Desmopressin for prevention of bleeding for thrombocytopenic, critically ill patients undergoing invasive procedures: A randomised, double‐blind, placebo‐controlled feasibility trial

Abstract Thrombocytopenic patients have an increased risk of bleeding when undergoing invasive procedures. In a multicentre, phase II, blinded, randomised, controlled feasibility trial, critically ill patients with platelet count 100 × 109/L or less were randomised 1:1 to intravenous desmopressin (0.3 µg/kg) or placebo before an invasive procedure. Forty‐three participants (18.8% of those eligible) were recruited, with 41 eligible for analysis. Post‐procedure bleeding occurred in one of 22 (4.5%) in the placebo arm and zero of 19 in the desmopressin arm. Despite liberal inclusion criteria, there were significant feasibility challenges recruiting patients in the critical care setting prior to invasive procedures.

Thrombocytopenic patients undergoing invasive procedures have an increased risk of bleeding [2], yet the role of platelet transfusions is unclear, and there is a need to consider alternatives given there are concerns about security of supply of platelets in many countries [3].
Desmopressin is a pro-haemostatic drug that is commonly used for patients with inherited bleeding disorders, and is under investigation for treatment of platelet dysfunction [4][5][6][7].There are currently no data evaluating the use of desmopressin in thrombocytopenic critically ill patients undergoing invasive procedures [8].
The aim of the study was to assess the feasibility of administering desmopressin or placebo to critically ill thrombocytopenic patients undergoing invasive procedures.

METHODS
We conducted a multicentre, phase II, double-blind, randomised, controlled feasibility trial across three intensive care units (ICUs) in the United Kingdom.Ethics approval was given by South Central-Oxford The procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation and with the Helsinki Declaration of 1975, as revised in 2013.
The allocation sequence was prepared using a fixed block size by an unblinded statistician independent from the trial statistician.Treatment allocation was with numbered sealed opaque envelopes.The members of the clinical team randomising participants and preparing the desmopressin/placebo were not blinded.They handed the reconstituted desmopressin/placebo to the patient's clinical team, and were not involved with the patient's trial-related care from this point and were not allowed to reveal the treatment allocation to other staff members.
Other clinical staff, the patient and outcome assessors were blinded to treatment allocation.
Participants were randomly allocated in a 1:1 ratio to either desmopressin 0.3 µg/kg in 50 mL 0.9% sodium chloride given intravenously over 20 min; or to 50 mL 0.9% sodium chloride given intravenously over 20 min.Desmopressin or placebo were administered prior to an interventional procedure (with a target of no more than 2 h before the procedure).
The primary outcome was the proportion of eligible patients meeting all inclusion criteria and none of the exclusion criteria, who were randomised to the trial and received the allocated intervention.A threshold of recruitment of 30% of eligible patients was used to define feasibility.
Clinical secondary outcomes were new onset of active bleeding up to 24 h after administration of desmopressin or placebo (measured using the HEmorrhage MEasurement [HEME] bleeding scale) [10]; exposure to blood components up to 24 h after desmopressin or placebo; thromboembolic events up to 28 days; and serious adverse events up to 28 days.Changes in measures of haemostasis were assessed as change in parameters from baseline to 30 and 120 min after infusion of desmopressin or placebo.
Confidence intervals presented are 95% and two-sided.The statistical package SAS version 7.13 was used to conduct analyses.The 95% confidence interval for the primary outcome was based on the binomial/Clopper-Pearson exact method.Continuous variables are reported as the median and interquartile range.When measurements were censored above a known value, the Kaplan-Meier estimator was used for reporting point-estimates for the percentiles of the empirical cumulative distribution.Discreet variables are reported as counts and %.Missing data are reported and no data were imputed.Outcomes were analysed using intention to treat or per protocol for adverse events.

RESULTS
Between a Missing data: age n = 1; renal failure n = 2; anticoagulant drugs given within 7 days of randomisation n = 2; procoagulant drugs given within 7 days of randomisation n = 2, procedure type n = 1.
Patient characteristics at baseline and invasive procedures were similar between groups (Table 1).The most common invasive procedures were CVC insertions or removals.Overall adherence to the protocol, including blood sampling and timing of interventional pro-cedures, was 25/42 (59.5%).The main barriers to adherence to the protocol were taking blood samples at the pre-specified times before procedures and ensuring that the procedures were performed within 2 h of the desmopressin or placebo infusion.
TA B L E 2 Outcomes.

Safety outcomes Placebo Desmopressin
Bleeding The participants allocated to desmopressin arm and who withdrew before the intervention are excluded.c No data available for one patient.d Thromboembolic events: In the desmopressin arm, there was one myocardial infarction (Day 1 after treatment).In the placebo arm, there was one myocardial infarction (Day 7), two deep vein thromboses (Day 9 and Day 20) and one ischaemic stroke (Day 6).
In the 24 h after the procedure, one participant in the placebo arm

DISCUSSION
The study successfully recruited and randomised patients to a comparison of desmopressin/placebo, but the overall proportion of potentially eligible patients recruited was low despite liberal inclusion criteria.The overall incidence of bleeding following invasive procedures in throm-bocytopenic patients in our study was low, as reported elsewhere [2].
No efficacy signal was seen for desmopressin on a range of outcomes nor markers of haemostasis.We had hypothesised that increasing VWF levels would compensate for thrombocytopenia [11].However, VWF levels did not increase, which may be due to baseline levels of VWF being very high because of acute illness [12].It is possible that desmopressin may still exert an effect in ways that were not assessed in this study such as enhancing the formation of procoagulant platelets [13].
In the study by van Baarle et al. [2], bleeding occurred predominantly in haematology ward patients undergoing subclavian CVC insertion, and not ICU [14].
Our study indicates that the area of research prophylaxis is challenging to deliver, as also found when studying plasma prophylaxis [15].
The number of eligible patients that could be randomised and received desmopressin/placebo as a proportion of those eligible was less than the 30%, anticipated.This did not meet the primary feasibility outcome, and suggests that it would be challenging to run a definitive efficacy trial on a larger scale.If a larger, definitive trial was conducted, approaches would be needed to address consenting and the urgent nature of many of the procedures that patients undergo, many of which are out of regular working hours.Centres with 24 h per day research cover are likely to capture more patients.Additionally, inclusion of procedures with a greater bleeding risk may increase the number of participants that could be recruited.
Strengths of our study were the randomised controlled trial methodology and the inclusion of a range of patients with thrombocytopenia.The rates of concordance with the clinical administration of desmopressin or placebo were high with the exception of timing of blood tests around the procedure and ensuring the procedure took place within 2 h of administration of desmopressin or placebo.
This trial demonstrated challenges in the running of a large definitive trial evaluating time-sensitive prophylactic interventions in critically ill thrombocytopenic patients undergoing invasive procedures.
Data presented as n (%). a Minor bleeding from a central venous catheter insertion site.b

( 1 /
22, 4.5%) experienced new active bleeding (minor bleeding from a central venous catheter insertion site) compared with none in the desmopressin arm.Four of 22 (18.2%) in the placebo arm received a red cell transfusion in the 24 h after randomisation, compared with three of 21 (14.3%) in the desmopressin arm.Thrombotic events within 28 days of randomisation occurred in four of 22 (18.2%) in the placebo arm and one of 19 (5.3%) in the desmopressin arm.The proportion of patients with at least one serious adverse event was 13/22 (59.1%) in the placebo arm and 11/21 (52.4%) in the desmopressin arm (Table2).There was one serious adverse reaction in the desmopressin arm where a transient unexpected episode of hypotension requiring increased inotrope requirements occurred during the infusion.Baseline von Willebrand factor (VWF) antigen levels in both armswere significantly elevated (up to 11.3 iu/mL) compared with the normal range for VWF antigen (0.5-2.0 iu/mL).There was no difference in VWF antigen, VWF activity, VWF collagen binding activity, platelet function analyser closure time and thrombin generation parameters for those treated with desmopressin or placebo (Supporting Information).
24.7%) of eligible patients (n = 213).This did not meet the primary feasibility outcome.The total number of patients randomised represented 54% (43/80) of all eligible patients when excluding out-of-hours cases.Three participants were lost to follow-up; one from the placebo group who was lost by Day 28 and two from the desmopressin group by Day 7 post treatment.TA B L E 1