OBJECTIVE: To investigate the relationship between CD59 and traditional hemolysis tests for paroxysmal nocturnal hemoglobinuria. METHODS: To make different concentrations of CD59-blocked erythrocytes by mixing with normal erythrocytes, and investigate the relationship between quantitative abnormalities of CD59 and hemolysis tests. RESULTS: The hemolysis rate was positively related to the percentage of CD59-blocked erythrocytes. The greatest hemolysis of CD59-blocked erythrocytes was found in the CoF-initiated hemolysis test when the CD59-defects less than 16%, and the greater in Ham's test. When the CD59 defects were more than 16%, CD59-blocked erythrocytes hemolyzed to a greater extent in Ham's test than in the others. CONCLUSIONS: Quantitative differences of CD59 underlies the variability in hemolysis tests, and the CoF-initiated hemolysis test is the most sensitive in detection of small CD59-defects.