Re-analysis of the PolyHeme Phase III trial dataset: Lessons for future haemoglobin-based oxygen carrier trauma trials

Abstract

Introduction: To assist design of future HBOC clinical trials for pre-hospital and prolonged field care, the haemoglobin-based-oxygen carrier (HBOC) Phase III trauma trial database comparing PolyHeme to blood transfusion was re-analysed to identify causes of adverse early outcomes versus the 30-day mortality outcome of the original trial. We questioned if failure of PolyHeme (10 g/dl) to increase haemoglobin concentration and dilutional coagulopathy versus blood, caused higher Day 1 mortality in the PolyHeme arm of the trial.

Methods: New analyses of the original trial database, including Fisher's exact test, examined impact of interval changes in total haemoglobin [THb], coagulation, fluid volumes administered and mortality on Day 1 in the Control (pre-hospital crystalloids, then blood after trauma centre admission) and PolyHeme arms of the trial.

Results: Admission [THb] was significantly greater (p<0.05) in PolyHeme (12.3 [SD = 1.8] g/dl) versus Control (11.5 [SD= 2.9] g/dl) patients. This early [THb] advantage was reversed within 6 h. Early mortality was negatively correlated with [THb] and maximum 1.4 h after hospital admission (17/365 for Control vs. 5/349 for PolyHeme). The mortality trend began reversing, when Control arm received blood. Coagulopathy was more common in the PolyHeme arm. Mortality rate was 2-fold greater for patients with coagulopathy in the control arm (18% vs. 9%, p = 0.1008) and 4-fold greater in PolyHeme arm (33% vs. 8.5%, p < 0.001). In a subgroup analysis of patients with major haemorrhage (n = 55), mortality was significantly higher in PolyHeme patients [12/26 (46.2%) versus 4/29 (13.8%) in control cohort (p = 0.018)], related to mean 10 liters more IV fluid administration and more severe anaemia (6.2 g/dL vs. 9.2 g/dL) in the PolyHeme cohort.

Conclusions: PolyHeme (10 g/dL) diminished pre-hospital anaemia. The inability of PolyHeme to reverse acute anaemia in a subset of major haemorrhage patients was due to volume overload secondary to high PolyHeme doses, resulting in dilution of clotting factors and low circulating THb (versus transfused controls) during the first 12 h of the trial. Haemodilution was associated with prolonged administration of PolyHeme, while blood transfusion was available to Control patients following hospital admission. Coagulopathy exacerbated bleeding, anaemia, contributing to excess mortality in the PolyHeme arm. Future trials for prolonged field care should evaluate HBOC with higher haemoglobin concentration, lower volume administration and transition upon trauma centre admission to blood plus coagulation factors or whole blood.

Keywords: Haemoglobin-based oxygen carrier; PolyHeme clinical yrial Re-analysis; Pre-Hospital; Trauma.