Prothrombin complex concentrate vs. fresh frozen plasma in adult patients undergoing heart surgery – a pilot randomised controlled trial (PROPHESY trial)

Summary There is equipoise regarding the use of prothrombin complex concentrate vs. fresh frozen plasma in bleeding patients undergoing cardiac surgery. We performed a pilot randomised controlled trial to determine the recruitment rate for a large trial, comparing the impact of prothrombin complex concentrate vs. fresh frozen plasma on haemostasis (1 h and 24 h post‐intervention), and assessing safety. Adult patients who developed bleeding within 24 h of cardiac surgery that required coagulation factor replacement were randomly allocated to receive prothrombin complex concentrate (15 IU.kg−1 based on factor IX) or fresh frozen plasma (15 ml.kg−1). If bleeding continued after the first administration of prothrombin complex concentrate or fresh frozen plasma administration, standard care was administered. From February 2019 to October 2019, 180 patients were screened, of which 134 (74.4% (95%CI 67–81%)) consented, 59 bled excessively and 50 were randomly allocated; 25 in each arm, recruitment rate 35% (95%CI 27–44%). There were 23 trial protocol deviations, 137 adverse events (75 prothrombin complex concentrate vs. 62 fresh frozen plasma) and 18 serious adverse events (5 prothrombin complex concentrate vs. 13 fresh frozen plasma). There was no increase in thromboembolic events with prothrombin complex concentrate. No patient withdrew from the study, four were lost to follow‐up and two died. At 1 h after administration of the intervention there was a significant increase in fibrinogen, Factor V, Factor XII, Factor XIII, α2‐antiplasmin and antithrombin levels in the fresh frozen plasma arm, while Factor II and Factor X were significantly higher in the prothrombin complex concentrate group. At 24 h, there were no significant differences in clotting factor levels. We conclude that recruitment to a larger study is feasible. Haemostatic tests have provided useful insight into the haemostatic changes following prothrombin complex concentrate or fresh frozen plasma administration. A definitive trial is needed to ascertain the benefits and safety for each.


Summary
There is equipoise regarding the use of prothrombin complex concentrate vs. fresh frozen plasma in bleeding patients undergoing cardiac surgery. We performed a pilot randomised controlled trial to determine the recruitment rate for a large trial, comparing the impact of prothrombin complex concentrate vs. fresh frozen plasma on haemostasis (1 h and 24 h post-intervention), and assessing safety. Adult patients who developed bleeding within 24 h of cardiac surgery that required coagulation factor replacement were randomly allocated to receive prothrombin complex concentrate (15 IU.kg À1 based on factor IX) or fresh frozen plasma (15 ml.kg À1 ). If bleeding continued after the first administration of prothrombin complex concentrate or fresh frozen plasma administration, standard care was administered. From February 2019 to October 2019, 180 patients were screened, of which 134 (74.4% (95%CI 67-81%)) consented, 59 bled excessively and 50 were randomly allocated; 25 in each arm, recruitment rate 35% (95%CI 27-44%). There were 23 trial protocol deviations, 137 adverse events (75 prothrombin complex concentrate vs. 62 fresh frozen plasma) and 18 serious adverse events (5 prothrombin complex concentrate vs. 13 fresh frozen plasma). There was no increase in thromboembolic events with prothrombin complex concentrate. No patient withdrew from the study, four were lost to follow-up and two died. At 1 h after administration of the intervention there was a significant increase in fibrinogen, Factor V, Factor XII, Factor XIII, a 2 -antiplasmin and antithrombin levels in the fresh frozen plasma arm, while Factor II and Factor X were significantly higher in the prothrombin complex concentrate group. At 24 h, there were no significant differences in clotting factor levels. We conclude that recruitment to a larger study is feasible. Haemostatic tests have provided useful insight into the haemostatic changes following prothrombin complex concentrate or fresh frozen plasma administration. A definitive trial is needed to ascertain the benefits and safety for each.

Introduction
Major bleeding that requires blood transfusion during cardiac surgery is associated with significant morbidity and mortality [1]. Standard care in the UK for replacement of clotting factors during bleeding in patients who are not taking vitamin K antagonists is fresh frozen plasma (FFP) [2].
The use of FFP in patients undergoing cardiac surgery is approximately 20-30% depending on the surgical procedure performed [3]. Evidence for the use of FFP in bleeding patients is lacking, and this was highlighted in a Cochrane systematic review [4]

Methods
The trial methodology and reason for its design have previously been published [6]. Inclusion criteria were adult patients (≥ 18 years), who were able to give consent and were undergoing elective or non-elective cardiac surgery (excluding procedures listed in online Supporting Information, Appendix S1) [6]. Written informed consent was obtained from all patients before surgery. For participants who were randomly allocated, clinical data were collected prospectively for up to 90 days after surgery (see online Supporting Information, Appendix S1), whereas for participants who were enrolled, but not randomly allocated (because they did not bleed), clinical data were collected prospectively for up to 24 h after surgery. bleeding. An FFP transfusion would be requested if thromboelastography was abnormal (see the algorithm in online Supporting Information, Appendix S1), and if thromboelastography or clotting test results were not available, it was administered in a 1:2 ratio with red cells, in line with national guidance [2]. Randomisation was by allocating participants in a 1:1 ratio to receive PCC or FFP using block randomisation, with block size varied randomly to ensure balance of treatments. The algorithm was written by the study statistician using the ralloc command in Stata.
Randomisation occurred via a web-based electronic database (REDCap) [7] for the first 5 months of the trial and was switched to manual randomisation envelopes for the next 3 months. Laboratory staff found paper randomisation easier and simpler to use during an emergency and the pilot study was designed to be responsive to learning and improvement ideas while the trial progressed.
Patients were randomly allocated to one of two groups: an intervention arm (four-factor PCC; Octaplex, Octapharma Ltd., Manchester, UK) or a standard arm (FFP).
Both products were stored in laboratories, but PCC was reconstituted at the bed-side. There is currently no agreed dose for PCC for the management of acquired bleeding disorders not related to vitamin K antagonists. The European Society of Anaesthesiology guidelines recommend a dose of 20-30 IU.kg À1 [8] and other observational studies have used doses ranging from 1000 IU (or 11.5 IU.kg À1 ) [9] to 1500 IU for three-factor PCC [10]. Based on these studies, and the current recommended dose for FFP in the UK ( The primary outcome was the proportion of eligible patients who consented and received the intervention within 24 h of surgery. Two months after the trial started, an amendment was made to the protocol to clarify the definition of 'recruitment' as those individuals who consented, were randomly allocated and received the intervention or control. For this trial to be successful, it was pre-specified that ≥ 30% of eligible patients must have agreed to participate, and ≥ 30% of consented patients must have been randomly allocated and received the intervention within 24 h of surgery in order to proceed to the full trial. Secondary outcomes included: proportion of patients where there was protocol adherence and protocol violation; difference in haemostatic capacity; time to administration of study drug; and safety up to 90 days after surgery (see definitions in online Supporting Information, Appendix S1). To assess haemostatic capacity, blood samples were taken at three time-points: before the intervention and during bleeding; within 1 h of the intervention being completed; and 24 h after the intervention (for sample handling, see online Supporting Information, Appendix S1). All samples were analysed by a biomedical scientist who was blinded to the patient group. The following tests were performed in accordance with the laboratory standard operating procedures: prothrombin time; activated partial thromboplastin time; Clauss fibrinogen; D-dimer; heparin levels (anti-10a and anti-2a); Factors II, V, VII, VIII, IX, X, XI, XII, XIII; von Willebrand antigen and activity; high molecular weight kininogen; prekallikrein; C1-inhibitor; antithrombin activity; protein C activity; free protein S antigen; a 2 -antiplasmin; plasminogen; tissue plasminogen activator; tissue factor activatable fibrinolysis inhibitor; prothrombin Factor I and II; thrombin-antithrombin complex; plasminantiplasmin complex; soluble endothelial protein C Receptor; thrombomodulin; tissue factor; and thrombin generation.
Sample size was based on estimating a consent rate of 30% of an expected 638 eligible participants over a 15-month period, with a 95%CI of AE3.5%. We also estimated that 30% of an expected 191 consented patients would go on to develop bleeding during, or within 24 h of, surgery that required FFP transfusion, allowing us to estimate a proportion of 30% with a 95%CI of AE6.5%. Based on the above, approximately 57 patients would be randomly allocated within 15 months, giving an expected final sample size of 51 patients completing the study after allowing for a 10% drop out or loss to follow-up. As this was a pilot study, no hypothesis testing was undertaken. Reasons for loss to follow-up were patients returned to their homes abroad and were not able to be contacted (n = 2), and patients were not reachable by telephone following several attempts (n = 2). The intention to treat group included 25 subjects in each group and per-protocol analysis included 21 subjects for both PCC and FFP arms.

Recruitment started in
Reasons for excluding eight patients from per-protocol analysis were due to protocol violations which included: three patients randomly allocated to FFP did not receive the intervention because bleeding had ceased by the time the products arrived; three had research blood sample violation; one was randomly allocated with a nonsequential envelope; and one randomly allocated to PCC had already received FFP. Baseline characteristics for enrolled and randomly allocated patients are presented in Table 1. Of the 50 randomly allocated patients, three in the FFP arm did not receive the intervention because they had stopped bleeding by the time the products had been Blood components received at 7 days; units † Blood components received at 30 days; units † We updated the most recent PCC systematic review [5] to assess the two safety outcomes (mortality and stroke) from this study, those from the review (all observational studies), and addition of any other studies (we identified one other observational study [12]) that have been published since the publication of the review (the search was updated on 27 June 2020). We observed no difference in all-cause mortality or stroke between people receiving PCC and those receiving FFP (online Supporting Information, Appendix S1, Figure S1).

Discussion
We have demonstrated that it is feasible to enroll patients  Other randomised controlled trials in the same setting of cardiac surgery have in some cases reported 50% violation rates [13], and this is partly due to study design, which in our trial sought to align interventions with clinicians' judgement on managing bleeding, rather than using a protocol-driven approach.   [12].
Antithrombin attenuates thrombin activity and its generation produces thrombin-antithrombin complexes, while plasminogen is the proenzyme of plasmin (regulated by a 2 -antiplasmin), the primary target of which is the degradation of cross-linked fibrin, with the end-product being dimeric D-domains (or D-dimers) [17]. No safety concerns were raised in our pilot trial. There is a long history of concerns regarding thrombotic events in association with PCC use, but this was not borne out in our trial when compared with FFP use. We updated the search report for the two safety outcomes (mortality and stroke) using the recent systematic review of PCC use in cardiac surgery and found no suggestion for increased rates of these adverse events.
In conclusion, our pilot study confirms that it is feasible to conduct a larger trial in the future. The doses selected for PCC and FFP in our trial appear well balanced as far as global haemostatic assays are concerned, but it remains to be determined whether these doses translate into different clinical benefits and safety for each of the interventions.