OBJECTIVE: To conduct an audit of the frequency of red cell concentrate transfusions (RCCTs) in infants of different weight categories, the donor exposure rate (DER), in these transfused infants and the volume of blood wasted during each transfusion, and to identify from this baseline information specific categories of infants who would benefit from the introduction of a limited donor exposure programme (LDEP). STUDY SETTING: Neonatal wards and neonatal intensive care unit (NICU), Tygerberg Hospital, Western Cape. STUDY DESIGN: A prospective descriptive study and comparison with a historic control group. SUBJECTS: Information on the birth weight, age at the time of each RCCT and number of blood donors to whom an infant was exposed were collected post factum for all infants admitted to the neonatal wards and NICU between May 1993 and May 1994. During this time, the red blood cell concentrate was supplied as single paediatric bags (180 ml) transfused within 14 days of donation. An LDEP was introduced in February 1995. With this system, red blood cells were supplied as triple packs: a main unit of 250 ml with three empty satellite packs allowing up to three separate transfusions. These were assigned to a specific infant and were to be transfused within 21 days of donation. A second system where one adult blood bag was divided into two 180 ml bags and assigned to one infant to be transfused within 35 days of donation was also assessed. RESULTS: Of the 7854 infants admitted during the first 12-month audit period, 387 (4.9%) received 977 RCCTs. Of these, 183 (47.3%) received one transfusion, 72 (18.6%) two transfusions, 51 (13.2%) three transfusions, 27 (7.0%) four transfusions and 54 (13.9%) five or more transfusions. Infants (N = 188) with a birth weight below 1500 g admitted to the NICU were identified as the group with the highest prevalence of RCCTs (68.6%), and it was therefore decided that in the prospective study such infants would qualify for the LDEP. A total of 81 infants was transfused with either the double (N = 47) or the triple bags (N = 34) over a 5-month period. The decrease in the mean DER (+/-SD) was clinically significant when the triple (1.9 +/- 0.8) (P = 0.0001) and the double bags (1.6 +/- 0.8) (P = 0.0001) were compared with the previous single-bag system (4.4 +/- 3.5). Of concern was the large mean volume of concentrated red cells (118.5 +/- 12.5 ml) wasted per transfusion with the single-bag system. CONCLUSIONS: This survey confirmed a high RCCT rate as well as a very high DER in very-low-birth-weight (VLBW) infants treated at a tertiary centre. By assigning a triple or double bag of red cells from one blood donor and extending the storage of blood for small-volume RCCTs in infants from 14 days to 35 days, donor exposure was reduced significantly. We urge the introduction of the multibag blood transfusion system and extended storage period of blood for small-volume RCCT for VLBW infants in South Africa.