Sperm DNA fragmentation (SDF) has been linked with male infertility,and previous studies suggest that SDF can have negative influence on pregnancy outcomes with assisted reproduction.We performed a retrospective review of consecutive couples with a high SDF level that had intracytoplasmic sperm injection (ICSI) using testicular sperm (T-ICSI).We compared the T-ICSI outcomes to that of two control groups:87 couples with failed first ICSI cycle and who had a second ICSI cycle using ejaculated sperm (Ej-ICSI),and 48 consecutive couples with high sperm chromatin structure assay (SCSA)-defined SDF (＞15％) that underwent an ICSI cycle using ejaculated sperm after one or more failed ICSI cycles (Ej-ICSI-high SDF).The mean number of oocytes that were retrieved and the total number of embryos were not different among the three groups.The mean number of transferred embryos in the T-ICSI group was higher than the Ej-ICSI group but not significantly different than the Ej-ICSI-high SDF group (1.4,1.2,and 1.3,respectively,P ＜ 0.05).Clinical pregnancy rate in the T-ICSI group was not significantly different than the Ej-ICSI and Ej-ICSI-high SDF groups (48.6％,48.2％,and 38.7％,respectively,P ＞ 0.05).No significant difference was found in live birth rate when comparing T-ICSI to Ej-ICSI and Ej-ICSI-high SDF groups.The results suggest that pregnancy outcomes and live birth rates with T-ICSI are not significantly superior to Ej-ICSI in patients with an elevated SCSA-defined sperm DNA fragmentation and prior ICSI failure(s).