Prospective randomised evaluation of a collagen/thrombin and autologous platelet haemostatic agent during cementless total hip arthroplasty.

Abstract

Total hip arthroplasty (THA) can be associated with substantial peri-operative blood loss which can negatively influence a patient's clinical outcome. Few haemostatic agents have been tested in THA. The aim of this study was to determine whether the use of a collagen/thrombin/ autologous platelet haemostatic agent would result in a significant decrease of blood transfusions for patients undergoing primary THA. THA patients meeting inclusion/exclusion criteria (n=109) were enrolled in this prospective, double-blind trial and randomised to a treatment arm (standard haemostatic methods plus haemostatic agent) or control arm (standard haemostatic methods only). The primary outcome was transfusion. Secondary outcome measures included peri-operative narcotic usage and post-operative haemoglobin levels, pain scores, function, and general health quality of life. Transfusions were required by 5/60 (8.3%) patients in the treatment group and 7/49 (14.3%) in the control group (p=0.33). The mean number of units transfused was not significantly different between the treatment group (2.2±1.3) and the control group (1.6±0.5) (p=0.36). Haemoglobin values on post-operative days 1, 2, and 3 were significantly higher in the treatment group (p=0.002, 0.04, and 0.02, respectively). Hip Disability and Osteoarthritis Outcome Score and Short Form-12 scores were not different between the two groups. In relatively healthy patients undergoing primary cementless THA there was no significant difference in number of transfusions or number of units transfused. It is unlikely that we will routinely use the investigated haemostatic agent to reduce blood loss in a healthy patient undergoing THA. The product may have some benefit in patients who refuse blood transfusions, have minimal ability to increase blood volume, are undergoing total joint revision, or have markedly low pre-operative haemoglobin levels, but this needs to be demonstrated.