A physician or group of physicians considers presentation and evolution of a real clinical case, reacting to clinical information and data.
Solving Clinical Problems
Venous limb gangrene and fatal hemorrhage: Adverse consequences of hit “overdiagnosis” in a patient with antiphospholipid syndrome†
Version of Record online: 21 JAN 2011
Copyright © 2010 Wiley-Liss, Inc.
American Journal of Hematology
Volume 86, Issue 2, pages 188–191, February 2011
How to Cite
Smythe, M. A., Warkentin, T. E., Woodhouse, A. L. and Zakalik, D. (2011), Venous limb gangrene and fatal hemorrhage: Adverse consequences of hit “overdiagnosis” in a patient with antiphospholipid syndrome. Am. J. Hematol., 86: 188–191. doi: 10.1002/ajh.21916
- Issue online: 21 JAN 2011
- Version of Record online: 21 JAN 2011
- Accepted manuscript online: 4 NOV 2010 01:05PM EST
- Manuscript Accepted: 27 OCT 2010
- Manuscript Revised: 25 OCT 2010
- Manuscript Received: 27 SEP 2010
A 58-year-old African-American female presented to the Emergency Department on April 6, 2007 with complaints of acute right upper-extremity (UE) pain and swelling. Past medical history included PE, DVT, and APS confirmed by repeated positive IgG and IgM anticardiolipin antibodies from 1998 through 2002; in July 2006, she also tested positive for nonspecific inhibitor (“lupus anticoagulant”). Other history included idiopathic thrombocytopenic purpura (ITP), autoimmune hemolytic anemia, hypertension, questionable systemic lupus erythematosus (SLE), steroid-induced diabetes mellitus, nonspecific liver dysfunction, and breast cancer (12 years prior). There was also a history of possible heparin-induced thrombocytopenia (HIT) based upon serologic investigations in a low pretest probability clinical setting of chronic thrombocytopenia that bore no clear relationship to preceding heparin therapy: (a) June 2006, 4Ts score = 0, enzyme-immunoassay (EIA) optical density (OD) = 1.763 U [normal < 0.400 U] by polyspecific assay that detects IgG, IgA, and IgM (GTI-PF4; GTI Diagnostics Inc., Brookfield, WI); and (b) October 2006, 4Ts score = 3, EIA OD = 2.304 U. Notably, both “positive” EIAs were reported as “equivocal” because high heparin concentrations failed to inhibit the positive reaction by ≥50% during the “confirmation” step. However, a platelet serotonin-release assay (SRA) in July 2006 was reported as “positive,” based upon 98% serotonin-release at low dose heparin and 12% serotonin-release at high dose heparin. She was on fondaparinux 10 mg daily as an outpatient since July 2006. The baseline platelet count before fondaparinux was 88,000/mm3 (normal: 150,000–400,000/mm3). The patient reported discontinuing therapy on her own accord 2 days prior to admission; additionally, poor compliance was suspected at times throughout therapy. Admission laboratory data included INR 3.0, aPTT 69 sec (using Actin FSL reagent, Dade Behring), platelet count 45,000/mm3, a hemoglobin of 9.9 g/dl, alkaline phosphatase 572 U/l, AST 100 U/l, ALT 99 U/l, total bilirubin 2.9 mg/dl, BUN 16 mg/dl, and serum creatinine 1.2 mg/dl (these results were similar to those 7 months earlier). An UE Doppler ultrasound showed DVT with the appearance of acute-on-chronic disease involving the right brachial veins.
This patient presents with acute DVT and thrombocytopenia 2 days after fondaparinux withdrawal. The history revealed questionable antibodies to platelet factor 4/heparin (PF4/H) complexes detected approximately 6 and 10 months earlier; moderately-to-strongly positive OD values of 1.763 and 2.304 by EIA reported as “equivocal” after the high-heparin “confirmatory” maneuver failed to inhibit reactivity. The confirmatory procedure is recommended on the presumption that it enhances diagnostic specificity . Indeed, platelet activation induced by HIT antibodies is inhibited in the presence of high heparin concentrations, which disrupt formation of the multimolecular PF4/H complexes . The failure to inhibit the reaction by ≥50% suggests that the antibodies might not be directed against PF4/heparin complexes, and therefore might not be “true” HIT antibodies. However, the reliability of this confirmation step to distinguish clinically-significant HIT antibodies is controversial, as patients with clinically-evident HIT and strongly-positive OD results may not show the requisite inhibition to “confirm” HIT [3, 4]. Moreover, this patient had a positive SRA. Although SRA specificity greatly exceeds that of the EIA, false-positive results can occur . The decision to perform HIT antibody testing was debatable as the 4Ts probability score for HIT was low (≤3 points) and such low scores are highly predictive (>98%) of a negative SRA .
Identifying the presence of true HIT in a patient with APS is challenging as the clinical presentation for both can involve the duad of thrombocytopenia and thrombosis. Further, APS patients often exhibit false-positive HIT serology, particularly with antigen assays . Pauzner et al. reported positive IgG EIA results for the PF4/H complexes in 12 of 14 patients with APS in whom functional assay testing for HIT was negative. The reaction was not inhibited in the presence of high dose heparin. The failure to inhibit is linked to the presence of antibodies to PF4 as opposed to antibodies against PF4/H complexes. Antibodies to PF4 are believed to contribute to the overdiagnosis of HIT in many patients with APS. Sikara et al. recently demonstrated that β2-glycoprotein I (the antigenic target of antiphospholipid antibodies) binds to PF4 forming stable complexes . This interaction is believed to be involved in the formation of the hypercoagulable state in APS.
Another consideration is the possibility of fondaparinux-induced HIT, which is thought to be very rare, limited to case reports at present [10–12]. However, the late occurrence of thrombocytopenia and its onset after stopping fondaparinux are not consistent with this diagnosis. Fondaparinux withdrawal in a patient with an underlying APS-associated hypercoagulable state is the most likely cause of the thrombosis and thrombocytopenia seen here.
She was admitted with UE DVT and argatroban 0.4μg/kg/min was started (hospital day 0). The pretreatment aPTT was 69 sec (1 month earlier it had been 85 sec). Warfarin 5 mg was also given upon admission. Approximately 5 hrs after argatroban initiation (and 2 hrs after receiving warfarin), argatroban was held for an aPTT of 102 sec and restarted 8 hrs later at 0.1 μμg/kg/min. Over the next 24 hrs, argatroban therapy was increased to 0.2 μμg/kg/min for an aPTT value of 54 sec.
Despite receiving argatroban, the patient developed progressive thrombosis, as shown by right UE swelling, with tight skin, and limb cool to touch from her mid-forearm distally; radial pulses remained present. On day 2, the INR rose to 3.8 and the patient developed a compartment syndrome. Argatroban was held and emergent fasciotomies of the right UE were performed. On day 3, fondaparinux 10 mg sq daily was started because of concerns of progressive right hand ischemia. A PF4/H-EIA was positive (OD, 1.587 U; confirmation step, 59% inhibition with high-dose heparin); however, the SRA yielded an “indeterminate” test result: 92% serotonin-release at low-dose heparin and 88% release at high-dose heparin (i.e., the confirmation step failed to demonstrate inhibition of platelet activation at high heparin concentrations). On day 3, two platelet doses were administered which raised the platelet count from 21,000/mm3 to 68,000/mm3. On day 4 (i.e., 3 days postwarfarin and while continuing off argatroban), the INR increased further to 4.7; the INR continued to increase through day 13 to 9.9. Vitamin K was never administered. Fondaparinux was discontinued and argatroban was restarted on day 4. A repeat UE Doppler on the right revealed DVT involving the proximal and distal subclavian, axillary, and brachial veins. Superficial vein thrombosis (SVT) was seen involving the right cephalic vein. On the left side, DVT was evident involving the distal subclavian, axillary, and brachial veins. SVT was also seen involving the left basilic vein.
As a result of the questionable history of HIT, and the exacerbation of thrombocytopenia, argatroban-rather than unfractionated heparin, low-molecular-weight heparin, or fondaparinux–was initiated. Unlike other parenteral anticoagulants, data supporting the efficacy of argatroban is limited to patients with suspected or confirmed HIT . In retrospect, argatroban may not have been a good choice of anticoagulant, as the diagnosis of HIT seemed unlikely, and literature supporting antithrombotic benefit of argatroban in non-HIT settings is sparse.
As a result of hepatic dysfunction and a prolonged baseline aPTT, the argatroban infusion was initiated with a low dose. The prolonged aPTT, which likely reflected the combination of APS and nonspecific hepatic dysfunction, presented a challenge with respect to argatroban monitoring. On admission, our patient's baseline was 69 sec—within the usual “therapeutic” range already before argatroban therapy. Using the aPTT to guide dosage adjustments in situations involving a prolonged baseline can be problematic. Cases of progressive thrombosis in patients with HIT following interruption and dose reduction of DTI therapy secondary to high aPTT values have been described; efforts to maintain a “therapeutic” aPTT (i.e. infusion holds and dose reductions) were thought to have resulted in subtherapeutic anticoagulation and, potentially, in rebound hypercoagulability as a result of the short half-lives of argatroban and lepirudin [13, 14] This scenario is similar to the case reported here, in that progressive thrombosis ensued following interruption of argatroban therapy for an elevated aPTT.
In our case, initiation of warfarin therapy likely further interfered with the ability to use the aPTT to monitor argatroban. Warfarin independently prolongs the aPTT . Warfarin-induced aPTT prolongation has been associated with DTI underdosing and progressive thrombosis . As a result, in patients with newly diagnosed HIT who are receiving warfarin, reversal with vitamin K is recommended . Although only one dose of warfarin was given to our patient, this may have contributed to the elevated aPTT. Warfarin is also a major risk factor in the pathogenesis of venous limb gangrene (VLG) complicating hypercoagulability states such as HIT or malignancy [17, 18].
In our patient, VLG likely resulted from the combination of warfarin and interruption of argatroban therapy. VLG is defined as acral necrosis in a limb affected by DVT despite the presence of palpable or Doppler-identifiable pulses. Patients typically demonstrate a supratherapeutic INR, microvascular thrombosis, and thrombocytopenia . Even a single dose of warfarin can precipitate VLG by promoting a reduction in protein C, creating a disturbance in procoagulant-anticoagulant balance. The half-life of protein C is short compared to that of prothrombin, the major procoagulant zymogen (8 vs. 60 hrs, respectively). As seen in this case, an interruption in parenteral anticoagulant therapy is often a contributing factor. The unopposed warfarin effect allows for ongoing thrombin generation and transformation of the DVT to VLG. Warfarin reversal with vitamin K should be considered in VLG regardless of the presence or absence of HIT .
Methylprednisolone 1 g IV daily for 3 days, then taper, and IVIG 500 mg/kg daily for 5 days was initiated for severe APS. Plasmapheresis was not undertaken secondary to lack of IV access. On day 4, hematology recommended anti-IIa levels using a chromogenic substrate assay (argatroban therapeutic range 0.4–1.2 μμg/ml) for analysis by an outside laboratory to assist in guiding dose adjustments. Therapy was adjusted to maintain an aPTT>100 sec while awaiting anti-IIa levels. Available paired aPTT and anti-IIa levels are shown in TableI.
|Hospital day||aPTT (sec)||Anti-IIa (μg/ml) therapeutic range = 0.4–1.2 μg/ml|
Less than 1% of patients with APS develop catastrophic APS (CAPS), which involves rapidly progressive multiple organ failure. In up to 60% of cases, CAPS is precipitated by a triggering event such as trauma, infection, minor surgery, or anticoagulation problems . Although our patient did not meet objective criteria for definite/probable CAPS, the possibility of this diagnosis, and its attendant high mortality rate, led to the addition of corticosteroids and immune globulin therapy.
Hematology recommended anti-IIa levels which have been suggested for DTI monitoring in patients with APS; however, concerns exist with reproducibility and linearity and clinical experience is limited . The reported therapeutic range for argatroban monitoring using an anti-IIa assay have varied with upper limits ranging from 1.1 to 2.0 μg/ml . The upper limit of the assay used here was 1.2 μg/ml which is potentially conservative. Godier et al. described the management of argatroban therapy with an anti-IIa assay (Hemoclot thrombin inhibitors, Hyphen BioMed, Neuville-sur-Oise, France) in an ICU patient on continuous renal replacement therapy in whom aPTT monitoring was difficult secondary to significant variability in the ratios obtained (target range, 0.4–0.8 μg/ml). Over a 23-day course, an aPTT ratio of ∼2 corresponded with anti-IIa activity less than 0.8 μg/ml . In the case we report, no relationship was evident between the aPTT values and anti-IIa results (Table I). For an aPTT of 93 sec, anti-IIa levels ranged from 0.5 to 2 μg/ml. This complicated the ability to adjust the argatroban infusion while awaiting anti-IIa results. An assay with a more rapid turnaround time would have been beneficial in our case.
On days 5 and 6, the patient was thought to have Evan's Syndrome (autoimmune hemolytic anemia plus ITP). A DIC screen was negative (fibrinogen 375 mg/dl, D-dimer << 0.5 μg/ml). The patient was transfused with packed red blood cells to maintain a hemoglobin >>8 g/dl. The reticulocyte count was 379/mm3 (25–85), immature reticulocyte fraction 0.29 (0.02–0.16), Coombs test positive, glucose 6-phosphate dehydrogenase 14.8 U/gHbg (5.7–14.7), lactate dehydrogenase 685 U/l (100–238), and the haptoglobin was 67 mg/dl (60–270).
Over the next several days the patient was continued on argatroban therapy. On hospital day 12, the patient was started on cefazolin for gangrenous fingers of the right hand. On day 15, a functional protein C assay result was 192% (70–140%) and a functional protein S assay returned at 67% (57–125%). On day 17, multiple anti-IIa levels drawn on days 13 through 16 (test not available on weekends) returned and ranged from 0.4 to 2.2 μμg/ml. During this time argatroban doses ranged from 1.0 to 1.2 μμg/kg/min and aPTT values ranged from 58 to >>124 sec. An UE Doppler on day 18 (left side) revealed continued presence of earlier thromboses plus an additional DVT in the proximal subclavian vein and a new cephalic SVT. On this day the right radial pulse became consistently absent by Doppler. The lower extremity Doppler examination was negative. On hospital day 19, the patient developed hypoxia and went into acute respiratory failure requiring intubation. At intubation, blood was noted in the oropharynx and bleeding persisted in the endotracheal tube after intubation and despite stopping argatroban. Bronchoscopy revealed bilateral pulmonary hemorrhage. The patient developed hypovolemic shock and difficulty with oxygenation and ventilation (arterial blood gases: pH 7.11, pCO2 64 mm Hg, pO2 48 mm Hg, and HCO3 19 Meq/l). The patient died of cardiopulmonary arrest 7 hrs postintubation (day 19). Laboratory data included BUN 21 mg/dl, Scr 0.8 mg/dl, alkaline phosphatase 244 U/l, AST 45 U/l, ALT 30 U/l, and total bilirubin of 0.7 mg/dl. The last anti-IIa level, drawn 1-day prior, was reported after the patient's death as 2 μμg/ml (supratherapeutic).
Monitoring parenteral anticoagulant therapy in patients with a prolonged aPTT from APS, liver dysfunction, DIC, or warfarin is challenging, with limited literature available to guide the clinician. Love et al. demonstrated the potential value of using a plasma-diluted thrombin time for DTI monitoring in patients with a prolonged aPTT. A 1:4 dilution of patient plasma with normal plasma was used and was not affected by lupus inhibitors . Another option is to use the ecarin clotting time (ECT). Ecarin is an enzyme from snake venom which activates the conversion of prothrombin to meizothrombin, an intermediate, in a phospholipid-independent manner. Perry et al.  reported the successful use of the ECT to monitor lepirudin therapy in a patient with HIT and APS with a baseline laboratory aPTT of 93.7 sec. The ECT, however, is not routinely available. Although anti-IIa levels were a reasonable approach in the case reported here, the turnaround time for results was problematic. Despite normal renal function and improving hepatic function, initial dosing well below the manufacturer's recommendation , and frequent monitoring, our patient developed a fatal bleed with elevated argatroban (anti-IIa) levels. Although our patient had a questionable history of SLE that can cause pulmonary hemorrhage, excessive argatroban anticoagulation was believed to explain the fatal hemorrhage .
This unfortunate patient case highlights the problems with “overdiagnosis” of HIT. Despite “positive” tests for HIT antibodies, the low pretest probability for HIT and the known propensity of patients with APS to yield false-positive HIT antibody results suggests that the patient did not have a true diagnosis of HIT. Moreover, the early administration of warfarin and the choice of argatroban for parenteral anticoagulation when monitoring was hindered by a prolonged baseline aPTT likely play a key factor in the progression of UE DVT to VLG. Ironically, the problems of anticoagulant monitoring posed by the prolonged baseline aPTT likely contributed to the subsequent overanticoagulation and fatal pulmonary hemorrhage. With benefit of hindsight, avoiding the temptation to test for HIT in a low pretest probability situation, and treatment with either heparin using anti-factor Xa monitoring or with non-aPTT-monitored therapy such as LMWH or fondaparinux would likely have resulted in a more favorable clinical course.
- 1Package insert, PF4 Enhanced®. GTI, Waukesha, WI, January 13, 2005.
- 24GlaxoSmithKline. Argatroban prescribing information. NC: Research Triangle Park; 2008.