A modified hypo-osmotic solution was used to select viable ejaculated and testicular spermatozoa to perform intracytoplasmic sperm injection (ICSI) in 27 treatment cycles from patients with total absence of sperm motility. The treatment cycles consisted of 15 cycles in which ejaculated spermatozoa were used and 12 cycles in which testicular spermatozoa were used. The hypo-osmotic solution consisted of 50% culture medium and 50% deionized water and was shown in previous in-vitro studies to be superior to the original solution used in the classical hypo-osmotic swelling test. Fertilization was achieved in 37.3% of the oocytes injected. Embryos were replaced in 70.4% of the cycles with a mean of 2.0 embryos per cycle. There were no statistically significant differences between the ejaculated sperm group and the testicular sperm group in the fertilization rate (42.7 versus 30.1%) or in the cleavage rate (92.7 versus 77.3%). Four pregnancies resulted, two in the ejaculated sperm group and two in the testicular sperm group, a pregnancy rate of 14.8%. All pregnancies were singletons but one pregnancy in each group had an early miscarriage. There were no statistically significant differences between both groups in the pregnancy rates (13.3 versus 16.7%), in the implantation rates (5.3 versus 11.8%) or in the delivery/ongoing pregnancy rates (6.7 versus 8.3%). It is concluded that the use of this solution to select viable but immotile spermatozoa for ICSI is a simple and practical method and is associated with acceptable fertilization and pregnancy rates.