Two years' experience with a hospital-based autologous blood donor pro
gramme is described. Of 511 patients enrolled, 86% underwent surgery w
ithout use of allogeneic blood, and 19% of collected units were not us
ed. For six common procedures, a 'target' autologous blood schedule of
2 or 3 units (depending on the procedure) was set. If an extra unit f
or each of 407 patients had been collected (i.e. 3 or 4), an additiona
l 20 patients would have been protected from exposure to allogeneic bl
ood at a cost of 387 unnecessary collections. The difference between t
he mean haemoglobin (Hb) concentration at entry to the programme and t
he mean pre-operative Hb concentration was not clinically important an
d only 25 of 511 patients were deferred on account of anaemia after on
e or more donations. The autologous blood used represents 3.8% of the
units of blood or packed red blood cells transfused during the 2 years
of study. Adverse (vasovagal) reactions were infrequent (2.2%) in spi
te of liberal acceptance criteria.