BACKGROUND Artificial insemination with sperm is used to improve the chances of conception for various causes of infertility. Traditionally, sperm is deposited in or around the endocervical canal (cervical insemination - CI). Some studies reported higher pregnancy rates if sperm was deposited in the uterine cavity itself (intrauterine insemination - IUI), but most were uncontrolled. However the cost and the risks (infection and anaphylaxis) of IUI may also be higher. OBJECTIVE The objective of this review was to assess the effects of depositing donor sperm in the uterine cavity (intrauterine insemination) compared to cervical insemination. METHODS The Cochrane Subfertility Review Group specialised register of controlled trials was searched. METHODS Randomised trials comparing intrauterine insemination and cervical insemination, using fresh or cryopreserved semen, with or without ovarian hyperstimulation. METHODS Trial quality assessment and data extraction were done independently by two reviewers. RESULTS Twelve studies were included. They comprised 697 patients undergoing 2215 treatment cycles. Ten trials used frozen semen, with three using ovarian hyperstimulation. Overall the methodological quality of the trials was low. The overall pregnancy rate per cycle in the intrauterine insemination group was 18% compared to 5% for cervical insemination. When cryopreserved donor sperm was used, the overall chance of pregnancy in spontaneous or clomiphene-corrected cycles was significantly higher with intrauterine insemination. This was irrespective of whether pregnancy rates were calculated on a per cycle (odds ratio 2.63, 95% confidence interval 1.85 to 3.73) or per patient (odds ratio 3.86, 95% confidence interval 1.81 to 8.25) basis. The greatest benefit appeared in trials with poor pregnancy rates (less than 6%) for cervical insemination. There was no difference in pregnancy rate between intrauterine and cervical insemination when fresh donor sperm was used (odds ratio 0.90, 95% confidence interval 0.36 to 2.24). CONCLUSIONS Intrauterine insemination appears to be beneficial when cervical insemination using cryopreserved donor sperm has had low pregnancy rates. This applies to spontaneous, clomiphene corrected and gonadotrophin stimulated cycles. However it may offer little benefit where high pregnancy rates have been achieved with cervical insemination. There appears to be no additional benefit from intrauterine insemination when fresh sperm is used for donor insemination.