Abstract

Background: Published formulae, frequently used to predict the volume of transfused red cells required to achieve a desired rise in haemoglobin (Hb) or haematocrit (Hct), do not appear to have been validated in clinical practice.

Aims: To examine the relation between transfusion volume and the resulting rise in Hb and Hct in critically ill children.

Methods: Phase 1: Sample of 50% of children admitted during 1997; 237 of these 495 patients received at least one packed red cell transfusion; 82 children were transfused without confounding factors that could influence the Hb/Hct response to transfusion and were analysed further. Actual rise in Hb concentration or haematocrit was compared to that expected from use of existing formulae. A new formula was developed. Phase 2: In 50 children receiving a packed red cell transfusion during 2001, actual rise in Hb concentration was compared to expected rise in Hb with use of the new formula.

Results: Phase 1: Existing formulae performed poorly; median ratio of actual/predicted rise in Hb or Hct ranged from 0.61 to 0.85. Using the regression coefficients new formulae were developed for both Hb and Hct. These formulae were applicable across all age and diagnostic groups. Phase 2: Median ratio of actual/predicted rise in Hb improved to 0.95 with use of the new formula.

Conclusions: Existing formulae underestimate the volume of packed red cells required to achieve a target Hb or Hct. Adoption of the new formulae could reduce the number of transfusion episodes in PICU, cutting costs and reducing risk.