Autotransfusion has been included in the routine protocol in some units as an effort towards blood conservation. In this study we aimed to measure the efficacy and limitations of autotransfusion and whether a heparin-bonded circuit had any advantage. One hundred five patients were randomised to one of three post-operative treatments. Group 1 (n = 34) was not autotransfused whereas groups 2 (n = 36) and 3 (n = 35) received autotransfusion with the circuit of group 3 coated with heparin. Homologous blood and blood products were given according to strict protocols identical for all groups. Transfused and circulating blood was analysed for haemostatic variables and the requirement for homologous blood was recorded. Autotransfused blood contained no intact platelets and very high levels of D-Dimers (a peptide fragment released when fibrin is lysed) which resulted in high levels of systemic D-Dimers in patients receiving autotransfusion. Flow cytometric analysis revealed that whilst platelet glycoprotein 1 b receptors were severely reduced immediately following surgery, there was no additional damage caused by autotransfusion. Furthermore, there was no difference in platelet aggregation, von Willebrand factor (vWF) multimetric analysis or clotting profiles between the groups. Median (interquartile range) blood loss was 898 ml (638-1195) in group 1, 853 ml (595-1348) in group 2 and 770 ml (615-1000) in group 3 (Kruskal-Wallis P = 0.46). Median transfusion requirements were 2 units in each group. Whilst auto-transfusion does not appear to compromise haemostasis, it does not reduce the requirement for homologous blood and heparin-bonding of the circuit has no impact.