Three or four factor prothrombin complex concentrate for emergency anticoagulation reversal? The true question is: what are we looking for?

Dear Sir,

We read with interest the letter by Marietta and colleagues¹ and agree that we should be looking at clinical instead of surrogate endpoints, such as the INR, when comparing the efficacy of 3 and 4 factor PCCs for warfarin reversal. The ideal trial to give this answer will be a large randomised double blind trial of 3 vs 4 factor PCC in patients on the same anticoagulant and where the dose, timing and route of intravenous vitamin K co-administration is the same, the PCC dose is the same and where randomisation is grouped according to INR, indication for anticoagulation and type of bleeding. Such a trial, however, is unlikely to ever be done.

In the absence of such a trial we are left with studies which vary enormously and where comparisons are fraught with danger. Because of the wide variability in patient factors and the way and timing that treatments are given, it is not easy to compare studies. As an alternative we would like the reader to consider the patient with an INR of 15.0 and a cerebral bleed. The patient is likely to have levels of factors II, VII, IX and X of <3–5%. Given at sufficiently high dosage both 3 and 4 factor PCCs will normalise the levels of factors II, IX and X, but with one product the factor VII will still be <5% whilst with the other will be >50%. It is unlikely that clinicians with expertise in managing patients with inherited FVII deficiency will be happy not to try and improve the FVII level of such a patient.

An axrgument for the safety and effectiveness of 3 factor PCCs despite incomplete correction of the INR is that it is the prothrombin fraction in PCC that is responsible for the correction of the bleeding tendency and that the FVII fraction is less important². The importance of prothrombin is shown by both in vitro³ and ex vivo⁴ experiments using thrombin generation showing a linear relationship between prothrombin and thrombin generation, whereas FVII has a hyperbolic relationship reaching normal thrombin generation at levels of around 5%. The prothrombin concentration after treatment with 3 factor PCC in the study by Holland et al. increased from 0.05 u/mL to 0.53 u/mL⁵. The latter is only at the lower limit of normal and given the linear relationship with thrombin generation, it is unlikely to fully compensate for the reduced levels of FVII (0.10 u/mL after treatment) that potentially continue to confer a bleeding risk.

Marietta and colleagues correctly raise the issue of safety of PCCs and suggest a reduced thrombotic tendency with 3 factor products based on the abstract by Dentali et al⁶. This meta-analysis showed a thrombotic incidence for 4 factor PCC of 2.3% versus 0.7% for 3 factor products but the 95% confidence intervals overlap (1.2–3.8% vs 0.0–2.4%) and the difference is not significant.

Whilst we agree with Marietta and colleagues¹ that we should concentrate on improving the use of PCC in patients with major bleeding in the context of coumarin anticoagulation, we remain convinced that for patients with very high INRs and truly life-threatening bleeding the 3 and 4 factor PCCs cannot be considered to be of equal efficacy at present. An approach whilst waiting for the evidence that 3 factor PCCs are clinically efficacious in patients with very high INRs is to also administer a small volume of fresh frozen plasma⁷.