Previous studies have suggested that plasmapheresis procedures using a separation membrane may activate the complement system and release anaphylatoxins. This study determines the content in C3a/C3a(des Arg) and C5a/C5a(des Arg) in plasma donations obtained by the new Haemonetics Filter Core (FC) procedure and compares it to Baxter Autopheresis C (Auto-C). FC performs sequential blood centrifugation and plasma filtration on a microporous polyethersulfone membrane, while Auto-C removes blood cells by simultaneous gravitation and filtration on a rotating nylon membrane. One group of 34 donors donated on FC and two groups of 30 and 10 donors on Auto-C. Plasma aliquots were taken from the plasma units within 30 min of the end of the collection procedures, frozen at < -30 degrees C and assessed for C3a and C5a at various time points of storage. Mean C3a/C3a(des Arg) in FC plasma (N = 34) was 1,151 (range: 526-2,991), 1,092 (range: 349-3498), and 507 (range: 307-815) ng/ml at time of collection and after 6 and 12 months of storage, respectively. Respective C5a/C5a(des Arg) was 26.6 (range 4.9-74), 18.9 (9.5-42.6), and 30.9 (range: 10.7-62.3) ng/ml. Mean C3a/C3a(des Arg) was higher in Auto-C (P < 0.001): 4,724 ng/ml (N = 10; range: 2,400-7 ,360) and > 4,149 ng/ml (N = 30; 2,408- > 6,430) after 3 and 18 months storage, respectively. Mean C5a/C5a(des Arg) was 32.1 ng/ml (N = 30; range: 10.6-57.2) after 18 months of storage. Complement activation in FC plasmas appears limited compared to Auto-C, suggesting better biocompatibility of this collection device and/or a favourable impact of the sequential cell centrifugation/filtration technology used. Further studies are needed to explain differences in complement activation between apheresis procedures and to assess clinical impacts, if any.