Mechanical trauma caused by centrifugal pumps is usually evaluated in terms of hemolysis. However, platelet damage caused by centrifugal pumps has not been studied well. We evaluated platelet damage in 2 different centrifugal pumps, the Medtronic BioMedicus BP-80 and the Terumo Capiox, in vitro and compared the results in terms of hemolysis. To evaluate platelet damage, the rate of increase (RI) for beta-thromboglobulin (beta-TG) and platelet factor-4 (PF-4) were measured by enzyme immunoassay. RI was defined as follows: RI for beta-TG is deltabeta-TG/deltaN and RI for PF-4 is deltaPF-4/deltaN where deltabeta-TG is the increase in beta-TG, deltaPF-4 is the increase in PF-4, and deltaN is the increase of the passing number, which is defined in the following equation: N = Qt/V (t, time; V, priming volume; Q, flow rate). Each pump was tested in a mock circuit for 3 h under a flow rate of 5 L/min and a pressure head of 100 mm Hg using fresh human heparinized blood (n = 5). For comparison, the normalized index of hemolysis (NIH) values were calculated for both pumps. The NIH values did not indicate a significant difference between the Capiox and the BP-80 pumps (Capiox vs. BP-80, 0.0021 +/- 0.0004 vs. 0.0034 +/- 0.0007, NS). However, the RI values for beta-TG and PF-4 in the Capiox were significantly lower than in the BP-80 (beta-TG, 0.198 +/- 0.047 vs. 0.376 +/- 0.049; PF-4, 0.080 +/- 0.014 vs. 0.268 +/- 0.043, p < 0.05). In conclusion, although there was no significant difference between the 2 pumps in terms of hemolysis, the Capiox centrifugal pump induced less platelet damage than the BP-80. The results suggest that measurements of RI for beta-TG and PF-4 are more sensitive parameters than NIH values for evaluating blood cell damage.