Background and Objectives Primary resuscitation for massive haemorrhage often occurs in emergency departments or operating theatres, with ongoing resuscitation in the intensive care unit (ICU). The aim of the study was to retrospectively review transfusion practice in the pre-ICU phase and ICU for patients with massive haemorrhage. Materials and Methods From 1998 to 2006, we developed an electronically linked database of blood and blood product usage and laboratory data with clinical outcome. All patients who received 10 or more units of red cells and required ICU admission were included. Results Of 238 patients who required massive transfusion, 40 died early (within 24h of massive transfusion), out of which 16 died in pre-ICU and 24 died in ICU. Comparatively this group of patients presented in the pre-ICU phase and on ICU admission, respectively, with coagulopathy (median international normalized ratio 1·6 and 2·1) and acidosis (median base deficit -11·5 and -14mmol/l). These patients had median ratios of fresh frozen plasma (FFP) to red blood cells of 1:3·3 and 1:1·3 in the pre-ICU and ICU phases, respectively. Severity of coagulopathy indicated by INR at ICU admission [P=0·04; area under receiver operator curve (ROC)=0·69] and RBC transfused (P=0·01) in 24h associated with mortality. Conclusions Patients who died early were coagulopathic before and on ICU admission and did not correct their coagulopathy. This study also shows that coagulopathy is associated with an increased risk of mortality. Early and aggressive correction of coagulopathy for patients presenting with coagulopathy may be effective in improving mortality.