Background: Intermittent infusion hemodiafiltration (I-HDF) using repeated infusion of ultrapure dialysis fluid through a dialysis membrane or sterile nonpyrogenic substitution fluid was developed to prevent a rapid decrease in blood pressure by increasing the patient’s circulating blood volume, to enhance the plasma refilling rate by improving peripheral circulation, and to enhance solute transfer from the extravascular space to the intravascular space by enhancing the plasma refilling rate. Furthermore, the effect of fouling caused by attachment of proteins to the membrane as a result of ultrafiltration can be reduced by backflushing of the membrane with the purified dialysate in I-HDF. Although there have been several clinical trials of I-HDF, there have been no comparisons of the clinical significance of and indications for I-HDF with those of conventional hemodialysis (HD). Objective: The aim of this multicenter randomized controlled crossover trial was to compare the clinical significance of I-HDF with that of HD in Japan. Method: Patients were randomized to receive HD, I-HDF, and HD (group A) or I-HDF, HD, and I-HDF (group B) in that order for 14 weeks each. The sample size of 70 was determined based on the operability and patient availability. Treatment outcomes were evaluated 5 and 14 weeks after the start of each treatment period. The patients received 4-h treatment sessions with no changes in session duration or anticoagulant therapy during the study. I-HDF was performed using a GC-110N dialysis machine. Two hundred milliliters of ultrapure dialysis fluid were infused at a rate of 150 mL/min by backfiltration every 30 min during treatment. The first and last infusions were performed 30 min after the start and 30 min before the end of treatment, respectively. The total estimated infusion volume per session was 1.4 L (i.e., 200 mL × 7 infusions). I-HDF is a type of online HDF with a small fluid replacement volume. An ABH-P polysulfone membrane hemodiafilter was used for I-HDF and a class 1 or 2 hemodialyzer with a polysulfone membrane not coated with vitamin E and approved by the Japanese reimbursement system was used for HD. The primary outcomes were the Short Form-36 version 2 summary scores for quality of life and the visual analog scale scores for clinical symptoms. Secondary outcomes were vital signs, number of interventions, and pre-treatment blood test results. These variables were evaluated 1 week before at the start of the study, and at 5 and 14 weeks after the start of each treatment period. The removal characteristics of the various solutes were evaluated when possible on the first day of each treatment period. All patients provided written informed consent to participate. Results: Thirty-two patients in group A and 32 patients in group B completed the trial. There were no differences in the primary or secondary outcomes between I-HDF and HD. Serum α<sub>1</sub>-microglobulin (MG) levels at 14 weeks were significantly lower for I-HDF than for HD. During treatment, the removal rates for urea and creatinine, which are low molecular weight substances, were significantly lower during I-HDF than during HD. In contrast, the β<sub>2</sub>-MG and α<sub>1</sub>-MG removal rates were significantly higher during I-HDF than during HD. Furthermore, there was significantly less albumin leak during I-HDF than during HD. The solute removal results reflect the difference in pore size between the hemodiafilter used for I-HDF and the hemodialyzer used for HD and the difference in convective transport attributable to filtration between the 2 methods. Conclusions: These findings show that the removal rates of low molecular weight substances are significantly lower and those of medium to high molecular weight substances are significantly higher with I-HDF than with HD. They also indicate that there is significantly less albumin leak during I-HDF than during HD, meaning that I-HDF may be a particularly suitable dialysis modality for patients with malnutrition and the elderly in Japan.