Objective: To evaluate IVF outcome after epididymal and testicular sperm retrieval in patients with obstructive or nonobstructive azoospermia. Design: Retrospective clinical analysis. Setting: Public university-affiliated IVF unit. Patient(s): One hundred twenty-three azoospermic patients (178 cycles). Intervention(s): Sixty-three patients (103 cycles) with obstructive azoospermia (group 1) underwent either epididymal or testicular sperm retrieval, and 60 patients (75 cycles) with nonobstructive azoospermia (group 2) underwent testicular sperm retrieval combined with IVF treatment. Mature oocytes were fertilized using intracytoplasmic sperm injection. After sperm preparation, supernumerary spermatozoa were cryopreserved. Main Outcome Measure(s): Oocyte fertilization rate and clinical pregnancy rate (PR). Result(s): The oocyte fertilization rate was 48.4% (534/1,104) in group 1 and 41.5% (312/751) in group 2 (not significant [NS] difference). A total of 100 cycles (97.1%) and 62 cycles (82.7%) in the obstructive and nonobstructive groups, respectively, had embryos for replacement (NS difference). The clinical PRs per ET cycle were 24% (24/100) and 17.7% (11/62) in the two groups, respectively. Oocyte fertilization rates, when fresh (46.4%) or frozen-thawed (41.8%) spermatozoa were used, were not significantly different in the two groups. The PR when fresh sperm were used was 23.6% (30/127), versus 14.3% (5/35) when frozen sperm were used (NS difference). The PR for women aged ≤35 years was similar to that for women >35 years of age (20.7% or 29/140 and 18.2% or 4/25, respectively). Conclusion(s): Epididymal and testicular sperm obtained in azoospermic patients can fertilize oocytes successfully and may lead to high fertilization rates and PRs. Freezing of these spermatozoa does not reduce the outcome of treatment significantly.