|Values are valid only on day of printing.|
Published: December 2008Print Record of Viewing
Dr. Stubbs discusses the use of
fresh frozen plasma for prevention of bleeding.
Presenter: James R. Stubbs, MD
Welcome to Mayo Medical Laboratories' Hot Topics. These presentations provide short discussions of current topics and may be helpful to you in your practice.
In the United States, the number of transfusions with FFP continues to steadily increase. In 1991, 2.3 million units were transfused, and by 2001 the number of transfused FFP units had increased to 3.9 million units which represents a 70% increase.
There is also evidence that FFP use is excessively high in US compared to other countries with similar advanced healthcare systems. In the US in 2001 the ratio of FFP units to red blood cell units transfused was 1:3.6. In contrast, in the United Kingdom in 2001, the ratio of FFP units to red cell ratio was 1:7.0.
This widespread use of FFP occurs despite the limited number of indications, outlined in published consensus documents. Despite the fact that the risks of such transfusions which include allergic reactions, transfusion-associated circulatory overload, and transfusion-related acute lung injury, are well known.
Because there are so few established indications for FFP transfusions listed in consensus guidelines, it is apparent based on the sheer number of FFP transfusions administered annually in the United States that a very large number of these transfusions are administered “outside these indications.” Many practitioners who prescribe FFP feel that such transfusions are required in situations other than those listed in established consensus guidelines.
What are the settings in which FFP is transfused outside of established guidelines?
At the Massachusetts General Hospital, a computerized physician-order entry system is used for all transfusions administered outside the operating rooms. The physician-order entry software requires that the person ordering the transfusion select the reason for the transfusion from a pick list. The reasons put forth by prescribing physicians for all FFP requests posted via the physician-order entry system were compiled during the period of February to April 2003.
During this time period, the most common reason for prescribing FFP was to “prepare a patient with an elevated INR for an invasive procedure.” This single indication represented 31% of all FFP ordered via the physician-order entry system, and 32% of all FFP units transfused outside the operating room.
So, what sort of evidence supports such a common practice?
Why do people give FFP prior to invasive procedures? The rationale underlying these transfusions is that even a mildly elevated INR is associated with excessive bleeding in the setting of an invasive procedure and that an intervention is needed for safety.
The assumptions driving pre-procedural use of FFP are:
Prophylactic FFP transfusions prevent or minimize the excessive bleeding that would have occurred in the absence of such transfusions.
The goal of this study was to systematically review the evidence supporting the assumption that an elevated PT/INR predicts bleeding in association with an invasive procedure.
Initially 682 abstracts were identified as candidate studies. From this, 25 studies ultimately survived the evaluation process and were included in this review. Only one of the included studies was a clinical trial, comparing transjugular liver biopsy to percutaneous liver biopsy with plugging.
The other studies were of two types. The studies were either case series in which patients were having a procedure of interest or they were case series in which the participants were enrolled because of abnormal coagulation test results.
Two articles evaluated the safety of bronchoscopy, three examined central vein cannulation, three evaluated femoral arteriography, thirteen evaluated liver biopsy, one studied paracentesis and thoracentesis, two studied renal biopsy, and one studied a mix of invasive procedures.
As an example let’s look at what investigators found when they looked into the data related to bronchoscopy. Two large studies looked at patients who underwent bronchoscopy procedures during which biopsies were performed.
Outcomes in patients with and without abnormal coagulation tests results were evaluated. The earlier of these two studies consisted of retrospective data collection covering the period from 1983 to 1991. The more recent study was based on prospectively collected data from the late 1990s.
Both studies identified rates of procedure-related bleeding that were similar in patients with normal and abnormal coagulation tests prior to the procedure. The differences in the risk of procedure-related bleeding were -2% and -3% respectively and interestingly, these results, on face value, although not statistically significant, show that the risk of procedure-related bleeding was lower in the patients with abnormal coagulation results.
The safety of central vein cannulation in the setting of abnormal coagulation testing results has been evaluated in three studies from the 1990s. The largest of these studies was published in 1999. This was a prospective evaluation of 580 patients with INR 1.5 or higher who had central lines places. Eighty-three percent of these patients also had a platelet count below 150,000/μL. Major bleeding occurred in one patient following central line placement for an overall event rate of 0.2%. Furthermore, this event was found to be due to inadvertent arterial puncture, therefore it was a technical performance issue rather than bleeding due to abnormal coagulation.
The other two studies evaluating central vein placement are consistent with the findings of the first study presented. Overall, 155 patients with abnormal coagulation testing results were evaluated in these two studies, and there were no episodes of major bleeding associated with the central line cannulation procedures.
It was concluded by the authors of these reports that bleeding complications associated with the placement of central lines in patients with abnormal coagulation testing results are rare, and that experienced physicians can safely perform such procedures in the face of these abnormal test results.
For the sake of time, I will not go through all the different invasive procedure settings evaluated by these authors individually, but will now look at the data from a more general perspective.
Overall, the authors identified 14 studies of patients undergoing invasive procedures that compared patients with and without abnormal coagulation testing results. In 12 studies, the difference in the risk of procedural-related bleeding ranged from -0.03 to +0.01. A negative number reflects a lower risk of bleeding for patients with abnormal coagulation test results, while a positive number reflects a lower risk of bleeding in patients with normal coagulation test results.
In two studies, the risk differences were +0.20 and -0.19 respectively. In all studies, the 95% confidence interval included zero, and all but one of the studies spanned from a negative to a positive number. Therefore, this data suggests that in these 14 studies, there was essentially no difference in the risk of bleeding between patients with and without abnormal coagulation testing in association with the invasive procedures that were performed.
The information in this study and others was used as the foundation for the development of guidelines on the assessment of bleeding risk prior to surgery or invasive procedures by the British Committee for Standards in Haematology.
I will quote some portions from the summary section of the document.
“Unselected coagulation testing is widely practiced in the process of assessing bleeding risk prior to surgery. This may delay surgery inappropriately and cause unnecessary concern in patients who are found to have ‘abnormal tests.’”
“Patients undergoing surgery should have a bleeding history taken. This should include detail of previous surgery and trauma, a family history and detail of anti-thrombotic medication. Pts with a negative bleeding history do not require routine coagulation screening prior to surgery.”
The following was stated in the recommendations portion of the document: “Routine coagulation testing to predict postoperative bleeding risk in unselected patients prior to surgery or other invasive procedures is not recommended.”
The take-home message is written on this slide. If History & Physical findings raise no index of suspicion regarding increased bleeding risk prior to invasive procedures – Do not perform coagulation testing!
This is an easy thing to say, however, I do believe it will be a hard instruction for many practitioners to follow because they want to believe that screening with coagulation tests, such as the PT/INR and aPTT actually help them identify patients with an increased risk of procedure-related bleeding.
You’ve seen this slide before, but now let us focus on the last two bullet points listed in bold print and pose the following questions:
Do FFP transfusions given in response to mild-to-moderate elevations of PT/INR actually correct the test result?
Do FFP transfusions given prior to procedures in response to mild-to-moderate elevations of the PT/INR actually prevent or minimize procedure-related bleeding?
Let’s get some historical prospective first. It’s clear that we’ve been scratching our heads about the value of pre-procedure FFP for a long time.
In 1966, Spector and colleagues published a study evaluating pre-procedure plasma transfusions in The New England Journal of Medicine. In this study, 13 patients with liver disease and elevated PT values ranging from 17- 26 seconds were transfused with 3-9 units of FFP.
In all cases, the PT failed to normalize, and the increase in coagulation factors achieved with the FFP transfusions were short-lived, with the levels declining by 50% within 4 hours of the transfusion (i.e., very little was achieved with these FFP transfusions in this patient population).
In a paper published in 2003, Youssef and co-workers conducted a two-part study, one part retrospective and one part prospective to evaluate the effectiveness of FFP in correcting the PT in patients with chronic liver disease.
In the retrospective analysis, a total of 80 patients were included. These patients had a pre-transfusion prolongation of the PT of more than 3 seconds above the control value, and this was despite having received parenteral vitamin K. All patients had evidence of cirrhosis of the liver, either by liver biopsy, or by clinical, biochemical and/or imaging criteria.
Twenty patients with liver disease and a prolonged PT were included in the prospective portion of the study. The indications for FFP transfusions in the prospective portion were: active bleeding, 4 patients; prior an invasive procedure, 11 patients; and concern for spontaneous bleeding, 5 patients.
In the retrospective portion of the trial, the mean PT prior to FFP transfusion was 17.5 seconds, the mean number of FFP units transfused was 3.75 units, and the mean PT following the FFP transfusion was 15.9 seconds. The mean change in PT following FFP, 1.5 seconds, was statistically significant, however, only 12.5% of the patients corrected their PT to within 3 seconds of their control value.
In the prospective portion of the trial, the mean PT prior to FFP transfusion was 20.0 seconds, the mean number of FFP units transfused was 2.9 units, and the mean PT following the FFP transfusion was 17.3 seconds. The mean change in PT following FFP, 3.75 seconds, was statistically significant, however, only 10% of the patients corrected their PT to within 3 seconds of the control value.
Now let’s turn our attention to a study from Massachusetts General Hospital published in 2006 in the journal of Transfusion. The authors stated that at the time of the conduct of their study, there were no studies examining the effect of FFP in hospitalized patients with mild prolongations of their PT.
For their purposes, this was defined as mildly prolonged PT up to 17 seconds or an equivalent mildly elevated INR up to 2.0. These mild abnormalities are very common among hospitalized patients and it is apparent that a large number of FFP transfusions are administered to this patient population. The investigators prospectively evaluated the effect of FFP on laboratory tests of hemostasis and whether there was any correlation between PT/INR and invasive procedure-related bleeding.
A prospective audit of all FFP transfusions from September 2, 2004 to June 30, 2005 was conducted at the Massachusetts General Hospital. During this time period, 1,091 units of FFP were transfused to patients with a PT of 13.1-17 seconds prior to transfusion. 471 of these 1,091 units were followed by a PT measurement within 8 hours of the transfusion. These 471 units were transfused to 121 different patients and 324 units were transfused during the first transfusion episode.
After the 324 transfusions that were part of the first transfusion episodes, the median decrease in PT was 0.20 seconds, while the median decrease in INR was 0.07 seconds. The proportion of patients achieving a normal PT, defined as less than 13.1 seconds or an INR of less than 1.1 following the FFP transfusions, was 0.0083. The PT/INR corrected to at least halfway to normal in only 15% of pts. No significant relationship was found regarding pre-transfusion PT value and the likelihood of correcting to at least halfway to normal following FFP. That is, there was no difference in the likelihood of correcting at least halfway to normal between patients with baseline INRs of 1.5 or lower and those with baseline INRs of more than 1.5.
A significant dose-response effect was not observed. Higher volumes of transfused FFP did not result in larger decreases in the PT post-transfusion. When patients receiving one unit versus two units of FFP were compared with regard to their likelihood of achieving a correction of their PT to at least halfway to normal, there was no significant difference between these two groups.
Also documented was the proportion of patients who achieved at least a 50% correction of their PT/INR, who received not more than one red cell unit, which was 15.9%, versus the proportion who received more than one red cell unit, which was 15.7%. This finding suggests that the effect of FFP on coagulation testing was essentially the same in stable and more actively bleeding pts.
The mean time to obtain a post-transfusion PT was 3.4 hours. No relationship between post-transfusion PT values and the time following FFP transfusions was detected. The likelihood of achieving a correction of the PT/INR of at least 50% towards normal was no different when looking at those patients who had a post-transfusion PT obtained less than 4 hours after transfusion compared to those who had a PT obtained 4 or more hours after transfusion.
No significant correlations were found between estimated red cell loss and pre-transfusion INR, APTT, platelet count or creatinine values. Interestingly, the correlation coefficient relating the PT/INR to the estimated red cell loss was -0.182 which indicates that increasing pre-transfusion PT values were associated with decreased estimated red cell losses. The take-home message is that there is not a significant association between pre-transfusion PT values and the extent of red cell loss.
In their discussion, the authors wrote, “Regardless of the number of units of FFP transfused or the number of hours after FFP transfusion, FFP resulted in only trivial decrements of the PT.” They also stated “Despite commonly-held views, we found that mild-to-moderate elevations of the PT was corrected by FFP in almost no pts.” Furthermore, a 50% normalization of the pre-transfusion PT was accomplished in a small number of pts, and there was no relationship between the extent of the PT prolongation and the likelihood of obtaining a 50% normalization of the PT.
The authors found no significant correlation between the pre-transfusion PT/INR value and the extent of bleeding. This finding is consistent with the concept that mild-to-moderately abnormal PT results are not predictive and do not correlate with the extent of blood loss.
In conclusion, the data from this study disclose a trivial effect of FFP on mild-to-moderately prolonged PT/INR values, and they suggest that there is no relationship between such PT elevations and the extent of blood loss.
This leads one to question the common practice of administering FFP transfusions to patients with mild-to-moderate PT elevations. In fact, the authors state, “FFP transfusion to nonbleeding patients in response to a mild-to-moderately prolonged PT value cannot be supported” and “[t]he role of FFP transfusion to bleeding patients with a PT of 13.1 to 17 seconds, or an INR, 1.1-1.85, is not entirely clear but results from this study question both goals and assumptions surrounding the PT as a guide to therapy.”