature and of a positive direct antiglobulin test (DAT), which was performed in almost all cases [40]. Drugs with the highest frequency of DIIHA include second- and third-generation cephalosporins, nonsteroidal anti-inflammatory drugs, in particular ibuprofen and diclofenac, and the chemotherapeutic oxaliplatin [40]. As far as the other type of anemias are concerned, the most important causative drugs for nonimmune-mediated hemolytic anemia are primaquine, sulphamethoxazole; for sideroblastic anemia, isoniazid and pyrazinamide; for megaloblastic anemia, methotrexate, trimethoprim/sulfamethoxazole, phenobarbital, and phenytoin; and for methemoglobinemia, dapsone, benzocaine, and prilocaine [41].
Finally, it is worth noting that some first-generation antiretroviral drugs have been reported as potential causes of dose- and time-dependent macrocytic anemia (zidovudine and to a lesser extent didanosine and stavudine) [42, 43].
Table 8.5 classifies drugs responsible for drug-induced anemia on the basis of the number of reports available in the literature [40].
Table 8.5
Most common drugs causing drug-induced anemia stratified according to the number of references available in the literature
Drug category
Stay updated, free articles. Join our Telegram channelFull access? Get Clinical TreeGet Clinical Tree app for offline access |
---|