Background. Transplant candidates frequently ask whether they should, based on information available at the time, accept a cadaver kidney or wait for a potentially better one. Methods. We analyzed 937 first and second cadaver transplants done between January 1, 1984 and December 31, 1997 to determine if information available at the time an offer is made could be used to predict long-term graft survival. Results. By Cox regression, risk factors for worse long-term graft survival were older donor age, cardiovascular or cerebrovascular cause of donor death, and delayed graft function (DGF). HLA-ABDR mismatch was marginally significant. Whether DGF will occur is not known at the time of an offer, but risk factors can be determined; we found these to be older donor age and > 10 panel-reactive antibodies (PRA) at transplantation (by Cox regression). Using these variables (PRA, ABDR mismatch, donor age, and donor cause of death) known at the time of an offer, we calculated the relative risk of worse longterm graft survival for each subgroup (Table 3 in manuscript). In general, older age and donor death from cardiovascular or cerebrovascular disease were associated with worse outcome. Kidneys from donors of < 50 yr had the best outcome, irrespective of match. Conclusion. The data provided can be used to help guide patients as to whether they are better off accepting an offered kidney or waiting for a potentially better one. If an offer is declined, the next kidney may have a potentially worse outcome.