To the Editor:—
Several case reports have suggested that recombinant factor VIIa (rFVIIa, Novoseven®; NovoNordisk A/S, Copenhagen, Denmark) would be an effective therapy to control severe postoperative bleeding for patients without preexisting coagulopathic disorder.1–3Furthermore, a recent placebo-controlled study has shown that prophylactic rFVIIa administration can reduce blood loss and erythrocyte transfusion for patients undergoing abdominal prostatectomy.4Although the optimal dose and the usefulness of additional bolus are not established for perioperative bleeding, most reports emphasized the efficacy and the safety issue after rFVIIa administration. However, treatment-related thrombotic complications have been suggested, but in the few cases reported, the cause-and-effect relation has never been clearly demonstrated.
A 69-yr-old man with a medical history of diabetes mellitus and arterial hypertension was hospitalized for hematuria and left lower limb lymphoedema. The scanner showed two bladder tumors and a left iliofemoral venous thrombosis. Doppler-echography confirmed a complete iliac, femoral, and popliteal venous thrombosis. The endoscopic bladder resection led to the diagnostic of high-grade multifocal urothelial carcinoma. The patient was discharged from the hospital with an anti–vitamin K therapy by fluindione.
Three months later, and 25 days before a second cystoscopy, another Doppler-echography was performed and showed venous permeability with only residual parietal change on the common femoral vein. A relay of fluindione by low-molecular-weight heparin (Tinzaparin; 12,000 UI/day) was initiated 5 days before surgery. Tinzaparin was stopped the day before cystoscopy. All biologic parameters before surgery were within the normal range, and at this time, there was no clinical evidence of deep venous thrombosis resurgence. The cystoscopy led to a second endoscopic bladder tumor resection, and continuous bladder irrigation with normal saline was instituted. Immediately after surgery, persistent blood loss through the bladder catheter was noticed. The persistent bleeding diathesis led to two additional hemostatic endoscopic procedures. However, the hemorrhagic diathesis remained refractory to surgery, and the patient was admitted to the intensive care unit. Despite persistent blood loss through the bladder catheter, the mean arterial pressure and heart rate remained stable without catecholamine infusion. Moderate hypothermia (35°C) was recorded. Hemoglobin decreased to 5.7 g/dl, reaching 9.1 g/dl after a transfusion of 3 units packed erythrocytes. Despite two additional fresh frozen plasma transfusions, the hemorrhagic diathesis persisted, and the hemoglobin concentration decreased approximately of 0.5 g/dl per hour. A moderate dilutional coagulopathic disorder occurred (prothrombin time, 17 s; activated partial thromboplastin time, 39 s; and platelet count, 147,000/mm3).
Ten hours after admission, a first dose of rFVIIa at 90 μg/kg was given. At the end of the infusion, bladder irrigation fluid became clear, and hemostasis was definitively achieved. Eight hours later, an additional 70-μg/kg rFVIIa bolus was given to prevent bleeding recurrence. The following day, systematic Doppler-echography revealed a femoral deep venous thrombosis recurrence, without any clinical symptom. A therapy by unfractionated heparin was initiated, without any hematuria resurgence. The patient was discharged from the hospital 10 days later with anti–vitamin K therapy, without any new Doppler investigation.
The extended use of rFVIIa in the perioperative period opens a new way in the management of refractory bleeding and has clearly generated enthusiasm, given the abundance of case reports. There is evidence that rFVIIa is a potent prohemostatic agent that may potentially lead to thromboembolic events. Experimental and clinical studies, however, support that rFVIIa enhances hemostasis at the site of injury without systemic coagulation activation, because thrombin is generated after rFVIIa infusion where tissue factor and activated platelets are localized. This explains that most of reports regarding rFVIIa efficacy underline the tolerance, with no or little evidence of treatment-related thrombotic events.5–7Indeed, very few thrombotic episodes are reported in the literature. However, coexisting risk factors contributing to thrombotic complications are often found, especially preexisting vessel wall damage such as atherosclerosis.8–10In our case, we cannot exclude that the deep vein thrombosis recurrence occurred between the second Doppler-echography and surgery, but the absence of any clinical evidence and the prescription of an effective anticoagulant treatment do not support this hypothesis. In this way, it could be suggested that endothelium healing was not achieved, leading to permanent tissue factor exposition and/or platelet activation and allowing the recurrence of thrombosis after rFVIIa infusion.
The recurrence of deep venous thrombosis may also be related to the rFVIIa dosage and to an adverse side effect of the additional prophylactic bolus 8 h later. The recommended dose of rFVIIa to stop bleeding for hemophilic patients seems to be well established, but the optimum dosage of rFVIIa for patient without preexisting coagulopathy remains unclear. The lack of controlled randomized trials, the small series or single cases reported, the variability of the hemostatic abnormality leading to the bleeding diathesis, and the nature of the surgical damage do not allow us to draw any standard guidelines. As a result, available data indicate the efficacy of one or two infusions of 20–120 μg/kg rFVIIa to stop bleeding.11Our patient received an initial 90-μg/kg rFVIIa bolus. No data in the literature suggest that a lower initial dose would have reduced the risk of thrombosis recurrence. Moreover, no data are available regarding appropriated rFVIIa strategy in the hours after the hemostasis achievement for nonhemophilic patients. The administration of an additional infusion of rFVIIa to prevent the hemorrhagic diathesis recurrence could only be supported by the short half-life of rFVIIa (approximately 2.9 h) and the lack of any available biologic parameters related to rFVIIa clinical effectiveness. In our case, rFVIIa reinjection may have played a role in the venous thrombosis recurrence. The usefulness and the safety of prophylactic rFVIIa reinjection to prevent bleeding resurgence still have to be demonstrated.
Therefore, rFVIIa is a potent hemostatic factor, but it should be used carefully to control postoperative bleeding for patients without preexisting coagulopathy. When hemorrhagic diathesis occurred in the postoperative period, a hemostatic surgical procedure should be first discussed, and the primary medical objective must be to correct the hemostatic parameters such as platelet count, anemia, dilutional coagulopathic disorders, acidosis, or hypothermia. Furthermore, before rFVIIa infusion, the benefit–risk balance should be evaluated, especially for patients with known atherosclerosis or thromboembolic disease. If bleeding remains uncontrolled, rFVIIa could be used as a last resort.
*Hôpital Saint-Louis, Assistance-Publique Hôpitaux de Paris, France. firstname.lastname@example.org