Recombinant factor VIIa for the prevention and treatment of bleeding in patients without haemophilia
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
The objective of this systematic review is to determine the optimal use of recombinant factor VIIa (rFVIIa) when used for the prophylactic or therapeutic management of haemorrhage in patients without haemophilia.
Background
Factor VII is one of the key initial components of the coagulation cascade. This process contributes to the maintenance of haemostasis via the sequential action of blood coagulation factors which, together with other components such as platelets, form blood clots in response to vessel injury. When exposed by injury, tissue factor (a lipoprotein present in the membrane of vessel wall endothelial cells) binds to factor VII, thereby activating the coagulation pathway. This complex of activated factor VII (FVIIa) and tissue factor plays a pivotal role in the generation of coagulation factor Xa and thrombin. Thrombin is the main enzyme active at the end of the coagulation sequence and converts fibrinogen into fibrin − the major constituent of blood clots. Factor VII is synthesised in the liver and, like all vitamin K‐dependent coagulation factors, undergoes a final synthetic step of carboxylation to generate the full enzyme. Activation of this molecule is believed to occur by the feedback action of other coagulation factors, including thrombin (Butenas 2002).
A recombinant form of activated factor VII (NovoSeven®, Novo Nordisk, Denmark) has been manufactured and in clinical use for a number of years for the treatment of bleeding in people with haemophilia and inhibitory antibodies to factor VIII (Lusher 1998). However, more recently, it is the potential of factor VIIa to minimise or control severe or refractory bleeding in a variety of different medical and surgical situations that has engendered considerable clinical interest (Hedner 2002). In particular, it has been hypothesised that high‐dose factor VIIa is capable of enhancing haemostasis at the local site of injury without systemically activating the coagulation cascade, thereby reducing the risk of widespread inappropriate thrombosis throughout the circulation (Key 2003a).
The existing literature is dominated by case reports and small case series (Key 2003b). One recent study from the UK, based on outcome data for a series of 40 patients with no previous coagulopathy receiving factor VIIa for uncontrolled bleeding, reported that bleeding stopped or decreased in 32 (80%) of patients (O'Connell 2003). This retrospective study outlined the variety of clinical scenarios in which factor VIIa might have a role, including excessive surgical bleeding, uncontrolled medical bleeding, and trauma. Bleeding in these clinical settings has multiple aetiologies, including diffuse small vessel oozing, dilution of clotting factors and platelets from massive transfusion, disseminated intravascular coagulation, hyperfibrinolysis, hypothermia (with slowing of the enzymatic reactions in coagulation), and acidosis.
Standard therapy in these settings would typically include transfusion of blood components, correction of hypovolaemia and any biochemical/ metabolic changes, and surgical intervention for large vessel haemorrhage (if suspected). In view of current concerns about the supply of blood components because of new measures aimed at minimising the transmission of variant Creutzfeldt‐Jakob Disease (vCJD) by transfusion, the potential for factor VIIa to reduce transfusion requirements in these patients could also be particularly pertinent. In one early small case series of trauma victims, transfusion requirements decreased from 30 ± 18 units before recombinant factor VIIa to 3 ± 3 units after the administration of 1−3 doses of recombinant factor VIIa (Martinowitz 2001).
Elective surgical procedures may also be amenable to the transfusion‐sparing effects of factor VIIa. The potential of recombinant factor VIIa for reducing surgical blood loss was demonstrated in one randomised trial involving 36 patients undergoing retropubic prostatectomy (Friederich 2003). The authors of this study reported that median blood loss in placebo‐treated patients was approximately 2.7 L, but was reduced by more than 50% in those treated with a single intraoperative dose of 20 or 40 µg/kg recombinant factor VIIa. However, other studies have suggested that the administration of recombinant factor VIIa is associated with a much less dramatic reduction in blood loss (e.g. Lodge 2002).
One of the concerns about extending the use of a coagulation factor such as factor VIIa to different patient groups is the potential for adverse effects, in particular the risk of venous thromboembolism. Even in people with haemophilia, a slightly increased risk of thrombosis has been noted and reported (Roberts 2001).
Several prospective studies using factor VIIa have been published (Planinsic 2002; Hendriks 2001), and several others are under way or planned (Erhardtsen 2002). A systematic review of the literature is therefore proposed, with the intention of qualitatively and, as appropriate, quantitatively analysing the results from all randomised controlled trials.
Objectives
The objective of this systematic review is to determine the optimal use of recombinant factor VIIa (rFVIIa) when used for the prophylactic or therapeutic management of haemorrhage in patients without haemophilia.
Methods
Criteria for considering studies for this review
Types of studies
Randomised controlled trials.
It is intended that quasi‐randomised trials will only be included for descriptive purposes should the numbers of identified randomised trials be limited to one or two studies.
Types of participants
Patients at risk of blood loss due to surgery or who receive specific additional agents in order to treat bleeding.
All age groups will be considered. Patients with haemophilia or other haemostatic defects will be excluded (e.g. Glanzman's thrombasthenia, inherited factor VII deficiency).
Types of interventions
-
Randomised controlled trials comparing rFVIIa to prevent bleeding (e.g. before surgery) versus no rFVIIa.
-
Randomised controlled trials comparing rFVIIa to treat severe intractable bleeding (e.g. in the context of medical or surgical bleeding, or trauma) versus no rFVIIa.
-
Randomised controlled trials comparing rFVIIa with alternative treatments for the prevention and/or treatment of haemorrhage.
-
Randomised controlled trials comparing different dose schedules for rFVIIa.
·
Co‐interventions aimed at managing bleeding will be carefully documented for all studies.
Types of outcome measures
-
Survival, at fixed relevant time periods (one day to one week), with mortality evaluated by cause (i.e. as either haemorrhagic, an adverse effect of the intervention, or non‐intervention related).
-
Bleeding (within a pre‐defined follow‐up period post‐intervention): measured as response of bleeding (e.g. prevented, stopped, decreased, increased, no change), mean number and/or duration of bleeding episodes, or severity of blood loss (e.g. by volume, rate or bleeding score).
-
Number of red cell transfusions and blood product transfusions required (whether as units transfused or episodes, in a follow‐up period relevant to the surgery or trauma).
-
Number of patients avoiding transfusions (for prophylactic studies).
-
Adverse effects of interventions (e.g. thrombosis, allergic reaction).
·
Information on additional costs of therapy will be reported, if presented in the included studies.
Search methods for identification of studies
We will search the following electronic databases:
Cochrane Injuries Group Specialised Register
CENTRAL (Cochrane Central Register of Controlled Trials, The Cochrane Library)
MEDLINE (1951‐date)
EMBASE (1974‐date)
LILACS
mRCT (meta‐database of ongoing trials, including Controlled Clinical Trials)
National Research Register
The following strategy will be used in CENTRAL, and modified as appropriate for the other databases:
#1 FACTOR VIIA single term (MeSH)
#2 factor viia OR factor 7a OR rfviia OR fviia
#3 (activated NEAR factor seven) OR (activated NEAR factor vii) OR (activated NEAR rfvii) OR (activated NEAR fvii)
#4 novoseven
#5 #1 or #2 or #3 or #4
#6 HEMORRHAGE explode all trees (MeSH)
#7 hemorrhag* or haemorrhag* or bleed* or bloodloss* or blood near los*
#8 HEMOSTASIS explode all trees (MeSH)
#9 hemosta* or haemosta*
#10 surg* or trauma* or transfus*
#11 #6 or #7 or #8 or #9 or #10
#12 #5 and #11
In MEDLINE, this search strategy will be combined with phases 1 and 2 of the optimal randomised controlled trial search strategy described in Appendix 5b of the Cochrane's Reviewer's Handbook (Alderson 2003).
Database searches will be augmented by checking reference lists of RCTs identified and of relevant narrative reviews. No language or publication restrictions will be applied. Authors of known trials will be contacted to request information on any further trials of which they may be aware, whether published, unpublished or ongoing. In addition, the company manufacturing recombinant FVIIa will be approached. A handsearch of the abstracts of the meetings of the International Society for Thrombosis and Haemostasis will be undertaken, in addition to those of the American Society of Hematology (already handsearched and submitted to CENTRAL).
Data collection and analysis
Selection for inclusion
Two reviewers (SS, CD) will screen all titles and abstracts of papers identified by the review search strategy for relevancy to the review question. Only studies clearly irrelevant will be excluded at this stage. All other studies will be assessed on the basis of their full text for inclusion/exclusion using the criteria indicated above (type of studies, participants, interventions and outcome measures). At this stage, two reviewers will independently assess eligibility and any discrepancies in their assessments noted. If inter‐rater reliability is poor (kappa value less than or equal to 0.75) the clarity of the inclusion and exclusion criteria will be examined and improved until a good level of inter‐rater reliability is achieved (kappa value greater than 0.75) (Abramson 2001). Disagreements will be resolved by consensus in discussion with a third reviewer. All eligibility criteria failed by excluded studies assessed in duplicate will be recorded. Studies in which important information is lacking (including foreign language studies, where a translation is awaited) will be clearly categorised and reported as studies pending an eligibility decision.
Evaluation of methodological quality of included studies
The following criteria will form the main evaluation of methodological quality (adapted from Schulz 1995):
-
generation of random sequence
-
concealment of treatment allocation schedule
-
blinding of clinician (person delivering treatment) to treatment allocation
-
blinding of participant to treatment allocation
-
blinding of outcome assessor to treatment allocation
-
a minimum proportion of randomised participants included in the main analysis
-
equal use of co‐interventions in each study arm.
·
Problems in respect of these issues will be recorded in full. Particular account will be taken of loss to follow‐up affecting the validity of the results for different outcomes to different degrees.
A narrative summary of each of these criteria for each study will be presented in tabular format within the review. In addition, the criteria of generation of random sequence and concealment of treatment allocation schedule will be rated according to criteria identified in the Cochrane Handbook (Clarke 2003). This assigns ratings of A (adequate), B (unclear) or C (clearly inadequate) to each specified methodological criteria.
Where information is unclear or missing, clarification will be sought from the author of the primary study. Two reviewers will undertake the evaluation of methodological quality independently.
Evaluation of the methodological quality of each included study will be used in the following ways within the review:
-
either as a possible explanation for differences in results between studies or to investigate heterogeneity
-
in sensitivity analyses, examining the effect on overall estimates of excluding studies of poor methodological quality.
·
No study will be automatically excluded from the review as a result of a rating of 'unclear' (B) or 'clearly inadequate' (C).
Data extraction
Aside from details relating to study quality, the following two types of data will be extracted:
Study characteristics ‐ place of publication, date of publication, population characteristics, setting, intervention, comparator, and outcomes. A key purpose of these data will be to examine clinical heterogeneity in the included studies independently from the analysis of results. Potential sources of heterogeneity in this specific review include details of intervention (dose, frequency) and participant group (clinical setting).
Results of included studies in respect of each of the main outcomes indicated in the review question. The reasons why an included study does not contribute data on a particular outcome will be carefully recorded and the possibility of the selective reporting of results on particular outcomes considered.
The reviewers undertaking the evaluation of the methodological quality of the included studies will extract the data independently. Data will be extracted using study‐specific data extraction forms which will be created and piloted and their repeatability assessed. If repeatability assessment is found to be poor (kappa value less than or equal to 0.75) the data extraction forms will be re‐designed until good levels of repeatability are achieved (kappa values greater than 0.75) (Abramson 2001). Missing data will be requested from the original investigators. Disagreements will be resolved by consensus between the reviewers. Once disagreements have been resolved, the consensus data extracted will be recorded onto a third data extraction form. One reviewer will transcribe this into the systematic review computer software Review Manager 4.2 (RevMan). Another reviewer will verify all data entry for discrepancies.
Analysis of data
Extracted data will be analysed using the most up‐to‐date version of RevMan available at the time of analysis.
The main method of analysis will be quantitative, but the overall interpretation will be based on a balanced assessment of the patterns of results identified across the included studies. Generally, the preferred form of summary result will be a risk ratio (RR) for binary data and a weighted mean difference (WMD) for continuous data, both with 95% confidence intervals. For dichotomous outcomes, the numbers of outcomes in treatment and control groups will be recorded. For continuous outcomes, means and standard deviations will be recorded. In both cases, the 'denominator' will be the numbers randomly allocated to treatment and control groups. Assumptions that the assignment of a denominator introduces with reference to missing data will be considered and discussed within the review, as appropriate. Statistical advice will be sought when included studies present results that depart from these two scenarios, for example survival analyses and randomised cross‐over studies.
Meta‐analysis will also be employed, using a fixed effect model in the first instance. The results from random effect models will also be examined, with summary measures interpreted cautiously. Clinical heterogeneity will be assessed by examining differences in study quality, in the type of illness or type of surgery for which the participant is receiving care, and in the definition or measurement of outcomes of each study. Statistical heterogeneity will be tested using the chi squared test and by visual inspection of graphs. A p value of less than 0.10 will be interpreted as evidence of heterogeneity. This will be discussed in the Results section of the review. In the event of significant heterogeneity, the results from the included studies will not be pooled. Even in the absence of statistical heterogeneity, the robustness of any summary measures will be explored particularly with respect to study quality.
The following sensitivity analyses will be undertaken:
-
the influence of the methodological quality of the trials, analysing separately RCTs in which the generation of random sequence was rated as adequate (A) and RCTs in which the concealment of allocation was rated as adequate (A)
-
for dichotomous data, the influence of participant drop‐out, analysing separately RCTs with less than 20% drop out, RCTs with 20 to 50% drop out and RCTs with greater than 50% drop‐out.
Only if there are sufficient data will the following sub‐group effects be examined for:
-
dose and frequency of recombinant factor VIIa administration
-
any other sub‐group effects identified will be clearly indicated as hypothesis‐generating.
Although it is believed that every effort will have been made to identify unpublished studies, publication bias will be assessed using funnel plots. It is acknowledged that asymmetry, of which publication bias is one cause, is difficult to detect with the small numbers of studies (i.e. less than 10) often encountered in systematic reviews.
Care will be taken in translating the results of the included studies into recommendations for action. We plan to involve all reviewers in drawing conclusions and making specific recommendations for future research.
