Systemic AL amyloidosis with acquired factor X deficiency: A study of perioperative bleeding risk and treatment outcomes in 60 patients

Abstract

Systemic light-chain (AL) amyloidosis may be associated with acquired factor X (FX) deficiency and optimal management of this coagulopathy is unknown. We reviewed our experience with 60 patients with isolated FX deficiency (≤50%) due to AL amyloidosis that underwent an invasive procedure between 1975 and 2007. They were classified as having severe (<10%; n = 6), moderate (10–25%; n = 15), or mild (26–50%; n = 39) FX deficiency. The patients underwent a total of 112 procedures, 19 (17%) of which were managed with peri-procedural treatment with one or more hemostatic agents. There were complications in 14 (13%) procedures (bleeding = 12, thrombosis = 1, death = 1). Baseline FX level was not predictive of bleeding risk; the only association with postintervention bleeding was central venous catheter placement. However, bleeding complications were relatively infrequent, particularly in patients with mild or moderate FX deficiency undergoing nonvascular procedures. Activated recombinant factor VII might be considered in patients undergoing major surgical procedures, but further experience is needed. Optimal management of AL patients with FX deficiency undergoing invasive procedures remains to be determined.

Introduction

Systemic light-chain (AL) amyloidosis is a clonal hematological disorder characterized by deposition of amyloid fibrils composed of immunoglobulin light chains, and can be associated with hemorrhagic [1] and thrombotic [2] disorders. Although the most common coagulopathy is the presence of a thrombin inhibitor [3], factor X (FX) deficiency causes the most frequent bleeding manifestations [1] and therefore, is the most clinically significant. The incidence of acquired FX deficiency associated with systemic AL amyloidosis varies from 8.7% [4] to 14% [1].

Acquired deficiency of FX is due to its adsorption by the amyloid fibrils [5,6], which leads to a greatly reduced circulating half-life. This can lead to serious hemorrhagic complications [4] and is a difficult bleeding disorder to treat, as many of the treatments for acute reversal of coagulopathy are ineffective. Published experience, limited to case reports, suggest that infusion of fresh frozen plasma (FFP) [5], platelets, plasma exchange [7], and vitamin K [8] are of limited hemostatic efficacy. Splenectomy has been used as a successful treatment of FX deficiency in patients with systemic AL amyloidosis, likely due to the removal of a large burden of amyloid deposits [9]. Autologous stem cell transplant has also been reported to raise the FX level [4]. However, both of these strategies are ineffective in the short term. Recently, there have been several case reports of activated prothrombin complex concentrates (APCC), including activated recombinant factor VII (rFVIIa), in the management of FX deficiency [10–12]. However, data on their role in management of patients with FX deficiency associated with AL amyloidosis is limited. We reviewed our cumulative experience of patients with systemic AL amyloidosis and associated FX deficiency that underwent an invasive procedure, with the following aims: (i) to determine the prevalence of bleeding complications; (ii) to identify preprocedural predictors of bleeding; and (iii) to identify optimal management strategies to attenuate bleeding risk in these patients.

Materials and Methods

Patients with systemic AL amyloidosis and isolated FX deficiency (≤50%) who underwent an invasive procedure between 1975 and 2007 were retrospectively identified from institutional databases. FX levels were measured by the prothrombin time-based one stage clot-based assay using factor X-deficient substrate from Precision Biologic. The lower limit of detection is 3%, and the assay is 100% specific for factor X activity. AL amyloidosis was diagnosed according to previously described methods [13]. Excluded were patients with other coagulopathies such as dysfibrinogenemia or multiple factor deficiencies, patients who underwent bone marrow biopsies, and those who had procedures performed >3 months before the diagnosis of FX deficiency. Endpoints of the study were bleeding, thrombosis, and death. The Institutional Review Board approved the study. Statistical analysis, including descriptive statistics, chi-square analysis, and logistic regression for associations between complications and explanatory variables was performed in STATA™ 8.2 (College Station, TX). P values < 0.05 were considered statistically significant.

Results

Sixty patients fit the study criteria; 39 (65%) men and 21 (35%) women. Median age was 60 years (range 35–86) and the median FX activity was 28.5% (range 3–50%). Based on FX activity, patients were classified as having severe (<10%, n = 6), moderate (10–25%, n = 15), or mild (26–50%, n = 39) FX deficiency. The patients underwent a total of 112 procedures, 13 (12%) of which were considered major procedures and 99 (88%) were considered minor procedures.

Major procedures (n = 13)

The major procedures included four pacemaker implantations, three splenectomies, one colon resection, one tracheostomy, one bladder repair, one renal transplant, one AV fistula placement, and one cholecystectomy. Table I includes details of treatment and complications. There were four patients with severe FX deficiency, two with moderate, and seven with mild. Of the four patients with severe deficiency, two had complications: hematoma in the splenic bed following splenectomy, and death in another patient. All four patients were treated with hemostatic agents in the perioperative period, including three who received rFVIIa. The death was in a patient who suffered a spontaneous splenic rupture and was initially treated with hemostatic therapies. Later, however, the patient became hemodynamically unstable and underwent emergent splenectomy, but died in the operating room due to bleeding complications. The two patients with moderate deficiency did not have a complication, and neither received hemostatic agents. Of the seven patients with mild deficiency, there were no complications, including the five patients who received no perioperative hemostatic agents.

TABLE I.

Major Invasive Procedures in Patients with FX Deficiency Due to AL Amyloidosis

Procedure	FX%	Year of procedure	Pre-op treatment	Post-op treatment	Complication
Severe FX deficiency (<10%)
Splenectomy	3	2006	FFP (2 units), rFVIIa (90 mcg/kg × 1)	FFP (4 units), rFVIIa (90 mcg/kg × 33 doses every 2–4 hr)	None
Splenectomy	3	1996	Plasma exchange	Plasma exchange, FFP (64 units), tranexamic acid, danazol, aminocaproic acid	Hematoma in splenic bed
Renal transplant	5	2003	rFVIIa (30 mcg/kg × 1)	rFVIIa (30 mcg/kg × 1)	None
Splenectomy (for splenic rupture)	6	2004	FFP (20 units), rFVIIa (90 mcg/kg × 9 doses every 3–4 hr), platelets (6 apheresis units)	Not applicable	Death due to bleeding
Moderate FX deficiency (10–25%)
Pacemaker placement	11	2000	None	None	None
Cholecystectomy	19	1990	None	None	None
Mild FX deficiency (26–50%)
Pacemaker placement	37	2004	None	None	None
Colon resection	40	2001	None	None	None
Bladder repair	43	2002	FFP (6 units), platelets (1 apheresis unit), DDAVP	FFP (22 units), platelets (5 apheresis units), DDAVP, vitamin K	None
Tracheostomy	44	2005	FFP (4 units), platelets (2 apheresis units)	None	None
AV fistula	46	2001	None	None	None
Pacemaker placement	46	1996	None	None	None
Pacemaker placement	49	2005	None	None	None

FX, factor X; FFP, fresh frozen plasma; rFVIIa, recombinant activated factor VIIa.

Minor procedures (n = 99)

Central venous catheter (CVC) placement was the most common minor procedure (n = 25). Other common minor procedures included right ventricular biopsy, upper and lower endoscopy with biopsy, and percutaneous renal biopsy. Of the six patients with severe deficiency, two had bleeding at the CVC insertion site; both patients received one dose of rFVIIa before the procedure (30–90 mcg/kg). One required an additional 8 doses of fFVIIa (90 mcg/kg) and 2 units of FFP, followed by eventual removal of the catheter due to ongoing bleeding at the site. The other had mild bleeding at the site that was treated conservatively with line reposition and pressure dressings. Of the 16 patients with moderate deficiency, there were two complications; one patient developed bleeding after CVC placement despite receiving rFVIIa (30–60 mcg/kg) pre and post-procedure, and another patient developed a subcutaneous hematoma following nerve biopsy who did not receive hemostatic agents. There were eight complications in the 39 patients with mild deficiency, including bleeding following CVC placement (n = 3), lower gastrointestinal bleeding following colonoscopy with biopsies (n = 1), hematuria following renal biopsy (n = 1), subcutaneous hematoma following nerve biopsy (n = 1), and hematuria following transurethral resection of a bladder mass (n = 1). The latter patient developed thrombosis of his arterio-venous fistula after 9 doses of rhFVIIa (90 mcg/kg).

Associations with postintervention complication/bleeding risk

In this series, the overall risk of complications was 13%. Twelve procedures were complicated by bleeding (11%), 1 by thrombosis (1%), and there was 1 death (1%). Of the 12 bleeding complications, half (n = 6, 50%) were seen in association with CVC placement, which is indicative of the potentially high risk of vascular procedures in this group of patients. Overall, rFVIIa was used preoperatively in nine procedures, 4 (44%) of which were complicated by bleeding, thrombosis, or death. Of the 93 procedures that were not managed with any perioperative procoagulants, 87 (78%) were uncomplicated.

Univariate analysis was performed to identify preprocedural variables that were predictive of complications (bleeding, thrombosis, or death) in this cohort. Age, gender, level of FX deficiency (mild/moderate/severe), and type of procedure (major vs. minor) were not predictive of complications; while CVC placement was significantly associated with complications (P = 0.03). In the subset of procedures that were not managed with perioperative procoagulant therapy (n = 93), FX level was not predictive of bleeding risk (P = 0.32), although there were only six events.

Discussion

In the current series of 60 patients with isolated FX deficiency due to AL amyloidosis, there was a 13% incidence of complications (bleeding, thrombosis, or death) associated with an invasive procedure. Aside from an increased risk of bleeding with insertion of a central venous catheter, there were no clear predictors of bleeding. In the current cohort, patients with severe FX deficiency were more likely than those with mild or moderate FX deficiency to receive preprocedure procoagulant therapy as prophylaxis (P < 0.001), which, in a retrospective analysis, makes it difficult to quantify the risk of bleeding as a function of baseline FX levels. However, in the subgroup of patients who did not receive periprocedure procoagulant therapy, bleeding risk was not predicted by the baseline FX level. The highest rate of complications in this series was after CVC placement (50%), which is in contrast to the 0% bleeding rate reported in pediatric hemophilia patients [14, 15] and bone marrow transplant patients [16] undergoing CVC placement. One patient did suffer a thrombosis after receiving repeated doses of rFVIIa, a known complication of treatment [17], including in patients with other bleeding disorders such as hemophilia [18]. Therefore, close monitoring is recommended while administering a course of rFVIIa or activated prothrombin complex concentrates (PCC). Concurrent use of fibrinolytic inhibitors in hemophilia patients with FVIII inhibitors receiving rFVIIa has been shown to be safe [19–21], however there is no such safety data in patients with amyloid receiving rFVIIa and their use in patients receiving PCC should be avoided.

Although patients with defined pre-existing coagulopathies were excluded from the present analysis, systemic AL amyloidosis patients have other predispositions to bleeding that are not easily quantified. These include thrombocytopenia, platelet dysfunction associated with renal insufficiency, vitamin K deficiency due to malabsorption, and/or decreased levels of coagulation factors due to defective hepatic synthetic function, which may contribute to bleeding. Amyloid deposition in the vessel wall is thought to result in vascular friability that may also contribute to bleeding in these patients. Therefore, patients with FX deficiency due to AL amyloidosis are not similar to a hemophilia patient, where the degree of factor deficiency is the major predictor of bleeding complications.

Although this study helps to estimate the risk of bleeding in AL patients with acquired FX deficiency undergoing operative procedures and describes their clinical outcome, it is limited in identifying the optimal periprocedural treatment regimen because of the wide variety of treatments rendered, relatively small number of patients, and significant selection bias in which patients received procoagulant therapy. Although this particular issue would be best addressed in a prospective fashion, it would be challenging given the rarity of FX deficiency in systemic AL amyloidosis. In practice, we favor use of rFVIIa, FFP, or aminocaproic acid in patients that require hemostatic agents, with vigilant monitoring for complications. It should be noted that 44% of procedures in which rFVIIa was used preoperatively were complicated by bleeding, thrombosis, or death; suggesting that it may not be particularly effective if there is a major challenge to hemostasis, although this conclusion is significantly limited by selection bias.

In conclusion, this is the largest series reviewing the experience of perioperative management and outcomes of patients with FX deficiency due to systemic AL amyloidosis. Our data suggests that there are no clear preoperative predictors of bleeding. Complications were infrequent, particularly for those with mild or moderate FX deficiency undergoing nonvascular procedures. Optimal management of these patients remains unclear.