Thursday, March 13, 2014

4-Factor PCC, Skepticism and Surrogate Endpoints

A guest post by Rory Spiegel (@CaptainBasilEM) who blogs on nihilism and the art of doing nothing at

Despite a lack of clinical data supporting their superiority, 4-factor PCCs have been universally accepted as the intervention of choice for the reversal of Warfarin induced bleeding. While PCCs have demonstrated rapid normalization of INR values, they have yet to show any added value over FFP in true clinically relevant endpoints. In the largest RCT to date, 4-factor PCC corrected INR values far faster than FFP but with a mortality rate that was almost double that of the FFP group. In a recent small RCT published in American Journal of Emergency Medicine, Karaca et al attempted to demonstrate that 4-factor PCCs provide more than just surrogate benefits when treating Warfarin induced gastrointestinal hemorrhage.

In this small unblinded trial, 40 patients with clinically suspected upper GI bleeds were randomized to have their coagulopathy reversed by either FFP or Cofact, a 4-factor PCC. The outcomes examined by the authors were INR values at 2 and 6 hours, time-to-endoscopy, and percentage of patients with active hemorrhage at time of endoscopy (based on the Forrest Classification). Patients were required to have their INR level reach 2.1 before undergoing definitive endoscopic interventions. As is typical in FFP vs PCC trials, the dose of FFP used was bordering on a straw man dose at 10-15 cc/kg.

To what should be no one’s surprise, 4-factor PCC reduced INR levels significantly faster than FFP, and thus patients in the PCC group underwent endoscopy earlier than their FFP counterparts. Patients in the PCC group received their endoscopy on average at 8 hours after admission, while the FFP group underwent endoscopy closer to the 12 hour mark. Endoscopy revealed more patients with active hemorrhage (Forrest Classification 1a or 1b) in the FFP group (7 patients vs 0) and sclerotherapy was performed in 10 patients in the FFP group, with 1 in the PCC group. Furthermore 3 patients in the FFP group required further therapy due to rebleeding, while no events of rebleeding occurred in the PCC group.

These superiorities in surrogate and pseudo-surrogate endpoints did not translate into the patient oriented endpoint of mortality, which was equivalent (one patient in each group). As far as the dreaded complication of “thrombolic events” that has been associated with PCC use it is somewhat unclear. One patient in the PCC group experienced a fatal IVC thrombosis but authors claim this condition was a presenting malady rather than an adverse event due to the administration of PCC.

Once again PCC has found success upon examination of soft endpoints. In an unblinded trial with surrogate and subjective endpoints, it is unclear if the PCC group’s performance was due to the medication’s efficacy or simply random chance and a small cohort. Until superiority in concrete clinically relevant outcomes is demonstrated, we should be wary of the crown of supremacy PCCs currently flaunt.

“Use and effectiveness of PCC vs FFP in gastrointestinal hemorrhage due to warfarin usage in the emergency department”


  1. I can't access this paper from our Turkish friends (merhaba ! ) but based on the abstract and Rory's comment :
    - was the endoscopic finding blinded (to the use of either FPP or PCC of course) ?
    - small isn't ugly: statistical tests takes into account the number of patients. Provided that the groups are truly similar after randomisation. Were they ? The odds that they aren't are high with this small number for a complex disease.
    - PCC does not do worse here it seems.
    - PCC has other benefits over FFP:
    - may be faster to administer (and prepare ) ?
    - viral inactivation
    - less volume ( usulally not an issue in the bleeding patient, but not all life threatening bleeds are active they may be impending and those patients are loikely to have some degree of heart failure

    But t'is more expensive

    1. Thanks for responding! As far as I can tell from the article there was no blinding to treatment groups and the physicians performing the endoscopy were aware of group allocation. The authors did not specify the particulars of how they selected groups but as per Table 2 the groups seemed reasonably balanced.

      You are right PCC in theory provides many advantages over FFP, but so far these advantages have yet to translate into clinically important benefits. In the largest RCT to date PCC corrected INR far faster than FFP and had far less cases of "fluid overload" (5.6% vs 12.8%) with an increase in thrombotic events (8.7% vs. 5.5%) and almost double the mortality rate (9.7% vs. 4.6%).
      I have written two posts discussing this paper in detail if you are interested ( and

      Your right these are all very small trials and these difference may in fact be due to chance alone. Presently 4-factor PCCs have yet to demonstrate any clinically relevant benefit when compared to FFP and as you said "t'is a bit more expensive"

      Thanks for writing!!

  2. I've seen them.
    But when there is a lot of physiologic logic and simple logic, not of the hair splitting type, and no decent clinical study , we can't infer the treatment does not work.
    Let's wait.
    Is there any study proving that reversing warfarin overdosing with vitamin K or dose adjustment reduces morbidity and mortality ?
    If not , why the hassle ?
    Just being skepticocynical ....

  3. I agree, there is a lack of definitive evidence supporting the superiority of any specific agent. I think that is my point. Until we have strong data supporting the added utility of PCC over FFP how can we claim its superiority with such certainty?

    I agree, there is a great deal of physiologic logic supporting PPC's use. Unfortunately this logic has yet to translate into clinical benefit. Conversely strong physiologic argument can be made as to why it may be harmful. Simply, it may cause an increase in thromboembolic events.

    As of know we don't have an answer and like you I prefer to be skepticocynical...

    Thanks you again for responding. I truly do appreciate the support

  4. Rory,

    Great analysis of the literature as always. I really enjoy your posts.

    I am a big fan of evidence based medicine... But perhaps the biggest challenge with EBM is when we have lack of good quality evidence to guide our practice. In the end, we have to make yes/no decisions with a patient in front of us.

    I currently practice in Australia and we have been using PCC's for many years now. (Certainly longer than the USA experience) When I am confronted with a patient on warfarin with an intracranial hemorrhage, I think immediate reversal is probably important. I know the best way to do this is with PCC's. Yes, the current data does not show any good patient oriented outcomes, but I still think it is a good thing to do in the current climate. I fully admit this is not an evidence based decision.

    PCC's are an order of magnitude cheaper than rFVIIa. In addition, there is better evidence to show tha rFVIIa is not a good idea. (i.e. more thrombotic complications and no decreased mortality despite smaller hematoma volumes).


    1. Hey Brian,
      Thanks again for writing! I agree EBM definitely has its own limitations and so much of our everyday practice is done on the peripheries of our current understanding of medicine. I'm sure in Australia and other countries which have had access to PCCs for many years have a far different perspective.

      I think my main opposition to PCCs is not whether they are a useful tool in warfarin reversal, but rather our tendency to declare their overwhelming superiority to FFP with no evidence supporting this declaration. As you mentioned, rFVII performs wonderfully if our goal is to lower a patient's INR and yet it has failed to demonstrating patient oriented benefits.

      Hopefully someone will publish a larger trial soon providing us with some guidance. Until then....

      Thanks again!!

  5. I am not sure mortality is the only "patient-centered" outcome here.

    While time to endoscopy is just a silly endpoint, there is certainly value in stopping bleeding and preventing "surgical interventions" (ie sclerotherapy).

    When a patient's chief complaint is vomiting blood and an intervention provides relief of that chief complaint (in fact, stops it completely), then I would say the intervention (PCC in this case) certainly had an effect on a patient centered complaint.

    If a patient presented with facial droop and tpa resulted in resolution of that facial droop (crazy, I know! -- but that's my point), would you not say that is a patient centered outcome even if mortality was unchanged?

    I am completely in the skeptics camp, but let's not let our rebellious nature overtake the primary focus of the rebellion.



Comments on posts > 10 days old will be moderated; blame spammers.