Oxaliplatin-Related Thrombocytopenia

D. L. Jardim; C. A. Rodrigues; Y. A. S. Novis; V. G. Rocha; P. M. Hoff


Ann Oncol. 2012;23(8):1937-1942. 

In This Article


We searched Medline (1998–2008), UpToDate database and Embase for case reports and up-to-date manuscripts about oxaliplatin and thrombocytopenia. We used the search terms "oxaliplatin" or "chemotherapy, colon cancer" in combination with the terms "thrombocytopenia", "myelossupression", "bone marrow suppression", "allergic reactions", "auto-immune reactions" "splenic sequestration" and "hemolytic uremic syndrome". We mostly selected publications from the past 6 years but did not exclude frequently referenced older publications or case reports. We also searched the reference lists of reports identified by this search strategy and selected those we judged relevant.

After reviewing all the data, the authors organized a practical review classifying the information by the different mechanisms of thrombocytopenia.

Thrombocytopenia Due to Sinusoidal Injury by Oxaliplatin

Prolonged thrombocytopenia and splenic enlargement can be observed in some patients receiving oxaliplatin-based treatments and a hypothesis to explain this association has been recently published.[7] Clinical and pathological analyses following oxaliplatin-based chemotherapy regimens used as a neoadjuvant treatment for resection of liver metastasis showed an association between hepatic sinusoidal injury, portal hypertension and thrombocytopenia.[8]

Oxaliplatin-induced sinusoidal injury was described as a disruption of the sinusoidal barrier leading to subsequent deposition of collagen in the perisinusoidal space and veno-occlusive fibrosis.[9] Some degree of sinusoidal injury can be observed in up to 78% of patients who receive oxaliplatin and these hepatic changes appear to be drug-specific, since patients who receive other agents, such as fluorouracil, or chemo-naive patients are clearly less affected by sinusoidal injury.[9,10] Perisinusoidal fibrosis and veno-occlusive lesion induced by oxaliplatin in the normal liver can lead to portal hypertension and its complications, including esophageal or hemorrhoidal varices with bleeding, ascites and splenomegaly with associated thrombocytopenia.[10] Studies with a series of patients indicate a median of 12 cycles of oxaliplatin before evidence of portal hypertension develops.[8]

The development of splenomegaly following oxaliplatin-based regimens is clearly more frequent in comparison to fluorouracil, and the splenic index (SI) can be a surrogate marker for portal hypertension.[11] The reported rate of spleen size increase after oxaliplatin treatment is up to 86%[7] and the mean increase in SI is 45.7%.[11] Single-agent fluoropyrimidine-treated patients have a mean increase in SI of only 16.3%, which is significantly lower. Increases in spleen size were directly correlated with the cumulative amount of oxaliplatin administered, and the degree of thrombocytopenia is significantly related to the increase in spleen size. The group from M. D. Anderson was able to detect thrombocytopenia in 28% of patients who developed splenomegaly and in only 5% of patients who did not.[7] As increases in spleen size also correlate with a histopathological grade of sinusoidal injury in this study, it is logical to view splenomegaly as a marker of portal hypertension secondary to oxaliplatin-induced sinusoidal injury and oxaliplatin-related thrombocytopenia resulting from splenic sequestration of platelets.

Oxaliplatin-related hepatic injury is associated with increased postsurgical morbidity, greater transfusion requirements and longer hospital stays.[10] A low-preoperative platelet count in this context and a high aspartate aminotransferase to platelet ratio index score were shown to be good indicators to predict sinusoidal obstruction syndrome severity.[12] Therefore, detection of thrombocytopenia in the preoperative setting after oxaliplatin chemotherapy might have a great clinical relevance, especially if concomitant with splenomegaly.

Thrombocytopenia due to splenic sequestration presents as a moderate prolonged thrombocytopenia after a median of 18 weeks of oxaliplatin-based chemotherapy.[7] The mean platelet count after splenic sequestration due to chemotherapy is 81 × 109/l,[13] and patients rarely present hemorrhagic manifestations. Splenomegaly may be associated with other features of portal hypertension, such as ascites and esophageal and hemorrhoidal varices, which may eventually lead to associated bleeding. Bone marrow aspiration should not be carried out routinely, but, if carried out, it will probably show normal trilineage hematopoiesis.[13] Imaging studies of the spleen have probably been already obtained for cancer evaluation but should be requested if not carried out before. Absence of splenomegaly should direct clinical judgment to other causes of thrombocytopenia, including bone marrow suppression and drug-induced immune thrombocytopenia (DIIT).

The management of these patients depends on the urge to continue systemic treatment. After discontinuation of oxaliplatin, platelet counts improve with time, although platelet recovery is slower for patients with associated splenomegaly. In this context, 23% of patients will maintain thrombocytopenia in the first year after completion of chemotherapy and platelet counts approach the baseline levels in ~2–3 years after oxaliplatin cessation.[7] There is no clinical treatment of oxaliplatin sinusoidal injury, although there are some suggestions that association of bevacizumab to the regimen might decrease the development of this complication.[7] When the administration of systemic therapy is needed and is precluded because of thrombocytopenia related to splenic sequestration, partial splenic embolization (PSE) is a safe and effective means to produce a rapid increase in platelet count. A report of the experience with PSE has shown an increase in mean platelet count from 81 × 109/l to 293 × 109/l and a mean time to a platelet count higher than 150 × 109/l of 10 days.[13] Most of the patients were submitted to 5-fluorouracil and oxaliplatin regimens before the procedure and the average time to resume systemic treatment was 32 days. However, the indication of PSE should be carefully considered since the platelets of these patients are usually normofunctional, and platelet counts >75 × 109/l should not preclude chemotherapy administration. Besides, PSE involves some grade of morbidity, such as infectious complications and prolonged hospital stays.

Other manifestations of portal hypertension should be screened and appropriately managed, especially through an upper gastrointestinal endoscopy to detect esophageal varices. Prophylactic use of nonselective beta-blockers could be considered when splenic enlargement and clinical consequences of portal hypertension are detected. Discontinuation of oxaliplatin also can be considered if high-risk varices or acute variceal bleeding develops.

Oxaliplatin-induced Immune Thrombocytopenia

Many drugs have been implicated in development of immune-induced thrombocytopenia, in which drug-dependent antibodies react to specific platelet membrane glycoproteins mediating accelerated platelet destruction. Although oxaliplatin more typically causes thrombocytopenia by suppressing hematopoiesis, this drug can also cause immune thrombocytopenia. There are a few case reports that describe this mechanism, and its exact incidence is not known. Immune-mediated thrombocytopenia has been identified in 3 of 42 patients presenting allergic reactions to oxaliplatin,[14] and considering that ~10% of patients exposed to oxaliplatin experience allergic reaction,[1] we can suppose that oxaliplatin-induced immune thrombocytopenia (OIIT) affects <1% of patients. It is noteworthy that this estimate is probably an underestimation since patients may present with immune-mediated thrombocytopenia in the absence of an allergic reaction.

The basic mechanism responsible for OIIT is the induction of antibodies that bind tightly to normal platelets only in the presence of the drug. The exact immune pathway mediating platelet destruction is not completely understood, but current hypotheses suggest that naturally occurring antibodies with weak affinity for platelet membrane antigens may be involved. In the presence of certain drugs, such as oxaliplatin, these antibodies can increase the strength of their interaction with such antigens, resulting in platelet destruction and thrombocytopenia.[15] Oxaliplatin-dependent antibodies against glycoprotein IIb/IIIa complex have been most commonly involved in immune-mediated thrombocytopenia,[6] although antibodies against other platelet surface glycoproteins, such as GP Ia/IIa and GP Ib/IX, have also been identified.[16,17] These antibodies react strongly with platelets only in the presence of oxaliplatin and weakly in the presence of cisplatin but neither react in the presence of carboplatin, 5-FU or folinic acid, indicating specificity for oxaliplatin.[6]

The main clinical characteristics of the oxaliplatin-induced thrombocytopenia case reports were reviewed and are summarized in supplemental Table S1 (available at Annals of Oncology online). The usual presentation is an acute and severe drop in platelet count leading to clinical bleeding or bruising within a few hours, but sometimes in up to 48 h, after the administration of a treatment cycle of oxaliplatin. The nadir of platelet count can be as low as 2 × 109/l. This complication most commonly affects female patients with advanced CRC and prior oxaliplatin exposure, usually occurring during retreatments. The majority of cases describe this complication after at least a total of 12 cycles of oxaliplatin. In some cases, thrombocytopenia can follow the onset of hypersensitivity reactions, such as skin rash, pruritis, chills and brochospasm.[5]

Although isolated thrombocytopenia is the most common finding in blood cell counts, acute hemolysis can be present (Evans syndrome). In these situations, the patient can also present back pain during infusion, chills and fever after each cycle, and the direct antiglobulin test is usually positive.[18–24] Oxaliplatin-induced thrombocytopenia can occur in the rare context of acute immune pancytopenia, which was described in two case reports.[16,25] In these reports, acute thrombocytopenia, hemolysis and neutropenia were mediated through production of oxaliplatin-dependent antibodies to platelets, red blood cells and neutrophils, respectively.

The clinical diagnosis of drug-induced thrombocytopenia is made based on established criteria.[26] Unlike others drugs that are used continuously, the causal relationship of oxaliplatin-induced thrombocytopenia is easily identified due to its intermittent use in the chemotherapy cycles, and a rechallenge is neither necessary nor recommended. Definitive laboratorial diagnosis can be established by detection of oxaliplatin-dependent antibodies in patients' sera carried out by flow cytometry. Binding of an IgG antibody is detected when the patient's serum is incubated with normal group O platelets in the presence of oxaliplatin and the fluorescence intensity is at least twice that of a normal control.[6] The reaction does not happen in the absence of the drug. This test in not generally available at the time of presentation, but it may be helpful to document oxaliplatin-induced acute thrombocytopenia and to discontinue the drug. When the test is not available, clinical judgment may be sufficient to establish the diagnosis. Additional tests such as peripheral blood smear examination, direct antiglobulin test and detection of antibodies to neutrophils are also helpful in assessing immune pancytopenia and Evans syndrome. Bone marrow aspirates, if carried out, will probably show the presence of adequate or increased numbers of megakaryocytes.[17]

There is no specific treatment for most of the patients with OITT. Platelet count usually recovers in a few days after discontinuation of the oxaliplatin, which is the main therapeutic maneuver. The patient should not receive oxaliplatin again and should carry a medical alert to remember this fact, since drug sensitivity usually persists indefinitely.[6] Subsequent treatments with other chemotherapy agents and biologic agents are possible. For instance, some case reports have shown that the administration of 5-FU, folinic acid, irinotecan and cetuximab could be carried out without consequences. For patients with major bleeding, appropriate hematologic support, including platelet transfusion, may be necessary to keep the platelet count at safe numbers. A posttransfusion platelet increase should be checked in ~1 h after the procedure to ensure that a satisfactory enhancement has occurred.

The value of steroids in this context is uncertain. The approaches described in previous cases varied between no steroid administration to pulses of methylprednisolone.[17] Whether steroids altered the clinical evolution of these cases is unknown. Since idiopathic immune thrombocytopenic purpura cannot initially be distinguished from OIIT, steroids may be justified. One case report describes a steadily increase in platelet counts after administration of i.v. immune globulin.[27] Unlike other causes of oxaliplatin-related thrombocytopenia, recovery is usually fast and complete after discontinuation of the drug, although many patients need repeated transfusions until they can maintain safe numbers on their own.

Thrombocytopenia Related to Oxaliplatin Myelossupression

Oxaliplatin is associated with mild to moderate bone marrow suppression similar to that observed with other platinum compounds. Undoubtedly, this is the main mechanism of oxaliplatin-related thrombocytopenia and has been reported to occur in 45%–77% of patients receiving oxaliplatin-based regimens for the treatment of CRC.[1,2,28] Decreases in platelet counts are more pronounced in patients receiving oxaliplatin as compared with those receiving other regimes containing 5-FU and folinic acid, with or without irinotecan (FOLFIRI). Although thrombocytopenia in this context can occur from direct tumor invasion in the marrow, from myelossupression caused by distant disease or from the combination of these mechanisms, myelossupression of oxaliplatin is assumed to be the main culprit since direct effects of CRC in the bone marrow are not common. The exact mechanism through which chemotherapeutic agents affect platelet production is not completely understood and neither is the reason why oxaliplatin potentially affects megacaryocytes more than other agents. Some investigators speculate that a direct toxicity to megakaryocytic progenitors might induce apoptosis of this population in an agent-sensitive manner.[29]

The patient typically presents platelets counts <75 × 109/l ~10 days after oxaliplatin administration. Symptomatic thrombocytopenia, manifested by petechiae, purpura and other mucosal bleeding such as gingival bleeding or epistaxis, is uncommon since patients rarely present grade 4 thrombocytopenia. Routine blood cell counts usually show concomitant anemia and neutropenia as a result of the bone marrow suppression.

Isolated thrombocytopenia in this context is not common and a falsely low-platelet count must be excluded. The specimen may have been improperly drawn or under-anticoagulated, leading to the presence of small clots and resultant thrombocytopenia. The blood smear must also be examined to exclude the possibility that platelet clumping occurred in an EDTA-containing blood sample. In both cases, it is recommended that the test be repeated in appropriate conditions. If an isolated thrombocytopenia is confirmed, one might consider other causes of oxaliplatin-related thrombocytopenia. Bone marrow aspirates should not be carried out routinely, unless tumor invasion or hematological malignancies are suspected on the basis of the complete blood count. Unlike what is seen with the other mechanisms of oxaliplatin-related thrombocytopenia, the bone marrow aspirates from patients with myelosuppression will show a decrease in the number of all hematological precursors, including megakaryocytes.

Most patients with thrombocytopenia related to bone marrow suppression will require no specific treatment, as their platelets counts will probably recover rapidly before the next cycle of chemotherapy. It may be necessary to delay the programmed treatment if a safe platelet count (75 × 109/l for FOLFOX) is not reached. Patients should be given platelet transfusions in the presence of bleeding manifestations assigned to a very low-platelet count. The threshold for prophylactic platelet transfusion is a common issue for discussion, due to an imprecise concept of a 'safe' platelet count, which may vary according to other conditions such as infection or other hemostatic abnormalities.[30] In the absence of risk factors for hemorrhage, a threshold of 10 × 109 platelets/l might be safely used without an increase in the bleeding risk and has been recommended by the American Society of Clinical Oncology guidelines.[31]

Currently, one area of great interest is the development of thrombopoietic growth factors for primary or secondary prophylaxis of chemotherapy-induced thrombocytopenia. In clinical studies, IL-11 (oprelvekin, Neumega®, Wyeth, Pharmaceuticals Inc, Philadelphia, PA) has modestly reduced the extend of thrombocytopenia and the need of platelet transfusion,[32,33] being the only agent currently approved by the USA FDA for chemotherapy-induced thrombocytopenia. However, the use of this agent is limited by its narrow therapeutic index. Administration of recombinant thrombopoietins (TPOs) to patients receiving chemotherapy with platinum compounds has been shown to induce enhanced platelet recovery,[34,35] but the development of these agents has been abandoned because a high incidence of antibody development leading to thrombocytopenia [36] and some concerns about hematological malignancies following administration of megakaryocyte growth factor to healthy volunteers.[37] First-generation TPO receptor agonists are currently being evaluated for chemotherapy-induced thrombocytopenia, including peptide (romiplostim) and nonpeptide (eltrombopag) mimetics. They might be a future option for oxaliplatin-treated patients, especially in cases when maintaining the dose intensity can translate into better clinical outcomes.

Speculative Mechanisms

Oxaliplatin treatment can lead to acute thrombocytopenia with hemolysis mimicking the hemolytic uremic syndrome (HUS).[38,39] In this context, patients also present acute onset of back pain, dark urine and oliguria. Investigations show an elevation of lactic dehydrogenase and indirect bilirubin levels, a drop in concentrations of haptoglobin and blood films commonly show circulating fragmented red blood cells (schizocytes). Drug-induced thrombotic microangiopathy may be related to an immune-mediated phenomenon involving ADAMTS13 metalloprotease or to a direct endothelial injury, both following certain drug exposures.[40] Thrombotic microangiopathies, such as HUS, can be present in patients with DIIT for reasons that are poorly understood. There are some associations of these complications in patients receiving quinine, a drug known to be related to DIIT.[41,42]

In the absence of oxaliplatin-induced antibodies, one may consider HUS as a distinct mechanism of oxaliplatin-related thrombocytopenia. In clinical practice, the differential diagnosis is troublesome. There are few reports of possible HUS following oxaliplatin administration. In one of them, neither kidney biopsy nor detection of oxaliplatin-dependent antibodies against erythrocytes and platelets were carried out and the diagnosis remains speculative.[38] In the other report, although hemolysis, thrombocytopenia and acute renal failure were present, kidney biopsy did not confirm thrombotic microangiopathy.[39] Fatal HUS has also been reported in a patient with metastatic pancreatic adenocarcinoma following administration of gemcitabine and oxaliplatin.[43] Other chemotherapy agents, such as mitomycin-C and gemcitabine, have been implicated as causal agents of thrombotic microangiopathy in a cumulative dose-dependent manner.[40,44] However, the relation of oxaliplatin to drug-induced thrombotic microangiopathy is not completely established. Therefore, thrombocytopenia in this context could be immune-mediated and a negative oxaliplatin antibody test and a kidney biopsy consistent with HUS are needed to begin treatment and that could include fresh frozen plasma and large-volume plasmapheresis.