Introduction
Recombinant factor VIIa (rFVIIa), originally developed for the treatment of acquired inhibitors associated with hemophilia [
1,
2], has been successfully used for bleeding due to acquired or congenital thrombocytopathy [
3,
4], extensive trauma and a variety of surgical procedures, including anecdotal reports of use in cardiac surgery patients [
5‐
9]. It reverses the effect of warfarin in healthy volunteers [
10], and corrects the prothrombin time in patients with hepatic failure [
11].
Most critically ill patients with trauma suffer profound coagulopathy. Coagulation abnormalities in these patients have been attributed to multiple factors including disseminated intravascular coagulation (DIC), excessive fibrinolysis due to release of tissue plasminogen activator (tPA), dilutional coagulopathy from fluid replacement and massive blood product transfusion, dysfunctional platelets, and metabolic abnormalities including acidosis and hypothermia [
12‐
18]. In addition, the use of cardiopulmonary bypass in patients undergoing cardiac surgical procedure may exacerbate the coagulopathy [
19].
Although the mechanism of action of rFVIIa remains unclear, many investigators have suggested that it binds to the surface of activated platelets and directly activates Factor X, thus bypassing the early steps of the coagulation cascade. Activated Factor X (Xa) then combines with activated Factor V (Va) on the platelet surface, leading to rapid conversion of prothrombin to thrombin [
20,
21]. Hemostasis is promoted through high concentrations of thrombin generated near activated platelets at the site of vascular injury. Based on its mechanism of action, rFVIIa may be effective in controlling hemorrhage due to trauma, surgery and other causes.
The present report is a retrospective review of 40 consecutive patients treated with rFVIIa (NovoSeven, Novo Nordisk, Denmark) for intractable hemorrhage. Twenty-four of these patients had bleeding associated with cardiac surgery, and the remaining 16 suffered with bleeding from other causes. The need for transfusion of blood products both before and after treatment and the mortality at 4 hour and 30 day were assessed. Patient characteristics were also assessed as risk factors for mortality.
Methods
Patients
The study was approved by the Institutional Board Review of the Washington Hospital Center, Washington DC, USA. The medical records of 40 consecutive patients treated with rFVIIa between July 2001 and June 2003 at the Washington Hospital Center, a tertiary care teaching hospital and level 1 trauma center, were reviewed. The patients were divided into two groups: 24 patients who had undergone cardiac surgical procedures (CSP), and 16 patients who had bleeding from other causes (OBP). There were more male patients in both groups (19 males and 5 females in CSP, 10 males and 6 females in OBP). The median age in the CSP was higher at 65 years (range 26–85 years) compared to that of OBP (median age 42; range 23–82 years). The type of surgery in the CSP group included 15 coronary artery bypass grafts (CABG), 3 aortic arch repairs either elective or for dissection, 2 aorta rupture repair (one also underwent CABG), 3 valve replacements and 1 ventricular rupture. In the OBP group, 5 patients had suffered either stab or gunshot wounds, 2 had traumatic liver lacerations, 2 had undergone joint replacement, 3 had clotting factor abnormalities (acquired Factor VIII inhibitor, congenital Factor VII deficiency, acquired VWF deficiency), and one each had post partum hemorrhage, warfarin overdose, bowel resection, and trauma from a motor vehicle accident.
Administration of rFVIIa
Commercially available rFVIIa, (Novo-Seven, Novo Nordisk, Denmark) was used in all patients. There were no predetermined criteria for administration. In most patients, treatment was given after extensive blood product support had failed to control hemorrhage, and the decision to treat was made jointly by the surgeon, hematologist and blood bank pathologist. rFVIIa was administered as a bolus dose of 90 mcg/kg. The first two patients received a second dose of 90 mcg/kg 6 hours after the first.
Endpoints and Assessment of Risk
The total number of products given before and up to twenty-four hours after rFVIIa administration, or until death, whichever came first, was quantified. Patients who died within 4 hours of treatment were not evaluable because blood product support was suspended shortly after rFVIIa administration in these hemodynamically unstable patients deemed non-salvageable. Also, given the short half-life of rFVIIa, this time frame did not allow for an accurate determination of efficacy. The number of patients surviving at 4 hours and 30 days after the administration of rFVIIa was compared. Thirty days was chosen as an end point since this is the operative mortality cut off point, as defined by the Society of Thoracic Surgeons, for inpatient and out of hospital death. For patients surviving more than 4 hours, charts were reviewed for thromboembolic complications during hospitalization.
In an attempt to determine risk factors for poor outcome, patients were assessed for age, gender, comorbid conditions, left ventricular ejection fraction (LVEF), use of cardiopulmonary bypass (CPB), hemoglobin, platelet count, arterial blood gas results and coagulation parameters including prothrombin time (PT), activated partial thromboplastin time (aPTT), fibrinogen, and D-Dimers. Laboratory parameters were not available for some patients.
Statistical Considerations
The Paired Student's t-Test was used to compare the number of blood product transfusions before and after treatment with rFVIIa, and, since the number of pair groups was less than 50, this result was confirmed with the Wilcoxon Matched-Pairs Signed-Ranks Test.
Discussion
This retrospective case series suggests that rFVIIa is effective in promoting hemostasis even when given as a last life-saving measure in poor prognosis patients with massive, transfusion-refractory hemorrhage. A statistically significant decrease in blood product transfusion was evident in 12 CSP and 11 OBP who survived more than 4 hours after rFVIIa infusion. Six of 12 CSP and 2 of 11 OBP died later between 3 and 30 days after rFVIIa infusion, but these deaths were due to causes other than hemorrhage.
Despite the obvious limitations of a retrospective analysis, this is the largest series of such patients reported in the literature to date. Al Douri et al [
8], described 5 patients who underwent open-heart surgery for valvular disease and experienced uncontrollable intraoperative or postoperative bleeding unresponsive to a blood product support protocol. The study was prospective, had well-defined criteria for "excessive bleeding" and excluded patients with a history of ischemic heart disease, stroke or venous thromboembolism. Hemostasis was achieved in all patients following a single dose of 30 ug/kg of rFVIIa, and there was no mortality from bleeding, although 1 patient died 3 days later of an unrelated cause. Overall, the 4-hour and 30 day mortality in the present series was high compared to that of Al Douri et al [
8], but it is difficult to compare these two very different patient populations.
Factors predictive of inferior outcome after cardiac surgery include age over 60, low LVEF, significant co-morbidity, the use of CPB and re-exploration after initial surgery [
22‐
24]. The use of CPB increases the rate of fibrinolysis and induces an inflammatory response [
19,
25]. Moreover, prolonged time on CPB, defined as 2.5 hours or more, is associated with increased bleeding and need for re-exploration. Two-6% of patients undergoing CABG require re-exploration for bleeding and is associated with a mortality rate of up to 22% [
19,
26,
27]. In this series, the CSP group had multiple poor prognostic factors. The median age was 65 years and in addition, all 12 of the CSP who died within 4 hours and 3 of 6 CSP who died later had preoperative LVEF of less than 25%. All of the 12 CSP who died within 4 hours in this series, required CPB and 8 of these patients were on CPB for more than 2.5 hours. More than half of our CSP (14 of 24) required re-exploration, which undoubtedly influenced the mortality rate.
Both, the CSP and OBP groups had a significant decrease in blood product use after rFVIIa treatment but the short and long-term survival appeared to be worse in CSP probably due to multiple poor prognostic factors as discussed above. Poor outcome has also been independently associated with peri-operative transfusion of more than 7 units of R [
28]. The CSP and OBP who survived received an average of 18 and 11 units of R in the peri-operative period, respectively. The R requirement of patients who died within 4 hours in both groups was probably higher but could not be reliably estimated. All 5 of the patients in the OBP group who died within 4 hours suffered from major blood vessel injury as a result of stabbing or gun shot wounds. These patients were unstable upon arrival to the emergency room with hemoglobins in the range of 3–6 g/dL.
Excessive hemorrhage requiring massive transfusion can lead to hypothermia, DIC, excessive fibrinolysis, dilutional coagulopathy, and metabolic acidosis, which may further exacerbate bleeding and morbidity. [
14‐
17] A reduction in the pH to 7.0 nearly abolishes rFVIIa activity as reported by Meng et al [
18]. In this present series, there was no obvious correlation between the coagulopathy and mortality in both the CSP and OBP groups. Five of the OBP who died within 4 hours had a pH of 7 or less. However, acidosis did not appear to play an important role in CSP.
The optimal dose and timing of administration of rFVIIa for these patients is unknown. The standard dose (90 mcg/kg) was first used in hemophiliacs and was based on the
in vitro dose-dependent reduction in aPTT in plasma from these patients [
1,
29]. The 90 mcg/kg dose of rFVIIa corresponds to concentration of 1.25 mcg/ml FVIIa in plasma. Martinowitz et al [
7], used a median dose of 120 mcg/kg (range 120–210 mcg/kg) to achieve hemostasis in 7 cases of trauma. In 2 patients, 120 mcg/kg was insufficient, and 2 additional doses of 60 mcg/kg were required to achieve hemostasis. Kenet et al. [
30], have shown better efficacy with a 300 mcg/kg bolus dose followed by continuous infusion. The standard therapeutic dose of 90 mcg/kg of rFVIIa was used in this series of patients although a lower dose of 30 mcg/kg dose has also been shown to be effective [
8]. However, in considering the data from the aforementioned studies, it would appear that higher initial doses, or additional doses, might improve outcome. It is also tempting to speculate that earlier treatment with rFVIIa may have prevented rapid clinical deterioration and complications associated with massive blood product transfusion in this series of patients. This question would best be addressed in controlled studies using standard dosing protocols and well defined criteria for intractable hemorrhage. Although rFVIIa is expensive, it would appear to be cost effective when compared with the combined cost of large amount of blood products.
Previous clinical experience with rFVIIa supports a good safety profile in patients with hemophilia and trauma. Less than 1% of patients receiving rFVIIa had thrombosis and thrombosis related complications [
21,
31]. In our series, only one patient had a deep vein thrombosis of the left arm associated with an indwelling line and not considered to be due to rFVIIa. There were no other documented cases of thrombosis or microemboli.