Introduction: Blood transfusion is not uncommon in patients following joint replacement surgery. However, allogeneic transfusion is associated with potential risks ranging from immune mediated allergic reactions to transmission of blood-borne organisms. Furthermore its use has a significant economical impact on healthcare services.

Methods: A retrospective audit of the case notes and haematological records of 196 patients who underwent primary total hip (THR), resurfacing hip (BHR) or total knee replacement (TKR) was carried out. Patients were routinely cross-matched for 2 units of blood if their pre-operative haemoglobin (Hb) levels were < 12g/dL in our cohort. Patient details including age, sex, BMI, comorbidities, and ASA grade were also collected. Surgical parameters such as operation time, tourniquet time (for TKR procedures) and trigger haemoglobin levels (Hb) with timing of transfusion were noted. Statistical analysis was undertaken to identify risk factors for transfusion.

Results: Forty-seven (24%) patients were transfused in our cohort. 78.7% of these patients were appropriately transfused with a trigger Hb of < 8g/dL. Patients transfused with a trigger Hb of > 8g/dL were no different with respect to demographics, procedure type and comorbidities. The average postoperative time interval to transfusion was 2.1 days. Five patients required an intra-operative transfusion. The mean pre-operative Hb levels in transfused and non transfused patients were 12.4 g/dL and 13.5 g/dL respectively. The mean number of units transfused was 2.31. A total of 109 units of blood were used: of these 70 units were cross-matched pre-operatively.

Univariate analysis established a significant relationship between the need for postoperative transfusion and preoperative Hb levels (p< 0.0001), length of surgery (p=0.01), age (p=0.03), history of respiratory disease (p=0.028) and hypertension (p=0.01). There was no significant relationship with respect to ASA grade and procedure type. Multivariate logistic regression analysis revealed pre-operative Hb (p< 0.0001) and age (p=0.015) as the strongest predictors of the need for post-operative transfusion. There is a strong correlation between length of surgery and time interval to transfusion (p=0.037).

Conclusion: Utilisation rates of blood products after primary hip and knee arthroplasty procedures exceed that of what is available from pre-operative cross-matching. In our institution more than 20% of patients may have been transfused inappropriately based upon local guidelines. The decision to cross-match blood pre-operatively for elective arthroplasty procedures should be re-evaluated in light of both patient age and pre-operative Hb levels. Postoperative monitoring of Hb levels should be undertaken early in patients with a prolonged duration of surgery.