BACKGROUND In vitro fertilisation (IVF) and embryo transfer as treatment for male factor infertility is associated with lower fertilisation and pregnancy rates than for other indications. Since the late 1980s several assisted fertilisation techniques have emerged and have been rapidly developed to try to enhance results for couples with male factor infertility, or to help couples with severe male factor for whom conventional IVF was not possible. The technique of partial zona dissection (PZD) was developed to increase the probability that a sperm capable of fertilisation comes in contact with the oocyte. Although this method improved conventional IVF results, the improvement was only marginal and relatively large numbers of sperm are still required. This drawback applied less to the subsequent technique of subzonal microinjection of spermatozoa into the perivitelline space (SUZI). However, for all of these techniques fertilisation rates remained low, rates of polyspermic fertilisation were increased, and cases with a very limited number of spermatozoa in the ejaculate could still not be treated. The advent of intra-cytoplasmatic sperm injection (ICSI) of a single sperm (or sperm head or nucleus) into the oocyte appears to be an important breakthrough. OBJECTIVE To investigate whether ICSI improves fertilisation and/or pregnancy rates in comparison to other fertilisation techniques. METHODS The Menstrual Disorders and Subfertility Group search strategy (see Review Group details) was used to identify trials that had compared ICSI with other infertility techniques, such as PZD, SUZI, conventional IVF and additional IVF. METHODS Trials were included if they compared the effects of these techniques on fertilisation and pregnancy outcomes. Only randomised studies were included in this review. METHODS Ten studies met the inclusion criteria for this review. Eight studies compared ICSI with conventional IVF. One study compared ICSI with SUZI and one study compared ICSI with additional IVF. Data was extracted independently by two reviewers. Where relevant data was missing or unclear, the authors had been consulted. Male participants were classified according to their semen parameters, i.e. normal semen (concentration >20 million per ml, motility >50%, morphology >14%), borderline semen (concentration 10-20 million per ml, motility 30-50%, morphology 4-14% normal forms) and very poor semen (concentration <10 million per ml, motility <30%, morphology <4% normal forms). RESULTS For couples with normal semen there is no evidence of a difference in fertilisation rates per retrieved oocyte or pregnancy rates between ICSI and conventional IVF. On the other hand, for fertilisation rate per inseminated oocyte, ICSI appears to result in better outcomes than IVF for normal semen. For couples with borderline semen ICSI results in higher fertilisation rates (all) than IVF. Couples with very poor semen will have better fertilisation outcomes with ICSI than with SUZI or additional IVF. CONCLUSIONS There is evidence from this systematic review that fertilisation rates are significantly better with ICSI than IVF in couples with borderline semen. When the semen parameters are normal there is insufficient evidence of a difference in effectiveness between ICSI and IVF when retrieved oocytes were the unit of randomisation. However, there was a small but statistically significant increase in fertilisation rate when inseminated oocytes were the unit of randomisation. Total fertilisation rates were significantly reduced in ICSI cycles than IVF but there were no damaged oocytes in IVF cycles regardless of the semen parameters.