The extent to which the use of different fission-product cross sections (FPCS) affects the neutron multiplication factor keff in high burnup cores of fast reactors is evaluated. It is found that discrepancies of the order of 2.5% exist when different FPCS are used to calculate keff in the same core. These discrepancies are due to the absence of data on a number of fission-product isotopes present in some of the nuclear data libraries on the one hand, and large differences in the capture cross sections of some of the isotopes on the other. A list of fission-product isotopes is proposed that, when used, reduces discrepancies in keff to < 1%. The important isotopes for fast-reactor burnup and keff calculations in which large discrepancies exist are identified, and it is suggested that they be subjected to further evaluation to close the discrepancy gap.