Over the past 16 years, 18 children under 2 years of age received chronic hemodialysis (HD) at our center. Five children were anuric at the start of HD and 6 had significant co-morbidity. The most common underlying diagnosis was posterior urethral valves. The median age at the start of HD was 12.2 months. A total of 39 episodes (defined as a discrete time period during which HD was the principle form of renal replacement therapy) of HD were performed, with a median duration of 7 months and 91.3 dialysis sessions per episode. Problems with vascular access were very common, with a revision ratio of 40%. Twenty-two line revisions were required for 36 episodes of line infection, with a median rate of line infection of 2.7 infections/patient years. The most commonly encountered organism was coagulase-negative Staphylococcus (69%). Twenty-three lines needed revision due to poor line function, despite the routine use of heparin. The effectiveness of HD was assessed in 11 patients who received HD for a continuous period of 3 or more months. The median urea reduction rate was 72%, while the parathyroid hormone levels improved to within twice the upper limit of the reference range in 69%. While there was no significant change in the median weight and height standard deviation score (SDS), the median SDS for head circumference showed significant improvement ( P=0.04). Both growth and developmental outcomes were strongly influenced by existing co-morbidity. Sixteen (89%) children were transplanted. Four (22%) children died, 3 after successful transplants. None of the deaths occurred on HD or resulted from its complications. In conclusion, HD in infants and small children is an effective and safe form of renal replacement therapy, but problems with vascular access limit its long-term use.