Pathology Consultation on Drug-induced Hemolytic Anemia

Arand Pierce, MD; Theresa Nester, MD


Am J Clin Pathol. 2011;136(1):7-12. 

In This Article


Drug-dependent Antibodies

Most drugs that cause hemolysis are mediated by DDABs. In the 30-year experience of 1 reference laboratory, cefotetan, ceftriaxone, and piperacillin were, in aggregate, responsible for 76% of all cases of DIIHA, with cefotetan accounting for the majority of cases.[3] Hemolysis typically appears within 2 weeks after starting the drug, and the patient has progressive anemia or other evidence of hemolysis. There is often no clinical suspicion of drug-mediated hemolysis because a routine request for RBC transfusion or preoperative screen typically initiates the evaluation. A positive DAT is the most reliable laboratory finding in DIIHA; if the DAT is negative, the likelihood of DIIHA is exceedingly small.

Following a positive DAT, an elution is performed to characterize the antibody coating the patient's RBCs. The RBCs are washed free of unbound antibody, followed by chemical modification (usually acid) to remove the attached antibodies; the resultant eluate is then characterized by testing against reagent RBCs. The eluate fails to react because the drug is not present in this in vitro testing. A careful clinical history, including medications and their temporal relationship with the onset of hemolysis, is critical because a positive DAT with a negative eluate is commonly associated with the passive transfer of ABO antibodies,[4] such as anti-A or anti-B from a prior out-of-group plasma or platelet transfusion,[5] intravenous immunoglobulin (IVIG) therapy,[6] or hemolytic disease of the newborn and fetus.[7] Nonimmunologic RBC adsorption of IgG due to hypergamma-globulinemia[8,9] is another cause to consider besides DDABs.

Because significant concentrations of residual drug or drug-antibody complexes may be present in the patient's serum,[10–12] DDABs will frequently react against non–drug-treated RBCs.[11] A broadly reactive antibody, including against the autocontrol, is found in half of cases,[12,13] suggestive of an autoantibody. Anticefotetan and anticeftriaxone antibodies may exhibit this reactivity.[12,14] This also occurs with piperacillin antibodies,[12,15] which may further mimic an autoantibody by showing an apparent preference for Rh(e)+ reagent cells.[16–19] These findings may be inconstant because they are highly contingent on the plasma drug concentration; thus, relating the laboratory results to the dosage and timing of administration is critical.

If the autocontrol is negative and all other panel RBCs are positive, an antibody against a high-prevalence antigen should alternatively be considered. A subsequent elution may help distinguish between WAIHA and DIIHA, but unfortunately, rare cases of DDABs cause positive elutions. This is more common with cefotetan and ceftriaxone[12] and with piperacillin[20] because they are most frequently positive in initial indirect antiglobulin test (IAT) testing, as described. When positive, the eluate in DIIHA typically reacts weakly compared with serum at initial antibody testing,[10] a clue to consider DIIHA rather than WAIHA. The clinical risk for a patient with DIIHA is that the finding of reactive antibodies in the serum or eluate will lead to an incorrect diagnosis of WAIHA and improper treatment with steroids but worse, continued administration of the offending drug that may prolong or worsen hemolysis. The only reliable confirmation of DDAB-mediated DIIHA over WAIHA requires testing patient serum after drug cessation and clearance of any circulating drug or drug-antibody complexes, indicated by a clear decrease in serologic reactivity. Figure 1 depicts the basic testing algorithm for diagnosing DIIHA, always keeping in mind that the drug history is key to this investigation.

Figure 1.

Algorithm for the evaluation of DIIHA. DAT, direct antiglobulin test; DIIHA, drug-induced immune hemolytic anemia; HDFN, hemolytic disease of the fetus and newborn; IVIG, intravenous immunoglobulin; NIPA, nonimmunologic protein adsorption; WAIHA, warm autoimmune hemolytic anemia.

Antibodies may be directed against the drug, its metabolites, a neoantigen formed by the drug and RBC membrane proteins, or any combination thereof.[21,22] The best-understood mechanisms of hemolysis involve penicillin and cefotetan, which covalently bind to RBC membranes, provoking IgG antibodies directed only to drug epitopes.[22] The coated RBCs undergo extravascular destruction via Fc-receptor recognition by splenic macrophages. Because drug-RBC binding is covalent, the drug may be bound firmly enough to RBCs to prevent clearance from the circulation. Extravascular hemolysis (albeit mild) may persist for months after drug cessation in this situation due to DDABs acting against the "stabilized" antigen. In vitro testing of these drug-specific antibodies is characterized by serum reactivity against drug-treated RBCs. It is interesting that many case reports of cefotetan and DIIHA involve a single surgical prophylactic dose in a patient who has never received the drug.[23] It is speculated that this is a sequela of widespread antibiotic use in US livestock[15,24] and supported by the high prevalence of anticefotetan antibody in US blood donors.[25] A single dose of prophylactic antibiotic is often overlooked when investigating a patient's drug history (especially when given in the operating room).

Most other drugs, including ceftriaxone and piperacillin, do not bind as strongly to RBC membranes, such that drug-coated RBCs are difficult to create for in vitro testing. These DDABs may interact noncovalently with the RBC membrane, forming a neoantigen variably composed of RBC membrane proteins and drug epitopes. These antibodies are associated with higher morbidity and mortality because they can fix complement and provoke precipitous intravascular hemolysis.

Ceftriaxone causes robust hemolysis with a disproportionately high fatality rate in pediatric patients (50%–100%), whereas adults are not as severely affected and fatality rarely occurs.[15,26] Piperacillin is frequently prescribed for patients with CF, and the associated DIIHA disproportionately affects this patient population; in fact, the only reported drugs associated with DIIHA in patients with CF are piperacillin and piperacillin-tazobactam.[18] Unlike cefotetan, all cases of piperacillin-associated DIIHA have been associated with prior drug exposure.

The treatment of DDABs DIIHA is straightforward: stop the drug. Hematologic recovery should occur within 2 weeks, although the DAT may remain positive for months. If WAIHA is presumed by the clinicians and treated with steroids, the patient is being placed at unnecessary risk because there is no compelling evidence that steroids are effective in treating DIIHA with DDABs. If the patient is immunosuppressed or infected, this is an especially important point for discussion with clinicians to avoid the potential harm of unnecessary steroid administration.

Drug-independent Antibodies

DIIHA is less commonly associated with DIABs. Most DIABs are associated with the generation of a true autoantibody-mediated WAIHA, often against a Rhesus antigen if any specificity is discernible.[27] Methyldopa, which is now used to treat gestational hypertension, was the historic cause of drug-induced WAIHA; fludarabine now accounts for the majority of drug-induced WAIHA among all drugs (Table 1). The mechanisms of DIAB formation are poorly understood and include molecular mimicry, drug adsorption causing altered RBC membrane protein antigens, and immune dysregulation. Because these drug-independent immunologic processes are essentially identical to WAIHA, in vitro testing cannot discriminate this type of DIIHA from WAIHA. The workup will show a positive DAT, positive antibody screen, and positive elution. Only hematologic improvement after stopping the drug can confirm the diagnosis of DIAB-mediated DIIHA. However, steroids are indicated for this type of DIIHA, as for any WAIHA; if hemolysis is persistent or severe, treatment with IVIG or immunosuppression should be considered, as for any non–drug-associated WAIHA.

Fludarabine is frequently used to treat chronic lymphocytic leukemia (CLL); thus, it may be impossible to differentiate a primary WAIHA associated with CLL from a DIIHA secondary to fludarabine. There is good evidence that fludarabine is independently causative of WAIHA in CLL,[28] and such cases should be scrutinized carefully because specific medical management may control or prevent hemolysis.[29]

A second drug-independent mechanism is gaining recognition: Nonimmunologic protein adsorption has been implicated in hemolysis associated with β-lactamase inhibitors, platinum-based chemotherapeutic agents, and cephalosporins, including cefotetan (Table 1). These drugs may alter the RBC membrane such that serum proteins (including immunoglobulins) are nonspecifically adherent, causing "spurious" positive DATs due to "bystander" antibodies present on the RBC membranes. This effect was previously considered to cause false-positive in vitro testing results, but there is now evidence that these adherent proteins may truly mediate RBC destruction in certain instances.[30–32] Of course, because β-lactamase inhibitors are commonly administered with antimicrobials such as piperacillin (eg, piperacillin-tazobactam), multiple hemolytic mechanisms may be operative.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: