PERSISTENT or uncontrolled bleeding is one of the most feared consequences of any surgical procedure. Numerous factors can contribute to its etiology. Among these, both inherited and acquired disorders of coagulation and/or thrombocyte function need to be taken into account.1,2Acquired hemophilia can be caused by antibodies to coagulation factors. Antibodies can arise in hemophilic or nonhemophilic patients. In the former case, patients are treated with Factor VIII concentrate and develop alloantibodies. Nonhemophilic patients produce autoantibodies (mostly against Factors VIII and IX), and this is believed to be an autoimmune reaction.3The estimated annual incidence rate of clinically apparent bleeding caused by autoantibodies is 1.48/million inhabitants,3or approximately 450 patients in the United States per year. The incidence of the disease seems to be increasing; it is more prevalent in elderly patients and it can be successfully treated if diagnosed adequately and early.3Here, we present a case of massive and long-lasting bleeding in a nonhemophilic patient undergoing extensive visceral surgery.
A 57-yr-old woman (158 cm, 74 kg) presented with a long-lasting history of upper abdominal pain and a diagnosis of chronic pancreatitis. On computed tomographic and nuclear magnetic resonance imaging the pancreatic duct was dilated and obstructed by a calcific structure. Endoscopic extraction of a stone had failed by endoscopic retrograde cholangiopancreatography. Coagulation studies were normal, with a mildly prolonged activated partial thromboplastin time (aPTT; 44 s, normal value 20–40 s).
Three months later, the patient decided to undergo surgery because of persistent pain. She was in good general health, employed full-time, and physically active. Her past medical history revealed mild hypertension, minor obstructive sleep apnea, and dyspnea on exertion without angina. She had a smoking history of 20 packyears. Her hypertension was treated with metoprolol and lisinopril. She had undergone several minor abdominal surgeries without complications. The interview and physical examination were otherwise unremarkable. She denied any signs of past or present abnormal bleeding or a family history thereof. Laboratory values were within normal range, with the exception of an aPTT of 78 s, which was not verified preoperatively. The patient underwent a laterolateral pancreaticojejunostomy as a drainage procedure, which was uneventful.
Postoperatively, a persistently elevated aPTT as well as a progressively more severe anemia (nadir, hemoglobin 4.9 mg/dl on postoperative Day [POD) 2) were noticed. Erythrocyte concentrates and fresh frozen plasma were transfused. Moderate bleeding persisted (two units of blood/day). Surgical reexploration demonstrated diffuse bleeding activity. On POD 4, an extensive coagulation workup demonstrated a severe Factor VIII deficiency (4%) as the most significant pathologic value (table 1). On direct questioning the patient recalled that she had developed an extensive hematoma on her lower leg after a minor trauma several weeks before surgery. The patient received Factor VIII with von Willebrand factor (12,000 units over 3 days; Hemate, CSL Behring, Marburg, Germany) and bleeding further increased. On POD 8, the diagnosis of an acquired Factor VIII deficiency was established when a Factor VIII inhibitor of 117 Bethesda units (fig. 1) was demonstrated. The patient was transferred to our intensive care unit. Once the diagnosis was established the patient was treated with activated prothrombin complex concentrate (aPCC; FEIBA [Factor VIII bypassing activity], Vienna, Austria) at a dose of 100 units/kg twice daily. Bleeding continued and treatment was switched to rFVIIa (up to 90 units/kg every 2 h; NovoSeven, Novo Nordisk, Bagsvaerd, Denmark; fig. 1A). Simultaneously, inhibitor elimination was started using corticosteroids (prednisolone 70 mg/day), plasmapheresis with adsorption of immunoglobulins (Therasorb-Ig; Miltenyi Biotec, Bergisch Gladbach, Germany) three days per week, followed by intravenous immunoglobulins (30 mg/day for 4 days; Biotest, Dreieich, Germany). The patient received a total of 50 units of packed red blood cell concentrations (fig. 1A), and bleeding was controlled at POD 14. Catastrophic bleeding started again on POD 29. Angiographic studies demonstrated a leak from the gastroduodenal artery that could not be controlled by endovascular coiling; it required surgical ligation and repeated abdominal packing. Persistent bleeding was temporarily controlled by high dose rFVIIa (fig. 1A), platelet transfusions, fibrinogen, and tranexamic acid. Inhibitor elimination as well as Factor VIII-c restitution was finally achieved on POD 100 (fig. 1B).
An enterocutaneous fistula had developed at the level of the transverse colon on POD 55. The abdominal cavity was partially open, and recurrent interenteric abscesses occurred that were accompanied by septic shock. Pathogens included K. pneumoniae, M. morganii, extended spectrum β lactamase E. coli, E. faecium, and P. aeruginosa. Drainage was achieved by repeated computer tomography–guided placement of pigtail catheters under antibiotic coverage. These complications required prolonged intensive care treatment after bleeding control.
On POD 200, the patient was again extubated, did not require ventilator support, was cardiovascularly stable, received full enteral nutrition, had normal kidney function, communicated freely, and spent several hours outside the hospital with her family. However, on POD 207 she developed another episode of septicemia. Despite aggressive treatment, this resulted in septic shock with multiorgan failure as a result of panantibiotic-resistant P. aeruginosa on POD 216. The patient died on POD 217.
Our case report demonstrates the catastrophic consequences of performing extensive visceral surgery in a nonhemophilic patient with a newly acquired Factor VIII inhibitor and insufficient preoperative preparation. If a patient is scheduled to undergo major surgery and presents with an abnormal coagulation study, the pathologic results should be verified preoperatively. Should the repeat study generate an abnormal value, the patient should be specifically asked about signs of a bleedings disorder (e.g., soft tissue hematoma) and a coagulation expert should then be consulted.
Patient Presentation and Demographic Characteristics
Our patient was 57-yr-old, previously healthy with no congenital hemophilia, and presented with progressively more severe postoperative bleeding. In a prospective cohort study, nonhemophilic patients with autoantibodies to coagulation factors (acquired hemophilia) were studied by the United Kingdom Hemophilia Centre Doctors’ Organization.3Over a 2-yr period they identified 178 patients. Patients were mostly older than 65 yr (85%; median age 78 yr) and had no coexisting disease (63%). Acquired hemophilia was associated with malignancy (15%), autoimmune or collagen vascular disease (15%), pregnancy (2%, mostly postpartum) or dermatologic diseases (3%), and these coexisting diseases are significantly more frequent in younger patients. Spontaneous bleeding frequently occurs subcutaneously (25%), in muscle (45%), or in the gastrointestinal tract (22%). Fatal bleeding is present in 9% (gastrointestinal, intracranial, retroperitoneal, and perioperative).
Establishing the Diagnosis of Acquired Hemophilia
If unexpected bleeding occurs, adequate first-line treatment should be given.#Our patient presented with diffuse bleeding and a prolonged aPTT. This can be because of factor deficiencies (VIII, IX, XI, XII, or fibrinogen), von Willebrand syndrome, lupus anticoagulants, medication effects (e.g. , heparin, hirudin, activated protein C), fibrinogen split products, or acquired hemophilia with antibodies against coagulation factors (treatment-related alloantibodies in hemophilic patients or autoantibodies in nonhemophilic patients). In the case of acquired Factor VIII inhibitors as a result of autoantibodies, a decreased Factor VIII is found and the pathologic aPTT is not corrected if aPTT is determined after mixing patient and pool plasma. Other coagulation studies such as international normalized ratio and factor concentrations (fibrinogen, Factors II, IX, XI, XII, XIII, and von Willebrand factor ristocetin-cofactor and antigen, table 1) are within normal limits. Lupus coagulants are not present. Thrombocyte count and function are also normal. Factor VIII inhibitor concentrations are determined by mixing patient and control serum in a dilution curve using an enzyme-linked immunosorbent assay. One Bethesda unit describes a 50% decrease in Factor VIII activity (Nijmegen-Bethesda protocol4). Importantly, neither Factor VIII levels nor inhibitor concentrations accurately predict bleeding intensity. A Factor VIII gene analysis is not indicated, since it is an acquired disease without a genetic background.
A number of case reports describe the clinical course and outcome of surgeries in patients with acquired Factor VIIII inhibitor but no accompanying congenital hemophilia A (table 2). Patients will sometimes not be identified preoperatively for several reasons (table 2). Coagulation studies can be near or even within normal limits before surgery. Alternatively, hospital policy or national guidelines do not demand these studies for minor surgical procedures, and these will thus not be available preoperatively. Mistakes can occur and surgery is performed despite abnormal values, or the risks are underestimated. However, even minor surgical procedures (e.g. , insertion of central lines or extraction of teeth) can have devastating consequences in these patients (table 2). Bleeding can occur during or immediately after surgery, but often involves a time delay of several days.
Treatment of Acute Bleeding
Our patient presented with two major episodes of acute bleeding. During the first episode, diffuse bleeding was noticed on surgical reexploration and was controlled by aFVIIa by POD 14. De novo and massive bleeding started again on POD 29. It originated from the gastroduodenal artery and was probably a result of an erosion caused by a drainage tube. It finally was controlled by surgical ligation, abdominal packing, and rFVIIa treatment.
While patients with congenital hemophilia A are treated with Factor VIII concentrates, this is not recommended in patients with acquired hemophilia. Our patient was initially treated with fresh frozen plasma and Factor VIII concentrates. Especially Factor VIII concentrates are ineffective, and might even increase bleeding since they can boost the formation of autoantibodies. Thus, two treatment options exist that are recommended in international consensus guidelines5: aPCC (50–100 units/kg every 8–12 h with a maximum daily dose of 200 units/kg), and rFVIIa (90 μg/kg, up to every 2 h because of its short half-life). If either aPCC or rFVIIa is ineffective, the other should be tried. aPCC is an activated prothrombin complex and bypasses the necessity of Factor VIII activation. rFVIIa induces supraphysiologic concentrations of Factor VIIa, which bind to thrombocytes and thereby directly activate Factor X and subsequently thrombin. Both aPCC and rFVIIa are very expensive, and daily treatment costs can exceed $20,000.
Importantly, bleeding can persist for days or even weeks. Furthermore, inhibitors can reoccur after successful elimination, leading to recurring bleeding episodes. In our patient, bleeding persisted for almost 10 weeks, leading to massive transfusion requirements (fig. 1), but was finally controlled.
Factor VIII Inhibitor Elimination and Surgery in Patients with Known Acquired Hemophilia
Inhibitor elimination aims to suppress autoantibody production using immunosuppressive agents. Treatment should involve steroids (prednisolone 1 mg · kg−1· day−1) with or without cyclophosphamide (1.5–2 mg · kg−1· day−1) for 4–6 weeks. Rituximab is considered a second-line treatment. For high-risk patients, plasmapheresis with immune adsorption and intravenous immunoglobulin treatment with or without Factor VIII substitution has been recommended.5,6Complete remission can be achieved in 70–80% of patients, and involves a median time of 40–60 days (range, 2–360 days).3,5Since acquired hemophilia A is more frequent in older patients, it is important to note that inhibitor elimination occurs faster in older patients than in the younger age group.3
We achieved complete remission in 70 days using corticosteroids, immune adsorption, and high-dose intravenous immunoglobulins. Cyclophosphamide was not used, and only a single dose of rituximab was given because of the recurrent infectious complications.
If patients with acquired hemophilia are identified preoperatively, surgery should be postponed except for life-threatening emergencies. Prior inhibitor elimination should be seriously considered in collaboration with coagulation experts. If successful, surgery has been reported to occur without bleeding complications (e.g. , lobectomy7). If surgery cannot be postponed, prophylactic treatment with FEIBA or rFVIIa should be seriously considered.
In summary, acquired hemophilia A is caused by autoantibodies (so-called inhibitors) to coagulation factors (mostly Factor VIII). Patient can be often identified by a history of unexplained bleeding episodes and by a prolonged aPTT. Before surgery, inhibitor elimination should be attempted. If surgery is urgent or unexpected bleeding occurs, treatment options include aPCC or rFVIIa.