Literature Review
Systematic Review: Benefits and Harms of In-Hospital Use of Recombinant Factor VIIa for Off-Label Indications
Veronica Yank, MD; C. Vaughan Tuohy, BS; Aaron C. Logan, MD, PhD et al
Annals of Internal Medicine. 2011; 154: 529-40.
Reviewers: Manchula Navaratnam, MBChB, FRCA and Glyn D Williams, MBChB, FFA
Lucile Packard Children’s Hospital and Stanford University Medical Center
Introduction
Complex congenital heart surgery, especially neonatal and ‘re-do’ operations, may be associated with significant bleeding, large volume blood product transfusion, delayed chest closure and prolonged intensive care stay. This has developed an impetus to try a wide variety of pharmacological agents for the prophylaxis or treatment of bleeding in these high–risk patients. One of these agents is recombinant activated factor 7 (rVIIa) (NovoSeven® RT, Novo Nordisk, Bagsvaerd, Denmark).
rVIIa is approved by the USA Food and Drug administration for the treatment of bleeding in patients with hemophilia A and B who have antibodies to factors 8 or 9 and in patients with congenital factor 7 deficiency. The last decade, however, has seen a widespread adoption of its ‘off-label’ use in hospitals across the USA despite any strong evidence to support its purported benefits.
Logan et al (1) recently highlighted the extent of this off-label prescribing by conducting a retrospective database review of the Premier Perspectives database, which holds records on 615 non-federal US hospitals throughout the USA. This study found a 140-fold increase in the off-label, in-hospital use of rVIIa from 2000 to 2008. By 2008 97% of all in hospital use of this drug was ‘off-label’, with the most common indications being adult and pediatric cardiac surgery, brain and body trauma and intra-cranial hemorrhage.
Despite the widespread off–label use of rVIIa to stop bleeding, there are increasing concerns about its efficacy and safety and recent adult studies suggest an increased risk of thromboembolism (2,3). Yank and colleagues conducted a systematic review to examine the benefits and harms of rVIIa for the following five off-label in-hospital indications:
cardiac surgery, intra-cranial hemorrhage, trauma, liver transplantation, prostatectomy.
Selection and ranking of studies
Their search methodology examined 10 widely used databases (including PubMed, EMBASE and Cochrane library) from inception until Dec 2010. Two independent reviewers screened titles and abstracts and subsequently undertook a full-text review of all randomized controlled trials (RCTs) and observational studies that compared rVIIa with placebo, alternative therapies or usual care for hospitalized patients for any of the 5 indications. Only studies that addressed direct (mortality or thromboembolic events) or surrogate (e.g. number of red blood cells transfused, length of intensive care unit stay) clinical outcomes were included. Study quality was assigned a score from poor, fair or good according to systematically applied criteria. For inclusion in meta-analysis RCTs had to be of good or fair quality and observational studies of good quality. Meta-analysis was only done if there were at least 2 studies of fair or better quality, one of which had to be a good quality study. Each study was also assigned a score for strength of evidence, ranging from insufficient, low, moderate or high. Citations reporting use in pediatric patients were excluded.
Sixty-two (out of 6191 potentially relevant) citations met inclusion criteria and of these, 26 were included in the meta-analyses (16 RCTs and 10 observational studies) and 38 additional studies were included in the harms analyses. The investigators extracted 5 types of data from each included study: study design, population evaluated, rFVIIa dosing and administration, outcomes assessed, and study funding.
Results
The results from this systematic review highlighted the wide variability of rFVIIa dosing for off-label use ranging from 5- 400mcg/kg. None of the studies compared rFVIIa with predecessor products - only with placebo or with usual in-hospital care and none of them were allocated a ‘high’ strength of evidence quality rating. It was noted that with regards to outcome analyses, most of the studies tracked patients for only 30-90 days following the administration of the drug.
Cardiac surgery
Two RCTs and 4 comparative observational studies of adult cardiac surgery patients receiving rFVIIa met inclusion criteria. 2 RCTs (one good and one fair quality) and only 2 of the observational studies were deemed to be of sufficient quality for meta-analysis.
The 'good quality' RCT assessed prophylactic use of rFVIIa at the end of complex, non coronary artery bypass grafting surgeries and used mortality as its primary endpoint. This study was assigned a ‘low’ rating for strength of evidence. The other RCT was deemed ‘fair’ quality and evaluated the use of rFVIIa for post-op bleeding with the primary end point being thromboembolic events. The observational studies evaluated treatment for postoperative bleeding. Since timing and context was similar for all studies and rFVIIa was given after cardiopulmonary bypass in all cases the authors considered it appropriate to combine these studies for meta-analysis.
The authors concluded that there was moderate strength of evidence indicating that rVIIa increased the risk of thromboembolism (arterial and venous) in cardiac surgery(risk difference of 0.05, confidence interval 0.01 to 0.1) and low strength of evidence suggesting no mortality benefit. They commented that red blood cell transfusion requirements were possibly reduced with rFVIIa, but the trend was apparent only across higher-quality studies and they did not find consistent results regarding length of stay in the intensive care unit.
Other indications
There was moderate strength of evidence suggesting that the use of rFVIIa for patients with intra-cranial hemorrhage who were not receiving oral anticoagulation had no significant benefit for mortality or functional outcome and an increased risk for arterial thromboembolism. Interestingly for body trauma, available evidence of moderate strength identified no increased risk for thromboembolic events, a reduced risk for ARDS (risk difference 0.05, confidence interval 0.02 to 0.08), and no difference in mortality. The authors concluded that the available evidence for the remaining indications was too limited for meta-analysis.
Harms analyses showed that mortality rates were lowest in RCTs and highest in observational studies. The authors felt this gave reason for additional concern because observational studies resembled unselected patient populations more closely than RCTs. There was no apparent relationship between rVIIa dose and harms outcomes.
Limitations of the review
This is a robust and well-conducted systematic review. One of its strengths lies in the fact that they did not pool data from heterogeneous studies and only studies of similar interventions and patient populations were included in meta-analysis. It may be argued that the use of quality criteria to determine which observational studies were included in the meta-analyses may have lead to some selection bias. However, after performing qualitative sensitivity testing, the investigators found that the poor-quality studies had similar findings to their better quality counterparts.
It may also be argued that any definitive conclusions from this meta-analyses is limited by the fact that the strength of evidence for all outcomes was only low or moderate. In addition, the studies comparing rVIIa with usual care may be susceptible to bias from inter-hospital variations and clinical advances in ‘usual care’ over time.
Due to the small number of available studies the authors admit they were unable to perform formal assessment techniques to exclude publication bias. The authors also pointed out that the manufacturer of rVIIa (Novo Nordisk) played a substantial role in much of the RCT evidence and although this does not necessarily equate with biased research it does require special attention in evaluating for the possibility of bias.
Conclusions
This systematic review provides compelling data that the off label use of rVIIa which costs around $10,000 per adult dose may not benefit adult patients undergoing cardiac surgery and may substantially increase the risk for thrombotic events. The factors influencing peri-operative bleeding after open-heart surgery differ between adult and pediatric patients. Therefore, extrapolating data from adult studies may be problematic. However, given the current paucity of pediatric data regarding rVIIa use for cardiac surgery, pediatric anesthesiologists should, at the minimum, be aware of adult evidence. In this regard there may be lessons to be learnt from the aprotinin saga. A CCAS taskforce is currently developing guidelines on the use of rVIIa for pediatric cardiac surgery.
References
- Logan A.C, Yank V, Stafford R.S. Off-Label Use of Recombinant Factor VIIa in U.S. Hospitals: Analysis of Hospital Records . Ann Intern Med April 19, 2011 154:516-522.
- Zangrillo A, Mizzi A, Biondi-Zoccai G, et al. Recombinant activated factor VII in cardiac surgery: a meta- analysis. J Cardiothorac Vasc Anesth. 2009; 23:34-40.
- Levi M, Levy JH, Andersen HF et al Safety of recombinant activated factor VII in randomized clinical trials. N Engl J Med. 2010; 363:1791-800.
- Yank V, Tuohy CV, Logan AC, et al. Systematic review: benefits and harms of in-hospital use of recombinant factor VIIa for off-label indications. Ann Intern Med. 2011;154: 529-40.
