Sperm DNA damage (fragmentation) is a recently discovered cause of male infertility for which no efficient treatment has yet been found. Previous findings have suggested that clinically relevant sperm DNA damage may occur at the post-testicular level. This study was undertaken to assess the clinical usefulness of ICSI with testicular spermatozoa in this indication. The percentage of spermatozoa with fragmented DNA, assessed by terminal deoxyribonucleotidyl transferase-mediated dUTP nick-end labelling assay, and ICSI outcomes were compared in two sequential attempts performed, respectively, with ejaculated and testicular spermatozoa in 18 men with increased sperm DNA fragmentation. The incidence of DNA fragmentation was markedly lower in testicular spermatozoa as compared with ejaculated spermatozoa. No differences in fertilization and cleavage rates and in embryo morphological grade were found between the ICSI attempts performed with ejaculated and with testicular spermatozoa. However, eight ongoing clinical pregnancies (four singleton and four twin) were achieved by ICSI with testicular spermatozoa (44.4% pregnancy rate; 20.7% implantation rate), whereas ICSI with ejaculated spermatozoa led to only one pregnancy which was spontaneously aborted. These data show that ICSI with testicular spermatozoa provides the first efficient assisted reproduction treatment option for men with high levels of sperm DNA damage.