Low-dose recombinant factor VIIa for massive bleeding

a Department of Haematology and Central Hematology Laboratory, Inselspital, University Hospital and University of Bern, Switzerland b Department of Anaesthesiology, Inselspital, University Hospital and University of Bern, Switzerland c Department of Intensive Care Medicine, Inselspital, University Hospital and University of Bern, Switzerland d Department of Emergency Medicine, Inselspital, University Hospital and University of Bern, Switzerland e Department of Visceral and Transplant Surgery, Inselspital, University Hospital and University of Bern, Switzerland f Department of Paediatrics, Inselspital, University Hospital and University of Bern, Switzerland


Introduction
Recombinant activated factor VII (rFVIIa) has been licensed in Europe for use in patients with haemophilia A or B with inhibitory alloantibodies, with acquired haemophilia, with congenital deficiency of coagulation factor VII, or with Glanzmann's thrombasthenia [1].Additionally, rFVIIa has been employed for prevention or treatment of massive bleeding in non-haemophilic patients as well [2,3].In contrast to the large number of published case reports and retrospective cohort studies on successful use of rFVIIa in various off-label situations, randomised placebo controlled trials report limited effects at best [4][5][6].According to the missing evidence for rFVIIa off-label use, a recently published European guideline on management of bleeding following major trauma states that rFVIIa should be considered only if first-line treatment with a combin-ation of surgical approaches, best-practice use of blood products, the use of antifibrinolytics and correction of severe acidosis, severe hypothermia and hypocalcaemia fail to control bleeding [7].Measurable effects of rFVIIa administration have been observed in several populations, such as reduced growth of haematoma in patients with intracerebral haemorrhage [8,9] and reduced need for blood products in trauma patients [10][11][12], in cardiac surgery [13], in obstetrics [14] or in urology [15].Furthermore, there are reports suggesting that early use of rFVIIa may prevent massive transfusion coagulopathy [16].Nevertheless, many studies failed to show a significant impact of rFVIIa on relevant clinical outcomes, in particular mortality [5,17].A meta-analysis of 22 randomised controlled trials in patients without haemophilia found a non-significant decrease in mortality and a significant reduction in the need for blood transfusion both in prophylactic and therapeutic trials [4].A recently published meta-analysis of 26 randomised controlled studies in non-haemophilic patients found in 12 therapeutic trials a non-significant decrease in mortality and no difference in control of bleeding or red blood cell administration [6].Side effects of rFVIIa, predominantly thromboembolic events, have been reported in off-label use, with various incidences ranging from <1% to 44% [18][19][20][21].While published meta-analyses did show no significant difference between patients who received rFVIIa and patients who received placebo with respect to the incidence of venous thromboembolic events [4,22], treatment with rFVIIa significantly increased the risk of arterial thromboembolic events [4,22], in particular among elderly patients [22].There was a significantly higher frequency of coronary events and a statistically non-significant trend toward an increased rate of cerebrovascular events among patients treated with rFVIIa compared to placebo [4,22].As for rFVIIa dose, among patients with central nervous system bleeding who received rFVIIa, there was a significant correlation between dose and rates of arterial thromboembolic events [22].Although many publications on off-label use of rFVIIa reported an effective median dose similar to that of standard indications, 90 μg/kg body weight [4,23], or even much higher, up to 200 μg/kg body weight [10], there are some studies indicating a therapeutic effect already at lower doses, in particular in certain indications, such as acute intracerebral haemorrhage [8], traumatic bleeding [11,12], retropubic prostatectomy [15] or cardiac surgery [24,25].In 2004, we had chosen to employ for off-label use at our institution a low rFVIIa dose of 60 μg/kg because it leads to a plasmatic rFVIIa concentration which still is 5-6 times higher than the minimal effective level [26], and because high rFVIIa doses could theoretically lead to excessive thrombin generation with consecutive activation of the anticoagulant protein C pathway by means of thrombin binding thrombomodulin [27].In order to regulate off-label use of rFVIIa at the University Hospital of Bern, an interdisciplinary working group was instituted in 2004.Representatives of the departments of anaesthesiology, emergency medicine, intensive care medicine, visceral surgery, paediatrics and haematology reached a consensus guideline for the administration of rFVIIa in patients with massive bleeding (table 1).This guideline is congruent with the recently published European guideline on management of bleeding after major trauma [7].The aim of the present study was to evaluate effect and safety of this consensus in-house guideline at the University Hospital of Bern.

Study design
This is a single-centre observational cohort study performed at the Inselspital, a tertiary care university hospital in Bern, Switzerland.Since it had been impossible to design and conduct a prospective randomised controlled study, the interdisciplinary working group agreed to prospectively document the experience collected since implementation of the institutional guideline for off-label use of rFVIIa (table 1).All patients treated with rFVIIa from January 1, 2005 to December 31, 2007 were included into this study.The treating physicians decided on the indication to administer rFVIIa.The haematologist on call verified the indication and provided the correct dose.This was possible since rFVIIa is stored at the Central Haematology Laboratory.Implementing a gatekeeper strategy, a haematology medical consultation is necessary in order to administer plasma derived or recombinant coagulation factor concentrates, including rFVIIa.

Study population
Patients from all clinical departments were included into this observational cohort study (table 2).This is a very heterogeneous population, exactly representing everyday practice at a tertiary care hospital.rFVIIa was also used off-label in 6 patients due to refusal of blood products by the patient; these patients were excluded from this analysis to avoid confounding of the data.The study was approved by the responsible Ethics committee (Kantonale Ethikkomission Bern) and conducted according to the institutional guidelines on observational studies.

Data collection
All clinical data were documented by the treating physician(s).A standardised data collection form was used.Missing clinical data were searched for in the original patient documentation, by the study group.Qualitative change of bleeding 30 minutes after rFVIIa administration (decrease, without change, increase) was assessed and documented by the treating physician.For some patients detailed quantitative data were available as bleeding rates (ml per hour) before and after rFVIIa administration.According to the current practice at our institution bleeding volume was estimated by means of the suction volume over time, the number of blood soaked gauze pads and serial haemoglobin measurements.Results of laboratory analyses were retrieved from the electronic data management system.Time and number of transfused packed red blood cells (RBC), fresh frozen plasma (FFP), and platelet concentrates (PLT) for all patients were available from transfusion service records.The end-points for follow up of the

Statistical methods
Clinical and laboratory data were collected in a database worksheet with Microsoft Excel (Redmond, USA).Endpoints were the number of transfused RBC, FFP and PLT before and after the application of rFVIIa, qualitative change of bleeding after rFVIIa administration (decrease, without change, increase), side effects of rFVIIa and inhospital mortality on days 1 and 30.Statistical analysis was performed with MedCalc (Mariakerke, Belgium).Due to the number of patients and the non-normal distribution of some parameters (D'Agostino-Pearson test), non-parametric tests were used for the statistical analysis of quantitative variables.The Kruskal-Wallis test was used for comparisons between various groups.A Wilcoxon signed rank test was used for paired samples, whilst a Mann-Whitney U test was used for independent samples.A p value of <0.05 was considered significant.

Figure 2
Transfusion requirement before and after rFVIIa.Change in transfusion rates of packed red blood cells (RBC), fresh frozen plasma (FFP), and platelet concentrates (PLT) from a time window of 4.9 hours (median; range: 2.5-8.8)before administration of rFVIIa compared to 24 hours thereafter for patients living more than one day after treatment (n = 63).3).

Effect of the use of rFVIIa
The severity of bleeding was reduced in 82% of the patients after application of rFVIIa (fig.1).The estimated bleeding rate decreased from 2750 mL/h (median; IQR 1000-4000) to 300 mL/h (IQR 50-2000; n = 14; p = 0.0005).The number of transfused RBC, FFP and PLT units (table 3) was significantly lower during the 24 hours after rFVIIa compared to the 4.9 hours (median; IQR 2.5-8.8)before rFVIIa (n = 73; p <0.0001, p <0.0001, p = 0.0094, respectively).The transfusion rates before and 24 hours after rFVIIa are shown in figure 2 for the patients still alive more than one day after treatment (n = 63).Despite greatly decreased transfusion requirements, haemoglobin and haematocrit levels remained stable during the 24 hours after rFVIIa administration (table 4).Coagulation studies show the typical rapid normalisation of PT due to high circulating rFVIIa levels.Of note, there is a spontaneous increase in "clot-

Side effects of the use of rFVIIa
A potential complication of rFVIIa was considered in 6/73 patients (8.2%).Two reports were related to patients who died within 24 hours after rFVIIa administration.One patient died from acute myocardial infarction (table 5, patient 10); although he had pre-existing massive left heart hypertrophy with a postulated reduced myocardial perfusion reserve, the acute myocardial infarction might be considered a rFVIIa side-effect.One patient died from massive pulmonary embolism after polytrauma (table 5, patient 4); postmortem examination revealed central and peripheral fat embolism, thus ruling out a side effect of rFVIIa.Four other patients with suspected rFVIIa related side-effects survived beyond the first 24 hours.In three instances, reported potential rFVIIa complications were either not confirmed or more likely to have occurred for other reasons.
In one patient the clinical suspicion of deep vein thrombosis was not confirmed by imaging studies.Another patient did not awake from anaesthesia after surgery of an extended adenocarcinoma of the distal oesophagus; prolonged hypotension to 80/45 mmHg was documented during the massive bleeding episode, suggesting hypoxic brain damage.Another patient developed critical ischaemia of a lower limb, probably related to infrarenal aortic dissection.Finally, one patient developed hemiparesis after ECC for aortic and mitral valve replacement; although stroke is not infrequent after open heart surgery [28], the aetiology of this hemiparesis could not conclusively be determined.In summary, we registered two arterial thromboembolic complications possibly related to the use of rFVIIa: a lethal myocardial infarction (table 5, patient 10) and a post ECCsurgery stroke.None of the 20 patients treated with tranexamic acid (table 3) developed thromboembolic events.

Discussion
The present observational cohort study in a mixed population of a tertiary care hospital shows that 60 μg/kg rFVIIa was effective and safe in massive bleeding.Clinically relevant bleeding before the use of rFVIIa was documented by an estimated bleeding rate of 1 L/h, a total volume replacement of median 11.9 L (table 2), and the number of administered blood products: median 14 RBC, 11 FFP and 2 PLT units (table 3).The majority of the patients (79%) received just one application of rFVIIa and only two patients received more than two doses (table 3).After rFVIIa the estimated bleeding was reduced in 82% (fig.1), significantly fewer blood products were applied (fig. 2 and table 3), haematocrit remained stable and fibrinogen levels increased (table 4), indicating that bleeding was satisfactorily controlled.Our results are consistent with former reports of rFVIIa in off-label indications, which however, used much higher doses [10,29,30].In addition, we observed an inhospital mortality of 14% during the first 24 hours and 32% at day 30, which compares very well with the hospital mor- tality rates reported by several recent studies in massive transfusion [31].
The incidence of possible rFVIIa related thromboembolic complications among the 73 patients in our study -one lethal acute myocardial infarction in a patient with massive Weight adjusted dose, μg/kg -64 (56-75) Time from 1 st to 2 nd application, h -  ).This is at the lower end of the reported incidence of thromboembolic complications after rFVIIa, ranging from <1% up to 44% [18][19][20][21], and compares very well with the rate of 3.2% arterial thromboembolic complications reported in placebo groups [22].Even if all six recorded complications were considered to be due to rFVIIa, their proportion (6/73, 8.2%) is well in line with the rate of 8.7% thromboembolic events, both venous and arterial, observed among patients who received placebo [22].Administration of 60 μg/kg rFVIIa theoretically leads to a plasmatic rFVIIa concentration which still is 5-6 times higher than the minimal effective level [26].A few studies have shown an effect of rFVIIa in even lower doses in specific populations: e.g.In addition, the risk of thromboembolic complications is likely to be lower at a lower rFVIIa dose, as indicated by the significant dose-effect correlation observed among patients treated with rFVIIa for central nervous system bleeding [22].Finally, rFVIIa efficacy may even be better with lower doses, as shown by a meta-analysis -reporting a significant reduction of transfusion requirements if the rFVIIa dose was ≤ 90 μg/kg but not if it exceeded 90 μg/kg [4] and suggested by a post-hoc analysis of patients receiving rFVIIa for bleeding after haematopoietic stem-cell transplantation [32].

Limitations of this study
The present study is an observational study and not a randomised placebo controlled trial.Therefore, the effectiveness of rFVIIa compared to placebo cannot be assessed.The number of investigated patients is small and represents a heterogeneous group.Complete laboratory values were available only for a subset of the patients.

Strengths of this study
The present study reflects "real-life" in a tertiary care hospital, at which it had been possible to develop an interdisciplinary consensus guideline.All patients receiving off-label rFVIIa during a three year period were included in our registry.Complete data for demographics, clinical charts, procoagulant-drug treatment, transfusion requirement, qualitative bleeding, observed complications, and mortality were available.Of particular note: 1.We extended the observational period till death, hospital-discharge or day 30, whichever came first; 2. we present details on all patients who died during the first 24 hours after administration of rFVIIa; and 3. we openly discuss all patients with perceived complications due to rFVIIa.
Hypotheses generated by this study Firstly, we observed that transfusion requirements significantly decreased after administration of rFVIIa.This suggests that early use of an effective procoagulant drug such as rFVIIa in massive bleeding may actually be cost-efficient -particularly if lower doses can be used [12,33] by preventing administration of additional blood products and procoagulant drugs, further surgical interventions, and/ or medical complications.Secondly, we have been able to show that rFVIIa could control clinically severe bleeding in patients with median fibrinogen levels of 1.48 g/L (IQR: 1.21-1.69).This suggests that rFVIIa can be effective without the need of aggressive fibrinogen substitution, as indicated by a recent study in a porcine model [34] .None of the 20/73 (27%) patients, who had received tranexamic acid developed thromboembolic events, suggesting that combining low dose rFVIIa with this antifibrinolytic drug is safe.All these hypotheses should be tested in appropriate prospective randomised trials.

Conclusions
This observational single-centre cohort study documents that 60 μg/kg rFVIIa, a significantly lower dose than commonly used, is effective in reducing massive bleeding and transfusion requirements in off-label use, with a low complication rate.The heterogeneous population, which represents everyday practice at a tertiary care hospital, may allow a wide applicability of our study results.However, there is a clear need for large controlled randomised trials powered to evaluate clinical relevant endpoints.rFVIIa should still be used restrictively in off-label indications before having more evidence on efficacy and safety.
PS created the database, collected data, performed the statistical analysis, analysed and interpreted the data, wrote and revised the paper; AM collected data, analysed and interpreted the data, wrote and revised the paper; ML, BR, HPK, HZ, DI, AH, and BL determined the guidelines (table 1) and revised the paper; LA conceived the study, determined the guidelines (table 1), analysed and interpreted the data, wrote and revised the paper; all authors gave final approval.
The authors would like to thank Gabriela Barizzi, Dr. Raffaella Hesse, Therese Jost and Valérie Kurta for the excellent support in data acquisition.The authors acknowledge the work of the staff physicians in returning clinical information with the registry forms.

Figure 1
Figure 1 Qualitative bleeding change after application of rFVIIa.(A) Qualitative change of bleeding 30 minutes after rFVIIa administration is shown for all patients with known information (n = 62), patients living longer than one day (n = 52), and patients deceased within 24 hours after application of rFVIIa (n = 10).(B) Same information according to patient category.Patients with decreased bleeding are shown in white, patients without change of bleeding are shown in grey, and patients with increased bleeding after rFVIIa are shown in black.

Figure 3
Figure 3 Survival at day 1 and day 30.Survival and in-hospital mortality data on day 1 and up to day 30 for all patients and for categories of patients after off-label use of recombinant factor VIIa (rFVIIa).Deceased patients are shown in black, living patients are shown in white.

Table 1 :
Guideline for off-label use of rFVIIa for treatment of massive bleeding at the Inselspital, University Hospital, Bern, Switzerland.
Exclusion criteriaPatient with a non-treatable condition likely to lead to death soon, such as non-treatable metastatic cancer, severe liver cirrhosis without indication for liver transplantation, very severe traumatic brain injury, pre-terminal chronic disease Absolute contraindications Non-compensated DIC or severe sepsis Acute myocardial infarction Acute ischaemic cerebral stroke Relative contraindications Severe coronary heart disease Previous ischaemic cerebral insult Antifibrinolytic drug The use of an antifibrinolytic drug like tranexamic acid is recommended before the use of rFVIIa Haematologist consultation No rFVIIa without consultation Dosing First dose 60 μg/kg body weight i.v.[35, 36] Second dose Repeat same dose after 60 to 120 minutes if the bleeding is persistent but has decreased; in case of unchanged bleeding a repeat dose of 90 μg/kg body weight is allowed Third dose No third dose is advised [37] Follow-up Laboratory Obtain blood sample for coagulation tests 1 h and 12 h after use Documentation Return completed data sheet with clinical data

Table 2 :
Baseline characteristics of the patients receiving recombinant factor VIIa for massive hemorrhage.
* All values represent no./ total no.(%) unless otherwise specified.† Estimated bleeding rate was documented in 23 patients ‡ Some patients had combinations of various anticoagulant / antiplatelet drugs.ECC = extracorporeal circulation; IQR = interquartile range; rFVIIa = recombinant factor VIIa Original article Swiss Med Wkly.2011;141:w13213 Swiss Medical Weekly • PDF of the online version • www.smw.ch Haemostatic drugs employed to control bleeding are shown in table3.Tranexamic acid was used in 27% of patients.Protamine sulphate and aprotinin (before retraction) were predominantly used in patients after extracorporeal circulation (ECC).Desmopressin was mainly used in hepatic surgery.The number of surgical interventions to control bleeding included currently on-going surgery when rFVIIa was used during the operation (table3).
Swiss Med Wkly.2011;141:w13213 Swiss Medical Weekly • PDF of the online version • www.smw.chTreatment before the use of rFVIIa

table "
In five instances, bleeding decreased after rFVIIa but patients died because of other reasons.In one patient with mediastinitis and valvular prosthesis infection (table 5, patient 10) no significant reduction of bleeding after 65 μg/ kg rFVIIa was documented and he died following acute myocardial infarction (see below).It is noteworthy that

Table 3 :
Treatments applied in 73 patients with massive bleeding.

Table 4 :
Key laboratory values before and after recombinant factor VIIa administration*.

values Closest to start of bleeding Before rFVIIa After rFVIIa Day 1 after rFVIIa
* All values represent median (IQR); no. of values is 23-71 dependent on parameter and time point aPTT = activated partial thromboplastin time; h = hours; IQR = interquartile range; n/a = not available; PT = prothrombin time; rFVIIa = recombinant factor VIIa Original article Swiss Med Wkly.2011;141:w13213 Swiss Medical Weekly • PDF of the online version • www.smw.chleft ventricular hypertrophy and a hemiparesis after open heart surgery -was low (2.7% [12]derich et al.with 20 or 40 μg/ kg in patients after prostatectomy[15], Mayer et al. with a minimum of 40 μg/kg in intracerebral haemorrhage[8]or Stein et al. with a fixed dose of 1.2 mg in trauma patients with coagulopathy[12].The results of the present study indicate that the frequently used dose of 90 μg/kg can be safely reduced to 60 μg/kg in off-label indications.In our experience this dose was able to control bleeding, as subjectively judged by the treating physicians and objectively documented by the number of transfused blood products.

Table 5 :
Detailed analysis of the 10 patients who died during the first 24h after use of recombinant factor VIIa.
* Evaluation codes: A = no effect of rFVIIa was possible, due to uncontrolled surgical or traumatic bleeding (n = 4); B = rFVIIa was effective, other cause of death than bleeding (n = 5), C = potential rFVIla side-effect (n = 1).rFVIIa,recombinant factor VIIa.Swiss Medical Weekly • PDF of the online version • www.smw.ch / potential competing interests: No funding.LA is a member of the Swiss Advisory Board for NovoSeven ® .There is no other conflict of interest relevant to this article.Professor Lorenzo Alberio, Department of Haematology and Central Haematology Laboratory, Inselspital, University Hospital and University of Bern, CH-3000 Bern, Switzerland, lorenzo.alberio@insel.ch Correspondence: