BACKGROUND: Transfusion guidelines have become increasingly restrictive. We investigated actual transfusion practices in geriatric hip fracture patients, why they differ from current guidelines and how this affects outcome. OBJECTIVES: The primary aim was to examine transfusion timing, evaluate how many red blood cell (RBC) transfusions are in keeping with guidelines and how this affects morbidity (infection, cardiac events and delirium), mortality and length of stay (LOS). Our secondary aim was to test the hypothesis that guidelines were more likely to be deviated from shortly before discharge and with consecutive transfusions. DESIGN: A retrospective observational study. SETTING: The Luzerner Kantonsspital, a major trauma centre, over a 12-month period from 1 February 2015 to 31 January 2016. PATIENTS: All patients over 70 years of age admitted to the Luzerner Kantonsspital with hip fractures over a 12-month period in 2015 to 2016 were included. RESULTS: 156 patients were included, to which 141 units of RBCs were transfused. All pre and intra-operative transfusions were according to guidelines; 110 transfusions were postoperative and 37 of these were not according to guidelines. Patients who were transfused had longer LOS in hospital (P = 0.002) and an odds ratio (OR) of 2.7 of contracting an infection (P = 0.04) in comparison with patients who were not transfused. No significant differences in mortality, LOS or morbidity were found between patients transfused according to guidelines and more liberal thresholds. Guidelines were more likely to be deviated from within the last 2 days before discharge than prior to this (58 vs. 24%, P = 0.03). Furthermore, 24 stable patients received two consecutive RBC units resulting in posttransfusion haemoglobin values of between 83 and 124 g l. CONCLUSION: Most RBC transfusions occur postoperatively, many still according to liberal transfusion thresholds, in particular shortly before discharge and as part of consecutive transfusions. Transfused patients had longer LOS and more infections than patients not transfused, but there was no difference in mortality, LOS or morbidity between patients transfused according to current guidelines and those where guidelines were deviated from.
BACKGROUND: Transfusion guidelines have become increasingly restrictive. We investigated actual transfusion practices in geriatric hip fracturepatients, why they differ from current guidelines and how this affects outcome. OBJECTIVES: The primary aim was to examine transfusion timing, evaluate how many red blood cell (RBC) transfusions are in keeping with guidelines and how this affects morbidity (infection, cardiac events and delirium), mortality and length of stay (LOS). Our secondary aim was to test the hypothesis that guidelines were more likely to be deviated from shortly before discharge and with consecutive transfusions. DESIGN: A retrospective observational study. SETTING: The Luzerner Kantonsspital, a major trauma centre, over a 12-month period from 1 February 2015 to 31 January 2016. PATIENTS: All patients over 70 years of age admitted to the Luzerner Kantonsspital with hip fractures over a 12-month period in 2015 to 2016 were included. RESULTS: 156 patients were included, to which 141 units of RBCs were transfused. All pre and intra-operative transfusions were according to guidelines; 110 transfusions were postoperative and 37 of these were not according to guidelines. Patients who were transfused had longer LOS in hospital (P = 0.002) and an odds ratio (OR) of 2.7 of contracting an infection (P = 0.04) in comparison with patients who were not transfused. No significant differences in mortality, LOS or morbidity were found between patients transfused according to guidelines and more liberal thresholds. Guidelines were more likely to be deviated from within the last 2 days before discharge than prior to this (58 vs. 24%, P = 0.03). Furthermore, 24 stable patients received two consecutive RBC units resulting in posttransfusion haemoglobin values of between 83 and 124 g l. CONCLUSION: Most RBC transfusions occur postoperatively, many still according to liberal transfusion thresholds, in particular shortly before discharge and as part of consecutive transfusions. Transfused patients had longer LOS and more infections than patients not transfused, but there was no difference in mortality, LOS or morbidity between patients transfused according to current guidelines and those where guidelines were deviated from.