Blood transfusion practices in children at a regional referral hospital in Kenya

Aims To describe key operational and clinical aspects of the blood transfusion process in children admitted to a regional referral hospital in Kenya. Methods Demographic and laboratory data were collected prospectively on all children (0-14 years) for whom blood was requested over a 2-year period. Clinical data were obtained by retrospective case note review. Results From May 2005 to April 2007, there were 2789 requests for blood for children with a median age of 1.8 years (IQR, 0.6-6.6). The blood bank crossmatched blood for 1950 (70%) of the requests and 1663 (85%) crossmatched blood requests were collected. Of these, 1505 (90%) were not returned and assumed transfused (54% of requests). The median volume of blood prescribed was 200 ml (IQR, 150-336) and that of blood requested was 400 ml (200-450). A blood volume of greater than one unit was requested in 12% of cases and crossmatched in 0.6%. Crossmatching took place within 2 hours of a request in 48% of cases; in 30% of cases blood was collected more than 2 hours after crossmatch; and median time from request to collection was 3.58 hours (1.35-12.83). Of the 590 children whose case notes were reviewed, 506 were aged greater than 6 months, median pre-transfusion Hb was 6.0 g/dL, 393 (67%) received a transfusion and 132 (22%) died. For those children with a pre-transfusion Hb recorded in the notes and who received a blood transfusion, the time interval from request to collection was significantly shorter in those children who met laboratory criteria for severe anaemia than those who did not (5.8 hours, p=0.01). Concomitant administration of furosemide was associated with death (OR 0.44; 95% CI 0.23-0.84). Conclusion Our data suggest that in a setting with a high demand for paediatric transfusion but historically poor supply, clinicians tend to over-order blood in terms of the number of requests and the volume of blood requested. This has implications for laboratory workload and the blood supply itself. The association of furosemide with transfusion and death requires further investigation. This study provides evidence for measures to improve the efficiency and safety of blood transfusions for children in this setting.