Presented at the American College of Emergency Physicians Scientific Assembly, Boston, MA, October 6, 2009.
The Efficacy of Factor VIIa in Emergency Department Patients With Warfarin Use and Traumatic Intracranial Hemorrhage
Article first published online: 1 MAR 2010
© 2010 by the Society for Academic Emergency Medicine
Academic Emergency Medicine
Volume 17, Issue 3, pages 244–251, March 2010
How to Cite
Nishijima, D. K., Dager, W. E., Schrot, R. J. and Holmes, J. F. (2010), The Efficacy of Factor VIIa in Emergency Department Patients With Warfarin Use and Traumatic Intracranial Hemorrhage. Academic Emergency Medicine, 17: 244–251. doi: 10.1111/j.1553-2712.2010.00666.x
- Issue published online: 1 MAR 2010
- Article first published online: 1 MAR 2010
- Received August 16, 2009; revision received September 1, 2009; accepted September 3, 2009.
- factor VII;
- traumatic intracranial hemorrhage;
Objectives: The objective was to compare outcomes in emergency department (ED) patients with preinjury warfarin use and traumatic intracranial hemorrhage (tICH) who did and did not receive recombinant activated factor VIIa (rFVIIa) for international normalized ratio (INR) reversal.
Methods: This was a retrospective before-and-after study conducted at a Level 1 trauma center, with data from 1999 to 2009. Eligible patients had preinjury warfarin use and tICH on cranial computed tomography (CT) scan. Patients before (standard cohort) and after (rFVIIa cohort) implementation of a protocol for administering 1.2 mg of rFVIIa in the ED were reviewed. Glasgow Coma Scale (GCS) score, Revised Trauma Score (RTS), Injury Severity Score (ISS), INR, and Marshall score were collected. Outcome measures included mortality, thromboembolic complications, and INR normalization.
Results: Forty patients (median age = 80.5 years, interquartile range [IQR] = 63.5–85) were included (20 in each cohort). Age, GCS score, ISS, RTS, initial INR, and Marshall score were similar (p > 0.05) between the two cohorts. Survival was identical between cohorts (13 of 20, or 65.0%, 95% confidence interval [CI] = 40.8% to 84.6%). There were no differences in rate of thromboembolic complications in the standard cohort (1 of 20, 5.0%, 95% CI = 0.1% to 24.9%) than the rFVIIa cohort (4 of 20, 20.0%, 95% CI = 5.7% to 43.7%; p = 0.34). Time to normal INR was earlier in the rFVIIa cohort (mean = 4.8 hours, 95% CI = 3.0 to 6.7 hours) than in the standard cohort (mean = 17.5 hours, 95% CI = 12.5 to 22.6; p < 0.001).
Conclusions: In patients with preinjury warfarin and tICH, use of rFVIIa was associated with a decreased time to normal INR. However, no difference in mortality was identified. Use of rFVIIa in patients on warfarin and tICH requires further study to demonstrate important patient-oriented outcomes.
ACADEMIC EMERGENCY MEDICINE 2010; 17:244–251 © 2010 by the Society for Academic Emergency Medicine
In 2000, an estimated 33 million persons over age 65 years lived in the United States, a number that is expected to reach 68 million by 2040.1 Traumatic brain injury is the fifth leading cause of death in patients aged 65 and over in the United States.2 With an increasing population of elders, it is estimated that the use of warfarin, the most common chronic anticoagulant prescribed in the United States, will increase as well.3,4 Multiple studies have associated warfarin use with an increased risk of traumatic intracranial hemorrhage (tICH) following blunt head trauma.5–11 Moreover, patients with preinjury warfarin use and tICH have higher mortality rates compared to patients with tICH not taking warfarin.9
Early detection and reversal of warfarin-induced coagulopathy is believed to improve outcomes in patients with tICH.12 In patients requiring immediate neurosurgical intervention, rapid and efficacious reversal to an appropriate international normalized ratio (INR) level is essential.13,14 Patients requiring neurosurgical procedures with preoperative INR levels of >1.25 have increased postoperative mortality.15 However, standard treatment such as vitamin K and fresh-frozen plasma (FFP) to reverse warfarin-induced coagulopathy are often problematic. Vitamin K requires 6 to 24 hours to reverse the coagulopathy and may fail to reverse warfarin coagulopathy in up to 40% of cases.15 FFP also fails to completely reverse anticoagulation in a substantial number of patients, largely because patients often require a large volume of FFP, which is both time-consuming and can create volume overload complications.16,17
Recombinant activated factor VII (rFVIIa, NovoSeven, Novo Nordisk A/S, Bagsvaerd, Denmark) was originally developed and approved for the treatment of bleeding in patients with hemophilia who developed antibodies to FVIII.16,17 Off-label use of rFVIIa has been documented in a number of bleeding conditions,18–22 including patients with severe blunt and penetrating trauma,23 head trauma,24,25 and spontaneous intracranial hemorrhage.26 Relatively small doses of 25 mcg/kg or less have been associated with reversing elevated INR values to below 1.3.27 Prior studies involving coagulopathic patients with tICH treated with rFVIIa have not shown any difference in mortality compared to standard treatment.25,28 However, the majority of patients enrolled in these studies were coagulopathic due to trauma rather than warfarin use. To our knowledge, there are no studies evaluating rFVIIa in coagulopathic patients with preinjury warfarin use and tICH.
The goal of this study is to compare outcomes in emergency department (ED) patients with preinjury warfarin use and tICH who did and did not receive rFVIIa for reversal of anticoagulation. We hypothesize that 1) rFVIIa will more rapidly reverse warfarin coagulopathy in patients with tICH compared to standard treatment and 2) patients treated with rFVIIa will have decreased mortality rates.
This is a before-and-after, retrospective cohort study of adult patients presenting to the ED with preinjury warfarin use and tICH. The institutional review board at the study site approved the study. The manufacturers of rFVIIa (NovoSeven, Novo Nordisk A/S) had no role in this study.
Study Setting and Population
The study was conducted at a single Level 1 trauma hospital with an annual ED volume of 65,000 patients. The ED evaluates approximately 85 adult patients with tICH per year. The hospital ED is staffed by board-certified emergency physicians and supports an emergency medicine residency program.
In November 2004, a protocol for patients on warfarin with blunt head trauma and tICH was implemented. This protocol included the use of a single vial of rFVIIa (1.2 mg) in addition to standard treatment in adult patients with the following criteria: 1) preinjury warfarin use, 2) confirmed tICH on cranial computed tomography (CT), and 3) an INR level of ≥1.3. This INR threshold was based on previous literature on this topic.13,29 Requests for rFVIIa are reviewed and monitored by the hospital pharmacy department, which maintains an electronic database of all requests for rFVIIa.
We searched the hospital trauma registry from January 1999 to March 2009 for eligible patients using the following criteria: 1) adult patient ≥ 18 years old and 2) International Classifications of Diseases, 9th revision (ICD-9) codes for tICH (codes 851–854). This search period included the earliest year ICD-9 codes were recorded in the trauma registry (January 1999) to March 2009.
We conducted a second search using the hospital pharmacy database to identify additional eligible patients. We searched for all patients with a request for rFVIIa through the hospital pharmacy. All rFVIIa requests are logged within a pharmacy computer database and approved by a pharmacist. The pharmacy database tracks requests and includes patient demographics, rFVIIa dosing, INR levels, and adjunctive treatment.
A list of patients identified from trauma registry and hospital pharmacy search criteria was compiled. Hospital medical records included a combination of electronic and hard copy patient information. Patient eligibility included 1) adult patients ≥ 18 years old, 2) the presence of tICH on cranial CT, 3) preinjury warfarin use, and 4) initial INR of ≥1.3. Patients eligible for rFVIIa in the postprotocol period but not receiving the medication were excluded from final analysis to focus on treatment effect. The presence of tICH was determined by documented attending radiology text report of the cranial CT. If the presence of tICH was read as questionable by the radiologist, patients would only meet eligibility if they had a hospital discharge diagnosis consistent with tICH. Warfarin use was determined on the medication list, ED physician note, or trauma physician admitting note. An initial INR of ≥1.3 was determined to be the first INR level recorded on the date of injury either at the study site or at the transferring hospital.
Data collection followed previously published guidelines on retrospective chart review.30 Data were collected on a standardized data collection form with predefined variables by two data abstractors trained to the methodology of data collection (DKN and JFH). Variables collected included age; sex; mechanism of injury; medical history; initial ED Glasgow Coma Scale (GCS) score; loss of consciousness; initial ED blood pressure; initial ED respiratory rate; alcohol intoxication; treatment with packed red blood cells (PRBCs), FFP, vitamin K, and rFVIIa; neurosurgical intervention and time to neurosurgical intervention; initial and repeat cranial CT results; initial platelet count; and the first five INR levels including date and time of each sample. Data were abstracted from the ED attending note if there were any discrepancies between physician and nursing notes. Alcohol intoxication was considered positive if documented in the physician note or if the laboratory ethanol level was ≥10 mg/dL. A unit of PRBCs equals 300 mL and a unit of FFP equals 600 mL. Cranial CT results were coded into anatomical categorical variables based on attending radiologist text report. Neurosurgical intervention was defined as the placement of an intracranial pressure monitor, burr hole, craniotomy, subdural drain, or intraventricular catheter. Time to neurosurgical intervention was defined as from the time of ED triage to the start of operation and was recorded in minutes. ED data were abstracted prior to knowledge of patient outcomes. Data collectors were not blinded to study hypothesis.
An Abbreviated Injury Score (AIS) for head and neck, face, chest, abdomen, extremities, and external body regions; overall Injury Severity Score (ISS); and a Marshall score were calculated and collected by data collectors who were trained in these calculations. The AIS and ISS are scoring systems developed to measure injury severity based on anatomical injuries divided by body regions.8,31 The Marshall score is a previously defined classification system that predicts patient outcome based on cranial CT findings (see Table 1 for classification of Marshall score).32
|Standard Cohort||rFVIIa Cohort|
|Age, yr (IQR)||81.5 (70.0–86.5)||71 (62.0–82.5)|
|Male (95% CI)||12/20, 60.0% (36.1–80.9)||14/20, 70% (45.7–88.1)|
|Initial GCS score (IQR)||14 (13–15)||14 (12–15)|
|Ethanol intoxication (95% CI)||3/20,15.0% (3.2–37.9)||1/20, 5.0% (0.1–24.9)|
|Isolated head injury (95% CI)||12/20, 60.0% (36.1–80.9)||11/20, 55.0% (31.5–76.9)|
|AIS head (IQR)||3 (3–4)||4 (3–5)|
|ISS (IQR)||16 (9–25)||16 (9–27)|
|RTS (IQR)||7.841 (7.841–7.841)||7.841 (7.006–7.841)|
|Initial INR (IQR)||2.51 (1.92–3.20)||2.87 (2.37–3.70)|
|Initial platelet count (IQR)||222 k/mm³ (162–246)||227 k/mm³ (188–267)|
|Marshall score* (IQR)||2 (2–2)||2 (2–5)|
Twenty-five percent of charts were randomly selected by computer program for duplicate data collection to measure interrater reliability.33 Both data collectors were blinded to results of data collection of the other data collector. Data missing from the medical record were coded as missing into the data set.
The primary outcome measure was in-hospital mortality. Secondary outcome measures included mortality at 48 hours and 30 days, successful reversal of anticoagulation, time to coagulopathy reversal (hours), worsening of intracranial hemorrhage, prevalence of in-hospital venous thromboembolism (VTE), hospital length of stay (days), intensive care unit length of stay (days), discharge Glasgow Outcome Score (GOS), and discharge GCS score. Mortality was based on review of hospital charts. Reversal of anticoagulation was considered successful if there was a documented INR level of <1.3 during hospitalization. This threshold was determined from prior literature.13,29 Time to reversal was calculated as the difference in time from the initial INR to the first INR recorded below 1.3. Worsening of intracranial hemorrhage was defined as an increase in Marshall score based on comparison of the first two cranial CT scans. VTE was defined as an imaging documentation of a pulmonary embolism or deep venous thrombosis during hospitalization. GOS and discharge GCS score were coded by data collectors through review of patient charts.
Data were entered into a spreadsheet and analyzed using STATA 10.0 statistical software (StataCorp, College Station, TX). Interval data were reported as the mean ± standard deviation (SDs) or median and interquartile range (IQRs). Proportions were presented with 95% confidence intervals (CIs). Categorical data were analyzed with chi-square test or Fisher’s exact test in cases of small cell size. Continuous data were analyzed with Student’s t-test if normally distributed data. Wilcoxon rank-sum test was used for nonparametric data or ordinal data. Interrater reliability was calculated by measuring the kappa statistic.
Forty-five eligible patients (median age = 80.5 years, IQR = 63.5–85.0) from January 1999 to March 2009 were reviewed. Five patients were identified after the rFVIIa protocol went into place and were eligible for rFVIIa, but did not receive rFVIIa. These patients were excluded from further analysis, leaving 20 patients in each cohort. No additional patients were identified through search of the hospital rFVIIa pharmacy database. Selection of eligible patients is described in Figure 1.
The most common mechanism of injury was fall from standing height or less (21 of 40 patients, 52.5%) and the most common indication for anticoagulation was a history of atrial fibrillation (31 of 40 patients, 77.5%). Patients had a median GCS score of 14 (IQR = 13–15), median ISS of 16 (IQR = 9–25) and median initial INR level of 2.72 (IQR = 2.17–3.51). The in-hospital mortality rate was 14 of 40, or 35.0% (95% CI = 20.6% to 51.7%). Patient characteristics, including mechanism of injury, initial INR, severity of injury (GCS, ISS), and severity of cranial CT findings (Marshall score) were similar (p > 0.05) between the rFVIIa and standard cohorts. All patients had a definitive read of tICH on cranial CT by the attending radiologist (see Table 1 for complete patient characteristics).
The mean (±SD) dosage of rFVIIa was 17.7 (±6.2) μg/kg. Patients were treated with rFVIIa a median of 2.1 (IQR = 1.7–3.3) hours after ED arrival. One patient received two doses of rFVIIa for persistently elevated INR. The remaining 19 patients received a single dose, and all doses were administered in the ED. All 40 patients received treatment with FFP; however, standard cohort patients received more units of FFP than rFVIIa cohort patients (standard group mean = 4.6 units, 95% CI = 3.4 to 5.7; rFVIIa group mean = 2.3 units, 95% CI = 1.7 to 2.9; p = 0.001). There was no difference in treatment with vitamin K between cohorts (see Table 2 for complete treatment characteristics).
|Standard Cohort||rFVIIa Cohort||Difference in Percentage or Means|
|PRBCs administered||7/20, 35.0% (15.4 to 59.2)||5/20, 25.0% (8.7 to 49.1)||−10.0% (−38.2 to 18.2)|
|PRBCs (units)||3.3 (1.8 to 4.8)||2.2 (0.8 to 3.5)||−1.1 (−2.9 to 0.8)|
|FFP administered||20/20, 100% (83.1 to 100)||20/20, 100% (83.1 to 100)||0% (−8.5 to 8.5)|
|FFP (units)||4.6 (3.4 to 5.7)||2.3 (1.7 to 2.9)||−2.3 (−3.5 to −1.0)|
|Vitamin K administered||16/20, 80.0% (56.3 to 94.3)||19/20, 95.0% (75.1 to 100)||15.0% (−5.0 to 35.0)|
|Neurosurgery||4/20, 20.0% (5.7 to 43.7)||7/20, 35.0% (15.4 to 59.2)||15.0% (−12.3 to 42.3)|
|Time to neurosurgery (hours)||74.6 (−70.5 to 219.7)||5.6 (2.1 to 9.2)||−69.0 (−143.7 to 5.8)|
No differences in in-hospital mortality were identified (7 of 20, 35.0%, 95% CI = 15.4% to 59.2% in both groups; p = 1.0). There was a trend to more thromboembolic complications in the rFVIIa cohort (4 of 20, 20.0%, 95% CI = 6.0% to 43.7%) than in the standard cohort (1 of 20, 5.0%, 95% CI = 0.1% to 24.9%; p = 0.34). No differences were observed in the other secondary outcomes of discharge GCS score and discharge GOS (Table 3).
|Standard Cohort||FVIIa Cohort||Difference in Percentage or Means||p-value|
|In-hospital mortality||7/20, 35.0% (15.4 to 59.2)||7/20, 35.0% (15.4 to 59.2)||0% (−29.6 to 29.6)||1.0|
|48-hour mortality||1/20, 5.0% (0.1 to 24.9)||1/20, 5.0% (0.1 to 24.9)||0% (−13.6 to 13.6)||1.0|
|30-day mortality||7/20, 35.0% (15.4 to 59.2)||7/20, 35.0% (15.4 to 59.2)||0% (−29.6 to 29.6)||1.0|
|Discharge home||6/20, 30.0% (11.9 to 54.3)||7/20, 35.0% (15.4 to 59.2)||5.0% (−24.0 to 34.0)||0.74|
|LOS (days)||16.0 (7.6 to 24.3)||15.4 (2.7 to 28.0)||−0.6 (−15.3 to 14.1)||0.93|
|ICU LOS (days)||11.5 (2.8 to 20.2)||11.8 (0.3 to 23.3)||0.3 (−13.7 to 14.3)||0.96|
|Mechanical ventilation (days)||8.3 (−0.7 to 17.2)||9.8 (−3.0 to 22.5)||1.5 (−13.6 to 16.6)||0.84|
|Thromboembolism||1/20, 5.0% (0.1 to 24.9)||4/20, 20.0% (5.7 to 43.7)||15.0% (−5.0 to 35.0)||0.15|
|Discharge GCS score||14 (IQR 3–15)||14 (IQR 3–15)||.||0.94|
|Correction to INR < 1.3||13/19, 68.4% (43.4 to 87.4)||19/19, 100% (82.4 to 100)||31.6% (10.7 to 52.5)||0.02|
|Hours to INR < 1.3||17.5 (12.5 to 22.6)||4.8 (3.0 to 6.7)||−12.8 (−17.2 to −8.3)||<0.001|
|Worsening Marshall* score||2/18, 11.1% (1.4 to 34.7)||5/16, 31.3% (11.0 to 58.7)||30.2% (−6.7 to 47.1)||0.53|
|GOS||3 (IQR = 1–5)||3 (IQR = 1–5)||1.0|
Thirty-eight of the 40 patients (95.0%) received a repeat INR measurement. Correction to normal (INR < 1.3) was more likely in the rFVIIa cohort (19 of 19, 100%, 95% CI = 82.4% to 100%) than in the standard cohort (13 of 19, 68.4%, 95% CI = 43.4% to 87.4%; p = 0.02). Time to normalization of INR was earlier in the rFVIIa cohort (mean = 4.8 hours, 95% CI = 3.0 to 6.7 hours) than in the standard cohort (mean = 17.5 hours, 95% CI = 12.5 to 22.6; p < 0.001; see Figure 2). The two patients that did not receive a repeat INR had severe tICH and died within 48 hours. Of the six patients in the standard cohort whose INR did not normalize, three patients died.
In patients who received a repeat cranial CT, there was no difference in progression of intracranial hemorrhage (defined as worsening Marshall score) between cohorts (standard cohort 2 of 18, 11.1%, 95% CI = 1.4% to 34.7% vs. rFVIIa cohort 5 of 16, 31.3%, 95% CI = 11.0% to 58.7%; p = 0.53).
There was a trend to more patients receiving neurosurgical intervention in the rFVII cohort (7 of 20, 35.0%, 95% CI = 15.4% to 59.2%) compared to the standard cohort (4 of 20, 20.0%, 95% CI = 5.7% to 43.7%; p = 0.29). There was also a trend to decreased time from ED arrival to neurosurgical intervention in the rFVIIa cohort (5.6 hours, 95% CI = 2.1 to 9.2 hours) compared to the standard cohort (74.6 hours, 95% CI = −70.5 to 219.7), p = 0.30. In patients with neurosurgical intervention, there was no difference in survival between cohorts.
There was good interrater agreement (κ > 0.60) in GCS score, AIS head, ISS, and Marshall score; neurosurgical intervention; use of FFP, RBCs, vitamin K, and rFVIIa; mortality; and GOS between the two abstractors.
To our knowledge, this is the first study that evaluated the efficacy of rFVIIa in patients with preinjury warfarin use and tICH. We compared a cohort of patients receiving a single dose of rFVIIa to patients not receiving this treatment. We demonstrated that administering rFVIIa significantly decreases time to correction of the patient’s INR.
Time to reversal of anticoagulation in patients with tICH is believed important to prevent ongoing hemorrhage. This significant difference in a laboratory outcome is consistent with prior studies on coagulopathy reversal with rFVIIa in nontrauma patients.34,35 These studies also demonstrated a rapid decrease in INR within 2–3 hours. While the reversal of anticoagulation is not a patient oriented outcome, many would argue that it is critical to reverse patients with tICH as soon as possible. Treatment of the patient’s coagulopathy aids in hemorrhage cessation and decreases the time for neurosurgical intervention.25
We were, however, unable to demonstrate any difference in patient-oriented outcomes (mortality or discharge GOS). The inability to translate normalization of INR to improved patient-oriented outcomes is consistent with other studies.25,36,37 The lack of improvement of mortality is also seen in studies involving much higher doses of rFVIIa24 and multiple doses of rFVIIa.23 One hypothesis is that normalization of INR does not correlate to the clinical hemostatic efficacy.34 With a number of studies showing no improvement in patient-oriented outcomes with the use of rFVIIa, the continued off-label use of this medication may not be warranted.
We found that there was a significant decrease in FFP utilization in the rFVIIa cohort compared to the standard cohort. The reduction in FFP is likely due to the rapid reversal of anticoagulation in the rFVIIa cohort, reducing the need for FFP. Decreased blood product utilization must be weighted against the costs of rFVIIa. One prior study found the use of rFVIIa to be more cost-effective than FFP in the reversal of coagulopathy in patients admitted to the intensive care unit with tICH.28
In the rFVIIa cohort, 20% of patients had a documented in-hospital VTE. While this was not statistically significant due to the small sample size, it is a potentially very important clinical outcome. The rate of VTE in the rFVIIa cohort is higher compared to that in prior studies evaluating the use of rFVIIa across a broad spectrum of indications.38 We hypothesize that due to the increased risk of VTE in trauma patients, administration of rFVIIa substantially increases the risk of VTE compared to those patients receiving rFVIIa without trauma.
The protocol called for a single vial of rFVIIa (1.2 mg), which translated to a mean dose of 17.7 (±6.2) μg/kg in the study population. Prior studies have administered larger doses and repeat dosages of rFVIIa.23,39 Prior literature has failed to document differences in mortality in low- versus high-dose rFVIIa across a broad range of clinical indications,39 as well as in patients with tICH.24 Other studies have shown limitation of the progression of hematoma size in patients with intracranial hemorrhage.24,26 Rather than measure the actual change in volume of hematoma, we instead evaluated the progression to a worse Marshall score. A change in a Marshall score would indicate a more substantial change in the progression of bleeding. We did not identify any change in Marshall score with use of rFVIIa.
Prior studies evaluating the use of rFVIIa in patients with coagulopathy following trauma have not shown any difference in mortality.23,25 However, no prior studies have evaluated the use of rFVIIa in patients with preinjury warfarin use and tICH. Our goal was to compare a single dose of rFVIIa to standard treatment to detect important patient-oriented outcomes. However, over a span of 10 years, we were able to enroll only 20 patients in each cohort. While there were equal rates of mortality in both groups, this small sample size may be underpowered to detect small differences in mortality. The study was also underpowered to detect other patient-oriented outcomes, specifically the rate of VTE. Post hoc power analysis of the rate of VTE shows that our study had a power of 0.29 and would require a sample size of 76 patients in each cohort to achieve a power of 0.80 to detect our observed difference of 15%.
While we collected various potential predictors of outcome such as ISS, GCS, Marshall score, age, and INR level, we did not collect all potential predictors. We did not collect information regarding preinjury aspirin and clopidogrel use, both of which potentially increase the risk of mortality in tICH. Moreover, there are other unknown potential confounders that were not collected for the study.
We identified five patients who were eligible for rFVIIa in the postprotocol period who did not receive rFVIIa. While there is no statistical significant difference in any of the markers of disease severity between the two cohorts, the exclusion of these patients may introduce selection bias. Patients who were eligible for rFVIIa, but did not receive it, may have been either too well appearing or too sick for clinicians to consider the use of rFVIIa to be of much benefit. We excluded these five patients from final analysis in the rFVIIa cohort to focus on the efficacy of treatment with rFVIIa. If we were analyzing the implementation of the protocol (effectiveness of rFVIIa), then we would have included these five patients in the rFVIIa cohort on the basis of intention to treat. We also did not include these five patients in the standard treatment cohort as these patients would have represented significant selection bias. Post hoc analysis of these five patients evaluated in the rFVIIa cohort on an intention-to-treat basis found similar results with our a priori analysis.
Finally, the study is retrospective and subject to the limitations of chart review methodology. We followed recommended guidelines for medial record reviews to minimize the limitations of this technique, and interrater reliability of the abstractors was good.30
In a patient population with preinjury warfarin use and traumatic intracranial hemorrhage, administration of recombinant activated factor VIIa is associated with a decrease in the time required to achieve normal international normalized ratio and with the administration of fresh-frozen plasma. However, there was no difference in mortality or other patient-oriented outcomes. Before widespread implementation of recombinant activated factor VIIa in patients with warfarin and traumatic intracranial hemorrhage on CT, a systematic study to demonstrate important patient-oriented outcomes is required.
- 1Age: 2000, Census 2000 Brief. Washington DC: US Census Bureau, 2001. Available at: http://www.census.gov/prod/2001pubs/c2kbr01-12.pdf. Accessed Nov 17, 2009..
- 2Geriatric trauma: patterns, care and outcome. In: MattoxKL, FelizianoDV, MooreEE (eds). Trauma. New York, NY: McGraw-Hill, 2000, pp 1099–114., , ,
- 39Recombinant factor VIIa for the prevention and treatment of bleeding in patients without haemophilia. Cochrane Database Syst Rev. 2007; 2:CD005011., , , .