Purpose: Despite positive responses in phase II clinical trials, the bioreductive prodrug apaziquone failed to achieve statistically significant activity in non-muscle invasive bladder cancer in phase III trials. Apaziquone was administered shortly after transurethral resection and here we test the hypothesis that haematuria inactivates apaziquone. Methods: HPLC analysis was used to determine the ability of human whole blood to metabolise apaziquone ex vivo. An in vitro model of haematuria was developed and the response of RT112 and EJ138 cells following a 1-h exposure to apaziquone was determined in the presence of urine plus or minus whole blood or lysed whole blood. Results: HPLC analysis demonstrated that apaziquone is metabolised by human whole blood with a half-life of 78.6 ± 23.0 min. As a model for haematuria, incubation of cells in media containing up to 75% buffered (pH 7.4) urine and 25% whole blood was not toxic to cells for a 1-h exposure period. Whole blood (5% v/v) significantly (p < 0.01) reduced the potency of apaziquone in this experimental model. Lysed whole blood also significantly (p < 0.05) reduced cell growth, although higher concentrations were required to achieve an effect (15% v/v). Conclusions: The results of this study demonstrate that haematuria can reduce the potency of apaziquone in this experimental model. These findings impact upon the design of further phase III clinical trials and strongly suggest that apaziquone should not be administered immediately after transurethral resection of non-muscle invasive bladder cancer when haematuria is common.