About ten years ago it was discovered that changes in filter design which increase passive filtration improved dialysis efficiency. Later, these modified membranes showed similar intra-dialytic efficiency when used in on-line hemodiafiltration or in bicarbonate dialysis, called internal hemodiafiltration. On the basis of these previous results, we studied the long-term effects of internal hemodiafiltration, in comparison with low-flux bicarbonate dialysis. The pre-dialysis beta2-microglobulin level was chosen as the primary outcome variable. A prospective multicenter study with a cross-over scheme, 2 treatments and 3 periods, was designed. Twenty-four patients, followed in two dialysis centers, were enrolled. Many other parameters were measured every month at the first dialysis session of the week. The intra-dialytic removal of urea, beta2-microglobulin and homocysteine was also calculated. The removal of uremic toxins was significantly higher in internal hemodiafiltration than in low-flux bicarbonate dialysis. The pre-dialysis value of urea, phosphorus, beta2-microglobulin and homocysteine was lower during internal hemodiafiltration as compared with low-flux bicarbonate dialysis. The mean pre-dialysis value of hemoglobin was significantly higher during internal hemodiafiltration than low-flux bicarbonate dialysis, with a trend towards a significantly lower consumption of erythropoiesis stimulating agents during internal hemodiafiltration as compared with low-flux bicarbonate dialysis. Long-term treatment with internal hemodiafiltration improves the removal of small molecules and stops the continuous increase of middle molecules as seen in low-flux bicarbonate dialysis. Internal hemodiafiltration may substitute low-flux bicarbonate dialysis, but we need new prospective studies about long-term hard end-points.