Severe thrombosis and acute thrombocytopenia in a cancer patient

Abstract

A 62-year-old Caucasian female who had recently been diagnosed with multiple myeloma with a sacral plasmacytoma received induction treatment with bortezomib, cyclophosphamide, and dexamethasone. She had a significant response to therapy and had just begun the process of mobilization and collection of peripheral blood progenitor cells for autologous stem cell transplantation when she presented to the emergency center (EC) with pain and swelling of the left upper extremity. Her blood pressure on presentation was 114/69 mmHg with a heart rate of 98 beats per minute. Serum sodium was 136 mEq/L, potassium 4.3 mEq/L, chloride 98 mEq/L, carbon dioxide 24 mEq/L, blood urea nitrogen 9 mg/dL, serum creatinine 0.6 mg/dL, glucose 119 mg/dL, calcium 8.6 mg/dL, albumin 3.9 mg/dL, phosphorus 2.5 mg/dL, and magnesium 2.1 mg/dL. Liver function tests were within normal limits, with the exception of lactate dehydrogenase that was elevated at 728 IU/L. Prothrombin time (PT) was 16.5 sec (normal 12.7–15 sec), international normalized ratio (INR) 1.24, and activated PTT (aPTT) was 38.3 sec (normal 24.7–35.9 sec). D-dimer was >20 μg/mL (normal 0–0.4 μg/dL) and fibrinogen was 424 mg/dL. A complete blood count (CBC) showed a white blood cell (WBC) count of 11,400/μL, hemoglobin of 11.6 g/dL, and platelet count of 14,000/μL. Review of the peripheral smear revealed the presence of anisocytosis and ovalocytes, and the absence of fragmented red blood cells (schistocytes). The platelet count had dropped from 316,000/μL to 14,000/μL in a 7-day period (see Fig. 1). Home medications included zolpidem, lorazepam, duloxetine, famotidine, fentanyl, hydromorphone, and ondansetron. Doppler ultrasound study showed a completely occlusive thrombus within the left internal jugular vein extending to the subclavian, axillary, brachial, and basilic veins. The patient had recently had a left subclavian central venous catheter (CVC) placed 11 days prior to presenting to the EC. The CVC was removed, but no anticoagulation was started secondary to severe thrombocytopenia. Clinical course of heparin-induced thrombocytopenia. The graph charts the time course of changes in platelet count after initial heparin exposure. On Day 11 after heparin exposure, magnetic resonance angiography (A: transverse view; B: coronal view; R: right; L: left) showed a lack of flow signal in the left sigmoid sinus, part of the left transverse sinus, and left inferior petrosal sinus. On Day 14, computed tomographic scan of the head (C) showed extensive left cerebral parenchymal hemorrhage causing a mass effect and midline shift. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.] There are many causes for acute thrombocytopenia in patients with malignancy. Chemotherapy-induced myelosuppression and involvement of the bone marrow by malignant cells are among common causes of thrombocytopenia in this patient population, and this patient has myeloma, a malignancy with a predilection to involve bones. Disseminated intravascular coagulation (DIC) caused by infections/sepsis or malignancy is an important differential diagnosis for thrombocytopenia in this group of patients. However, this patient had no signs or symptoms of infection, and the normal fibrinogen and aPTT are not consistent with DIC. Chronic malignancy-induced DIC cannot be excluded, but the acute presentation of thrombocytopenia makes this a less likely diagnosis. Thrombotic microangiopathy (e.g., thrombotic thrombocytopenic purpura) has been reported as a complication of chemotherapy or advanced malignancy. Lack of schistocytes in the peripheral blood smear of this patient made the diagnosis of DIC and thrombotic microangiopathy unlikely. An important cause of thrombocytopenia in patients with malignancy is drug-induced thrombocytopenia. This patient was not started on any new medication in the weeks prior to cause a decline in her platelet count. Presence of deep venous thrombosis in a cancer patient with a CVC is not unusual (about 4% of patients) [1]. Consumptive thrombocytopenia can be in the differential diagnosis. For instance, pulmonary emboli can be associated with about a 20% drop from the baseline platelet counts [2]. The optimal treatment of catheter-related thrombosis is not clear. Treatment of upper extremity catheter-related thrombosis in cancer patients with anticoagulation can allow the line to remain in place [1]. Anticoagulant therapy was contraindicated by thrombocytopenia in this patient; therefore, the CVC was removed. The patient returned to the EC the following day complaining of worsening left arm pain and severe headache and dizziness. Laboratory results were similar to the previous day results with the D-dimer remaining >20 μg/mL, PT was 17.5 sec, INR 1.34, and aPTT 40.5 sec. CBC showed a WBC of 15,100/μL, hemoglobin of 11,200/μL, and platelet count of 22,000/μL. Brain magnetic resonance imaging and magnetic resonance angiography were obtained and showed a lack of flow signal in the left sigmoid sinus, part of the left transverse sinus, left inferior petrosal sinus, and left sphenoparietal sinus (see Fig. 1). Hematology was consulted for diagnosis and management of propagating severe thrombosis in the setting of thrombocytopenia. Heparin-induced thrombocytopenia (HIT) was suspected and a HIT antibody test was ordered, because it was discovered that heparin flushes were started when the patient had her CVC placed. Worsening of patient's symptoms and presence of radiologic and clinical evidence of intracranial venous thrombosis were extremely worrisome and required immediate attention. Despite the presence of severe thrombocytopenia, anticoagulation therapy was indicated at this time [3]. The possibility of HIT was taken into consideration while choosing the anticoagulant. For the initial therapy, an anticoagulant with a short half-life might be preferable in the event of hemorrhagic complications in the setting of thrombocytopenia. Whether or not this patient should receive platelet transfusion was another important question. In a patient with severe thrombocytopenia and an intracranial pathology, full dose anticoagulation could carry a significant risk of bleeding. Would transfusion of platelets alleviate the risk of intracranial hemorrhage or worsen the thrombosis? Platelet transfusion in HIT is controversial. On one side, platelet transfusion is contraindicated in HIT [4, 5] because it may precipitate further thrombosis; on the other side, therapeutic or prophylactic transfusion of platelet in patients with HIT have been performed without any complication and with appropriate rise in platelet counts [6]. Upon admission to the hospital the patient received a transfusion of five units of platelets for thrombocytopenia and was started on a continuous intravenous infusion of argatroban at 2 μg/kg/min. Subsequently, the PF4/heparin antibody titer was reported as strongly positive with an optical density of 3.033, confirming the diagnosis of HIT. The argatroban infusion rate remained within therapeutic range for the next 48 hr, and the platelet count gradually began to rise (see Fig. 1). The patient was stable except for a complaint of persistent headache. WBC was 4,800/μL, hemoglobin of 9,100/μL, and platelet count of 63,000/μL. At that point, a decision was made to transition the therapy from argatroban to fondaparinux, and one dose of fondaparinux 7.5 mg was administered subcutaneously. The argatroban infusion was discontinued 1 hr after the subcutaneous injection. Argatroban is a highly selective direct thrombin inhibitor (DTI) approved for treatment of HIT that reversibly binds to the active site of thrombin and exerts its anticoagulant effects by inhibiting thrombin-catalyzed or -induced reactions, including fibrin formation, activation of factors (V, VIII, and XIII), activation of protein C, and platelet aggregation. It is currently FDA approved for the prophylaxis or treatment of thrombosis in patients with HIT and for patients with or at high risk for HIT undergoing percutanous coronary intervention. Argatroban has been shown to significantly improve clinical outcomes such as all-cause death, all-cause amputation, or new thrombosis compared to historical controls in treatment trials of HIT. Argatroban is administered via continuous infusion and has a half-life (t1/2) of 39–51 min. Argatroban is cleared largely via the biliary system and dose reduction is required in patients with hepatic impairment and in the critically ill. Argatroban prolongs aPTT, PT, and INR. Since argatroban can falsely elevate the PT/INR, monitoring of aPTT is imperative in monitoring efficacy and safety of argatroban [7]. HIT-related antigen production depends on the molecular weight and length of polysaccharide residue present in heparins (>2.4 kDa and >10 saccharide units respectively). Fondaparinux, a novel selective factor Xa inhibitor, is considered a pentasaccharide with molecular weight of 1.7 kDa and five saccharide units. Studies have suggested that fondaparinux saccharide chain is too short to induce an antibody response; therefore, it can be used in the treatment of HIT-associated thrombosis [8]. Although fondaparinux is FDA approved for the prevention and the treatment of venous thromboembolism, it does not have an FDA approved indication for the treatment of HIT. Despite the presence of several case reports and case series in the literature indicating successful treatment of HIT-associated thrombosis with fondaparinux, two case reports have hypothesized the induction of PF4/heparin antibodies associated with fondaparinux [9]. One prospective open-label trial of 7 patients with diagnosis of HIT treated with fondaparinux and 10 historical control HIT patients treated with lepirudin and argatroban, found no different in platelet recovery; however, more successful bridging with warfarin was achieved in fondaparinux group [10]. The safety and efficacy of fondaparinux in the acute phase of HIT is unknown. Recently, the American College of Chest Physicians suggested that warfarin not be initiated in HIT until the platelet count has recovered to at least 150 × 109/L, and that the transition from therapeutic doses of a non-heparin anticoagulant to warfarin is easier with fondaparinux than with DTI's such as argatroban, lepirudin, or bivalirudin. This is because fondaparinux does not significantly interfere with the INR, whereas DTI's prolong the INR. Fondaparinux can be stopped after a minimum of 5 days overlap with warfarin if the INR has reached therapeutic range [5]. To date, there are no randomized clinical trials comparing DTIs versus factor Xa inhibitors in the treatment of HIT. Further research is required to investigate role of fondaparinux in HIT. The following day the patient was found comatose and transferred to the intensive care unit. A computed tomographic scan of the head revealed a large left cerebral parenchymal hemorrhage causing mass effect and midline shift (see Fig. 1). Despite treatment with mannitol and transfusions of platelets, fresh frozen plasma (FFP), and a 2-mg intravenous dose of recombinant activated factor VII (rFVIIa), the patients' neurological status did not improve, and the patient passed away shortly afterward. As the unfortunate hospital course of this patient demonstrates, anticoagulation carries the risk of catastrophic intracerebral bleeding. The patient had no evidence of renal dysfunction, which may have resulted in the accumulation of fondaparinux and measurement of antifactor Xa level was not completed because the patient had only received one dose and was not at steady state. Management of a bleeding patient on anticoagulation comprises of general measures and using specific antidotes [11]. General measures include transfusion of blood products (platelet, red blood cells, and FFP). There is no specific antidote available for fondaparinux; however, rFVIIa has been shown to normalize the coagulation parameters in patients receiving fondaparinux [12]. The severe complications that this patient experienced during her hospital course stemmed from the administration of heparin flushes. Although most consider heparin flushes to be benign, our case illustrates that fatal consequences can result from very small doses of this medication. Heparin is commonly used to maintain the patency of indwelling CVCs in cancer patients. A rare but serious consequence of heparin administration is HIT. This immune-mediated adverse event is usually associated with a platelet count drop of 50% or more from baseline about 5–10 days after the initiation of heparin. HIT is caused by heparin-dependent antibodies that act against complexes of platelet factor 4 (PF4) bound to heparin. Formation of these immune complexes leads to platelet activation, thrombin generation, and formation of thrombi [13]. Patients with HIT, regardless of the presence or absence of thrombosis should receive non-heparin anticoagulants [5]. Anticoagulation with DTIs (lepirudin, argatroban, or bivalirudin) or fondaparinux are among possible therapeutic options. The type of anticoagulant and its dose should be tailored to patient's kidney and liver function status [5]. Warfarin is contraindicated as the initial anticoagulant for treatment of HIT and must be avoided until platelet count is recovered on DTIs. Venous limb gangrene, a devastating limb necrosis, is associated with warfarin administration during acute HIT. Warfarin can accelerate the progressive microvascular thrombosis by depletion of protein C (a natural anticoagulant). Therefore, warfarin use in acute phase of HIT should be avoided until significant resolution of thrombocytopenia [14]. The reported incidence of HIT related to heparin flushes is unknown, but there is enough data available to alert clinicians of the potential serious sequela from small doses of heparin. Nand et al., reported that out of 108 patients diagnosed with HIT at their institution, 15 patients only received heparin flushes prior to diagnosis [15]. Several case reports have been published describing HIT secondary to heparin flushes to maintain catheter line patency [16-19]. One report describes thrombocytopenia that was not evident until larger doses of heparin were administered subsequently after a thrombotic complication [16]. Kadidal et al. described case reports of three cancer patients, with both clinical and laboratory confirmation of HIT secondary to daily heparin flushes of CVCs over a 2-year period. Each of the patients had complex presentations, which delayed the recognition of HIT [17]. Interestingly, a study by Gettings et al. showed that heparin flushes were the most common cause of HIT a group of critically ill postoperative patients [19]. A similar case to our patient was reported by Refaai et al. in which a patient who only received 3 days of heparin flushes developed delay-onset HIT with a nadir platelet count of 43 × 109/L complicated by deep vein thrombosis, pulmonary embolism, and cerebral vein thrombosis. The patient was started on lepirudin and transitioned to warfarin after platelet count recovery [20]. Another report describes a patient with a nadir platelet count of 7 × 109/L who developed cerebral thrombosis from HIT after one dose of 5,000 units of unfractionated heparin. The authors choose not to anticoagulate the patient secondary to severe thrombocytopenia and the patient subsequently developed additional thrombotic complications. In conclusion, the authors mention that anticoagulation should be considered even in the presence of severe thrombocytopenia [21]. Although it is difficult to determine the true incidence of HIT secondary to heparin flushes, it remains a serious medical concern due to the widespread use of heparin. CVCs are a risk factor for the development of venous thromboembolism. Peripherally inserted central venous catheters (PICCs) carry a greater risk of thrombosis compared to subclavian vein or internal jugular vein lines [5]. Occlusion of catheters occurs mainly during the first weeks after the CVC is inserted. Many institutions have evaluated the use of heparin flushes, and opted for a seemingly safer alternative, normal saline. An evaluation of complication rates of PICCs at a single institution showed that the rate of vein thrombosis and line occlusion was 3.4 and 4%, respectively. Interestingly, the lines were only flushed with normal saline once weekly or after intermittent use [22]. Four other trials, two in pediatric cancer patients [23, 24], one in apheresis [25], and one in critically ill patients with triple lumen CVCs [26] showed no differences in catheter patency between use of heparin flushes and saline flushes. Rabe et al. did show a significant difference in catheter patency with heparin vs. saline; however, the dose of heparin used in this trial was 5,000 units [27]. The authors mention that a subgroup of patients receiving heparin flushes at the standard dose of 200 IU/mL did not differ significantly from the normal saline group. The use of normal saline flushes appears to be as effective as heparin in maintaining peripheral line patency [28], but there is still a lack of large randomized controlled trials demonstrating the equivalence of normal saline flushes to heparin flushes in CVCs. Heparin flushes put patients at risk for HIT, bleeding complications, medication dosing errors, drug interactions, and are associated with a higher cost of therapy. Based on these potential complications, we feel that the use of normal saline flushes should be considered for all patients with CVCs. Most of the data for prevention of catheter-related thrombosis deal with systemic thromboprophylaxis rather than line flushing; however, may studies have failed to prove a difference in outcomes. The 8th Edition of the American College of Chest Physicians guidelines on the prevention of venous thromboembolism recommends that neither prophylactic doses of LMWH nor low-dose warfarin should be used to prevent catheter-related thrombosis [5]. Our patient case illustrates that even small doses of heparin can result in potentially fatal consequences. The question of whether or not anticoagulation is appropriate for patients with HIT and severe thrombocytopenia remains unanswered. The patient we present did suffer a fatal consequence of anticoagulation; however, the platelet count had recovered significantly prior to this event. It appears there is no clinically significant difference between heparin flushes and normal saline flushes in maintaining CVC patency; therefore, we recommend avoiding heparin flushes if possible to prevent the unfortunate consequences that occurred in our patient case.