Summary. In a retrospective review of transfusion errors in a large teaching hospital, we found the true incidence of errors to be at least four times the actual mistransfusion events detected. Seventy-®ve per cent of the errors were detected as near misses. The mistransfusions equated to 1/8610 compatibility procedures, and 1/27 007 units of blood issued, whereas the number of true transfusion errors equates to 1/2153 compatibility procedures and 1/6752 units of blood issued. The major error-prone activities included patient identi®cation at phlebotomy and the ®nal infusion of the blood product at the bedside. Of the cases, 95´2% were due to non-compliance with existing guidelines. Potential disasters were avoided only by the vigilance of the blood bank staff and the systems in place to detect errors.

This study identifies the benefit of using donor exposure rate (DER) to transfusion rate (TR) ratio as a discriminative index for assessing improvement in practice pattern in multiple-transfused neonates. It provides a methodology to demonstrate reduction in donor exposure that is not evident from the use of DER alone. Two time points, one 12-month period (1996-1997) before and one 12-month period (1999-2000) following the introduction of a paedipack system, were reviewed. Blood issued and wasted was quantified. The 1994 BSCH guidelines to define transfusion were used for both time periods, and recombinant erythropoietin (EPO) was not used. Following implementation of paedipack system, 186 donor units were made into satellite bags and kept for 35 days. A dramatic decrease in DER : TR ratio was noted for 79 transfused infants. The DER : TR ratio was 1 : 1 before and 1 : 3.2 after introduction of paedipacks, giving a 70.5% reduction in donor exposure risk. This was not evident from the use of DER alone, which remained the same (2.4) in the historical and study groups. High transfusions per donor unit (TPDU) correlated with the reduction in DER : TR ratio. Red cell wastage per transfusion was 190 +/- 30 mL before and 24.5 +/- 10 mL after intervention.

This article demonstrates a 62% reduction in the number of febrile nonhaemolytic transfusion reactions (FNHTRs) and 50% reduction in febrile reaction rate associated with red cell transfusions following graded introduction of universal leucodepletion. Though this is a statistically significant reduction (P = 0.009), it shows limited efficacy in abrogating this complication. We also found a reduction in the proportion of cases of FNHTRs with lymphocytotoxic antibodies over the period studied from 54% in 1998, 28% in 1999 to 23% in 2000. This corresponds to a relative increase in the number of febrile reactions without human leucocyte antigen (HLA) antibodies following full implementation of universal leucodepletion, as the total number of reported reactions actually fell considerably during the period. The increase in the number of cases without HLA antibodies was directly proportional to the increase in the number of leucodepleted units used.