In a series of 296 patients treated in our centre by dialysis and/or renal transplantation over the last 15 years, the actuarial survival rates at 5, 10 and 15 years were 79, 68 and 57 percent respectively. Comparing the influence on survival of each of these two treatments separately should avoid two methodological biases represented by the sequential risk due to the succession of treatments and by different pretherapeutic situations. We therefore analysed these patients' survivals by the unbiased Mantel-Byar method, using a comparison of multiple survival factors (Cox's technique). We showed that treatment was an independent factor of survival and that transplantation had a more beneficial effect than dialysis. However, these replacement techniques seem to have less influence on survival than the pretherapeutic situations.